National digital taxes – Lessons from Europe

Abstract

Digitalisation has done more to shape the 21st century than virtually any other phenomenon. However, international tax law has seemingly failed to keep pace with rapid technological developments, which has likely led to inequalities between the tax burden of traditional and digital business models. Thus far, there has been no consensus regarding the issue of fair taxation of the digital economy at the international and EU level. As European policymakers have begun to experience noticeable amounts of pressure to act, several EU countries have pushed forward and introduced unilateral measures to ensure they receive a fair share of the tax revenues pie.

However, it is unclear whether national digital taxes can overcome the tax challenges stemming from the increasing digitalisation of the economy. Thus, newly proposed and implemented national digital taxes in Europe are thoroughly elaborated in the context of their relationship with double tax treaty law, the perils of double/multiple taxation, their coherence with European law, their global and regional impact on competition and competitiveness, their contribution to tax revenues and the establishment of fair taxation conditions. The analysis concludes with a presentation of pertinent suggestions regarding national and supranational tax policy.

Keywords:

• Digital Tax
• Digital Economy
• BEPS
• Unilateral Measures
• Global Competition
• Double Tax Treaties

1. Introduction

Unlike any other phenomenon, increasing digitalisation has contributed to shaping the 21st century, which has inevitably led to meaningful changes in our everyday lives. Technological progress has created opportunities to offer and provide internet-based services to customers around the world without having to establish any type of physical presence in other countries (see Schön, 2018; Olbert & Spengel, 2019; Hidien, 2019, p. 268).

It is evident that international income and corporate taxation regimes have not followed suit with these developments. Moreover, the increasing tax inequality between traditional and digital business models and the peril of diminishing tax revenues both loom with a greater sense of urgency. Current statistical data support these assertions: Within the territory of the European Union (EU), traditional international business models are burdened with an average effective tax rate of 23.2%, whereas international digital business models are taxed at an average rate of only 10.1% (B2C models) and 8.9% (B2B models) (see European Commission, 2017; PwC, 2017, p. 6 et seq.).

Over the past few years, the Organisation for Economic Cooperation and Development (OECD) and the EU have made several attempts to rebalance the relationship between digitalised and traditional businesses. However, neither has succeeded in carving out proposals that would be acknowledged by a broad consensus. The ineffectiveness of ongoing discussions led European policymakers to unilaterally introduce national measures primarily aimed at levelling the playing field and grabbing their fair share of the tax revenues pie (see Schön, 2018; European Commission, 2017; Scientific Advisory Board of the German Ministry of Finance, 2018). Consequently, these unilateral measures must be scrutinised and validated to ensure they achieve their intended purposes.

The aim of this paper is to shed light on the three types of national efforts undertaken by certain EU member states to address the issue of fair taxation of the digital economy. First, a comparative analysis of the three categories of national digital taxes is presented, which elucidates the similarities and differences related to the scope (single vs. multiple digital business models), the base (gross revenues vs. net profits) and qualification for double tax treaty law purposes (income tax vs. consumption/hybrid tax). Five national measures serve as examples, which are provided to sufficiently support the arguments presented in the subsequent sections. The chosen unilateral measures are then explained in the context of their relationship with double tax treaty law, the peril of double/multiple taxation, their coherence with European law, as well as their impact on global and regional competition and competitiveness, their contribution to tax revenues and the establishment of fair taxation conditions. Thereafter, suggestions regarding national and supranational tax policies are proposed to address effectively the issues surrounding the complexities of the digital economy.

2. Taxation trends in Europe’s digital economy

2.1 Background

The challenges associated with the fair taxation of the digital economy were identified as a primary focus of the OECD’s base erosion and profit shifting (BEPS) project. The steps taken by the Task Force for the Digital Economy (TFDE) culminated in the publication of the OECD BEPS Action 1 Report in 2015, which mapped the importance of digital technologies for businesses, the characteristics of digital businesses and the challenges associated with direct and indirect taxation (see OECD, 2015). The TFDE used the Action 1 Report to illustrate options specifically tailored for the digital economy, though they did not recommend their implementation as they expected a substantial impact from the implementation of general BEPS measures (see OECD, 2015, nr. 383; see also Wünnemann, 2019, p. 135). The countries involved in the OECD’s BEPS project agreed to further consultations and follow-up work that should result in a solution supported by a broad political consensus by 2020 (see OECD, 2015, nr. 385; see also Wünnemann, 2019, p. 135; Pinkernell, 2019, p. 339 et seq.).

In March 2018, the TFDE issued an interim report that primarily focused on the implications of digital transformation, such as the blurred boundaries between traditional and digital business models and the increasing significance of intangible assets, data and user-created value. The interim report also described the predominant digital business models and their fundamental characteristics (see OECD, 2018b, chapter 2).

In their January 2019 policy note, the TFDE scheduled further discussions among participating countries regarding their draft proposal, which was based on two pillars. More specifically, it addressed the modification of profit allocation and nexus (pillar I) and targeted the remaining BEPS issues associated with the digital economy (pillar II; see OECD, 2019a, p. 1). A detailed work programme was then issued in May 2019 for the implementation of three proposed approaches designed to tackle pillar I concerns: the “user participation” approach (preferred by the UK and France), the “market intangibles” approach (preferred by the United States) and the “significant economic presence” approach (preferred by emerging countries like India; see OECD, 2019b; see also Daurer, 2019, p. 555). Most recently, the TFDE published two public consultation documents containing the details of a proposed “unified approach” under pillar I (see OECD, 2019c; see also Mayr, 2019c) and the global anti-base erosion (GloBE) proposal under pillar II (see OECD, 2019d). As of the time of the publication of this paper, it is unclear whether the OECD will achieve its ambitious goal to provide the participating countries with a consensus-based solution by January 2020 (see Mayr, 2019c).

Rooted in the preparatory work at the OECD level, the European Commission (EC) issued two proposed directives in 2018. If they were to become law, they would be incorporated into the national tax laws of EU member states (see Kofler, 2018, p. 57).

The first directive centred around the concept of a “significant digital presence” designed to adjust permanent establishment (PE) criteria to make them fit, which could effectively address cross-border situations between EU member states in the age of digitalisation (see European Commission, 2018a). A company’s digital presence would be regarded as “significant” if one or more of the following are true:

• The company generates annual revenues of more than 7 million euros in a single EU member state;

• The company has more than 1,000 users in a single EU member state; or

• The company has more than 3,000 signed agreements for digital services per year in a single EU member state (see Article 4 of the draft, European Commission, 2018a, p. 17).

Apparently, the idea of a digital PE would collide with provisions in existing double tax treaties. Thus, the EC recommended the member states to modify the relevant double tax treaties accordingly (see European Commission, 2018a, p. 4; see also Kofler, 2018, p. 57).

The draft of the directive regarding the creation of a “significant digital presence” as an additional anchor for the allocation of revenues was accompanied by a proposed digital services tax on revenues generated from providing certain digital services (see European Commission, 2018b). Regarding its design, the EU digital services tax (EU DST) can be characterized as a “hybrid tax” that embodies elements of an income tax and a consumption tax. While the tax addresses revenues gained from value creation in a particular member state, it would be prone to being passed on to consumers and is triggered by providing services. Three types of digital services would fall under the scope of the EU DST:

• The placement of advertising on a digital interface that targets users of that interface;

• Making a multi-sided digital interface available to users that allows users to find other users and interact with them, which may also facilitate the provision of underlying supplies of goods or services directly between users;

• The transmission of data collected about users and generated from users’ activities on digital interfaces (see Article 3 of the draft, European Commission, 2018b, p. 28 et seq.).

Accordingly, the EU DST is clearly focused on large corporations, as only those EU companies with worldwide revenues exceeding 750 million euros and taxable revenues of more than 50 million euros would be subject to the EU DST (see Article 4 of the draft, European Commission, 2018b, p. 31).

The tax rate is set at 3% of the in-scope gross revenues, excluding the value-added tax (VAT) (see Articles 7 and 8 of the draft, European Commission, 2018b, p. 34).

The EU DST gathered considerable political momentum as a “quick fix”, whereas the integration of “significant digital presences” into EU member states’ double tax treaties was acknowledged as the preferable mechanism over the long term (see Kofler, Mayr, & Schlager, 2017; Bendlinger, 2018, p. 268 et seq.; Kofler, 2018, p. 57).

However, the implementation of either (or both) of these measures failed in the face of resistance from certain member states, most notably Ireland, which is where Google’s and Facebook’s EU headquarters are located (see Leigh, 2018). Even the proposed tax solely on the provision of targeted advertising services (i.e., the EU digital advertising tax [EU DAT]; see Council of the European Union, 2019a) could not secure approval from all member states (see Council of the European Union, 2019b, p. 8; Mayr, 2019a, p. I; Mayr, 2019b, p. 264). Against this background, several EU member states decided to introduce national solutions.

2.2 Grouping of national legislation by tax design

As of the time of the publication of this paper, several EU member states have introduced or proposed national measures regarding the taxation of the digital economy (for a digital tax map, see Tax Foundation, 2019b). These proposed and implemented unilateral tax measures can be categorised as one of the following three types:

• Digital services taxes;

• Digital advertising taxes;

• Unilateral adjustments to PE definitions in the national income tax code.

Digital services taxes (DSTs) address multiple digital business models. Generally, they are influenced by the draft of the EU DST and mostly mimic its primary characteristics (e.g., wording, tax scope, tax base, tax rate and revenues thresholds). Therefore, national DSTs usually refer to the same three kinds of digital business models defined in the EU DST, namely targeted advertising, the processing of user data and the provision of online marketplaces (see above, Subsection 2.1.). Policymakers and scholars mostly classify these DSTs as “hybrid taxes” because they combine elements of income and consumption taxes. DSTs generally help to level the playing field and function as a substitute for corporate taxation. By contrast, the national DSTs are linked to the provision of digital services, and since consumption is calculated using gross revenues excluding the VAT, they are prone to be passed on to customers (see Scientific Advisory Board of the German Ministry of Finance, 2018, p. 3; Pinkernell, 2019, p. 361). Accordingly, DSTs are not covered by the scope of existing double tax treaties, which only applies to taxes on income and capital (see Kofler et al., 2017; Schön, 2018).

Most unilateral measures taken by EU member states resemble DSTs. Given the broad media coverage of the French DST, it might be considered the most famous example. Apart from that, the UK (which is an EU member state for now), Italy, Spain and most recently (September 2019) the Czech Republic have drafted DST proposals (see HM Treasury, 2018b; Obuoforibo, 2019; Gazzetta Ufficiale, 2018; Nocete Correa, 2019; Kleinová, 2019). Except for the UK DST, which is covered in Subsection 2.3.1., all national DSTs have been designed in a manner similar to the EU DST.

In contrast to DSTs, digital advertising taxes (DATs) exclusively target online advertising services. Thus, the scope of DATs is significantly reduced to address only a single digital business model. Like national DSTs, the national DATs are based on gross revenues excluding the VAT. Notably, DATs do not fall under the scope of double tax treaties because they are regarded as consumption or transaction taxes (see Kofler et al., 2017).

As of the time of the publication of this paper, only Austria and Hungary have implemented national DATs (see also Tax Foundation, 2019b). The Austrian DAT mainly incorporates the principles of the proposed EU DAT (see Austrian Parliament, 2019b, p. 25). The Hungarian DAT has been in effect since 2014, making it one of the first digital economy tax measures instituted anywhere in the world. In contrast to the EU DAT, both the Austrian and the Hungarian DAT cover both the revenues from the provision of target advertising services (see Article 3 nr. 1 of the EU DAT draft, Council of the European Union, 2019a, p. 17) and all forms of digital advertising services (see Subsections 2.4.1. and 2.4.2.).

Adjustments to the PE definition may refer to a small or vast array of digital business models depending on the design of the legal amendment. If a digital business model meets the criteria of the digital PE, the earnings in a particular country are subject to the national income tax in the same manner as traditional businesses. In contrast to DSTs and DATs, income taxes are based on net profits. Amendments to the national PE definition indisputably underlie double tax treaty provisions (see Kokott, 2019, p. 129; Cibuľa & Kačaljak, 2018, p. 85).

Only a single European country, namely Slovakia, has unilaterally modified the PE definition in its national income tax code (see also KPMG, 2019).

For the purpose of this paper, five national measures were selected for closer examination, including the French and the UK DSTs, the Austrian and the Hungarian DATs and the amendments to the Slovakian income tax code, which are presented in the following sections. From the range of similar national DSTs, France functions as the representative model for EU DST-inspired national DSTs because it is the only one that has already been implemented. Moreover, from the range of EU DST-inspired national DSTs, the French DST is likely to affect digital businesses the most because France’s economy is the third largest in the EU (before Brexit; see Eurostat, 2018). The UK DST is also presented since this legal draft incorporates significant distinctions from the EU DST, which are important to conducting the analysis of national digital taxes presented in Section 3. Moreover, the UK represents the second largest economy in the EU (before Brexit; see Eurostat, 2018). Therefore, the implementation of the UK DST is likely to have a significant and probably even larger impact on the digital economy compared to the French DST. The Austrian and Hungarian DATs differ considerably in relation to certain aspects that are also important to the analysis presented in Section 3. Thus, both are further explained in Subsection 2.4. Lastly, the amendments in the Slovakian income tax code are presented to illustrate a unique and bold move in national tax policy. These five examples are only described to the extent necessary to support the aim of this paper.

2.3 First group: Digital services taxes

2.3.1 The UK digital services tax

In contrast to comparable unilateral measures within the EU territory, the UK DST supplements previous measures that have predominantly addressed undesirable arrangements of digital businesses. The UK DST’s predecessor, the UK diverted profits tax (“UK DPT”), was primarily enacted to re-allocate revenues from abusive tax structures to the UK affiliate, which was the case with Google’s tax scheme, the so-called “Double Irish Dutch Sandwich” (i.e., the “Google Tax”). However, the purpose of the UK DPT has not been narrowed just to rebalance the tax inequality between traditional and digital business models (see Avi-Yonah, 2019, p. 59; Sinnig, 2017, p. 410; Houlder, 2017). Indeed, the UK DPT tackles any kind of profit shifting that exploits “tax mismatches” (e.g. shifting profits to a low-tax jurisdiction) or avoids a PE in the UK. If such an abusive arrangement is discovered, the UK DPT of 25% is applied to the diverted UK-related profits (see HM Revenue & Customs, 2018, p. 3; Sinnig, 2017, p. 410 et seq.). Thus, the UK DPT can be classified as a “classic” anti-tax avoidance provision.

By contrast, the UK DST was proposed in November 2018, and it was designed to ensure that the local share of digital businesses’ created value is taxed in the UK (see Obuoforibo, 2019). The UK DST aims to address services, which are characterised by the value created from the participation of their users in accordance with the OECD “user participation” approach that the UK has promoted (see Wünnemann, 2019, p. 137; see also above, Subsection 2.1.). Hence, if enacted, the UK DST would be applied to revenues generated through the provision of social media platforms, search engines and online marketplaces (see HM Treasury, 2018b, subsections 1.16 to 1.18). These digital services must be understood in their broadest sense. For instance, a “provision of a social media platform” may resemble online advertising, subscription fees or sales of data, to name a few (see HM Treasury, 2018b, subsections 4.1 and 4.2). Consequently, the UK government chose not to take the EU DST as a pattern, but instead, defined the tax scope autonomously.

Similar to the EU DST, the draft of the UK DST implies two thresholds for global annual revenue and UK-related annual revenue, which must be exceeded to trigger the DST (see HM Treasury, 2018b, subsection 1.18). The revenue thresholds are set at 500 million pounds of global in-scope revenues and 25 million pounds of UK-sourced in-scope revenues, respectively (429.9 million euros and 21.5 million euros according to the exchange rate of the European Central Bank [ECB] of November 22, 2019; see HM Treasury, 2018b, subsection 6.4). Compared to the thresholds of the EU DST, which have been set at 750 million euros of overall global revenues and 50 million euros of in-scope revenues generated within the EU territory (see above, Subsection 2.1.), the UK DST is more likely to target fewer corporations. In contrast to the EU DST, which would have been 3%, the UK DST rate has been set at 2%. The UK government justifies the lower tax rate by arguing that the UK DST is calculated from gross revenues (instead of net profits) (see HM Treasury, 2018b, subsections 6.1. and 6.3; Obuoforibo, 2019).

2.3.2 The French digital services tax

The French DST (“Taxe sur les services numériques”, which is usually referred to as “Taxe GAFA”) could undoubtedly be attributed to the national digital tax that has received the broadest media coverage of all proposals by a European country. The controversial arguments surrounding its implementation by the French National Assembly in July 2019 culminated in Donald Trump threatening Emmanuel Macron with the imposition of tariffs on a variety of French goods, most notably French wine (see Tankersley, Nicas, & Swanson, 2019). Notwithstanding, Emmanuel Macron signed the bill on July 24, 2019, which put the French DST retroactively into effect as of January 1, 2019.

Concerning its main characteristics, the French DST virtually mirrors the draft of the EU DST. The French DST centres around two digital business models in accordance with the EU DST proposal. Hence, the French DST is applicable to revenues generated by targeted advertising and the provision of online marketplaces, but, unlike the EU DST proposal, not the processing of user data (see Article 299 II. of the French Tax Code, Legifrance, 2019; see also Hidien, 2019, p. 270 et seq.). The nexus to France is given as soon as a digital interface is accessible by users residing in France. In contrast to the EU DST that refers to the global and EU-wide revenues (see above, Subsection 2.1.), French lawmakers restricted the scope of the French DST by setting the second threshold at 25 million euros of annual revenues generation from the provision of digital services solely within French territory (see Article 299 III. of the French Tax Code, Legifrance, 2019). Although the French DST cannot be offset against the corporate tax, it can be deducted as a business expense (see Hidien, 2019, p. 270).

2.4 Second group: Digital advertising taxes

2.4.1 The Austrian digital advertising tax

Before implementing its DAT, a tax on traditional advertising services had been in force in Austria for over 18 years (“Werbeabgabe”). However, this advertising tax did not include online advertising services for historical reasons. Against this background, Austrian policymakers decided to extend the existing advertising tax to establish a fair taxation scheme for the digital economy (see Austrian Parliament, 2019b, p. 26 et seq.).

The Austrian DAT (“Digitalsteuer”) was broadly designed in accordance with the draft of the EU DAT. However, the Austrian DAT goes beyond the scope of the latter, as it does not solely refer to targeted advertising, but to any kind of online advertising services (see also Mayr, 2019b, p. 266). Moreover, Austrian lawmakers have narrowed the scope of the DAT, which is similar to French policymakers (see above, Subsection 2.3.2.). Hence, such companies are only considered taxable persons for the purposes of the Austrian DAT when

• The total amount of worldwide revenues for the relevant financial year exceeds 750 million euros; and

• The total amount of taxable revenues in Austria during the relevant financial year exceeds 25 million euros (see Article 2 of the Austrian DAT, Austrian Parliament, 2019b, p. 2).

By contrast, the existing Austrian advertising tax does not apply a similar exemption for small and medium-sized enterprises (SMEs) (only a de minimis rule for revenues which do not exceed 10,000 euros; see also Geringer, 2019, p. 395).

The Austrian DAT was set at 5% of the in-scope revenues to determine a uniform tax rate for the existing Austrian advertising tax and the Austrian DAT (see Mayr, 2019b, p. 266; Geringer, 2019, p. 399).

In Austria, tax revenues are usually not earmarked. Hence, it is remarkable that 15 million euros of the annual tax revenues from the Austrian DAT will be explicitly dedicated to financing of the digital transformation process of Austrian media companies (see Article 8 paragraph 4 of the Austrian DAT, Austrian Parliament, 2019, p. 3; see also Mayr, 2019b, p. 266; see on the issues arising from this earmarking below, Subsection 3.3.).

2.4.2 The Hungarian digital advertising tax

Hungary introduced an advertisement tax in 2014 to address revenues from traditional as well as online advertising services. The current tax rate is set at 7.5%, and smaller advertising companies that do not raise revenues of more than 100 million Hungarian forints (299,141 euros according to the ECB exchange rate of November 22, 2019) are exempt from the Hungarian DAT (see Sinnig, 2017, p. 412 et seq.).

In contrast to the Austrian DAT, the Hungarian tax is uniformly applicable to offline and online advertising services. Most notably, smaller advertising companies are indiscriminately exempt from the tax, even if they provide traditional or online advertising services.

2.5 Third group: Extensions of the PE definition in the national income tax code

2.5.1 Amendments to the Slovakian income tax code

Slovakia is the only European country that has broadened the scope of the PE definition in the national income tax act (“Zákon o dani z príjmov”; see also KPMG, 2019). In 2018, the requisitions for limited tax liability in Article 16 of the Slovakian income tax code were amended so that a PE was established in Slovakia by providing recurring mediation services related to transportation and accommodations, even if they are solely performed through digital platforms (see Zákony pre ľudí, 2019). The term “digital platforms” is defined in Article 2 ag) of the Slovakian income tax code as hardware or software platforms, which are necessary to create and manage applications (see Zákony pre ľudí, 2019; see also Cibuľa & Kačaljak, 2018, p. 80 et seq.; European Parliament, 2019a, p. 40).

3. Critical analysis of national digital taxes in Europe

3.1 Introduction

At first glance, the arguments justifying the introduction of national measures to tax the digital economy (see above, Subsection 2.1.) appear convincing. As of the time of the publication of this paper, both the OECD and the EU have not yet delivered recommendations to tackle the peril of increasing tax inequality between traditional and digital business models and diminishing tax revenues due to ongoing digitalisation. On the other hand, politicians are exposed to a noticeable amount of pressure to counteract these developments. Notwithstanding, national policymakers should be careful not to rush into implementing short-sighted and ill-thought-out measures which can predominantly entail disadvantageous effects.

Therefore, the three groups of national efforts designed to get a fair share of the tax revenues pie and rebalance existing tax inequality are thoroughly examined in the following sections. Each section is dedicated to one of the following five aspects:

• Relationship with double tax treaty law/peril of double and multiple taxation;

• Coherence with European law;

• Impact on (global and regional) competition and competitiveness;

• Contribution to tax revenues;

• Establishment of fair taxation statuses.

3.2 Relationship with double tax treaty law/peril of double and multiple taxation

Both DSTs and DATs do not fall under the scope of current double tax treaties because double tax treaties currently only cover taxes on income and capital (see also above, Subsection 2.2.). Consequently, residence states are not obliged to grant tax credits for DSTs and DATs paid in EU countries. In case residence states decide not to accept the deduction of DSTs and DATs, revenues are prone to underlie double or multiple taxation (depending on the company group’s structure) (see Hidien, 2019, p. 270; Kofler et al., 2017). Since the largest digital businesses are US-based, the 80% foreign tax credit of the GILTI scheme is notably only applicable to income and corporate taxes (thus not the DSTs and DATs; see Schildgen, 2019, p. 370 et seq.). The deductibility of the French DST as business expenses (see above, Subsection 2.3.2.) may not serve as a meaningful device to avoid double taxation because the targeted businesses mostly do not pay any corporate tax in the source states, which is why the discussion regarding fair taxation of the digital economy emerged in the first place (see also Pinkernell, 2019, p. 363).

On the other hand, amendments to national PE definitions are undoubtedly covered by existing double tax treaties (see also above, Subsection 2.2.). Apparently, Slovakia’s unilateral extension of the PE definition is not consistent with existing provisions of the Slovakian double tax treaties (see Kokott, 2018, p. 129; Cibuľa & Kačaljak, 2018, p. 85). Generally, double tax treaties must be primarily applied in case of incompatibility, as long as they are not superseded by national law. Currently, the Slovakian legal principles and local laws do not allow treaty overrides (see Cibuľa & Kačaljak, 2018, p. 85 et seq.). Therefore, the broader PE definition in the Slovakian income tax code cannot come into effect in cross-border situations. Accordingly, the new digital PE may not have any practical significance. Thus, Slovakia’s unilateral amendments must be considered as a toothless measure as long as no corresponding amendments are made to the referring double tax treaties (see also Cibuľa & Kačaljak, 2018, p. 86 et seq.; Schön, 2018).

3.3 Coherence with European law

All national DSTs and DATs in EU countries have implemented revenues thresholds that must be exceeded for a company to be subject to the national DSTs and DATs. By law, these thresholds do not differentiate between domestic and foreign businesses. However, the DSTs and DATs will mostly burden foreign digital companies (see Schön, 2018; Becker & Englisch, 2018). Consequently, they are prone to indirectly discriminate against foreign companies, which is a violation of the fundamental freedoms of the EU, particularly the freedom of establishment and the free movement of persons (see also Ehrke-Rabel, 2019, nr. 1146f; Kokott, 2018, nrs. 102 et seq.; Schaumburg, 2019, nrs. 4.64 et seq.).

National tax laws of EU member states must also comply with EU state aid law. Accordingly, EU member states are generally prohibited from granting positive state aid (e.g., subsidies) or negative state aid (e.g., favourable tax treatments) to a selected group of companies (see Micheau, 2011, p. 196; Kokott, 2019, nr. 141; Englisch, 2019, nrs. 9.8 and 9.10.). However, state aid can be justified if they differentiate between undertakings in a similar legal and factual situation, and this differentiation can be explained by the nature and logic of the reference system (i.e., the reference system that usually includes all regulations related to a particular tax such as the corporate tax). Thus, unequal treatment must be rooted in the particular tax regime (see also Micheau, 2011, p. 203 et seq.; Kokott, 2019, nr. 215; Geringer, 2019, p. 396; Englisch, 2019, nr. 9.23). Enactment of the Hungarian DAT was previously tried before the General Court of the European Union, with its de minimis rule (see Subsection 2.4.2.) being considered unlawful state aid for smaller advertising companies. However, the General Court did not share the EC’s concerns: Since the Hungarian DAT was enacted for redistribution purposes, the differentiation between advertising companies with revenues below and above the set threshold was justified (see General Court of the European Union, 2019). On the other hand, the Austrian DAT was primarily introduced to establish a fair taxation of the digital economy. Yet, only large (foreign) companies exceeding the revenues thresholds (see Subsection 2.4.1.) are subject to the Austrian DAT, which leads to de facto “fair taxation” of just parts of the digital economy. Thus, the exemption cannot be explained by the nature and logic of the system, and the ECJ will likely characterise the use of revenues thresholds as unlawful negative state aid (see Geringer, 2019, p. 395 et seq.). Scrutinising national DSTs in other EU member states leads to the same result, when the comparable regulations on revenues thresholds and the purposes of the DSTs are taken into account. Apart from that, the Austrian DAT also entails direct state aid, with its revenues being primarily used to fund the digital transformation of domestic media companies, which cannot be explained by the nature and logic of the reference system (see Geringer, 2019, p. 401 et seq.; Ehrke-Rabel, 2019, nr. 1146j). Another state aid issue concerning the Austrian DAT can be found in the unequal treatment of SMEs that provide traditional advertising services and those that provide digital advertising services because digital advertising companies are widely exempt from the DAT (see above, Subsection 2.4.1.; see also Geringer, 2019, p. 399 et seq.).

In contrast to the national DSTs and DATs, national amendments to the PE definition, as in the case of Slovakia, do rather not raise any issues regarding their compatibility with European law, most notably because it is not necessary to apply the modified PE definition in cross-border situations (see above, Subsection 3.2.).

3.4 Impact on (global and regional) competition and competitiveness

The estimated tax burden resembles one of the key factors in business decision-making concerning the choice of location and cross-border activities. Among other things, (digital) taxes should be designed such that companies can easily determine which tax burden will apply if they provide services in a particular country (see Zöchling, Plott, Rosar, & Dziurdź, 2018, p. 8 et seq.). DATs benefit from their well-defined scope, as companies would rather not face difficulties in the evaluation of whether their services constitute digital advertising. The Slovakian definition of a digital PE is also prone to putting companies in a position where the assertion of the overall tax burden linked to business activities in Slovakia should not cause any obstacles. On the other hand, the proposed and implemented national DSTs seemingly create a certain degree of legal uncertainty. Scholars and practitioners agree that the definitions of the taxable digital services in the existing DST models entail grey areas, at national as well as at EU level, because the boundaries between analogue and digital business activities have become increasingly blurred. This circumstance might hamper the economic prosperity of countries implementing DSTs (see Pinkernell, 2019, p. 363; Becker & Englisch, 2018; Wünnemann, 2019, p. 139; EY, 2018, p. 3; Scientific Advisory Board of the German Ministry of Finance, 2018, p. 5). The intensity of their capability to hamper economic prosperity, which is generally associated with DSTs, is possibly amplified in the case of the UK DST. It includes its own catalogue of definitions that differ significantly from EU DST wording (see above, Subsection 2.3.1.). Hence, the travaux preparatoires of the EU DST may not be utilised to better understand the UK DST regulations (see also Obuoforibo, 2019).

As stated above, both national DSTs and DATs trigger double (and multiple) taxation (see Subsection 3.2.). Therefore, these taxes add up to the existing (corporate) tax burden in the residence states. Hence, it seems appropriate to assume that this circumstance hampers the competitiveness of digital companies located in countries with national DSTs and DATs. Most notably, this can stifle emerging digital businesses and start-ups in the marketplace (see also Scientific Advisory Board of the German Ministry of Finance, 2018, p. 3 et seq.).

Accordingly, the introduction of additional national taxes on the revenues generated from the use of digital business models could discourage traditional businesses from launching digital transformation processes in their businesses (see Schön, 2018). Because the grade of digitalisation is likely to determine the future existence of businesses in the globalised economy (see Zöchling et al., p. 3 et seq.), European companies may forfeit competitiveness by losing ground to competitors from other countries and continents (see Becker & Englisch, 2018; Schön, 2018; European Parliament, 2019a, p. 49; IFO, 2018, p. 16). Moreover, a race concerning the attractiveness as a location for digitalised businesses might be spurred by the introduction of unilateral measures by only a number of EU member states, thereby toxifying the EU’s political and economic climates.

Another important point that warrants discussion is the likelihood of unilateral measures triggering retaliation by residence states. The headquarters of the largest IT companies are located predominantly within the territory of the United States (see Pinkernell, 2019, p. 364; Scientific Advisory Board of the German Ministry of Finance, 2018, p. 2). Hence, the introduction of national digital taxes by other countries heavily affects the US sphere of interest. The United States has clearly indicated that they will not accept any “unilateral and unfair […] tax that targets our technology and internet companies” (US Department of the Treasury, 2018; see also EY, 2018, p. 4). In reality, Donald Trump threatened Emmanuel Macron with the placement of tariffs on French goods, and referring investigations were initiated (see above, Subsection 2.3.2.). If the results were positive, the United States may decide to introduce penalties against France and other European countries that implement national digital taxes (see also European Parliament, 2019a, p. 49). If so, they would be equal to the penalties introduced in October 2019 as a result of European countries granting subsidies to Airbus (see POLITICO, 2019). Alternatively (or additionally), the United States could introduce national taxes that would primarily address foreign business branches (e.g., the German automobile industry; see Scientific Advisory Board of the German Ministry of Finance, 2018, p. 5; Pinkernell, 2019, p. 363 et seq.). Therefore, any unilateral measures that tax the digital economy risk triggering retaliation by digital businesses’ residence states. Such a legal arm’s race re-establishes borders and introduces penalties at the expense of the targeted businesses, which could potentially discourage business owners from providing cross-border services (see Becker & Englisch, 2018; Kokott, 2019, p. 125).

3.5 Contribution to overall tax revenues

Sustainable tax design considers the contribution of a measure to the tax revenues (see Hey, 2019, p. 13 et seq.). Accordingly, it needs to be elaborated whether the unilateral measures taken by certain EU member states are likely to meet this criterion. To provide a uniform comparative method, the total tax revenues in 2016, as stated in the OECD Revenue Statistics 2018 (see OECD, 2018b), underlaid this comparison.

According to the UK government, the UK DST will raise between 275 million pounds (2020/2021) and 440 million pounds (2023/2024) (see HM Treasury, 2018a, p. 2). In 2016, the UK reported total tax revenues of 644,700 million pounds (see OECD, 2018b, p. 151). Putting these numbers into perspective, the UK DST would represent between 0.04% (2020/2021) and 0.07% (2023/2024) of the 2016 overall tax revenues.

The French DST would add an estimated 400 to 650 million euros to the 2019 overall tax revenues (see Hidien, 2019, p. 272). By comparison, the 2016 total tax revenues in France amounted to 1,013,100 million euros (see OECD, 2018b, p. 102). The French DST thereby reflects 0.04% to 0.06% of France’s 2016 overall tax revenues.

The Austrian Ministry of Finance assumed the annual tax revenues derived from the DAT to increase annually from 25 million euros in 2020 to 34 million euros in 2023 (see Austrian Parliament, 2019a, p. 1). Austria’s 2016 total tax revenues amounted to 149,200 million euros (see OECD, 2018b, p. 86). Accordingly, the Austrian DAT represents 0.02% of Austria’s 2016 tax revenues.

The actual tax revenues from the Hungarian DAT were apparently not available in English. The German Institute for Economic Research (IFO) estimated Hungary’s share of revenues from the EU DST to be approximately 31.6 million euros (see IFO, 2018, p. 27). In 2016, Hungary reported total tax revenues of 13,887,800 million Hungarian forints (approximately 41,544 million euros according to the ECB exchange rate of November 22, 2019; see OECD, 2018b, p. 108). Therefore, the implementation of the EU DST would have reflected 0.07% of Hungary’s 2016 total tax revenues. Considering that the scope of the EU DST would have been significantly broader than the scope of the existing Hungarian DAT, it may be assumed through an argumentum a maiore ad minus that the effective revenues from the Hungarian DAT are estimated to be significantly less the 0.07%.

The Slovakian measure is not prone to contribute in any way to the country’s total tax revenues because the modified PE definition lacks an effective scope due to the existing double tax treaty provisions (see above, Subsection 3.2.).

In conclusion, the contribution from implemented and proposed national digital taxes to the overall tax revenues appears to be comparatively miniscule. Hence, it is unlikely that these measures will translate into substantial fiscal effects (see also Pinkernell, 2019, p. 363; Scientific Advisory Board of the German Ministry of Finance, 2018, p. 4).

3.6 Establishment of fair taxation statuses

The unilateral measures in EU member states have been introduced with the aim to contribute to “levelling the playing field” by ensuring the fair taxation of revenues/profits realised from the use of digital business models (see Obuoforibo, 2019; Hidien, 2019, p. 268; Austrian Parliament, 2019b, p. 25; Sinnig, ISR 2017, p. 412; Cibuľa & Kačaljak, 2018, p. 80 et seq.; see also European Commission, 2017, p. 6; European Commission, 2018b, p. 3).

However, neither of the national DSTs or DATs includes a provision that would only trigger the tax if the corresponding revenues underlay insufficient (corporate) taxation. Thus, DSTs or DATs are levied regardless of whether these revenues are “undertaxed” (see also German Ministry of Finance, 2018, p. 5, on the EU DST). The significance of this circumstance has been amplified since the 2017 Tax Cuts and Jobs Act was signed into law, while the US-based companies that have dominated the digital economy ever since have been burdened with a 12.3% combined tax rate in Ireland and the United States since 2018 (see Pinkernell, 2019, p. 338). Compared to the standard corporate tax rate for (domestic) businesses in Hungary (9%) and Ireland (12.5%) (see Tax Foundation, 2019a), making a case for the “undertaxation” of US-based companies appears difficult as of 2018. Yet, US-based companies are predominantly burdened with national DSTs and DATs (see Kokott, 2019, p. 125; Becker & Englisch, 2018).

On the other hand, the indifferent tax design of the national DSTs and DATs may backfire on businesses that were previously not considered to be potentially subject to these taxes (particularly traditional business models with additional digital content). For example, the “Axel Springer” publishing house resides in Germany, where the company is subject to one of the highest corporate tax rates among European countries (see Tax Foundation, 2019a). This notwithstanding, it is subject to the French DST (see Hidien, 2019, p. 272). Thus, traditional businesses could be subject to a higher tax rate as after they undergo digital transformation, and the implementation of DSTs and DATs could discourage traditional businesses from digitalising their work processes and services, which could have tremendous effects on their competitiveness in the future (see Becker & Englisch, 2018; Schön, 2018; see also above, Subsection 3.3.).

Both the national DSTs and DATs discriminate against certain digital business models because they do not cover all of them, with DATs discriminating to a greater extent than DSTs due to the limitation of the scope to digital advertising services. For instance, national DSTs and DATs do not cover streaming services (e.g., Netflix, Hulu, Spotify) or electronic payment services (e.g., Paypal) (see Brameshuber & Franke, 2018, p. 287). Moreover, national DSTs and DATs solely address large companies that exceed the established revenues thresholds (see above, Subsections 2.3. to 2.5.). Therefore, DSTs and DATs will likely add another layer of unfair taxation to the existing inequalities between traditional and digital business models by introducing differing tax treatment of particular digital business models and companies of different sizes.

Against this background, national DSTs and DATs are apparently not capable of fulfilling their purpose of levelling the playing field. The EU member states’ rationale for not linking the application of DSTs and DATs to the premise of an insufficient tax burden in the residence state (similar to anti-avoidance rules; see also below, Section 4.) and not introducing the digital taxes for all digital businesses regardless of the actual sum of their annual revenues remains unclear.

In contrast to DSTs and DATs, modifications of the PE definition can contribute to a fair and equal taxation of both traditional and digital business models, as companies would underlie the national income tax schemes under the same conditions. Moreover, the adapted PE definition is applicable to all digital businesses underlying its scope, regardless of the sum of their annual revenues. However, if the broadening of the PE definition is narrowed down to two digital business models (online platforms for the mediation of transport and accommodation; see above, Subsection 2.5.1.), as in the case of Slovakia, all the other digital business models remain “unfairly taxed”. Indeed, the Slovakian model also adds another layer of unequal tax treatment to the existing inequalities. Therefore, the explicit national measure implemented by Slovakian lawmakers insufficiently contributes to the establishment of the fair taxation of the digital economy.

4. Suggestions for national and supranational tax policy

The implementation of national digital taxes is likely to be accompanied by numerous undesirable side-effects (see also Kofler, Mayr, & Schlager, 2018). Therefore, it is important to determine which policy tools might better level the playing field and ensure a fair share of the tax revenues from digital businesses.

Apparently, some digital businesses are structured so that company groups can illegitimately exploit favourable tax regulations of the destination countries, which significantly reduces their tax burden. The UK DPT was designed to target such legal, yet undesirable arrangements (see above, Subsection 2.3.1.). However, such arrangements cannot be considered a phenomenon that exclusively appears among businesses engaged in the digital economy. Indeed, BEPS has been practiced by companies regardless of their business models. Thus, it is important to determine whether the tax issues related to the digital economy could be sufficiently tackled by the implementation and rigid execution of anti-tax avoidance regulations, inter alia transfer pricing rules (see Schön, 2018). Such measures would entail the upside of being indiscriminately applicable to all sizes of traditional and digital foreign and domestic businesses. Hence, they would rather not violate European law or serve to promote indiscriminate tax treatment (see also above, Subsections 3.3. and 3.6.). As anti-tax avoidance provisions have been globally recognised as acceptable national measures in the age of BEPS, the likelihood that they would trigger retaliation by the residence states appears rather limited (see also above, Subsection 3.4.). Lastly, the TFDE backed its refusal to recommend unilateral measures in the OECD BEPS Action 1 report by arguing that the implementation of BEPS measures might already lead to an adequate mitigation of tax challenges associated with the digital economy (see OECD, 2015, nr. 383; see also above, Subsection 2.1.).

Undoubtedly, countries may also consider re-negotiating double tax treaties with residence states engaged in digital businesses to ensure the introduction of digital PEs. However, it seems highly doubtful that source states will reach an agreement on such amendments in bilateral situations.

At the EU level, the EC should more extensively use its powers associated with the “Guardian of the Tax Treaties” (see also European Parliament, 2019b), and combat more determinately arrangements between EU member states and (digital) businesses which undermine the purposes of the single market. As a famous example, the EC urged the Republic of Ireland to collect 13 billion euros in taxes from Apple in 2016 after the EC had condemned this sweetheart tax deal as an unlawful state aid (see European Commission, 2016; Farrell & McDonald, 2016; Kanter & Scott, 2016). Hence, regular and strict application of state aid laws may serve as an effective deterrent against unfair favourable treatment and tax incentives, which can lead to a distortion of a single market.

Apart from national, bilateral and EU-wide solutions presented herein, policymakers could take steps to finally reach a broad consensus for income tax purposes at the international level, particularly by employing the ongoing process at the OECD (see above, Subsection 2.1.). Settling on a compromise among a large number of countries would ensure wide-spread acceptance and application of nexus approaches for digital businesses. Hence, traditional and digital businesses would underlie tax regimes under comparable conditions (see also above, Subsection 3.6.). Moreover, broadly-accepted measures would minimise the risk of creating disadvantages in regional and global competition and provoking retaliatory measures (see above, Subsection 3.4.; see also Becker & Englisch, 2018).

However, the existing measures to curb BEPS within the EU (anti-tax provisions and state aid law, see above) may already provide a sufficient arsenal of meaningful devices to tackle the most serious forms of tax inequality (see also Staringer, 2017, p. 348; Schön, 2018). After all, the central issues concerning a fair taxation of the digital economy, most notably the peril of double and multiple taxation (see above, Subsection 3.2.) and the probability of distortions of the regional and global market (see above, Subsection 3.4.), have long been well-known, and, thus have accompanied discussions among scholars and policymakers long before the rise of the digital economy.

5. Conclusion

In response to the deadlock at the OECD and EU levels regarding the issue of a fair taxation of the digital economy, several EU countries have moved forward with the introduction or proposal of unilateral measures. This legislation has primarily served two aims, namely allocating an appropriate share of the tax revenues from digital services to the particular country and reducing tax inequality between domestic and digital business models (Subsection 2.1.).

The national digital taxes implemented and proposed in the EU territory can be allocated to three groups, including the tax type (income vs. consumption/hybrid taxes), the tax scope (single vs. multiple digital business models) and the tax base (gross revenues vs. net profits) (Subsection 2.2.). Five of these national digital taxes have been presented in greater detail (Subsections 2.3. to 2.5.) for the critical analysis described in subsequent section (Section 3.). Accordingly, the following conclusions were drawn:

• The three types of unilateral approaches to the fair taxation of the digital economy (i.e., DSTs, DATs and unilateral amendments to PE definitions) risk triggering double and multiple taxation (DSTs and DATs) or violating existing double tax treaty provisions (unilateral amendments to PE definitions; Subsection 3.2.);

• In addition to the issues concerning international tax law, DSTs and DATs may also violate the EU treaty provisions, particularly the fundamental freedoms and state aid law. Due to the ineffectiveness of unilateral amendments to PE definitions, these measures do rather not raise any concerns at the EU level (Subsection 3.3.);

• DSTs and DATs may serve as a deterrent against the digital transformation of traditional businesses and weaken a country’s attractiveness as a business place. Moreover, DSTs and DATs are also likely to provoke retaliatory measures, such as the introduction of tariffs and taxes tailored to harm other countries’ predominant business branches. National amendments to national PEs will not affect competitiveness and international trade, as long as they are not capable of overriding tax treaties, as in the case of Slovakia (Subsection 3.4.);

• Neither of these measures is capable of meaningfully contributing to the countries’ tax revenues (Subsection 3.5.);

• All the national digital taxes in the EU fail to deliver on their promise of establishing tax equality between traditional and digital businesses. In fact, they add another layer of unequal treatment to the existing inconsistences (Subsection 3.6.).

Given these issues, the national DATs and DSTs and the extension of the PE definition in Slovakian income tax law entail significant negative side-effects. Instead of implementing similar measures, source states may attempt to re-negotiate double tax treaties with residence states. However, the likelihood of residence states, particularly the United States, backing down in bilateral situations appears to be infinitesimally small, and reaching a compromise in the recent OECD negotiations may be more likely. Notwithstanding, countries should not overlook and/or underestimate the existing toolkit of anti-tax avoidance regulations that can be used to tackle the most distorting tax schemes of digital businesses. Additionally, the EC could effectively secure the functioning of the single market among EU member states by rigidly proceeding against harmful tax arrangements between businesses and particular countries. Lastly, the main issues associated with the fair taxation of the digital economy (e.g., tax-avoidance schemes, double taxation and trade wars) have yet to manifest from the emergence of digitalisation, though such issues continue to be contemplated in discussions on cross-border economic activities.