2
ACKNOWLEDGEMENTS
I would like to thank my classmates who took part to the Research Project ‘New Forms of Governance: From the EU to the World?’, especially for contributing to a great ‘preparation stage’ before starting my research. We had interesting discussions during our meetings which significantly contributed to my understanding of the literature. It was also very nice to keep in touch after the literature meetings and to talk about our experiences during the research process.
In order to collect all the necessary data for answering my research question as good as possible, I thank the interviewees, who despite their busy schedules made time for me. The interviewees provided useful in-depth insights regarding EU initiatives that aim to combat tax avoidance.
Furthermore, I would like to thank my housemates and foremost my sister for reviewing my thesis, thereby sharing their critiques and asking the right questions, which contributed to the quality of my analysis.
Last but not least, I thank my supervisor Jonathan Zeitlin, for his support and for providing me with the right directions as I needed them.
Iris Visser
3
ABBREVIATIONS
APAs Advanced Price Agreements ATAP Anti-Tax Avoidance Package ATAD Anti-Tax Avoidance Directive BEPS Base-Erosion and Profit-Shifting CBCR Country-by-Country reporting
CCCTB Common Consolidated Corporate Tax Base CFC Controlled Foreign Corporations
CoC Code of Conduct on Business Taxation CoCG Code of Conduct Group on Business Taxation DG COMP Directorate-General Competition
DG TAXUD Directorate-General Taxation and Customs Union ECOFIN Economic and Financial Affairs Council
EU European Union
FWP Future Work Package
MEP Member of European Parliament MNCs Multinational Corporations or; MNEs Multinational Enterprises
OECD Organization for Economic Co-operation and Development UNCTAD United Nations Conference on Trade and Development
4
LIST OF FIGURES AND TABLES
Table 3.1: Characteristics of forms of governance 22, 23 Figure 5.1: Timeline EU developments in EU tax policy 1996-2012 42 Figure 5.2: Timeline EU developments in EU tax policy 2013-2016 43
5
Content
1. Introduction ... 8
1.1 From public pressure to a political answer ... 8
1.2 Research “problem”: ‘but, it is legal right?’ ... 9
1.3 Relevance ... 11
2. Literature Review: the nature of the problem of global tax regulation and how to solve it ... 12
2.1 The struggle for global tax regulation ... 12
2.2 What kind of governance is needed to solve the problem ... 13
2.3 Contribution of this thesis to the literature ... 13
3. Theoretical Approach: the forms of governance that potentially fit to EU tax policy in combating tax avoidance... 15
3.1 New forms of governance vs. Hierarchical forms of governance ... 15
3.2 Network governance ... 17
3.3 Experimentalist governance ... 18
3.3.1 The concept of experimentalist governance ... 18
3.3.2 The Open Method of Coordination ... 20
3.4 Governance criteria ... 21
3.5 Do the governance theories fit EU tax policy? ... 24
4. Methodology ... 28
4.1 Case selection ... 28
4.2 Type of research and methods ... 29
4.3 Data collection ... 30
4.3.1 Documentary analysis ... 30
4.3.2 Elite interviews ... 30
4.4 Reliability and validity ... 32
4.5 Research framework ... 33
5. The EU’s approach to combating tax avoidance over the last two decades... 34
5.1 Overview of EU company taxation ... 34
5.2 The Code of Conduct on business taxation ... 35
5.3 The Common Consolidated Corporate Tax Base (CCCTB) ... 35
5.4 The OECD BEPS project ... 36
5.5 State aid investigations of tax deals ... 37
5.6 The Anti-Tax Avoidance Package (ATAP) ... 39
5.6.1 The Anti-Tax Avoidance Directive ... 39
5.6.2 The ‘Accounting’ Directive... 40
6
5.8 ... 43
6. EU’s main tax policy initiatives in governance-theoretic terms ... 44
6.1 Code of Conduct on business taxation ... 44
6.1.1 Main instrument ... 44
6.1.2 Interaction and participation... 44
6.1.3 Goal-setting ... 45
6.1.4 Implementation ... 46
6.1.5 Revision ... 49
6.1.6 Learning from diversity and local experience ... 50
6.1.7 Sanction/penalty/destabilisation mechanism ... 51
6.2 State aid investigations ... 51
6.2.1 Main instrument ... 52
6.2.2 Interaction and Participation ... 53
6.2.3 Goal-setting ... 54
6.2.4 Sanction/penalty/destabilisation mechanism ... 54
6.3 The Anti-Tax Avoidance Package (ATAP) ... 55
6.3.1 Main instrument ... 55
6.3.2 Interaction and participation... 55
6.3.3 Goal-setting ... 56
6.3.4 Implementation ... 58
6.3.5 Revision ... 59
6.3.6 Learning from diversity and local experience ... 60
6.3.7 Sanction/penalty/destabilisation mechanism ... 61
6.4 EU main policy initiatives characterized ... 61
7. The strengths and weaknesses of the Code of Conduct Group and how new initiatives respond 64 7.1 ‘the Governance’ ... 64
7.1.1 Identifying harmful tax practices ... 64
7.1.2 The criteria of the Code of Conduct ... 65
7.1.3 Differences in implementation and monitoring difficulties ... 66
7.2 Transparency ... 66
7.3 Unanimity & Mandate ... 68
7.4 The Code of Conduct Group: more weaknesses than strengths? ... 69
7.5 No obvious relationship between the Code of Conduct Group and how new initiatives respond ... 71
7
7.5.2 Code of Conduct Group and ATAP ... 72
8. State aid investigations and the Anti-Tax Avoidance Package: more promising? ... 74
8.1 State aid: one of the few options ... 74
8.1.1 Understanding for using the State aid instrument ... 74
8.1.2 Focus on US multinational corporations ... 75
8.1.3 State aid: could it be effective? ... 75
8.2 ATAP ... 76
8.2.1 ATAD ... 77
8.2.2 The ‘accounting’ Directive ... 80
8.2.3 ATAP: could it be effective? ... 81
9. Conclusion ... 84 10. Discussion ... 86 11. Bibliography ... 87 12. Appendix ... 94 12.1 Interview questions ... 94 12.2 Interviewees ... 95
8
1. Introduction
As Dietsch and Rixen (2016) start their recently published book by saying: ‘Until quite recently, a book with the title Global Tax Governance would have been unthinkable’, illustrates that global governance-thinking on taxation is something new. This introduction chapter shows that, anno 2016, it is not only ‘thinkable’ to write about global tax governance, but that there are good reasons for it. This thesis will particularly focus on EU tax governance, and therefore the introduction, the first section, focuses on the public pressure in the EU to formulate a political answer to the problem of tax avoidance. The second section elaborates on the research problem and introduces the research question, and the last section offers an explanation for the relevance of this thesis.
1.1 From public pressure to a political answer
LuxLeaks, Starbucks, Fiat, the Belgian ‘excess profit’ scheme, Panama Papers. These tax scandals raised media attention and subsequently public awareness, causing a public debate on international tax issues. There is a difference between tax evasion and tax avoidance. Tax evasion is illegal, and these structures are used by both individuals and corporations, whereas tax avoidance is legal, and these tax constructions are primarily used by multinational corporations (MNCs). As the Panama Papers reveal, most cases are illegal, but the media (including popular persons as Joris Luyendijk (writer), Diederik Samsom (Dutch Parliamentarian, Labour party) and Rick Grafhoff (Dutch Parliamentarian, The Greens)) focused more on the cases which are about tax avoidance (De Correspondent, 2016). Though the Panama Papers reveal more about individuals that illegally evade taxes, this pressure can still be used for reaching political agreement on a topic which is about tax avoidance: 9 March 2016 (after the Panama Papers), EU finance ministers reached a political agreement to exchange information on multinationals’ tax affairs by (non-public) country-by-country reporting (EUobserver 2016b).
This political agreement is just a small part of EU’s attempts to combat tax avoidance. The Code of Conduct Group is one initiative, which already exists since 1998. However, a leaked document obtained by Sven Giegold, a Member of the European Parliament (MEP)(on March 8, 2016), considers the Code of Conduct Group as a weak response against tax avoidance (Guerrera, 2016a). Another initiative is the Commission’s State aid investigations, known by the Starbucks, Fiat and the Belgian ‘Excess Profit’ cases. Some might say that, since the EU lacks formal powers to have a say in domestic tax policies, the Commission is forced to use its State aid powers when aggressive tax planning disrupts the internal market (for instance: Financieele Dagblad 2016). The new Anti-Tax Avoidance Package (ATAP), presented in January 2016, brought about change in the Commission’s formal powers, by introducing the Anti-Tax Avoidance Directive (ATAD) as the Member States reached an agreement in June 2016 (press release IP/16/1886).
9
The last two initiatives make one assume that the Commission is seeking formal powers or hard law to find a solution for the problem of tax avoidance by MNCs. Since there might be a spill-over effect when tax regimes of one country (potentially) negatively influence another country, there is a need for international coordination. The EU’s initiatives raise interesting questions about what forms of governance can be identified and whether they are effective. Can hard law be a solution? Is the work of the Code of Conduct Group truly that weak? Or should an appropriate solution be found in a combination of forms of governance? How can the initiatives be understood in their (political) context? These are important questions to understand EU’s tax governance.
1.2 Research “problem”: ‘but, it is legal right?’
The quotation marks around the word “problem” show doubt about whether we should consider it a problem. Tax avoidance is legal, as mentioned in the introduction. Tax advisors tend to avoid the word ‘tax avoidance’. Instead, they are talking in terms of ‘tax planning’. Tax planning does not have to be harmful, because it aims to avoid companies to pay double taxes when they operate in more countries. But both terms refer to the possibilities for MNCs to lower their tax costs which can in some cases be considered as tax avoidance. The Commission itself says it is not against ‘tax competition’; it just wants to make sure this tax competition is fair (for instance: Council Directive, COM/2016/026). Rixen defines tax competition as independent governments which can adopt their tax policies strategically, like designing their tax systems in order to attract new investments. These decisions can be reflected in tax rates, tax bases and the enforcement of tax laws (Rixen, 2011: 449).
The question is, who should decide on what can be considered as harmful and not harmful? Ronzoni describes two different perspectives on what can be expected on justice beyond borders: that of the cosmopolitans and the non-cosmopolitans (in Dietsch and Rixen, 2016: 201). The cosmopolitans worry about tax competition because it would exacerbate global inequalities between people. On the contrary, non-cosmopolitans are concerned about the problem-solving threats, which would mean that the State needs to intervene.
Moreover, cultural perspectives between countries should be taken into account. Small countries are more likely to be strongly attached to their national tax policies than bigger countries. Some countries don not mind being open about their taxation information, whereas other countries consider privacy more important. In addition, from a historical perspective, countries were always responsible for their own tax systems.
The different frames of the problem might as well influence the willingness of Member States to shift their powers to the level of the EU, which subsequently probably influences the choice for certain forms of governance as well.
10
When doing research in forms of governance, in this case, for the EU policy field of taxation, and to find out whether the EU initiatives are effective, it is important to take into account the different problem perceptions. The following research questions and sub-questions, which are split up in theoretical and empirical questions, all include the focus areas of this thesis.
Main research question:
Is the European Union developing a potentially effective framework to combat tax avoidance by multinational corporations? If so, based on what combination of forms of governance?
Theoretical questions:
- What forms of governance can be expected to be found in EU Tax Policy in combating tax avoidance?
- What would be an potentially effective combination of forms of governance to combat tax avoidance?
Empirical questions:
- How has the EU’s approach in combating tax avoidance by multinational corporations evolved over the last two decades?
- How can the EU’s main policy initiatives in this field be characterized in governance-theoretic terms?
- What have been the main strengths and weaknesses of the Code of Conduct Group (CoCG)? How do the Commission’s State aid investigations and ATAP proposal respond to the perceived weaknesses of the CoCG?
- Are the Commission’s new initiatives (and their forms of governance) likely to prove more effective than the CoCG in combating tax avoidance?
By answering these research questions, this thesis aims to first of all, describe the developments in the EU tax policy field. Second, an analysis will be provided on how EU’s main policy initiatives can be characterized in governance-theoretic terms. Third, it will analyse the different perspectives on whether the EU’s initiatives do effectively solve the problem of tax avoidance. By analysing these initiatives on their effectiveness and being aware of their governance structures, another objective is to find out what combination of forms of governance are appropriate for the EU to combat tax avoidance.
11
1.3 Relevance
The societal relevance of this thesis can be explained by the interest of both developing and developed countries, which have their own small and medium enterprises (SMEs) and citizens. Comparing to big multinational corporations that operate internationally, SMEs and citizens do pay considerably higher amount of taxes, because they are not able to make use of the same possibilities as multinationals can. Moreover, there is the argument that governments lose a lot of tax income, which could be used for public services like education, health care and infrastructure. According to some NGOs, like ActionAid, the poorest people in developing countries would suffer from tax avoidance significantly. Furthermore, differences in national tax policy systems between countries might cause unfair competition (by so-called spill-over effects) or result in Member States influencing each other’s tax policy systems with its risk for ‘a race to the bottom’. Research on how the problem of tax avoidance is governed by focusing on EU’s main initiatives, will offer a better understanding on what the EU can, or cannot do to put an end to harmful and unfair tax competition.
Global or EU governance in the field of taxation is relatively a new domain of academic research, as the quote of Dietsch and Rixen in the beginning of this chapter already illustrates. Since the topic has been raised on the political agenda, it also becomes relevant to research the new initiatives the EU presents, and simultaneously to put these new initiatives in the light of their predecessors. This thesis aims to contribute to the little existing literature in the field. The scientific relevance can also be found in the fact that there has not been much attention yet within the research on experimentalist governance on the topic of global or EU tax governance. As an answer to today’s complex problems, experimentalist governance architectures become more resilient in EU governance processes (Sabel and Zeitlin 2008: 280). Considering the tax avoidance problem is one of the most complex problems the EU faces, it could possibly benefit greatly from more research into experimentalist governance solutions.
12
2. Literature Review: the nature of the problem of global tax
regulation and how to solve it
This chapter offers a brief exploration of previous literature on global tax regulation and governance. As mentioned in the introduction, global governance-thinking on taxation is relatively new. In the academic field, more and more attention is paid on investigating tax avoidance, tax evasion, tax dodging, tax havens, or likewise chosen definitions which somehow relate to what is actually happening in the field of global tax (for example Kornhauser 1996 from a law perspective and Janeba and Peters 1999 from an economic perspective). Nevertheless, the governance perspective on these developments is less ubiquitous.
2.1 The struggle for global tax regulation
Global tax governance moved high up the political agenda since the financial crisis, resulting in governments starting to look more closely at tax receipts of multinationals (Politico Europe 2016a). According to Sharman, ‘tax is an area where States are the most in need of international cooperation but least able to achieve it’ (2006: 1). The rise in cross-border trade and financial flows, has far outpaced States’ capacities to control it and governments are losing power to corporations and impersonal market forces. Governments must adopt market-friendly or investor-friendly policies to compete with one another and thereby undermine States’ power (Sharman 2006: 3). Put differently, globalization is – at first glance – a cause for the failure to regulate tax competition.
Furthermore, ‘taxation is crucial to economic sovereignty’. Sharman understands the resistance of the Member States: the problem of implementing international regulation will create losers as well as winners anyhow. ‘Even if a group, or the majority of the world’s States, cooperate to tame the demands of mobile capital, the remaining free-riders will benefit by attracting more capital with the resulting economic benefits’ (Sharman, 2006: 4). Because of these free-riders, every State would be better off ‘by “giving in” to investors’ demands, particularly by cutting taxes’. This illustrates a collection action problem.
The title of Sharman’s book – Havens in Storm: the Struggle for Global Tax Regulation – indicates it is seeks to explain the nature and outcome of the struggle in the field of global taxation. The focus is rather on the problem than on providing ideas or recommendations how to govern global tax. As the title of the book of Dietsch and Rixen (2016) – Global Tax Governance: What is Wrong with it and How to Fix it – indicates it take into account both the nature of the problem and provides solutions for solving the problem. The latter is discussed in the next section.
13
2.2 What kind of governance is needed to solve the problem
The nature of the problem – including globalization developments and a collective action problem – of harmful tax competition, and thus tax avoidance, influences the types of governance that are needed to tackle it. Rixen provides preferred forms of governance to organize global tax policy. Rixen aims to answer the question ‘What are the optimal international institutions in the area of direct taxation’? Therefore, one has to deal with the following paradox: on the one hand preference should be given to institutional solutions that leave national tax sovereignty untouched to the largest extent as possible and on the other hand, to safeguard effective national sovereignty, it actually requires strong institutions that restrict nation States tax policy choices (Rixen in Dietsch and Rixen, 2016: 325-26).
As a solution for this dilemma, Rixen proposes the International Tax Organization (ITO). The aim of ITO is to find a cooperative solution, with inclusive membership (thus, all countries in the world), and that it needs strong monitoring and enforcement mechanisms. This ITO will function as a multilateral forum for governments to negotiate about the concrete rules. Also, it should entail what Rixen calls a ‘legalised dispute-settlement’-procedure, in which countries can be forced to abandon tax policies that constitute harmful tax competition (Rixen in Rixen and Dietsch, 2016: 326).
Because ‘policy changes are not legitimately chosen by the governments involved but forced upon them by the pressure of tax competition’, Rixen refers to two types of sovereignty. De jure sovereignty, which is about the legal freedom to come in action, and de facto sovereignty, which is about the ability to effectively achieve the desired goals of tax policy. Consequently, de jure sovereignty can be constrained by formal, legal measures, whereas de facto sovereignty can also be constrained by social material and other factors (Rixen in Rixen and Dietsch, 2016: 329-30). When talking about sovereignty, Rixen notes that effective protection of self-determination is only possible when self-determination has its limits. This comes from the idea that individual freedom must be constrained if it endangers the freedom of others. According to Rixen, this limitation on self-determination is required to overcome the collective-action problem which is inherent to tax competition.
2.3 Contribution of this thesis to the literature
Regarding the nature of the problem, an institution such as an ITO might be considered as the only way to solve the problem, since uniform rules (applying to every country in the world) are in favour, that need to be backed up by hierarchical governance. However, Rixen’s idea for ITO is quite ambitious, because it is hard to find an agreement on this already in the EU (which will be explained in chapter 5), let alone in the world.
One should take into account that every country has its own tax system and there is a wide variety in how these tax policies are designed. Therefore, in a situation in which some kind of hierarchical
14
governance would be possible, rules should be balanced in such a way that they are uniform enough to be effective, but it should also be possible to implement these rules in national tax policies.
These overarching rules cannot be perfectly uniform and thus imply some room for interpretation. Therefore strong monitoring and enforcement mechanisms – that are required for Rixen’s ITO – are another big challenge.
Considering the difficulties that need to be overcome in order to create an effective ITO, but at the same time being aware of the nature of the problem, this thesis aims to contribute by focusing on a less ambitious form of governance than ITO. Instead, today’s important EU tax policy initiatives are prioritized, by analysing how they can be described in governance-theoretic terms in their current form. By doing so, the EU tax policy initiatives can be better understood from a governance perspective, and at the same time, it contributes to the literature of new forms of governance – and in particular experimentalist governance – to what extent these are a manifestation of EU tax policy.
15
3. Theoretical Approach: the forms of governance that potentially fit
to EU tax policy in combating tax avoidance
In this chapter the most important theoretical concepts and how they relate to each other, will be introduced. First the new forms of governance that are developing next, and/or within hierarchical forms of governance will be discussed. After that, two new forms of governance, network governance and experimentalist governance, are elaborated further. One example of experimentalist governance, the Open Method of Coordination (OMC) is especially important as a potential manifestation of the Code of Conduct Group – one of the main EU initiatives.
3.1 New forms of governance vs. Hierarchical forms of governance
New modes of governance (NMGs) are seen as a ‘natural and successful alternative to traditional hierarchical forms of governance’ (Bartolini 2011: 1). NMGs are being understood as ‘flexible forms of governance’, as Craig and de Búrca (2011) state it. The process of policy-making (the shaping and application) of these new modes of flexible forms involve those whom the policies are meant to apply to. Some examples are the ‘new approach to harmonisation’, the ‘Lisbon Agenda’ and its Open Method of Coordination (OMC) (which will be elaborated upon in 3.3) and initiatives such as the principles of subsidiarity and proportionality and the Commission’s ‘better regulation’ initiative. Those initiatives show the European institutional actors as aware of the importance to improve effective methods of regulating and governance, ‘against the backdrop of an assumption that some of the preferred policy methods and legal approaches in the past had not delivered their intended results’ (Craig and de Búrca 2011: 161).
There has been some critique on using the word ‘new’ because new governance practices might already exist under ‘old’ and traditional forms of governance (Bartolini 2011: 13). Bartolini argues that there are some reasons that justify using the word ‘new’. For instance, ‘traditional governance’ would not clearly point to ‘the discovery, elaboration of and experimentation with forms of co-production’, indicating ‘modalities of actors’ coordination as sources of social order and collective decisions’, as well as a growing importance of a ‘decisional network’ is unemphasized (Börzel and Risse 2005 in Bartolini 2011: 15).
Instead, the traditional or classic government model starts with the idea that demands from citizens, voters, consumers and taxpayers, are transferred and aggregated in a political system, which generates policy responses. The public administration is responsible for the implementation (Bartolini 2011: 7). Hierarchical governance relies on top-down, binding or (detailed) legal measures (traditional hard law), deriving from the classic principal-agent way of rule-making and by a command-and-control approach
16
(Héritier and Lehmkuhl 2008; Craig and de Búrca 2011). Principal-agent theory is based on the assumption of perfect rationality and that responsibility and control needs to be clear. It seeks to prevent hold-ups by the agent by drawing up a complete contract with in-built incentives to prevent shirking on the part of the agent (Karagiannis 2007 in Héritier and Lehmkuhl 2008: 4). Such a complete contract requires limited room for discretion and thus has to be prescriptive.
An advantage of hierarchical governance is that if the principal is able to apply more control mechanisms, the less agency shirking there will be, leading to a better policy performance. Another advantage for hierarchical governance, is in situations that a supranational principal will be better able to control the agent at the supranational level, than is possible for multiple national principals at the national level (Héritier and Lehmkuhl 2008: 6).
However, such an ‘ideal’ type of hierarchical governance cannot always be applied, because of bounded rationality results in incomplete contracts between the principal and the agent. Contracts might not include ‘provisions for all possible contingencies and events of the future’. As a consequence, contracting partners are challenged to draw up on a basis where conditions can change, resolve conflicts in the application, and ensure compliance with the contract (Héritier and Lehmkuhl 2008: 4).
The practice of incomplete contracts between the principal and agent can be conceived at the level of the EU as well. On a EU level, the classic government can be translated into the ‘classical Community method’. This method shows a strong role for the European Commission (being the principal) for designing legislation and executing policies, which are mediated and modified by the inter-governmental process and applied by national administrations (the agents) (Kohler-Koch and Rittberger 2006: 34; Bartolini 2011: 7; Wallace and Reh 2015: 99). Though, according to Wallace and Reh, it was expected that the ‘functionalist pressures’ would be strong in the EU, in practice they find very few examples of such a centralized and hierarchical institutional process, where national policies are replaced for collective ones (2015: 102). The fact that the European Commission may prefer the Community Method whereas Member States may prefer to keep their direct competences, and both realize they cannot always achieve this, causes different levels of involvement of stakeholders (Bartolini 2011: 9). Therefore, a pure principal-agent relation is often not feasible, given the great chance for ‘incomplete contracts’ and uncertainty about the role of governmental actors.
These shortcomings show that the European Commission might have several reasons for delegation: lacking expertise, time and political attention. Nevertheless, delegation is not without risks, as Héritier and Lehmkuhl State: there is the risk of adverse selection and the risk of moral hazard (when the principal loses control over the agent; also causing problems in accountability)(2008: 5). In addition, when there is a variety of agents (like several regulatory networks), this might also ask for more control
17
by the principal (Thatcher and Coen 2008). Unclear distinctions between the principal(s) and the agent(s), mean that standard mechanisms of accountability would not work as well.
Although NMGs are – according to Bartolini – seen as a ‘natural and successful alternative to traditional hierarchical forms of governance’, others argue that new modes of governance are linked to hierarchy (Héritier and Lehmkuhl 2008: 7), or that NMGs are often introduced after legislative deadlocks (Kohler-Koch and Rittberger 2006: 37). There might be an observed ‘move away’, implying less reliance, but that does not mean that hierarchical forms of governance are disappearing (Craig and de Búrca 2011). Moreover, theories of experimentalist governance and network governance consider their form of governance as a contribution next to the existing/older/hierarchical forms of governance. As Héritier (2003) States it: ‘redistributive policies and “deeply entrenched problems” are unlikely candidates for the successful application of NMGs, although “distributive, coordinative network goods problems as well as diverse, discrete, and high complexity/uncertainty issues are more amendable to a “treatment” by new modes of governance’ (in Kohler-Koch and Rittberger 2006: 37).
According to the governance literature, a rise of NMGs can be observed. The shortcomings of hierarchical forms of governance show a need for new forms of governance, but NMGs ‘alone’ should not be regarded as some kind of magical solution to complex problems. Therefore, one should look at how these different forms of governance exist next, or within each other, to explore how they contribute in realizing policy goals. In the next two sections, network governance and experimentalist governance will be discussed further in detail, to see what the NMG’s look like more specifically.
3.2 Network governance
One new mode of governance is network governance. Network governance can be recognized by several characteristics. An important feature is its multi-level nature. Multi-level implies an interdependence of actors operating at different levels: local, regional, national and supranational (Kohler-Koch and Rittberger 2006: 34). Another feature is the growing importance of private actors in the policy process. Because of these features, the State needs to play another role. The State is vertically and horizontally segmented and it rather ‘activates’ actors than it is being authoritative; networks engage in policy formulation and implementation without clearly delegated authorisation, but have to take into account pre-existing legislative or institutional rules (Héritier and Lehmkuhl 2008: 5). Governing the (EU) network means that relevant State and societal actors have to be brought together around specific topics (Eising and Kohler-Koch 1995: 5 in Kohler-Koch and Rittberger 2006: 34). Furthermore, the orientation is towards problem-solving rather than individual utility maximization (Eising and Kohler-Koch 1995: 6 in Kohler-Koch and Rittberger 2006: 35). Therefore, trust and shared values have a key role in network governance.
18
The multi-level dimension of some of the EU networks Héritier and Lehmkuhl (2008) describe, facilitate co-operation, information exchange and implementation of supranational provisions. A large variety of actors are present, based on voluntary participation. What is missing are certain systematic mechanisms for monitoring the implementation processes, peer review, learning, and revision – which are present in experimentalist governance architectures, as will be discussed in the next section.
Even though there is broad agreement in the literature that networks play a dominant role in several stages in the policy process, the actual decisions are still taken accordingly to the hierarchical mode of decision-making (Börzel, 2005: 87 in Kohler-Koch and Rittberger, 2006: 36). Similarly, although Coen and Thatcher (2008) describe a growing in network governance by exploring European networks of regulators that have been created in key industries, the European Commission has also stressed their intention to retain control as principal over regulatory processes (Héritier and Lehmkuhl, 2008: 5).
3.3 Experimentalist governance
3.3.1 The concept of experimentalist governance
Another new mode of governance is experimentalist governance (XG), also being called as experimentalist governance architectures. These ‘XG architectures’ become more resilient in today’s EU governance processes as an answer to complex problems with a high level of strategic uncertainty, just as network governance. Policy makers are aware of their limited capacity to rely on strategic dispositions to guide action, or simply even don not know how to achieve goals (Sabel and Zeitlin 2008: 280). To respond, a joint exploration of the situation and the possibilities can be organized, assuming that joint and continuous learning makes the risks associated with persistent uncertainty more manageable (Sabel and Zeitlin 2012: 411).
Another characteristic of XG is the ‘multi-polar or polyarchic distribution of power’, which implies no single actor has the capacity to impose her own preferred solution without taking the views of others into account (Sabel and Zeitlin 2008: 280, Zeitlin 2016: 2). In addition, both network governance and experimentalist governance share the view that an exclusive focus on the interactions between nation States and supranational entities they create is ‘unduly restrictive’ (de Búrca et al. 2014: 479).
The decision-making process of the EU experimentalist architecture is described as a recursive process of framework goal-setting and revision, and happens through comparative review of implementation experience, in diverse local contexts and which is well adapted to the Union’s turbulent and polyarchic environment.
Sabel and Zeitlin argue that, although this EU decision-making architecture is not explicitly stated in their topmost Directives and Treaty provisions, this underlying architecture is regularly and decisively shaping EU governance. They recognize the framework goals and measures for gauging their
19
achievement which are established by joint action of the Member States and the EU institutions. Freedom and sufficient autonomy is given to level units to implement rules. In return, the lower-level units have to report regularly on their performance, measured by agreed indicators, and participate in a peer review to compare the results. The framework goals, metrics and procedures are periodically revised by both the actors that initially established them and new participants ‘whose views come to be seen as indispensable to full and fair deliberation (Sabel and Zeitlin 2008: 273-274). The iterative process allows for correcting problems, which were identified in a previous phase (Zeitlin 2016).
As mentioned above, new forms of governance form a challenge for accountability mechanisms. For instance, when peer review happens in isolation, it can be ineffective because its deliberation might seem to only yield recommendations that can be ignored without a penalty to whom they are addressed (Sabel and Zeitlin 2008: 305). Therefore, Sabel and Zeitlin argue that new governance mechanisms, having soft law without sanctions, risk to achieve only the very modest effects of which they are capable; because sometimes actors can be moved by ‘moral suasion or the baser fear of public embarrassment’ (2008: 305).
However, directly deliberative polyarchy (DDP) as a new form of dynamic accountability and peer review emerge under recursive framework making and revision. This mechanism disciplines the State and protects the rights of citizens without freezing the institutions of decision-making. ‘Directly deliberative’ refers to the concrete experience of actors’ differing reactions to current problems, to generate possibilities for consideration that is used. It is also polyarchic because local units learn from, discipline and set goals for each other. Therefore, DDP can be considered as ‘a machine for learning from diversity’ (Sabel and Zeitlin 2008: 276).
In addition, to overcome the risk of less effective soft law – not limited to fields where the EU has weak competences, but also where there is room for discretion under well-developed domains where the Union has extensive legislative and regulatory powers (Zeitlin 2016: 4) – than it is capable of, Sabel and Zeitlin introduce the so-called ‘penalty default’. In situations in which information, expertise and knowledge is shared by different involved actors within one network, the actors perceive freedom, to a certain extent, in finding ways to deal with a particular problem in a particular policy domain. In return for their autonomy, as said, they must regularly report on their performance and have to participate in peer reviews. But next to that, the penalty default functions as a destabilisation regime, to make sure actors will comply with the ‘rules that allow some room for interpretation’ as efficient as possible.
20
The penalty default should not be confused with the shadow of hierarchy. It differs in origin: the shadow of hierarchy arose due to the roll back of sectoral governance by shifting tasks to non-legislative and private actors policymaking bodies (Héritier and Lehmkuhl 2008: 4), because the State or public hierarchy more generally is limited, state Sabel and Zeitlin. Because of this limitation, non-State actors (which can be organized in networks) are trusted with some authority, but that they have to ‘respect the conditions of the founding bargain itself’, when they are ‘bargaining in the shadow of the State’ (Sabel and Zeitlin 2008: 307). The concept of the penalty default comes from the idea that incentives need to be created in order that involved parties produce a rule that approximates the optimum outcome, rather than trying to approximate the optimum outcome itself (Sabel and Zeitlin 2008: 306). Put differently, parties are stimulated to produce their own ‘bargaining conditions’ in the form of rules . If the involved actors do not establish a system for warranting to one another the information they disclose, and then act on what they currently know to regulate their interdependence fairly and efficiently, then a court or administrative agency has to impose a penalty default on the actors. Sabel and Zeitlin State that such a penalty default could take a form of a rule that everyone fears more than if all the actors regulate according their established warranting system (and no party might independently choose)(2008: 306). For instance, penalty defaults involve judgements by the European Courts or Commission decisions, which oblige Member States to explore how to pursue their preferred goals in such a way that it is compatible with the principles of EU law.
3.3.2 The Open Method of Coordination
Sabel and Zeitlin recognize the EU’s Open Method of Coordination (OMC) as a new experimentalist governance architectures that most clearly displays its ‘recursive properties’ (2008: 289). It is an EU policy-making process, or regulatory instrument, that is formally initiated by the Lisbon Council in 2000. The OMC has the following elements: (1) fixing guidelines combined with specific timetables for achieving (short, medium and long term) goals, (2) appropriate, quantitative and qualitative indicators and benchmarks, tailored to the needs of different Member States and sectors, as a means of comparing best practices, (3) translations of the European guidelines into national and regional policies by setting specific targets and adopting measures, which take into account national and regional differences, and (4) periodic monitoring, evaluation and peer review organized as mutual learning processes (Presidency Conclusions, Lisbon European Council, 23-24 March 2000 in Sabel and Zeitlin 2008: 290-291). However, Sabel and Zeitlin argue that actual OMC processes vary considerably in their modalities and procedures.
‘The OMC does not result in EU legislation, but is a method of soft governance which aims to spread best practice and achieve convergence towards EU goals in those policy areas which fall under the partial or full competence of Member States’. The fourth element described above, aims to help
21
Member States learn from one another and consequently improve their domestic policies. This learning is not as soft as one might think: processes as ‘peer pressure’ and ‘naming and shaming’ are terms that are often used for the OMC (European Parliament, 2014).
Radaelli (2003) focused on the Code of Conduct against harmful tax competition, talking about an OMC ‘in disguise’. Applying the OMC to the Code of Conduct, Radaelli uses slightly different characteristics including: (1) guidelines, (2) peer review, (3) best practice, (4) benchmarking and (5) learning and diffusion of shared beliefs among policy makers. Although no reference to the OMC was made by the Council’s Code of Conduct Group, Radaelli recognizes – in his article in 2003 – the first three (basic) characteristics, thus once can talk of an OMC in disguise. However, Radaelli concludes that ‘the logic of the Code of Conduct does not fit in well among the OMC aims of participatory governance and social learning’. Whereas convergence is certainly there at the level of discourse, Radaelli argues that this should not be confused with the convergence of actual tax policies in the Member States (2003: 513).
According to Radaelli, claims about learning, participatory governance, and ‘better regulation’ are applicable to the most sophisticated forms of OMC. Therefore, when referring to these properties, one should be conscious of this when using them as a benchmark. Obviously, these properties can vary considerably, from – for instance – a minimum extent of participation to a maximum.
As Radaelli applies the properties to the code, he argues that, firstly, the aim of better regulation cannot explicitly be observed in terms of a code that improves on traditional EU law and to deliver more flexible regulation. Secondly, for the aim of participatory governance, Radaelli conceives the business community and non-governmental actors as – so far – outsiders in the deliberations concerning the code, although an increasing number of actors are into international business taxation (2003: 515). For the property of learning, Radaelli only observes that it is limited to the level of discourse.
3.4 Governance criteria
From the governance theories discussed here, one can derive several categories to compare them with each other, which can be useful when analysing EU tax policy in governance-theoretic terms. Table 3.1 presents this overview.
The overview includes the following categories: ‘main instrument’, interaction and participation, goal-setting, implementation, revision, how diversity and local experience is dealt with and the sanction/penalty/destabilisation mechanism. By comparing the different forms of governance with the help of these categories, it becomes easier to apply it for EU tax policy.
22 Hierarchical governance Network governance Experimentalist governance Open Method of Coordination (OMC) Main instrument Top-down, binding or (detailed) legal measures (traditional law) (Horizontal) networks with a multi-level nature ‘governance method’ that includes experimentalist features ‘governance method’ that includes experimentalist features Interaction and participation Principal-agent relation and ‘command and control’- approach Co-operation and information exchange depending on trust and shared values
Joint exploration or ‘directly deliberative polyarchy’; taking into consideration
concrete actors’ experiences and local units learn, discipline and set goals for each other, multi-polar distribution of power Participatory governance Goal-setting Central; principal draws ‘complete contracts’, prescriptive ‘incomplete contracts’ ‘incomplete contracts’, recursive process of framework goal-setting, taking into account revision results Fixing guidelines with specific timetables for achieving goals, ‘better regulation’; improving traditional law by providing flexible regulation Implemen-tation Clear responsibilities for the agent, State is authoritive Actors have freedom to operate by taking into account pre-existing legislation and institutional rules, State ‘activates’
Lower-level units are given substantial discretion, but in return: regularly report on performance (measured on agreed indicators) and have to take into account what is learned from joint and learning experience (through the revision) and solve earlier identified problems Periodic monitoring with the help of appropriate, quantitative and qualitative indicators and benchmarks
Revision Limited to the principal owning control mechanisms No clear revision and learning processes Peer-review, comparative review of information experience, coordinated learning from local experience
Best practices are compared, periodic evaluation and peer review as mutual learning processes; peer pressure and naming and shaming
23 Hierarchical governance Network governance Experimentalist governance Open Method of Coordination (OMC) How diversity and
local experience is dealt with One-size-fits-all solutions A large variety of actors is present in the networks, voluntary participation
Goals are adapted to various local contexts and coordinated learning from local experience is taken into account in the revision process. EU guidelines are translated into national and regional policies, indicators and benchmarks are tailored to the needs of different Member States Sanction/penalty/ destabilisation mechanism Hierarchical solutions, on the EU level: sanctions as fines and infringement proceedings when MS fail to adopt EU legislation
None Penalty default; destabilisation mechanism to force reluctant actors to cooperate in frame-work rulemaking, by threatening them to reduce their control, on the EU level: judgments by EU courts or Commission decisions
None
Table 3.1 Characteristics of forms of governance
The ‘main instrument’ is a bit of a controversial category, since the ‘main instrument’ of some of the forms of governance is a bit vague. For experimentalist governance and the OMC, its powers rely in their ‘governance method’, which start to make sense when looking at its experimentalist features. For network governance, obviously, its power lies in their networks to find local solutions. Nevertheless, when comparing the forms of governance, it becomes clear that for hierarchical governance, it is not its method, but the fact it can produce legally-binding measures, that makes it a powerful form of governance. Although the non-hierarchical governance forms itself do not provide legislation, it does not mean they are only producing soft law. Rather they are organized next to hierarchical governance and their strength can be observed in how soft law is implemented, or how experimentalist features contribute to the implementation of hard law. Of course, the governance methods of hierarchical governance and network governance matter as well. But comparing to experimentalist governance and OMC as an example of experimentalist governance, the main strength of hierarchical governance and network governance can be founded in respectively their capability to provide hard law, and the room for discretion that networks have to solve local problems.
24
The interaction and participation category is especially relevant to make a distinction between hierarchical governance and forms of non-hierarchical governance by looking at how (by first knowing who) relevant actors interact. Whereas hierarchical governance can be recognized by a strict principal-agent relationship, the forms of non-hierarchical governance can be characterized in terms of cooperation. Experimentalist governance goes even further than classic cooperation in networks, by introducing the directly deliberative polyarchy. The goal-setting, implementation and revision are important categories to understand how the forms of governance design policies and regulations. Moreover, it is relevant to know whether diverse local contexts and experiences of different actors are taken into account by these forms of government, and how the prefer to implement those goals.
The last category should be added because it distinguishes the forms of governance whether they have the means to apply some kind of sanction, penalty of destabilisation mechanism (like in the form of a penalty default), in case goals are not implemented in the preferred or most efficient way.
3.5 Do the governance theories fit EU tax policy?
This section aims to answer the theoretical research questions:- What forms of governance can be expected to be found in EU Tax Policy in combating tax avoidance?
- What would be a reasonable combination of forms of governance to address the problem of tax avoidance?
For EU tax policy, there are several reasons for analysing hierarchical governance, network governance and experimentalist governance. From the governance literature in general one can conclude that different forms of governance can complement each other, varying in composition depending on the problem structure, previous governance structures and historical context. The article of Radaelli illustrates that it is interesting to look to what extent the Code of Conduct is a manifestation of OMC. However, Radaelli emphasizes that the Code of Conduct is embedded in a wider context and therefore cannot be assessed without the consideration of other interdependent initiatives to crack down harmful tax competition (2003: 515). Therefore, it is relevant to look at the initiatives that are in ‘the Code’s context’, and what forms of governance fit those initiatives. The introduction of this thesis showed that that EU tax policy consists on more initiatives than only the Code of Conduct. Though there is no clear evidence for relationships between the EU tax policy initiatives that aim to fight harmful tax competition/combat tax avoidance, based on the literature, some assumptions can be made, based on the theoretical framework.
25
Soft law is used in the EU where the lack of political consensus blocks the road to traditional law-making processes (Radaelli 2004: 514). The same might be apply for the Code of Conduct (and the fact it is still existing) as a response to consensus blocks for traditional law. Although Radaelli’s conceives in his empirical evidence no reason for the latter being the case, it does not mean this argument is out of question, considering that the Code of Conduct is not a very transparent body.
Such legislative deadlocks are the result of the policy field of taxation standing closely to national sovereignty. In addition, unanimity and the lack of a specific Treaty base for direct taxation (Radaelli 2003: 516) make the path to legislation even narrower. Still, coming up with new strategies as new forms of governance, one continuous to need some basis of a shared perception for the need to cooperate (Lavenex 2015: 28), since (some) actors might draw meaningful short term benefits from non-cooperation. Especially one might have doubts of a shared perception on taxation, since there is a fine line between ‘acceptable’ and ‘harmful’ tax practices.
As a consequence, ‘the existence of strong incentives to cheat or free-ride on cooperative arrangements created the need for hierarchy, including precise commitments and strong enforcement mechanisms, in order to reduce the potential advantages from non-cooperation’ (Koremos et al. 2001: 773ff., 776ff. cited in Lavenex 2015: 28). In EU taxation, some Member States as well might see potential advantages from non-cooperation, relying heavily on attractive corporate tax regulations, since it can be a significant component of a Member States’ economy. It is likely that for some Member States the possible outcome of cooperation on international taxation might not compensate for the benefits of non-cooperation. Therefore, reasoning in reverse, a disappointment with soft law, could also lead to willingness to solve problems with hard law. The Anti-Tax Avoidance Directive (ATAD), as part of the Anti-Tax Avoidance Package (ATAP) and/or the State aid investigations, that refer to EU State aid rules, as forms of hard law, might potentially be an example of a response to disappointing Code of Conduct results. This response however does not certainly have to do anything with the concrete tax topics the initiatives are dealing with – as they might differ – but at least it can be a response to the fact to find better ways to realize the objective of combating tax avoidance.
Nevertheless, the risk of non-cooperation remains as well for hierarchical governance. And even if there are strong demand-side pressures for centralisation and regulation, existing institutional arrangements and organizations limit and shape the supply of new institutions (Thatcher and Coen, 2008: 830). However, the serious negotiations on ATAD during this research, and the fact agreement is found with all Member States on 21 June 2016, it turns out that there is willingness of the Member States for designing hard law (European Commission 2016c).
26
A reason for looking more closely to the Code of Conduct and to follow-up on Radaelli is because his article is published in 2003, and now, thirteen years later, the Code still exists, might be developed and/or more documents might be available that were not yet accessible in 2003. Radaelli’s empirical evidence shows that Member States share more information than in the past and that they have not introduced aggressive tax schemes, but that it is too early to say whether convergence of tax policy beliefs has produced domestic policy change (2003: 516). Therefore it is interesting to find out to what extent this is the case today. Moreover, one might wonder if the Code of Conduct is best characterized in terms of an OMC, regarding Radaelli’s conclusions about what is lacking, and whether the Code of Conduct might rather be characterized as a form of network governance. Since certain characteristics of an OMC might be missing, such as participatory governance and learning, one might rather choose a form of governance that is primarily based on dialogue and cooperation depending on trust and shared values such as network governance. On the contrary, there might be reasons for considering experimentalist governance, like an OMC, as being present in the Code of Conduct, as there might be more information available and the Code might have been developed since 2003.
Next to the Code of Conduct, it could also be relevant to explore the presence of experimentalist governance architectures for the Anti-Tax Avoidance Package, because there is a chance it appears under more centralised and hierarchical institutions. The Single Supervisory Mechanism (SSM) for Eurozone banks for example, is designed as a relatively centralised and hierarchical institution. But despite the emphasis on harmonisation, Zeitlin observes that the SSM is not seeking to impose a single ‘one-size-fits-all’-approach to supervision of banks or that they have to homogenise their business models. Rather, the SSM has the aim to ‘ensure consistency across institutions and supervision tailored to [their] specificities… by balancing uniform supervisory anchor points with constrained supervisory judgment’, by doing so, the SSM accommodates banking diversity, which would remain ‘very desirable for financial stability (Nouy 2015b cited in Zeitlin 2016: 11). As national fiscal systems also differ just as national banking systems do, it would be interesting to find out whether everybody is being treated the same, by recognizing every Member State is different. Moreover, one should note that the ATAP is more than just a Directive, and that it is accompanied with another ‘amending’ Directive, a communication, recommendation and a staff working document. The focus on this thesis will be on the ‘hard law’ part of ATAP; the two Directives.
Whether experimentalist governance features can be recognized for the Code of Conduct and/or ATAP or not, it might both be interesting to explore the added value of experimentalist governance and, if it is limited or not present at all, to look for possible gaps where it could fit in in order to realise a more powerful strategy.
27
Whereas the Code of Conduct and ATAP initiatives are subject for exploring to what extent hierarchical, network and experimentalist governance are apparent, this is not self-evident for the State aid investigations. Rather, the State aid investigations can – at first sight – be considered as some kind of support instrument next to other EU tax initiatives, aiming to raise awareness about certain tax agreements being reprehensible. The cases about multinationals like Starbucks and Fiat in respectively the Netherlands and Luxembourg, revealed they have tax agreements that are not in line with EU State aid rules. These cases have received quite some media attention and subsequently raised awareness about what is fair taxation. Next to the public pressure, the multinational has to pay back its taxes to the concerning Member State and the Member State has to put an end to the harmful tax agreement. By doing so, the State aid investigation might be regarded as some kind of destabilisation mechanism in the form of a penalty default, in order to stimulate Member States to draw up rules that are compatible with the EU State aid rules, because they fear a State aid investigation. When it appears that this expectation is true, the State aid investigations might be a part of a wider experimentalist governance strategy.
The pros and cons of different forms of governance become apparent when looking more closely at the EU tax policy initiatives, which motivates the choice – next to Radaelli’s context argument – even more to investigate how different forms of governance, through different initiatives, are ‘combined’ in order to solve the same problem of tax avoidance. The relationship between hierarchical governance and new forms of governance vary. If a relationship can be found, that makes it all the more important to analyse, so that it can be determined to what extent the different forms of governance are present. However if no relation is found, it remains useful to explore what EU tax policy looks like in governance-theoretic terms and whether these initiatives and their governance characteristics can be considered as effective. Nevertheless, based on the theory one might expect that relationships between initiatives ‘representing’ different forms of governance, are present.
There is little evidence to justify whether network governance, experimentalist governance and OMC are forms of governance as relevant enough to apply to EU tax policy. The EU tax initiatives have to be further examined thoroughly before determining the presence of a form of governance, which therefore is done in chapter 6. However, based on the problem structure of tax avoidance and the emerging initiatives to do something against it, and their structures that can be observed at first sight, it is likely that different forms of governance will be combined, because if this were not the case, it would make EU tax initiatives less effective. This argument also works the other way around: if the different forms of governance are not strategically combined, this might be one of the reasons that the EU is not capable for developing an effective framework to combat tax avoidance.
28
4. Methodology
In this chapter an explanation is given on the case selection, which methods will be used and how the data is collected.
4.1 Case selection
To find out whether the European Union is developing a potentially effective framework to fight tax avoidance by multinational corporations, the research focuses on the three main EU initiatives:
1. The Code of Conduct Group on business taxation 2. The State aid investigations
3. The Commission’s Tax Package 2016 (ATAP)
The case-selection is be considered as information oriented: to maximize the utility of information from small samples and cases are selected on the basis of expectation about their information content (Flyvberg, 2011). It is expected that these three cases are most interesting to research EU tax policy on combating tax avoidance. One reason is that these initiatives can be regarded as simply the initiatives that have recently made the most impact, or are expected to be an important part of EU’s tax policy strategy. Another reason is the differences in approaches of these initiatives, which make it particularly interesting to analyse the initiatives in their governance-theoretic terms, but also how these different initiatives relate to each other.
The Code of Conduct Group as a case is relevant because it already exists for more than eighteen years and it is basically the first attempt to get rid of harmful tax practices. Therefore, it becomes interesting to look at its learning process over time. Furthermore, as explained in the previous chapters, there are some reasons to expect the CoCG to contain experimentalist governance features. There have been several critiques of the CoCG for its slow decision-making process and for not being transparent. New initiatives – the State aid investigations and ATAP – are being introduced by the European Commission, which makes one suppose that since the CoCG is not sufficient in combating tax avoidance, new initiatives are necessary.
The State aid investigations – Starbucks in the Netherlands, Fiat in Luxembourg and the Belgian Excess Profit case – have received extensive media attention. In this way, the State aid investigations can be considered as a supporting vehicle to make the tax avoidance topic rise on the political agenda, for both European as national agendas. Furthermore, describing this initiative in governance-theoretic terms, it might be useful if it can function as some kind of penalty default, since it might influence Member States to make up tax rules that are compatible with the EU State aid rules.
29
The Anti-Tax Avoidance Package, including the Anti-Tax Avoidance Directive, is relevant in the sense that is a first initiative that seriously addresses the problem of tax avoidance. In the period the research for this thesis took place, it did not only appear that ATAD had a serious chance to become hard law. As a matter of fact, all Member States reached an agreement on ATAD, which makes it even more relevant to focus on the willingness for legally-binding measures for EU tax policy. In governance-theoretic terms, it is relevant to find out whether ATAD is better described in terms of hierarchical governance or whether experimentalist governance characteristics can be recognized as well, as the Directive might offer some room for interpretation for the implementation of the anti-abuse measures on the Member State level.
4.2 Type of research and methods
This thesis applies for a qualitative research approach. This choice depends on what fits best to the nature of the ‘variables’ – in this regards better be described as ‘characteristics’ of different forms of governance – that are researched, as well as the aim of this thesis (Nyikos and Pollack, 2003: 7). The nature of the characteristics (that are as well dependent and independent, or it is not even clear whether they can be clearly separated) of the (multi-level) governance concepts, and the aim of this thesis (to investigate to what extent this forms of governance are recognized and who these actually contribute to addressing the problem), do require a qualitative research method, which involves ‘the examination of one or a few cases by descriptive methods’ ((Nyikos and Pollack, 2003: 6).
An important tool in qualitative and ‘within-case’ research is process tracing. The method of process tracing is especially well-suited for exploring causal processes and analysing complex decision-making (George and Bennett, 2005 in Tansey, 2007: 765). This thesis will not necessarily focus on causal processes. Rather, it will use the variant of process-tracing of explaining outcomes that happen in practice. The ambition for this variant is ‘to craft a minimally sufficient explanation of a particular outcome’ (Beach and Pedersen, 2013: 18), by explorative research and providing detailed descriptions. By doing so, it allows the researcher to characterize EU tax policy on combating tax avoidance in governance-theoretic terms and how this influences particular outcomes.
Moreover, explaining-outcome studies often have theoretical ambitions that reach beyond a single case (Beach and Pedersen, 2013: 19). For this thesis, for instance, one of its ambitions is to show to what extent experimentalist governance architectures can help addressing problems in combination with the existence of some kind of shadow of hierarchy or penalty default, when it is not self-evident that a perception of all Member States is shared that cooperation in a particular (financial) policy field is needed.
30
Unsurprisingly, an understanding of how the EU policy process works in order to do a solid process-tracing analysis is essential (Nyikos and Pollack, 2003). For this reason, I will take into account the operation of EU policy processes when adopting a process tracing analysis.
4.3 Data collection
Two most important ways of data collection of this thesis are documentary analysis and elite interviewing.
4.3.1 Documentary analysis
For the documentary analysis, both official documents as media articles have been analysed. Official documents from EU institutions are (amending) Directives or proposals for Directives, (Staff) Working Documents, Recommendations, Communications (from the European Commission to the European Parliament and the Council) memos and press releases. For the media articles, there is a particular focus on the EUobserver and Politico Europe and in addition, some Dutch media articles. Furthermore, some reports are analysed, such as reports by the European Parliament, and some reports of NGOs.
4.3.2 Elite interviews
Elite interviews are an essential source of data for this thesis. In this section the sampling method, elite interview as a relevant form of interviewing and the choice for semi-structured interviews and its advantages are elaborated.
Sampling method
The sampling method for identifying and accessing the interviewees are non-probability. This fits a process-tracing study in which the ultimate goal is to reduce randomness as much as possible (Tansey, 2007: 765). One non-probability approach is convenience sampling: interview the ones that are willing to talk (Tansey, 2007: 769). Another approach is snowball-sampling: every interviewee is asked at the end of the interview if he or she can recommend someone to talk with (Tansey, 2007: 770). The sampling will also be purposive: the knowledge of the researches has a guiding function in identifying particular respondents (Tansey, 2007: 770). For this thesis, a particular focus was on finding respondents that have different perspectives, in order to provide more complete answer to the sub-questions about the shortcomings of the CoCG and whether the State aid investigations and ATAP are more promising.
Elite interviewing and its uses
The most relevant form of interviewing for process tracing approaches is elite interviewing, since elite actors are probably critical sources of information about the political processes to be analysed (Tansey, 2007: 766). The most evident uses of elite interviews are (1) to corroborate what has been established from other sources and (2) to reconstruct an event or set of events (Tansey, 2007: 766). Next to the
