Keywords

Introduction

This book answers a puzzle which has emerged within the European Union (EU) in the twenty-first century: the EU’s importance has unequivocally grown for decision-making, both in politics and policies following a decade of concatenated crises, from the Eurozone in 2009, the refugee crisis in 2013, Brexit in 2016 to the covid crisis in 2019. At the same time, accountability mechanisms at the EU level have remained largely unchanged since the signature of the Treaty of Lisbon in 2007, even though support for the integration process has varied and currently seems to have largely returned to pre-crises levels. Most of the studies who attempt to understand the way in which the EU is being legitimised focus on the European Parliament elections. Indeed, the issue of EU accountability has led to a large literature around the concept of second-order elections, and how the preferences about the EU are channelled through these Europe-wide elections (Gattermann et al., 2021; Hix & Marsh, 2011; Reif & Schmitt, 1980).

Instead, we argue that to fully understand how EU accountability works, it is necessary to focus on the national political environment and national elections. This channel of accountability has been established (de Vries, 2007), but is still poorly understood. In this book, our goal is to determine the EU’s national channel of accountability in two steps: first, through an extensive analysis of politicisation of the EU in media and parliamentary debates, to show the media and party context which provides cues for voters to formulate preferences about the EU. Second, establishing the importance of the EU for voting in legislative elections through both experimental and observational analysis, and examining, in detail, the contexts which condition it.

There have been a number of studies which focus on EU politicisation, as well as its importance for EU issue voting. Yet, we believe our study is timely due to two new developments in the nature of the EU that arguably warrant more focus on this topic. First, the onset of the Eurozone crisis has been seen as an important moment for EU politicisation at the domestic level (De Vries, 2018; Hoeglinger, 2016; Hooghe & Marks, 2018; Hutter & Kriesi, 2019; Ruiz-Rufino & Alonso, 2017; Schäfer & Gross, 2020), and it has given rise to a large debate on the consequences of “Europe” for electoral behaviour (De Vries & Hobolt, 2016; Lobo & Lewis-Beck, 2012, Lobo & Pannico, 2020). Second, the nature of the EU has evolved from a regulatory to a distributional power (Börzel & Risse, 2018). Whereas previously it could be argued that the pursuit of electoral legitimacy by the EU was unnecessary, since it consisted of a regulatory power, the monetary union and the Eurozone crisis have made clear that the decisions taken at the EU level actually have important distributional consequences, not only between member-states, but also within countries (Börzel, 2016). Both these trends intersect with national politics, more precisely through the media, parliamentary, and electoral arenas.

Our goals determined our choice of country cases: namely Belgium, Germany, Greece, Ireland, Portugal, and Spain. These countries were selected on the basis of their different trajectories within the EU both in political and economic terms. Whereas Belgium and Germany are founding members, the others joined at later moments. Economically, the countries differ in performance and underwent different paths following the onset of the Eurozone crisis, with Germany and Belgium performing relatively well and the other countries having to implement bailouts.

The findings have implications for the ongoing debates on EU democracy, since they underline the importance that national institutions—media, parliamentary debates, and national elections—have in enabling citizens to hold politicians accountable for their EU positions. This suggests that when discussing EU quality of democracy and how it can be improved, it is necessary to take into account national political institutions.

The Book’s Argument

The EU’s importance is undeniable in member-states’ policy-making. From the euro to agriculture, from competition to the single-market, European-level decision-making influences the design of member-states’ policies. The reality of the EU is everywhere, and it has gained in importance in the last decades, both in level and in scope (Börzel, 2005), as a result of the successive crises (Genschel & Jachtenfuchs, 2016), and may even have undergone a paradigmatic shift in policies and processes in the health and economic domains following the covid-19 crisis (Schmidt, 2020).

Concerning policy-making, the Maastricht Treaty and the implementation of European Monetary Union was a turning point for European integration, as it opened the way for further deepening. The introduction of the euro as well as free movement of citizens made it clear that the EU had become a part of citizens’ everyday life. Hooghe and Marks’ (2009) seminal contribution posited that, since Maastricht, the EU had shifted from a “permissive consensus” to a “constraining dissensus”, where citizens and their views on Europe had to be taken into account when considering further steps in European integration.

In the last 15 years, marked by successive crises, the EU has become more political, as its institutions have often intervened in a discretionary way to respond to the multiple challenges that have arisen (Middelaar, 2019). During the Eurozone crisis, in the response to covid-19, as well as following Russia’s invasion of Ukraine, the EU has responded in a purposeful way, rather than simply following rules and norms set out by the Treaties.

Yet, the deepening of EU integration through the effective supranationalisation of policies and politics has not been accompanied by a change in the Treaties, which would formally increase the level of accountability of the EU, as would perhaps have been expected by neo-functionalists and federalists alike. This, therefore, leads us to pose the following question: Where are then, the deepened channels of accountability to accompany the renewed importance in decision-making which the EU has acquired? While many scholars continue to monitor the functioning of EP elections to detect a Europeanisation of political behaviour at that level, this book argues that accountability is occurring at the national level. It follows this argument through an examination of politicisation at the media and parliamentary levels, and then on the importance of EU issue voting at the national level and how it is shaped by the informational and party contexts.

The Book’s Contribution

Given that the successive crises have increased EU politicisation, and that these have asymmetric distributional consequences, it is necessary, beyond establishing its existence, to understand whether EU accountability works similarly in countries at opposite ends of these consequences. In this study, we characterise EU politicisation in a multidimensional fashion. Then we establish EU issue voting and the factors which shape it, focusing on the relation of EU issue voting with left–right, the media and the parliamentary debates context in each of the countries included in the study.

Our multidimensional and multimethod analysis of politicisation across two arenas (media and parliamentary debates) shows that politicisation is present in them, with the share of negative tone in media being considerably higher than in parliamentary debates. Also, in both arenas, the main focus of EU articles and debates is “policies” rather than the “polity” dimension. These findings are common to all countries, and suggest that both arenas are contributing to enable citizens to form opinions on the EU. Yet, differences emerge in the types of policies being politicised in each country both in the media and parliamentary debates, suggesting EU differentiation between countries. Namely, in some countries domesticated EU policies are being discussed, rather than EU policies strictu sensu or any institutional EU issues.

Turning to our voting behaviour analysis we find that European citizens do take into account their attitudes towards the EU when making electoral choices. This finding is robust to different methods, namely using experimental and observational data. Moreover, EU issue voting is asymmetric, with parties that are more Eurosceptic than the respondent being punished on average more than parties that are more pro-EU. While the importance of EU issue voting is established in very clear terms, and we find it does cross-cut the left–right issue, it is not replacing the left–right issue in the vote calculus.

As regards the contexts which shape EU issue voting, we find that the media plays a role in shaping the relative electoral chances of pro-EU and Eurosceptic parties, with low salience and high negativity contributing to undoing the asymmetry detected, on average, in EU issue voting. Regarding parliamentary debates, they are a source of information especially for citizens who vote for Eurosceptic parties. This suggests that in a post-crisis context, the politicisation in the media and parliamentary debates contributes to the strengthening of EU issue voting, thus underlining the national channels of European accountability which exist in Europe today.

We turn now, in this introduction to the book to a brief survey of the concept of EU politicisation and how it was measured both for media and parliamentary debates. Next, we explain the two main channels of EU accountability, at the EU and national level, and how we have focused on the characterisation of national channels of EU accountability in this book. After describing the various datasets which were collected in preparation of the book, the final section outlines the book chapters.

EU Politicisation

Politicisation refers to “an increase in polarization of opinions, interests or values and the extent to which they are publicly advanced towards the process of policy formulation” (de Wilde, 2011, p. 559). Scholars of EU politicisation have focused on how contestation over regional integration connects to domestic conflict, and on how this contestation influences the speed and direction of regional integration (Hooghe & Marks, 2009). By stating that since Maastricht, public opinion had evolved from a “permissive consensus” to a “constraining dissensus”, the post-funcionalist theory, proposed by Hooghe and Marks (2009), was fundamental to turn the focus on public opinion towards the EU.

Measuring EU politicisation has not been systematically done in similar ways, and thus, the evidence which has been accumulating is not totally comparable (Hutter & Kriesi, 2019; Rauh, 2015; Risse, 2015). Studies differ in the way they measure politicisation, in the way they collect the data, and in the forum where the data is collected. More concretely, some studies focus mostly on salience, while others include not only salience but also polarisation (Lobo & Karremans, 2018). Yet, there is an issue of whether it would have been important to include actors too (Hutter & Kriesi, 2019). Namely, for any issue to be politicised, it needs to be salient, it has to be divisive, and it needs to be divisive among a large number of political actors.

Considering media studies, research is accumulating in favour of an increase in EU politicisation, even though there are some sceptics (De Wilde & Zürn, 2012; Green-Pedersen, 2012; Hutter & Grande, 2014; Kriesi et al., 2013; Statham & Trenz, 2013). Green-Pedersen (2012) concludes that there has been little EU politicisation, by looking at both media and party programmes in Denmark. Similarly, Hoeglinger argues that the EU politicisation has been limited (2016, p. 146). On the other hand, other studies find that EU politicisation has been increasing (Hutter & Grande, 2014; Statham & Trenz, 2013, p. 169).

In their comprehensive comparative media study, Hutter and Kriesi (2019) find evidence that the Eurozone and refugees crises contributed to the party-in-the-media politicisation of the EU on a regional basis. These authors do not find a linear increase in EU politicisation. On the contrary, they point to a punctuated politicisation, which increases both due to anticipated (e.g. EP elections) or unexpected (e.g. crisis) events. More recently, Silva et al. (2022), using automated text analyses, show that the Eurozone crisis increased the media politicisation of the EU, in Greece, Portugal, Spain, and Ireland vis-à-vis Germany and Belgium.

Other literature on politicisation has moved beyond the salience and polarisation issue. Silva et al. (2022) highlight that the crisis brought a greater emphasis by the media to supranational institutions vis-à-vis intergovernmental ones and an increasingly negative tone of articles. Hurrelmann et al. (2015) distinguish between EU dimensions in all arenas. Namely, membership—costs and benefits of being in the EU; constitutional structure—institutions, decision-making processes; EU policy issues—European policies on the agenda; and domesticated issues—national issues deriving from one’s own country membership. Whereas the first two dimensions pertain to polity, the last two concern policy issues. This differentiation highlights the need to evaluate if the EU polity itself is being contested, or whether only policies which would suggest a Europeanisation of political debates.

In relation to EU politicisation in parliamentary debates, there has been considerably fewer studies. When considering parliaments, the literature on EU politicisation has focused mainly on parliaments’ legislative and oversight roles, rather than its communicative function. In order to study the latter, it is necessary to concentrate on plenary debates (Auel & Raunio, 2014a, 2014b). Compared to media, parliamentary debates offer an unmediated access to party positions on issues. Certainly, there are issues of comparison, as parliamentary speeches are constrained by different institutional rules across countries. Still, they remain an important and relatively untapped resource to understand party positions.

The existing studies point to divergences concerning the importance of parliamentary debates. On the one hand, it has been perceived that the EU has been politicised in the parliamentary arena (Hurrelmann et al., 2015). On the other hand, some studies have found that there is a lack of debates on the EU (Rauh & Wilde, 2018) and that the level of politicisation has been generally low (Auel & Raunio, 2014b). Nevertheless, the effect of the crisis seems to have contributed to an increase in EU politicisation, which tends to depend on periods around important EU events. Several studies have pointed to the fact that, beyond institutional differences between parliaments that may affect comparability, the party system and the existence of Eurosceptic parties in Parliament are the most important predictors of EU politicisation in plenary debates (Auel & Raunio, 2014a, 2014b; Wendler, 2016).

In this book, politicisation was measured both in automated fashion, to measure salience and polarisation, as well as through manual content analysis to probe into the dimensions of EU topics discussed in the media and parliamentary debates. Each of the measurements is explained in detail in Chapters 2 and 3, and here we detail how these measures were conceptualised and how they complement each other to form a complex perspective of EU politicisation.

Concerning the measurement of EU politicisation in the media, the focus was not, as previous studies had done, on political actors’ statements covered by the media (e.g. Hutter & Kriesi, 2019; Hutter et al., 2016). Instead, the book considers all articles from the relevant (political, economic, and business) sections of each newspaper 30 days before each election date. As in Schmidtke (2016), EU salience is operationalised as the percentage of articles, given our sample, that were about the EU topic, based on a list of EU terms. When a newspaper article contained more than one EU-related term, or if one term appeared in the title, the article was coded as being about the EU. The same principle applied to parliamentary speeches, using the same list of terms. Yet, for parliamentary speeches, the data corpus included all years from 2002 to 2017.

The concept of polarisation or contestation in the media was examined by looking at the proportion of articles, in a year/newspaper, with a negative tone towards the EU. This was done using sentiment analysis after translating all the corpus to the English language, using the English language sentiment dictionary. The same method was applied to the parliamentary debates. These measures of salience and tone were then used to compare trends in politicisation across arenas in Chapter 4, as independent exogenous variables to explain EU issue voting in Chapters 5 and 6, and as media and party contexts to the in-depth case studies covered in Chapters 7–12.

Further, manual coding was employed to probe deeper the EU dimensions which were being politicised in the media, following Hurrelmann et al.’s (2015) typology. Native language researchers were recruited to code, for each article about the EU, if the EU issue mainly dealt with in the article concerned (1) membership, (2) constitutional structure, (3) EU policies, or (4) domesticated policies as explained above, as well as the general topic of the policies discussed. The same frame and method was applied to a sample of parliamentary debates held in the year prior to each national election, to understand the nature of the EU topics which are being debated.

Our analysis of media and parliamentary debates politicisation shows that it is present in both arenas, and increased on average following the crisis in the media but not to the same extent in parliamentary debates. Second, when we distinguish between politicisation in news articles vs. op-eds in the media, differences in EU stances between left and right in the mainstream media can be observed, and these increased following 2009. While politicisation increased in all countries’ media following the Eurozone crisis (especially if we consider op-eds), it did so focusing on different EU topics. In Spain, domesticated policies have always been the most salient dimension in articles about the EU, with Portugal following its neighbour, from 2011 onwards. Ireland stands out since, in all of the four elections analysed, the EU was mostly debated in terms of EU policies, until 2016 where the membership dimension became dominant. Finally, in the Greek case, we find the only instance where, at the peak of the crisis, the membership dimension became the most important dimension. Finally, Germany tends to discuss European policies to a greater extent than the other countries in the sample.

When comparing the two arenas, we find that policy-related issues are, in general, more addressed than polity-related ones, both in the media and parliamentary debates, which is positive from the perspective of EU accountability. Within policies, it is economic- and financial-related matters (“Economy and Work” and “Finances and Taxes”) which dominate the European debate, both in the media and parliamentary debates.

A major important difference emerges between parliaments and media, however. Namely, the share of negative tone in media is considerably higher than in parliaments, and the Eurozone crisis reinforced that trend. This multidimensional and multimethod analysis of politicisation across two arenas (media and parliamentary debates) suggests the following: politicisation is present, and therefore fosters EU accountability, despite working differently depending on the arena. Yet, when we deepen the analysis, differences emerge in the dimensions of the EU being politicised in each country, suggesting EU differentiation, with Germany debating more EU policies than Southern European countries.

Holding the EU Accountable?

The chain of responsiveness—what occurs when the democratic process induces the government to form and implement policies that the citizens want—is perhaps one of the most important elements of the quality of democracy (Powell, 2004). Powell established the main elements of this chain, which include the establishment of voters’ preferences, the polarisation of parties on key issues, the ability of voters to select parties based on different policies, as well as the ability of politicians to implement the preferred policies. In each step of this causal chain, there are numerous factors which may break or strengthen it. While it is exceedingly difficult to verify how the chain of responsiveness is working, Powell argues that vertical accountability, i.e. the ability to hold politicians responsible for their preferred policies at elections, plays a big role in ensuring the chain’s strength and thus the quality of democracy.

Indeed, traditionally, the main focus of studies which considered EU chain of responsiveness and quality of democracy, have focused on vertical accountability, processed through European Parliament (EP) elections. Many studies have criticised the EU for lacking democratic legitimacy (Follesdal & Hix, 2006; Hix, 2013). At the centre of the “EU democratic deficit” critique has been the second-order nature of the European Parliament elections. As the only EU institution which is directly elected, scholars have established that these elections are not really about European issues, but rather national ones (Reif & Schmitt, 1980). The concept of European elections being second-order follows from a number of principles about the EU. While EP competences in decision-making have increased substantially, there is still no direct link between voting for the EP and the formation of the European Commission, which is largely still decided by national governments. Also, despite the increased powers of the EP, the European Commission and Council still wield dominant power, and the successive crises since 2009 have only served to reinforce this trend. Due to the nature of the electoral system, at present, voters in each country select national parties to represent them. As such, they tend to consider the national political context and use the EP election to punish or reward the national incumbent and mainstream parties. Following from these principles, second-order elections such as the EP elections are supposed to have higher abstention levels, higher votes for extreme or protest parties, and fewer votes for incumbent or mainstream parties, vis-à-vis the national legislative elections (Schmitt, 2005). Indeed, a lot of research on European Parliament elections has found that these remain second-order national elections (Nielsen et al., 2017).

Nevertheless, the European Parliament elections in 2019 did not follow strictly the expectations of a second-order election (Gattermann et al., 2021). For the first time since the 2004 enlargement, EP elections saw an increase in turnout, which surpassed the symbolic 50% threshold. Yet, overall, in terms of individual vote choice, Ehin and Talving (2021) have shown that the second-order model is still supported for those elections.

Thus, by and large, EP elections continue to be second-order, and this makes it even more obvious that we need to consider the way European issues are processed through national channels of accountability during national elections. Indeed, EU accountability is a multi-level system of governance that enjoys dual electoral legitimacy. On the one hand, citizens can express preferences through a direct vote at European Parliament elections. On the other hand, in national legislative elections, they also select representatives who express EU preferences that are congruent with their own (De Vries & Hobolt, 2016). These electoral choices, at the national level, then contribute to the formation of governments who make decisions at EU-level bodies, namely the Council of Ministers and the European Council (Schneider, 2019). This dual legitimacy system can only work, however, if we establish the existence of EU issue voting at the national level. Beyond establishing its importance, in a comparative perspective, this book explores the national contexts which foster or discourage the expression of EU preferences at the national ballot box. Thus, the chain of responsiveness and ensuing quality of EU democracy will depend on the existence of EU issue voting at the national level.

When considering EU issue voting, it has been studied at the national level mainly in three ways. Namely, identifying its existence, explaining its determinants, and, to a lesser extent, examining its consequences. De Vries (2007) was the first to establish the evidence of EU issue voting at the national level, while more recently, Hobolt and De Vries (2016) confirmed its importance in national as well as in EP elections, using observational data collected after the European Parliament elections in 2009. Further, there is research on the determinants of EU issue voting, and in particular the media environment. The degree to which EU-related information is available is also important to determine EU issue voting (De Vries et al., 2011) and dependent on the degree of EU politicisation (Hobolt & Spoon, 2012). Previous studies have focused on the characteristics of the political competition on EU issues, namely when the EU is salient and the parties have different positions on the process of European integration (De Vries, 2007).

There is an emerging debate on the degree to which European issue voting relates to other factors of individual voting behaviour. Indeed, most of the research concerning the importance of European integration for national politics have focused on the party system and its changes (Hooghe et al., 2002; Hutter & Kriesi, 2019; Marks & Steenbergen, 2002). Hooghe et al. (2002) were path-breaking in explaining the way Europe relates to the left–right dimension, through the introduction of the concept of GAL-TAN. GAL parties are in favour of European integration, have a cosmopolitan outlook, and also share pro-Green policies. The TAN category aggregates those parties who are Eurosceptic, have a nationalist outlook, and share traditional values. TAN parties tend to be radical right parties, and this pole of the GAL-TAN axis was responsible for the politicisation of EU integration. These authors posited that the GAL-TAN axis was orthogonal to the left–right axis, and that the party systems tended to form a curvilinear pattern, with parties on the extremes of the left and of the right tending to be more Eurosceptic, and those on the centre tending to be more pro-EU.

Other authors which have focused on party systems and voting choices, place parties’ positioning on the EU, as well as citizens’ attitudes, as part of a “globalization cleavage” (Kriesi et al., 2008), which distinguishes between citizens who tend to be pro-EU, pro-migration, and cosmopolitan, vs. those who are Eurosceptic, anti-immigration, and nationalist (Bornschier, 2010; Dalton, 2018; Norris & Inglehart, 2019). In certain elections, notably in France, this cleavage has gained prominence in relation to the left–right dimension in structuring vote choices (Lachat & Michel, 2020; Schön-Quinlivan, 2017). The eurozone crisis also leads to studies which analyse the importance of the EU issue for voting in Southern Europe. Extant research shows that challenger parties continue to be the ones which politicise the EU issue to a greater extent and, therefore, the EU issue may still cross-cut the left–right axis in these countries (Santana & Rama, 2018). Others state that there has not been an increase in the importance of EU for voting, since austerity has served to reinforce the left–right dimension, rather than replace it (Hooghe & Marks, 2018; Otjes & Katsanidou, 2017).

In this book, we show that European citizens do take into account their attitudes towards the EU when making electoral choices. Through an experimental analysis (Chapter 5) fielded in our six countries, we are able to state with confidence that the proximity between parties’ and voters’ positions on the EU is the cause, rather than the consequence, of respondents’ electoral choice. Further, we show that EU issue voting is asymmetric, with parties that are more Eurosceptic than the respondent being punished on average more than parties that are more pro-EU.

In addition, through observational data collected in Greece, Ireland, Portugal, and Spain, we confirm that EU issue voting cross-cuts the left–right issue, and in every country analysed (with the exception of Ireland), left–right proximity has a higher impact on likelihood to vote for a party than proximity on EU issues. Even when we distinguish between mainstream and challenger parties, left–right is still more important than EU issue proximity for the vote. In our country case studies, where EU issue voting is assessed in comparison to other political issues in each of the six countries, we find that EU issue voting tends to matter more for right wing challenger parties, but in certain countries, namely Portugal and Greece, it works principally on the left, and also serves to distinguish between the two principal mainstream parties.

Concerning the way media and parliamentary debates impact the importance of the proximity between parties’ and voters’ positions on the EU for the vote, a distinction was made between the EU salience and tone. As regards the media context, in the experimental study we find there is a connection between the availability of EU information and the strength of EU issues voting, while the tone of EU news appears less relevant. Turning to the impact of parliamentary context, we found that, for parties with a more negative tone—or Eurosceptic discourse—EU proximity matters more in determining the voters’ choice than for parties with a positive or neutral EU tone in parliament.

Data Collection

Our substantial data collection underpins and strengthens our research goals and scope. The final media dataset includes 165,341 print newspaper articles, of two mainstream newspapers per country, of all relevant articles published one month before each national legislative election between 2002 and 2017. Each of these articles were also manually coded to consider the EU topics discussed. In addition, online media in each of the mainstream newspapers was scraped one month before each 2019–2021 webpanel (115,876 articles).

As regards parliamentary debates, the dataset includes all plenary debates in the six countries (2002–2019) (Kartalis & Lobo, 2021). All speeches “about the EU” in the year preceding the legislative elections were identified in automated fashion. In addition, a representative sample of these have been manually coded on EU content. A total of 10,514 speeches were randomly selected from all speeches delivered in parliament in the year preceding each of the 26 national legislative elections that took place in our countries of choice.

Further to the parliamentary and media data collection, 22 representative online surveys were carried out in the six countries concerned. We ran one simultaneous two-wave panel in January and June 2019, which functions as an EP election panel, and where the conjoint experiment was embedded. Next, we ran panel surveys when national legislative elections were held (Greece, July 2019; Portugal, October 2019; Spain, April 2019, Ireland, January 2020, Germany, September 2021).Footnote 1 This array of different types of data strengthens our research findings.

We now turn to outlining in detail the book chapters, which are organised in two parts. In the first part, the importance of the EU for mainstream media and parliamentary debates in Belgium, Germany, Greece, Ireland, Portugal, and Spain is established. Thus, the first part includes three chapters devoted to setting up the longitudinal trends of EU politicisation in each country. In the second part, there are eight chapters, delving on the importance of the EU for electoral behaviour in 2019/2020 national elections in these countries, both from a comparative and single case-study perspective.

The Book Chapters

The first three chapters chart the level of EU politicisation in the media and parliamentary arenas, from a longitudinal perspective and across the six case studies. Besides revealing temporal and spatial variance in how the EU penetrated the national arenas, they investigate the factors that account for such variation.

In Chapter 2, Tiago Silva and Yani Kartalis measure EU politicisation in mainstream media, between 2002 and 2017, in Belgium, Germany, Greece, Ireland, Portugal, and Spain. The focus is on the distinction between news and opinion articles, as well as on the dimensions of EU in the media. The chapter shows that, after the crisis, the EU media politicisation diverged more in tone than salience, even if these two indicators are quite correlated. When news and opinion articles are compared, it is clear that the latter diverge more across newspapers in the same country. When considering the EU dimensions portrayed in the media, while all countries have increasingly focused on policies, Spanish and Portuguese media focus on domesticated policies, whereas the rest of the countries tend to debate European policies to a larger degree.

In Chapter 3, by Yani Kartalis and Tiago Silva, automated measures of salience and tone are employed to examine EU politicisation in parliamentary debates. The Chapter also analyses the four distinct dimensions of EU, and the extent to which these are discussed in each Parliament by different categories of parties. The findings show that the EU has not been very politicised in national parliaments, and this did not change considerably with the Eurozone crisis. As regards the determinants of parliamentary politicisation, the EU position of parties was the strongest predictor of EU politicisation, being statistically significant for both EU salience and contestation. Also, in the parliamentary speeches of the six countries concerned, the EU was likelier to be contested by the parties on the left and the parties not in the government.

Chapter 4 complements the previous analysis of EU politicisation through a comparison of trends in media and parliamentary debates, between 2002 and 2017. While acknowledging the inherent differences between the two arenas, Nelson Santos and Susana Rogeiro Nina compare salience and tone within EU dimensions, as well as the policies which are discussed. Politicisation is present in both arenas, albeit negative tone is much more prevalent in the media than in parliaments. In both arenas, (economic) policy-related issues are predominant, which may foster EU accountability. A more fine-grained dimensional analysis shows differences between countries, suggesting EU differentiation, with Germany debating more EU policies than Southern European countries.

The second section of the book is composed of two comparative chapters and six single case studies. It establishes EU issue voting, and develops fully the analysis of the contexts which may help or hinder the expression of EU preferences by citizens, thus holding national politicians accountable.

In Chapter 5, Roberto Pannico and Marina Costa Lobo approach the issue of the importance of the EU for voting in an experimental setting. This model enables a truly exogenous estimation of the importance of EU attitudes for voting in the six countries. The results confirm that EU preferences are the cause not the consequence of electoral choices. It also shows that EU issue voting is asymmetric, with parties that are more Eurosceptic than respondents being on average more punished than parties which are more pro-EU than the respondent. However, this asymmetry only seems to hold in contexts where EU information is more available and positive.

In the following chapter, Lea Heyne, Marina Costa Lobo, and Roberto Pannico make use of post-election surveys carried out in 2019–2020 in Greece, Ireland, Portugal, and Spain, to consider the importance of the EU vis-à-vis left–right issue proximity on the likelihood to vote for different parties. Our results show that, in all countries surveyed, EU issue voting does cross-cut the left–right dimension. Yet, left–right proximity has a higher impact on likelihood to vote for a party than EU issue proximity, in all countries, both for mainstream and challenger parties. The interaction of parliamentary tone employed by parties in EU debates and EU proximity shows that, for parties which have a more negative tone, EU proximity determines vote choice to a larger extent than for other parties.

Having established EU issue voting and its determinants, in a comparative perspective, using experimental and observational data, the rest of the book is composed of six single case studies. In each of the country chapters, media, parliamentary, and survey data are combined to explain EU politicisation and EU issue voting in comparison to competing relevant political issues.

In Chapter 7, Dieter Stiers examines the case of Belgium, which is, as he demonstrates, a singular case on different levels. As the author explains, the legislative and European elections in Belgium are held simultaneously, and voting is mandatory. Combined, these factors make it unlikely to find EU issue voting in Belgium since voters can express their European preferences on the same day using the EP ballot, and a lot of voters show very little interest in politics. Further, the data on media shows that, overall, the European issue is not very salient in Belgium, since most parties share a positive consensus towards the EU. Yet, EU issue voting is detected at the national level. The author distinguishes between Wallonia and Flanders, as well as between electors with different levels of political sophistication to nuance the results.

In Chapter 8, Marc Debus and Rosa Navarrete explain the way in which key economic issues, as well as immigration, compare to EU issue voting in Germany. In this case, there is a sort of paradox: while Germany is decisively important for decision-making processes in the EU, the German media and parliamentary debates tend not to politicise the EU, with the exception of the (AfD—Alternative für Deutschland). The authors find that parties behave strategically regarding the EU in the plenary debates. There is also evidence of EU issue voting in the last federal elections, and the authors further show how that it interacts with individuals’ perceptions on immigration policy.

In Chapter 9, Roula Nezi examines the case of Greece which is the country in our sample where EU salience and polarisation increased most dramatically, both in the media and in parliamentary debates, following the onset of the Eurozone crisis and the successive bailouts which the country endured. The chapter analyses the importance of EU issue voting in two models, one where that variable is included alongside ideological self-placement and economic perceptions, and another contrasting the EU issue with two other relevant political issues. Those relevant issues were the Prespa agreement which had been signed by the Tsipras government to resolve the “Macedonia” question and proved highly controversial, as well as the asylum law, with both issues serving to differentiate between the two main parties—Syriza and ND (Nea Demokratia). Results show that the EU issue, as well as perceptions of the economy, are the two most important issues differentiating the vote between Syriza and all other parties in the party system, with those holding Eurosceptic perceptions voting for Syriza, and pro-EU attitudes reinforcing the vote for the centre-right party of ND.

The Irish case is analysed by Lea Heyne in Chapter 10. She begins by highlighting the importance of the 2020 Dáil elections, where, for the first time ever, Sinn Féin was the most voted party. The analysis of EU politicisation shows that it has increased in both media and parliamentary debates in the last decades, and the media has become more negative since the Eurozone crisis and Brexit. In parliamentary debates no such negativity is found, with parties adopting a more positive tone. The multivariate analysis of electoral behaviour shows that there is no strong impact of EU issue voting, with the exception of voting for Sinn Fein.

In the Portuguese case, covered in Chapter 11, Marina Costa Lobo compares EU issue voting to socio-economic attitudes during the 2019 legislative elections. As in the Irish case, there are increasing trends of EU issue politicisation from 2002 to 2019, both in the media and parliamentary debates, particularly noticeable after the onset of the Eurozone crisis in 2009. Yet, in a post-bailout stage, and especially since the left coalition government took office in 2015, a degree of depoliticisation of the EU issue is detected in parliament. Concerning EU issue voting the data show that it continues to be relevant among Communist voters, but also for the mainstream centre-right PSD (Partido Social Democrata), signalling that in Portugal it matters not only for voters of challenger but also mainstream parties.

In Chapter 12, Hugo Marcos-Marne examines the politicisation of the EU in media and by different political parties in Spain as well as the dynamics of EU issue voting in the May 2019 legislative elections. Results show that there was an increase in the salience of EU issues after 2008 that subsides after 2012, and an unclear pattern concerning tone, even if, overall, it seems that salience has not been accompanied by more contestation of the EU. Concerning EU issue voting, it predicts only votes for the radical-right VOX both in general and within-ideological-blocks comparisons.

In the Conclusion, we examine the evidence provided in favour of holding Europe accountable at the national level. The comparison of media and parliamentary debates trends longitudinally make clear that politicisation is higher in the former than the latter. In the media, there is overall growth in salience and tone turning negative. In Parliaments, parties tend to behave strategically, depending mainly on their EU stance. As regards, the consequences of this politicisation for voting behaviour, it is demonstrated through the experimental analysis as well as observationally. Furthermore, both the EU media as well as parliamentary politicisation appear to be associated with the strength of EU voting. Indeed, both voters of mainstream and challenger parties use the vote to express EU preferences in national legislative elections.

Our analysis has consequences for the debates about EU legitimacy today. Most of the efforts for democratising Europe involve supranational reforms, at the level of the European Parliament, or citizens’ initiatives at the EU level. Yet, the important vertical accountability which is occurring in national legislative elections regarding the EU is being largely ignored. The national channel of EU accountability, and namely the domestic institutions which contribute to it, must become a central part of the debate on EU legitimacy, for this issue to be properly addressed.