The differentiated politicisation of European tax governance

Abstract This article addresses the politicisation of European tax governance in mass media. Although taxation is commonly assumed to remain a national prerogative, European institutions have extended their reach far into national tax regimes. Whether this expansion of EU authority towards one of the core functions of the nation state is accompanied by politicisation is the question to be addressed in this article. To analyse the causal nexus between growing EU authority – measured in terms of formal authority and tax legislation – and the politicisation of EU tax governance, the paper compares public debates in Germany, Ireland and Switzerland over a time period of 30 years between 1981 and 2011. Based on a newspaper content analysis, the empirical results demonstrate that EU tax governance is increasingly politicised in all three countries. Varying levels and content of politicisation across time and space are explained by national economic and socio-cultural conditions.

The first account centres on the authority transfer from the national to the European level, the second on the distributional effects of EU tax governance and the third on socio-cultural conditions. My main claim is that by integrating these approaches into a combined explanatory framework rather than treating them as competing explanations, it becomes possible to explain not only varying levels of politicisation across countries but also its specific timing and content. I suggest a framework which views authority transfer as the driving force of politicisation (De Wilde and Zürn 2012;Zürn et al. 2012) and distributive effects (Anderson and Reichert 1998;Eichenberg and Dalton 2007) and national socio-cultural contexts (Ecker-Ehrhardt 2012; Hooghe and Marks 2009) as the conditioning variables which shape the country-specific timing, strength and direction of the authority transfer effect. In this view, high levels of politicisation in media debates are most likely, when EU institutions extend their competences to issue areas of wide national concern and thus create a strong conflict between national economic or socio-cultural preferences and European aspirations. The content and framing of these debates is expected to be shaped by the economic/socio-cultural nature of conflict and countryspecific socio-cultural conditions.
To analyse this causal nexus, the paper compares public debates in Germany, Ireland and Switzerland between 1981 and 2011. Based on a newspaper content analysis, the empirical results suggest that EU tax governance has become politicised in all three countries. Driven by an interaction of increasing formal authority in neighbouring issue areas and a growing frequency of EU tax legislation, and conditioned by varying national contexts, politicisation follows substantially different trajectories in the three analysed publics. The data show that cross-national variance is shaped by national economic and socio-cultural contexts. The more national economic growth models are put into question by EU tax governance and the stronger the conflict between exclusive nationalists and cosmopolitans in a given country, the stronger is the politicisation of EU tax governance. Economic and socio-cultural conditions shape the content and evaluative thrust of public debates.
The paper proceeds in six steps. First, the theoretical argument is introduced. Second, the research design, methods and data are presented. The third, fourth and fifth section detail the empirical findings for the individual country cases. The conclusion compares results across countries and discusses their implications for further research.

Explaining politicisation: an integrated explanatory framework
This contribution aims to explain the timing and level of politicisation in mass media and the specific content of these debates. I draw on three strands of literature to derive an integrated explanatory framework, which is fine-grained enough to account for these different dimensions of politicisation (for an encompassing literature review, see De . The first explanation centres on authority transfer from the national to the European level as the main driver of politicisation. The second approach focuses on the distributive consequences of European integration and the third perspective emphasises the influence of national socio-cultural conditions -in particular exclusive nationalism and cosmopolitanism. A common claim in the European integration literature holds that the transfer of political authority to the EU is the driving force of politicisation (De Wilde and Zürn 2012;Zürn et al. 2012). It is argued that the growing scope (Pollack 1994) and level (Börzel 2005;Leuffen et al. 2013) of EU authority accounts for politicisation, because citizens are more frequently confronted with the effects of European governance. Citizens consider the EU no longer to be the sum of national preferences, but as an important independent actor, whose governance has direct implications for their lives. This growing importance and visibility of European governance is expected to provide the basis for politicisation and shape the levels of politicisation (Grande and Hutter 2015).
Starting from this premise -although often implicitly -a second strand of literature holds that the economic effects of European integration shape the level and content of politicisation at the national level. Because authority transfer alone cannot explain varying patterns of politicisation across countries and over time, this political economy literature argues that individuals who personally expect to gain from European integration will support it, while potential losers will oppose it (Anderson and Reichert 1998;Eichenberg and Dalton 2007).
In the absence of an agreed-upon economic model of EU tax governance's individual distributive effects, I turn to sociotropic variants. These models suggest that international distributive concerns, i.e. the question of how burdensome national publics perceive EU tax governance to be for the national economy as a whole, explain the cross-national variance of politicisation (Bechtel et al. 2014;Mansfield and Mutz 2009). From this perspective, growing levels of economically framed politicisation can be expected when EU tax governance creates distributive effects between countries. Theories on international income tax competition predict such distributional effects between small, low-tax and large, high-tax countries due to trade liberalisation (Kanbur and Keen 1993). If the difference in country size is large enough, the smaller country is better off under tax competition because the revenue loss from a tax cut is relatively minor compared with the major revenue gain from the inflow of the foreign tax base of the larger country (Genschel and Schwarz 2011). Consequently, small states are likely to profit from international tax competition whereas large countries are potential losers (Kemmerling and Seils 2009). Under these conditions, politicisation should grow and take substantially different forms across countries when EU tax governance infringes national economic growth models based on income tax competition. Public debates should intensify and focus on economic implications as soon as EU tax governance targets income taxation.
A third perspective claims that identity rather than economic concerns structures how citizens react to authority transfer. One camp of scholars (Ecker-Ehrhardt 2012;Inglehart and Welzel 2005) argues that cosmopolitan values and the so-called skill revolution (Hainmueller and Hiscox 2006;Koehn and Rosenau 2002) shape the levels and content of politicisation (Rixen and Zangl 2013). Cosmopolitan competencies and values provide individuals with the necessary cognitive abilities and identity to grasp the complexity of international politics and to engage with them proactively. Cognitive mobilisation increases the public awareness of and identification with European institutions and thus contributes to growing levels of politicisation framed in supportive cosmopolitan terms.
A second camp emphasises that citizens holding an exclusive nationalist identity feel threatened by the growing political authority of international institutions and therefore engage in more critical politicisation framed in nationalist terms (Ecker-Ehrhardt 2014; Hooghe and Marks 2009). However, contrary to cosmopolitan identity, exclusive nationalism is expected to lie dormant as a catalyser of politicisation as long as it is not mobilised by populist right-wing parties (De Vries and Edwards 2009;Hooghe and Marks 2004: 417). Only if right-wing parties tap into the feeling of threat to national identity, triggered by authority transfer, are exclusive nationalists mobilised and contribute to politicisation on the basis of nationalist justifications (see also Kriesi et al. 2012).
The institutionalist and the political economy perspective and both variants of the socio-cultural explanation have already been demonstrated to enhance our understanding of the politicisation of European governance. Yet individually all three strands of literature have a hard time explaining the country-specific timing, level and framing of politicisation. This more specific understanding of the differentiated nature of politicisation, however, is crucial if we want to gauge its consequences for EU decision-making Schimmelfennig et al. 2015). Consequently, I suggest an explanatory framework which views authority transfer as the driving force of politicisation and country-specific economic and socio-cultural context as two potentially conditioning variables. In this view, high levels of politicisation are most likely when EU institutions extend their competences to issue areas of wide national concern and thus create strong conflict between national economic or socio-cultural preferences and European aspirations. The content and framing of these debates is expected to be shaped by the nature of this conflict and country-specific socio-cultural conditions. The more conflict is driven by economic/socio-cultural preferences, the more intense should be the focus of public debates on the respective aspects of authority transfer. Whether media debates are framed in supportive cosmopolitan terms or result in nationalist backlash is finally expected to be conditioned by the presence and mobilisation of cosmopolitans and exclusive nationalist.

Research design and data
To test the integrated framework, I provide three in-depth country case studies, which trace politicisation processes over 30 years , and a crossnational comparison. In the spirit of the most similar systems design (Gerring 2007: 131-9), I selected countries that are similar in many important respects and differ in terms of the explanatory factors. Germany, Ireland and Switzerland are established liberal democracies with consolidated political institutions and party systems. They belong to the economically most developed and internationally integrated countries in the world and are thus strongly affected by the globalisation of taxation (Paris 2003).
As regards the institutionalist perspective, I selected countries which vary in terms of the authority transferred to the EU (for a similar approach see Hutter and Grande 2014: 6). The selection of Germany as a founding member of the EU, which has not opted out of major steps of European integration (Katzenstein 1997), Ireland as an accession country, which has opted out of a number of integration steps (Laffan and O'Mahony 2008), and Switzerland, which is not a formal member of the Union but is considerably integrated with the EU via bilateral treaties (Kriesi and Trechsel 2008: Ch. 11), represent the entire spectrum of levels of integration. Authority transfer is measured with the help of three indicators: first, the formal centralisation of political authority at the European level over all issue areas of European governance (Börzel 2005;Leuffen et al. 2013); second, the formal centralisation of political authority in the field of taxation; third, the scope and level of EU tax legislation as a more issue-specific indicator (see the online appendix). This threefold measure goes beyond the current focus in much of the literature on formal authority transfer. It is intended to capture the informal dynamics that shape political outcomes and potentially their politicisation but which are frequently underestimated in much of the current literature (Kleine 2013).
To account for the political economy perspective, the sample includes three countries holding different positions on tax competition. Germany represents large, high-tax countries, which potentially lose in asymmetric tax competition. Ireland's 'Celtic Tiger' revolution was not least based on its low corporate income tax regime and Switzerland has built a strong reputation not only for its low corporate tax rates but also for its banks' discreet handling of individual income.
As regards socio-cultural contexts, the sample provides for varying shares of cosmopolitans and exclusive nationalists across countries and over time. It includes countries with high and medium levels of mobilisation, which differ in the distribution of both groups. This variable is operationalised with the help of two proxies (see the online appendix). First, the share of exclusive nationalists and cosmopolitans is measured with the help of World Value Survey and European Values Study data on the feeling for territorial belonging.
Respondents who see themselves as belonging primarily to their continent or the world are classified as cosmopolitans. Exclusive nationalists are identified by subtracting from the group of respondents, who see themselves as belonging primarily to the nation, those who choose continent or world as their second territorial identity (Jung 2008). Second, the mobilisation of populist right-wing parties is gauged on the basis of cumulative vote shares in national elections of those parties classified as populist right by Norris (2005: 54-7).
The politicisation of EU tax governance is defined as a process which increases the salience and polarisation of EU tax legislation among an expanding set of actors in public debates . To gauge the level of politicisation and its content, I analyse mass media debates. Two opinion-leading centre-left and centre-right quality newspapers per country are examined. 1 I apply political valuation analysis (Schmidtke and Nullmeier 2011) to code evaluations of EU tax governance and elements of political claims analysis (Koopmans and Statham 1999) to identify demands, calls to action or proposals. In combination, both approaches target those elements of public debate that form the core of politicisation. They identify and describe public speech acts that problematise and contest a given issue by looking at either diagnostic assessments or prescriptive statements (Statham and Trenz 2015). The analysis identifies the evaluative thrust of statements and how evaluations and demands are justified. I distinguish four types of justifications to gauge the effect of economic and socio-cultural conditions: economic-nationalist, economic-cosmopolitan, cultural-nationalist and cultural-cosmopolitan.
The analysis was conducted for four randomly constructed weeks per newspaper and year. This stratified random sampling technique helps to create a manageable and representative text corpus from which inferences can be drawn. The corpus consists of 578 newspaper articles that contain 707 evaluations and claims (for the coding scheme see the online appendix).
I operationalise politicisation with the help of the three core components of the concept: salience, polarisation and actor expansion Hutter and Grande 2014: 2;Zürn et al. 2012); and an index integrating all three variables. Salience deals with the level of public debate and is measured by the share of newspaper articles on EU tax governance of the total number of published newspaper articles. Polarisation pertains to the plurality of evaluations and demands and is measured by the share of articles that contain a minimum of two evaluations or claims of the total amount of articles on EU tax governance. This minimum threshold builds on the idea that a polarisation of opinions can only be identified when there is more than one opinion on an issue. Actor expansion refers to the share of non-executive participants in mass media debates. It is measured by the share of evaluations and demands uttered by non-executive speakers, like economic actors, party politicians, representatives of NGOs and unions and journalists. To compute a politicisation index, I follow Hutter and Grande (2014: 3-4) and multiply salience by the sum of polarisation and actor expansion. The index can reach a maximum value of 20,000, if EU tax governance was critically discussed by only non-executive speakers in all articles published in the analysed newspapers.

Germany: economic pressures and cosmopolitan replies
The German case is characterised by strong authority transfer. Germany is the most Europeanised country in the sample. Both overall European formal authority and tax legislation grow over time. They limit German tax sovereignty to an ever-increasing extent, despite the EU's constantly low formal authority in the field of taxation (see upper right-hand plot in Figure 1). Formal authority transfer is particularly strong in the aftermath of the Single European Act (SEA) and the Maastricht Treaty and less pronounced during the later revisions of Amsterdam, Nice and Lisbon. The frequency and scope of EU tax legislation grow more slowly but constantly over time. During the 1980s, EU tax legislation was limited to a low number of legislative acts primarily targeted at the coordination of VAT systems and excises. Subsequently, the EU became more active in these fields and expanded the scope of legislation to other major   areas of taxation. Since the end of the 1990s even income taxation is a regular -albeit still not very frequent -target of European tax legislation. Taken together, formal authority transfer and legislative output should contribute to a constant growth of politicisation levels since the 1980s, which should be most pronounced during the Maastricht period.
To what extent is the effect of authority transfer shaped by national economic and socio-cultural conditions? To address this question, the case study traces politicisation over time. If politicisation is mainly structured along the economic dimension, we should observe growing levels of politicisation since the 1980s and a pronounced intensification of the debate since the 1990s, when EU market integration began to accelerate European tax competition ) and when the EU started targeting national income taxation. Due to Germany's role as a high-tax country, public debate should be largely supportive and focus on the economic consequences of European integration. If the German debate is more strongly structured along the socio-cultural dimension, levels of politicisation should rather stagnate during the 2000s and debates should be mainly framed in a cosmopolitan fashion. The share of both exclusive nationalists and cosmopolitans decreased during the 2000s, as the lower left-hand plot in Figure 1 indicates. Furthermore, the vote share of populist right-wing parties in national elections is low over the entire investigation period, indicating a weak mobilisation of exclusive nationalists.
The quantitative indicators show that the politicisation of EU tax governance in Germany can be broken down into three main sequences, which suggest that both economic interests and socio-cultural concerns structure the debates. During the first period, EU tax governance received increasing media attention but public debate remained volatile and weakly polarised (1981)(1982)(1983)(1984)(1985). During this period, the German debate was marginal and concerned with a limited number of specific policies, like the harmonisation of VAT or excises on liquid gas. News coverage was largely descriptive and evaluations and demands were invariably brought forward by German national executives from both centre-right and centre-left parties. During these years, the scattered debate resembled a foreign policy discussion, which cannot be classified as politicised.
This situation changed after the enactment of the SEA. Already in 1986, but even more substantially between 1987 and 1989, attention to European tax governance increased. In essence this is supposed to say that the SEA had the goal of harmonizing VAT and excises, aimed for the provisions of the SEA, the issue of withholding taxes on capital income took centre stage. These discussions appear to be shaped by cosmopolitan thinking. Nationalist claims to protect national tax sovereignty against EU interference were absent. The debate among a small group of national executives, EU representatives and economic actors was mainly focused on how the single European market can be furthered without risking Europe's and Germany's competitiveness.
After the enactment of the Maastricht treaty, growing salience and actor expansion were increasingly accompanied by polarisation. All three components of politicisation intertwined and formed a fully politicised debate, which was characterised by a diversification of issues under scrutiny and a largely supportive discussion of European tax harmonisation in general and of a variety of specific taxes. Criticism was mainly fed in by foreign executives, especially from Great Britain and Switzerland, who occasionally employed nationalist justifications to warn against the European threat to tax sovereignty. In less nationalist terms, these speakers also argued that Europe's competitiveness could be called into question by too far-reaching tax harmonisation.
By contrast, the relatively few critical German voices mainly called for improving the current system and claimed that EU tax legislation did not go far enough. An exclusive nationalist defence of tax sovereignty by German commentators was largely absent and even weaker nationalist claims were rare. Strong support for harmonising, for instance, excises on energy, capital income taxation and VAT was primarily justified in economic-cosmopolitan terms. In these discussions between representatives of German banks and other economic actors, German and foreign executives and German journalists European tax legislation was frequently seen as a solution to common European problems, like unfair tax competition, economic stability or even climate change.
The final sequence during the 2000s was characterised by stagnating or even decreasing levels of politicisation. This development seems to result partly from the resolution of one of the most controversial issues of the previous period: the EU savings directive, which deals with the cross-border taxation of capital income. Although the savings directive and its weaknesses remained an important issue for national executives, EU representatives and German banks during the subsequent years, the final decision on this directive slowed down the politicisation of EU tax governance. During the years to come, new issues like the harmonisation of corporate taxation, the introduction of a financial transaction tax and a more ecological tax system dominated the discussions among a broad variety of actors but did not generate levels of politicisation similar to those surrounding the savings directive. Although German speakers were still largely in favour of more EU tax governance, the number of sceptical voices grew during this period. Due to the alleged failure of the savings directive and slow progress in other areas, like corporate taxation, even non-executive speakers, such as tax lawyers, NGOs or laypeople, became more critical. In the same vein, more general demands for a harmonised European tax policy, as a means to stabilise the EU and the monetary union, were less frequent. The final period appears to be characterised by some disenchantment among the German public. However, there was also no strong nationalist backlash against EU tax governance, demanding a reduction of EU tax authority.
Overall, the threefold sequence of the German debate suggests two core results. First, EU tax governance has become more politicised in German public debates during the past three decades. Although it can certainly not be considered to be a top priority of the German public, the data show that the topic has become more salient and polarised among an expanding set of actors.
Second, the timing, intensity and specific content of the debate appear to support the integrated explanatory framework. The interplay between authority transfer and Germany's position as a potential loser of tax competition is reflected in the timing of politicisation. Because the continuing integration of the European market and the respective authority transfer in neighbouring issue areas makes Germany more vulnerable to external economic pressures, demand for more European tax governance emerges. Socio-cultural concerns are of limited relevance in explaining the timing of German debates. Yet content and framing suggest that the German politicisation process is not exclusively structured by economic preferences. It is also shaped by socio-cultural conditions. Economic-cosmopolitan justifications dominate the German discussion during the entire period of investigation. Most nationalist contributions fall short of a strong nationalist backlash and are mainly focused on the questions of German competitiveness and a level playing field in taxation. Economic concerns about the potentially negative consequence of tax competition are rarely formulated in strong nationalist terms, asking for instance for less European economic integration to protect the German tax system from European competition. Rather, these considerations are predominantly channelled into demand for more European tax governance. Instead of unfolding in a nationalist backlash against European integration, Germans opt -in a cosmopolitan fashion -for more and tighter European tax integration.
These results suggest that economic concerns play an important role for the timing of politicisation. The content and framing of the debates are shaped by socio-cultural context. The discussed solutions to economic pressures display a cosmopolitan bias, which mirrors the relatively weak position of exclusive nationalists and the comparable strength of cosmopolitans. Instead of strong nationalist criticism and demand for less European integration to protect the German economic model, German speakers support further integration of income taxation, VAT and excise duties. Frequently, these statements are justified in a cosmopolitan rather than a nationalist manner. Speakers from centre-left and centre-right parties and civil society but notably not from populist right-wing parties or the economy are often more concerned with the viability of the European integration project than with the effects of tax competition on the German economy.

Ireland: rebellion of the Celtic Tiger
Ireland is also strongly affected by authority transfer to the EU. Both formal authority and tax legislation grow over time and limit Irish national sovereignty to an ever-increasing extent. The growth of formal authority was marked by strong increases during the first two periods of treaty revision and less tremendous changes during the later revision periods of Amsterdam, Nice and Lisbon. The opt-out of the Schengen acquis in the context of the Amsterdam Treaty and some exceptions in the field of criminal and domestic security cooperation integrated in the Lisbon Treaty contributed to a slightly weaker growth of overall formal authority than in the German case. EU formal authority in taxation remains constant. The development of EU tax legislation follows a continuous trajectory and exerts effects similar to the case of Germany. The interaction of formal authority and tax legislation should contribute to patterns of politicisation similar to the German case.
How do Ireland's economic situation and socio-cultural conditions shape this effect? Tax competition among European states began to accelerate during the 1990s. Ireland's low corporate tax regime is frequently considered a poster child of this development. Profiting from European markets and funds, improved education and its low corporate tax regime, the Irish economy experienced unprecedented success during the 1990s, becoming known as the 'Celtic Tiger' (Barry et al. 2001). Consequently, economically structured politicisation should emerge as soon as EU tax governance targets the Irish corporate tax system and it should be framed quite critically, defending the country's growth model.
However, the results of the German case study suggest that the framing of economic concerns in a more nationalist or a more cosmopolitan fashion is shaped by socio-cultural context. The lower left-hand plot in Figure 2 shows that exclusive nationalists dominate (see also O'Kelly 2004). Despite some fluctuations during the 1980s and 1990s, the share of exclusive nationalists is substantial, whereas the share of Irish cosmopolitans is marginal over the entire   investigation period. Populist right-wing parties are absent in Ireland. These socio-cultural conditions suggest that Irish public debates should focus on the economic consequences of EU tax governance. Demands and evaluations are likely to be justified in more nationalist terms than in Germany.
The quantitative indicators appear to support these expectations. Between 1981 and 1998 the level of politicisation does not correspond to growing EU authority. Although the strong growth of overall formal authority and tax legislation took place in the context of a relatively strong presence of exclusive nationalists, the Irish public showed little interest, because the economic consequences of EU tax governance during this period were marginal. Up until 1996, salience, polarisation and actor expansion were very low. Even the enactment of the Maastricht Treaty, which had a substantial impact on the German debate, and the respective Irish referendum in 1992, did not spark politicisation. It took until the mid-1990s for Irish journalists, politicians and representatives of the financial industry to discover EU tax governance and until the enactment of the Treaty of Amsterdam for the debate to become more politicised among an expanding set of actors, including campaigners against the Lisbon Treaty and prominent intellectuals.
A more qualitative reading of the material reveals that Irish debates focused primarily on income taxation. The major bone of contention is to be found in Ireland's preferential corporate tax rate of 10 per cent for manufacturing and manufacturing services. Established as a temporary measure to stimulate the Republic's industry in 1980, this policy became a cornerstone of Ireland's economic growth model. It was a key to the Republic's competitive position in the globalising economy and was thus 'deemed politically beyond challenge' by large parts of the population (Hardiman 2000: 828). During the late 1990s the Irish corporate tax system came under increasing scrutiny from European institutions. Referring to the newly established code of conduct for corporate taxation and European regulations on state aid, in 1998, the Commission urged Ireland to bring its corporate tax regime into conformity with EU regulations. This expansion of EU authority to corporate taxation marked the beginning of what De Wilde and Lord (2015) call an international conflict between Irish governments, European institutions and high-tax countries like Germany. It constituted a strong conflict with Irish economic preferences and set the tone for the politicisation of EU tax governance: Attempts to harmonise corporate taxation were regarded as an affront to the Irish nation (Allen 2000: 84) and proposals to increase EU formal authority in taxation or to abolish the national veto on tax matters in the EU Council were frequently attacked by a broad coalition of Irish politicians and civil society representatives. In sum, the Irish debate is characterised by an economic-nationalist backlash, which is not only justified in economic terms but which is also shaped by more cultural concerns about national sovereignty and self-determination.
The timing, intensity and content of Irish public debate lend further support to the integrated explanatory framework. The interplay of formal authority transfer in neighbouring policy areas, the expansion of EU tax governance to corporate taxation and Ireland's position as a winner of corporate tax competition explains why politicisation has taken off almost one decade later than in Germany. While growing EU formal authority and tax policy output faced little politicisation during the 1980s, the expansion of EU tax governance to the economically sensitive issue of corporate taxation clearly constitutes the trigger for politicisation in Ireland. The nationalist potential lay dormant as long as EU tax legislation did not affect issues of crucial economic importance, only to unfold rapidly when an issue widely recognised as being of national interest was targeted. Neither authority transfer nor national context alone can explain this trajectory of the Irish debate.
As regards the content and framing of politicisation, the data suggest that the distribution of exclusive nationalists and cosmopolitans has an effect. Due to the dominance of nationalists, the share of nationalist justifications in Irish debates is almost 20 percentage points higher than in Germany. Consequently, the interplay of authority transfer and economic concerns results in an economic-nationalist backlash.

Switzerland: the critical outsider
Switzerland is the least Europeanised country in the sample. While both EU formal authority and tax legislation grow over time and limit Swiss sovereignty,   this effect is substantially weaker than in Germany and Ireland. Formal authority transfer was marked by a long period of stagnation between 1957 and 1992 and a remarkable but still comparatively weak growth with the enactment of Bilateral I and II during the early 2000s (Vahl and Grolimund 2006). The expansion of EU tax legislation to Switzerland is directly linked to these bilateral treaties. It is limited to a few but important legislative acts in the area of direct taxation and tax fraud. The interaction of formal authority transfer and tax legislation should result in a mild politicisation during the early 2000s. The highest levels should be observable around 2005, when Bilateral II, which includes the majority of EU tax legislation, was enacted.
How is the authority transfer effect shaped by the Swiss national context? Similar to Ireland, the Swiss economic model is characterised by low corporate tax rates. In addition, individual income tax rates are low and, due to its well-developed financial service sector, including strict banking secrecy, Switzerland was highly successful in attracting foreign capital (Emmenegger 2014). This economic context should result in politicisation as soon as the EU attempts to interfere with these elements of the Swiss economic model. The Swiss socio-cultural context suggests that this politicisation should be framed in strong nationalist terms. Although the share of cosmopolitans is substantial and even slightly higher than the share of exclusive nationalists, the strong presence of populist right-wing parties SVP (Schweizerische Volkspartei; Swiss People's Party), EDU (Eidgenössische Demokratische Union; Union of Federal Democrats), LdT (Lega die Ticinesi; Ticino League), SD (Schweizer Demokraten, Swiss Democrats), FPS (Freiheitspartei der Schweizer, Freedom Party) implies that exclusive nationalists are highly mobilised and thus likely to engage in politicisation (see Figure 3).
The quantitative indicators suggest that these expectations are only partially met: While the data indicate growing politicisation with the enactment of Bilateral I, the expected intensification of public debate after 2005 cannot be observed. In addition, the relatively steady debate on EU tax governance before Bilateral I is striking. During this period, EU formal authority is marginal and tax legislation does not directly affect Switzerland. A more qualitative reading of the material shows that the early years of the investigation period (1994)(1995)(1996)(1997)(1998)(1999) were characterised by non-specific news coverage, which mostly assumed the perspective of an outside observer. The majority of articles described potential implications of the Maastricht Treaty and of the EMU for European tax governance. These articles frequently reflected the content of controversial debates within European member states. Swiss actors rarely contribute their own evaluations or demands. However, those few statements by Swiss journalists which take a position had a quite critical or even fearful tone. They postulated that Swiss tax sovereignty might fall victim to the efforts of European high-tax countries to harmonise taxation within the EU and even beyond its borders. Despite these few critical voices, the unexpectedly strong politicisation during this period resulted mainly from an observer perspective, which described conflict between EU members, not from the controversial examination of EU tax governance by the Swiss public.
This picture has changed since 2000. Salience and polarisation grew and the set of actors involved in the debate expanded. The reason for this politicisation of EU tax governance lies in the events in Feira, where the European Council was able to resolve a conflict on individual income taxation. To enable the taxation of non-resident savings, EU member states agreed to exchange information on cross-border interest payments, on the condition that equivalent measures were adopted by six non-member states (Cattoir 2007: 7-10). The inclusion of Switzerland in this group set off alarm bells among Swiss executives, the press and bankers. It triggered a politicisation strongly shaped by economic-nationalist concerns. A broad national coalition argued that the EU savings directive must not be allowed to restrict Swiss banking secrecy. Between 2001 and 2004 the Swiss debate was clearly dominated by the controversy over the savings directive and the defence of Swiss banking secrecy. Evaluations and demands were mainly framed in economic-nationalist terms and cosmopolitan arguments were marginalised, despite the relatively strong presence of cosmopolitans. Some speakers even claimed that banking secrecy and the commitment to tax competition -as a means to prevent the bloating of political institutions -are not only economic assets but essential components of Swiss national identity.
The timing, level and content of Swiss public debates lend further support to the integrated explanatory framework: The overall level of politicisation is lower than in Germany and Ireland, where European formal authority and the frequency of European tax legislation are much higher than in Switzerland. Although the quantitative indicators suggest that politicisation levels are quite similar across countries, a qualitative analysis shows that a large share of the Swiss debate follows the trajectory of what De Wilde and Lord (2015) call a foreign conflict.
The timing of politicisation in the context of the savings directive indicates the expected conditioning effect of economic context on authority transfer. Similar to the Irish case, politicisation was triggered by the interplay of authority transfer and Switzerland's position as a winner in tax competition. Both cases demonstrate that politicisation took off as soon as EU tax governance expanded to areas of wide economic concern. This conclusion is underlined by the fact that the overarching context of the enactment of Bilateral I and negotiations on Bilateral II -i.e. the formal transfer of authority -was rarely mentioned in the debate. Similarly, Swiss participation in the directive on taxation of interest and royalty payments between companies and cooperation with European institutions in cases of VAT fraud -both negotiated in the context of Bilateral II -received no attention in the Swiss media. The effect of the interaction of authority transfer and national economic context, triggering politicisation, is further demonstrated by the development of the Swiss debate between 2004 and 2011. After years of negotiations on the savings directiveaccompanied by a heated public debate -the Swiss government achieved an important breakthrough in June 2003 by having a withholding tax instead of information exchange formally accepted by the EU. This translated quickly into less politicisation as soon as information exchange and respective restrictions of banking secrecy were off the table.
As regards content and framing, the data suggest the expected effect of the presence and mobilisation of nationalists and cosmopolitans. While the early years of Swiss discussions mirrored the economic-cosmopolitan justifications of the German debate, the controversy on the savings directive was dominated by nationalist concerns about Swiss banking secrecy as a matter of economic considerations and national identity. This distribution of justifications is shaped by the strong mobilisation of exclusive nationalists and populist right-wing parties. In addition, the Swiss case seems to suggest that socio-cultural concerns do not only shape the content of politicisation but also affect its timing and level. The almost 20 percentage point higher share of cultural frames in the Swiss debate compared to Germany and Ireland and the politicisation of the savings directive suggest that this extension of EU competencies was not only considered a threat to the Swiss economic model but also to national identity. Clearly, banking secrecy is an important element of the Swiss financial industry's success (Palan 2002). However, the Swiss debate was also shaped by a cultural-nationalist framing, focusing on national sovereignty and identity. As many Swiss consider banking secrecy a part of their national identity (Guex 2000), the savings directive not only interfered with the national growth model but also with national identity to trigger a strong nationalist backlash.

Conclusion
The empirical findings largely support the explanatory framework suggested in the introduction. The main claim to be tested in this contribution pertains to the conditioning effects of national economic and socio-cultural contexts on the politicisation of authority transfer. While the political economy literature suggests that authority transfer becomes politicised due to its economic implications, more society-centred approaches claim that socio-cultural concerns -i.e. exclusive nationalism and cosmopolitanism -shape country-specific patterns of politicisation. The combination of these three perspectives helps us understand why, when and how EU tax governance has become politicised over the past decades.
The results of all three case studies show that the recipe for a strong politicisation of European tax governance consists of formal authority transfer in neighbouring issue areas, creating public concerns about taxation, broad and frequent EU tax legislation, which targets economically sensitive areas and a high share of cosmopolitans and exclusive nationalists in a given population. Formal authority transfer especially in the area of monetary policy in the case of Germany and industrial policy and state aid in the case of Ireland created expectations and concerns in the field of taxation and shaped discursive spillover effects, driving the politicisation of EU tax governance. This effect is intertwined with EU tax legislation. While much of the current literature on the politicisation of EU governance focuses exclusively on formal authority transfer, my analysis suggests that national mass media do not only react to major treaty revisions but that politicisation is also affected by EU legislative output. Even in issue areas like taxation, in which EU formal authority is still low, we may observe politicisation driven by the interaction of formal authority transfer in neighbouring issue areas and tax legislation under specific conditions. As the cases of Ireland and Switzerland suggest, one important condition for this kind of politicisation is that EU tax legislation targets areas which are of pronounced national economic or cultural importance.
The growing political authority of European institutions has different implications for country-specific patterns of politicisation. Although formal authority transfer and EU tax legislation have similar effects on the German and Irish tax systems and to some extent also on Switzerland, the varying timing, level and content of politicisation in the three countries can only be understood with a view to national economic and socio-cultural contexts. Consequently, further research on the implications of politicisation needs to take into account its differentiated nature. Only by analysing the country-specific patterns of politicisation will future research be able to shed light on how politicisation affects national governments and European institutions.