Euroscepticism in 2019 : from the numbers to the roots : a quantitative and qualitative analysis of Euroscepticism in EU member states in 2019

(1)

Euroscepticism in 2019: From the numbers to the roots.

A quantitative and qualitative analysis of Euroscepticism in EU member states in 2019.

Master Thesis

Author:

BSc. MN Phernambucq

Student number:

10560602

E-mail:

matthijsphernambucq@hotmail.com

University of Amsterdam

MSc Human Geography (Political Geography)

Lecturers:

dr. V.D. Mamadouh

dr. Ir. C.J.M. Karsten

Amsterdam, August 12 2019.

(2)

ACKNOWLEDGEMENTS

The past 7 months have been a true rollercoaster for me. The primary goal during these months was

writing my master’s thesis as the final chapter of, on the one hand, the master’s degree Human

Geography with the sidetrack Political Geography, and, on the other hand, my University career.

Little did I realise at the start of the thesis project what I was heading into and which distractions

would be on my way. I would like to express my genuine appreciation for those who helped my

through these tumultuous times.

First of all, I would like to thank my supervisor Virginie Mamadouh. Her guidance through the project

was beyond useful. Not only did she give me plentiful feedback during the thesis project and did she

help me through tough times, she was also thinking along with my personal situation and able to

adapt to my, admittedly often changeable, planning.

Secondly, I would like to thank Inge van der Welle for inspiring me during the tough start of this

project. Starting a project from zero has always been difficult for me and this time it was no

different. She was the one who referred to the European Parliamentary elections which gave me the

first inspiration and the starting point for this research.

Finally, I would like to thank the people closest to me. You have been of great help and without your

support I would never have been able to finish my thesis.

Thank you,

Matthijs Phernambucq

Gent

(3)

ABSTRACT

The aim of this research is to gain better insight in the causes, the extent, the characteristics and the

dynamics of Euroscepticism in the EU. This research covers a time frame of 5 years, from 2014 until

the latest European Parliamentary elections from the 23

rd

to the 26

th

of May 2019. The unexpected

result of the Brexit referendum of 2016 is the clearest indication that there is a poignant need for

better understanding of Euroscepticism in the EU. This research aims to do its share in reducing this

lack of understanding.

Based on the analysis of three open data surveys and the results of the European Parliamentary

elections of 2014 and 2019, this research presents in-depth information about different aspects of

Euroscepticism. Furthermore, the data is analysed with a geographical point-of-view, giving

much-needed insight in the spatial dynamics of Euroscepticism.

(4)

INTRODUCTION

This chapter outlines a motivation for this research and explains the academic and societal needs for

this research. The objectives of this research are elaborated and this chapter will introduce the

research methods of this research.

MOTIVATION

The winning leave-vote at the Brexit referendum on the 23

rd

of June 2016 managed to shock the

European Union (hereafter EU) and the rest of the world (BBC, 2016; BBC, 2016). The morning after

the Brexit referendum, newspapers around the world mentioned a future of insecurity, doubt,

possibilities of future referenda and crashing financial markets. Despite the importance of the

referendum, the results of the referendum were unpredicted by both elections polls and financial

markets (Economist, 2016). This surprise caused major shocks to financial markets, international

relations, and created yet another crisis for the EU. Since the economical crisis in 2008, the news

coverage around the EU seems to be constantly related to crises (Taggart & Szczerbiak, 2018). The

economical crisis in 2008 pushed the EU to change their usual and comfortable regulating role for a

more proactive, redistributional role (Scipioni, 2017). For the first time since the foundation of the

EU and its predecessors at the end of the Second World War, the EU played a major role in EU

citizen’s everyday life. Obviously, becoming more important and influential naturally comes with

increased support as well as resistance (van Spanje, & de Vreese, 2011, Taggart & Szczerbiak, 2018).

The results of the increased importance and the foreshadowing of the EU’s next major crisis, the

migrant crisis in 2015, were clearly visible in the 2014 European Parliamentary elections (hereafter

EP elections). The results of these elections are specifically notable for the rise of Eurosceptic

political parties in comparison to the 2009 EP elections (Taggart & Szczerbiak, 2018; Thielemann,

2017; Dennison & Geddis, 2018; De Sio et al., 2014).

In the light of the growth of Euroscepticism and the continuous crises drawing a disturbing narrative

around the EU, the leave-vote in the Brexit referendum in 2016 could have been seen as a continuity

of a trend rather than a shocking surprise. In fact, it is rather concerning that academics nor election

polls nor financial markets managed to predict the referendum results. The surprising referendum

results have called for, beyond other things, genuine questions about our understanding of the EU

and their citizens. The continuous flow of crises that the EU is trying to solve, rather strikingly shows

that the EU is acting constantly one step behind events. It is solving its own problems rather than

acting before these problems evolve. One way to predict future problems before they occur is a

better understanding of which thoughts live inside EU citizens. More specifically, signalising and

understanding Eurosceptic thoughts is crucial in order to prevent future crises like the Brexit

referendum. Only by understanding these thoughts, the EU is able to act before Eurosceptic citizens

do. This research aims to use the 2019 EP elections to dig into the problem of understanding

Euroscepticism under EU citizens.

OBJECTIVES

The increased interest in the EU and the increased Euroscepticism that came with it pushed the

importance of Euroscepticism as an academic concept (Taggart & Szczerbiak, 2018). The research on

this concept differs from discovering the roots of Euroscepticism to the importance of the EP

(7)

elections for the EU citizens. The academic concept of Euroscepticism is being shaped, but the Brexit

referendum showed the poignant lack of knowledge and understanding. The new results of the 2019

EP elections provide a valuable opportunity to conduct new analyses on Euroscepticism. Not only

will extra research improve the understanding of the quantity of Euroscepticism and help predict the

results of possible future referenda better, it will also create in-depth understanding of

Euroscepticism. This research will use the opportunity that the EP elections of 2019 provide, to join

in the academic discussion around Euroscepticism. It will not only aim to research how many EU

citizens are Eurosceptic, but also discover who these citizens are, what they are sceptic to, and

where they live. Only with profound understanding of the EU citizens, the EU will be able to listen

and answer to the, apparently, increased scepticism towards the EU. The concept of Euroscepticism

needs to be elaborated further and faster than before to release the EU from its narrative around

crises and to answer to the calls of EU citizens for change.

This research will aim to investigate three aspects of Euroscepticism. First of all, possible predictors

for Euroscepticism will be elaborated and tested. This part aims to discover the characteristics of

those with a Eurosceptic vision. It aims to find distinctive characteristics of EU citizens who are more

likely to be Eurosceptic than others. Secondly, using a combination of election results and social

surveys, this research aims to discover how Euroscepticism has developed within the EU and into

which extend Euroscepticism in the European Parliament represents the feelings of EU citizens.

Thirdly, this research aims to discover the geographical characteristics of Euroscepticism. The

following side-questions will be answered:

-

Which indicators can be linked with individual Euroscepticism?

-

Does Euroscepticism in the European Parliament represent Euroscepticism in EU member

states?

-

What are the geographical characteristics of Euroscepticism in the EU and how has

Euroscepticism developed since 2014?

With analysing these side-questions, I aim to answer the main research question of this research:

“How has Euroscepticism developed in the EU between 2014 and 2019 and how can this be

explained?”

The research will consist of different types of data. Quantitative analyses based on open data will

give insight in the indicators of Euroscepticism, which is needed for the first sub-question and

contains geographical information that can be used for the geographical analyses for sub-question

three. The results of the EP elections from 2014 and 2019 will be used, together with the open data,

to answer the second sub-question. Lastly, literature on Euroscepticism will be used to create an

overview of all European Eurosceptic political parties.

In order to optimize the validity of the comparing analysis between the electorate of the European

Parliamentary elections and the respondents of the social surveys used in the analyses, some groups

have to be left out of the analysis. The respondents of the surveys should match the electorate of

the elections. In other words, all respondents that are not allowed to vote during the European

Parliamentary elections should be removed from the surveys. All respondents younger than 18 years

old will be removed from the analyses. In addition, all respondents from outside de EU will be

(8)

removed as well. Since the analyses will be based on continental- and national scale and the

elections are nationally organized as well, Great Britain and Germany are considered as one country.

RESEARCH OUTLINE

The research will start with an extensive elaboration of the existing theoretical framework

concerning Euroscepticism and the EP elections. This framework includes relevant theories, findings

and aspects that could bias the analysis, especially that of the EP election results. This bias could for

instance include the role of the political parties in the national government, the status of the EP

elections for the EU citizens, in other words, the motives of voters to vote, and the voting conditions

in the EU member states. Secondly, the methodology will explain how this research will aim to

answer the side-questions. As stated above, this research consists of three parts that do not only

differ in research questions, but also in methodology and even the type of data involved. Based on

the theoretical framework and the methodology, an extensive conceptual scheme will be presented

which will be used as the visual representation of the set-up of the analysis. This conceptual scheme

can be considered a detailed elaboration of the expected causalities and results. Afterwards, the

results will be given which will be followed by the conclusion and discussion. In the conclusion, I will

aim to answer the side-questions and eventually the main research questions. The discussion will

reflect on the role of this research on the existing academic literature and will focus on the societal

and academic relevance of the findings of this thesis. Depending on the findings of the analysis,

certain recommendations will be expressed.

(9)

THEORETICAL FRAMEWORK

This chapter will cover the existing academic literature on the concepts in this research. The two most important concepts are Euroscepticism and the EP elections. This research deals into a great extent with the concept of Euroscepticism. All current academic discussions around this concept are relevant and thus will be elaborated here. Firstly, this chapter will discuss the history and the rise of Euroscepticism as an academic concept. Afterwards, this chapter will make a distinction between individual Euroscepticism and party-based Euroscepticism. The first part goes deeper into the roots and causes of Euroscepticism while the second part deals with the way in which this Euroscepticism expresses itself in political parties.

Lastly, the concept of the European Parliament elections will be discussed. This paragraph deals specifically with the trends and habits that influence the electorate during the EP elections. This part is particularly important for this research for it expresses which external factors could influence the findings later on.

EUROSCEPTICISM

This section introduces the main academic concept of this research, Euroscepticism. This includes the meaning of Euroscepticism, an overview of the main academic findings concerning Euroscepticism and the evolution of Euroscepticism as an academic concept. Further on, the analysis of Euroscepticism will shift towards the way in which Euroscepticism is translated into politics, and, more specifically, the European Parliamentary elections.

THE CONCEPT EUROSCEPTICISM

While, recently, the reasoning behind Euroscepticism on an individual basis is a widely researched theme, little has been written about the exact meaning of Euroscepticism (van Spanje & de Vreese, 2011). Taggart and Szczerbiak (2018) aimed to determine which political parties in de EU member states can be considered Eurosceptic. As research method, they asked experts to determine Eurosceptical political parties. This method implies that there is always a level of subjectivity concerning the image around these political parties that determines whether they are Eurosceptical or not. In their research, the question was not so much a question of whether these parties are Eurosceptical as it was about the question whether they tend to talk positive or negative about the EU. As Krouwel and Abts (2007) state: “Euroscepticism as commonly referred to, however, may mean not only scepticism but also cynicism or outright opposition” (Krouwel and Abts, 2007: 259). Taggart and Szczerbiak (2004) refer to Euroscepticism as a ‘flurry’ academic concept. The interpretation of Euroscepticism seems to change between countries and scholars. Existing literature around Euroscepticism consist of both qualitative and quantitative Euroscepticism with little comparative analyses between member states (Taggart & Szczerbiak, 2004). In their attempt to create a definition of the concept of Euroscepticism, Taggart and Szczerbiak created a distinction between hard- and soft Euroscepticism. Hard Euroscepticism is referred to as outright negativity towards the concept of the EU as a whole, whereas soft Euroscepticism collects feelings of critisicm towards the EU, but not disliking the concept as a whole (Taggart & Szczerbiak, 2004, Taggart & Szczerbiak 2018). With their distinction between hard- and soft Euroscepticism, they aim to solve an academic discussion that keeps returning, even in more recent works. Krouwel and Abts (2007) and van Spane & de Vreese (2011) refer to the same discussion and state that a0lthough often misinterpreted, being Eurosceptical does not per definition imply a negative opinion towards the EU. Being sceptical could imply a certain amount of distrust, but not disliking the concept as a whole. Besides, the EU consists of many aspects someone could be sceptical to. Lubbers and Scheepers (2005) have made the distinction between instrumental Euroscepticism and political Euroscepticism. The first one implying doubts about whether membership of the EU is beneficial for the member state, and the latter implying a stance where political power should be returned to the national level instead of the EU (van Spanje & de Vreese, 2011; Lubbers & Scheepers, 2005). Wessels (2007) and Krouwel and Abts (2007) base their analysis on Easton’s political support

(10)

classification (Wessels, 2007; Krouwel and Abts, 2007; Easton, 1975; van Spanje & de Vreese, 2011). In this classification, different aspects of the EU have been distinguished which include the EU’s effectiveness and responsiveness as well as support for the political regime and support for membership (van Spanje & de Vreese, 2011). Wessels (2007) further includes the dimension of EU identity whereas Krouwel and Abts focus on the image of the EU (van Spanje & de Vreese, 2011; Wessels, 2007; Krouwel & Abts, 2007).

The current academic theories around the concept Euroscepticism show that there is not only a wide set of distinctions of the concept Euroscepticism and streams within Euroscepticism, it also shows a wide variety of possible predictors for Euroscepticism, but also that there is a broad distinction of issues that these

Eurosceptic EU citizens could be sceptic to.

In an attempt to get deeper insight into the flurry concept of Euroscepticism, the next section focuses on the history of the concept.

THE EVOLUTION OF EUROSCEPTICISM

The academic history of Euroscepticism as a concept is rather short. After the Second World War, European integration has been an elite-driven project about which citizens had ‘non-attitudes’; they had no strong opinion, neither positive nor negative (van Spanje & de Vreese, 2011; Converse 1964; Taggart & Szczerbiak, 2018). Recently, however, the interests of EU citizens in the EU have increased, which, as a result, creates stronger attitudes (van Spanje & de Vreese, 2011). The EU has managed to become a major theme in national politics, which naturally comes with resistance. As a subject becomes more important, so does the resistance towards it increase as well. This new wave of interest and the rise of disagreement towards the EU project are referred to as ‘constraining dissensus’, a decline of consensus under the electorate (van Spanje & de Vreese, 2011; Hooghe & Marks, 2009) or even a new social cleavage; a politicised subject that divides the electorate (De Sio et al., 2014). The EU has become a major political subject that divides the electorate not only during the EP elections, but also during national elections (De Sio et al., 2014). According to Caporaso et al. (2015), the rise of interest and disagreement in the EU is stirred up by the economical crisis back in 2008. Before this crisis, the EU played a regulating role, which usually kept them on the background. The economical crisis, however, pushed the EU to act in a redistributional role. Both the EU institutions and the EU citizens were not used to this new role (Scipioni, 2017; Caporaso et al., 2015). The EU institutions were structurally, financially and legally incapable to fulfil this proactive, redistributional role in the extent to which was requested from them (Scipioni, 2017; Caporaso, 2015; Berger and Heinemann, 2016; European Parliament, 2016; European Parliament, 2016). Besides, it is unlikely that the EU citizens are currently willing to provide the EU institutions with the financial means and political power that these institutions need for this role (Scipioni, 2017). The evolution of Euroscepticism did not stop after the economical crisis. In contrary, since the economical crisis in 2008, the news coverage around the EU seems to be constantly related to crises (Taggart & Szczerbiak, 2018). The migrant crisis in 2015 and the Brexit referendum in 2016 are crises that still dominate the news. Apart from European referenda and national elections, this growth of interest in the EU and Euroscepticism is especially visible in the EP elections (van Spanje & de Vreese, 2011). Euroscepticism in the EU has grown rapidly since the economical crisis, and the EP elections of 2014 have shown an exponential growth of Euroscepticism in comparison with the 2009 EP elections (Taggart & Szczerbiak, 2018; Thielemann, 2017; Dennison & Geddis, 2018; De Sio et al., 2014). This implies that Euroscepticism is not only a feeling, but is also exploited by political parties. The next part section will focus on the characteristics of Euroscepticism as an idea, here referred to as individual Euroscepticism, whereas the following part will analyze how this ideology has found its way into politics, referred to as party-based Euroscepticism.

(11)

INDIVIDUAL EUROSCEPTICISM

This part will discuss the academic discussion around Euroscepticism on an individual basis. The previous sections on Euroscepticism have already explained the rise of individual and party-based Euroscepticism in the EU since the Economic crisis. In this part, the academic analysis focuses on people who have become

Eurosceptic. This section includes a description of the indicators that have been linked to Euroscepticism by scholars.

PREDICTORS OF EUROSCEPTICISM

With the growth of interest in the EU, scholars started to investigate the predictors that can be linked with Eurosceptic attitudes. Hobolt (2009) has found that EU attitudes, political awareness, government satisfaction, subjective social class, age and party recommendations are predictors for Eurosceptic ideas (van Spanje & de Vreese, 2011; Hobolt, 2009). Schuck and de Vreese (2008) found in their analysis that a negative attitude towards the EU is the main predictor for Euroscepticism (van Spanje & de Vreese, 2011; Schuck & de Vreese, 2008). In addition, de Vreese & Tobiasen (2007) also mention the perception of the EU identity as a predictor for Euroscepticism. A strong perception of feeling European results in a more positive vote towards the EU. On the contrary, a low level of trust in the national government, negative attitudes towards immigrants and extremist ideological ideas result in a more negative vote towards the EU (de Vreese & Tobiase, 2007; van Spanje & de Vreese, 2011). Besides, more recent works have focussed on the link between crises and Euroscepticism (Taggart & Szczerbiak, 2018). Euroscepticism increased as a result of the Economical crisis, whereas signs have been found that the Brexit vote and the problematic procedure that follows, have created uncertainty under Eurosceptic citizens, which improved pro-EU attitudes (Taggart & Szczerbiak, 2018; Gomez, 2015; de Vries, 2017). In the meantime, links have been found between the migrant crisis and increased Euroscepticism, taking in mind the vast geographical differences within the EU (Taggart & Szczerbiak, 2018). In order to answer the first side-question of this analysis, “Which indicators can be linked with individual Euroscepticism?”, I have decided to test the predictors that have been found in previous analyses as described in this section. This will be further elaborated in the methodology.

PARTY-BASED EUROSCEPTICISM

The way in which Euroscepticism evolves and presents itself in political parties is a broad academic discussion on itself. Taggart & Szczerbiak note three distinctive features of Euroscepticism. Firstly, although

Euroscepticism is often referred to as an (extreme) right wing concept, for example by De Sio et al. (2014), Euroscepticism does not necessarily imply a certain political orientation. Euroscepticism has been found on all sides of the political spectrum (Taggart & Szczerbiak, 2018). Secondly, Euroscepticism can be expressed in a variety of ways (Taggart & Szczerbiak, 2018). Thirdly, of all EU member states, during the EP elections of 2014, only Lithuania and Malta did not have a Eurosceptical political party (Taggart & Szczerbiak, 2018). The

widespread presence of Eurosceptical political parties in EU member states can be seen as a sign of strong opposition towards the EU. However, Taggart & Szczerbiak (2018) conclude that this presence of party-based Euroscepticism is a natural response to the normalisation of European politics in both national as international politics. Party-based Euroscepticism, in turn, has broadened the social cleavage. The electorate only becomes politically divided once political parties exploit a subject (De Sio et al., 2014).

DETERMINING EUROSCEPTIC PARTIES

As the first two features of Euroscepticism, as stated by Taggart and Szczerbiak (2018), explain, is the distinction of Euroscepticism in political parties not a straightforward process. Especially the fact that

(12)

Euroscepticism is expressed in a variety of ways makes it difficult to determine which political parties can be considered Eurosceptic. There are two schools of research on Euroscepticism in European political parties; the North Carolina School and the Sussex School (Mudde, 2012). The North Carolina School’s theoretical

foundation is the cleavage theory (Mudde, 2012). This theory splits up political parties on their position towards European integration. This split of parties determines which parties can be considered Eurosceptic parties. On the one hand, the green, alternative and liberal parties are considered as pro-European, while on the other hand, the traditional, authoritarian and nationalistic parties are considered Eurosceptic parties (Mudde, 2012). With this, rather simplistic, foundation, only extremist parties show signs of Euroscepticism, which contradicts with Taggart & Sczcerbiak’s findings that Euroscepticism is expressed in a variety of ways and is found all over the political spectrum (Mudde, 2012; Taggart & Szczerbiak, 2018). Besides, this school considers Euroscepticism as a concept that ignores the complex concept of scepticism and misinterprets it as direct opposition towards European integration. Only hard Euroscepticism has been analyzed. Lastly, apart from the discussion about the meaning of scepticism, Euroscepticism goes beyond the party’s position towards European integration. European integration in one aspect of the EU and Euroscepticism does not stop there. A political party that is not content with the democracy in the EU, but supports further European integration is Eurosceptical.

Taggart & Szczerbiak’s ideas are the base of the Sussex School to determine Euroscepticism (Mudde, 2012). This school does make a distinction between hard and soft Euroscepticism. This distinction, although it might be more accurate than the North Carolina School’s theory, makes it difficult to frame political parties. They even find themselves in difficulty to distinguish political parties as either hard or soft Eurosceptic, primarily due to the limited information of the parties on European integration, which, in term, could be interpret as

reduced Euroscepticism (Mudde, 2012). At the same time, the North Carolina school asks experts to determine a party’s position towards European integration (Mudde, 2012). This creates a subjective distinction based upon reputation rather than actual position.

The social cleavage from the North Carolina School as described in literature, with the green, alternative and liberal parties on one side and the traditional, authoritarian and nationalist parties on the other, has reduced the complex concept of Euroscepticism in political parties solely to the question whether a party supports or resists further European integration. On the other hand, the Sussex School seems to be a more accurate way of determining Euroscepticism at political parties, but it suffers from the lack of information of these parties in their stance towards the EU which decreases the practicality of this method (Mudde, 2012). All in all, both schools of determining political parties as Eurosceptic parties face major problems and require expert input to be successful. Different methodologies have been found inadequate, mainly due to language limitations. I have decided to use the Sussex School to determine Euroscepticism, mostly due to the fact that analyses with this method are more valid and do not exclude soft Euroscepticism which is a major part of Eurosceptiicsm and therefore this research. The main problem with this school is that expert interpretation is required. In this research, I aim to reduce this problem by using the same distinction of Eurosceptic political parties as presented in the research of Taggart and Szczerbiak (2018). Their research was based on the period 2015 and 2017 and I do realize that copying an expired analysis for the 2019 EP elections will reduce validity of the analysis. However, creating a new distinction for the 2019 EP elections is due to the need of experts and language issues beyond my abilities.

This chapter so far has explained the rise of Eurosceptic ideas and that Euroscepticism has already been exploited by a variety of political parties in the EU. However, assuming that a vote on a political party that shows signs of Euroscepticism implies that the electorate shares the same Eurosceptic feelings is wrong. In order to reduce the bias created by wrong assumptions, I have included a section that lines out the existing theories around the EP elections.

(13)

THE EP ELECTIONS

This paragraph will explain the existing theories around the EP elections and discuss which external factors could influence the results of the Eurosceptic parties in the 2019 EP elections.

THE STRUCTURE OF THE EP ELECTIONS

The European Parliament is, together with the Council of the European Union, the legislative power of the European Union and is the only European institution that is directly elected by the EU electorate (European Parliament, n.d.). The EP elections are held every five years since the creation of the European Parliament in 1979. The elections take four days and are the largest transnational elections at the same time in the world (European Parliament, n.d.). During the EP elections, EU citizens will vote on national parties after which the Members of the European Parliament, often referred to as MEPs, will be elected (European Parliament, n.d.). Once elected, these MEPs can choose to work together with similar minded MEPs from other EU member states and become a member of a European political group (European Parliament, n.d.).

SECOND-ORDER ELECTIONS

The EP elections are by many scholars considered as second-order elections (Reif & Schmitt, 1980; Schmitt & Toygür, 2016; Hobolt & Spoon, 2012; Marsh & Mikhaylov, 2010; Schmitt & Toygür, 2016; De Sio et al., 2014). This theory implies that the EP elections are considered less important by the electorate than national elections. As a result, citizens voting for the EP elections base these votes on national scale issues. Reif and Schmitt were the first to name the Parliamentary elections within the EU and its predecessors second-order elections (Reif & Schmitt, 1980).

The importance of the second-order perception of EU voters is that they tend to change their voting behaviour between national and EP elections (Hobolt & Spoon, 2012; Marsh & Mikhaylov, 2010; Schmitt & Toygür, 2016; De Sio et al., 2014). This is caused by several factors. Firstly, the majority of the EU voters make their decision based upon domestic arguments rather than European issues (De Sio et al., 2014). Secondly, the level of voter turnout is generally lower for EP elections than for national elections (Schmitt & Toygur, 2016; De Sio et al., 2014). Thirdly, nationally ruling parties generally experience a drawback in terms of electoral performance (De Sio et al., 2014). In the 2014 EP elections, only Matteo Renzi’s Democratic Party in Italy managed to increase its results in comparison with the national elections while being in the national government. This is reflected in the EP elections, which results in a high level of votes that are considered as a protest vote; a vote against a ruling party or an event (Hobolt & Spoon, 2012; Schmitt & Toygür, 2016; De Sio et al., 2014). Schmitt & Toygür analysed a possible growth in results for Eurosceptic political parties in the EP elections of 2014 in comparison with the national elections. Their analysis showed no statistically significant proof that Eurosceptic political parties perform better in EP elections. This could imply that Eurosceptic parties are not a target for these protest votes. However, the fact that the second-order theory for the EP elections has been proven in all previous editions of the EP elections means that it will probably play a role on the results of Eurosceptic parties. Fourthly, as Hobolt & Spoon (2012) mention, national political parties behave different in the European field. They have different priorities and perhaps different ideas about European issues than their national scale colleagues (Hobolt & Spoon, 2012).

However, on the contrary, national elections’ votes can also be influenced by European issues (Tillman, 2004; van Spanje & de Vreese, 2011). Tillman (2004) refers to this phenomenon as EU voting. He found evidence that the national political parties’ stance towards European integration should correspond with voters’ stance to increase the likelihood of them voting for that party.

(14)

In more recent years, scholars have aimed to find changes in these theories on Euroscepticism. Schmitt & Toygür (2016) analyzed whether the increased interest and news coverage about the EU, fed by the impact of the financial crisis and the start of the migrant crisis, could have changed the perception of the EP elections as secondary elections. However, Schmitt & Toygür’s research has proven that even after this crisis, the EP elections of 2014 were still considered second-order elections. It is likely that the results of the 2019 EP elections also show signs of second-order elections. This increases the likelihood of protest voting, or, in other words, a vote for a party without sharing the same ideology. This could potentially mean that, in some EU member states, the results of Eurosceptic parties is influenced by the voters’ dissatisfaction with the national government rather than genuine Eurosceptic feelings. In this analysis, the status of the political parties in their national government will be taken into account to offset this possible bias. However, there are strong signs that the EP election results do not accurately represent the Eurosceptic feelings within the matching EU member state. There are simply too many factors that could influence the electorate’s decision making. For this reason, I have decided to not only use EP election results as indicators for Euroscepticism in the EU, this research will also use social surveys of open data to get an indication of the level of Euroscepticism in the electorate. For the second side-question of this research, an analysis will be done into which extent the EP election results represent Eurosceptic feelings in the EU member states.

THE GEOGRAPHY OF EUROSCEPTICISM

The EU consists of 28 member states and all of them have a different position within the EU. As Taggart and Szczerbiak (2018) imply, the three main crises that the EU has faced since 2008 will have influenced the perception of the EU under EU citizens heavily, but with strong geographical differences. Oddly, these

geographical differences are sparsely researched. Scholars tend to focus on one EU member state in particular. Perhaps, Euroscepticism is a concept that is primarily covered by political scientists whose focus is not on geography. As a thesis in the direction of political geography, this study creates the ideal opportunity to create further insight in the geographical differences on different scales. The open data surveys provide information about the urban-rural differences. The combination of the lack of existing research on the role of geographical differences on Euroscepticism and the likelihood of the existence of these differences, and the opportunity to avoid linguistic problems using the translated open data surveys, provides enough reason to implement a third side-question in the analysis: “What are the geographical characteristics of Euroscepticism in the EU and how has it developed since 2014?” Apart from the fact that geographical differences in Euroscepticism are very interesting to analyze, it also helps to answer the main research question: “How has Euroscepticism developed in the EU between 2014 and 2019 and how can this be explained?” Development always has a geographical dimension.

This last part of the analysis will focus on Euroscepticism as a quantitative actor. In-depth analyses of how Euroscepticism is shaped differently between EU member states are definitely interesting and require more research, but it is beyond the scope of this research.

The exact methodology that will be used to answer the research questions, will follow in the next chapter, the methodology.

(15)

METHODOLOGY

This chapter describes the research methods that are used in the analysis to get answers to the research questions. This includes the type of data that is used in the analysis, the gathering methods of this data, the characteristics of this data, the operationalisation of the research question and side-questions, the way in which this data is processed and the consequences of this procession to the liability and the validity of the data. In the end, a conceptual scheme will be presented to clarify the structure of the research.

DATA

For the analysis of Euroscepticism in the EU, two methods will be analysed. First of all, the opinions of EU citizens towards the EU and the drivers of these opinions will be analysed using quantitative research. Secondly, Euroscepticism in the results of the 2014 and 2019 EP elections will be analysed based on the framework provided by Taggart and Szczerbiak.

EUROSCEPTICISM BASED ON OPEN DATA SURVEYS

Two open data sources based on the results of surveys will be used: the Standard Eurobarometer and the European Social Survey (ESS). The European Social Survey (ESS) is a survey funded by the participating EU member states. Its aim is to monitor social change in Europe. The first round of the ESS was conducted in 2002, and it has been repeated every two years. Every round of the social survey consists of two parts. The first part covers general subjects that are repeated in every version of the survey. The second part focuses on one particular subject, usually subjects that are often in the news at the time of the surveys. The seventh round of the ESS was in 2014 and focused on the respondents’ attitudes towards migrants, a hot topic back in 2014 and it still is. The rise of radical right political parties in the EU member states, the high level of labour migration from Eastern European countries to Western European countries and the vast increase of asylum seekers in the EU have made this the right opportunity for the EU to try to understand these issues (ESS, 2016). The extensive surveys of the ESS 7, one single survey took on average almost an hour to complete, have been collected via face-to-face CAPI interviews (ESS, 2016). CAPI interviews are face-to-face interviews with the help of an electronic device such as a laptop or tablet.

The Standard Eurobarometer is a social survey established in 1974 (European Commission, n.y.). With two publications every year since 1974, the European Commission has provided itself a vast amount of data about the public opinion of EU citizens. The questions in the Standard Eurobarometer focus primarily on the respondents’ opinion towards the EU and its institutions. For this reason, it can be seen as an evaluation for the EU. The repetitive nature of the Standard Eurobarometer makes it specifically useful for analysing trends in the opinion of EU citizens. For this analysis, the Standard Eurobarometer 81 and 90 will be used. Both these versions of the Standard Eurobarometer were conducted at the same time as respectively the 2014- and the 2019 EP elections. The Standard Eurobarometer 81 from 2014 misses some relevant immigration related questions that are available in the Standard Eurobarometer 90. I have decided to implement some findings from the ESS7 to make up for these missing questions. The Standard Eurobarometer 81 and the ESS7 were created at the same time, 2014, and the focus on immigration makes the ESS7 particularly interesting for the analysis.

Since the 62nd version of the Standard Eurobarometer, CAPI interviews are used to collect data (GESIS, n.y.). Although the Standard Eurobarometer does not explain its detailed research method itself, only referring to face-to-face surveys, I have been able to find that the Standard Eurobarometer has a contract with Kantar since 2004, a research company using CAPI interviews for the Standard Eurobarometer (Kantar, 2019). This gives me enough information to assume that the ESS7 and the Standard Eurobarometer 81 used the same

(16)

research method. This means that no bias can be expected from the use of two separate surveys due to different research methods.

The databases from the Eurobarometer have been downloaded from GESIS. The permission for these downloads have been granted based on the reason; “for my final exam (e.g. bachelor or master)”. The database from the European Social Survey has been downloaded from their own website with permission due to the analytic nature of the study.

EUROSCEPTICISM BASED ON THE EP ELECTIONS

The second part of the analysis on Euroscepticism in the EU will make use of the 2014 and 2019 EP election results. Taggart and Szczerbiak have used the Sussex School of determining Euroscepticism in political parties to present an overview of Eurosceptic political parties in the EU member states between 2015 and 2017. They used surveys to get information about expert interpretations of the political parties. Mainly due to the lack of information given by political parties in their stance towards the EU, a more objective research method was not possible (Taggart & Szczerbiak, 2018). It is not in my means to create a similar distinction of Euroscepticism at political parties for the 2019 EP elections, partly due to the fact that expert surveys are required to create such an analysis and partly due to language limitations. For this reason, I have decided to use the political party distinction that covers the political atmosphere between 2015 and 2017 also on the 2019 EP elections, with the addition of the obvious Brexit party in the UK elections. The difference between the results of these parties between the 2014 and the 2019 EP elections will give a clear overview of the dynamics of

Euroscepticism. However, I do realise that using old data for analyses of 2019 reduces the validity of this analysis.

CHARACTERISTICS OF THE DATA

This paragraph gives an overview of the characteristics of the data used in the analysis. It also explains how the data has been changed in order to make it more suitable for the analysis.

CHARACTERISTICS OF THE DATABASES

The two databases from the Eurobarometer are extensive surveys with respondents from all EU member states. The ESS7 consists of respondents from 20 EU member states. Besides, ESS7 also implements research data from respondents from Israel whereas Eurobarometer 90 includes data from Turkish Cypriot Community, Montenegro, North Macedonia, Russia and Turkey. In the meantime, both databases from the Eurobarometer also separate Eastern Germany from Western Germany and Great Britain from Northern Ireland. This analysis focuses partly on causal relations between two variables within the EU and partly on differences between EU member states. Therefore, the focus of this research is solely on EU citizens meaning that respondents from outside the EU have been removed from the analysis. The UK has been implemented since, at the moment of writing, they are still part of the European Union. Obviously, their special position within the EU has a big influence on the British respondents in the more recent Eurobarometer 91. The influence of the Brexit will, obviously, be discussed in the analysis, without going too deep into this relatively recent chapter of the EU. The role of the Brexit on the EU citizens is another, very interesting subject to analyse, but is beyond the scope of my research. Yet, I understand that the Brexit vote could bias the findings in this analysis and thus certain precaution has to be implemented. The British respondents as well of those from Germany are considered as one country since the analysis focuses on national scale geographical differences.

All databases have respondents from 15 years or older. ESS7 even has two respondents aged 14. In this analysis, I aim to link the findings from the surveys to the EP elections. For a more accurate analysis, I have decided to include only the respondents that are allowed to vote, aged 18 or higher.

(17)

After removing the respondents from outside the EU and those younger than 18 years old, the number of respondents in the Eurobarometer 81 is n=27327. Since the extra number of respondents from both Germany and the UK will be compensated with the weighting, no more respondents will be removed from the analysis. The Eurobarometer 90 originally consists of n=32600 respondents. After removing respondents from outside the EU and those younger than 18 years old, the database consists of n=26959 respondents. The ESS7 originally has n=40185 respondents which drops to n=36471 respondents after removing Israel and those younger than 18 years from the analysis. Here, no separation between Eastern- and Western Germany has been made. Neither is there a distinction between Great Britain and Northern Ireland.

The amount of variables in the survey of the Eurobarometer 81 is 405. 5 years later, in the Eurobarometer 90, this number has almost doubled to 800. The ESS7 differentiates 601 variables. A large difference between the Eurobarometer 81 and the Eurobarometer 90 is the self-reflective nature of the latter survey. Back in 2014, the Eurobarometer seemed to be mainly used to get an impression of what was going on in the minds of the European citizens while the Eurobaremeter 90 seems to be used as a reflection report of the latest half a year or even an indication of what the EU should decide in the near future in the minds of the respondents. The focus has shifted from towards more contemporary issues, such as migration and fake news. Besides, there are more personal questions such as whether a respondent has internet or a tablet. I am not familiar with the incentives of these new questions. Perhaps they are used as an indication of economic growth or increased welfare. The questions in the ESS7 are more sensitive than those of the Eurobarometer and include detailed questions about the preferred political parties, medical conditions and the level of education of family members.

CHARACTERISTICS OF THE EP ELECTION RESUL TS

The EP elections results are, due to the national scale nature of the elections, a collection of political parties from 28 EU member states. Obviously, this results in a very large number of political parties. After the elections, these parties can form political groups with other parties with similar ideas. In table 2, the overview of party-based Euroscepticism based on the framework provided by Taggart and Szczerbiak (2018), I have added three columns: these political groups, the political orientation of the political parties and whether these parties are part of the national government. The data for these additional columns derives from extensive literature research, especially for the analysis of the political orientation of these political parties. As the table shows, there is no obvious Eurosceptic political group visible. It seems that political parties do not cooperate based upon Eurosceptic ideas. The second column, political orientation, confirms the theory from Taggart & Szczerbiak that Euroscepticism can be found on all sides of the political spectrum. However, Euroscepticism seems to be more of an extremist issue, meaning on the far left and far right of the spectrum, than a centre-issue. The third column, that shows if the party is part of the national government, will be used in the analysis between Euroscepticism in the election results and Euroscepticism in the social surveys. This method builds on the findings of De Sio et al. (2014) who found that ruling national parties generally encounter an electoral drawback in the following elections. This effect also counts for EP elections. Besides, EP elections are, partly due to their second order status, an event where protest voting is used often by EU citizens (Hobolt & Spoon, 2012; Schmitt & Toygür, 2016; De Sio et al., 2014). Voters, who are disappointed in their ruling national parties, use protest votes during the EP elections to express these disappointments. This often results in a relatively high turnout for smaller political parties in the EP elections (Schmitt & Toygür, 2016). With these theories in mind, it is worth considering the national status of the analysed political parties as a possible reason for changing results between the 2015 and the 2019 EP elections.

(18)

RESEARCH METHODS

This section describes how the data is used to create the results of the analyses.

PREDICTORS FOR EUROSCEPTICISM

The first side question of the analysis will focus on the causes of Euroscepticism on individual basis. Both open data sources, the Standard Eurobarometer and the ESS, focus specifically on the respondents’ attitudes towards the EU. For this reason, there are many survey questions that, together, could imply Euroscepticism. Based on the findings in the literature, the dependent variable called “Euroscepticism based on analysis” has been produced that combines a set of variables that are considered indicators for Euroscepticism. This variable is a result of the merged variables “EU Statements: More decisions at EU level”, “EU Statements: Better future outside EU”, “Democracy Satisfaction: European Union”, “EU Meaning: Waste of Money”, “Trust in

Institutions: European Union”, “EU Future: Optimistic/Pessimistic” and “Present Direction: European Union”. With this new dependent variable, I have aim to cover all aspects of Euroscepticism. These include the general opinion of the respondent towards the EU based on trust, his vision towards the EU and whether he or she wants to leave the EU. In the meantime, all aspects of Euroscepticism onto which the respondent could be sceptical such as further integration, the size of the EU and the level of democracy in the EU are included. In order to create this new variable, the data had to be processed. The step-by-step explanations of this data procession can be found in Appendix 1: “Changelog SPSS”.

For the analysis of which factors influence Euroscepticism, a wide variety of independent variables have been selected, also based on the findings and assumptions in existing literature. These independent variables include ‘Age (categorised)’, ‘Subjective Social Class’, ‘Trust in National Government’, ‘Image of the EU’, ‘Attitude towards immigrants’, ‘Political Awareness’, ‘Perception of EU identity’ and ‘Political Orientation’. Some of these variables have been found to influence Euroscepticism by other scholars (political awareness, level of satisfaction with and trust in the national government and subjective social class by Hobolt (2009), EU identity and negativity towards immigrants by de Vreese & Tobiasen (2007)), whereas the three recent crises in the EU are expected to influence Euroscepticism as well. Besides, these crises seem to have increased geographical differences in Euroscepticism (Taggart & Szczerbiak, 2018). For this reason, I have decided to also include geography-related independent variables that determine first of all, whether the respondent lives in the centre or the outskirts of the EU. Secondly, they examine the role of an urban or rural community on Euroscepticism. Thirdly, national scale differences will be analysed as well. Party recommendations and extreme ideological ideas have also been found to influence Euroscepticism, but these variables cannot be tested with the variables provided in the databases and will therefore be left out of the analysis. Yet, the column political orientation of table 2 seems to confirm the link between Euroscepticism and extreme ideological ideas.

Table 1: An overview of the used variables and survey questions from the surveys.

Variables Survey questions

Independent variables:

- Trust in national government Trust in Institutions: National Government - Political awareness EU Informedness: Respondent

Political Interest Index

- Perception of EU identity EU Citizenship: Feel to be EU citizen Attachment to: European Union - Subjective social class Social class: Self assessment

- Age Age 4 categories

- Political orientation Left-right placement - Rural-Urban Type of Community

(19)

- Centre-Outskirts Centre or Outskirts of the EU - Image of the EU Image of the EU: positive/negative - Attitudes towards Immigrants Important Issues EU: Immigration

EU Meaning: Not enough Frontier Control Feeling Towards Immigration: From Outside EU Statements: Immigrants contribute a lot Immigrants make country better/worse place to

live*

Allow many/few immigrants from poorer countries in EU*

Immigrants take jobs away or create new jobs*

Dependent variables:

- Euroscepticism based on analysis EU Statements: More decisions at EU level Democracy Satisfaction: European Union EU Statements: Better Future Outside EU EU Meaning: Waste of Money

Trust in Institutions: European Union EU Future: Optimistic/Pessimistic Present Direction: European Union

European Unification go further or gone too far* *From European Social Survey 7

Source: Author

The influence of these possible predictors on Euroscepticism will be tested with the use of bivariate

correlations. Since both of the variables in the analysis are ordinal, Spearman’s rank-order correlation will be used to indicate the strength and significance of the relationship.

REPRESENTATION OF EUROSCEPTICISM IN THE EUROPEAN PARLIAMENT

The second part of this thesis covers the representation of the Eurosceptic feelings in the European

Parliament. Firstly, the development of party-based Euroscepticism in the EU between 2014 and 2019 will be elaborated. This will be done by expanding the work on party-based Euroscepticism in the EU between 2015 and 2017 as created by Taggart & Szczerbiak (2018). The data of the results of these parties are provided by the European Parliament and will be compared to the previous election results. Secondly, the focus shifts to national scale Euroscepticism. This part will analyse whether the Eurosceptic political parties are

representative of the Euroscepticism as found in the surveys from the same country. This will show whether the rise of Euroscepticism in the European Parliament in 2014 and perhaps again in 2019 is a direct result of the rise of Euroscepticism in the EU member states.

The share of votes on Eurosceptic parties in EU member states during the EP elections will be compared with the share of the respondents from that country that has Eurosceptic feelings.

EUROPEAN ELECTIONS AND NATIONALITY

According to the treaty of Maastricht in 1998, EU citizens are free to move and reside in any EU member state (Marzocchi, 2018). This means that many EU citizens do not live in the country of their nationality. According to Eurostat, 3.8% of the EU citizens of working age live in another EU member state (Eurostat, 2019). This could influence the results of the EP elections. The rules for EU citizens who live abroad are that someone can only vote once and chooses whether to vote in their country of residence or their country of nationality (Council directive 1993/109/EC, 1993). For the results of the EP elections, it should not matter in which country the EU citizens will vote. However, for this analysis, that compares national scale surveys with EP election results from the corresponding country, it is worth taking into account. The surveys themselves also

(20)

distinguish country of residence and country of nationality for the respondents. However, the rest of the analyses in the surveys use the country of residence to separate the EU member states. Since it, to me, seems more likely that the respondents’ votes can be influenced by factors concerning the country of residence than the country of nationality, I have decided to stick to the country of residence as variable for the analysis.

GEOGRAPHICAL ASPECT OF EUROSCEPTICISM

The great increase of Euroscepticism in the European Parliament after the EP elections of 2014 is often linked by scholars to the economic crisis. In addition, Taggart & Szczerbiak (2018) found signs that the migrant crisis of 2015 has a strong link with Euroscepticism, be it with great geographical differences. In the meantime, much is uncertain about the role of the Brexit on Euroscepticism. The one likely true fact about this relationship is that it will influence Euroscepticism differently between EU member states. This last side-question will focus on these geographical differences between EU member states concerning Euroscepticism. The geographical differences will be analysed in three different categories: national scale differences,

differences between the centre of the EU and the outskirts of the EU, and lastly differences between urban and rural neighbourhoods.

Besides the description of percentages of Euroscepticism in EU member states, this analysis will focus on the changes between 2014 and 2019. There have always been differences between EU member states in their vision towards the EU and only differences in time will tell information about the dynamics in these visions. For this chapter, the surveys will be used because they, more accurately than the election results, explain the fundamental reasoning that creates these dynamics. The differences between the constructed dependent Eurosceptic variable in the 2014 and the 2019 surveys will be elaborated for each EU member state. The results will be presented with GIS. The same analysis will be done with the centre-outskirts analysis, using the newly constructed variable CENTROUTSK. In this variable, the EU member states have been separated between ‘Centre’ and ‘Outskirts’ based on their geographical location within the EU. The central countries in this analysis are The Netherlands, Belgium, Luxembourg, Germany, Denmark, Great Britain, Ireland, France, Sweden, Czech Republic, Slovenia and Austria. The other countries, Romania, Bulgaria, Estonia, Latvia, Lithuania, Greece, Malta, Cyprus, Italy, Spain, Portugal, Croatia, Slovakia, Hungary and Finland are considered the outskirts of the EU. Lastly, the differences between urban and rural Euroscepticism will be elaborated using the variable ‘Type of Community’ in both the Eurobarometer 81 and Eurobarometer 90.

LIABILITY AND VALIDITY OF DATABASES AND RESEARCH METHODS

Taggart & Szczerbiak (2018) have created an overview of Eurosceptical parties in the EU member states during the 2014 EP elections with the use of experts. Because it is not in my abilities to conduct a similar analysis for the 2019 EP elections, I have decided to use their analysis for both the 2014 and the 2019 EP elections. I have not been able to analyse the plans and ideas of these parties that are considered Eurosceptic. Therefore, I cannot guarantee that none of these parties have changed their minds towards the EU since 2017. Watching the news taught me that at least the British Eurosceptic Brexit Party has been created since 2017. This party has been added to the framework. However, any additional changes have not been made.

WEIGHTING DATA

Weighting data is used to correct for errors such as sampling errors and non-response errors (ESS, 2014). While doing analysis with surveys, it is important to be sure that the respondents of the survey properly represent the researched population. Every respondent must have had exactly the same chance to be in the sampling group (ESS, 2014). Since this is very difficult to achieve, especially when the researched population is as big as the EU, weighting is used to compensate errors that have occurred during the sampling. The

(21)

Eurobarometer and the European Social Survey present a wide set of weights that can be used by the users, depending on what the population of their research is. There are two different types: post-stratification weights and population size weights. Post-stratification weights are produced after the data gathering. They contain design weights and together they compensate for differences between the respondents and the population caused by non-response errors or sampling errors. The role of underrepresented groups in the survey will be enhanced by higher weights, while overrepresented groups are weighted down. Population size weights are used in analyses which use data from multiple EU member states. Respondents from Luxembourg represent less people than respondents from Germany or France due to the differences in population. For this reason, German respondents are weighed stronger than those from Luxembourg in analyses on this database. National scale differences between EU member states are an important part in this analysis. Therefore, I have decided to use both the post-stratification weights and the population size weights. In the Eurobarometer databases only population size weights will be used, because “By general rule the population size weights in the Eurobarometer include the post-stratification weighting factors as available” (GESIS, n.y.). In the database for the European Social Survey however, the post-stratification weights and the population size weights are given separately in the database. This is problematic for SPSS, since the program can only weigh the data based on one weight at a time. For this reason an additional weighting variable has been produced that combines the post-stratification weights and the population weights into one. The exact method of this procedure is given in Appendix 1, the SPSS change log. For the databases of the Eurobarometer, the EU28 population size weight has been chosen which includes all current EU member states.

POLITICAL ORIENTATION

The political orientation of the Eurosceptic political parties are based on news papers, academic articles and books. Although I have aimed to analyse the parties as accurate as possible, however, labeling a political party with a political orientation is always a matter of interpretation and therefore often disputed. The label will be only used as indicators for unity between Eurosceptic political parties to test Taggart and Szczerbiak’s theory that Euroscepticism is found in all sides of the political spectrum. Considering the wide variety of political orientations found in table 2, it is highly unlikely that the findings of this analysis are wrong due to using the wrong information.

(22)

CONCEPTUAL SCHEME

Possible indicators for Euroscepticism: Euroscepticism:

(Independent variables) (dependent variables)

Source: Author

The conceptual scheme is a visual representation of the expected causal relations in this analysis. It will help to interpret the methodology and the structure of this research. The independent variables are the possible indicators that are expected to influence Euroscepticism in the analysis. The dependent variables are a set of concepts that relate to Euroscepticism and focus on a specific side-question or the main research question. The individual predictors are individual characteristics that are expected to influence the individual’s voting decision. These expected predictors will be based on existing literature on Euroscepticism and will be retrieved from the databases that are used in this analysis.

Last of all, this analysis will cover possible factors that could cause bias in the analysis of the predictors of Euroscepticism. In other words, factors that could influence a respondent’s decision making or situations in

Individual Eurosceptisicm

Party-based Euroscepticism Individual predictors

- Level of trust in national government

- Negativity towards immigrants - Political awareness

- Government satisfaction - Perception of EU identity - Subjective social class - Attitude towards EU - Age

Geographical aspects

- Centre or outskirts of the EU - Nationality

- Living in rural or urban area

Party-based Euroscepticism based on framework of Taggart & Szczerbiak (2018)

EU crises

Possible bias

(moderating and/or mediating) - Status of EP elections - Position of parties in

national government - Voting conditions

(23)

which there are strong indications of a correlation between an indicator and a political party, but where the respondent votes for the party for other reasons than Euroscepticism. For example, the vote is a protest vote against the current national government or the role of the political party’s national scale counterpart has its influence on the respondents. These factors, if they are found, are moderators; variables that influence the correlation between the indicator and the dependent variable either positively or negatively.

RESULTS

The chapter will present the results of the research and answer the side-questions and eventually the main research question: How has Euroscepticism developed in the EU between 2014 and 2019 and how can this be explained?

INDICATORS OF INDIVIDUAL EUROSCEPTICISM

This section will show the results for the first side-question: Which possible indicators can be linked with individual Euroscepticism? The analyses have been conducted with bivariate correlations between, on the one hand, the possible indicator and, on the other hand, the constructed independent variable Euroscepticism (4 categories). Spearman’s rank-order correlation has been used to indicate the strength of the correlation. Table 2: Correlation between Euroscepticism and possible predictors.

Eurobarometer 81 Eurobarometer 90 Trust in National Government .317*** .196***

Political Awareness .154*** .065***

Perception of EU identity .396*** .239*** Subjective Social Class -.130*** -.042***

Age .117*** .100***

Political Orientation -.003 .015*

Rural-Urban -.022*** -.031***

Centre-Outskirts -.105*** -.096***

Image of the EU .628*** .385***

Attitude towards Immigrants .316***(ESS7) .187*** *P ≤ 0.05 **P ≤ 0.01 ***P ≤ 0.001

Source: Author

Of the 10 analysed possible indicators for Euroscepticism, 9 indicators showed significant correlation with Euroscepticism in both 2014 and 2019. Only political orientation does not seem to be a valid indicator for Euroscepticism. Perhaps this is because Euroscepticism is not expected to be influenced by one particular political orientation, but the extent of this orientation; extreme. Also the bivariate colleration between the original, not recoded, variable political orientation and Euroscepticism does not give a significant correlation. The findings of de Vreese and Tobiase (2007) and van Spanje and de Vreese (2011) will not have to be rejected by this analysis, because the results do not prove that there is not a link between political orientation and Euroscepticism either. In fact, it confirms the theory from Taggart and Szczerbiak (2018) which states that Euroscepticism can be found all over the political spectrum.

All other possible indicators which have been derived from theories and findings by other scholars, trust in the national government, political awareness, perception of EU identity, subjective social class, age and the image of the EU have been confirmed to influence Euroscepticism. Notable, also the geography related variables influence Euroscepticism significantly, be it very limited. There is no existing literature yet that brings up these correlations. Respondents who live in a rural neighbourhood are slightly more likely to be Eurosceptic than those living in urban regions. Besides, respondents from the outskirts of the EU are less Eurosceptic than those

(24)

living in the centre of the EU. This is also visible on figure 1 and 2, which show Euroscepticism in respectively 2014 and 2019 for each EU member state. Also, this difference seems to be decreased in 2019, which is also visible on the maps.

REPRESENTATIVENESS OF EUROSCEPTICISM IN EUROPEAN PARLIAMENT

In this section, the results of the analysis for the second question will be presented. The second side-question is: Does Euroscepticism in the European Parliament represent Euroscepticism in EU member states? In the ninth European Parliament, which is constructed after the EP elections of 2019, will be 235 MEP’s from Eurosceptic parties, as shown in table 2. The European Parliament has 751 MEP’s in total meaning that 31,3% of the European Parliament can be considered Eurosceptic. Back in 2014, the 220 Eurosceptic MEP’s covered 29,3% of the European Parliament.

Table 2: The evolution of Party-based Euroscepticism in the EU between 2014 and 2019

Country Eurosceptic parties Political group Political orientation In national government Results in 2014 EP elections (in %) Results in 2019 EP elections (in %) Results in latest national elections (in %) Number of MEP’s 2014 2019 Austria Freedom Party of Austria Europe of Nations and Freedom Centre right to far right (1) Yes 19,7 17,2 20,5 4 3 Social Democratic Party of Austria Progressive Alliance of Socialists and Democrats Centre left (1) No 24,1 23,9 26,8 5 5 Austrian People's Party European People's Party Centre right (1) Yes 27,0 34,6 24,0 5 7 Alliance for the Future of Austria

None Centre right (2) No _* _* _* 0 0 Belgium Flemish Interest Europe of Nations and Freedom Centre right to far right (3) No 4,3 11,7 12,0 1 3 Workers' Party of Belgium

None Far left (4) No 1,5 3,0 8,6 0 1

New Flemish Alliance European Conservatives and Reformists right wing (4) No 16,8 13,7 16,0 4 3

Bulgaria Attack None Far right (5) No 3,0 1,1 9,1*** 0 0 National

Front for the Salvation of Bulgaria Europe of Freedom and Democracy Right wing (5) No 3,1 1,2 _* 0 0 Bulgarian National Movement European Conservatives and Reformists Right wing (5) No 10,7*** 7,4 4,2*** 2 2 Bulgarian Socialist Party Progressive Alliance of Socialists and Democrats Left wing (5) No 18,9 24,3 27,2 4 5

Croatia Human Shield None Eurosceptic (6) No 0,5 5,7 6,2*** 0 1

Cyprus Progressive Party of Working People European United Left - Nordic Green Left Left wing (7) No 27,0 27,4 25,7 2 2 Ecologist Movement

Euroscepticism in 2019 : from the numbers to the roots : a quantitative and qualitative analysis of Euroscepticism in EU member states in 2019