Thesis

La reconfiguration de la politique de la clase ouvrière

GALINDO ALFONSO, Jorge

Abstract

Pour beaucoup de personnes dans les pays occidentaux, avoir un emploi à vie ne peut plus être considéré comme sûre. Au cours des deux dernières décennies, la précarité de l'emploi a augmenté de manière prononcée et constante. La crise économique et institutionnelle, provoquée par ces facteurs structurels, a porté la démocratie à ce qu'on comprend maintenant comme un moment politique et institutionnel déterminant. Les principaux partis qui ont traditionnellement soutenu l'élaboration des politiques en Europe ont perdu rapidement du terrain au profit d'un groupe très hétérogène de nouvelles formations anti-établissement. La crise est particulièrement aiguë pour les partis sociaux-démocrates: ils perdent leur espace en tant que négociateurs politiques parmi groupes de classe. La présente thèse examine comment et pourquoi ces changements se produisent en termes de modèles de vote, de préférences politiques et de reformes de politiques.

Reference

GALINDO ALFONSO, Jorge. La reconfiguration de la politique de la clase ouvrière. Thèse de doctorat : Univ. Genève, 2019, no. SdS 109

URN : urn:nbn:ch:unige-1173060 DOI : 10.13097/archive-ouverte/unige:117306

Available at: http://archive-ouverte.unige.ch/unige:117306

Disclaimer: layout of this document may differ from the published version.

1 / 1 The reconfiguration of working class politics THÈSE présentée à la Faculté des sciences de la société de l’Université de Genève par Jorge Galindo Alfonso

sous la direction de prof. Lucio Baccaro

pour l’obtention du grade de Docteur en sciences de la société mention sociologie

Membres du jury de thèse:

M. Bruno AMABLE, Professeur M. Lucio BACCARO, Professeur Mme. Silja HAUSERMANN, Professeur M. Harry Jonas PONTUSSON, Professeur, président du jury

Thèse no 104 Genève, 25 février 2019 ii The reconfiguration of working class politics

La Faculté des sciences de la société, sur préavis du jury, a autorisé l’impression de la présente thèse, sans entendre, par-là, émettre aucune opinion sur les propositions qui s’y trouvent énoncées et qui n’engagent que la responsabilité de leur auteur.

Genève, le 25 Février 2019

Le doyen Bernard DEBARBIEUX

Impression d'après le manuscrit de l'auteur

Table of contents Résumé ...... v Abstract ...... vii Remerciements - Acknowledgements ...... ix Chapter 1. Introduction ...... 11 1.1 Part I. New trajectories and determinants of working class vote ...... 11 1.1.1 A new notion of the working class? ...... 12 1.1.2 Rising working class divisions ...... 13 1.2 Part II. The new fight for labour market policies among workers ...... 15 1.3 Outline of the dissertation ...... 16 Chapter 2. Where Have All the Working Class Voters Gone? ...... 21 2.1 The new faces of working class vote ...... 22 2.2 Why these changes? ...... 24 2.2.1 A matter of mistrust? ...... 24 2.2.2 Economic anxiety vs identity? ...... 25 2.3 Analysis: setup and results ...... 26 2.3.1 Measuring working class vote propensity ...... 26 2.3.2 Explaining working class vote choices ...... 29 2.3.3 Cross-factor comparison and summing up results ...... 34 2.4 Interpreting and concluding ...... 34 Chapter 3. Dualism, Welfare State, and the Populist Radical Right ...... 41 3.1 Populism and the Transformation of European Party Systems ...... 42 3.1.1 Exclusion (or the fear of it) as a determinant of PRR vote ...... 42 3.1.2 Insiders, outsiders and the exclusionary far-right ...... 44 3.2 Data, methodology and results ...... 46 3.2.1 Dualisation and PRR vote ...... 46 3.2.2 Pinning down the mechanism ...... 48 3.2.3 Country-level effects ...... 50 3.3 Concluding remarks ...... 51 Chapter 4. Workers’ preferences on flexicurity ...... 59 4.1 The insider/outsider framework and its main critiques ...... 60 4.1.1 The basic framework ...... 60 4.1.2 The class/occupation critique ...... 60 4.1.3 Combined approaches ...... 61 4.2 What works best? ...... 62 4.3 Operationalisation & methods ...... 64 4.4 Results ...... 67 4.5 Interpreting and concluding ...... 69 Chapter 5. A (social democratic) escape from dualisation? ...... 75 5.1 When elite battles take it from the base ...... 76 5.2 Dualisation and its consequences over social democratic parties: the emergence of an internal battle ...... 77 5.2.1 The origins of dualisation ...... 77 5.2.2 The battle over dualisation within party structures ...... 78 5.3 When structural factors meet the internal divisions ...... 80 5.3.1 The crisis in Spain and Zapatero’s labour market reform ...... 81 5.3.2 The crisis in Italy: Fornero’s and Renzi’s reform ...... 83 5.4 Reconsidering alternative explanations ...... 87 5.5 Concluding ...... 88 Chapter 6. Concluding remarks ...... 97 6.1 Voting choices: class matters (somehow) ...... 97 6.2 Policy choices: class vs job position ...... 98 6.3 Cross-chapter potential connections: paths for future research ...... 98 6.4 Beyond research ...... 99 Appendix ...... 101 iv The reconfiguration of working class politics

Résumé v

Résumé

Pour beaucoup de personnes dans les pays occidentaux, avoir un emploi à vie ne peut plus être considéré comme sûre. Au cours des deux dernières décennies, la précarité de l'emploi a augmenté de manière prononcée et constante. Les syndicats, les travailleurs privés de leurs droits, les experts et les politiciens soulignent tous que le travail précaire est un problème majeur fortement lié à certaines dispositions légales qui permettent et même renforcent son existence. Dans ce contexte, la récession européenne du période 2008-2013 a eu d'énormes conséquences.

La crise économique et institutionnelle, provoquée par ces facteurs structurels, a porté la démocratie à ce qu’on comprend maintenant comme un moment politique et institutionnel déterminant. Les principaux partis qui ont traditionnellement soutenu l'élaboration des politiques en Europe ont perdu rapidement du terrain au profit d'un groupe très hétérogène de nouvelles formations anti- établissement. La crise est particulièrement aiguë pour les partis sociaux-démocrates: ils perdent leur espace en tant que négociateurs politiques parmi groupes de classe.

Je considère que la disparition des partis traditionnels, en tant que responsables directs de la tendance enregistrée dans la réglementation du marché du travail (et d'autres domaines) depuis 1973, ne peut pas être comprise isolement de l'élaboration des politiques. L'évolution des préférences politiques des électeurs compte tenu de leur position au sein de la structure économique et de son lien avec l'offre de politique constitue le fil conducteur de la présente thèse. Compte tenu de sa grande portée, allant depuis préférences réglementaires et sociopolitiques générales jusqu’à des choix et des résultats spécifiques, je suis une structure à quatre papiers regroupés en deux parties, ce qui me permet de tester quatre arguments différents. Les résultats combinés nous aident à mieux comprendre la vague actuelle de changement politique.

La première partie traite du comportement du vote. La conclusion principale du chapitre deux est que le nouveau choix politique de la classe ouvrière est l'abstention, bien que, comparée à d'autres groupes de classe, elle est plus susceptible d'opter pour l'extrême droite. De manière significative, les deux tendances sont devenues comparativement plus intenses après 2008. À l’heure de préciser les facteurs derrière ces décisions, les données montrent que les électeurs économiquement inquiets de la classe ouvrière sont poussés vers l'abstention et pas vers l'extrême droite, tandis que les électeurs anti-immigration de classe ouvrière choisissent l'extrême droite. Dans les deux cas, la classe ouvrière européenne post-2008 est beaucoup plus influencée par ces facteurs.

Le chapitre trois analyse la classe ouvrière en fonction du poste occupé. On constate que les travailleurs sont plus enclins à choisir des partis de droite exclusifs (PRR, ou partis de droite radicalement populiste) chaque fois qu’ils jouissent d’un statut acquis qui pourrait être attaqué, c’est-à- dire s’ils sont devenus des travailleurs permanents (initiés). Ces segments sont plus susceptibles de voter pour le PRR d'exclusion par rapport aux outsiders. On considère que la plate-forme entièrement protectionniste de PRR présente un intérêt particulier pour la protection de son statut économique. Ainsi, les initiés confrontés à une concurrence accrue au sein de son segment du marché du travail et ceux dont la situation matérielle serait la plus touchée en cas de perte de leur statut auraient encore plus de chances de choisir une plate-forme entièrement protectionniste. Au même temps, la tendance est plus marquée dans les pays où les systèmes de protection sociale sont fondés sur le statut (par exemple, les États-providence bismarckiens) et offrent donc une protection plus large aux initiés par rapport aux étrangers.

L'idée à retenir jusqu’ici c’est que, dans certains pays, la structure de la demande politique ainsi que la répartition des risques économiques et du marché du travail conditionnent l'émergence d'une nouvelle offre de partis hors du centre politique. Il y a toutefois une conclusion plus simple: les politiques spécifiques sont importantes.

La flexicurité, définie comme une manière de « liquide en échange flexibilité » a été présentée comme une solution politique visant à réduire les divisions grandissantes au sein de la classe ouvrière. Il s’agit toutefois d’une proposition qu’on voit rarement promulguée. Quelle est la raison ? Répondre à cette énigme est pertinent pour comprendre les déterminants des réformes sous-optimales du marché du travail. Il est aussi potentiellement significatif dans un sens plus large: il présente un micro-cosmos vi The reconfiguration of working class politics

expliquant pourquoi et comment les électeurs sont désenchantés des partis traditionnels dans les situations de crise. La deuxième partie de la thèse est consacrée à cette question. Le chapitre quatre mène une analyse préliminaire concluant que les différences de postes (employés / chômeurs) donnent des indices vérifiables sur l’idée de savoir pourquoi et comment une coalition pro-flexicurité pourrait être créée, mais les différences de classe confèrent un pouvoir explicatif similaire, voire supérieur. Les chômeurs sont plus susceptibles que les employés de préférer des réductions de la protection de l'emploi, un investissement accru des politiques actives du marché du travail et le maintien du niveau des allocations de chômage. Ainsi, ils ont tendance à préférer la flexicurité à son alternative opposée. Mais ils ne préfèrent pas une libéralisation complète (réduisant les trois systèmes de protection) à une protection complète (maintien de la protection de l'emploi, des avantages non liés à l'emploi et des politiques actives du marché du travail à des niveaux élevés). Au même temps, les employeurs et les professionnels montrent des préférences plus fortes que les travailleurs manuels peu ou moyennement qualifiés en ce qui concerne la flexicurité et la libéralisation. Bien que ce soit une approche exploratoire qui évite de tirer des conclusions solides, le cadre professionnel offre des perspectives un peu plus prometteuses, mais la perspective du poste reste clairement pertinente pour expliquer le soutien variable apporté aux différentes options de réforme du travail.

Le chapitre cinq, cherche à comprendre pourquoi certains pays démocratiques poursuivent réformes du marché du travail alors que d’autres ne le font pas. Cette vaste question est au cœur de l’économie politique comparée. Il est bien établi que les parties prenantes forment et consolident les préférences politiques en fonction de quatre facteurs: coalitions disponibles, incitations générées par des mesures antérieures, idées disponibles et conditions structurelles actuelles qui déterminent à la fois leurs possibilités de croissance et leur répartition.

Les gouvernants sont en tout état de cause des acteurs essentiels dans l’élaboration des politiques. Si la path dependence et les déterminants structurels restent constantes, on pourrait alors soutenir que la composition idéologique du gouvernement devient la clé pour comprendre la variation des réformes du marché du travail. Si bien on attend généralement que les partis de droite produisent des réformes de libéralisation, il peut être plus difficile de prédire les résultats lorsque les réformes sont entre les mains de partis de gauche constitués d'un noyau constitué de travailleurs. Les cas de l'Italie et de l'Espagne pendant la grande récession (2008-2013) sont un bon exemple. Les deux pays ont des structures de marché du travail profondément dualisées, avec différents niveaux de protection efficace contre le risque de chômage accordés à différents travailleurs en fonction de leurs positions contractuelles, divisant la population active entre les personnes protégées (c.-à-d. insiders) et les non protégées (c.-à-d. outsiders). Les deux pays ont porsuivi les mêmes solutions politiques à plusieurs reprises pour résoudre la dualisation, par la réduction de l’écart de protection entre types de contrats. Au cours de la Récession, les deux pays ont été soumis à des pressions structurelles étrangères similaires pour libéraliser leurs lois du travail. Dans les deux cas, un parti de centre-gauche était en charge. Malgré toutes ces coïncidences, l’Italie a poursuivi une réforme beaucoup plus ambitieuse. Le but de ce chapitre est d’examiner pourquoi elle a choisi de la faire.

Je soutiens que, bien que le parti italien Partito Democratico (PD) et l’espagnol Partido Socialista Obrero Español (PSOE) aient une aile réformiste et une autre favorable au statu quo, le front réformiste ne parvient qu'à gagner la bataille interne dans le PD. Les deux ailes mènent la même bataille depuis des décennies dans les deux pays, et chaque chef de parti est arrivé au pouvoir avec un compromis entre les deux ailes. À l'exception du Premier Ministre italien , qui a accédé au pouvoir en 2013 avec un discours clairement réformiste et pro-changement qui décrivait les syndicats et ses représentants comme des bloqueurs intéressés. Sa victoire au sein du parti a été obtenue grâce à une stratégie «confrontationiste» soutenue par les réformistes, qui jouissent d'une présence assez importante au sein du PD depuis que le parti a été conçu dès le début comme une alliance entre centristes et socialistes. Lorsque Renzi a associé sa victoire interne aux élections primaires de 2013 au résultat favorable du parti aux élections européennes de 2014, il a affirmé qu'il avait reçu un mandat de changement. Ceci (avec la charge de problèmes, la pression externe) était la fenêtre d'opportunité des réformistes pour faire pression en faveur d'une réforme à grande échelle. En Espagne, les réformistes ont utilisé le même argument, mais l'absence d'un chef de parti capable d'utiliser un mandat politique pour faire pression en faveur de la réforme a entravé leurs tentatives.

Dans son ensemble, ces quatre chapitres offrent une vue en mosaïque de la reconfiguration de la politique de la classe ouvrière. Abstract vii

Abstract

For many people in Western countries, having a job for life cannot be taken for granted any more. In the last two decades there has been a steep and steady increase in job insecurity. Unions, disenfranchised workers, experts and politicians, all highlight precarious work as a major problem that is intimately related to certain legal provisions that allow and even foster its existence. Within this context, the 2008-2013 European recession had enormous consequences. The economic and institutional crisis that ensued from these structural factors brought democracy to what is understood now as a defining political and institutional moment. The mainstream parties that traditionally sustained policymaking in Europe are rapidly losing ground in favour of a very heterogeneous group of new anti-establishment formations. The crisis is particularly acute for social democratic parties: They are losing their space as a political dealmaker between class groups. I consider that the demise of mainstream parties, as chief architects of the trend registered in labour market regulation (and other fields) since 1973, cannot be understood in isolation from policymaking. The changing political preferences of voters given their positions within the economic structure, and how does it relate to policy supply, constitutes the common thread of the present dissertation. Given such a broad scope, going from broad regulatory and socio-political preferences to specific choices and outcomes, I follow a four-paper structure grouped in two parts, which allows me to test four separate arguments. The combined results help us better understand the current tide of political change. The first part deals with voting behavior. Chapter two's main finding is that the new political home of the working class is abstention, although, when compared to other class groups, it is also more likely to opt for the far-right. Significantly, both trends have become comparatively more intense after 2008. When it comes to further pin down the factors behind these decisions, data shows that economically anxious working class voters are being driven towards abstention and not towards the far-right, while anti-immigration working class voters are choosing the far-right. In both cases, the post-2008 European working class is much more determined by these factors. Chapter three breaks down the working class depending on job position. It is found that workers are more prone to choosing exclusionary right wing parties (PRR, or Populist Radical Right parties) whenever they have a gained status that might be attacked, i.e. when they have become permanent workers (insiders). These segments are more likely to vote for the exclusionary PRR when compared to outsiders. It is considered that PRR's all-protectionist platform is of special interest for protecting their economic status. Thus, insiders facing more competition within its segment of the labour market, and those whose material situation would be most affected in case of losing their status, would be even more likely to choose an all-protectionist platform. At the same time, the trend is steeper in countries where welfare systems are status-based (i.e. Bismarckian welfare states), and therefore offer a larger protection to insiders compared to outsiders. The takeaway idea of the first and second paper is that, in some countries, the structure of political demand plus the distribution of economic and labour market risks conditions the emergence of new party supply away from the political centre. There is a simpler bottom line, however: Policies matter. Flexicurity, defined as some form of “cash in exchange for flexibility”, has been portrayed as a policy solution to reduce the growing divisions within the working class. It is, however, a proposal that you rarely see enacted. Why is that so? Answering to this puzzle is relevant for understanding the determinants behind suboptimal labour market reforms. It is also potentially significant in a broader sense: It presents a micro-cosmos on why and how do voters become disenchanted with traditional parties in crisis situations. The second part of the dissertation is devoted to it. Chapter four conducts a preliminary analysis finding that differences in job position (employed/unemployed) give testable clues to the idea of why and how a pro-flexicurity coalition might be crafted, but class differences convey similar, if not superior, explanatory power. The unemployed are more likely than the employed to prefer employment protection reductions, increased investment in active labour market policies and maintaining levels of unemployment benefits. Thus, they tend to prefer flexicurity over its opposite alternative. But they do not prefer complete liberalisation (reducing all three protection systems) over full-scale protection (keeping employment protection, non- employment benefits and active labour market policies at high levels). At the same time, employers and professionals do display more intense preferences than mid- and low-skilled manual workers towards both flexicurity and liberalisation. If anything, and although this is only an exploratory viii The reconfiguration of working class politics

approach that refrains from establishing strong conclusions, the occupational framework delivers somewhat more promising insights, but the job position perspective remains clearly relevant to explain varying support for different options of labour reform. Chapter five, finally, aims to explore why some democratic countries pursue such labour market reforms while others do not. This admittedly broad question lies at the heart of comparative political economy. It is well established that stakeholders form and consolidate political preferences based on four factors: Available coalitions, incentives generated by past measures, available ideas, and present structural conditions which determine both the opportunities for growth and distributing it. Governing parties are in any case pivotal actors in policymaking. If path dependency and structural determinants are left as constants, then it could be argued that the government’s ideological composition becomes the key to understanding variation in labour market reforms. While it is generally expected for right-leaning parties to produce liberalizing reforms, it can be more difficult to predict the results when reforms are in hands of left-leaning parties with a core constituency formed by workers. A good instance is provided by the cases of Italy and Spain during the Great Recession (2008-2013). Both countries have deeply dualised labour market structures, with different levels of effective protection against the risk of unemployment given to different workers depending on their contractual positions, dividing working population between the protected (i.e. insiders) and the unprotected (i.e. outsiders). Both countries have had the same policy solutions repeatedly floated to solve dualisation through the reduction of the breach of protection between contractual types. During the Recession, both countries faced similar structural pressures from abroad to liberalise their labour laws. In both cases there was a centre-left party in charge. Yet, in spite of all these coincidences, Italy pursued a much more ambitious reform. The aim of this chapter is to examine why it elected to do so. I argue that, while in both the Italian Partito Democratico (PD) and the Spanish Socialist Party (PSOE) there is a reformist wing and a pro-status quo wing, the reformist front only manages to win the internal battle in the PD. Both sides have been fighting this same battle for decades in both countries, and every single party leader has come to power with a compromise between both wings. Except for Italian PM Matteo Renzi, who in 2013, came to power with a clear reformist, pro-change discourse that portrayed unions and its representatives as self-interested blockers. His victory within the party was achieved thanks to a 'confrontationalist' strategy supported on reformists, who enjoy a somewhat prominent presence within the PD since the party was conceived from the beginning as an alliance between centrists and socialists. When Renzi coupled his internal victory in the 2013 primary election with the party's favourable result in the 2014 European election, he argued that he had received a mandate for change. This (together with problem load, external pressure) was the reformists' window of opportunity to push for a full-scale reform. In Spain, reformists used the same argument set but the lack of a party leader who was able to use a mandate for change to push for reform hampered their attempts. Put together, these four chapters offer a mosaic view on the reconfiguration of undergoing dynamics within working class politics.

Remerciements - Acknowledgements ix

Remerciements - Acknowledgements

First and foremost, I would like to express my deep appreciation towards my supervisor, Professor Lucio Baccaro, for his strong support and good advice, as well as for his patience, during all the stages of the present dissertation. I would also like to thank the members of my jury for their disinterested commitment and for their valuable feedback and fruitful criticism, opening my eyes to previously unconsidered angles. At the same time, I must send a thank you to the Swiss National Science Foundation. The SNF gave me essential funding and the initial opportunity to develop the project that culminates in this dissertation. Beyond that, I hold a very important intellectual and personal debt with Professor Pepe Fernández- Albertos, who opened the door for this project. He was the one who made me aware of the SNF's opening, and kindly supported my candidacy. Also, I consider Gonzalo Rivero to be my informal academic mentor. Without his experienced support I would have never been able to navigate the academic world. A huge thanks goes for him as well. The list of people whose comments, debates and points of view have greatly enriched my perspectives is quite long. But I must include at the very least those who have accompanied me as members or collaborators of Politikon during all these years. Aside from Gonzalo himself, that list includes Luis Abenza, Berta Barbet, Sílvia Clavería, Elena Costas, Juan Font, Lucas Gortázar, Kantor, Kiko Llaneras, Ramón Mateo, María Ramos, Octavio Medina, Jorge San Miguel, Roger Senserrich, and Pablo Simón. I could have never dreamt of a better group of people to debate with for eight years now (and counting). Guillem Vidal, a person who is not only my co-author in one of the papers presented here, but also a brilliant political scientist who never studied political science, deserves all recognition for any good thing springing from chapter three. I reserve a special place in this list to my closest family. My parents, Geles and Andrés, my uncles and my cousin (Aurora, Toni and Andrea). If I have any love and respect for knowledge, I owe it to them. If I have been able to come this far without getting crazy in the process, I owe it to them as well. But, most of all, I am infinitely thankful for having come to Geneva following your path, Catalina, and for the relentless, unconditional support that you have given me since that moment, almost six years ago. Wife, friend, partner, companion... I don't care that much about the word. What I care about is our common life project. I hope this dissertation enables and enhances it in the same way you empower the best part of me every single day. I hope, and I promise, that the culmination of this long trip will impulse us to further discover the world together. x

Introduction xi

Chapter 1. Introduction

The mainstream parties that used to sustain policy-making in Europe are now losing ground to a heterogeneous group of new formations with their common trait (if they share any) of being self- proclaimed anti-establishment platforms. They appear on the left (e.g. Podemos in Spain, Syriza in Greece), on the right (e.g. Front National in France, Jobbik in Hungary), but also on the political centre (e.g. En Marche in France, Ciudadanos in Spain, M5S in Italy). Political scientists have given a great deal of attention to both new right-wing (e.g. Kitschelt 2007; Rydgren 2007 among many others) and new left-wing (Lavelle 2008; March and Rommerskirchen 2015) parties, as well as to the most affected among traditional political families. On the demand side, the working class is accumulating most of the attention in the search for who is responsible of the unravelling of Post-war Western party systems. This crisis has been particularly acute for social democratic parties, which seem to have lost their space as deal-makers between the different groups involved in political conflict within Western capitalist democracies, losing the favour of the working class. Thus, both researchers and the public are wondering what happened with the voting coalitions that sustained traditional parties and, most importantly, the key policies they brokered (e.g. Pierson 1996; Bonoli 2005; Gringich and Häusermann 2015). As existing welfare systems come under strain and pressure to reform while party constituencies are being reconfigured, a peremptory question emerges: who are the people pushing for reform in either one direction or another, and who are the ones resisting changes? More specifically, where do preferences of workers lie regarding new policy packages to confront the rising of new risks? These two questions are best considered together for a simple reason: parties (regardless of whether they choose to call themselves “parties”, “platforms, “movements” or some variation) remain the ultimate decision makers, and sooner or later voters hold them accountable. The vote-party- government-policy continuum is of course not entirely straightforward, and has been challenged by functionalist approaches that treat conflict as a sub-optimal outcome. However, the revival of dissent in Western politics arguably demonstrates that the ‘civilised conflict’ remains the basic channel for decision making. Within this (admittedly broad) framework, I tackle four separate, more specific questions: (1) where are working class voters' party preferences right now?; (2) how do new divisions across workers affect voting decisions, particularly regarding the most distinctly new platform (i.e. far-right parties)?; (3) where is the support for new policy packages that aim to tackle rising risks, and how does this support distribute itself according to class and job status divisions?; (4) why do these packages get themselves enacted into law in some places while not in others, and how does that relate to old and new party platforms with underlying supporting coalitions? The resulting four-paper structure allows me to test different arguments and explore various fronts that will hopefully help us better understand the current tide of political change. I believe these four (admittedly distinct) chapters are best when divided into two separate parts (questions 1 and 2; questions 3 and 4).

1.1 Part I. New trajectories and determinants of working class vote used to be a political home open to almost everyone. It was a home built by the working class, for the working class, but access was not restricted to them. That is no longer the case, however. The demise of social democratic parties has ranged from significant to dramatic across Western European countries, where the centre-left used to dominate. British Political Scientist Chris Hanretty wrote that “[e]lectorally, West European social democrats are at their lowest point for forty years” (Hanretty 2015): using 1970 as a base year, he shows how Western European social democratic parties have lost enough vote to reach 1960s levels, with the mid-2000s being a crucial turning point. Most significantly, working-class vote, once the force behind all-inclusive reforming coalitions, is no longer captivated by social democrats. According to ESS data, 2002 and 2016, social democratic voting intention among manual workers in Europe went down from 24.5% to 15.8% (see Table 2.1 in Chapter 2). Such a drop raises the question of why it decreased so dramatically and what took these two (parties and constituencies) away from each other. For many observers, especially those positioned furthest of the ideological spectrum, social democracy had been losing touch with its traditional voting base. As Jeremy Corbyn put it during his campaign to become the new Labour Party leader: "I think our problem as a party is we weren't 12 The reconfiguration of working class politics

offering a clear enough alternative to the Conservatives” (Beattie 2015), referencing the ‘’ centrist strategy put forward by Tony Blair during the 1990s. He is, of course, not alone in this belief. Nichi Vendola, a notable leftist politician from Southern Italy, accused centrist Matteo Renzi of “killing” the centre-left when he took control of the Partito Democratico (Vendola 2015). Pablo Iglesias, co- founder and head of the emerging leftist party Podemos in Spain, has built all his strategy around the idea that the traditional centre-left party PSOE prefers to reach agreements with the conservative PP (“Iglesias acusa al PSOE de preferir a Rajoy [Iglesias Accuses PSOE of Preferring Rajoy], 2018, February 5). Class affiliation has been part of these accusations as well. Jean-Luc Mélenchon, leader of the leftist platform La France Insoumise (roughly translating to Unsubordinated France), has accused Emmanuel Macron of being a “a bourgeoise who knows nothing about life” ("Un grand bourgeois qui ne connaît rien à la vie”, 2017, January 4). Macron, having a socialist origin, won the 2017 French presidential election with a centrist platform that was heavily criticized by more left-leaning politicians. These debates have their academic correlates. The idea of party convergence traces back to the very foundations of contemporary political science. Early spatial models of party competition (Downs 1957; Stokes 1963) suggested that agents competing for maximizing vote support had incentives to converge towards the median voter’s political preferences. Such equilibrium opened up the space for new competitive strategies. For instance: if the population remaining in one of the political extremes is numerous enough, a political entrepreneur might have incentives to create a corresponding platform. Say, social democracy moving to the centre in several Western democracies during the 1990s under the ‘third way’ strategic proposition commanded by Tony Blair and US President Bill Clinton, and therefore leaving enough room for more leftist proposals to thrive. If, also, the working class is over- represented in that ideological extreme, the appeal would be specifically directed to them. Structural, conflict-based understandings of the demise of social democracy might be broadly read under the light of this (very schematic) model: a vacuum left by traditional centre-left platform on economic policy that becomes filled by someone else (e.g. Patton 2006 focusing on the German case). There has not been, however, a corresponding rise of far-left or new-left vote among European workers, something that could be expected from such line of argumentation. In consequence, other authors have seen a different ‘vacuum’ opening up, now in the identity-cultural axis. Donald Trump’s 2016 presidential victory was regarded by many as the answer given by a different kind of convergence: that of ‘identity ’ (Lilla 2016) or ‘cultural globalization’ (Croucher 2018). Under this perspective, the growing consensus among mainstream parties on multiculturalism and open border policies left a space to be filled by proponents of closed borders and homogenous national identities. Moreover, the demand of these ‘nationalist’ policies would mostly come from manual workers, and would have a particular focus on anti-immigration positions (Oesch 2008; Kriesi 2012). A multi-spatial view of political competition that joins economic and cultural dimensions was then proposed by some authors (Kitschelt 2012). As a matter of fact, some centrist politicians such as Emmanuel Macron (France), Albert Rivera (Spain), Matteo Renzi (Italy) and Guy Verhofstadt (leader at the European Parliament of the Alliance of Liberals and Democrats for Europe) have argued that the new divide defining Western politics prioritises pro-openness platforms (open borders, open markets and open identities) versus pro-closedness proposals. Pro-closedness would mean left- authoritarian, or ‘welfare chauvinist’, policy packages, something several researchers have already explored (Schumacher and van Kebergsen 2016; Afonso and Rennwald 2018). Whether for economic or for cultural reasons, all these perspectives highlight a growing distance between the working class and social-democratic parties, and the subsequent search of new options. However, there is no clear-cut conclusion, neither on the direction nor on the determinants for the search. Since my goal in this first two chapters is to contribute to this debate, I believe I should begin by questioning the core concept of “working class” itself as it is classically understood.

1.1.1 A new notion of the working class? Until now, I have not problematized the concept of “working class”, implicitly identifying it (or leaving the reader to identify it) with workers in the industrial sector, or workers whose tasks mostly comprise manual actions. More specifically, the most common, unstructured consideration of the concept still traces itself back to the original Marxian notion: anyone who does not possess capital, and therefore must sell her labour power to survive, is part of the working class. Introduction xiii

It has been suggested, however, that such a notion, albeit structurally relevant even nowadays, is insufficient to explain complex multi-platform political choices in pluralist democracies. Moreover, the decoupling of management and capital ownership and the tertiarisation of the economy during the second half of the 20th century demands that we reconsider the notion of class. Summarising a very prolix and complex subgenre of the sociological literature, John Goldthorpe and his co-authors came up with what became the most widely used class classification in the last quarter of the century (Goldthorpe et al 1980). It was based upon two variables: the difficulty of monitorisation of tasks performed by the individual, and the level of human asset specificity required to develop them. Goldthorpe et al advanced a crucial idea: in a post-industrial economy, considering the task composition of occupations and, consequently, the amount and the type of human capital needed to develop them becomes a crucial variable to distinguish among workers, and particularly among their exposition to situations of economic adversity. Oesch (2006) built upon these ideas to propose his now ubiquitous occupation-based class division. Its main traits are (1) the combination of sector-based and skill-based aspects of work life; (2) the distinction between types of tasks (and the kind human capital required to execute them) within both the industrial and within the service sectors, with particular importance given to the latter. In the last decade, Oesch has widely demonstrated the relevance of his approach to define political behaviour (e.g. Oesch 2008a, Oesch 2008b, Oesch 2012). Therefore, departing from Oesch's work, it might be worth assuming a broader and, at the same time, more nuanced idea of the working class. More specifically, as the share of manual work diminishes in Western economies, it makes sense to consider semi- and non-qualified service workers whose tasks fall outside the desk/office area, along with non-supervisor manual workers as part of the working class. These segments share key traits that link them with the more classical notions of the working class, namely the relative lower reliability on human asset specificity, a comparatively high possibility of monitorisation by supervisors, and, as a consequence of both, a higher exposition to economic risks due to labour market woes. In Marxian terms, it could be said as well that these workers are further away from the possession of capital, particularly if we accept a definition of capital that goes beyond its physical/monetary notion and includes social, cultural, educational and relational dimensions (Bourdieu 1986). Such is the approach I take in the two first chapters of the present dissertation. My goal with the two first chapters that comprise the present dissertation is to contribute to these debates by focusing on the following specific questions: (1) if the working class is no longer a social democratic monopoly, where is it heading now? (2) is the working class making a united or a divided trajectory? (3) what are the determinants of workers’ voting choice? I don’t expect to find a single answer to these questions, such as “the working class is heading in bloc to the far-left due to identity issues”, or “manual workers now prefer the far-left because of economic grievances”, or even “workers are completely de-aligned and there is no clear voting pattern”. Instead, I foresee a complex map of multiple realignment that is defined along new lines of political competition. Studying voting choices is arguably best done through survey analysis, thanks to its capacity to aggregate individual moves in an efficient, representative manner. In this regard, I find the European Social Survey (ESS) to be the best source thanks to its broad regional scope (covering most European countries), its coherence across waves, and its long-time reach (2002-2016). I apply logistic and multinomial logistic regression models over ESS data to find out voting patterns and determinants, restricting my analysis to the aforementined definition of the working class. This explains why I find in Chapter 2 that while manual workers seem to be indeed more likely than other class groups to vote for the far-right (and having disappeared any significant link between them and social democrats once several control variables are thrown in), this does not imply that it manages to keep a statistically significant differentiated majority of manual workers: instead, they tend to distribute across party groups mirroring other class segments, with the notable exception of a substantial overrepresentation in abstention. Moreover, the drivers of working class vote are multiple: mistrust in existing institutions and economic anxiety particularly after the Great Recession move voters towards abstention, while the link between anti-immigration preferences and far-right vote has become even more intense than it already was after 2008. The onset of the crisis marks therefore an interesting juncture that might be linked to the greater prevalence of economic insecurity.

1.1.2 Rising working class divisions Any consideration of the reshaping of the working class after the advent of tertiarisation is however incomplete without taking into account the idea of dualisation. It was the relatively unexpected rise of 14 The reconfiguration of working class politics

structural unemployment after the supply shocks of the 1970s that suggested to labour economists the growing importance of job status, i.e. being employed or unemployed, to define political preferences. Insiders (those sheltered against the consequences of unemployment) and outsiders (those who were not) became commonplace notions in the field of labour economics during the 1980s (e.g. Lindbeck and Snower 1985), and then made their entrance into the world of political economy by the 1990s under the basic idea of insiders being more interested in institutions that keep their status, such as employment protection legislation, than outsiders (Saint-Paul 1993). It was David Rueda who took this to the arena of party politics, suggesting that insiders (now not only the employed, but also, or more particularly, those who enjoyed a lower exposition to the risk of unemployment) and outsiders (not only the unemployed, but those at a higher risk of becoming so, e.g. temporary workers with lower severance payments attached to their contracts) had different political preferences, which in turn might create a dilemma to traditional parties, and particularly to social democrats: if there are emerging divisions within their natural constituency, does this mean they have to choose which constituents they will cater to? (Rueda 2005, 2007). From this point of view, the insider-outsider division is particularly problematic for social democratic parties, since they would face a dilemma between defending existing institutions and regulations, and the emerging demand for reform. Note that there is a three-stage incremental consideration implied in the stream of the dualisation literature. The basis of everything is the notion of differentiated risk exposure, a socio-economic finding that seems to be relatively well-established (although with important cross-country differences). From there, different preferences on policy packages are derived. The last step is the choice of political platforms, which is based on policy preferences and, in turn, on different exposures to risk. In Chapters 3 and 4 of the present dissertation I tackle the two last stages of this continuum. These assertions are, however, debated within the literature. Beginning with the latter, while the weakening of the link between outsiders and social democracy is not particularly disputed, its destination has recently sparked some interesting, yet preliminary findings. While, for instance, Marx and Picot (2013) have identified a link between temporary workers and the new left, Rovny and Rovny (2017) complement this with the connection between insiders and the far-right. There remains, however, work to be done, particularly regarding the better identification of specific mechanisms. This is why, in the third chapter, studying the labour market dimension as a possible determinant of radical- right voting shows that insiders tend to prefer the far-right, while outsiders have a negative relationship with them. Moreover, working-class (manual and service), mid- and low-skilled insiders display a particularly pronounced move in this direction. The exclusionary nature of these parties seems to appeal to insiders, building an equivalence between their particular interest and that of ‘the people’ (in populist terms). Most crucially, when looking for more specific mechanisms, such tendency can be at least taken closer to the dimension of perceived risk exposure: it looks like it is the combined presence of higher unemployment rates and migrant workers' rates within occupation-country clusters which determines insiders' choice for the far-right. Using again European Social Survey data in order to model far-right voting across Western European countries where the platform has an established presence, insider over-representation among the Populist Radical Right (PRR) seems well-proven, particularly within Bismarckian welfare states where existing conditions might reinforce protectionist perspectives among insiders. Also, insider voters who belong to working class segments or whose position is otherwise more vulnerable (e.g. mid- and low- skill levels) have a stronger association with PRR parties, thus suggesting that the combination of having something to lose and the idea of an alternative world on which the imagined protection or status quo is actually lost is a strong determinant for workers choosing the far-right. Last but not least, union strength at the country level seems to deter insiders to join PRR parties, underlining the importance of organizing the working class to stop the advancement of the far-right across Europe. The combination of both chapters and its mixed conclusions serve as a reminder of the fact that vote is among the most complex political phenomena to be explained, since it summarises in a single, easily measurable variable all the possible factors that define a person’s political position. Applied to the problem at hand, it means that the divorce between social democracy and the working class cannot be attributed to a single culprit, and that it is a phenomenon composed by several complementary dynamics. Introduction xv

1.2 Part II. The new fight for labour market policies among workers Moving on now to the second half of my dissertation, it is a well-established fact that policies matter to define the relationship between the voters and the elected. Among them, labour market policies might matter with particular intensity from the point of view of class voting patterns. Or, at least, these seem to be at the core of both lines of argumentation. For the insider/outsider framework, the key difference between the former and the latter group comes with the choice of preferred policies to protect them against the risk of unemployment: while insiders choose EPL, outsiders have a bias towards unemployment benefits (NEB) and active labour market policies (ALMP) (Rueda 2005). On the other hand, labour market liberalisation constitutes a cornerstone of party convergence arguments (McBride & Williams 2001; Palier 2010). Also, it seems logical to expect that the change in support for the chief architects of the trend registered in labour market (de)regulation since 1973 cannot be understood in isolation from their policy-making. Therefore, the second common thread of the present dissertation is the policy preferences of voters given their positions within the economic structure, and how this relates to the actual policy supply. As a matter of fact, one of the most frequent critiques to the insider/outsider framework departs from what could be seen as a party convergence basis, in the sense that it focuses on the coherence and completeness protectionist policy packages (Emmennegger 2009), thus forcing outsiders to look for alternatives to the left of social democracy (Marx and Picot 2013). From this point of view, the classical left-right opposition (translated into protection-deregulation for labour market policies) would still be very much the way to interpret policy fights underlying coalition change among workers. Dualisation, i.e. the selective deregulation of EPL (but also cuts in NEB and ALMP) that mostly affect non- permanent workers and that has been particularly widespread among European countries since the 1980s (Emmennegger and Häusermann 2012; Häusermann and Schwander 2012), should be therefore regarded as a form of partial liberalisation. In that respect, flexicurity reforms that advocate for an even decrease in EPL in exchange for increasing spending and resources for NEB and ALMP could be portrayed as a way of achieving a compromise between employers and (different types of) workers in advanced capitalist economies. Or, at the very least, it should be the preferred option for those who are unprotected by existing policy arrangements, especially where EPL is prevalent. Proponents of the insider/outsider framework imply as much. Maintaining our position within this perspective for the sake of the argument, it should be therefore somewhat surprising for the proponents how uncommon flexicurity reforms are. In the last two decades, since the rise of the concept during the 1990s, only a few European countries are considered as having fallen under this paradigm, most notably Denmark (Viebrock and Clasen 2009) and the Netherlands (Van Oorschot 2004). The insider/outsider perspective would correspond to this mentioning of the blocking of those who benefit from current protectionist/dual regulations (i.e. insiders), who in turn are at the core of centre-left party coalitions. It is also possible to fit the role of employers within this framework, who in principle should prefer full-scale liberalisation but whose situation in segmented labour markets might be seen as a second-best, since it gives to them access to ‘cheap labour’ (King and Rueda 2008). Alternatively, flexicurity proposals might constitute an incoherent, incomplete policy package that does not correspond with voters’ fault lines since preferences organise along the single protection- liberalisation line, which in turn would find more correspondence with a classic occupational division. High-skilled occupations, or occupations that accumulate more (financial, social, cultural, human) capital would favour liberalisation more than low-skill occupations whose occupants gather less resources to confront labour market risks on their own. Flexicurity would here fit better with the former group (Boeri et al. 2012). This is an empirical question at heart, one that could be solved if the appropriate data on individuals’ preferences were available. Unfortunately, there is no readily accessible source that perfectly fits this objective. There is, however, a good-enough source: Eurobarometer special surveys in all EU countries conducted in 2006, 2009 and 2011 that feature three key questions on preferences on whether or not EPL, ALMP and NEB ought to be reformed. I use logistic models based on these data sources to approximate an answer to the question of who supports deregulation, and in particular, who supports flexicurity (decreasing EPL, but not NEB and increasing ALMP). My findings indicate that flexicurity, indeed, has very little support across all types of labour market participants. Moreover, while the insider/outsider framework helps to explain the varying preferences for EPL reduction and for 16 The reconfiguration of working class politics

flexicurity over dualisation, it does not seem to apply well to divisions over NEB/ALMP, or over liberalisation vs protection. Instead, the occupational class division conveys explanatory power all across the policy spectrum, and particularly in the latter group. All in all, the emerging picture gives some partial support to both the insider/outsider perspective and its critiques, producing a combined result that politicians (particularly social democratic politicians) should take into account. Indeed, there seems to be some potential tensions within broad voting coalitions that have traditionally sustained centre-left parties. Which in turn gives way to the last stage of the vote-party-government-policy link: the moment of decision making, and thus the moment on which the need of reform should be faced by ruling parties. The Great Recession that began in 2008, and that turned into a European Monetary Union (EMU) debt-focused crisis by 2010, offers a perfect context for analysing the possible answers to this question. In its second phase, the crisis exerted considerable pressure over a selected group of European economies. Cyprus, Greece, Ireland, Italy, Portugal, Spain (and even Belgium and France to a lesser extent) all came under the scrutiny and pressure of its creditors given its apparent incapacity to generate enough growth to repay its public and private debts. International organisations such as the IMF, the OECD and the European Commission itself recommended, or directly agreed on, deep structural reforms. Labour markets were one of the key areas under consideration, particularly for Southern European countries. All of them were considered to be plagued by productivity problems, hampered by stringent and segmented EPL. The recommendation of EPL streamlining then became general. Structural and external factors are widely regarded as forces pushing for reform (Carlin and Soskice 2008; Lodovici 2000; Meardi 2012), especially those coming from the EMU institutional setup (Alesina et al. 2008; Calmfors 2001; Duvel and Elmeskov 2006). However, existing institutions are considered by many authors as fundamental determinants of future policy paths (Ebbinghaus 2005; Thelen 2009; Thelen 2014). This would make a comparison between countries sharing both a common past and present structural pressures particularly informative for my goal of understanding how parties facing similar conditions and have similar ideological backgrounds, approach labour market reform. In the 2010-2015 crisis cycle, both Spain and Italy were debtor countries within the EMU featuring strained labour markets and facing strong demands to streamline regulations. Also, both countries have a history of segmented labour markets (Mingione 1995). Therefore, history, context and derived constraints were shared to a large extent. Most crucially for my analytical goals, in both countries there was a mainstream centre-left party in charge when key decisions to improve labour market performance were demanded both from abroad and from within. In other words, comparing the Italian and the Spanish approach to the EMU crisis with regards of labour market reform makes a good ‘most similar cases’ approach, in which all but one key determinant variable match between cases, while the outcomes differ. The differing outcomes reflect the distinct reform paths taken by the Spanish Socialist Party (PSOE by its Spanish name) in 2010 and by the (PD by its Italian name) in 2014-2015. While the PSOE enacted what could be seen as a partial, rather timid reform that went somewhat in the direction of segmented liberalisation, the PD went forward to implement a full-scale de-dualising reform much more in line with external and internal demands for streamlining EPL. Through a qualitative comparative analysis that takes into account (1) the countries’ recent history of labour market reforms under social democratic rule; (2) the constraints faced during the EMU crisis; (3) the different political strategies pursued by the key centre-left parties, this chapter emphasises the last factor. Without being able to make a fully-generalisable argument (qualitative analysis provides nuance in exchange of generalisation capacity), I propose that the key difference between PD’s and PSOE’s reform choices come, at least partially, from the different resolution of long-standing internal debates among party elites on how to tackle labour market inefficiencies. In both cases there was a more reformist wing (mostly composed by members with a background in Economics) of pro-flexible agents within the platform. The PD's reformist wing enjoyed a window of opportunity that was not there for the PSOE's, and managed to take advantage of it.

1.3 Outline of the dissertation As explained above, the present dissertation is divided into two parts, each of them comprising two different chapters. In the first part, chapter two deals with the changing relations between new and old parties and different segments of voters, considered under an occupational/class perspective. It Introduction xvii

tackles the diverse competition for the working class vote making dichotomy analyses that compare voting patterns among workers when choosing between social democrats and the far-left, social democrats and the far-right, and most significantly, abstention. Chapter three focuses on the three- way relationship between dualism, Welfare State, and the Populist Radical Right (PRR), and pins down the voting preferences of insiders and outsiders towards PRR parties, depending on their labour market conditions and the structural, regulatory context. Both chapters together offer a nuanced, complex view on where working class voters are heading, and why. The second part takes on the diverging preferences for different labour market policies among workers, and how these translate into actual reforming patterns. Chapter four explores the question of who wants flexicurity, using Eurobarometer data to define insider, outsider, high- and low-skilled workers’ preferences on labour market change. Finally, chapter five puts the results into a real-world reform process, conducting a qualitative comparative case study between Spain and Italy to discover under what conditions social democracy finds a way out of dualisation. Employment protection reforms in Spain and Italy during the EMU crisis are carefully considered to offer a possible answer. Behind the Italy-Spain comparison lies the simple yet powerful notion that social democracy does not have a fixed path. Instead, in line with Rueda (2007), it should be emphasised that mainstream parties can actually choose their path. Which, in turn, should put into perspective the insights obtained from all other chapters. Since (as self-contained research projects) these are rather lousily related to each other, I devote chapter six to defining common threats, inserting some more general concluding remarks that aim to align the partial conclusions of each chapter while also defining possible avenues for further research that mostly (but not solely) relate to the idea of dynamic supply-sided actors.

18 The reconfiguration of working class politics

References Afonso, Alexandre, and Linne Rennwald (2018) 'Social Class and the Changing Welfare State Agenda of Radical Right Parties in Europe', in Philip Manow, Bruno Palier and Hannah Schwander (eds.) Welfare Democracies and Party Politics: Explaining Electoral Dynamics in Times of Changing Welfare Capitalism (171), Oxford: Oxford University Press. Alesina, Alberto, Stefano Ardagna, and Vicenzo Galasso (2008). 'The euro and structural reforms', National Bureau of Economic Research working papers, No. 14479. Arzheimer, Kai, and Elisabeth Carter (2006). 'Political opportunity structures and right-wing extremist party success', European Journal of Political Research, 45:3, 419-443. Bonoli, Giuliano (2005). 'The politics of the new social policies: providing coverage against new social risks in mature welfare states', Policy & politics, 33:3, 431-449. Bourdieu, Pierre (1986). 'The forms of capital.' In Cultural theory: An anthology: 81-93. Calmfors, Lars (2001). 'Unemployment, labor market reform, and monetary union', Journal of labor Economics, 19:2, 265-289. Carlin, Wendy and David Soskice (2008). 'German economic performance: disentangling the role of supply-side reforms, macroeconomic policy and coordinated economy institutions', Socio- Economic Review, 7:1, 67-99. Crouch, Colin and Wolfgang Streeck (eds.) (1997). Political economy of modern capitalism: Mapping convergence and diversity. New York: Sage. Croucher, Sheila (2018). Globalization and belonging: The politics of identity in a changing world. Lanham: Rowman & Littlefield. Downs, Anthony (1957). 'An economic theory of political action in a democracy'. Journal of political economy, 65:2, 135-150. Duval, Romain and Jörgen Elmeskov (2006). 'The effects of EMU on structural reforms in labour and product markets', ECB Working Paper, No. 596 Ebbinghaus, Bernhard (2005). 'Can path dependence explain institutional change? Two approaches applied to welfare state reform', MPIfG Discussion Paper, No. 05/2. Emmenegger, Patrick (2009). 'Barriers to entry: Insider/outsider politics and the political determinants of job security regulations'. Journal of European Social Policy, 19:2, 131-146. Emmenegger, Patrick, Silja Häusermann, Bruno Palier, and Martin Seeleib-Kaiser (eds.) (2012). The age of dualization: the changing face of inequality in deindustrializing societies. New York: Oxford University Press USA. Gingrich, Jane, and Silja Häusermann (2015). 'The decline of the working-class vote, the reconfiguration of the welfare support coalition and consequences for the welfare state'. Journal of European Social Policy, 25:1, 50-75. Goldthorpe, John H., Catriona Lewellyn, and Clive Payne (1980). 'Social mobility and class structure in modern Britain', The British Journal of Sociology, 32:1, 137-139 Häusermann, Silja, & Hannah Schwander (2012). 'Varieties of dualization? Labor market segmentation and insider-outsider divides across regimes'. In Emmenegger, Patrick, Silja Häusermann, Bruno Palier, and Martin Seeleib-Kaiser (eds.) (2012) The age of dualization: the changing face of inequality in deindustrializing societies (27-51), New York: Oxford University Press USA. Inglehart, Ronald F. (1971). 'The silent revolution in Europe: Intergenerational change in post- industrial societies', American political science review, 65:4, 991-1017. Inglehart, Ronald F. (2008). 'Changing values among western publics from 1970 to 2006', West european politics, 31:1-2, 130-146. King, Dave, and David Rueda (2008). 'Cheap labor: the new politics of “bread and roses” in industrial democracies', Perspectives on Politics, 6:2, 279-297. Introduction xix

Kitschelt, Herbert (1995). ‘Formation of Party Cleavages in Post-Communist Democracies: Theoretical Propositions’, Party Politics, 1:4, 447-472. Kitschelt, Herbert (2007). ‘Growth and Persistence of the Radical Right in Postindustrial Democracies: Advances and Challenges in Comparative Research’, West European Politics, 30:5, 1176-1206. Kriesi, Hanspeter, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier, Timotheos Frey (2006). ‘Globalization and the transformation of the national political space: Six European countries compared’, European Journal of Political Research, 45:6, 921-956. Kriesi, Hanspeter (ed.) (2008). West European politics in the age of globalization. Cambridge: Cambridge University Press. Kriesi, Hanspeter (ed.) (2012). Political conflict in Western Europe. Cambridge, Cambridge University Press. Kriesi, Hanspeter (2014). ‘The Populist Challenge’, West European Politics, 37:2, 361-378. Kriesi, Hanspeter and Simon Bornschier (2012). 'The populist right, the working class, and the changing face of class politics', in Jens Rydgren (ed.), Class politics and the radical right (28- 48), New York: Routledge. Lavelle, Ashley (2016). The death of social democracy: Political consequences in the 21st century, New York: Routledge. Lodovici, Manuela S. (2000). 'The dynamics of labour market reform in European countries', in Gøsta Esping-Andersen and Marino Regini, Why deregulate labour markets (30-65), Oxford: Oxford University Press. Lucassen, Geertje, and Marcel Lubbers (2012). ‘Who fears what? Explaining far-right-wing preference in Europe by distinguishing perceived cultural and economic ethnic threats’, Comparative Political Studies, 45:5, 547-574. Marx, Paul and George Picot (2013). 'The party preferences of atypical workers in Germany', Journal of European Social Policy, 23:2, 164-178. McBride, Stephen and Russell A. Williams (2001). 'Globalization, the restructuring of labour markets and policy convergence: The OECD ‘jobs strategy’'. Global Social Policy, 1:3, 281-309. Meardi, Guglielmo (2012). 'Employment relations under external pressure: Italian and Spanish reforms in 2010-12', International Labour Process Conference, Stockholm, 27-29. Mewes, Jan and Steffen Mau (2012). 'Unraveling working-class welfare chauvinism', in Stefan Svallfors (ed.), Contested welfare states: Welfare attitudes in Europe and beyond (119-57), Stanford: Stanford University Press. Mingione, Enzio (1995). 'Labour market segmentation and informal work in Southern Europe', European Urban and Regional Studies, 2:2, 121-143. Oesch, Daniel (2003). 'Labour market trends and the Goldthorpe class schema: a conceptual reassessment', Revue suisse de sociologie, 2:2, 241-262. Oesch, Daniel (2008). 'Explaining workers' support for right-wing populist parties in Western Europe: Evidence from Austria, Belgium, France, Norway, and Switzerland'. International Political Science Review, 29:3, 349-373. van Oorschot, Wim (2004). 'Balancing work and welfare: activation and flexicurity policies in The Netherlands, 1980-2000'. International Journal of Social Welfare, 13:1, 15-27. "Pablo Iglesias acusa al PSOE de preferir que siga Rajoy en La Moncloa a gobernar con Podemos [Pablo Iglesias Accuses PSOE of Preferring Rajoy as a Prime Minister over Ruling with Podemos]" (2018, February 5). Europa Press. Retrieved from http://www.europapress.es/nacional/noticia-pablo-iglesias-acusa-psoe-preferir-siga-rajoy- moncloa-gobernar-podemos-20180205001500.html (Accessed 27 June 2018). Palier, Bruno (ed.). (2010). A long goodbye to Bismarck?: the politics of welfare reforms in continental Europe. Amsterdam: Amsterdam University Press. 20 The reconfiguration of working class politics

Patton, David F. (2006). 'Germany's Left Party. PDS and the ‘Vacuum Thesis’: From regional milieu party to left alternative?', Journal of Communist Studies and Transition Politics, 22:2, 206-227. Pierson, Paul (1996). 'The new politics of the welfare state', World politics, 48:2, 143-179. Pontusson, Jonas (1995). 'Explaining the decline of European social democracy: The role of structural economic change?', World Politics, 47:4, 495-533. Prosser, Tomas (2016). 'Dualization or liberalization? Investigating precarious work in eight European countries', Work, employment and society, 30:6, 949-965. Rueda, David (2005). 'Insider-outsider politics in industrialized democracies: the challenge to social democratic parties', American Political Science Review, 99:1, 61-74. Rydgren, Jens (2007). 'The sociology of the radical right', Annual Review of Sociolology, 33, 241-262. Schumacher, Gijs and Kees van Kersbergen (2016). 'Do mainstream parties adapt to the welfare chauvinism of populist parties?', Party Politics, 22:3, 300-312. Stokes, Donald E. (1963). 'Spatial models of party competition', American political science review, 57:2, 368-377. Thelen, Kathleen (2009). 'Institutional change in advanced political economies', British Journal of Industrial Relations, 47:3, 471-498. Thelen, Kathleen (2014). Varieties of liberalization and the new politics of social solidarity. Cambridge: Cambridge University Press. "Un grand bourgeois qui ne connaît rien à la vie [A Bourgeoise who Knows Nothing about Life]”. (2017, January 4). Marianne Magazine, retrieved from https://www.marianne.net/politique/melenchon-sur-macron-un-grand-bourgeois-qui-ne-connait- rien-la-vie Viebrock, Elke and Jochen Clasen (2009). 'Flexicurity and welfare reform: a review'. Socio-Economic Review, 7:2, 305-331. Visser, Mark, Marcel Lubbers, Gerbert Kraaykamp, and Eva Jaspers (2014). 'Support for radical left ideologies in Europe', European journal of political research, 53:3, 541-558.

Where Have All the Working Class Voters Gone? xxi

Chapter 2. Where Have All the Working Class Voters Gone?

Traditionally, left-leaning parties have confronted a dilemma between trying to enlarge their coalition including middle classes through more moderate policy proposals (and therefore renouncing to revolutionary politics) and risking the support of their working class base (Przeworski 1985). Post-war social democratic parties (henceforth SDP) were to a large extent a solution to the dilemma: their platform struck an equilibrium within the left that undoubtedly renounced revolution, but did maintain substantial redistributive efforts. For at least three decades, the strategy seemed to work. SDP managed to sustain a cross-class coalition that gave the left enough electoral capacity to become all but hegemonic in Western Europe. Lately, however, SDP seem to be losing some of its power. In aggregate terms, support obtained by SDP has been decreasing steadily for the last decade and a half (Hanretty 2015). Working class voters seem to be responsible for a substantial part of these losses: According to ESS data, between 2002 and 2016 being a working class voter in Europe no longer means an increased likelihood of voting for SDP. At the same time, manual workers' support for far-right parties is clearly on the rise (Betz and Meret 2012, Bornschier and Kriesi 2012). These changes in voting patterns have led to a number of discussions in the literature on the determinants of working class vote. Both socio-demographical and attitudinal factors have been identified as culprits of this process: gender (Givens 2004, Spierings and Zaslove 2017), age, educational levels come along with attitudinal factors that, lately, have captured most of the attention. Along with the comparatively well-established determinants of institutional mistrust (Billiet and de Witte 1995; Rydrgen 2005) and preference for authoritarian leaders (Choma and Hanoch 2017) a lively debate has sparked over the possibility of voters (and especially working class voters) turning away from SDP and towards the far-right due to either economic anxiety (Dippel et al. 2015) or cultural- related preferences, particularly related to immigration (Inglehart and Norris 2016). On the supply side, other authors speculate with a move in search of a new left-authoritarian platform able to put together protectionist policies (Afonso and Rennwald 2018), possibly following a growing economic policy convergence among mainstream centre-left and centre-right parties (Abedi 2002), one that might turn itself towards the far-right (Meguid 2005; Van der Brug et al. 2005) or towards the left (Patton 2006). In trying to cut a clearer path through these views, my empirical analysis shows that the working class’s new main destination is, as a matter of fact, abstention. When compared to other class groups, this group is also significantly more likely than others to vote for the far-right, although it receives a relatively small portion of overall working class vote. More significantly, both trends have become comparatively more intense after 2008. Interestingly, this adding of intensity is less so in countries with a far-right party with parliamentary presence, suggesting that the supply-sided effect of these kind of organisations goes towards increasing (rather than decreasing) political alignment. An important exception is noted related to far- left organisations, displaying the opposite behaviour, and thus giving some ground to arguments suggesting that these two might act as each other's antidote. These findings elicit the next-stage question of what explains the shift in voters' preferences. Following the aforementioned attitudinal factors highlighted by previous studies, I use a series of logistic regressions applied to European Social Survey data (ESS) using manual and service workers as my universe of reference to find that, in essence, working class economically anxious voters are being driven towards abstention and not towards the far-right, while anti-immigration working class voters are the ones choosing the far-right. In other words: while their move towards abstention is mostly determined by mistrust in existing institutions and the feeling of economic anxiety (the latter being particularly intense, especially after 2008), the choice for the far-right appears to be more dramatically linked to preferences for anti-immigration preferences. To previous studies on far-right vote that took this same line I add a fundamental finding: the fact that the post-2008 European working class is much more determined by it. Section 1 begins by illustrating some descriptive figures of the evolution of working class vote among Europeans, and summarises a brief discussion of the relevant literature leading to my first set of hypotheses that solely focus on connections between class groups and voting decisions. Section 2 instead elaborates on the determinants of vote, adding my second strand of hypotheses to them, and section 3 describes my methodological approach to test them. Section 4 provides the results of my 22 The reconfiguration of working class politics

core analysis, while section 5 complements them with an extra exploratory analysis. Section 6 concludes by wrapping up the key insights and suggesting where and why further research may be needed.

2.1 The new faces of working class vote Beginning with a kosher approach to the links between parties and class groups, and leaving a deeper debate on the determinants of demand to be discussed in the next section, the first thing to note is that there are growing indications of a new demise of SDP under an updated version of its original dilemma, on which SDP seem to be unable to retain the majorities they held in recent decades. The phenomenon is taking place among all voters (e.g. Hanretty 2015), but it seems to be particularly intense among working class voters (e.g. Gringich and Hausermann 2015), as it is appreciable in the following table.

Table 2.1. Share of manual workers’ vote for each party family across the EU, 2002-2016.

2002 2004 2006 2008 2010 2012 2014 2016

Abstention 22.8 % 26.2 % 25.6 % 25.5 % 27.5 % 27.8 % 28.4 % 28.3 %

Far-left 3.0 % 3.8 % 2.4 % 2.7 % 2.8 % 3.3 % 3.7 % 3.2 %

Social democrats 24.5 % 21.1 % 23.0 % 22.9 % 18.1 % 20.5 % 17.6 % 15.8 %

Green/new left 1.4 % 1.1 % 0.9 % 1.6 % 1.5 % 1.0 % 1.3 % 1.4 %

Agrarian 1.5 % 1.7 % 2.0 % 0.9 % 0.4 % 1.0 % 1.0 % 2.1 %

Specific-issue 0.3 % 0.2 % 0.2 % 0.4 % 0.9 % 0.9 % 0.8 % 1.1 %

Liberals 5.6 % 4.4 % 5.1 % 4.3 % 4.5 % 4.4 % 5.6 % 3.5 %

Christian- 7.4 % 6.3 % 6.2 % 4.6 % 4.9 % 4.7 % 5.1 % 5.2 % democrats

Conservatives 8.4 % 10.0 % 10.1 % 12.2 % 14.1 % 9.4 % 9.7 % 9.5 %

Far-right 3.6 % 3.6 % 3.7 % 3.5 % 4.2 % 3.9 % 6.0 % 4.7 %

Other 21.4 % 21.7 % 20.9 % 21.5 % 21.1 % 23.0 % 20.8 % 25.2 %

Source: European Social Survey, 2002-2016

This first approach to the data comes to reiterate the hypothesis of SDP-working class estrangement. Moreover, it points to a post-2008 juncture, which would suggest a further deterioration of the relationship after the Great Recession. Hence my first hypothesis, H1a. Working class voters (both manual and service) are less likely than all other employee class groups to choose SDP. These trends have become more pronounced after 2008. When looking for where the working class is heading now, the latter trend in the literature is the far- right, which would be somewhat confirmed by data on Table 2.1. The connection that disadvantaged workers seem to have with the far-right has been widely explored indeed (Betz 1993, Mughan et al 2003). Bornschier and Kriesi (2012) offer the most compelling example of pinning down this relationship to specific class groups (skilled production workers, routine manual and non-skilled service workers) using 2010 pan-European data from the European Social Survey, noting that the root of these in the extreme right is quite a robust finding. These findings nicely match with Ares' (2017) recent account of political similarities between manual and non-manual members of what we could Where Have All the Working Class Voters Gone? xxiii

label as the 'wider' working class. In consequence, I henceforth consider the working class to be formed by these groups. There are as well strong indications of the Great Recession being an activator or at the very least a catalyst of far-right support (Bermeo and Bartels 2013, Kriesi and Pappas 2015). Preliminary analyses bring some evidence of an increase in support for far-right parties in the post-crisis period (Funke and Trebesch 2017). Hence my next hypothesis, that only intends to bring to a broader test what has been already partially proven. H1b. Working class voters (both manual and service) are more likely than all other class groups to choose the far-right. These trends have become more pronounced after 2008. Borschner and Kriesi choose to focus their analysis only where the far-right has a relevant presence, and rightly so, since their interest lies primarily in these parties. Mine is, however, broader. And Table 2.1 points to a trend that is clearest among all: a growing preference for abstention. The 'class bias' on turnout has been widely noted by the literature, and mostly proxied through education (e.g. Verba and Nie 1972, Wolfinger and Rosenstone 1980, Leighley and Nagler 1992, Lijphart 1997). Studies focused on more complete and up-to-date class group classification are substantially less numerous. Oesch (2006) constitutes an exceptional departing point, since it not only updates the class schema to the current world of work, but tests its political implications and finds negative effects on turnout for at least Sweden and Switzerland. There is no readily available reason not to expect its wider generalisation, especially after 2008: it would make sense to expect the Great Recession to have a negative effect on working class turnout, through either sociotropic (Kinder and Kiewelt 1981) or pocketbook (Rosenstone 1982, Carreras and Castañeda 2016) mechanisms. Therefore, and taking the preliminary evidence from Table 2.1 into a harder test, I would expect: H1c. Working class voters (both manual and service) are more likely than all other class groups to abstain. These trends have become more pronounced after 2008. H1d. Working class voters (both manual and service) are more likely to abstain than to choose to vote any party. These trends have become more pronounced after 2008. Another fact stemming from Table 2.1 is the trend towards the far-left. Historically, the link between the working class and more leftist parties has been considered to be a necessary part of the SDP dilemma, in virtue of which (particularly in hard times) the most disadvantaged workers always had an 'exit' option ready should they have wanted to find more radical positions. Even more: on the supply side, that SDP have moved too far to the centre on economic issues has become a commonplace idea. Such is the argument often made by newly successful far-left parties such as Syriza in Greece, Bloco d'Esquerda in Portugal, or Podemos in Spain. Challengers to mainstream SDP leaders within their own parties have also sustained this argument (e.g. Jeremy Corbyn in the UK's Labour Party or Pedro Sánchez within the Spanish PSOE). The idea of a 'neoliberal social democracy' strongly resonates with the partisan convergence hypothesis, which states that when mainstream parties converge on policy proposals and government, they open up the space for the far-right (Meguid 2005; Van der Brug et al. 2005) or for the far-left (e.g. Patton 2006). More decisively, Spies and Franzmann (2011) find that less polarisation (that is, more convergence) on economic-related issues opens up the space for alternative parties. Indeed, if the turn away from classic social democratic parties is mostly due to socio-economic reasons and following the standard assumptions of class voting (e.g. Lipset 1960), it would make sense to observe the working class vote shifting further to the left. As a matter of fact, a new pattern of working class voting for non-SDP left parties may stem from economic concerns. According to Achterberg and Houtman (2006), following Lipset (1959), blue collar workers are comfortable voting for old far-left parties because they offer a combination of cultural authoritarianism and economic interventionism. However, during the Great Recession, a new class of left parties emerged. There are some indications of these parties absorbing part of the working-class vote. For instance, in the 2012 Greek general election, working class neighbourhoods were more likely to yield good results for SYRIZA (Voulgaris & Nikolakopoulos 2014). The January 2015 elections saw this pattern repeated four years on (Tsirbas 2016). There is, however, little proof of a clear or stronger link between the far-left and the working class. On the contrary, most recent works have found that it is not the working class, but rather upscale professionals as well as socio-cultural workers (e.g. Oesch 2008) who choose the most progressive 24 The reconfiguration of working class politics

options, particularly those ascribed to the so-called New Left that display culturally liberal positions. The link with the working class has been instead hypothesised under the condition of the aforementioned left-authoritarian platform. Afonso and Rennwald (2018) trace it among certain recent instances of far-right strategies, but it could be argued as well that the more traditional far-left is at least partially fitting for the working class searching new alternatives. Since, according to Table 2.1, that does not seem to be the case, it would be helpful to find out why. A possible alternative stems from the realisation that, although neither far-right parties nor the far-left are perfect left-authoritarian platforms, there seems to be arguments in favour of both being at least potentially able to serve as a second-best for those who seek an alternative of the sort. Pablo Iglesias, Podemos' leader, has said as much, when he asserts that "there is no far-right party in Spain thanks to Podemos" (Machuca 2016). One could take this argument, reverse it, and test it to explain the apparent lack of aggregate interest on far-left alternatives: it might be the case that there where a far- right alternative is available, working class voters are less prone to choosing far-left options. Thus, my next hypothesis: H1e. If there is no far-right party with parliamentary representation, the probability of the working class voting for the far-left increases after 2008. If there is, it decreases. Following this supply-sided analysis, one could wonder that the other two expected trends (higher working class abstention, and lower voting for SDP) would be equally affected by the presence/absence of a far-right party, i.e. it would make sense to foresee that wherever there is a relevant far-right competitor, the SDP and turnout would see their presences comparatively further diminished. However, the opposite guess could be equally justified. At the end of the day, according to some accounts that portray the SDP as the parties more clearly associated with the existing policy equilibrium (and the far-right as its clearest antagonist), it would make sense to expect the presence of the far-right to decrease, rather than increase, the fall of SDP. Regarding turnout, a growing capacity of the far-right to dominate the public debate might help expelling faster certain workers disenchanted with the absence of other alternatives. Or it could be that there is some other hidden (structural) variable that explains simultaneously the rise of the far-right and of abstention, as well as a slower decline curve for SDP. Based on these guesses, I hold a (weak) expectation of the following: H1f. If there is no far-right party with parliamentary representation, the probability of the working class not voting or voting for SDP decreases to a larger extent after 2008 than if there is. H1g. The probability of abstaining increases further after 2008 in presence of a far-right party than it does in absence of it.

2.2 Why these changes? The question of why the working class would abandon its traditional political home for new options, and particularly towards the far-right or away from voting, has sparked a great deal of conversation and controversy that goes well beyond the supply-sided debates and has a particular focus on demand-sided determinants. Across Europe, SDP leaders and base members were engaged in intense debates, especially after the Great Recession and beyond: the crisis cycle that began with the Lehman Brothers crash in 2008 and which took a particularly high toll on European mid- and low- income workers. These debates are being matched at the academic level, where three key demand- sided factors have been identified both in the far-right and in the non-voting strands of the literature: mistrust towards existing institutions, the effect of economic hardship (normally dubbed as economic anxiety), and identitarian aspects (or, more specifically, preferences on immigration), with the latter more usually associated with far-right voting patterns. I take them on to extract empirically testable hypotheses applied to my question of interest: why the working class is choosing abstention and the far-right, why is it being driven away from SDP, and why these trends have become more pronounced after the Great Recession?

2.2.1 A matter of mistrust? If we see the party system as an oligopolistic equilibrium between supply and demand, volatility (i.e. change in votes outside the existing equilibrium; for a complete literature review on the issue see Powell and Tucker 2012) is by definition the process of abandoning mainstream parties towards new options, or towards no option at all. It thus makes sense to expect those voters exiting the existing equilibrium to be critical of the institutions that support it. Such a link has been documented for the far- Where Have All the Working Class Voters Gone? xxv

right (Billiet and de Witte 1995; Rydrgen 2005) and for the far-left (Lubbers and Scheepers 2007), as well as for lower turnout (Cox 2003, Grönlund and Setälä 2007). These insights might very well apply to the working class, particularly through the breakdown of class mobilization by traditional parties and the subsequent lower turnout (Gray and Caul 2000). But also, to the choice of far-right parties, as the work of Pippa Norris has shown (Norris 2011, Norris and Inglehart 2016). Although, to the best of my knowledge, there is no specific test of this factor applied to working class voters, there appears to be no reason to expect otherwise. Moreover, since it could be said that SDP epitomises the traditional partisan institution among these voting segments, it would make sense to expect support for it to correlate with trust in political institutions. Regarding the effect of the Great Recession, there are two (contradictory) ways to anticipate it. There is an extensive body of literature pointing to the effect of economic crises on all versions of political trust. To point to a recent and comprehensive example, Van Erkel and Van Der Meer (2016) conduct a multilevel study around the Great Recession, finding not only that macroeconomic performance affects political trust, but also that it does so to a larger extent among the less educated. However, the present analysis does not consider trust and the shock to be a dependent plus an independent variable, but instead they are seen as co-independent explaining working class vote decisions. In that sense, they 'compete' with other factors, and it would make sense to expect some of them (in particular, the following two featured in this section) to convey more explanatory power. Thus, although I do expect (mis)trust to play a role, I don't expect it to necessarily increase after 2008. H2a. Mistrust towards political institutions drives working class members to non-voting and to the far- right, while trust leads them to the SDP. However, its effect is expected to be either constant or diminished after 2008.

2.2.2 Economic anxiety vs identity? The displacement towards abstention of disenchanted voters whose perceived economic situation does not correspond with any of the dominant policy platforms is relatively unambiguous, and has been documented, demonstrating that both party convergence (Thurner and Eymann 2000; Kurella 2017) economic hardship (Rosenstone 1982) and economic judgement (Tillman 2007) affect turnout. This is a relatively well-established fact, and there is no reason not to expect it to reproduce itself among the working class. On the contrary, given that the spread within this category has been arguably growing in recent times, it would make sense to expect that a subjective, individual measure of economic anxiety captures part of this variance, particularly after the Great Recession and its possibly uneven effect among workers. Given this accumulated evidence, it is possible that, due to academia's bias of prioritizing the search for an explanation to puzzling outcomes, most of the most recent literature looking for an economic- based shift in working class vote has focused on the far-right. The idea of ‘losers of globalisation’ was advanced by Kriesi et al. (2006): globalisation provokes the emergence of a separation between winners and losers from economic and cultural threats. The concept of globalisation's 'winners and losers' would become material for political entrepreneurs ready to build a protectionist coalition. The links between the notion of ‘losers’ and the working class come from the fact that these segments are the ones who have potentially more to lose from the economic forces driving globalisation. Arzheimer and Carter (2006), for instance, offer an early, positive test on how being a manual worker might be driving the vote to the far-right, as the political wing that holds the most clearly anti- globalisation positions. Dippel et al. (2015) accurately links trade shocks, deregulation & voting behaviour in Germany to find that trade exposure in the labour market drives low-skill manufacturing voters towards the populist political right. Outside of these works, the fact that working class and far-right voters are linked through economic factors (and, more generally, the idea of connecting economic anxiety with the far-right vote) is contested by those who assert instead that the roots of the link between these issues stem from identitarian factors. Lipset (1959) argued early on that low socio-economic status and low educational attainment led to "extremist, intolerant and transvaluational forms of political behaviour" (Lipset 1959:1). Back in the 1950s and 1960s, communist parties fulfilled this demand for a left-authoritarian alternative among workers. At the same time, the same logic of party convergence mentioned in regard to the economic axis could be applied to identitarian issues to explain the decision not to vote. Thurner and Eymann (2000) 26 The reconfiguration of working class politics

demonstrated that, at least for Germany, immigration could be a policy issue powerful enough to drive voters to or away from existing party options. Research linking working class identity issues with voting for the far-right are currently more prominent than those following the classical Lipset route. This literature has mostly been constructed as a battle between socio-economic and cultural factors in order to prove the reigning importance of the latter in explaining far-right vote (e.g. Lucassen & Lubbers 2012; Borschnier and Kriesi 2012; Oesch 2013). Hainmueller & Hopkins (2014) conducted a literature review that assessed that sociotropic cultural worries are the main driver towards anti-immigration attitudes. Norris & Inglehart (2016) comes across as an attempt to settle the question of the influence of sociotropic cultural worries. It uses multinomial logistic regression applied to a pooled dataset of all ESS rounds to estimate that cultural grievances, and particularly those linked to immigration, are more important than socio-economic factors on defining far-right vote. As such, it stands in stark contradiction with the most stripped-down versions of the 'losers of globalisation' and ‘economic anxiety as driver towards the far-right’ hypotheses, against which is gaining ground. There is much less insight coming from existing studies on how this factor might have evolved over time. It makes sense, however, to expect a deepening of its importance due to two reasons. One is exogenous: The Great Recession might have increased the sensation of being a 'loser of cultural modernisation'. Another one, endogenous: since the most distinctive trait of the European far-right's policy platform is anti-immigration, it would make sense to expect its electorate to grow more convinced of these positions. To sum up, the 'economy vs identity' debate as it stands should lead us to expect: (1) identitarian (particularly immigration-related) factors playing a bigger role on explaining far-right vote; and (2) economic anxiety being able to better explain displacement towards non-voting. Both trends would have deepened after 2008 due to the shock-like nature of the Great Recession. Hence the following hypothesis. H2b. Economic anxiety drives working class members to non-voting, while the contrary leads them to the SDP. Its effect is expected to be either constant or increased after 2008. H2c. Anti-immigration sentiment drives working class to the far-right, while pro-immigration sentiment leads them to the SDP. Significantly, its effect is expected to be either constant or diminished after 2008.

2.3 Analysis: setup and results The literature offers us interesting clues on where the working-class realignment is headed in the 21st century as well as an explanation why. However, it does so in a disjointed manner. To the best of my knowledge, a joint consideration of the 'where' question has received less attention; there has been even less exploration of a combined test that attempts to give a complex, nuanced and multi-polar answer to the 'why' and 'where' questions simultaneously. In my understanding, there should be three pillars of such a test. First, the disjointed character of existing evidence shows that a multi-dimensional perspective is needed to fully understand the changing patterns of working class voters. Abstention should be placed at the forefront of these considerations, due to its apparently growing capacity to inhibit workers from voting. Second, the attempt has to be comprehensive. The question under consideration is inherently European. Single country analyses are more nuanced, but less generalizable; single year tests are similar. Both types of analyses make it much harder to find common European patterns. A multi-year, pan-European analysis restricted to the manual working class is not only recommended to find these patterns, but it is also much less common in the literature. Third, some sort of time-based perspective must be included. I have pointed to the Great Recession as a reasonably significant anchoring point as its massive economic, social and political consequences were felt in party systems across the continent.

2.3.1 Measuring working class vote propensity To test my first set of hypotheses I pool all ESS waves (2002-2016), featuring all EU countries plus Norway and Switzerland while removing Croatia and Cyprus for their very particular, historically- related party systems that allow for little (if any) cross-country comparison with the rest of the EU Where Have All the Working Class Voters Gone? xxvii

members1. Unlike past similar works, such as Borschnier and Kriesi (2012), I do not exclude countries with absence of far-right parties since I am interested on more than one voting option. I run a series of logistic regressions aimed at identifying which class groups are more likely or less likely to support certain parties. In this, I choose the approach of Inglehart and Norris (2016) that somehow follows (and re-sets) the standard for the field. Borschnier and Kriesi opt for a two-stage Heckman selection model, but the way I am considering choices is better portrayed as binomial, since I do not contemplate turnout as a separate, first-degree choice as Borschnier and Kriesi do. I should reiterate that my goal is not only to rank party support within classes, but to rank class support within party families as well. By taking this approach I intend to minimise obvious structural and supply-sided factors that might confuse the analysis: the fact that a worker is more likely to choose a SDP over other parties does not necessarily reflect class effect, but rather that SDP remain large enough across Europe as to coordinate the vote from different class groups. Returning to my model, the dependent variable is therefore individual vote choice re-classified by party family in order to conduct a multi-country analysis. To define which parties are part of a specific family I use the Chapel Hill expert surveys (CHES) criteria. This framework recognises nine different families and places all European parties within them, according to the opinion of several leading experts (CHES 2016). The original families are: • com (far-left) • soc (centre-left) • eco (green) • agr (agrarian/rural) • spec (specific issue) • lib (liberals) • chr (Christian Democrats) • con (conservatives) • right (far-right) Apart from non-voting, there are three families of interest for my research purposes: com - soc - right. Table A.2 in the Appendix summarises which parties are included within each family of interest (far- left, social democrats, far-right), as well as the number of available observations per family and country. Still, I run this first analysis for all party families except for agrarian and specific issue parties, due to the regional variability and small size of these two families. Each model is run twice: once for the 2002-2008 period, and then another time for the 2010-2016 period. Using a dummy control would yield nearly identical effects. My key independent variable (IV) in this first stage is class belonging. Among the many available options to operationalise it, I choose Oesch (2006). Its eight-level class schema has proven solid and reliable. It offers a clear-cut distinction of manual workers that relies on a nuanced consideration of the NACE standard (the Statistical Classification of Economic Activities in the European Community). Table A.3 of the Appendix gives the number of observations per class group for each country included in the sample. Along with class there is a set of relatively conventional control variables: • gender: as a general rule, men are more likely to vote to the far-right (Givens 2004, Spierings and Zaslove 2017) and to the far-left (Inglehart and Norris 2000) • age • Whether the person was born in the country where she currently resides or not. • Finally, I introduce fixed effect controls for the different ESS waves (2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016) and countries featured in the analysis.

1 Austria, Belgium, Bulgaria, Switzerland, Czech Republic, Germany, Denmark, Estonia, Spain, Finland, France, United Kingdom, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Norway, Poland, Portugal, Sweden, Slovenia, Slovakia 28 The reconfiguration of working class politics

Table 2.2. Marginal probabilities of voting for far-left, social democratic and far-right parties depending on class group.

Far left SDP Far right Abstention

2002- 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2008

Employers and 0.0260 0.0251 0.1880 0.1390 0.0198 0.0378 0.1200 0.1300 managers -0.0021 -0.0046 -0.0119 -0.0121 -0.0033 -0.0061 -0.0076 -0.0059

Small business 0.0214 0.0233 0.1670 0.1250 0.0405 0.0649 0.1940 0.2100 owners -0.0010 -0.0012 -0.0067 -0.0072 -0.0022 -0.0020 -0.0067 -0.0058

Technical (semi-) 0.0324 0.0371 0.2950 0.2100 0.0309 0.0619 0.1250 0.1580 professionals -0.0018 -0.0024 -0.0159 -0.0069 -0.0029 -0.0042 -0.0082 -0.0063

Production workers 0.0336 0.0349 0.2460 0.1860 0.0410 0.0718 0.2550 0.2790

-0.0021 -0.0024 -0.0095 -0.0076 -0.0038 -0.0060 -0.0082 -0.0065

(Associate) managers 0.0301 0.0333 0.2530 0.2060 0.0332 0.0505 0.1360 0.1340

-0.0026 -0.0013 -0.0121 -0.0105 -0.0026 -0.0048 -0.0123 -0.0076

Clerks 0.0328 0.0298 0.2370 0.2130 0.0406 0.0550 0.1700 0.1810

-0.0023 -0.0026 -0.0099 -0.0100 -0.0015 -0.0028 -0.0034 -0.0036

Socio-cultural (semi-) 0.0492 0.0521 0.2820 0.2250 0.0266 0.0352 0.1180 0.1340 professionals -0.0029 -0.0031 -0.0115 -0.0120 -0.0054 -0.0071 -0.0036 -0.0043

Service workers 0.0316 0.0359 0.2470 0.1900 0.0393 0.0689 0.2250 0.2490

-0.0019 -0.0010 -0.0071 -0.0056 -0.0019 -0.0045 -0.0059 -0.0075 Full results on Table A.5, based on logistic regression model from Table A.4. All results are statistically significant at p>.01.

The results of the logistic regressions displayed in Table A.4 of the Appendix confirms H1a, H1b, H1c, H1d. Coefficients for production workers are significant at p=0.00. More tellingly, marginal effects (Table 2.2 above) demonstrate that the manual and service working classes are (1) more likely to abstain than to do anything else, and is also the most likely group to abstain; (2) production workers are only exceeded by the self-employed and the business-owners in their tendency to vote for the social democrats, and they display a lower likelihood of choosing other party families as well except for the far-right; (3) manual workers are the most likely to choose far-right parties. Moreover, all three results have become more pronounced in the 2010-2016 period. Abstention is the most likely destination for manual workers, and manual workers are the most likely class group to abstain. At the same time, although workers are still more likely to vote for dominant parties (including SDP) than to vote far-left or far-right, belonging to the manual working class makes an individual more likely to choose the extreme right (and less likely to opt for SDP) than other individuals. Therefore, while different classes tend to vote, approximately, for the same parties (i.e. class is not a huge discriminant of the vote, which might be seen as a proof of party convergence as well), class belonging remains a statistically significant factor to determine the probability of each individual to support a certain type of party. Both production and service workers display fairly similar vote tendencies. Moving now to check the effect of far-right presence, I run again the same model structure but now only for abs, soc and com as DVs, and I divide countries into two groups: those with a far-right party that holds parliamentary presence at the time of running the survey, and those without. I therefore Where Have All the Working Class Voters Gone? xxix

have a 2x2x3 model set, dividing by period, far-right presence and the actual DV. Table 2.3 displays the key results.

Table 2.3. Marginal probabilities of voting for far-left, social democratic and non-voting depending on class group and on the presence/absence of a far-right party

Far left SDP Abstention

No far right Far right No far right Far right No far right Far right

2002- 2010- 2002- 2010- 2002- 2010- 2002- 2010- 2002- 2010- 2002- 2010- 2008 2016 2008 2016 2008 2016 2008 2016 2008 2016 2008 2016

Employers and 0.026 0.045 0.027 0.014 0.220 0.154 0.164 0.129 0.099 0.104 0.136 0.145 managers -0.003 -0.010 -0.002 -0.003 -0.012 -0.003 -0.020 -0.020 -0.007 -0.004 -0.011 -0.009

0.023 0.036 Small business 0.021 0.016 0.196 0.140 0.147 0.116 0.180 0.177 0.203 0.228 owners -0.002 -0.003 -0.001 -0.001 -0.011 -0.013 -0.009 -0.009 -0.017 -0.015 -0.004 -0.004

0.038 0.062 Technical (semi-) 0.028 0.023 0.354 0.236 0.255 0.196 0.097 0.125 0.143 0.176 professionals -0.004 -0.006 -0.001 -0.001 -0.019 -0.010 -0.016 -0.009 -0.012 -0.009 -0.005 -0.007

0.036 0.056 Production 0.032 0.023 0.280 0.207 0.225 0.174 0.243 0.248 0.261 0.296 workers -0.001 -0.005 -0.004 -0.002 -0.007 -0.016 -0.017 -0.008 -0.012 -0.013 -0.008 -0.007

(Associate) 0.040 0.050 0.023 0.023 0.296 0.226 0.225 0.194 0.096 0.105 0.159 0.150 managers -0.001 -0.002 -0.003 -0.002 -0.016 -0.024 -0.015 -0.009 -0.013 -0.011 -0.011 -0.008

0.033 0.050 Clerks 0.032 0.018 0.272 0.260 0.214 0.187 0.153 0.162 0.182 0.193

-0.003 -0.007 -0.002 -0.003 -0.007 -0.012 -0.014 -0.004 -0.005 -0.008 -0.005 -0.004

0.057 0.076 Socio-cultural 0.043 0.038 0.313 0.222 0.259 0.225 0.101 0.122 0.129 0.140 (semi-) professionals -0.003 -0.007 -0.004 -0.003 -0.007 -0.024 -0.018 -0.012 -0.006 -0.004 -0.004 -0.007

0.036 0.057 Service workers 0.028 0.024 0.270 0.208 0.232 0.180 0.213 0.225 0.233 0.261

-0.001 -0.002 -0.003 -0.001 -0.010 -0.012 -0.007 -0.007 -0.012 -0.009 -0.003 -0.010 Full results on Table A.5, based on logistic regression model from Table A.4. All results are statistically significant at p>.01.

In essence, H1e and H1f are confirmed: first, far-left parties grow more in absence of a far-right rival. Second, non-voting increases substantially more in places where the far-right is present. Third, SDP decrease faster if there is no far-right party available (which might have to do with the fact that the far- left is stronger there as well). Since all three phenomena take place not only among my class groups of interest but also among all other segments, joint structural factors should be considered before drawing conclusions solely based on supply-sided considerations. However, the evidence is nonetheless significant.

2.3.2 Explaining working class vote choices In order to move on to the next set of hypotheses, I need to reduce my analysis to a subset of our original ESS data pool. Thus, I return to Oesch's class schema and use the ‘manual production workers’ and 'service workers' categories as a filter. All following analyses are therefore restricted to this group, that I equate to the broad working class (to see a study of similarities, check Ares 2017). Thanks to the massive size of the dataset, sample sizes remain large and solid. My dependent variables are once again soc, right, abs. I also reuse the basic controls from my previous logistic regression: gender, age, and whether the voter was born in her current country of 30 The reconfiguration of working class politics

residence or not. I add a basic control for education attainment, another to check the preference for strong leaders to take authoritarianism sentiment into account (following the growing body of literature that points to it as a factor for vote choices, particularly among non-college voters, e.g. Choma and Hanoch 2017) as well as the key independent variables. Following my previously stated hypotheses, there are three sets of variables to test, all of them normalised into a 0-1 scale in order to obtain smoother, more readable results. (1) Trust in political institutions (0=no trust; 1=full trust). As I do in the next chapter, I proceed by normalizing factor loadings of three ESS questions about how much respondents trust parties, politicians and the national parliament. (2) Economic anxiety (0=low; 1=high). I introduce the answers to the ESS question “Which of the descriptions on this card comes closest to how you feel about your household's income nowadays?”. The question has four possible answers: the respondent might be - living comfortably on present income (=0), - coping on present income (=.33), - finding it difficult on present income (=.66), or - finding it very difficult on present income (=1). (3) Views on immigration (0=strongly for; 1=strongly against). Following Polavieja (2016), I factor and then cluster all immigration questions common to all ESS waves. These are: a. Allow many/few immigrants of same race/ethnic group as majority, ranging from “allow many” to “allow none”. b. Allow many/few immigrants of different race/ethnic group as majority, ranging from “allow many” to “allow none”. c. Allow many/few immigrants from poorer countries outside Europe, ranging from “allow many” to “allow none”. d. Immigration bad or good for the country’s economy, on a ten-point scale. e. Cultural life of the country is undermined or enriched by immigration, in a ten-point scale. f. Immigrants make country a better or worse place to live, on a ten-point scale. As before, I run each one of the three models twice: once for the 2002-2008 period, and again for the 2010-2016 period. Putting both together gives a sense of the general trend and at the same time a sense of the effect of each factor of interest, as well as its change before and after the Great Recession. Before moving on, I should reiterate that my analysis is not actually capable to show shift of voters’ decisions, but rather aggregate trends over time. Table A.6 in the Appendix displays the regressions’ results. However, I will limit my considerations in this section to marginal probabilities, which is a more intuitive and adequate way to test our hypotheses. To begin with, it seems quite apparent from Figure 2.1 that mistrust in political institutions draws voters away from SDP and towards non-voting: The least trusting working class voters are substantially less likely to vote for them than the most trusting ones. Largely, these voters clearly move towards abstention.

Where Have All the Working Class Voters Gone? xxxi

Figure 2.1. Marginal probabilities - choosing among parties for working class voters depending on trust in political institutions(0: very low trust - 1: very high trust)

Full results on Table A.7, based on logistic regression model from Table A.6.

It is also remarkable not only that distrust is a driver towards the far-left and the far-right, as expected, but that the slope of the curve for both (and particularly for the SDP) is less steep after 2008, indicating that the effect has decreased, and has nearly disappeared for the far-right. It only retains a 32 The reconfiguration of working class politics

strong power for non-voting. After the crisis, working class voters further duplicate their probability of abstaining if they hold a very low level of trust towards the country’s institutions. The second factor considered is economic anxiety. First, it should be said that its effect for the far-right is essentially irrelevant according to Figure 2.2. Instead, its clearest effect goes for abstention: working class voters going through economic hardships move toward non-voting, especially after the crisis.

Figure 2.2. Marginal probabilities - choosing among parties for working class voters depending on income pressure (0: no pressure at all - 1: high pressures)

Full results on Table A.7, based on logistic regression models from Table A.6. Where Have All the Working Class Voters Gone? xxxiii

Non-voters are also driven by anti-immigration sentiment to a significant extent, but clearly the steepest effect corresponds to the working class move towards the far-right. The extent to which anti- immigration views affects it is striking, particularly after 2008, when it turns into a five-fold increase across the whole curve that becomes more intense as we consider more extreme positions.

Figure 2.3. Marginal probabilities of choosing among parties for working class voters depending on preferences regarding immigration (0: very much for - 1: very much against)

Full results on Table A.7, based on logistic regression models from Table A.6. 34 The reconfiguration of working class politics

It is equally significant how pro-immigration views have increased its effect as a driver of SDP vote after 2008, highlighting the increased relevance for the political battle within the working class that this particular issue has acquired over the last decade and a half.

2.3.3 Cross-factor comparison and summing up results Table 2.4 summarises the results obtained. Overall, my core hypotheses seem to be substantially confirmed. Mistrust in institutions is a common driver for manual workers abandoning SDP and choosing both the far-right and not voting, albeit with decreased effect after 2008. Aside from mistrust, non-voters display economic anxiety and an inclination towards anti-immigration views, the latter trait being shared with far-right parties, again with steeper effects since the Great Recession.

Table 2.4. Summarizing the results for voting choices of working class in Europe: direction, significance and change between (2002-2008) and (2010-2016). Trusting institutions Being anti- Suffering economic immigration anxiety SDP Positive. Decreased. Negative. Increased. Change from slightly positive to slightly negative. Far-right Negative. Decreased. Positive. Increased. No clear relationship. Abstention Negative. Decreased. Positive. Decreased. Negative. Increased

Although the focus of the present work is not necessarily to pitch these factors against each other, checking which ones bear the highest degree of explanatory power is nonetheless informative.

Table 2.5. Most sizeable factors behind vote choices for working class in Europe.

2002-2008 2010-2016

Choosing SDP over the rest Trust in institutions Views on immigration

Choosing far-right over Views on immigration Views on immigration SDP

Abstaining over the rest Trust in institutions Trust in institutions

Immigration-related issues seem to have become dominant among working class voters. Views on immigration helps to maintain voters within SDP but it also takes them away, towards far-right parties; trust/mistrust determines the move to abstention, along with economic anxiety.

2.4 Interpreting and concluding The present work is an attempt to map out the political demand of the European working class in the present century. Drawing from the existing literature, I have restricted my 'where' question to abstention as well as to three party families of particular interest for this class group: the far-left, the social democrats (SDP) and the far-right. Also based on previous works, the 'why' question has been restricted here to three key dimensions: trust in political institutions, economic anxiety, and sentiment towards migration. Unlike those previous works, mine attempts to (1) enlarge the scope; (2) focus on the (broad) working class; (3) use larger samples that make it possible to introduce a time dimension in the analysis. The results lead us to four key conclusions. First, abstention is the political position that has absorbed the majority of working-class vote, especially in recent years. It is also remarkable that status as a manual worker increases the Where Have All the Working Class Voters Gone? xxxv

probability of voting for both the far-right and the far-left compared to belonging to most other class groups (but not all of them). However, in absolute terms, abstention beats all other options by a substantial margin. Additionally, all these trends have become more acute after the Great Recession. Moreover, the presence/absence of viable far-right platforms might be having an impact on the evolution across time of turnout (negative), the performance of far-left options (negative) and SDP (positive). Second, the drivers of the three key choices (non-voting, far-right, SDP) associated with the working class display an interesting pattern. Workers turning away from SDP and towards the far-right share with non-voters the mistrust towards political institutions, but not feelings of economic anxiety, which greatly affect the decision to turn out and vote but not to do so for the far-right. And although there does not seem to be any strong indication of economic anxiety being a significant driver, there is a very clear indication of anti-immigration being the key factor of choosing the far-right. Most interestingly, those choosing abstention also share the trait, while SDP working class vote is positively affected with pro-immigration views. These results are substantially aligned with existing research for both all voters (Inglehart and Norris 2016) and working-class voters (Borschnier and Kriesi 2012). However, now I am able to confirm it for a broader period of time and a larger number of countries. More significantly, I can trace some time-related evolution. Because, and this is the third key finding, these effects have become more intense since 2008, pointing to a possible ‘crisis effect’. Fourth, while working class voters who remain with SDP trust mainstream political institutions and are, in general, pro-immigration rights (thus have a culturally liberal preference set), none of these factors bears an overwhelming degree of intensity, therefore suggesting that SDP working class vote is not overwhelmingly ideological, with the possible exception of an upward trend of the importance of immigration-related perspectives. What (if anything), motivates workers to remain within social democracy is something that requires further research, perhaps following the structural divisions within the working class (Rovny and Rovny 2017 is a relevant first step in this direction). It appears as though classical right-leaning preferences on economic and social policy (i.e. being anti- redistribution and anti-immigration) move workers towards the far-right and towards abstention, while economic anxiety seems to be determinant for the growing trend against turnout. At the same time, mistrust in existing institutions takes working class voters in all directions (but mostly towards abstention). Among these, the far-right would attract working class members who are warier of immigration, with these platforms owning the issue. Meanwhile, economically pressed workers would continue to abstain, and the far-left would be more effective in its capacity to grow among these segments there where the far-right is missing. Would this suggest that, wherever the far-right is present, the identitarian-related dimension is taking over the political conflict (particularly after 2008) while economically pressed workers are being left at the margins? The evidence shown here does not go far enough to support this claim, but it certainly suggests that this is a hypothesis worth exploring in further research. 36 The reconfiguration of working class politics

References Abedi, Arim. "Challenges to established parties: The effects of party system features on the electoral fortunes of anti-political-establishment parties." European Journal of Political Research 41.4 (2002): 551-583. Achterberg, Peter, and Dick Houtman (2006). "Why do so many people vote ‘unnaturally’? A cultural explanation for voting behaviour." European Journal of Political Research 45.1: 75-92. Afonso, Alexandre, and Linne Rennwald (2018) 'Social Class and the Changing Welfare State Agenda of Radical Right Parties in Europe', in Philip Manow, Bruno Palier and Hannah Schwander (eds.) Welfare Democracies and Party Politics: Explaining Electoral Dynamics in Times of Changing Welfare Capitalism (171), Oxford: Oxford University Press. Aguilar, Edwin E. and Alexander C. Pacek (2000). 'Macroeconomic conditions, voter turnout, and the working-class/economically disadvantaged party vote in developing countries'. Comparative Political Studies, 33(8), 995-1017. Arzheimer, Kai (2009a). ‘Contextual factors and the extreme right vote in Western Europe, 1980- 2002’, American Journal of Political Science, 53:2, 259-275. Arzheimer, Kai (2009b). ‘Contextual Factors and the Extreme Right Vote in Western Europe, 1980- 2002’, American Journal of Political Science, 53:2, 259-275. Arzheimer, Kai (2013). 'Working-class parties 2.0? competition between centre left and extreme right parties', retrieved from http://www.kai-arzheimer.com/working-class-parties-extreme- right.pdf.pagespeed.ce.KYZzrQNtTJ.pdf (Accessed June 29, 2017). Arzheimer, Kai (2018). 'Explaining electoral support for the radical right'. The Oxford Handbook of the Radical Right, 143. Arzheimer, Kai, and Elisabeth Carter (2006). ‘Political opportunity structures and right-wing extremist party success’, European Journal of Political Research, 45:3, 419-443. Bakker, Bert N., Matthijs Rooduijn, and Gijs Schumacher (2015). ‘Following the leader or following your ideology? The case of populist radical right voting’, Unpublished article, http://www.bertbakker.com/wp- content/uploads/2015/01/Bakker_Rooduijn_Schumacher_PopulistVoting.pdf (accessed June 29, 2017). Bartels, Larry and Nancy Bermeo (Eds.). (2013). Mass politics in tough times: opinions, votes and protest in the Great recession. Oxford: Oxford University Press. Betz, Hans-Georg, and Carol Johnson (2004). ‘Against the current—stemming the tide: the nostalgic ideology of the contemporary radical populist right’, Journal of Political Ideologies, 9:3, 311-327. Betz, Hans-Georg, and Susi Meret (2012). ‘Right wing populist parties and the working class vote: what have you done for us lately?' on Rydgren, Jens, ed. Class politics and the radical right. Routledge, 2012. Biggs, M, and S Knauss (2012). ‘Explaining Membership in the British National Party: A Multilevel Analysis of Contact and Threat’, European Sociological Review, 28:5, 633-646. Billiet, Jaak & Hans de Witte. (1995). Attitudinal dispositions to vote for a new extreme right- wing party: The case of the ‘Vlaams Blok’. European Journal of Political Research, 27, 2, 181-202. Bornschier, Simon, and Hanspeter Kriesi (2012). ‘The populist right, the working class, and the changing face of class politics’, Class politics and the radical right, 10-29. Burgoon, Brian (2013). ‘Inequality and anti-globalization backlash by political parties’, European Union Politics, 14:3, 408-435. Carreras, Miguel y Néstor Castaneda (2016). 'Economic Adversity and Voter Turnout: A Multilevel Analysis'. Paper presented at the IPSA 2016 meeting. Clarke, Harold, Paul Whiteley, Walter Borges, David Sanders, and Marianne Stewart (2016). ‘Modelling the dynamics of support for a right-wing populist party: the case of UKIP’, Journal of Elections, Public Opinion and Parties, 26:2, 135-154. Where Have All the Working Class Voters Gone? xxxvii

Cox, Michalene (2003). 'When trust matters: Explaining differences in voter turnout'. JCMS: Journal of Common Market Studies, 41(4), 757-770. Dippel, Christian, Robert Gold, and Stephan Heblich, (2015). 'Globalization and its (dis-) content: Trade shocks and voting behavior', National Bureau of Economic Research working papers, No. 21812. Dunn, Kris (2015). ‘Preference for radical right-wing populist parties among exclusive-nationalists and authoritarians’, Party Politics, 21:3, 367-380. Eatwell, Roger, and Cas Mudde (2004). Western democracies and the new extreme right challenge. London; New York: Routledge. Evans, Geoffrey, and Jon Mellon (2016). ‘Working Class Votes and Conservative Losses: Solving the UKIP Puzzle’, Parliamentary Affairs, 69:2, 464-479. Evans, Jocelyn A.J. (2005). ‘The dynamics of social change in radical right-wing populist party support’, Comparative European Politics, 3:1, 76-101. Funke, Manuel and Christoph Trebesch (2017). 'Financial crises and the populist right'. DICE Report 4 / 2017 December Volume 15 Gingrich, Jane, and Silja Häuserman (2015). 'The decline of the working-class vote, the reconfiguration of the welfare support coalition and consequences for the welfare state', Journal of European Social Policy, 25:1, 50-75. Givens, Terri E (2004). ‘The Radical Right Gender Gap’, Comparative Political Studies, 37:1, 30-54. Gidron, Noam, and Peter Hall (2017). ‘Populism as a Problem of Social Integration’, Annual Meeting of the American Political Science Association, San Francisco, September. Golder, Matt (2003). ‘Explaining Variation In The Success Of Extreme Right Parties In Western Europe’, Comparative Political Studies, 36:4, 432-466. Golder, Matt (2016). ‘Far-right Parties in Europe’, Annual Review of Political Science, 19:1, 477-497. Goodwin, Matthew J., David Cutts, and Laurence Janta-Lipinski (2016). ‘Economic losers, protestors, islamophobes or xenophobes? Predicting public support for a counter-Jihad movement’, Political Studies, 64:1, 4-26. Gray, Mark and Miki Caul (2000). 'Declining voter turnout in advanced industrial democracies, 1950 to 1997: The effects of declining group mobilization', Comparative political studies, 33(9), 1091- 1122. Grönlund, K., & Setälä, M. (2007). 'Political trust, satisfaction and voter turnout', Comparative European Politics, 5(4), 400-422. Hainmueller, Jens, and Daniel J. Hopkins (2014). 'Public attitudes toward immigration', Annual Review of Political Science 17: 225-249. Harteveld, Eelco (2016). ‘Winning the losers but losing the winners? The electoral consequences of the radical right moving to the economic left’, Electoral Studies, 44, 225-234. Harteveld, Eelco, and Elisabeth Ivarsflaten (2016). ‘Why Women Avoid the Radical Right: Internalized Norms and Party Reputations’, British Journal of Political Science, 1-16. Hatton, Timothy J (2016). ‘Immigration, public opinion and the recession in Europe’, Economic Policy, 31:86, 205-246. Hanretty, Chris (2015). 'Electorally, West European social democrats are at their lowest point for forty years', retrieved from https://medium.com/@chrishanretty/electorally-west-european-social- democrats-are-at-their-lowest-point-for-forty-years-ac7ae3d8ddb7 (accessed March 29, 2018). Hernández, Enrique, and Hanspeter Kriesi (2016). ‘The electoral consequences of the financial and economic crisis in Europe’, European Journal of Political Research, 55:2, 203-224. Hooghe, Liesbet, Gary Marks, and Carole J. Wilson (2002). ‘Does Left/Right Structure Party Positions on European Integration?’, Comparative Political Studies, 35:8, 965-989. 38 The reconfiguration of working class politics

Inglehart, Ronald (1977). The silent revolution: Changing values and political styles among Western publics. Princeton: Princeton University Press. Inglehart, Ronald, and Pippa Norris (2016). ‘Trump, Brexit, and the Rise of Populism: Economic Have- Nots and Cultural Backlash’, Social Science Research Network working papers, retrieved from https://papers.ssrn.com/abstract=2818659 (Accessed December 3, 2016). Kinder, Donald R. and D. Roderick Kiewiet (1981). 'Sociotropic politics: the American case', British Journal of Political Science, 11(2), 129-161. Machuca, Pablo (2016). "Pablo Iglesias: "En España no existe una formación de extrema derecha porque existe Podemos" [Pablo Iglesias: "In Spain there is no far-right party thanks to Podemos"]". Huffington Post. Retrieved from: https://www.huffingtonpost.es/2016/11/11/iglesias-extrema-derecha_n_12911376.html (accessed 28 November 2018) Ivarsflaten, Elisabeth (2005a). ‘The vulnerable populist right parties: No economic realignment fuelling their electoral success’, European Journal of Political Research, 44:3, 465-492. Jackman, Robert W, and Karin Volpert (1996). ‘Conditions favouring parties of the extreme right in Western Europe’, British Journal of Political Science, 26:4, 501-521. van Kessel, Stijn (2015). Populist Parties in Europe. London: Palgrave Macmillan UK. Kitschelt, Herbert (1988). ‘Left-Libertarian Parties: Explaining Innovation in Competitive Party Systems', World Politics, 40:2, 194 -234. Kitschelt, Herbert (1995). ‘Formation of Party Cleavages in Post-Communist Democracies: Theoretical Propositions’, Party Politics, 1:4, 447-472. Kitschelt, Herbert (2007). ‘Growth and Persistence of the Radical Right in Postindustrial Democracies: Advances and Challenges in Comparative Research’, West European Politics, 30:5, 1176-1206. Kriesi, Hanspeter (2014). ‘The Populist Challenge’, West European Politics, 37:2, 361-378. Kriesi, Hanspeter, Edgar Grande, Romain Lachat, Martin Dolezal, Simon Bornschier, Timotheos Frey (2006). ‘Globalization and the transformation of the national political space: Six European countries compared’, European Journal of Political Research, 45:6, 921-956. Kriesi, Hanspeter and Simon Hug (eds.) (2010). Value Change in Switzerland. Lanham: Lexington Press. Kriesi, Hanspeter and Takis S. Pappas (Eds.). (2015). European populism in the shadow of the great recession. Colchester: ECPR Press. Lefkofridi, Zoe, Markus Wagner, and Johanna E. Willmann (2014). ‘Left-Authoritarians and Policy Representation in Western Europe: Electoral Choice across Ideological Dimensions’, West European Politics, 37:1, 65-90. Leighley, Jan E. and Jonathan Nagler (1992). "Individual and systemic influences on turnout: Who votes? 1984", The Journal of Politics, 54(3), 718-740. Lijphart, Arend (1997). "Unequal participation: Democracy's unresolved dilemma presidential address, American Political Science Association, 1996", American political science review, 91(1), 1-14. Lipset, Seymour M. (1960). 'Party systems and the representation of social groups', European Journal of Sociology/Archives Européennes de Sociologie, 1:1, 50-85. Loxbo, Karl (2014). ‘Voters’ Perceptions of Policy Convergence and the Short-term Opportunities of Anti-immigrant Parties: Examples from Sweden: Voters’ Perceptions of Policy Convergence and the Short-term Opportunities of Anti-immigrant Parties’, Scandinavian Political Studies, 37:3, 239-262. Lubbers, Marcel, Mérove Gijsberts, and Peer Scheepers (2002). ‘Extreme right-wing voting in Western Europe’, European Journal of Political Research, 41:3, 345-378. Lubbers, Marcel, and Jochem Tolsma (2011). ‘Education’s impact on explanations of radical right-wing voting’, Interdisciplinary Conference on Migration, Economic Change, Social Challenge, April 6th-9th, 2011 University College London. Where Have All the Working Class Voters Gone? xxxix

Lucassen, Geertje, and Marcel Lubbers (2012). ‘Who fears what? Explaining far-right-wing preference in Europe by distinguishing perceived cultural and economic ethnic threats’, Comparative Political Studies, 45:5, 547-574. Manatschal, Anita (2015). ‘Switzerland - Really Europe’s Heart of Darkness?’, Swiss Political Science Review, 21:1, 23-35. Mayer, N. (2013). ‘From Jean-Marie to Marine Le Pen: Electoral Change on the Far-right’, Parliamentary Affairs, 66:1, 160-178. MacWilliams, Matthew C. 'Who decides when the party doesn’t? Authoritarian voters and the rise of Donald Trump', PS: Political Science & Politics 49.4 (2016): 716-721. Meguid, Bonnie M. 'Competition between unequals: The role of mainstream party strategy in niche party success', American Political Science Review 99.3 (2005): 347-359. Middendorp, Cees P. (1993). ‘Authoritarianism: Personality and ideology’, European Journal of Political Research, 24:2, 211-228. Mierina, Inta, and Ilze Koroleva (2015). ‘Support for Far-right Ideology and Anti-Migrant Attitudes among Youth in Europe: A Comparative Analysis’, The Sociological Review, 63:2, 183-205. Mudde, Cas (2007). Populist radical right parties in Europe. Cambridge: Cambridge University Press. Mudde, Cas (2010). ‘The Populist Radical Right: A Pathological Normalcy’, West European Politics, 33:6, 1167-1186. Mudde, Cas, and Cristóbal Rovira Kaltwasser (2013). ‘Exclusionary vs. Inclusionary Populism: Comparing Contemporary Europe and Latin America’, Government and Opposition, 48:2, 147- 174. Nickell, Stephen, and Sushil Wadhwani (1990). ‘Insider Forces and Wage Determination’, The Economic Journal, 100:401, 496. Norris, Pippa (2005). Radical Right: Voters and Parties in the Electoral Market. Cambridge: Cambridge University Press. Oesch, Daniel (2006). Redrawing the Class Map. Stratification and institutions in Germany, Britain, Sweden and Switzerland. London: Palgrave Macmillan UK. Oesch, Daniel (2008). ‘Explaining Workers’ Support for Right-Wing Populist Parties in Western Europe: Evidence from Austria, Belgium, France, Norway, and Switzerland’, International Political Science Review, 29:3, 349-373. Patton, David F. (2006). 'Germany's Left Party. PDS and the ‘Vacuum Thesis’: From regional milieu party to left alternative?' Journal of Communist Studies and Transition Politics, 22(2), 206-227. Pardos-Prado, Sergi (2015). ‘How Can Mainstream Parties Prevent Niche Party Success? Center- Right Parties and the Immigration Issue’, The Journal of Politics, 77:2, 352-367. Polavieja, Javier G. (2016). ‘Labour-market competition, recession and anti-immigrant sentiments in Europe: occupational and environmental drivers of competitive threat’, Socio-Economic Review, 14:3, 395-417. Przeworski, A. (1985). Capitalism and social democracy. Cambridge: Cambridge University Press. Rooduijn, Matthijs (2015). ‘The rise of the populist radical right in Western Europe’, European View, 14:1, 3-11. Rooduijn, Matthijs (2017). ‘What unites the voter bases of populist parties? Comparing the electorates of 15 populist parties’, European Political Science Review, 10:3, 1-18. Rosenstone, Steven J. (1982). 'Economic adversity and voter turnout', American Journal of Political Science, 25-46. Rueda, David (2005). "Insider-outsider politics in industrialized democracies: the challenge to social democratic parties", American Political Science Review, 99:1:,61-74. 40 The reconfiguration of working class politics

Rydgren, Jens (2008). ‘Immigration sceptics, xenophobes or racists? Radical right-wing voting in six West European countries’, European Journal of Political Research, 47:6, 737-765. Rydgren, Jens, and Patrick Ruth (2011). ‘Voting for the Radical Right in Swedish Municipalities: Social Marginality and Ethnic Competition?: Voting for the Radical Right in Swedish Municipalities’, Scandinavian Political Studies, 34:3, 202-225. Spies, Dennis, and Simon T. Franzmann (2011). ‘A Two-Dimensional Approach to the Political Opportunity Structure of Extreme Right Parties in Western Europe’, West European Politics, 34:5, 1044-1069. Steinmo, S. (1994). Rethinking American exceptionalism: culture or institutions?, in L. Dodd & C. Jillson (eds.), The dynamics of American politics. Boulder: Westview Press. Steinmo, S. & Watts, J. (1995). 'It’s the institutions, stupid! Why comprehensive national health insurance always fails in America', Journal of Health Politics, 20, 329-372. Stockemer, Daniel (2012). ‘The Swiss Radical Right: Who are the (new) voters of the Swiss People Party?’, Representation, 48:2, 197-208. Swank, Duane, and Hans-Georg Betz (2003). ‘Globalization, the welfare state and right-wing populism in Western Europe’, Socio-Economic Review, 1:2, 215-245. Tsirbas, Yannis (2016). 'The January 2015 parliamentary election in Greece: government change, partial punishment and hesitant stabilisation', South European Society and Politics, 21:4, 407- 426. Van der Brug, Wouter, Meindert Fennema, and Jean Tillie (2005). 'Why some anti-immigrant parties fail and others succeed: A two-step model of aggregate electoral support', Comparative Political Studies, 38:5, 537-573. Van Erkel, Patrick F. and Thomas W. Van Der Meer (2016). Macroeconomic performance, political trust and the Great Recession: A multilevel analysis of the effects of within-country fluctuations in macroeconomic performance on political trust in 15 EU countries, 1999-2011. European Journal of Political Research, 55(1), 177-197. Verba, Sidney and Norman H. Nie (1972). Participation in America. New York: Harper & Row. Voulgaris, Yannis and Ilias Nikolakopoulos (2014), ‘Εισαγωγή: ο εκλογικός σεισμός του 2012» [Introduction: The 2012 Electoral Earthquake], in Yannis Voulgaris and Ilias Nikolakopoulos (eds.) 2012: Ο Διπλός Εκλογικός Σεισμός [2012: The Twin Earthquake Elections] (1-23), Athens: Themelio. Werts, Han, Peer Scheepers, and Marcel Lubbers (2013). ‘Euro-scepticism and radical right-wing voting in Europe, 2002-2008: Social cleavages, socio-political attitudes and contextual characteristics determining voting for the radical right’, European Union Politics, 14:2, 183-205. Wolfinger, Raymond E. and Steven J. Rosenstone (1980). Who votes? New Haven: Yale University Press.

Dualism, Welfare State, and the Populist Radical Right xli

Chapter 3. Dualism, Welfare State, and the Populist Radical Right with Guillem Vidal, PhD candidate at the European University Institute

Why do people vote for the populist radical right (PRR)? More specifically, why do these parties disproportionately attract voters from disadvantaged backgrounds? The link between the PRR and the working class is relatively well-established (Scheepers, Eisinga & Lammers 1993; Kitschelt 1995; Lubbers et al. 2002; Arzheimer and Carter 2006; chapter two of the present work), but there is a heated debate on the determinants of this link. In its most simplified version, the debate boils down to a dichotomy between cultural and economic explanations. While the advocates of cultural explanations argue that the rise of populism is best explained by a cultural backlash and value orientations (Norris 2005; Kriesi 2013; Inglehart and Norris 2016), the proponents of the economic argument put forward materialistic explanations as a key driver for populist support (Harteveld 2015; Dippel et al. 2015). Lately, some authors (notably, Rovny and Rovny 2017) have introduced a crucial nuance on this debate: the distinction between so-called outsiders and insiders (i.e. those facing worse or better prospects of labour market protection and security. PRR parties have been attributed the label of 'exclusionary populism' (Mudde 2013) insofar they propose a clear-cut division between 'the elite' and 'the people' and, at the same time, define the latter in an exclusionary way, i.e. bound by a specific national, ethnic, cultural or social group which they claim to represent. Exclusionary populism is increasingly featuring an all-protectionist platform that encompasses not only anti-immigration policies, but also economic nationalism (Kriesi et al. 2006) as well as welfare chauvinism (Alonso and Rennwald 2018). The combination of cultural and economic protectionism arguably matches the preferences of the working-class electorate. Yet, a growing strain of political economy literature suggests that the distinction between insiders and outsiders also constitutes and important structural divide amongst workers that influences voting behaviour (Saint- Paul 1993, Rueda 2005, Marx 2013). Since the insider/outsider division cuts across classic class divisions, we are interested in studying the effect of labour market vulnerability on the vote for PRR and the extent to which the working class shifts to the PRR may be explained by it. It has been shown that workers in occupations associated with greater risk exposure are more likely to choose the far- right, while that is not the case for those whose contract situation points in that same direction (Rovny and Rovny 2017). Our aim here is to sharpen our understanding of this relationship. In a nutshell, we expect workers to be more prone to choosing exclusionary parties whenever they have a gained status that might be attacked. Following that logic, insiders will more likely to vote for the exclusionary PRR when compared to outsiders because their all-protectionist platform is of special interest for protecting their economic status. In consequence, we also pose that not all insiders are equal before labour market risks. If the reason behind insiders choosing the exclusionary far-right were to protect their status, it would make sense that those facing more competition within its segment of the labour market, and those whose material situation would be most affected in case of losing their status, would be even more likely to choose an all-protectionist platform. At the same time, the trend should be comparatively more visible in countries where welfare systems are status-based (i.e. Bismarckian welfare states), and therefore offer a larger protection to insiders compared to outsiders. Our empirical inquiry into the voting behaviour of European PRR parties using comparative survey data (European Social Survey, henceforth ESS, 2002-2016) suggests that voters of the PRR are, on average, more likely to be labour market insiders. Our findings show that this effect is, as expected, more pronounced for insiders whose occupations are particularly affected by both unemployment and migrant rates, as well as in countries with Bismarckian welfare states, lower unionization rates, or that belong to certain occupational and low-skilled categories. These results represent a new contribution that will hopefully help advance the puzzle, providing a more structural, comparative perspective. The chapter proceeds as follows. The following section contextualises the emergence of populist parties in Europe and discusses the nature of populism. Next, we briefly introduce a review of the literature on the determinants of PRR vote and labour market divisions on voting behaviour. The theoretical section concludes with the discussion of our hypotheses concerning the labour market insider-outsider divide on the vote for PRR parties. The data and methodology are then briefly discussed before introducing the results. The last section offers some concluding remarks. 42 The reconfiguration of working class politics

3.1 Populism and the Transformation of European Party Systems The transformations in the electoral alignments in Europe since the 1960s have been a widely studied subject in contemporary political science. Kriesi et al. (2006; 2008; 2012) identify two waves of transformation that reshaped North-European party systems - a first wave led by the new left in the 1970s and early 1980s, and a second wave driven by the new right since the 1990s. The term ‘new’ underscores that these actors challenged the political order at the time. Most importantly, the challenge arose from new issues and demands that these actors brought into the political process. The new left advocated individual autonomy, the free choice of lifestyles and other universalistic values (e.g., della Porta and Rucht 1991; Kitschelt 1988). The new right, by contrast, focused on immigration and European integration as threats to the homogeneous nation-state. The so-called second wave took place under a context marked by a deepening of the globalisation and technification process, the fall of the USSR -and thus the demise of classic far-left authoritarian platforms-, and rising inequality within Western countries. The 2008 Great Recession gave a new impulse to the paradox: precisely when structural processes (i.e. globalisation and its side effects within Western workforces and rising inequality) gave a more propitious context to protectionist and redistributive platforms, the traditional left was in a deep political crisis. Many have seen populism as being at the forefront of the new parties' answer to the aforementioned paradox, and therefore as the key feature of this second wave. As Mudde (2013) argues within his comprehensive review of the concept, the label 'populist' should be used with caution, since it might quickly become a slippery concept unless properly defined. Following Mudde and Kaltwasser (2013) among others, we opt for what he descibes as a thin-centred ideological approach to define populism. From this point of view, populism would have anti-elitism as its key feature since it would consider society as essentially divided on moral grounds between two antagonistic camps: the good people vs the corrupt elite (Mudde and Kaltwasser 2013: 1). Populists project themselves as the bearers of the people's true will and the defenders of 'the good people'. Everything that falls beyond the good people thus becomes the ‘enemy’, which can take different forms. The new European far-right breeds a particular strand of exclusionary populism that contrasts with the inclusive populism typical of Latin American left-leaning parties (Mudde and Kaltwasser 2013). While inclusive populism is meant to define the people as outsiders in a fight against the elite for a more encompassing set of policies and institutions, exclusionary populists actively define a number of groups beyond the 'corrupt elite' as not deserving equal treatment. In a way, it could be said that the key difference between both types of populism is how the enemy is defined depending on the relation vis a vis the system. If one is (or feels) out of 'the system' and has always been so, the enemy is limited to those who organize 'the system' against you. However, if one is (or feels like) part of a 'system' that was working well but now it is not, then there are enemies beyond the elite. The populist right's discourse towards immigration offers a crystalline example, since they are often used by it as scapegoats for many of the malaises affecting 'the people'. The standard structure of the argument matches with the idea of exclusionary populism, pointing (explicitly or not) to an alliance between the elites and the newcomers in detriment of nationals. This dual enemy (i.e. immigrants and the elites) is what has been labelled as the tripartite structure of the PRR (Damhuis 2018). The division between inclusionary and exclusionary populism squares well with the distinction between populist left and populist right. While the former tends to focus on economic fairness and redistribution, the latter puts more attention on protectionist and even authoritarian solutions. Without advancing any formal hypothesis yet, it therefore makes sense to expect that the latter will be supported by a specific kind of voter inclined towards these proposals.

3.1.1 Exclusion (or the fear of it) as a determinant of PRR vote PRR parties are born and raised in the context of crises of political representation, and according to several works, they take advantage of the working class’ search for a new political home. It is what Kitschelt (1995) defined early on as "occupational risk". Scheepers, Eisinga and Lammers (1993); Scheepers, Billiet and De Witte (1995); Eisinga et al. (1998) all find a nexus between either manual workers or the less educated and far-right vote. More recently, Arzheimer and Carter (2006) conduct one of the first comprehensive cross-country analyses on the determinants of the far-right vote and find that, as they expect, being a manual worker significantly increases the probability of voting for the PRR. Moreover, Lubbers et al (2002) complementarily observe that those in a professional occupation Dualism, Welfare State, and the Populist Radical Right xliii

are significantly less likely to vote for the PRR. This consistent finding, replicated in many comparative case studies as well as single case studies, crucially informs the debate on what determines the demand for PRR parties and poses a question that is particularly central to the intentions of the present work: why do those who are closer to the margins of society get associated with an exclusionary movement? The earliest and most straightforward answer comes from the so-called Realistic Conflict Theory (Coser 1956). In its most stripped-down version, the theory poses that if a subset of a given population perceives itself as competitor for a limited set of resources with another subset that can be labelled as 'outsider', conflict is likely to follow. The analogy with the way the exclusionary populist right exploits immigration as a threat is self-evident. Lubbers and Scheepers (2000) are among the first to find support for this interpretation, at least for the German case. Traditionally, the notions of inclusion and exclusion within a capitalist society have been linked to the idea of class: certain class groups enjoyed a higher social status than others and fought to defend it. In the last two decades, however a new, cross-class division has been highlighted by sociologists and economists: dualisation, or the increasing breach between insider (protected) and outsider (precarious, unprotected) workers. Developed first by labour economists whose focus was to highlight the higher bargaining power of permanent employees thanks to employment protection legislation (Blanchard and Summers 1986; Lindbeck and Snower 1988), political economy took on the issue early on to explain policy equilibriums in favour of high employment protection first (Saint-Paul 1993), and the subsequent dilemma confronted by social democratic parties who aspire to attract both insiders and outsiders due to their different policy preferences (Rueda 2005). Insiders, regardless of their class belonging, were now seen as the group fighting to maintain their gained status. In consequence, if social democratic parties focus on employment protection legislation and dismiss passive and active employment policies, outsiders might turn their back to them in search for some alternative that facilitates a policy set to improve their situation. It is important now to pause to highlight that alternative definitions to the notion of outsiderness (and therefore insiderness) have been proposed. Probably Hausermann and Schwander (2011) bring the most influential example to the table: departing from a multi-dimensional conceptualisation that is not restricted to contract type and that draws important significance from an occupational division, they find a highly relevant way to classify workers according to their exposure to potential exclusion, with consequences for their political preferences. While the contract-based definition of dualisation has taken to identify outsiders with abstention (Rueda and Lindvall 2014) and the New Left (Marx 2013; Rovny and Rovny 2017), the occupation- based classification helps to find a link between outsiderness and far-right vote (Rovny and Rovny 2017). The latter finding is in line with a small but growing breed of nuanced studies that prove the importance of structural elements. Lubbers & Tolsma (2011), for instance, pool 2002-2008 ESS data to establish that, although migrant-related fear and authoritarian attitudes do explain why people with lower education levels tend to vote for the far-right compared to the higher educated, attitudes end up having a much weaker predicting capacity among the former group. As Dippel et al. (2015) suggest, it might be the case that these segments of the population feel the pressure of potential exclusion through the labour market and respond through exclusionary vote patterns to further shelter their position against external threats. Dippel et al. apply a causal mediation analysis to find out that at least two thirds of the voting effects provoked by trade exposure in Germany can be explained through its impact on labour markets. When they take the analysis to the individual level, it turns out that low- skilled manual workers are the ones turning towards the far-right following the trade-on-labour effect. In a neat correlate to Dippel et al, van der Waal & de Koster (2017) find that protectionist positions diminish the weight of cultural variables in voting decisions towards the far-right, which means that economic proposals to attract voters towards the far-right seem to matter. Summing up, we know that (1) contract-based outsiders are more likely to move away from PRR parties; (2) but occupation-based outsiders are more likely to get close to them; (3) and there seems to be some preliminary indications of competition for material-based benefits (and the loss of them) driving PRR vote. In consequence, our core hypotheses (laid out in the next section) will revolve around these departing assumptions. However, the working class - PRR link might very well have a non-material basis. Lipset (1959) already assessed that those with 'low status' or education tended to espouse more extremist, intolerant forms of political behaviour. The author goes as far as to assert that "the lower-class individual is more likely to have been exposed to punishment, lack of love, and a general atmosphere 44 The reconfiguration of working class politics

of tension and aggression since early childhood, experiences which often produce deep-rooted hostilities expressed by ethnic prejudice [and] political authoritarianism" (Lipset 1959: 495). What lies behind Lipset's idea is the concept of identity. The vote for PRR parties might be understood as a search for a new set of values to relate to (Lubbers and Scheepers 2000) of a class group that feels that they exist in a world whose previously established certainties are now crumbling, or from segments of society who face problems of social integration (Gidron and Hall 2017). In its most essential version, PRR vote would be a backlash against cultural change. In his seminal work, Inglehart (1977) found that a profound turn in dominating values was taking place in Western societies, away from materialism worries and towards post-materialistic perspectives such as social justice or environmentalism. There was a new generation of middle-class professional workers who was leading the way and thus bringing a change in values that would end up feeding the New Left, as well as deep crises within classic social democratic parties. The far-right surge might be read as a backlash to this process, one that would be particularly appealing for those who are further away from the groups involved in the post-material revolution, normally manual and under-qualified workers (Ignazi 1992; Kitschelt 1995). The link between illiberal cultural preferences and far-right vote has been well established both for the whole population (e.g. Inglehart and Norris 2016) and for the working class (Bornschier and Kriesi 2012, as well as chapter 2 of this dissertation), so it would be a rather slim contribution to bring it again to test. Our goal, rather than adding to the existing dichotomical debate between economic and non- economic factors of far-right vote, is to deepen on the former through the exploration of the insider/outsider distinction. In that sense, there is a potential dimension to the non-material side of voting for exclusionary populism to avoid potential exclusion: the loss of societal status. The fear of 'class loss' (more commonly referred by its French name, déclassement) is defined as much more than the mere erosion of income and wealth (Bourdieu 1978), featuring in essence the displacement to a lower status level under the regards of your peers. It could be said that those who have achieved the status of insiders have gained with it a sheltering against this risk, and that any threat to its status goes beyond its economic dimension (Santos Ortega 2004). There is, however, little evidence on the influence of this particular factor as a driver of PRR vote in existing literature. Thus, it is hard to hold any strong expectation. A test should be nonetheless included in order to distinguish a purely material effect from other dimensions of labour-linked status.

3.1.2 Insiders, outsiders and the exclusionary far-right If there seems to be an emerging consensus on outsiders leaving the social democratic (SD) coalition towards either the left or abstention, but rarely or never to the right side of the spectrum (Schwander 2017), this begs the question: would that mean that insiders are more likely to make the opposite move? There seems some preliminary evidence in that direction (see for instance Hausermann and Walter 2010 for the Swiss case, as well as Rovny and Rovny 2017). The logic behind the proposition mirrors Rueda's core contribution: if, whenever social democratic parties stick to insider interest, outsiders move away, would it not logically follow that insiders are equally prone to finding parties that are more suitable in case social democrats turn towards the right-liberal side of the spectrum, making insiders fear exclusion? The party convergence thesis is then a somewhat necessary companion for this reasoning and one that is at least partially supported by the events in the last decade (e.g. Magalhaes 2014; Hernández and Kriesi 2016). Hence our preliminary hypothesis: H1: Insiders tend to vote for PRR parties compared to outsiders, who tend to separate themselves from PRR parties. With this first hypothesis, we have established the base of our theory: insiders are more interested in exclusionary populist parties than outsiders. Going beyond previous works in a similar direction (particularly Rovny and Rovny 2017), we pose that it will be mostly because they risk more in the gamble in terms of losing economic status. However, it might very well be argued that, even if we control for cultural attitudes and preferences, being a permanent worker might be capturing something else. To pin down our mechanism (i.e. loss of economic status), we propose three measures. The first measure is arguably the one closest to our proposed mechanism, and it introduces a dimension of potentially perceived competition for jobs. It does so by combining the unemployment rate and the rate of non-native workers within the occupation-country of the individual, to expect that: H2a: Insiders in occupations displaying comparatively higher levels of unemployment and immigrant workers will be more likely to vote for PRR parties. Dualism, Welfare State, and the Populist Radical Right xlv

At the same time, we assume that not all workers face similar labour market risks of vulnerability. In that sense, combining the status-based measure of dualisation and the occupation-based measure produces an interesting possibility: to measure whether a particular permanent worker can see her status more or less threatened depending on how much competition she faces within her current job position, and how much could her material situation be damaged in the case of a more open labour market. Although the insider/outsider division cuts through traditional class and occupational classifications, it does not make them disappear. More specifically, it could be argued that manual and service workers are more exposed to labour market risks, displaying higher unemployment rates as well as facing longer unemployment spells across Europe in recent years. As a matter of fact, chapter two showed that manual and service workers are more likely to choose far-right parties than their counterparts. It would therefore make sense then to expect manual and service insiders to display an even deeper inclination to vote for PRR. H2b: Insiders in working class occupations will be more likely to vote for the PRR than outsiders. Another way to look at the same phenomenon would be achieved by creating a two-by-two classification to distinguish between high- and mid- or low-skilled insiders and outsiders. Skill level is a good predictor of exposure to the consequences of labour market risks by itself, with high-skilled workers normally being more prepared to cope with the woes that might come with unemployment. Following the logic of H1 and H2a, it would therefore make sense to expect it to have an additional effect over PRR vote. As a matter of fact, it could be argued that skill interacts with job status on the distribution of risk exposure, particularly for low-skilled outsiders who might face a lower probability of building a stable career path, with temporary jobs becoming some sort of dead ends for certain profiles (Booth et al 2002, Autor and Houseman 2005). Low-skilled insiders, instead, may perceive the risk they face as higher since the gap between the status they hold as insiders and the one they could fall into is wider. Moreover, if job status and occupational/skill dimensions come to interact to define PRR vote probability, it would make sense to expect a combined category to be more accurate. H2c: Mid- and low-skilled insiders are more likely to vote for the PRR than any other group. A broader comparative perspective suggests that labour market and welfare systems under different institutional setups may well condition the exclusionary populist framing of “the people” and the “other”. These structural and institutional attributes of distinct welfare states not only influence levels of social stratification, or inequality; more important for an understanding of populism, perhaps, they influence the degrees of social integration, cohesion, and segmentation. Bluntly, what makes an insider has very much to do with the regulatory setup. Esping-Andersen (1990) delineated a division between the 'three worlds of welfare capitalism': Nordic states featuring inclusive institutions (i.e. encompassing trade unions and a universalistic welfare safety net); Anglo-Saxon (or liberal) economies favour flexibility (i.e. free choice and competition to the detriment of equality); and Continental (also called Bismarckian) states displaying a corporatist approach to welfare that tends to protect active workers with long careers through the generation of rights throughout their working lives. Essentially, Nordic countries countervail segmentation and inequality while liberal economies prioritize individual and firm decisions, and Bismarckian states tend to favour dualisation since it distributes protection to its citizens depending on their work status and occupational profile. Hausermann and Schwander (2012), for instance, finds stronger insider-outsider divide in Southern welfare models. Afonso and Rennwald (2018) suggest (as an intuition) that PRR electorate contains an oversized portion of insiders in Bismarckian welfare systems who are afraid of sharing their safety net with outsiders. Their work focuses on how welfare chauvinism fits well with the preferences of these groups, normally composed by 'male breadwinners'. Complementarily, Swank and Betz (2002) prove that universal welfare states diminish the direct effects of globalization (in particular, the increase in people-flows and the decline of manufacturing employment) while wiping out the effect of other variables (i.e. rises in trade openness and capital mobility). Thus, it makes sense to expect that, if insiders are in a country where they have a lot to lose (status-based), they will vote for the exclusionary populist right with even more intensity since they will be particularly interested in protecting their own positions (Roberts 2017). A variation of our first hypothesis can thus be formulated: H3a: Insiders in status-based countries (with Bismarckian welfare states) will be more likely to vote for the PRR than in non-status-based countries. 46 The reconfiguration of working class politics

Voting is by no means the only political activity taken by workers to shelter themselves against labour market risks. Probably the main (and historically the earliest) political decision taken by any employee to pursue a better job status is to join a union. Unionised workers tend to be less exposed to unemployment and to the risk of income loss (Johnson and Mieszowski 1970). Moreover, it makes sense to expect that unions serve as a mechanism to homogenise working class vote and keep it within the left realm, since there seems to be a correlation between unionisation levels and the strength of social democracy in Western European countries that goes beyond the natural individual link between belonging to a union and voting centre-left: strong unions are more likely to effectively pressure dominant parties into policymaking that is favourable to insiders. While it is true that they might face a dilemma between appealing to their most populous constituencies (i.e. insiders) and searching for new support among outsiders (Polavieja and Richards 1997), since most of its base is constituted by insiders, it makes sense to expect unions to prioritise insiders’ interests. The implication of all these factors for our argument is clear: if unions are stronger in a given country, insiders have less incentives to choose PRR parties. On the contrary, wherever they are weaker, insiders will have more incentives to opt for the far-right. H3b: In countries where unions have less of a presence (measured with union density) insiders will be more likely to vote for the PRR than outsiders.

3.2 Data, methodology and results To test the hypotheses outlined above we rely on all available rounds (1-8) of the ESS. The ESS is a comprehensive European-wide survey that includes a detailed classification of the labour market position of the respondent, vote choice in the previous national elections, and a wide range of ideological and socio-demographic variables on a biyearly basis from 2002 to 2016. This allows us to create detailed measures of our variables of interests and control for other explanatory factors in a cross-national comparative perspective. Unlike in the previous chapter, since now our sole point of interest is PRR vote, we restrict our analysis to these Western European countries were PRR parties are significant enough (i.e. have parliamentary representation): Austria, Belgium, Switzerland, Denmark, Finland, France, United Kingdom, Netherlands, Norway and Sweden. Our dependent variable measures the vote choice by party family in the last national elections right before the survey took place. We coded each individual party for each of the ESS rounds following the standard CHES party family classification. We employ a simplified version that renders seven categories2: • Mainstream Left - Social-Democratic parties. • Mainstream Right - Liberal, Christian-Democrats and Conservative parties. • Radical Left - Communist and other Radical Left parties. • Greens • Populist Radical Right - See Table B.2 in the Appendix for a list of the selected parties by country and the sample size in our dataset. The selection of PRR parties is based on the studies of van Kessel (2015) and Mudde (2007)3. • Others - Agrarian, Confessional, Regional/Ethnic or without family. • Abstention - Those who abstained in last national elections.

3.2.1 Dualisation and PRR vote Our core independent variable (IV) is labour market status (or dualisation). Here, it is operationalised in three categories following Rueda (2005; 2007), namely outsiders, insiders and upscale. Outsiders consist of those individuals who are unemployed or are in part-time or fixed-term employed, including housekeepers. The upscale group is composed of self-employed and large employers, the upper

2 See Table B.2 in the Appendix for the frequencies and percentages of the party families by country. 3 See also Guiso et al. (2017: 62-63). Dualism, Welfare State, and the Populist Radical Right xlvii

middle-class, managers and small business owners with employees4. Insiders are the remaining individuals whom are neither outsiders nor upscale5 (see Figure A.1 in the Appendix to see the percentages of each category by country in the sample). To test our hypotheses, we run a series of logistic and multinomial logistic models using country-fixed effects and clustering standard errors at the country-level. A series of socio-structural (age, gender, education except for the skill level test) and ideological (anti-immigration, anti-redistribution, political trust) variables are included in the model as controls, running one model with only the structural set of controls and another model with all control variables. To recapitulate, our first hypothesis suggested that voters of the PRR would tend to be insiders (H1). The underlying assumption is that the nature of the exclusionary populist discourse will appeal to different socioeconomic groups depending on their labour market position. Since PRR parties highlight outside threats to the nationals’ life style and job security, we expect their discourse to appeal to insiders because they fear competition. Figure 3.1 shows the average marginal effects of labour market status on vote choice for the different party families. The results show that voters of PRR parties tend to be, on average, overrepresented by insiders as opposed to outsiders. While this effect is significant and in the expected direction, it is also significantly smaller than for voters of SD parties. Consistent with previous work (Rovny and Rovny 2017), our results also suggest that outsiders are overrepresented amongst those who did not vote (abstained in the past elections) -especially in South European countries (Hausermann and Schwander 2012: 45). Although to a lesser degree, outsiders are also overrepresented in the voters of the green or new left parties in our findings, which again is consistent with existing literature that looks at the effect of temporary contracts on voting behaviour (Marx 2015: 74, 2014: 147).

Figure 3.1. Marginal effects of insiders vs outsiders by party family

Results on Table B.4, based on models from Table B.3.

4 Respondent's occupational code 11, 12, 13 in in ISCO88-2d (ISCO88) plus self- employed professionals and large employers. 5 See Figure B.1 in the Appendix to see the n of each category by country in the sample. 48 The reconfiguration of working class politics

Returning to our case of interest, Tables B.3 and B.4 in the Appendix show the marginal effects of voting for the PRR of different multinomial logistic models. In all models the coefficients for insider with outsider as a reference category range from .004 to .007, where the effect is statistically significant at p<.015 in all models. This means that, on average, insiders are about 0.47pp and 0.6pp more likely than outsiders to vote for the PRR (in the case for SD parties, for instance, it is about 2.5pp). Even if we find the expected direction of the coefficient, the results are not of an impressive magnitude, suggesting that labour market position and the appeal to national protectionism is only one of the many reasons why voters might opt for these parties. If we pay attention to the other variables in the models, we find that the labour market dimension does not seem the most relevant variable for explaining the vote. For instance, a one-point increase in the anti-immigration attitudes scale leads to an increase of between 12 and 17.4 percentage points of likelihood to vote for the PRR. This is by far the variable included in the model with the most substantial effect. It should be noted that the higher the anti-immigrant attitudes, the more insiders (as opposed to outsiders) are likely to vote for radical right, especially for men (results not shown). Commenting on other control variables, looking at the predicted probabilities of social class on the PRR vote (see Table B.5 in the Appendix) we find that the coefficients of service workers, small business owners, and to a lesser degree clerks, are higher than average (though the latter two are not statistically significant at p>0.05). Moreover, we also find that females are strongly underrepresented in PRR parties (ranging between 1 and 2 percentage points lower), as already discussed in the literature (Spierings and Zaslove 2017). Our results in terms of education indicate that it is not the least educated that tend to vote for the PRR, but those with middle levels of education (other studies have found similar results, see Arzheimer and Carter 2006). Our variable to measure economic anxiety does not render clear results: we only find a significant and negative effect on voting for the PRR amongst those who claim to have severe difficulties with income. In terms of political variables, it is worth highlighting that PRR are, on average, less likely to trust political institutions, more dissatisfied with the way democracy works, and show higher levels of political interest. Regarding ideology, besides the aforementioned effect we find for anti-immigration attitudes, we also find that authoritarian attitudes are also important in predicting the vote. In fact, together with anti-immigration, authoritarian attitudes are one of the best predictors of PRR vote.

3.2.2 Pinning down the mechanism While our preliminary tests come to ratify recent findings from the literature (Rovny and Rovny 2017), our second set of hypotheses aims to deepen the analysis. In a quest to pin down a mechanism that explains the aforementioned relationship, we point to two competing explanations. It might be that insiders feel a threat to their earned material status and therefore opt for an exclusionary platform. Or, it might be that the threat is rather non-material, and that 'insider' is serving as a proxy of other dimensions of societal status. We put both to test. For the material status hypothesis, which is the one we choose to defend in H2a, we have calculated the average rates of non-natives and unemployment levels for each pair of class group (following Oesch's schema) and country, then multiplied them to get a measure of the joint presence of immigrant workers and unemployment in the individual’s work environment. For the societal status hypothesis, the ESS provides a question on how important it is for the individual to maintain respect from her peers. Both are included in two separate models as interactions with our core dependent variable (insider vs outsider), while maintaining them as simple controls when testing for the other. Last but not least, we also include an additional model to see if different levels of employment protection (EPL) affect insiders' proneness to vote for PRR parties, in order to distinguish the purely institutional effect. All tests include as well socio-demographic controls and those indicating a certain degree of anti-immigration sentiment, either for economic or cultural reasons. Results in Tables B.5 to B.10 of the Appendix come to confirm H2a while discarding the 'non-material status' hypothesis and being relatively inconclusive on the effect of actual EPL levels. Figure 3.2 below shows that while there is no difference on the marginal probability of voting for PRR between insiders and outsiders at the lowest level of the combined migration-unemployment variable, at the highest level the gap doubles. At the same time, Table B.10 indicates virtually no difference in the probability distribution when using the 'respect' variable.

Dualism, Welfare State, and the Populist Radical Right xlix

Figure 3.2. Marginal effects of insiders vs outsiders of PRR voters by immigration*unemployment rate/position interaction

Results on Table B.6, based on models from Table B.5.

To reinforce the picture of economic status being the actual factor behind the relationship between insiders and exclusionary populism, H2b considered that working class insiders would have a higher probability to vote for PRR parties. Insiders who belong to socioeconomic groups which are more likely to be in vulnerable positions are more attracted to the protectionist and exclusionary discourse of PRR parties given that they would benefit most from being sheltered from further competition. Social class categories are defined by Oesch (2006) class schema, whose connection with voting patterns has already been proven (e.g. Oesch 2013). In order to test the combined effect of being an insider and in a position of labour market vulnerability on the effect to vote for the PRR, we run models testing the interaction effect between the two variables. Figure 3.3 shows the average marginal effects.

Figure 3.3. Marginal effects of insiders vs outsiders of PRR voters by social class

Results on Table B.12, based on models from Table B.11. 50 The reconfiguration of working class politics

The interaction effects show that working class (service and production workers) insiders are more likely to vote for PRR parties. Differences are statistically significant (at p<0.05). These results suggest that the effect of being an insider on voting for the PRR is mediated by the degree of risk of status loss faced by the individual in the labour market, attributing the PRR insider vote to working class insiders. It follows that, in line with our expectations in H2c, results show that low-skilled manual and service insiders, as well as technicians and clerks, are much more likely than their skilled counterparts to vote for PRR parties (see Tables B.7 and B.8 in the Appendix). The combination of skill level and job status stems from crossing our main IV with the Oesch simplified schema, in which mid- and low-skilled segments are defined as small business, production workers, clerks and service workers. High-skilled segments would be large employers, professionals, technicians, and socio-cultural professionals. As a matter of fact, this reinforces our quest for a mechanism to better explain our first hypothesis, since it looks like it is the insiders in the working classes, both production workers and service workers, who display a larger tendency to vote more for the radical right.

3.2.3 Country-level effects Moving now into country-level effects, we advanced that insiders would be overrepresented amongst PRR voters especially in Bismarckian welfare states (H3a). If, as widely argued, these models of capitalism produce a larger risk differential between insiders and outsiders (Hausermann et al. 2013), then the search for security hypothesised in H1a should play a larger role in mobilising insiders towards the PRR. Figure 3.4 illustrates that classic Western Bismarckian countries with strong PRR parties (Austria, Belgium, France, Netherlands, Switzerland) tend to display the expected effect, albeit with variations. In Germany, the paramount example of a Bismarckian state, the insider/outsider division seems to play a smaller role than in all its category partners, being at the level of the UK or Finland. This can be explained by the fact that the sample includes a relatively small amount of PRR votes in the country. Since the rise of the AfD in Germany in the last general election of September 2017, trends should be clearer. Finland, on the other hand, offers a counter-example: generally seen as a member of the Nordic (and thus, egalitarian) family, insiders feel particularly strong about PRR parties. However, it should be noted that results for Finland and Sweden are non-significant, while its Scandinavian neighbours (Denmark and Norway) feature an overrepresentation of outsiders amongst the voters of PRR parties. Figure 3.4. Marginal effects of insiders vs outsiders of PRR voters by country

Results on Table B.13, based on models from Table B.12. Dualism, Welfare State, and the Populist Radical Right li

To finish our tests, we check whether there is a country-level link correlation between the strength of unions and how deep the connection between insiders and PRR vote goes. Strength of unions in any given country is measured by the unionisation rate, since it is a relatively parsimonious and universal proxy widely used in the specialised literature. Figure 3.5 shows that there seems to be a correlation between the level of union density in a country and the strength of the coefficient that indicates the proportion of insiders vs outsiders that vote for the radical right. Although not a lot can be confirmed by this, it does seem to indicate that in countries where workers are more organised, the insiders tend to vote less for the radical right. Figure 3.5. Correlation between the level of union density in a country and the strength of the coefficient that indicates the proportion of insiders vs outsiders that vote for the radical right

This is just a preliminary approach, although a promising one, to a question that begs further tests: whether workers’ self-organisation can prevent PRR parties’ appeal among the working class to grow beyond a certain threshold.

3.3 Concluding remarks In this chapter, we addressed the labour market dimension behind the PRR vote. Given the exclusivist nature of the discourse of PRR parties, we hypothesised that labour market insiders who defend the welfare status quo against external threats are overrepresented in these parties. Specifically, we expected this effect to be strongest in countries with Bismarckian types of welfare state and lower unionisation rates, amongst those individuals at higher risk of labour market vulnerability, who belonged to the manual and service working classes, and especially in those labour segments facing an apparently tougher competition for jobs. We used a large survey dataset (ESS) covering eight rounds over fourteen years to model voting behaviour of the PRR within European countries. The findings can be summarised in four points. First, we found that, on average, labour market insiders tend to be overrepresented amongst voters of PRR parties across Europe. While the magnitude of the effect we found is rather small and suggests that this divide is not a decisive element of the PRR vote, the results confirmed our expectations. Second, and trying to at least approach the mechanism behind these results, we found that those insiders facing combined higher levels of non- native workers and unemployment within their occupations and countries duplicated their probability of voting for PRR parties compared to both insiders facing lower levels and outsiders facing any levels. 52 The reconfiguration of working class politics

At the same time, we established that it is (manual and service) working class insiders (especially low- skilled) who display a clearer inclination to vote for PRR parties. Third, we observed that in countries with Bismarckian types of welfare states the voters of PRR tend to be more insiders than in countries with universalistic types of welfare states (i.e. North European countries, where the result is either non-significant as it happens for Sweden and Finland, or outsiders are the overrepresented group, as is the case for Denmark and Norway). This is because insiders in Bismarckian types of welfare state have more to lose. Fourth, we checked that there are some preliminary indications of union presence mitigating the link between insiders and the PRR within a given country. The general picture left by the data is one that at the same time confirms and expands former research (particularly Rovny and Rovny 2017) and, simultaneously opens up the space for further research. It does so by departing from the apparent fact that insiders who face tougher labour market conditions (due to either personal or structural reasons) are more likely to vote for populist radical right parties. This might be an interestingly new, more nuanced line of research for those who defend the ‘economic anxiety’ argument to explain PRR vote; one that does not try to pitch itself against cultural or identity-related arguments, but that instead considers the phenomenon as necessarily complex and multi-causal.

Dualism, Welfare State, and the Populist Radical Right liii

References Abou-Chadi, Tarik (2016). ‘Niche Party Success and Mainstream Party Policy Shifts? How Green and Radical Right Parties Differ in Their Impact’, British Journal of Political Science, 46:2, 417-436. Aichholzer, Julian, Sylvia Kritzinger, Markus Wagner, and Eva Zeglovits (2014). ‘How has Radical Right Support Transformed Established Political Conflicts? The Case of Austria’, West European Politics, 37:1, 113-137. Akkerman, Tjitske (2012). ‘Comparing Radical Right Parties in Government: Immigration and Integration Policies in Nine Countries (1996?2010)’, West European Politics, 35:3, 511-529. Alonso, Sonia, and Sara Claro da Fonseca (2012). ‘Immigration, left and right’, Party Politics, 18:6, 865-884. Ansell, Ben, and David Art (2010). ‘Membership Matters: Radical Right Party Composition in Comparative Perspective’, in Annual Conference of the American Political Science Association, Washington DC, vol., 3. http://users.polisci.umn.edu/~ansell/papers/Ansell%20Art%20Membership%20Matters%20APS A.pdf (Accessed June 29, 2017). Arzheimer, Kai (2009a). ‘Contextual factors and the extreme right vote in Western Europe, 1980- 2002’, American Journal of Political Science, 53:2, 259-275. Arzheimer, Kai (2009b). ‘Contextual Factors and the Extreme Right Vote in Western Europe, 1980- 2002’, American Journal of Political Science, 53:2, 259-275. Arzheimer, Kai (2012). Working-class parties 2.0? competition between centre left and extreme right parties. na http://www.kai-arzheimer.com/working-class-parties-extreme- right.pdf.pagespeed.ce.KYZzrQNtTJ.pdf (Accessed June 29, 2017). Arzheimer, Kai (2015). ‘The AfD: Finally a Successful Right-Wing Populist Eurosceptic Party for Germany?’, West European Politics, 38:3, 535-556. Arzheimer, Kai, and Elisabeth Carter (2006). ‘Political opportunity structures and right-wing extremist party success’, European Journal of Political Research, 45:3, 419-443. Bakker, Bert N, Matthijs Rooduijn, and Gijs Schumacher (2015). ‘Following the leader or following your ideology? The case of populist radical right voting’, Unpublished article, http://www.bertbakker.com/wp- content/uploads/2015/01/Bakker_Rooduijn_Schumacher_PopulistVoting.pdf (Accessed June 29, 2017). Bentolila, Samuel, and Giuseppe Bertola (1990). ‘Firing Costs and Labour Demand: How Bad is Eurosclerosis?’, The Review of Economic Studies, 57:3, 381-402. Bentolila, Samuel, and Juan J Dolado (1994). ‘Labour flexibility and wages: lessons from Spain’, Economic policy, 9:18, 53-99. Betz, Hans-Georg, and Carol Johnson (2004). ‘Against the current—stemming the tide: the nostalgic ideology of the contemporary radical populist right’, Journal of Political Ideologies, 9:3, 311-327. Biggs, M, and S Knauss (2012). ‘Explaining Membership in the British National Party: A Multilevel Analysis of Contact and Threat’, European Sociological Review, 28:5, 633-646. Blanchard, Olivier, and Juan F Jimeno (1995). ‘Structural Unemployment: Spain versus Portugal’, The American Economic Review, 85:2, 212-218. Blinder, Scott, Robert Ford, and Elisabeth Ivarsflaten (2013). ‘The Better Angels of Our Nature: How the Antiprejudice Norm Affects Policy and Party Preferences in Great Britain and Germany, American Journal of Political Science 57.4, 841-857. Bolin, Niklas, Gustav Lidén, and Jon Nyhlén (2013). ‘Political Parties and Local Refugee Reception in Sweden’, in XXII Nordiska kommunforskarkonferensen, \AAbo, 21-23 November, 2013., vol. http://www.diva-portal.org/smash/record.jsf?pid=diva2:666562 (Accessed June 29, 2017). 54 The reconfiguration of working class politics

Bornschier, Simon, and Hanspeter Kriesi (2012). ‘The populist right, the working class, and the changing face of class politics’, in Rydgren, Jens (ed): Class politics and the radical right. New York: Routledge. Bowles, Samuel (1985). ‘The production process in a competitive economy: Walrasian, neo- Hobbesian, and Marxian models’, The American economic review, 75:1, 16-36. Bulow, Jeremy I, and Lawrence H Summers (1986). ‘A theory of dual labor markets with application to industrial policy, discrimination, and Keynesian unemployment’, Journal of labor Economics, 4:3, Part 1, 376-414. Burgoon, Brian (2013). ‘Inequality and anti-globalization backlash by political parties’, European Union Politics, 14:3, 408-435. Clarke, Harold, Paul Whiteley, Walter Borges, David Sanders, and Marianne Stewart (2016). ‘Modelling the dynamics of support for a right-wing populist party: the case of UKIP’, Journal of Elections, Public Opinion and Parties, 26:2, 135-154. Coffé, Hilde, and Marieke Voorpostel (2010). ‘Young people, parents and radical right voting. The Case of the Swiss People’s Party’, Electoral Studies, 29:3, 435-443. Dahlstrom, Carl, and Anders Sundell (2012). ‘A losing gamble. How mainstream parties facilitate anti- immigrant party success’, Electoral Studies, 31:2, 353-363. Drazen, Allan, and Nils Gottfries (1994a). ‘Seniority rules and the persistence of unemployment’, Oxford Economic Papers, 228-244. Drazen, Allan, and Nils Gottfries (1994b). ‘Seniority Rules and the Persistence of Unemployment’, Oxford Economic Papers, 46:2, 228-244. Dunn, Kris (2015). ‘Preference for radical right-wing populist parties among exclusive-nationalists and authoritarians’, Party Politics, 21:3, 367-380. Eatwell, Roger, and Cas Mudde (2004). Western democracies and the new extreme right challenge. London; New York: Routledge. Ebert, K, and D Okamoto (2015). ‘Legitimating Contexts, Immigrant Power, and Exclusionary Actions’, Social Problems, 62:1, 40-67. Emmenegger, Patrick (2009). ‘Barriers to entry: insider/outsider politics and the political determinants of job security regulations’, Journal of European Social Policy, 19:2, 131-146. Esping-Andersen, Gosta (2013). The Three Worlds of Welfare Capitalism. London: John Wiley & Sons. Evans, Geoffrey, and Jon Mellon (2016a). ‘Working Class Votes and Conservative Losses: Solving the UKIP Puzzle’, Parliamentary Affairs, 69:2, 464-479. Evans, Jocelyn AJ (2005). ‘The dynamics of social change in radical right-wing populist party support’, Comparative European Politics, 3:1, 76-101. Evans, Jocelyn, and Gilles Ivaldi (2010). ‘Comparing forecast models of Radical Right voting in four European countries (1973-2008)’, International Journal of Forecasting, 26:1, 82-97. Fischer, Stanley, and National Bureau of Economic Research (1986). NBER macroeconomics annual 1986. Cambridge, Massachusetts; London: MIT Press. Fitzgerald, Jennifer, David Leblang, and Jessica C Teets (2014). ‘Defying the Law of Gravity: The Political Economy of International Migration’, World Politics, 66:3, 406-445. Ford, Robert, and Matthew J Goodwin (2010). ‘Angry White Men: Individual and Contextual Predictors of Support for the British National Party’, Political Studies, 58:1, 1-25. Givens, Terri E (2004). ‘The Radical Right Gender Gap’, Comparative Political Studies, 37:1, 30-54. Gidron, Noam, and Peter Hall (2017). ‘Populism as a Problem of Social Integration’, Annual Meeting of the American Political Science Association, San Francisco, September. Vol. 1. 2017.. Gómez-Reino, Margarita, and Iván Llamazares (2013). ‘The Populist Radical Right and European Integration: A Comparative Analysis of Party?Voter Links’, West European Politics, 36:4, 789- Dualism, Welfare State, and the Populist Radical Right lv

816. Golder, Matt (2003). ‘Explaining Variation In The Success Of Extreme Right Parties In Western Europe’, Comparative Political Studies, 36:4, 432-466. Golder, Matt (2016). ‘Far-right Parties in Europe’, Annual Review of Political Science, 19:1, 477-497. Goodwin, Matthew J, David Cutts, and Laurence Janta-Lipinski (2016). ‘Economic losers, protestors, islamophobes or xenophobes? Predicting public support for a counter-Jihad movement’, Political Studies, 64:1, 4-26. Grubb, David, and William Wells (1993). ‘Employment regulation and patterns of work in EC countries’, OECD Economic studies, 7-7. Guiso, Luigi, Helios Herrera, and Massimo Morelli (2017). ‘Demand and Supply of Populism’, , Rochester, NY: Social Science Research Network https://papers.ssrn.com/abstract=2924731 (Accessed October 22, 2017). Hainmueller, Jens, and Daniel J Hopkins (2014). ‘Public Attitudes Toward Immigration’, Annual Review of Political Science, 17:1, 225-249. Hajnal, Zoltan, and Michael U Rivera (2014). ‘Immigration, Latinos, and White Partisan Politics: The New Democratic Defection: Immigration, Latinos, and White Partisan Politics, American Journal of Political Science, 58:4, 773-789. Harmon, Nikolaj A (2017). ‘Immigration, Ethnic Diversity, and Political Outcomes: Evidence from Denmark’, The Scandinavian Journal of Economics, http://onlinelibrary.wiley.com/doi/10.1111/sjoe.12239/full (Accessed June 29, 2017). Harteveld, Eelco (2016). ‘Winning the losers but losing the winners? The electoral consequences of the radical right moving to the economic left’, Electoral Studies, 44, 225-234. Harteveld, Eelco, and Elisabeth Ivarsflaten (2016). ‘Why Women Avoid the Radical Right: Internalized Norms and Party Reputations’, British Journal of Political Science, 1-16. Hatton, Timothy J (2016). ‘Immigration, public opinion and the recession in Europe’, Economic Policy, 31:86, 205-246. Hausermann, Silja, and Hanna Schwander (2012). ‘Varieties Of Dualization? Labor Market Segmentation and Insider-Outsider Divides Across Regimes’, in Patrick Emmenegger,Silja Hausermann,Bruno Palier,and Martin Seeleib-Kaiser (eds.), The Age of Dualization: The Changing Face of Inequality in Deindustrializing Societies, vol., International Policy Exchange Series. Oxford: Oxford University Press, 27-51. Häusermann, Silja, Hanna Schwander, and T Kurer (2014). ‘Explaining welfare state preferences in dualized societies’, Unpublished Manuscript. Presented at ECPR Joint Sessions, Mainz. Häusermann, Silja, and Stefanie Walter (2010). "Restructuring Swiss welfare politics: post-industrial labor markets, globalization, and Welfare Values" in Simon Hug and Hanspeter Kriesi (eds.) Value Change in Switzerland. Lanham: Lexington Press, 143-68, Hernández, Enrique, and Hanspeter Kriesi (2016). ‘The electoral consequences of the financial and economic crisis in Europe’, European Journal of Political Research, 55:2, 203-224. Hooghe, Liesbet, Gary Marks, and Carole J Wilson (2002). ‘Does Left/Right Structure Party Positions on European Integration?’, Comparative Political Studies, 35:8, 965-989. Inglehart, Ronald, and Pippa Norris (2016). ‘Trump, Brexit, and the Rise of Populism: Economic Have- Nots and Cultural Backlash’, , Rochester, NY: Social Science Research Network https://papers.ssrn.com/abstract=2818659 (Accessed December 3, 2016). Ivarsflaten, Elisabeth (2005). ‘The vulnerable populist right parties: No economic realignment fuelling their electoral success’, European Journal of Political Research, 44:3, 465-492. Jackman, Robert W, and Karin Volpert (1996). ‘Conditions favouring parties of the extreme right in Western Europe’, British Journal of Political Science, 26:4, 501-521. Jenkins, J Craig (n.d.). The Politics of Social Protest: Comparative Perspectives on States and Social Movements. University of Minnesota Press. Jensen, Mads Dagnis, and Holly Snaith (2016). ‘When politics prevails: the political economy of a Brexit’, Journal of European Public Policy, 23:9, 1302-1310. van Kessel, Stijn (2015). Populist Parties in Europe. London: Palgrave Macmillan UK. 56 The reconfiguration of working class politics

Kestilä, Elina, and Peter SöDerlund (2007). ‘Subnational political opportunity structures and the success of the radical right: Evidence from the March 2004 regional elections in France’, European Journal of Political Research, 46:6, 773-796. Kitschelt, Herbert (1988.). ‘Left-Libertarian Parties: Explaining Innovation in Competitive Party Systems', World Politics, 40:2. Kitschelt, Herbert (1995). ‘Formation of Party Cleavages in Post-Communist Democracies: Theoretical Propositions’, Party Politics, 1:4, 447-472. Kitschelt, Herbert (2007). ‘Growth and Persistence of the Radical Right in Postindustrial Democracies: Advances and Challenges in Comparative Research’, West European Politics, 30:5, 1176-1206. Kriesi, Hanspeter et al. (2006). ‘Globalization and the transformation of the national political space: Six European countries compared’, European Journal of Political Research, 45:6, 921-956. Kriesi, Hanspeter, ed. (2008). West European politics in the age of globalization. Cambridge, UK ; New York: Cambridge University Press. Kriesi, Hanspeter, ed. (2012). Political conflict in Western Europe. Cambridge ; New York: Cambridge University Press. Kriesi, Hanspeter (2014). ‘The Populist Challenge’, West European Politics, 37:2, 361-378. Lefkofridi, Zoe, Markus Wagner, and Johanna E Willmann (2014). ‘Left-Authoritarians and Policy Representation in Western Europe: Electoral Choice across Ideological Dimensions’, West European Politics, 37:1, 65-90. Lindbeck, Assar, and Dennis J. Snower (1985). ‘Explanations of unemployment, Oxford Review of Economic Policy, 1:2, 34-59. Lindbeck, Assar, and Dennis J Snower (1988a). ‘Cooperation, Harassment, and Involuntary Unemployment: An Insider-Outsider Approach’, The American Economic Review, 78:1, 167- 188. Lindbeck, Assar, and Dennis J Snower (1988b). ‘Cooperation, Harassment, and Involuntary Unemployment: An Insider-Outsider Approach’, The American Economic Review, 78:1, 167- 188. Lindvall, Johannes, and David Rueda (2014). ‘The Insider-Outsider Dilemma’, British Journal of Political Science, 44:02, 460-475. Lipset, Seymour Martin (1959). ‘Democracy and Working-Class Authoritarianism’, American Sociological Review, 24:4, 482. Loxbo, Karl (2014). ‘Voters’ Perceptions of Policy Convergence and the Short-term Opportunities of Anti-immigrant Parties: Examples from Sweden: Voters’ Perceptions of Policy Convergence and the Short-term Opportunities of Anti-immigrant Parties’, Scandinavian Political Studies, 37:3, 239-262. Lubbers, Marcel, Mérove Gijsberts, and Peer Scheepers (2002). ‘Extreme right-wing voting in Western Europe’, European Journal of Political Research, 41:3, 345-378. Lubbers, Marcel, and Jochem Tolsma (2011). ‘Education’s impact on explanations of radical right-wing voting’, in Interdisciplinary Conference on Migration, Economic Change, Social Challenge, April 6th-9th, 2011 University College London, vol. London: University College London http://cream.conference-services.net/resources/952/2371/pdf/MECSC2011_0218_paper.pdf (Accessed June 29, 2017). Lucassen, Geertje, and Marcel Lubbers (2012). ‘Who fears what? Explaining far-right-wing preference in Europe by distinguishing perceived cultural and economic ethnic threats’, Comparative Political Studies, 45:5, 547-574. Magalhães, Pedro C (2014). ‘Introduction - Financial Crisis, Austerity, and Electoral Politics’, Journal of Elections, Public Opinion and Parties, 24:2, 125-133. Mair, Peter (2013). Ruling the void: the hollowing of Western democracy. Verso. Manatschal, Anita (2015). ‘Switzerland - Really Europe’s Heart of Darkness?’, Swiss Political Science Review, 21:1, 23-35. Dualism, Welfare State, and the Populist Radical Right lvii

Marx, Paul (2014). ‘Labour market risks and political preferences: The case of temporary employment: Labour market risks and political preferences’, European Journal of Political Research, 53:1, 136-159. Marx, Paul (2015). The political behaviour of temporary workers. Springer. Marx, Paul, and Georg Picot (2013). ‘The party preferences of atypical workers in Germany’, Journal of European Social Policy, 23:2, 164-178. Mayer, N (2013). ‘From Jean-Marie to Marine Le Pen: Electoral Change on the Far-right’, Parliamentary Affairs, 66:1, 160-178. Middendorp, Cees P (1993). ‘Authoritarianism: Personality and ideology’, European Journal of Political Research, 24:2, 211-228. Mierina, Inta, and Ilze Koroleva (2015). ‘Support for Far-right Ideology and Anti-Migrant Attitudes among Youth in Europe: A Comparative Analysis’, The Sociological Review, 63:2_suppl, 183- 205. Mudde, Cas (2007). Populist radical right parties in Europe. Cambridge, UK; New York: Cambridge University Press. Mudde, Cas (2010). ‘The Populist Radical Right: A Pathological Normalcy’, West European Politics, 33:6, 1167-1186. Mudde, Cas, and Cristóbal Rovira Kaltwasser (2013). ‘Exclusionary vs. Inclusionary Populism: Comparing Contemporary Europe and Latin America’, Government and Opposition, 48:2, 147- 174. Nickell, Stephen, and Sushil Wadhwani (1990). ‘Insider Forces and Wage Determination’, The Economic Journal, 100:401, 496. Norris, Pippa (2005). Radical Right: Voters and Parties in the Electoral Market. Cambridge University Press. Oesch, Daniel (2006). Redrawing the Class Map. Stratification and institutions in Germany, Britain, Sweden and Switzerland. London: Palgrave Macmillan UK. Oesch, Daniel (2008). ‘Explaining Workers’ Support for Right-Wing Populist Parties in Western Europe: Evidence from Austria, Belgium, France, Norway, and Switzerland’, International Political Science Review, 29:3, 349-373. Pardos-Prado, Sergi (2015). ‘How Can Mainstream Parties Prevent Niche Party Success? Center- Right Parties and the Immigration Issue’, The Journal of Politics, 77:2, 352-367. Piore, Michael J (1978). ‘Dualism in the Labor Market : A Response to Uncertainty and Flux. The Case of France’, Revue économique, 29:1, 26-48. Polavieja, Javier G (2016). ‘Labour-market competition, recession and anti-immigrant sentiments in Europe: occupational and environmental drivers of competitive threat’, Socio-Economic Review, 14:3, 395. Richards, Andrew, and Javier García de Polavieja (1997). Trade Unions, Unemployment and Working Class Fragmentation in Spain. Instituto Juan March de estudios e investigaciones. Rooduijn, Matthijs (2015). ‘The rise of the populist radical right in Western Europe’, European View, 14:1, 3-11. Rooduijn, Matthijs (2017). ‘What unites the voter bases of populist parties? Comparing the electorates of 15 populist parties’, European Political Science Review, 1-18. Rovny, Allison E, and Jan Rovny (2017). ‘Outsiders at the ballot box: operationalizations and political consequences of the insider-outsider dualism’, Socio-Economic Review, mww039. Rueda, David (2005). ‘Insider-Outsider Politics in Industrialized Democracies: The Challenge to Social Democratic Parties’, American Political Science Review, 99:01. Rueda, David (2006). ‘Social Democracy and Active Labour-Market Policies: Insiders, Outsiders and the Politics of Employment Promotion’, British Journal of Political Science, 36:03, 385. Rueda, David (2007). Social democracy inside out: partisanship and labor market policy in industrialized democracies. Oxford: Oxford University Press. Rydgren, Jens (2008). ‘Immigration sceptics, xenophobes or racists? Radical right-wing voting in six West European countries’, European Journal of Political Research, 47:6, 737-765. Rydgren, Jens, and Patrick Ruth (2011). ‘Voting for the Radical Right in Swedish Municipalities: Social Marginality and Ethnic Competition?: Voting for the Radical Right in Swedish Municipalities’, 58 The reconfiguration of working class politics

Scandinavian Political Studies, 34:3, 202-225. Saint-Paul, Gilles (1993). ‘On the Political Economy of Labor Market Flexibility’, NBER Macroeconomics Annual, 8, 151-187. Saint-Paul, Gilles (1997a). ‘The Rise and Persistence of Rigidities’, The American Economic Review, 87:2, 290-294. Saint-Paul, Gilles (2002). ‘The Political Economy of Employment Protection’, Journal of Political Economy, 110:3, 672-704. Saint-Paul, Gilles, Charles R Bean, and Giuseppe Bertola (1996). ‘Exploring the Political Economy of Labour Market Institutions’, Economic Policy, 11:23, 263. Scheepers, Peer, Rob Eisinga, and Jan Lammers (1993a). ‘Het electoraat van de Centrum Partij/Centrum Democraten in de periode 1982-1992’, Mens en Maatschappij, 68:4, 362-385. Scheepers, PLH, Jaak Billiet, and H de Witte (1995). ‘Het electoraat van het Vlaams Blok: De kiezers en hun opvattingen’, http://repository.ubn.ru.nl/bitstream/handle/2066/14396/3312.pdf (Accessed December 8, 2017). Schwander, Hanna, and Silja Häusermann (2013). ‘Who is in and who is out? A risk-based conceptualization of insiders and outsiders’, Journal of European Social Policy, 23:3, 248-269. Shapiro, Carl, and Joseph E Stiglitz (1984). ‘Equilibrium Unemployment as a Worker Discipline Device’, The American Economic Review, 74:3, 433-444. Shapiro, Carl, and Joseph E Stiglitz (1985). ‘Can Unemployment Be Involuntary? Reply’, The American Economic Review, 75:5, 1215-1217. van Spanje, Joost, and Claes de Vreese (2015). ‘The good, the bad and the voter: The impact of hate speech prosecution of a politician on electoral support for his party’, Party Politics, 21:1, 115- 130. Spierings, Niels, and Andrej Zaslove (2017). ‘Gender, populist attitudes, and voting: explaining the gender gap in voting for populist radical right and populist radical left parties’, West European Politics, 40:4, 821-847. Spies, Dennis, and Simon T Franzmann (2011). ‘A Two-Dimensional Approach to the Political Opportunity Structure of Extreme Right Parties in Western Europe’, West European Politics, 34:5, 1044-1069. Stockemer, Daniel (2012). ‘The Swiss Radical Right: Who are the (new) voters of the Swiss People Party?’, Representation, 48:2, 197-208. Swank, Duane, and Hans-Georg Betz (2003). ‘Globalization, the welfare state and right-wing populism in Western Europe’, Socio-Economic Review, 1:2, 215-245. Taggart, Paul (1995). ‘New populist parties in Western Europe’, West European Politics, 18:1, 34-51. Werts, Han, Peer Scheepers, and Marcel Lubbers (2013). ‘Euro-scepticism and radical right-wing voting in Europe, 2002-2008: Social cleavages, socio-political attitudes and contextual characteristics determining voting for the radical right’, European Union Politics, 14:2, 183-205.

Workers’ preferences on flexicurity lix

Chapter 4. Workers’ preferences on flexicurity

While voting decisions is of course a crucial dimension of changing political dynamics, the policy side cannot and should not be dismissed. Too often we tend to divorce both aspects of political reality as if they belonged to completely separated analytical worlds. That is why I decided to devote a chapter within the present work to preferences over a key policy field for my segment of interest: labour market regulation. According to many of its advocates, labour market reforms have a compromised policy agreement at hand, ready to use to avoid extreme outsets: Flexicurity, defined as some form of cash (non- employment benefits or NEB, active labour market policies or ALMP) in exchange for flexibility (on firing and hiring norms, i.e. employment protection legislation or EPL). Instead, liberalisation (reducing EPL, ALMP, NEB across the board) is normally seen as a more confrontational project, one led by employers’ interests. Since flexicurity arrangements are touted by its advocates as an available, well- balanced equilibrium, there remains a question on why flexicurity is not more enshrined in current policy in advanced capitalist societies: composite indexes such as Tangian (2004) and Maselli (2010) based upon OECD data tend to show that only Scandinavian countries (Denmark, Sweden, Norway) plus the Netherlands display clear-cut flexicurity structures. Although there might be supply-sided answers to the puzzle (see next chapter for an angle on this), this chapter is devoted to the demand side of policymaking and will examine the question of where exactly the demand for a flexicurity reform is, and how it compares to its much more prevalent opposite, i.e. dualisation. The insider/outsider framework (e.g. Saint-Paul 1992; Rueda 2005) would explain pro-flexicurity support in terms of labour market position: Workers under permanent contracts (insiders) will be against it as it seriously harms their acquired status, by decreasing their protection — most of which was obtained thanks to EPL’s imposed cost on hiring and firing that reduces job rotation — by increasing it in other fronts that do not directly benefit them. Conversely, outsiders (the unemployed and workers under non-standard contracts) would have no interest in EPL and may very well favour its demise as slowing job rotation conflicts with their interests, but they would very much favour NEB and ALMP hikes. From this point of view, flexicurity does not dominate the policy landscape because outsiders hold less political power than their insider counterparts. An alternative view holds that, if there is a boundary across which workers can be divided amongst themselves, it tends to follow a rather classical occupation/skill split. In other words: it is the more permanent feature of socio-economic class rather than mere job status that determines preferences on labour reform. Correspondingly, preferences on labour market policy follow a more standard division between full-scale protection (increasing state intervention through all three dimensions) and liberalisation (the reverse strategy). Thus, flexicurity is not more common because there is little demand for it. In an attempt to overcome the debate, some views have defended that both skill/occupation and the job status perspective should be combined to fully grasp who might be in favour of flexicurity (Boeri et al. 2012). I conduct a preliminary analysis that, to my knowledge, is the first one to systematically feature all three key dimensions defining flexicurity (EPL, NEB, ALMP) for a wide range of Continental European countries, pitching at the same time main explanatory frameworks against each other. Thus, the main contribution of the present chapter is empirical rather than theoretical. I come to find that, indeed, the employed/unemployed divide gives testable clues to the idea of why and how a pro-flexicurity coalition might be crafted, but occupational differences convey similar explanatory power. As expected by the insider/outsider framework, the unemployed are more likely than the employed to prefer EPL reductions, ALMP/NEB increases, and thus flexicurity over dualisation. But they do not prefer liberalisation over full-scale protection (keeping EPL, NEB and ALMP at high levels) compared to the employed. At the same time, employers and professionals do display more intense preferences than mid- and low-skilled manual workers towards both flexicurity and liberalisation. If anything, and although this is only an exploratory approach that refrains from establishing strong conclusions, the occupational framework delivers somewhat more promising insights, but the insider/outsider perspective remains clearly relevant to answer the question at hand. The remainder of the chapter is divided in five parts. The next section surveys the literature departing from the notion of dualisation. After, I proceed to describe my argument and hypotheses. Section 3 specifies my methodological approach, section 4 offers the results, and section 5 concludes. 60 The reconfiguration of working class politics

4.1 The insider/outsider framework and its main critiques

4.1.1 The basic framework The idea of labour as a quasi-fixed factor has been with us since at least Oi (1962). However, it seems fair to say that the literature on dualisation and segmented labour markets began with the work of Michael Piore and associated authors (prominently Berger and Piore 1980). Doeringer and Piore (1971) defined the notion of ‘Internal Labor Markets’ (ILMs): within-firm institutional arrangements that sheltered the labour force from external (i.e. general) labour market pressures. As a partial derivation of this literature, since the 1980s a whole strand of labour economics was developed, devoted to analysing the differentiated bargaining power of workers depending on their degree of exposure to market risks. Pioneering work by Solow (1985), Blanchard and Summers (1986), Lindbeck and Snower (1986; 1988), Nickell and Whadwhani (1990) or Jimeno and Toharia (1993), among others, pointed to the bargaining power enjoyed by employees under a permanent contract due to their degree of job protection, setting wages among the market-clearing level (Drazen and Gottfries 1994). Saint-Paul (1993) gave the literature a political-economy spin when he took it to the political arena, posing that insiders pushed for maintaining a high level of employment protection legislation and blocked liberalising reform, whereas outsiders would be more interested in a more liberalising framework that allowed them to compete in equal conditions. Saint-Paul (1993) and his subsequent work (Saint-Paul 1997; Saint-Paul 2002), as well as that of Boeri, Borsch-Supan and Tabellini (2001), provided Rueda (2005) and Rueda (2007) with a core insight to consolidate the insider-outsider framework: due to economic factors, and because unemployment disproportionately effects a part of labour, interests of insiders and outsiders are not aligned. Rueda comes to argue that, while insiders hold a stronger preference for EPL, outsiders prefer to push for ALMP and NEB, and may be ready to accept a lower level of employment protection. Due to the differences in their level of exposure to the risk of unemployment and the subsequent differences in policy preferences, social democratic parties face a dilemma between catering to the preferences of their core constituencies (which are traditionally understood to be insiders) and moving towards a policy platform that leans more towards outsiders. It is also suggested that liberal and conservative parties could in turn craft an outsider-upscale coalition6. Although this is only a possibility, for which Rueda (2005) finds no support as conservative parties do not seem willing to provide higher active or passive labour market policies, it is left open within the theoretical proposition. In any case, outsiders might turn to options from the far-left to the far-right (Rueda 2007) in case social democrats do not provide them with the policy changes they demand. Lindvall and Rueda (2013) follow the interactions between party platforms and differentiated electoral support in 1990s’ Sweden to argue that outsiders respond to social democrats’ strategies and vice versa: if parties come close to insiders, outsiders will either leave the political process or go to the extremes of the spectrum. Conversely, if they come closer to outsiders, insiders will punish them.

4.1.2 The class/occupation critique Contrary to the insider/outsider framework, some argue that the classical view stemming from power resource theory still holds: division does not cut across the working class so starkly. Instead, they argue that the working class remains united in its demand for increased protection on all fronts. The literature on how different forms of risk affect political behaviour comes mostly to substantiate the correlation between risk exposure and increasing preferences towards protection and redistribution both at a macro (Cusack et al. 2006) and a micro level (Mughan 2007; Rehm 2009; Marx 2013), even when accounting for income (Rehm, Hacker and Schlesinger 2012). More specifically, there is a certain consensus on considering job insecurity as a factor explaining pro-protection, left-leaning political preferences (e.g. Mughan and Lacy 2002; Cusack et al. 2006; Hausermann 2010; Rehm 2011). Departing from such premise, Emmenegger (2009) offers a compelling critique of the insider/outsider framework. He poses that the crucial (ever-used) distinction between employers and labour still holds as the main cleavage across the individuals’ political preferences is based upon the person’s position in the production structure. Among the reasons supporting this evaluation two stand out due to their

6 I use 'upscale' in the Rueda (2005) sense here: as a high-skilled worker in a managerial, high-paying job. Workers’ preferences on flexicurity lxi

theoretical relevance. First and foremost, Emmenegger (2009) poses that electoral platforms constitute closed packages, and one cannot pick and choose among the different dimensions, separating for instance higher unemployment benefits from higher employment protection or taxes. Moreover, Emmenegger argues, outsiders may be well-aware of the danger of transferring power from the political left to the right side of the spectrum. Such a change in the balance of power would help upscale groups and other constituencies belonging to liberal and conservative parties to advance their policies in other fields, potentially clashing with outsiders’ preferences. Acknowledging this fact would limit the possibilities of conservative powers to craft cross-class coalitions. This potential contradiction is also addressed tangentially by Rueda (2007) when discussing the role of corporatism and unions: It remains unclear which would be the choice of both insiders and outsiders if we alert the actors to the balance of powers as well as make other policy issues relevant for the model. Regional variability should be considered as well. For instance, King and Rueda (2008) expose the existence of different sources of what they call ‘cheap labour’ in distinct Western countries. Emmenegger et al. (2012) make different proposals to reconsider what an outsider is and how we shall classify it. These and other works discuss whether the insider/outsider division always corresponds to the lines delineated by the original framework: while in some countries the working class breach will be marked by non-standard workers through dual employment protection legislation, in others employees with standard contracts will have a much wider wage distribution, restricting the creation and maintenance of cheap labour on the lowest tier of such distribution. There is a key insight at the core of this critique: flexicurity is a proposal that hardly fits the left-right ideological axis, since it implies a compromise featuring a policy that is usually associated with right- leaning positions (EPL reduction) and a much more left-wing policy of increasing public expenditure (NEB/ALMP expansion). The consequences of flexicurity for the balance of power might therefore not be clearly understood by the actors involved, and if a risk-averse position is taken towards the uncertain outcome of a reform process, each one of them might instead resort to more traditional platforms.

4.1.3 Combined approaches Another way to look at the ‘insider/outsider vs occupation’ debate comes by piecing out the different dimensions that naturally compose any workers’ identity. For instance, Boeri et al. (2012) predict two conflicts on how to determine EPL and NEB. One occurs between insiders/outsiders, while the other between high and low skilled workers. With that in mind, the authors speculate (and find) that flexicurity emerges in countries with a high share of high-skilled workers, which falls in line with more recent findings using novel datasets (most significantly Busemeyer et al. 2018) pointing to these segments of the workforce supporting ALMP and, in general, so-called social investment models that feature not only ALMP but also higher education or childcare expenditure (Garritzmann et al. 2018). At the same time, their macro hypothesis is based on a theoretical model that departs from the individual level, asserting that low-skilled workers will be interested in EPL as it helps them to protect their positions. Since Boeri et al. do not test individual preferences, much less combined policy preferences, their work begs the question of which dimension would it be prioritised by individuals, or what would be the role played by each of these and other determinants that define any persons’ position within the labour market. The work of Hausermann and Schwander (2011) as well as Hausermann and Schwander (2012) addresses this issue by using multi-dimensional categorization of labour market positions in order to expose where the fault lines are by testing individual level preferences on labour market policies. Hausermann and Schwander (2011) use three variables to distinguish between outsiders, insiders and, also between high-skilled and low-skilled workers within each group: age, gender, skill level and a comprised version of Oesch (2006) class scheme, featuring five groups: capital accumulators, mixed service public servants, low service public servants, blue collar workers and socio-cultural professionals. They run a series of cross-tabulation analyses in a wide sample of European countries classified in the four classical welfare regimes, since they consider context to be essential to determine ‘outsiderness’: liberal welfare states, Nordic welfare states, Continental welfare states and Southern European welfare states. For them, 'outsider' is a label that corresponds to those social groups (created by crossing all the aforementioned variables) whose proportion of non-standard employment, part-time employment and unemployment is among the average, 'insiders' being the 62 The reconfiguration of working class politics

rest. The 'status' variable typical of the insider/outsider classical framework is intentionally excluded in order to compare its explanatory power with the authors’ proposed operationalisation. Hausermann and Schwander also follow a complex approach to build their dependent variable, comprising several policy dimensions operationalised through survey questions on social policy preferences, grouped according to factor analysis in four different protection models: liberal, protective, pro-investment and pro-redistribution. Their findings demonstrate that outsiders generally want more protective state intervention, while high- skilled workers in general are interested in increased state’s efforts on education and employability. More specifically, high-skilled insiders are in favour of liberalisation (specifically on NEB and pensions), while low-skilled outsiders generally prefer higher protection (NEB, interventionism, and redistributive measures). High-skilled outsiders are in favour of social investment models (featuring high ALMP investment and low NEB spending). Also, when they compare the explanatory power of their definition of insider/outsider with the standard, status-based definition, theirs emerges as victorious. Interestingly, though, there is no indicator referring to EPL, which is at the heart of Rueda’s argument. Its absence comes with some other methodological issues that complicate the interpretation of Hausermann and Schwander (2009). First, the dataset they use (International Social Survey Programme 2006) is limited to a single year and to an incomplete list of countries (Australia, Canada, Denmark, Finland, France, Germany, Great Britain, Ireland, New Zealand, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland and the United States) lacking several key European countries and including instead Australia, New Zealand or the US. This matters because the discussion on dualisation and its political effects has been usually limited to Europe. Second, there is a trade-off inherent to building complex, multi-dimensional variables: as variables are added to the right and left side of any equation, the explanatory power of the model logically increases. However, the equation becomes over-charged to the point that its explanatory power becomes also self-evident. In other words, there should be an equilibrium to be found between explanatory power and parsimony. To overcome the lack of a direct test on EPL preferences, Amable (2013) analyses the proposal for a single contract in France (contrat de travail unique or CTU in French). The author uses both the status-based insider/outsider operationalisation and the one proposed by Hausermann and Schwander (2009). He finds service workers and to some extent the unemployed to be for a CTU reform, the evidence being stronger for the former group. Nonetheless, Amable underlines that the groups most in favour of an EPL reduction could be seen as insiders from both perspectives, since they are “wealthy, senior, have medium or high education/skill levels or responsibility positions” (Amable 2013: 30). This contribution, although limited to a single country, is very valuable by itself, and should be regarded as a single case-based test of the competing validity of different factors towards drastic EPL reforms.

4.2 What works best? The literature leaves the reader with conclusions that are somewhat mixed; the only commonality is that workers do not seem to be united on their policy preferences, but the question of what the key dividing lines across these preferences are persists. To add a bit of clarity, I intend to focus the present contribution on empirical analysis. On the side of the dependent variable, a reasonable first step is to delimitate the essential components of a flexicurity regime or, for that matter of any other labour market regulation regime. Following Tangian (2004), Maselli (2010), as much as the earliest definitions of the concept, a reduction of employment protection legislation (EPL) along with an increase of unemployment benefits (NEB) and active labour market policies (ALMP) is widely accepted as comprising the basic traits of flexicurity. Note that all three are also chosen by Rueda’s foundational work, as well as by his critics. These could therefore be considered the key dimensions of interest. Before identifying support or rejection towards complex policy packages, one should be able to piece down the preferences towards the components themselves. Here is where my first set of hypothesis comes into play. Before specifying them, I should choose a way to model my independent variables. The most essential version of the core debate on the determinants of labour market regulation preferences pitches the status-based insider/outsider framework against a more classical division based on skill or occupation. Attempts at finding common ground are undoubtedly valuable, and though they might be a Workers’ preferences on flexicurity lxiii

promising route, it is also one to be taken with caution, given the risk of over-defining the independent variable. Thus, I opt here to take a step back and to use the most naked, simplified versions of the contending sides of the argument. According to the insider/outsider framework, it should be expected that outsiders, and most especially the unemployed (whose position as job seekers is similar at its essence across countries), are more interested in all the traits defining flexicurity. At the same time, both the class critique and the combined perspectives hypothesise a positive relationship between cutting down passive protection mechanisms (EPL, NEB) and stepping up active scheme, unlike Rueda, who would expect (and found in his classic work) that interest on ALMP was less intense among the employed. On this matter, I side with the aforementioned critique as it seems plausible for the employed to find a more immediate way to capitalise on further skill training. Therefore, I expect that: H1a. The unemployed are more likely than the employed to prefer EPL reductions, but less likely to prefer NEB reductions and ALMP increases. H1b. Employers and high-level occupations are more likely to prefer ALMP increases and NEB/EPL reductions than mid- and low-level occupations. Moving now to the ultimate focus of the present chapter, a necessary second step is to establish a proper definition that can be easily generalised and operationalised. In truth, despite several attempts, there is hardly a standard operationalisation of concepts like ‘flexicurity’ or ‘liberalisation’, which becomes evident when one realises that all the referenced works hardly share a common dependent variable. I propose here an operational framework, based on Thelen's (2014) definition of labour market regimes, that also matches the most clear-cut attempts to measure flexicurity (Tangian 2004, Maselli 2010). Figure 4.1. Spectrum of labour market policy packages.

Own elaboration based on Thelen (2014). 64 The reconfiguration of working class politics

There are two key dimensions of protection when it comes to labour market policies. EPL corresponds to the sphere of regulation, focusing on protecting a job position by making it harder to fire whoever occupies it. NEB and ALMP, conversely, protect the worker especially when she is in the process of finding a new job. If both dimensions are maximised, then we could talk about a ‘protectionist’ approach to labour market policies. If neither is, ‘liberalisation’ is in place because supply and demand of labour flows freely. These two should therefore be understood as two extremes of a continuum, usually referred to in the most classic works of labour market regulation, or in those that take a power resources approach that tend to expect upscale workers and especially employers to support full-scale deregulation. Workers, on the other hand, (and most significantly low-skilled workers) would prefer the opposite, protectionist option to shelter themselves against the risk of unemployment. The insider/outsider literature, instead, highlights the distinction between those policy packages that protect the job position at the expense of protecting the worker (considered as segmenting or dualising since they are supposed to enlarge the gap between protected and unprotected, between insiders and outsiders), and those that protect the worker at the cost of reducing EPL. In consequence, these perspectives highlight the difference between stable and precarious workers when choosing policy packages. From this point of view, each of these two extreme perspectives best explain a different dilemma: one is better to expose an eventual choice between all or nothing, while the other serves best to highlight the distinction between those who prefer protecting the job position, and those who would rather move towards an individual-based protection system. Based on Boeri et al., it could be however argued that the occupational dimension also relates to this latter dilemma. Therefore, my expectations regarding broad policy packages are the following: H2a. The unemployed are more likely than the employed to prefer flexicurity arrangements, while there should not be significant differences between the two groups on their preferences regarding liberalisation. H2b. High-level occupations and employers are more likely to prefer liberalisation than low- level occupations, than they are to prefer flexicurity.

4.3 Operationalisation & methods To test my hypotheses, I combine several dimensions in a parsimonious way to build both my dependent and my independent variable (DV-IV). First, I use a very limited number of dimensions. Second, in an attempt to maximise analytical transparency and robustness I provide results not only for the relations between the final, combined variables, but also for its components. I also restrict the analysis to countries where the policy choices I am considering carry particular relevance, i.e. Scandinavian, Continental (Esping-Andersen 1990) and Southern/Mediterranean (Ferrera 1996) welfare regimes, where the choice between dualization and flexicurity has historically been at the forefront of the labour regulation debate7. To operationalise my dependent variable, I use three Eurobarometer waves (for 2006, 2009 & 2011) that contain three specific questions on EPL, ALMP and NEB. To my knowledge, mine is the first attempt to use a single cross-country data source to investigate these preferences since David Rueda’s 2005 seminal work. Although using the same source (Eurobarometer 1995), the instrument to measure EPL was much more imprecise then, since it consisted of the degree of agreement with the sentence “Job security is very important”. Mine includes a direct question on the matter: Whether the interviewed person would be willing to increase the flexibilisation of contract norms to foster employment creation ("very much agree", "tend to agree", "tend to disagree", "fully disagree"). Admittedly, this question introduces a pro-liberalisation bias in the sample, since it somehow links reducing EPL with a desirable outcome, such as creating new jobs. I acknowledge this fact and take into account that I therefore assume that levels of support for EPL (but also for NEB and ALMP) should be interpreted in a relative fashion, not in absolute terms, and focusing on the most extreme positions. I henceforth avoid drawing absolute conclusions and focus on inter-group comparisons. However, there are simply no other recent, reliable and sample-strong sources to compare EPL support featuring a wide array of countries. Also, given that precedent operationalizations of cross-

7 Austria, Belgium, Switzerland, Germany, Spain, Finland, France, Italy, Netherlands, Norway, Portugal, Sweden. Workers’ preferences on flexicurity lxv

country analysis are arguably weaker, I proceed with the analysis although cautioning against reading its results like nothing but an exploratory, preliminary take on the issue that will need dedicated field research to be further confirmed. Moving on to the NEB variable, it measures the agreement with the proposal of increasing income differences between employed and unemployed (again: ’very much agree’, ‘tend to agree’, ‘tend to disagree’, ‘fully disagree’). The ALMP variable does the same with increasing ALMP expenditure. To increase results’ relevance and minimise pro-liberalising biases as much as possible, when testing preferences for single policies, I isolate the most positive answer ('very much agree' or 'fully disagree') to all three questions to generate a proxy for preference salience. Moreover, I take advantage of the new INVEDUC dataset collected by Busemeyer et al. (2018) to run consistency tests for ALMP and NEB policy preference structures. Independent variables are built by combining the Eurobarometer’s occupational classification and the division drawn by classical political economy of labour market institutions between insiders and outsiders. The class schema, roughly based upon Oesch (2006), comprises ten different groups of active workers/job seekers (I exclude the retired since they are not directly affected by labour market regulation): *1 Large employers and higher-grade managers *2 Professionals *3 Small owners *4 Farmer & fisherman *5 Employed professionals and technical experts *6 Middle and lower-grade managers *7 Clerks, sellers, transporters (mid-level service workers) *8 General service workers *9 Skilled manual workers *10 Unskilled manual workers These are provided for currently employed individuals, as well as for the unemployed. As for job market position, I am unable to exactly reproduce Rueda’s classic operationalization of insiders and outsiders since the Eurobarometer survey has no additional information that would allow me to draw lines among the employed. Instead, I substitute 'insiders' with those employed, while 'outsiders' are proxied by those without a job and actively seeking a position. First, I conduct the analysis using occupation and labour market position as independent variables for all three policy fields (EPL, NEB, ALMP). I use logistic regressions since the possibility of running marginal effects tests perfectly fits my ultimate goal of comparing different probabilities of supporting policy positions within occupational and job position groups. I feature one model specification per IV- DV couple. Aside from my variable of interest, the model is completed with controls for gender, age, country and year fixed effects. Given the idiosyncrasy of the questions used to operationalise my key variables, it seems necessary to run some sort of parallel test to broadly confirm or discard the findings obtained through the Eurobarometer. As stated above, there is no other source of equal quality that features fitting questions for all three policy dimensions, including at the same time so many countries while also giving the possibility to build IVs that match the frameworks to be tested. There are, however, some sources that do include at least NEB and ALMP (surveys on EPL preferences seem to be very uncommon for no apparent reason). Of these, I choose the most recent wave of the European Social Survey (ESS 8th wave - 2016). A special module with questions on NEB and ALMP has been included for a subset of countries that, although does not fully match my broad Eurobarometer selection, is broad enough to grant some insights through comparison. The selected NEB question investigates the degree of agreement with the following proposal: "Unemployment benefit should be cut if an individual turns down a job that requires a lower level of education of that held by the individual", answers ranging from "1: Should lose all unemployment benefit" to "4: Should keep all unemployment benefit". The ALMP question inquiries about the degree of agreement (from 1="strongly against" to 4="strongly 66 The reconfiguration of working class politics

in favour") with the following proposal: "Spend more on education for unemployed at cost of unemployment benefit". Both do differ from the original Eurobarometer questions, but that is part of the reason to choose them: by using different operationalisations for the same DVs, findings should be strengthened and enriched. In a second phase, the operation is repeated but the dependent variable is changed. Now by using the combined version to determine the support for flexicurity and for liberalisation, I build two IVs. ‘Flexicurity’ measures the likeliness of the individual to support a policy package that would include EPL reduction, ALMP increase and the maintenance of NEB levels. ‘Liberalisation’, instead, refers to a reduction in all three dimensions. Recalling that the chosen Eurobarometer questions have four possible positions (1 = “strongly agree” and 4 = “strongly disagree”), I combine them into a five-level ordered variable for each policy package explained in the following table.

Table 4.1. Operationalising preferences for flexicurity and liberalisation: variable definition.

Flexicurity Liberalisation

5 Strongly agreeing with EPL reduction, ALMP increase; Strongly agreeing with EPL/NEB reduction; strongly disagreeing with NEB reduction. strongly disagreeing with ALMP increase.

4 Two out of three of the following: Strongly agreeing with Two out of three of the following: Strongly agreeing EPL reduction/ALMP increase, strongly disagreeing with with EPL/NEB reduction; strongly disagreeing with NEB reduction. The remaining should be “mildly ALMP increase. The remaining should be “mildly agreeing (EPL, ALMP)/disagreeing (NEB)”. agreeing (EPL, NEB)/disagreeing (ALMP)”.

3 One out of three: Either strongly agreeing with EPL One out of three of the following: Strongly agreeing reduction/ALMP increase, or strongly disagreeing with with EPL/NEB reduction; strongly disagreeing with NEB reduction, while mildly agreeing on the other two. ALMP increase. The remaining should be “mildly The remaining should be “mildly agreeing (EPL, agreeing (EPL, NEB)/disagreeing (ALMP)”. ALMP)/disagreeing (NEB)”.

2 Mildly agreeing with EPL reduction, ALMP increase; Mildly agreeing with EPL/NEB reduction; mildly mildly disagreeing with NEB reduction. disagreeing with ALMP increase.

1 Not agreeing with EPL reduction, ALMP increase; and Not agreeing with EPL/NEB reduction; and agreeing with NEB reduction. agreeing with ALMP increase.

This combined approach provides the possibility of obtaining a gradual result. Also, note that the opposite policy position of flexicurity is fairly concurrent with stringent dualisation (maintaining EPL, reducing NEB, not increasing ALMP), while liberalisation would be countered with full-scale protection (maintaining EPL and NEB, increasing ALMP). However, since this approach omits all those positions that do not identify with either of the respective extremes, the sample gets significantly reduced. This, together with the already mentioned pro-liberalisation bias introduced by the formulation of the EPL question, I must insist on the fact that the values that it produces are only relevant in relative terms (i.e. to be compared internally), and therefore should be read as such. Results of all specifications are included in the Appendix. In the following Section I focus on interpreting predicted probabilities since its interpretation is much more straightforward as well as appropriate for my hypotheses.

Workers’ preferences on flexicurity lxvii

4.4 Results Beginning with the preliminary tests applied to single policy fields, Figure 4.2 shows that being unemployed increases the predicted probability of strongly agreeing with an EPL liberalising reform in 7pp, from 0.26 to 0.338 — while not a huge difference, it is a significant one according to Tables C.3 and C.4 in the Appendix. Similar differences appear related to NEB and ALMP reforms, all of them statistically significant both at the original logistic regression and at the average marginal effects test.

Figure 4.2. Marginal probabilities of strongly agreeing on labour market reforms, by job situation, and occupation

Full results on Table C.4, based on logistic regression models from Table C.3.

Differences on all three dimensions are, however, more notable when looking at the occupational division, although its significance on parent regressions is murkier and only goes below p<.1 at the categories with most extreme results (significance on marginal effects tests are, nonetheless, always significant to p<.01). Regarding EPL, there is a range of nearly 12pp in marginal probabilities between mid-range service workers (low-level managers, clerks) and large employers. Also, manual or semi- manual service workers seem to be particularly sceptical. But, significantly enough, no category of employed worker nears the favourability of EPL reduction displayed by those who are either owners or self-employed. General service workers and unskilled manual workers seem to be at par with

8 I mostly focus on interpreting marginal effects since both the value and the p-test of the regression itself are hard to read, as they do not refer to straightforward differences but to odd ratios, and the selection of the base category determines significance to a large extent. 68 The reconfiguration of working class politics

employed professionals on their likelihood to strongly support EPL cuts, albeit far from owners and the self-employed. Mid-range service workers appear, somehow surprisingly, as the least appealed by the prospect of reducing EPL. The picture on NEB reform is a bit more compressed, although the range covers 10pp and is mostly defined by the inclination of small owners to prefer reductions. It is nonetheless interesting to see that employed and self-employed professionals are less likely to strongly want them than skilled manual workers. Mid-level service workers are again at the lowest part in the distribution. Regarding the employed/unemployed divide, it is again significant, but also shows a smaller spread: 6pp between employed and unemployed. At the same time, high-skilled workers and large owners are much more likely to be strongly pro- ALMP. Here the division is more strongly occupation-driven, since job position marks only a 4pp difference on the likelihood to support ALMP increases, with somewhat lower significance in the parent logistic regression. Nonetheless, my hypothesis is confirmed on this as well. In contrast, there is a big gap (18pp) in marginal probabilities between professionals (employed or self-employed) and unskilled manual workers for strongly preferring ALMP hikes. Figures C.7 and C.8 display the result for the alternative operationalisations using the ESS questions. There are several coincidences to be highlighted. First, significance and direction of the results (both in the parent regression and in the marginal effects test) for the employed/unemployed dichotomous IV remains constant. Second, being unemployed clearly marks a more intense effect on preferences over NEB than on preferences over ALMP. Third, significance for the occupation IV in the parent regression is less clear, while the range of variation on marginal probabilities is wider (and overall significant). Fourth, the high-to-low linear relationship between occupational scale and intensity of pro- ALMP sentiment is broadly confirmed, albeit a bit less intense than in the Eurobarometer data; this is probably due to the trade-off nature of the ESS question. Fifth, among top occupations, business owners and high-level managers appear to be more likely than professionals (particularly those under a dependent employment regime) to favour NEB cuts, and skilled manual workers are more prone to them than their unskilled counterparts. There are, however, some interesting differences as well. Especially, the fact that mid-range service workers appear to be more likely than almost any other category to prefer NEB cuts. Moreover, general service and manual workers seem to join employers and self-employed professionals at the same level. This result is however counter-intuitive, not having clear correlates among the main referents of the literature studying welfare preferences. It should therefore be taken with caution. Before further commenting on policy packages, this is a good moment to pause and recall the fact that these values do not necessarily convey the overall likelihood of supporting either policy package among the sample. Since both variables are restricted to those who take a position in either the dualisation-flexicurity axis or the protection-liberalisation axis, many individuals are left out of the analysis, as shown by Tables C.10 to C.14 in the Appendix. Returning to the analysis, as shown by Figure 4.3 (as well as by the full results included in Tables C.5 and C.6 in the Appendix) the distinction between employed and unemployed carries a bit less explanatory power than social class when considering the preferences for flexicurity (making the unemployed more likely to strongly support it, as expected). More specifically, self-employed professionals and employers head the classification of pro-flexicurity groups, nearly tripling unskilled manual workers on the marginal probability of strongly preferring flexicurity over dualisation. Employed professionals are, instead, somewhat left behind on this matter. The job status distinction retains significant power as well to explain different preference intensities: as expected, the employed are twice as likely than the unemployed to strongly prefer dualisation, and the contrary is true regarding flexicurity.

Workers’ preferences on flexicurity lxix

Figure 4.3. Marginal probabilities of supporting liberalisation vs protection and flexicurity vs dualisation, by job situation, and occupation (extreme positions)

Full results on Table C.6, based on ordered logistic regression models from Table C.5.

Preferences on the liberalization-protectionist dilemma display a somewhat lower strength and are generally non-significant in the original regression (albeit retaining significance on the test for average marginal probabilities). Its variance is non-existent when using the simplified version of the insider- outsider framework: both the employed and the unemployed display the same pattern of preferences along this axis. There are nonetheless interesting differences when looking at the occupational dimension. Again, professionals and employers are the most favourable to flexibilising reforms. However, now clerks and skilled manual workers are more inhibited than unskilled workers, somehow paralleling the results obtained for EPL reduction.

4.5 Interpreting and concluding Where, then, does the demand for a flexicurity reform lie? Rather than taking an open-ended approach to this question, I departed from a debate that has been ongoing in the literature for the last couple of decades: According to its proponents, a flexicurity policy package would offer a compromise that moves away from the usual left-right/protectionist-liberal axis. This argument however, leaves the question of why it is so difficult to find a constituency for it if it is in fact the compromise it is touted to be. There are two very different ways to approach this. While the insider/outsider framework poses that there is a pro-flexicurity group but one which is too politically weak to impose its preferences, the skill/occupation perspective over the issue argues that there is no such constituency, and that outsiders would rather have protection or liberalisation depending on their occupational position or skill level. Finally, in the combined framework proposed by other authors, both status and skill/occupation would play a role in determining the support for different labour market policies. Using Eurobarometer data from 2006-2011 with a broad case base that includes all Western European continental countries, results offer new insights on the issue. First, both the insider/outsider and the occupational perspective appear relevant to the issue, and therefore considering both at the same time seems to be advised for further research (either in line with Hausermann and Schwander, 70 The reconfiguration of working class politics

or with other operationalisations) as well as for policy entrepreneurs looking for opportunities and coalition formation. As expected by theory, the unemployed are more likely than the employed to support EPL cuts, while the employed display more intense preferences towards NEB cuts. Regarding ALMP, contra Rueda, it looks like the latter are slightly more inclined than the former to favour increased spending. This is however no impediment for the unemployed to prefer protective flexicurity proposals to a larger extent than the employed, who would instead favour dualisation with further intensity. On the protectionist- liberalisation axis, instead, there is no apparent difference between both categories. The occupational scheme yields other interesting findings. Chiefly, there is indeed a general distinction between employers and high-level positions on one side, and mid- and low-level workers on the other, with the former preferring flexicurity and liberalisation over its alternatives to a larger extent than the latter, while also being more pro-cutting NEB and EPL, but also pro-ALMP. This general picture is greatly nuanced by results. First, the strongest opposition to EPL-NEB cuts as well as to full-scale liberalisation is found among mid-range service workers (middle management, clerks, sellers, transporters) according to Eurobarometer data (not to ESS data), while it is indeed unskilled workers who display a weakest preference towards ALMP hikes (with consensus across sources). Second, professionals do not show a strong appetite for NEB cuts, and only the self-employed seem to look at EPL reductions with comparatively higher interest, as they do (along with employers and again, to an extent confirmed by both ESS and Eurobarometer sources) for flexicurity. Both do, however, show the highest support for ALMP. Third, there is a distinction between employers’ preferences regarding EPL and NEB reduction: while large employers and high-level managers agree to a larger extent than other groups with the former, small business-owners differ significantly when it comes to NEB cuts (particularly when looking at Eurobarometer data), perhaps reflecting the different tax burdens that both active and passive policies impose on each kind of employer. This is indeed reproduced on the ALMP dimension, where those running small companies are substantially less likely than their counterparts in larger ones to agree on its increase. These caveating results suggest an interesting possibility that will have to be further tested with ad- hoc field research that improves external sources giving somewhat conflicting results to this particular regard. It might be that while the ALMP dimension of labour market policy follows a more linear distinction on its preference ordering within different occupation categories, EPL and NEB are more of a U-shaped pattern, with mid-level and even high-skilled professionals being more wary of reduction than unskilled workers, and all of them being far from the levels displayed by employers. In consequence, the potential coalitions for flexicurity appear quite murky, and mostly rely on a possibly thin grouping of large employers, high-level managers, self-employed professionals and the unemployed. The alternative, displayed by ESS data, indicates a possible agreement on NEB cuts between employers and mid- and low-skilled workers, which would be even more puzzling from a political standpoint. Aside from trying and reproducing the current result with different, ad-hoc sources, a possible hypothesis to guide further research departs from the fact that flexicurity consists of a two-step reform. When reducing EPL for permanent workers, those who do not hold a standard contract still stand to lose something, and that loss depends on the probability they attach to obtaining one in the future. Such loss has to be compensated for with an increase in employability and expected income throughout along their lifespan. For that to happen, the second step (expenditure) is a necessary condition. Therefore, for outsiders to agree to flexicurity they have to trust that both steps will be fulfilled. This becomes truer as outsiders' skills are worse for competing in the market: paradoxically, those who would benefit more from eradicating dualisation are also those who are more cautious about eliminating it since they need the additional spending, unlike outsiders with highly-skilled profiles. Then, ‘pure’ options that are less politically complex are preferred by those who hold a strong preference for deregulation. This is a dynamic argument, one that would necessarily require a series of country cases with panel references in order to be tested. For the time being, what I can assert with the data in hand is that flexicurity policies might indeed face a problem when trying to raise coherent, enthusiast and broad coalitions of voters.

Workers’ preferences on flexicurity lxxi

References Bentolila, Samuel and Giuseppe Bertola (1990). 'Firing costs and labour demand. How bad is Eurosclerooerisis?', Review of Economic Studies, 57:3, 381-402. Bentolila, Samuel and Juan J. Dolado (1994). 'Labour flexibility and wages: Lessons from Spain', Economic Policy, 9:18, 53-99. Berger, Samuel and Michael Piore (1980). Dualism and discontinuity in industrial societies. Cambridge: Cambridge University Press. Blanchard, Olivier J. and Lawrence H. Summers (1986). 'Hysteresis and the European unemployment problem', NBER Macroeconomics Annual 1986, V.1, 15-90. Boeri, Tito, Jose I. Conde-Ruiz, and Vincenzo Galasso (2012). 'The political economy of flexicurity', Journal of the European Economic Association, 10:4, 684-715. Burgoon, Brian and Fabian Dekker (2010). 'Flexible employment, economic insecurity and social policy preferences in Europe', Journal of European Social Policy 20:2, 126-141. Busemeyer, Marius R., Philipp Lergetporer, and Ludger Woessmann (2018). "Public opinion and the political economy of educational reforms: A survey." European Journal of Political Economy 53 161-185. Cusack, Thomas, Torben Iversen and Philipp Rehm (2006). 'Risks at work: The demand and supply sides of government redistribution', Oxford Review of Economic Policy, 22:3, 365-389. Doeringer, Peter B. and Michael J. Piore (1971). Internal Labor Markets and Manpower Analysis. Lexington: Heath. Dolado, Juan J., Carlos Garcia-Serrano and Juan F. Jimeno, JF (2002). 'Drawing lessons from the boom of temporary jobs in Spain', The Economic Journal, 112:480, F270-F295. Elmeskov, Jörgen, John P. Martin, and Stefano Scarpetta (1998). 'Key lessons for labour market reforms: Evidence from OECD countries' experience'm Swedish Economic Policy Review, 5:2. Emmenegger, Patrick (2009). 'Barriers to entry: Insider/outsider politics and the political determinants of job security regulations', Journal of European Social Policy, 19:2, 131-146. Emmenegger, Patrick, Silja Hausermann, Bruno Palier, and Martin Seeleib-Kaiser (2012). 'How we grow unequal', in Emmenegger, Patrick, Silja Hausermann, Bruno Palier, and Martin Seeleib- Kaiser (eds). The age of dualization: The changing face of inequality in deindustrializing societies (3-27). Oxford: Oxford University Press. European Commission (1995). Eurobarometer 43. Available at http://ec.europa.eu/commfrontoffice/publicopinion/archives/eb/eb43/eb43_en.htm (accessed 5 July 2016). European Commission (2006). Eurobarometer 65. Available at http://ec.europa.eu/commfrontoffice/publicopinion/archives/eb/eb65/eb65_en.htm (accessed 5 March 2017). European Commission (2009). Eurobarometer 71. Available at http://ec.europa.eu/commfrontoffice/publicopinion/archives/eb/eb71/eb71_en.htm (accessed 5 March 2017). European Commission (2011). Eurobarometer 76. Available at http://ec.europa.eu/commfrontoffice/publicopinion/archives/eb/eb76/eb76_en.htm (accessed 5 March 2017). Garritzmann, Julian L., Marius R. Busemeyer and Erik Neimanns ‘Public demand for social investment: new supporting coalitions for welfare state reform in Western Europe?’, Journal of European Public Policy, 25:6, 844-861 Güell-Rotllan, Maia and Barbara Petrongolo (2007). 'Workers’ Transitions from Temporary to Permanent Employment: the Spanish Case', LSE Centre for Economic Performance Discussion Papers, No. 438. 72 The reconfiguration of working class politics

Hall, Peter A. and David Soskice (eds.) (2001). Varieties of capitalism: The institutional foundations of comparative advantage. Oxford: Oxford University Press. Häusermann, Silja (2010). The politics of welfare state reform in continental Europe: Modernization in hard times. Cambridge: Cambridge University Press. Häusermann, Silja and Hanna Schwander (2011). 'Explaining welfare state preferences in dualized societies'. Les Cahiers Européens de Sciences Po, No. 01. Häusermann, Silja and Hanna Schwander (2011). 'Varieties of dualization? Labor market segmentation and insider-outsider divides across regimes', in Emmenegger, Patrick, Silja Hausermann, Bruno Palier, and Martin Seeleib-Kaiser (eds). The age of dualization: The changing face of inequality in deindustrializing societies (27-52). Oxford: Oxford University Press. Iversen, Torben and David Soskice (2001). 'An asset theory of social policy preferences', American Political Science Review, 95:4, 875-893. Jimeno, Juan F. and Luis Toharia (1993). 'The effects of fixed-term employment on wages: theory and evidence from Spain', Investigaciones económicas 17:3, 475-494. King, D and Rueda, D (2008). 'Cheap labor: The new politics of ‘bread and roses’ in industrial democracies', Perspectives on Politics, 6:2, 279-297. Koopmans, Ruud, and Hanspeter Kriesi (1997). Citizenship, National Identity and the Mobilisation of the Extreme Right: A Comparison of France, Germany, the Netherlands and Switzerland. Berlin: Wissenschaftszentrum für Sozialforschung. Lindbeck, Assar and Dennis J. Snower (1986). 'Wage setting, unemployment, and insider-outsider relations', American Economic Review, 76:2, 235-239. Lindbeck, Assar and Dennis J. Snower (1987). 'Efficiency wages versus insiders and outsiders', European Economic Review, 31:1, 407-416. Lindbeck, Assar and Dennis J. Snower (1988). 'Cooperation, harassment, and involuntary unemployment: an insider-outsider approach', American Economic Review, 1988: 167-188. Lindbeck, Assar and Dennis J. Snower (2001). 'Insiders versus outsiders', Journal of Economic Perspectives, 15:1, 165-88. Lindvall, Johaness and David Rueda (2014). 'The Insider-Outsider Dilemma', British Journal of Political Science, 44:2, 460-475. Marx, Paul and George Picot (2013), 'The party preferences of atypical workers in Germany', Journal of European Social Policy, 23:2, 164-178. Maselli, Ilaria (2010). ‘Beyond Flexibility and Security: A composite indicator of flexicurity’ CEPS Working Document No. 329/May 2010 Maurin, Eric and Fabien Postel-Vinay (2005). 'The European job security gap', Work and Occupations, 32:2, 229-252. Mughan, Anthony (2007). 'Economic insecurity and welfare preferences: A micro-level analysis', Comparative Politics, 39:3, 293-310. Mughan, Anthony and Dean Lacy (2002). 'Economic performance, job insecurity and electoral choice', British Journal of Political Science, 32:3, 513-533. Mughan, Anthony, Clive Bean, and Ian McAllister (2003). 'Economic globalization, job insecurity and the populist reaction', Electoral Studies, 22:4, 617-633. Nickell, Stephen, and Sushil Wadhwani (1990). 'Insider forces and wage determination', The Economic Journal, 100:401, 496-509. Piore, Michael J. (1980). 'Economic fluctuation, job security, and labor-market duality in Italy, France, and the United States', Politics and Society 9(4, 379-407. Przeworski, Adam (1985). Capitalism and social democracy. Cambridge: Cambridge University Press. Workers’ preferences on flexicurity lxxiii

Przeworski, Adam, and John D. Sprague (1986). Paper stones: A history of electoral socialism. Chicago: University of Chicago Press. Rehm, Philipp (2009). 'Risks and redistribution: An individual-level analysis', Comparative Political Studies, 42:7, 855-881. Rehm, Philipp (2011). 'Social policy by popular demand', World Politics, 63:2, 271-299. Rueda, David (2005). 'Insider-outsider politics in industrialized democracies: The challenge to social democratic parties', American Political Science Review, 99:1, 61-74. Rueda, David (2007). Social democracy inside out: Partisanship and labour market policy in industrialized democracies. Oxford: Oxford University Press. Saint-Paul, Gilles (1993). 'On the political economy of labor market flexibility', NBER Macroeconomics Annual 1993, Vol 8, 151-196. Saint-Paul, Gilles (1996). 'Exploring the political economy of labour market institutions', Economic Policy, 11:23, 265-315. Saint-Paul, Gilles (2002). 'The political economy of employment protection', Journal of Political Economy, 110:3, 672-704. Tangian, Andranik S. (2004). Defining the flexicurity index in application to European countries. WSI- Diskussionspapier No. 122 Thelen, Kathleen (2014). Varieties of liberalisation and the new politics of social solidarity. Cambridge University Press.

74 The reconfiguration of working class politics

A (social democratic) escape from dualisation? lxxv

Chapter 5. A (social democratic) escape from dualisation?

The present chapter aims to explore why some democratic countries pursue liberalising labour market reforms while others do not. This admittedly broad question lies at the heart of comparative political economy. To summarise decades of research in a single sentence: Actors form and consolidate political preferences based on four factors: available coalitions, incentives generated by past measures, available ideas, and present structural conditions which determine both the opportunities for growth and distributing it. If path dependency and structural determinants are left as constants, then it could be argued that the government’s ideological composition becomes the key to understanding variation in labour market reforms. While it is generally expected for right-leaning parties to produce liberalising reforms, it can be more difficult to predict the results when reforms are in hands of left-leaning parties with a core constituency formed by workers. A good instance is provided by the cases of Italy and Spain during the Great Recession (2008-2014). Both countries have deeply-dualised labour market structures, with different levels of effective protection against the risk of unemployment given to different workers depending on their contractual positions, dividing the working population between insiders (i.e. protected) and outsiders (i.e. unprotected). Both countries have had the same policy solutions repeatedly floated to solve dualisation through the reduction of the breach of protection between contractual types. During the Recession, both countries faced similar structural pressures from abroad to liberalise their labour laws. In both cases there was a centre-left party in charge (until 2011 in Spain and from 2013 in Italy). Yet, despite all these coincidences, Italy pursued a much more ambitious form of de-dualising liberalization. One of the aims of this work is to examine why it elected to do so. I argue that, while in both the Italian Partito Democratico (PD) and the Spanish Socialist Party (PSOE) there is a reformist wing and a pro-status quo wing (the former mostly headed by economists, the latter populated by 'union people'), the reformist front only manages to win the internal battle in the PD. Both sides have been fighting this same battle for decades in both countries, and every single party leader has come to power with a compromise between both wings. Except for Italian PM Matteo Renzi, who in 2013, came to power with a clear reformist, pro-change discourse that portrayed unions and its representatives as obstacles who only thought of their own interests. Basing his arguments on a political, normative version of the insider/outsider framework, his victory within the party was achieved thanks to a 'confrontationalist' strategy supported by reformists, who enjoy a stronger presence within the PD since the party was conceived from the beginning as an alliance between centrists and socialists. The PSOE is instead a classical socialist party. When Renzi coupled his internal victory in the 2013 primary election with the party's favourable result in the 2014 European election, he argued that he had received a mandate for change. This (together with problem load, external pressure) was the reformists' window of opportunity to push for a full-scale reform. In Spain, reformists used the same argument set but the lack of a party leader who was able to use a mandate for change to push for reform hampered their attempts. The pro-status quo wing put Spanish Socialist PM Jose Luis Rodriguez Zapatero under an ultimatum at a crucial decision-making juncture, with his Minister of Labour threatening to resign if a full-scale EPL reform were to be enacted. And Zapatero chose to maintain the coalition. Four years later, several pro-reform economists would find a more suitable space for their demands in Cs. The comparison sheds light on why some countries with rigid labour markets pursue liberalisation while others do not. Taking a two-case study approach arguably restricts the generalisability of my hypothesis, but in exchange it greatly helps to expose the centrality of elite coalitions and how they interact with historical factors, structural determinants, and ideas. The remainder of the chapter is organised as follows: Section 1 reviews the literature and proposes an analytical framework and then section 2 develops the main arguments. Section 3 compares the emergence of dual labour markets in both countries and tackles the debate that sparked. Section 4 goes on to analyse how each country confronted demands to de-dualise the labour market during the Great Recession, and section 5 concludes by highlighting again the varying mechanism while the sixth section closes by suggesting some possible avenues for further research.

76 The reconfiguration of working class politics

5.1 When elite battles take it from the base Discussing the influence of within-party coalitions on the decision to liberalise labour markets has been quite fruitful for comparative political economy. Classic power resources theory posed that the degree of protection offered by the state to workers is a more or less direct function of the power enjoyed by the working class (Korpi 1985). It thus follows that such power, understood as turnout and mobilisation, significantly decreases the probability of liberalisation due to workers’ preference for protective arrangements. Following this argument and taking into consideration more recent arguments from Saint-Paul (2002) or Gallasso (2010), it can be seen that labour market reforms are more likely under right-leaning governments. However, the standard power resource perspective is insufficient to explain why some centre-left governments apply pro-liberalization reforms. Consider what Gilles Saint-Paul called the ‘constituency effect’ (Saint-Paul 2000): institutions create rents that might be threatened with any eventual reform. Those whose rents are threatened tend to introduce a ‘status quo bias’ in the decision-making process, that will be greater depending on their number or their capacity to self-organise and mobilise, i.e. the constituency’s capacity to exert influence on decision makers. This could be seen as an adaptation of the power resources hypothesis through a weak form of path dependency, one that becomes particularly interesting with the following question: What happens if the left party has a pro-status quo and a pro-reform wing? This resulting divide is somewhat foreseeable: Rueda (2005, 2007) poses that social democratic parties face a dilemma regarding labour market policy. Either they please those who are protected through existing regulations (i.e. the insiders), who prefer strong employment protection legislation (EPL), or they give space to outsiders’ preferences (i.e. those not protected by existing regulations, mainly workers under non-permanent contract and unemployed people seeking jobs), which are more oriented towards active and passive labour market policies (unemployment benefits, training) and give less importance to EPL. At this point, the notion of self-interested constituencies blocking reforms is quite evident. What becomes interesting and somehow novel is how political entrepreneurs with particular agendas might take advantage of the argument to advance their own policy agendas, picking up ideas and putting them in motion (Hogan and Feeney 2012). Economic crises might very well act as catalysers for reform. However, the (supposedly) ’functional’ demands will have to face a given structure of political incentives. The executors, i.e. politicians and policymakers, will have to manage such structure, therefore creating its own dilemma. Although one could consider the government/party elite as a unified actor weighing pros and cons of each strategy in a purely rational way, it might be more realistic to see it as a divided entity that faces its own internal conflict. A logical consequence of these lines of argumentation would be that as long as social democratic parties come from an inclusive tradition (and therefore have both insiders and outsiders, high-skilled and low-skilled workers within its coalition), it would make sense to have a mixed elite within itself, one that includes representatives of both constituencies. Some of its leading members might be closer to unionised, low-skilled or insider workers. Others, in turn, may come from a more socio-liberal background, having a closer link with high-skilled newcomers to the labour market. Under a presumably divided party elite, it is reasonable to expect that once the pressure for reform arises, the latent conflict within the party elite would become activated. For the reformist side, the insider/outsider framework might prove to be a useful resource: properly framed, it provides a narrative with a culprit (insiders) and a victim (outsiders), as well as a clear-cut solution (reform). The production of this narrative will however have to meet with a proper window of opportunity to make it victorious. Problem load itself is only a part of it. The other part is the capacity of the reformist side to impose itself in the internal battle, which might be facilitated by a range of historical and contextual factors. The 2008 Great Recession, and particularly its aftermath in Europe (the European Monetary Union — EMU — crisis), offers an exceptionally good field of study to sharpen and test the argument, since it constituted a common shock that affected a range of countries, some of which party is in charge and intends to support itself though both high-skilled voters and outsiders. Such is the case for Spain and Italy: two intensely dualised economies (Molina and Rhodes 2007) facing intense debt troubles, fiscally limited, and with widespread external pressure to reform what was seen as sclerotic labour A (social democratic) escape from dualisation? lxxvii markets by a number of international and national actors (e.g. OECD 2012; OECD 2015), who usually recommended a liberal-leaning reduction of the breach of protection between permanent and non- permanent workers to fix them. In other words, the ‘problem-load’ argumentations that point to the importance of objectively measurable factors as reform-detonating (e.g. Hancke and Rhodes 2005) as well as those that put macroeconomic elements along with pressure from foreign actors might help explaining why both countries begin looking for reform options, but do not seem capable to explain variation: both Italy and Spain experienced the most difficult part of the crisis with a considerable problem load and external pressure, as well as centre-left governments, but each took different paths from the beginning (Picot and Tassinari 2014). Taking down my hypothesis to the cases of interest, I pose that the different paths of labour market reform followed by centre-left governments in Italy and Spain during the EMU crisis are linked to the structure of opportunities faced by reformist factions. While pro-status quo fronts were strong enough within both parties, reformists managed to get control of the PD through a primary election. Then, the mandate they claimed was reinforced thanks to beyond-expectations result in a European election. The resulting window of opportunity gave more space to pro-reform ideas within the PD, while they could only advance to a certain point within the PSOE. In turn, Italy responded to structural pressure by applying a substantial liberalization to its labour market regulation, while Spain fell short. In other words: while the PD had its interests aligned with external pressure, the PSOE had to balance out costs at home and abroad. Both countries had to confront these demands with pro-status quo groups, but these groups could pose a challenge within the party elite only in Spain. From a methodological standpoint, an Italy-Spain comparison provides a fruitful ‘most similar cases’ approach (Anckar 2008), one in which other key factors (past structural conditions, problem load and corresponding external pressure, the availability of ideas as solutions, the presence of a centre-left party in power at the time of reform) can and will be considered one by one, since they are kept constant, while one (the space, position and strength for reformist political entrepreneurs) does change and thus provide variability. While (as with any case study) it is hard to argue for a complete generalization of any results obtained, hopefully they will provide insights that help to shed some light over the central question. The case study in this chapter utilized two different sources. First, there are secondary sources (press records) dating from the time when the reforms were enacted. These help to correctly sequence the events and document the exchange of actors’ positions. Second, I have conducted thorough field research in Rome and Madrid, focused on interviewing relevant policymakers to confirm the sequence of events and to pin down the internal dynamics of the reform process. Considering that a chronological exposition of facts that keeps a semi-parallel analysis of the two cases at hand serves best the purpose of drawing a clear layout of the relevant facts, I begin by exploring how the origins of dualisation created similarly divided constituencies and party elites in both countries.

5.2 Dualisation and its consequences over social democratic parties: the emergence of an internal battle

5.2.1 The origins of dualisation Workers, employers and the state (and the sub-groups forming them since actors are not homogeneous) have something in common: they hold a double preference for growth and for increased power leading to a larger share in income redistribution. Coalitions happen to maximise both dimensions, although non-coordinated equilibria may and do arise as well. In the Spanish and Italian cases, alternatives for grouping around certain policy preferences regarding labour regulation have been historically conditioned by two strong restrictions: 1) an inward-looking economy with stable and predictable demand following a development model based on state intervention and captured internal markets; 2) a scarcity of taxable resources for strong demand-driven policies, the main restrictions being low productivity levels as well as low employment rates traditionally focused on ‘male breadwinners’. Under such conditions, the original path towards dualisation was created in both Italy and Spain: A coalition between core workers and sheltered employers is feasible and would use regulatory measures featuring a one-off tax on firings. The alternative would be less profitable since it would consist of extended social contributions to finance unemployment insurance and a training system of

78 The reconfiguration of working class politics

little use and excessive cost. Not offering any protection at all was not an option either as labour had organising capacity to ask for protection. These factors have led to an over-use of contract, hiring and dismissal regulation as the default tool to protect against the risks of income loss for workers in both countries. At the same time, expenditure and active measures have been under-used, as well as parcelled through contribution-based entitlements. In Spain, dualisation was consolidated in the 1980s. The 1982 Workers' Charter established a system of judicially reviewed dismissal, with severance payment fixed at 45 days per year worked (capped at 48 months) in case of unjust dismissal and 20 days (capped at 12 months) in case of firing due to objective/economic causes. In the Charter, the possibility of fixed-term and other forms of temporary contracts were contemplated. As in all the previous legislation since the 1920s, it was explicitly established that the by-default modality was open-ended, and that there had to be a clear indicator of impermanence in the nature of the job position for allowing the use of temporary modalities. However, two backdoors were left open: the possibility of continuation of temporary contracts was left open in the Workers’ Charter, to be defined by future law or decree; and the use of temporary contracts to foster employment creation was allowed ‘under special circumstances to be determined by the Government in further legislation’. That is precisely what happened between 1981 and 1983. A series of three law-decrees which allowed the use of temporary contracts regardless of the nature of the job (‘to foster employment’) and the establishment of training and internship contracts followed. Limits in the maximum use of temporary contracts were established (a percentage of total employees that decreased as the size of the firm became larger). By the end of 1984, further changes were introduced through a new Royal Decree, destined to develop what had been established in the Charter’s reform. In Italy, the Workers' Charter passed in May 1970 with the acquiescence of the then-hegemonic Christian Democratic party (Democrazia Cristiana, DC). The Charter, which was quite protective with workers’ rights in most of its provisions, featured a rule that would later become particularly controversial. Article 18 (Articolo 18) stated that, if a dismissal was deemed as invalid by a judge, the worker had the right to demand his reinstatement with compensation, or an indemnity replacing the reinstatement. Article 18 has been since seen by both its defendants and its detractors as the keystone of labour protection in Italy. Meanwhile, the use of non-standard contracts was only fully allowed nearly three decades later. Under a left-leaning government, the 1998 ‘Treu package’ (Pachetto Treu, named after Tiziano Treu) was agreed upon by social agents after intense negotiation. The reform greatly enlarged the possibilities to use temporary contracts in all sectors, increasing its maximum duration through prorogation. It also potentiated new atypical contract forms, such as the ‘co.co.co' (collaborazione coordinata e continuativa, continuous and coordinated collaboration). This contractual form had existed since the 1970s, but thanks to Treu’s reform its use became widespread. At the turn of the 21st century, a new centre-right government put Roberto Maroni, from the far-right Lega Nord, in the Ministry of Labour. His advisor Marco Biagi, professor of Labour Law, began to draft a reform proposal aimed at intensifying the use of flexible contracts. Biagi was then assassinated by the Political-Military Communist Party (PCPM), but the reform passed nonetheless. It introduced contractual forms that allowed for job sharing and on-call jobs, and it complemented the co.co.co with a new format called 'co.co.pro', a project-based contract that gave substantial leverage to companies looking to hire workers for short periods. As assessed by the European Commission (2017), the combination of the Treu and Biagi reforms clearly relaxed the use of standard temporary contracts and simultaneously introduced new forms of non-permanent arrangements, all the while it maintained the level of protection in permanent contracts. The 1997-2003 reform cycle consolidated dualisation into Italian work regulation, creating a much more complex picture than previous regular-irregular, North-South divisions.

5.2.2 The battle over dualisation within party structures Once the concept of dualisation became common among evaluators of economic and labour market performance in both Italy and Spain, ideas to counteract it began to flourish. In the early 1990s, labour market segmentation was pinned down as a major problem in Spain both by the specialised literature (e.g. Bentolila and Saint-Paul 1992, Bentolila et al. 1994) and, more importantly so, by the involved actors. At that time, the PSOE had been in power for a decade as the first left-leaning government in half a century after Franco’s authoritarian regime. Top PSOE government officials, including the A (social democratic) escape from dualisation? lxxix

Minister of Labour himself, pointed to dualisation as a first item on their list of priorities (Cebrián and Parra 1993). Leftist PSOE militants and rank-and-file members agreed (Díez 1994). Unions were by no means far from this view, denouncing an increase in precarious job positions (“UGT detecta aumento de la precariedad [UGT Detects an Increase on Preciariousness]”, 1992, August 16). The situation was so dire that even employers accepted that job insecurity was a serious issue to address. While an agreement was reached that this situation needed to be resolved, those involved remained deeply divided and could not reach an agreement on the identification of its origins, let alone its possible solutions. Broadly speaking, two blocks could be identified among those dominating the debate within the Spanish left-leaning political elite. The majority bloc tended to link the increase of outsiders with the so-called ‘decausalisation’ of temporary and short-term contracts, implicitly arguing that the backdoor for precariousness is the low protection level for temporary workers. There was a tendency to relate the over-use of those contract forms to the Spanish productive structure and a ‘lack of investment’ on more quality-oriented sectors. Conversely, a contrarian group contended that it was the protection gap between standard and non-standard contracts that provoked segmentation, emphasising the high dismissal costs for the former. Those coming from Economics departments, international organizations (IMF, World Bank, OECD), related to the Ministry of Finance, the Spanish Central Bank or Moncloa’s Economic Cabinet for the Prime Minister (a unit of economic advisors that directly responds to the Office of the Prime Minister), as well as more centrist, liberal-leaning politicians, tended to agree with the minority position. Individuals with closer links to Labour Law departments, the Ministry of Labour and less liberal politicians leaned towards the majority view. Regarding social partners, unions tended to side with the majority whereas employers either did not take any position (simply complaining about excessive rigidity) or timidly sided with the latter. The structure of this debate reproduced itself in all reform attempts under PSOE rule, beginning with the party’s last efforts right before Felipe González lost for the first time in fourteen years, in 1996. In subsequent years, there were a couple of marginal changes that could have been regarded as victories by de-dualising reformists. Under the rule of the liberal-conservative Popular Party (Partido Popular, PP), a new contract form with reduced severance payment (33 days per year worked) was born, to be applied to specific groups who faced particular difficulties to get permanent contracts. Also, the main forms of temporary contracts had their own severance payment (eight days per year worked) attached, to be applied regardless of whether the worker was fired or the whole period stipulated was fully completed. However, neither of these changes brought a significant change to the share of the Spanish working population on temporary contracts, stalling at one third in a decade of unparalleled growth between 1998 and 2007, which was fuelled by foreign credit and construction-focused internal consumption. In consequence, reformists were never fully satisfied with these marginal reforms, and instead began to discuss the possibility of full-scale reform that would substitute all contract formats with a single permanent contract. The idea, taken from French economists (Blanchard and Tirole 2003), would be fully supported by reformists in Spain once the Great Recession set in. The situation in Italy was similar. From the early 2000s on, economists identified dualisation as a potential source of precarious labour (e.g. Barbieri and Scherer 2005). However, at the political level, proposals on how to tackle dualisation differed. On the most liberal side, Article 18 attracted critics’ views from quite early on. As Ichino (2015) recalls, PSI senator Gino Giuni proposed reform in the Senate floor back in 1985. That same year, the Labour Law specialist Luigi Mengoni advocated for limiting its scope in a paper sanctioned by the National Council for Economics and Labour (CNEL by its Italian name). During the 1990s, members of the centrist Radical Party pushed unsuccessfully for a referendum to eliminate both Article 18 and temporary contracts (Ichino 2015: 40). Left-leaning politicians such as Pietro Ichino, or Ignazio Marino represented these positions during the 2000s, to the point of taking up the proposals for a single contract that came from reformist economists, most notably Tito Boeri and Pietro Garibaldi (Boeri and Garibaldi 2008). This single contract format would have eliminated temporary arrangements in exchange for flexibilizing permanent contracts. As it will be demonstrated below, certain actors in Spain took on and adapted this proposal during the Great Recession, which stresses the coincidences in proposals and strategies between reformers across borders. Meanwhile, unions and further leftist politicians, along with certain economists and labour law experts, shared the worry over dualisation but greatly differed on reform proposal. In an interesting turn, the radical left party Communist Refoundation (Rifondazione Communista, RC) successfully pushed for a referendum to “defend" Article 18. The Italian Labour General Confederation (Confederazione generale italiana dei lavoratori, CGIL, Italy’s main union) and the left wing of then-dominant social

80 The reconfiguration of working class politics

democratic platform Left Democrats (Democratici di sinistra, DS) called to support it, while the more centrist Italian Confederation of Workers' Unions (Confederazione italiana sindicati lavoratori, CSIL) and most DS moderates proposed abstaining from it. Among members of the left, the abolition of Article 18 was only favoured by a minority of DS members. The CGIL in particular called for numerous protests against what they saw as growing precarisation, defending either a reversion to pre-reform regulation or the enforcement of a more restrictive use of non-standard contracts Still, neither the advocates of increased protection for all, nor those who pushed for a flexibilisation of permanent contracts to replace temporary arrangements managed to advance their proposals in the political field. Instead, dualisation remained as a lock-in equilibrium until the Great Recession acted as an external shock that stimulated reform discussions. But to fully understand how the battles between both sides evolved differentially within each party, it should be taken into account its varying historical and organisational characteristics. While it is true that both the PSOE and the PD featured strong bonds with the union movement from their respective beginnings, and while it is also true that the PSOE had a moderate element within itself that could be traced back to the Second Spanish Republic years (Juliá 2017), the PSOE was founded and grew under the premise of being a classical socialist party that then followed a relatively standard evolution towards social democracy (Mateos 1993). In other words, its ideological journey was from the left and towards the center, but it never fully reached it, and it never stopped having a strong socialist legacy. When, in 1974, Felipe González was chosen as the new leader that would take the party into the reinstauration of Spanish democracy during a clandestine congress, its platform was at the same time built around the idea of renouncing to Marxism but not to socialism (Santesmases 1993). This trait, an attempt to reach the center from the left, was maintained by all subsequent party leaders. The PD, instead, was born as an explicit and relatively equilibrated coalition between the centre and the left (Ventura 2018). Its goal, from the beginning, was to keep that equilibrium, and thus it featured members of the old DS (classical social democrats) and The Daisy (La Margherita, reformist centrists) and some commentators tended to equate the PD with its American homonym, the Democratic Party. As a matter of fact, the two party leaders and prime ministers who could command and execute the reform processes during the crisis corresponded with this difference of party profiles: while José Luis Rodríguez Zapatero came to power in 2003-2004 with the explicit goal of enlarging the social democratic base through the inclusion of younger, urban, progressive voters with a multiculturalist agenda (González and Caínzos 2011), but making sure that they would not travel too far since the PP was at that moment somehow skewed to the right through a polarising strategy (Bernecker and Mainhold 2007), Matteo Renzi rose through a radical reformist, anti-old left discourse (Pasquino and Valbruzzi 2017).

5.3 When structural factors meet the internal divisions Between 2010 and 2015, as debtor countries with strained labour markets and high external and structural pressures, both governments were considerable pressure to change their labour regulations. It is worth highlighting that there appears to be no way to discern any higher pressure (neither through the intensity of the problem itself nor through its interpretation and how it was translated into external pressure) that might help explaining different reform paths. Going through different indicators, Spain marked higher in unemployment rate during the whole 2009-2015 period, and while its 2009 yearly GDP loss was slightly more contained (-3.9pp vs 5.5pp for Italy), the Italian recovery in 2010 would be quicker and steeper. Measures of labour marked dysfunctionality such as the share of non-standard contracts, the rate of firings, or inequality of income before taxes were all higher in Spain than in Italy between 2008 and 2015. Finding now a proxy for creditor’s (and in general external) pressure, the spread between interest rates for the Spanish and the German 10-year bond was at 170 basic points in May 2010, a few days before Spanish PM Zapatero launched his reforming efforts. In August 2014, while Renzi was plotting his, the Italy-Germany spread was at 150 points, and had been going down since the beginning of the year, while in Spain the trend was upwards during the first months of 2010. On the more qualitative/perception dimension, it will be shown in the following sub-sections that calls for reform were intense and constant regarding both countries. In sum, while the context and the constraints were shared, the response was nonetheless different. The variation on their response seems to correlate with the different reconfigurations undergone by centre-left parties of both countries. A (social democratic) escape from dualisation? lxxxi

5.3.1 The crisis in Spain and Zapatero’s labour market reform Most accounts of the Great Recession establish the fall of Lehman Brothers in September 2008 as the starting point of the turmoil. However, job losses in Spain had begun shorty before this date. The deceleration of the housing boom was already noticeable by the beginning of the year, as early as Autumn 2007 according to some market analysts. Temporary jobs were disproportionately affected by the recession: according to data from the National Statistics Institute (Instituto Nacional de Estadística, INE) between 2008 and 2011, more than a third of those were wiped out, but less than 10% of permanent positions disappeared. This effect was particularly harsh among the youth: More than 60% of individuals below 25 years old with non-standard contracts lost their jobs in that period. The PSOE initially reacted to the crisis with a stimulus-based set of measures to fight the slump. The so-called ‘Plan E’ brought €8bn to be spent by local authorities in more than thirty thousand different projects, with the majority oriented towards improving infrastructure or civil buildings. This initiative employed 410,000 people during 2009, but most of the positions were low-skilled, and attached to short-term contracts that would end sooner than later. On the passive side, a new, non-contributory program gave 426€ per month over six months to every unemployed person who had lost access to contribution-based insurance after the 1st of August 2009. The scheme covered up to 825,000 individuals until February 2011, when it was discontinued and substituted by a slightly more austere support allowance that provided a reduced benefit of 400€ and lasted for six months. Plan E was discontinued around the same date as the allowances. The two only attempts to offer expenditure-based coverage against the risk of unemployment during the Great Recession fell after less than three years due to a mutation in the essential nature of the crisis. By the end of 2009 Europe was entering a new phase of the crisis, one that produced a spike in sovereign debt and uncertainty about its repayment. Counter-cyclical Keynesian stimulus became therefore suspicious for creditors. This new situation displaced the centre of gravity towards the so-called competitive constraint as creditors began to increase their pressure. All the current account deficit and export losses accumulated during the credit boom years (1998-2006) moved to the forefront of the debate. The strategy of growing by leveraging, with a particular focus on sheltered sectors that were labour- intensive rather than productivity-oriented, was deemed as a failure by experts, by international organizations and by the PSOE government itself, who advocated for a "change in the productive model" (Garea 2009). A simultaneous second constraint that came with the sovereign debt crisis was a severe limit on budget capabilities. Since debt both public and private and the uncertainty to repay it was now a central matter for international creditors, expenditure was considerably limited. The situation somewhat resembled that of the early 1990s Spanish crisis: a decade of debt-driven growth with a focus on construction came to a sudden end. Only now Spain found itself without the capacity to use monetary policy since it had been transferred to the ECB. Moreover, to get any sort of fiscal stimuli it would have to undergo a negotiation with its fellow European countries, while each member state had incentives to listen only to their national electorate’s demands. Additionally, those voters who held the capacity to accept fiscal transfers towards indebted countries were not only the actual creditors, but also people whose systemic risk was deeply linked to that of debtor countries, creating even more incentives to force the latter to apply fiscal consolidation and to undergo reforms. Since the beginning of the crisis, organised employers saw the window of opportunity that this situation presented to align their interests with external pressures for increased flexibility. In consequence, the CEOE (Spain’s main employers’ association) insisted on reforming the labour market, with a gradual increase in intensity (Díaz Ferrán 2008; Díez 2009; Gómez 2009; Gómez 2009b); unions acknowledged the offensive (Noceda 2009). By the beginning of 2010 the PSOE government began by explicitly making dualisation a top issue (“El Gobierno centra la reforma en los jóvenes [The Government Focuses on Young People]”, 2010, 6 February). Meanwhile, the cost constraint became clearer than ever by May 2010. At the beginning of that month the government (more specifically, Prime Minister Rodríguez Zapatero) staged what was understood as a U-turn on its economic policy at the Parliament and the media. The Executive shifted from an expenditure-based, demand-oriented plan to fight the crisis of austerity and supply-sided reforms. When doing so, Zapatero expressed external, cost and debt-related constraints. He made clear that this was not his preferred choice, but that of ‘markets’ and ‘investors’, referring to the high leveraging of the Spanish economy. Zapatero’s speech underlined that Spain’s commitment with the EMU required this extra effort given the difficult situation for the Eurozone after Greece’s first bailout

82 The reconfiguration of working class politics

programme was approved. The specific measures included a 5% cut on public sector salaries, freezing old age pensions, eliminating one-off cash grants for new parents, cutting €6bn from public investment projects, reducing expenditure on subsidised pharmaceuticals and on grants for households with dependent elders, and increasing taxes for high income earners. It is within such context that the debate within the left on the labour market reforms should be understood. Union opposition was almost taken for granted due to the aforementioned measures, since the reform was presented as part of the package. Still, they expressed their concerns on the usual matters: enlarging just dismissal causes, the degree of judicial intervention of dismissal and, to a lesser extent, the possibility of enlarging the 33-day-per-year-worked severance payment (Noceda and Gómez 2010). All of these ideas were floated by the PSOE government as possible routes for reform at one point or another (Abellán 2010). Employers instead recognised the convenience of the situation for their interests, and henceforth expressed the intention to reduce dismissal costs even further (Abellán and González 2010). Behind these discussions laid the same dilemma that both employees and employers had been confronting for the last two decades. The IMF, along with the European and Spanish Central Banks, saw a dysfunctional labour market as part of the explanation for Spain’s crisis, and thus demanded more EPL flexibility, considering the idea of a single contract proposed by several economists (Pozzi 2010). A manifesto put forward by one hundred top Spanish economists in early 2009 (FEDEA 2009) promoted by one of the main Spanish think tanks (the Spanish Foundation for Applied Economic Studies, or FEDEA), served as a landmark of this proposal, which consisted of the elimination of all forms of temporary contracts and the imposition of a single, indefinite duration contract for all new hires. The new, single contract would increase firing costs of current temporary contracts, but would also mean lower costs compared to existing permanent contracts. Such a proposal had its supporters within the Zapatero government. Pedro Solbes, Minister of Finance until April 2009, defended the proposition both in public and in private. According to his memoirs, he went so far as to discuss it with the Prime Minister in January 2010 (Solbes 2013). According to his own account of the discussion with Zapatero, Solbes maintains that it was rejected because his chief estimated that a single contract would have brought “two general strikes”, thus anticipating unions’ rejection of a single contract. Solbes' successor, Elena Salgado, instead gave clear statements against the single contract on numerous occasions (“Salgado rechaza un contrato único [Salgado Rejects a Single Contract]”, 2010, 5 April). However, her deputy in the Ministry, the General Secretary José Manuel Campa, was seen as an advocate for this transition (Carreño 2010). Nor was her the only one; another proponent was José Ignacio Conde-Ruiz, a then member of the Prime Minister’s Economic Office (in Spanish, the Oficina Económica del Presidente: a small group of economic advisors not dependent on the Minister of Finance but directly responding to the executive chief), who held the position of Head of Economic Policy. Conde-Ruiz, an adamant advocate of a single contract (e.g. Conde-Ruiz 2012), left Zapatero’s team two weeks after the final labour market reform was approved (“El Gobierno cesa al director de Política Económica [The Government Fires Its Head of Economic Policy]”, 2010, July 3). Proponents of a de-dualising reform insisted on the need to protect outsiders (Conde-Ruiz 2012), attaching it as a duty to a self-described progressive, leftist party. Thus, they effectively used the political economy insider-outsider framework to advance their arguments. On the other side, the single contract proposal (and, in general, demands for heightened flexibility) faced the decided opposition of unions from early on (El Periódico, 20 June 2009). In fact, negotiations between UGT-CCOO and the PSOE government failed to produce an agreement (Ximénez de Sandoval and Garea 2010), and were so unproductive that one of the PSOE deputies who also was a former Secretary General of CCOO failed to vote for the reform at the Spanish parliament (breaking vote discipline is extremely uncommon in Spain due to its closed list election systems). Within the PSOE, the then-Minister of Labour Celestino Corbacho began by arguing that reducing dismissal costs would be useless unless temporary contracts were completely eliminated (“Reducir el coste del despido no servirá de nada [Reducing Firing Costs Will Have No Effect]”, 2010, January 25). He also accused the Spanish Central Bank of having an obsession with the issue (“Corbacho acusa al Banco de España [Corbacho Points to the Spanish Central Bank]”, 2009, November 2) while arguing that the crisis and the unemployment rate did not have its origins in the labour market (“La crisis no tiene origen laboral [The Origins of the Crisis Are Not Labour-Related]”, 2009, February 2). However, his position evolved towards the need of enlarging the scope of cheaper permanent contracts, although never referring to the possibility of a single contract. Instead, the Minister bought into the unions’ A (social democratic) escape from dualisation? lxxxiii argument of improper use of temporary contracts while defending its existence in line with employers’ demands (“Corbacho culpa a la crisis [Corbacho Blames the Crisis]”, 2010, May 28; “La referencia es el contrato de 33 días [The 33-Day Contract Is the Reference)]”, 2010, 1 June), effectively blocking the way for proponents of a single contract. A key moment of Minister Corbacho’s opposition came when he let PM Zapatero know that he would step down as a Minister of Labour if anything resembling a single contract was proposed and passed. This was confirmed separately by two interviewees who at the time occupied senior advisory positions within the Zapatero government, and who declared their support for (and internal involvement with) single contract proposals. Nonetheless, Corbacho was viewed by some PSOE members as bending too much to reformists' demands. In an interview a PSOE MP close to the union movement and with intense involvement in labour policy issues highlighted the difference between him and his predecessor, Jesús Caldera, who was seen as someone “who carried political weight, who was not easy to contradict and who combined a reformist aura with links within unions (…) That is why Jesús [Caldera] was not ‘killed’ by Zapatero, but by [Pedro] Solbes and [Alfredo Pérez] Rubalcaba [Minister of Interior, normally associated with pragmatist positions]”. From this source’s point of view, Corbacho was brought in so the Ministry of Labour could be managed from the Ministry of Finance. However, the MP admits that the final reform did not violate too many of the boundaries established by unions. Thus, both the public and private accounts of the process leave the impression of an intense battle between the reformist faction and the pro-status quo faction and which had its correlation at the party’s voting base. This would help explain why Corbacho did not provide enough space for a full-scale reform, and why Zapatero feared having to face “two general strikes” as per Pedro Solbes’ account. In Saint-Paul’s words, these voters might have introduced a status quo bias in the government’s decision making process. However, if neither a single contract (or any convergence between permanent and non- permanent contracts) nor a full-scale liberalisation was politically viable, which reform could the government produce to meet external pressure? The outcome was therefore similar to that of 1993- 1994: a liberal-dual reform at its core, despite featuring some de-dualising changes. The most contentious aspect of the new law passed in June 2010 was the facilitation of fair dismissal due to economic causes. The firm could now fire employees when it suffers a (as cited in the law) “negative economic situation, in cases such as the existence of present or expected losses, or the persistent decrease of the income level”. Unlike previous modifications of the norm, this new wording was specific enough to actually reduce the percentage of legal cases over unjust dismissals. The scope for the open-ended contract with a reduced severance payment of 33 days per year worked was increased to: people who had been unemployed for more than three months, unemployed between 31 and 44 who had been fired from an open-ended contract, and the recently unemployed who had been under temporary work contracts in the last two years. Severance payment for most temporary contracts was progressively increased from 8 to 12 days per year worked. Also, the maximum duration of the short-term contrato por obra y servicio (contract for a specific service or task), which was widely used as a shortcut for flexibilisation, was limited to three years (four if agreed through collective bargaining). After that period the contract would automatically become open-ended. However, concessions to reduce dualisation ended here: training contracts could be used with youngsters up to 25 years old instead of 21. From 2011 on, the limit of chaining three temporary contracts for the same employee in the same budget, a limited that was established in 2006, was now removed until 2013 as a way to foster employment creation. Also, the use of training contracts was further loosened, and could now be applied to young people up to 30 years old. These modifications eliminated the de-dualising effort of increasing temporary contracts’ severance payment. The aftermath of the reform demonstrated that the PSOE choosing the middle ground left everyone unhappy: the IMF and the OECD asked for “braver” (Jiménez, 2011) or “additional” (“La OCDE pide a Rajoy reformas adicionales [OECD asks Rajoy for additional reforms]”, 2013, December 18) actions. Unions called for a general strike, and the PSOE ended up losing four million votes between the 2008 and 2011 general elections. It retained, however, much of its appeal among insiders. The situation for Italy’s PD however, was quite different.

5.3.2 The crisis in Italy: Fornero’s and Renzi’s reform The PD is a relative newcomer to Italian politics. It was founded in late 2007 as an attempt to recreate old, catch-all social democratic parties but, as stated above, was somehow modelled upon the

84 The reconfiguration of working class politics

American Democratic Party. Merging the heirs of the communists with pro-social Christian-democrats and socio-liberals with the aim of creating a unified, electorally powerful platform was deemed to provoke intense internal tensions in certain policy areas; labour market policy was one of them. The PD was founded under ’s mandate, who had returned to the Prime Minister’s office in 2006, but whose two-year government failed to produce any significant reforms despite growing internal and external concerns over the country’s productivity performance (Luciditi and Kleinknetch, 2009). In May 2008, Prodi lost the premiership against Silvio Berlusconi, who inaugurated his third strand at the Chigi palace just before Lehman Brothers fell. Three years and a half after his electoral victory, Berlusconi lost his position in what he portrayed as an external plot (Mackenzie 2014), although he was at the time facing serious corruption scandals. Setting aside the timing of these events, what is true is that days before the PM’s resignation, interest rates on Italian bonds were skyrocketing (Bowley 2011). Markets’ pressure prompted European leaders to demand decided action from Berlusconi, whose power rested on a centre-right coalition that was, at that particular moment, trying to get the 2012 budget approved. The key vote took place on November the 7th, and Berlusconi’s favoured proposal fell short by eight votes (Moddy and Mackenzie 2011). This was a notable setback, since his centre-right coalition had managed to obtain 344 seats in the past election. In other words: The European debt crisis broke the platform. Mario Monti, the former European Commissioner (1995-2004) not affiliated with any party, managed to obtain a parliamentary majority with centre-left, centrist and centre-right votes. Monti’s two years in power offered the first window of opportunity for those who aimed to end dualisation through the flexibilisation of permanent contracts and the elimination or reduction of non- standard contracts. His Minister of Labour was Elsa Fornero, while Monti kept the Ministry of Finance for himself. Fornero was quick to talk about the need to reduce “contract fragmentation”, betting on “the good flexibility” in a clear reference to flexicurity models (Fornero 2012). After a brief conversation with social agents, one in which CGIL’s leader emphasised that reforming firing conditions was a boundary that could not be crossed (Mania 2012) and thus the need to respect Article 18 (Camusso 2012), the Ministry of Labor then put forward a proposal to substitute Italy’s 48 types of contract with a single contract which was more or less in line with the original Boeri-Garibaldi proposal. These measures matched with those being pushed forward by her economist colleagues in Spain. To restrict Article 18 without derogating it, the single contract would have a first three-year trial phase with cheap firing conditions (a severance pay equivalent to five days of wage per month). A wage floor of 25'000€ / year was also predicted for new hires (i.e. those below the three-year mark), rising to 30'000€ for co.co.pro hirings whose two-thirds of personal income comes from a single firm (Grisseri 2012). The CGIL was quick to react negatively to the proposal, with its youth section calling it a “scam” (Palombi 2012). CSIL and UIL were less confrontational, but did not accept the proposal right away. February and March were months of conversations between the government and social agents (“Fornero rinvia l’incontro con le parti sociali [Fornero Retakes Contacts with Social Partners]”, 2012, February 29), as well as among parties supporting the Monti government. And while Fornero expected the PD to come round and support the reform, she informed them that the reform could go forward nonetheless (“Il Governo procederá [Government Will Proceed]”, 2012, February 22). However, Fornero’s declarations could easily be seen as a bluff by the PD, since it was already clear that unified party support was needed to get the necessary parliamentary votes: The centre-right Berlusconi coalition had already exploded, retaining only a fraction of the parliamentary power it once had. The party’s then-leader, , quickly responded that the PD would support a “good” reform (“Scontro Fornero-Bersani sulla riforma del lavoro [Fornero-Bersani Standoff on Labour Market Reform]”, 2012, February 23), which was a way of acknowledging the different sensibilities within the party regarding the proposed changes. Some key references such as Walter Veltroni, former PD leader, came out to support Fornero’s reformist efforts calling to end the taboo surrounding Article 18, saying that “we need to change a labour market that keeps dramatically marginalizing the young, the precarious, women and the South” (Maltese 2012) in a clear signal towards a pro-outsider flexibilising reform. Such words provoked surprise among some of his fellow party members. For instance, Antonio Bocuzzi, manual worker turned MP, considered Veltroni’s words to be “unbelievable” (Bocuzzi, 2012). Bersani, then, had to find an equilibrium between these two extremes. After months of conversation, Fornero and Monti presented a new reform proposal. Although it was no longer a single contract, it did affect Article 18 by aiming to restrict its effects to discriminatory firings, leaving instead “economic” firings (i.e. firings due to economic reasons) to be resolved through a severance pay system. The CGIL responded rather negatively (“Lavoro, sindacati divisi [Divided A (social democratic) escape from dualisation? lxxxv

Unions]”, 2012, March 20), which in turn took far-left platforms out of any possible agreement and compounded the PD’s dilemma. One of the clearest voices among those supporting the reform came from MP Francesco Boccia, who highlighted the need to “cut the umbilical cord between the PD and the CGIL” (De Marchis 2012). Senator Pietro Ichino had a more moderate tone, but nonetheless called for a positive vote. Ichino’s figure is particularly significant: during the last decade he has been quite involved in single contract proposals (see Ichino 2015 for a synthesis), facing critical comments from many of his partners. One of them being Stefano Fassina, head of the PD’s economic program during the years of the Monti government. Fassina, who would leave the party in 2015 due to disagreements over labour policy with Matteo Renzi, was then involved in a party fight with the socio-liberal wing of the party (“Il caso Fassina [The Fassina Issue]” 2011, November 24), to the point that Bersani had to intervene to resolve the dispute (“Bersani frena [Bersani Stops It]”, 2011, November 23). Bersani then expressed the PD’s intention to amend the reform proposal before it was approved (Nalbone 2012). Three months later, the reform was finally approved with the PD voting in favour. Parliamentary negotiations toned down the initial proposal: article 18 was rewritten to allow some space for severance pay in case of unjustified economic firing, while always leaving the option of reinstatement open depending on a judge’s decision. On the precarious side of contracts, the results were mixed: while the use of co.co.pro was restricted (asking for more detailed and specific definitions of the projects and introducing the need of a base wage) and there were new measures to combat false self-employment, the use of temporary contracts was eased (de-causalising them and slightly extending time limits before its conversion to a permanent contract). While CGIL remained against the reform and initiated a mobilization campaign, the PD divisions kept re-emerging after the final parliamentary vote. For instance: Fornero was not invited to the party’s annual party in what Stefano Fassina defined as a politically motivated decision (Sarti 2012); and Bersani criticised the flexibility aspects of the reform (“Bersani e Fornero a confronto su riforma [Bersani and Fornero Fight Over the Reform]”, 2012, September 19). Moreover, in September 2012 leftist Puglia governor Nichi Vendola proposed a referendum to revert the law. The CGIL, along with numerous far-left and alternative left platforms, sustained the proposal. Meanwhile, the PD remained internally divided on whether the party should be more critical of the reform in particular and of Fornero in general. More significantly, the debate was taking place in a pre-electoral climate. In November, the PD would hold a primary election with two frontrunners. Pierluigi Bersani represented the traditional, unifying soul of the party, also retaining within itself the most leftist currents. Conversely, Matteo Renzi advertised himself not only as the candidate of change, but also as that of open and flexible regulations. As a matter of fact, he considered Fornero's reform to be “timid and inefficient” (Costamagna 2015), thus willing to go even further, as he would demonstrate during his premiership. For now, Bersani would win the primary election 61%-39%, but the party's internal conflict was anything but closed. Before moving on, it is worth remarking on the resemblance between Fornero's and Zapatero's reforms. Both somehow reduce the degree of protection of permanent workers to the point of angering organised permanent workers (i.e. unions) but not to fully satisfy proponents of full-scale liberalisation. Both send mixed signals on protection for non-standard workers, mixing dualising and de-dualising measures. In summary: both politicians' reforms represent a similar compromise. Neither compromise managed to create the expected stimulus in the labour market in either country. In the two years following the reform the unemployment rate kept rising until it reached a plateau in mid-2014. By then, Matteo Renzi had become . He did not get the position through an election. The one celebrated in 2013 left the PD as the most voted party by a small margin. While Bersani’s platform reached 29.55% (Camera), Berlusconi fell 0.3pp short of a victory. Newcomers Five Star Movement (Movimento 5 Stelle, M5S) stopped at 25.56%. Beppe Grillo’s self- labelled populist party became a parliamentary veto player, one that would support neither a centre- left nor a centre-right government. This in turn provoked the end of Bersani’s position as PD general secretary and forced a grand coalition under the rule of moderate PD member , who lasted in power for only 300 days before an internal coup gave the PM office to Matteo Renzi in February 2014. It had only been three months since he won the party’s second primary election in a year. One of Renzi’s key priorities was labour market reform. According to the account of one of his high- ranked economic advisors within the PD, it was clear to Renzi from the beginning that there was a "political window of opportunity” opening up, and that it was relatively narrow. He identifies this "window" with Renzi's triple victory: on the primary election, on the internal battle for the PM office, and

86 The reconfiguration of working class politics

finally, on the good result obtained by the PD in the 2014 European election, which was quickly framed by Renzi as a popular backing to his reformist attempts (Pasquino and Valbruzzo 2017). Since his initial challenge to the party's highest office, Renzi himself made his intentions fully public, referring to the clear need to promote a single contract à la Boeri-Garibaldi both during and after the primary elections campaign (Portanova 2013). His plan was not dissimilar from Fornero's original proposal, but he carried it on further. The aforementioned source marks the PD's inner debate on the Jobs Act that took place in early November as a crucial turning point. The first battle, according to the advisor, was to get the key topics into the party’s elite debating agenda. Once that was achieved, and even after taking all perspectives into account (and, according to the source, there was a significant degree of plurality), the party takes Renzi's route although it does so without a clear majority. In other words: According to Turco (2014), Renzi managed to impose himself by dividing the party's internal opposition. This impression matches with that given by a gubernatorial advisor who was involved in the reform process. He cites the need to act fast and decidedly on labour, a topic which according to him was deemed a government priority from early on, due to this lack of organization and the “weakness” of the party’s most leftist wing. The party advisor stresses the timing: after the European election victory, the reformist wing moved fast and consolidated its internal efforts, culminating in an internal meeting that took place on December 2014. By then, the advisor points, eleven quarters of GDP recession had passed by, and thus Renzi could put together his recently earned political capital with an "economic emergency" argument to impose the so-called Jobs Act. Renzi's Jobs Act consisted of a series of law decrees passed between mid-2014 and early 2015. The Jobs Act left Article 18 limited to discriminatory reasons, essentially as Fornero and Monti intended. Alternatively, in case of unfair dismissal for other more common causes, compensation was set for one month (30 days) per year worked, with a minimum of 2 months and a maximum of 18 months, all in a new standard contract with a three-year trial period. All in all, the reform was seen as “radical” by several observers favouring flexibilisation (OECD, 2015; Benassi and Durazzi 2015), with Pietro Garibaldi calling it “probably the most important reform in the last twenty years […] could change the labour market in a decisive manner” (Garibaldi 2015). Early proponents of single contract models were involved in the reform (Ichino 2015), with economist Tito Boeri in fact holding an appointed position within the Renzi government as a president of the National Institute of Social Security (INPS by its Italian name). But internal opposition did eventually arise, despite its lack of organization. A group of PD MPs headed it, especially Stefano Fassina, whose role within the party became diminished after Renzi’s victory in December 2013, and , who fought and lost the primary election against the current Prime Minister. Both represented more leftist, protectionist sensibilities within the party. Fassina was absent during the first Jobs Act vote, as were another 39 of his fellow party members (“Jobs act, Camera approva testo [Jobs Act: Parliament Approves]”, 2014, November 25), while Civati was one of two PD MPs voting against it. Both, along with a substantial number of former PD members, would end up leaving the PD to join alt-left projects. The Jobs Act was not the only dividing issue within the party, but it was one of the key factors. Not the least because the CGIL was very vocally against it (CGIL 2015). The conflicts that arose around the Fornero reform were now multiplied (“Partito spaccato ma il premier ha i numeri [Broken Party but PM Has the Numbers]”, 2014, September 29; “Il Jobs act fa litigare il Pd [Jobs Act Provokes PD Infighting”, 2014, December 26). Opposition went as far as to make Filippo Taddei, the PD’s head economic advisor, need an escort (Huffington Post Italia, 14 November 2014), sparking some remembrances of Marco Biagi’s fate. A graphic way to delineate the conflict between Renzi’s PD and the CGIL comes by referring to an allegorical character that was defined by Renzi himself. In several moments before and during the reform process, the Prime Minister mentioned someone called ‘Marta’. He defined her as a 28-year- old who does not have the right to be a mother despite waiting for a child due to her job position: unlike “her friends, public servants, she [Marta] has no guarantee, because during the last years Italy has created first-class and second-class citizens” (Renzi 2014). Marta was supposed to epitomise the figure of the young, female outsider. The response from Susanna Camusso (CGIL’s leader) was clear: She appeared with an “I am Marta” t-shirt in a rally, implying that it was the union who could defend outsiders’ interests, and who was aware of the fact that precariousness was reaching all levels. A high-ranked CGIL official explained in an interview the way he saw Renzi’s argument over Marta. According to him, Renzi thought that unions should only take care of workers who are within a given firm, i.e. taking unions’ representative capacity to the firm level and out from the national realm. That is way, the source says, Renzi does not see with good eyes the societal project of CGIL, and why he A (social democratic) escape from dualisation? lxxxvii must argue that unions do not represent Marta, but he does. The view of this official was that by insisting on unions’ incapacity to represent outsiders, Renzi could restrict their actual area of influence to the firm level. However, the source added, Renzi was unable to give any real answer or solution to Marta, he could only provide resolutions to her employers. In other words: what this union official saw in Renzi was a project that sided with capital and divided the working class. In contrast, Renzi and his allies used the insider/outsider framework as a political/normative argument in a way that looked even clearer in the Italian case. "Marta" was, after all, intended to epitomise the disadvantaged worker. And, unlike in Spain, here it was successful.

5.4 Reconsidering alternative explanations After going through the evolution of reform paths in both Italy and Spain, it should be possible to pin down the varying factor that helps to understand why the former pursued a full-scale de-dualising reform. First, it is useful to go again, but in a more systematic way, through alternative explanations that are commonplace in the literature about policy changes and that have been referred to throughout the analysis, compiling its referred common points. Historical paths. Dualisation is a product of historical factors that have created a very similar labour market structure in both countries, at least for what it refers to its segmented nature and the way it is being identified as a problem by both participating actors and external observers. While there are undoubtedly important differences on how dualisation operates in both countries (beginning with a much heavier regional and black-market component in Italy than in Spain, although both dimensions also feature in most descriptions of the Spanish situation), its essential form is either the same or at the very least widely identified in a similar way, for what the insider/outsider framework is widely applied to both countries by decision-makers. Problem load. Labour market segmentation is not only similarly identified as a problem in both Italy and Spain. The way it is portrayed as a heavy burden on growth and productivity is fairly similar across countries, especially after the 2008 shock. There are no appreciable differences on neither the actual degree of problem load that might help to explain a deeper intervention in Italy (if anything, Spain had higher unemployment rates and GDP contraction), nor in the way analysts portrayed the degree of problems faced by each country. It could be even said that Spain was commonly seen as the next most problematic case within the Eurozone after Greece, Portugal and Ireland, and that Italy was mentioned either at a similar level or as next in line (e.g. Roubini and Mihm 2011). Lastly, measures of income variance also point to a higher difficulty in Italy, with Spain becoming the most unequal country in Western Europe during the crisis (Maqueda and Gómez 2017) mostly due to its inability to improve its labour market structure. External pressure. Once again, Spain displayed, if anything, higher levels of interest rates on public bonds at the peak of the crisis and was at least as prominently featured in arguments by international organisations and creditors as Italy was. It is therefore hard to argue that a differential degree of external pressure moved the Italian government beyond the Spanish government. Moreover, when the PP came to power in 2011 and enacted its own labour market reform under an arguably higher degree of external pressure since the banking bailout agreed with creditors was already in place, its format was not yet comparable to Renzi’s job act. Availability of ideas. As it has been widely referred above, there was a considerable degree of coincidence and even intertwining between Italian and Spanish policy entrepreneurs on proposals to reduce dualisation. If anything, the similarity in this dimension is even higher than in the rest, since initial proponents of, say, single contract reforms were well connected and even co-authored papers on the matter (e.g. Boeri et al 2010). Ideology of the government party/coalition. Here is where variance begins playing an interesting role. In principle, as it has been explained, both the PSOE and the PD occupy a centre-left space and belong to the European social democratic tradition. At the same time, a conflict between a more pro- union, pro-status quo side and a more centrist, liberal and reformist front is present in both parties. These are the coincidences, but the conflict has two different outcomes. Thus, the real varying condition comes with intra-party division, how it gets resolved in each of the two cases, and how a specific window of opportunity facilitates the reformist agenda in the Italian case. In both Italy and Spain there is a historical path that leads to a dual labour market, an economic shock that greatly increases the problem load along with external pressure to solve it, and there are reformist solutions available in the world of ideas. In both countries, the reformist faction within a centre-left

88 The reconfiguration of working class politics

government tries to advance these ideas, but only in Italy they achieve a full-scale de-dualising reform. They do it through a key centrist ally being in power, one who advanced within the party ranks by pushing an explicitly reformist and anti-union agenda, and who obtained an important extra amount of political capital through an unexpectedly good result in a second-order (European) election. In Spain, in contrast, the party/government leader was committed with pro-status quo factions as much as he was with reformists, and the latter had no leverage available to impose its preferences.

5.5 Concluding The present chapter had the goal of exploring a possible explanation for diverging paths to reform the labour market under a centre-left rule. Quantitative analysis can only provide a limited view on the dynamics of coalition reconfiguration within parties and its possible consequences on reform paths. In other words, these methods and the available data can only go so far towards unpack the way in which policies are considered, discussed, and finally discarded or approved. With the goal of fully spelling out the mechanisms advanced by my argumentation, I analysed the evolution of reforms in two Southern European economies: Spain and Italy. Both countries share a common past of labour market segmentation that created an insider/outsider divide with potential consequences for centre-left parties. In both cases, these parties hosted different ideological factions with regard to labour market regulation, among other economic policy issues: one, more liberal-oriented with connections to orthodox economics, and another, more protectionist and linked to trade unions. Between 2010 and 2015, as debtor countries with strained labour markets and high external and structural pressures, both governments were under considerable pressure to change their labour regulations. Context and constraints were shared, as well as the ideological family of the main party within the ruling coalition, but the response was nonetheless different, and I see these different paths as linked to winning and losing sides of internal party battles, which in turn gave or withdrew space for internal ideological strands. Hopefully, the present exercise could illuminate future research options, particularly with regards to the interplay between labour market policy decisions and the dynamics of party competition. It does come to align with a growing body of literature that couples the importance of macroeconomic/problem load considerations with political incentives, strategies and calculations (Baccaro and Lim 2007, Avdagic 2010). In recent years, substantial portions of the electorates of core ruling parties in both Spain and Italy have reconsidered their political orientations. New formations have emerged, and old formations are splitting and/or changing their support base. In Spain, the PSOE (social democrats) struggled and eventually failed to maintain an emerging, broad coalition that was not yet consolidated when the crisis hit. Its 2010 labour market reform consisted of an impossible exercise of reconciling new context with old constituencies based on low-skilled workers. It could be argued that the political relatives of those reforms were one of the key underlying causes for the emergence of two new parties in Spain. Both share a negative perspective regarding labour market segmentation, but have nearly opposing views regarding the solutions, perfectly reflecting their new core constituencies. Under similar structural conditions, the Italian PD made a rather bold move away from dualisation, a decision that contrasts strongly with the one taken by its Spanish counterpart. Comparing the evolution of policymaking in Spain and Italy during those two periods illuminates the interplay between a conditioning context and parties' core constituencies. Renzi seemed to decide to build a platform based upon high-skilled workers. The PD move towards the centre matched well with the emergence of a party of young outsider and low-skilled workers, the anti-establishment M5S, who then proceeded to build a left- authoritarian platform. While there remains much work to be done in order to pin down the causal relationships between party decline and emergence, and labour market policy decisions, the Spain-Italy comparison during the Great Recession, the EMU crisis and beyond, seems to offer a promising departing point.

A (social democratic) escape from dualisation? lxxxix

References Abellán, Lucía (2010, April 26). El Gobierno plantea extender a todos los parados el contrato de 33 días [The Government Suggest Extending the 33-Day Contract to All Unemployed People]. El País. Retrieved from https://elpais.com/diario/2010/04/26/economia/1272232806_850215.html (Accessed 3 June 2017). Abellán, Lucía and Miguel González (2010). El Gobierno responsabiliza a CEOE del fracaso en la negociación laboral [The Government Blames CEOE for the Failure on Labour Negotiations]. El País. Retrieved from https://elpais.com/diario/2010/06/11/economia/1276207202_850215.html (Accessed 3 June 2017). Al via la direzione Pd sul Job Acts. Partito spaccato ma il premier ha i numeri [PD Secretariat Begins on the Jobs Act. Party split but the premier has the numbers]. (2014, September 29). Il Sole 24 Ore. Retrieved from http://www.ilsole24ore.com/art/notizie/2014-09-29/renzi-napolitano-oggi- direzione-pd-partito-spaccato-ma-premier-ha-numeri-121814.shtml?uuid=ABnirDyB (Accessed 12 May 2016). Avdagic, Sabrina (2010). When are concerted reforms feasible? Explaining the emergence of social pacts in Western Europe. Comparative political studies, 43(5), 628-657. Baccaro, Lucio (2002). 'Negotiating the Italian pension reform with the unions: lessons for corporatist theory', ILR Review, 55:3, 413-431. Baccaro, Lucio and Sang-Hoom Lin (2007). "Social pacts as coalitions of the weak and moderate: Ireland, Italy and South Korea in comparative perspective”. European Journal of Industrial Relations, 13(1), 27-46. Bentolila, Samuel, Juan J. Dolado, and Juan F. Jimeno (2012). 'Reforming an insider-outsider labor market: the Spanish experience', IZA Journal of European Labor Studies, 1:1, 4. Bernecker, Walther L., and Günther Maihold (2007). España, del consenso a la polarización: cambios en la democracia española. Chicago: Iberoamericana. Bersani e Fornero a confronto su riforma del lavoro, esodati e abolizione Imu (video). [Bersani and Fornero Fight over Labour Market Reform]. (2012, September 18). Ilsussi Diario. Retrieved from http://www.ilsussidiario.net/News/Cinema-Televisione-e-Media/2012/9/19/BALLARO-Bersani-e- Fornero-a-confronto-su-riforma-del-lavoro-esodati-e-abolizione-Imu-video-18-settembre- 2012/322052/ (Accessed 12 May 2016). Blanchard, Olivier and Jean Tirole (2003). 'Protection de l'emploi et procédures de licenciement [Employment protection and firing procedures]'. La documentation française. Blyth, M. (2002). Great transformations: Economic ideas and institutional change in the twentieth century. Cambridge: Cambridge University Press. Boeri, Tito, and Pietro Garibaldi (2008). Un nuovo contratto per tutti. Milano: Chiarelettere. Boeri, Tito., Jose I. Conde-Ruiz, and Vincenzo Galasso (2003). 'Protecting against labour market risk: employment protection or unemployment benefits?' IZA Discussion paper series, No. 834. Boeri, Tito., Jose I. Conde-Ruiz, and Vincenzo Galasso (2012). 'The political economy of flexicurit', Journal of the European Economic Association, 10:4, 684-715. Bowley, Graham (2011, November 7). Interest Rates on Italian Bonds Pushed to New Levels. The New York Times. Retrieved from https://www.nytimes.com/2011/11/08/business/global/italy- bonds-push-higher.html (Accessed 9 May 2016). Camusso, Susanna (2012, January 5). “L’articolo 1 va bene così com’è [Article 18 is OK as It Is]”. . Retrieved from https://video.repubblica.it/economia-e-finanza/camusso--l-articolo- 18-va-bene-cosi-com-e/85658/84047 (Accessed 9 May 2016). CCOO (2010). "Las reformas en las políticas laborales por sí solas no generan empleo [Labour market reforms do not create employment by themselves]”. Retrieved from http://www2.ccoo.es/csccoo/menu.do?Areas:Empleo:Actualidad:77660 (Accessed 3 June 2017).

90 The reconfiguration of working class politics

Carreño, Belén (2009). Salgado ficha a un defensor del contrato único [Salgado Hires a Single Contract Advocate]. Público. Retrieved from http://www.publico.es/actualidad/salgado-ficha- defensordel-contrato-unico.html (Accessed 2 February 2017). Cebrián, Belén, and Carmen Parra (1993, April 2). El Gobierno propone un vuelco en las relaciones laborales, los salarios y los contratos [The Government Proposes a Radical Change in Labour Relations, Wages and Contracts]. El País. Retrieved from https://elpais.com/diario/1993/04/02/economia/733701620_850215.html (accessed 23 September 2015). Celestino Corbacho defiende abaratar el despido 'sin restar derechos a los trabajadores’ [Celestino Corbacho Defends Cheapening the Firing Process ‘Without Decreasing Workers’ Rights’]. (2010, April 12). El Mundo. Retrieved from http://www.elmundo.es/mundodinero/2010/04/12/economia/1271072041.html (Accessed 3 June 2017). Conde-Ruiz, Jose I. (2012). 'Por un contrato de igualdad de oportunidades [In Favor of an Equal Opportunity Contract]'. eldiario.es. Retrieved from https://www.eldiario.es/zonacritica/contrato- igualdad-oportunidades_6_79502069.html (Accessed 4 June 2017). Corbacho acusa al Banco de España de tener la "monomanía" de proponer abaratar el despido [Corbacho Blames Spanish Central Bank of Being “Obsessed” with Lower Firing Costs]. (2009, 2 November) Europa Press. Retrieved from http://www.europapress.es/economia/laboral- 00346/noticia-economia-laboral-corbacho-acusa-banco-espana-tener-monomania-proponer- abaratar-despido-20091102113042.html (Accessed 3 June 2017). Corbacho afirma que reducir el coste del despido "no serviría de nada" para frenar el paro [Corbacho Asserts that Reducing Firing Costs “Won’t Serve” to Decrease Unemployment Rate]. (2010, January 25). Europa Press. Retrieved from http://www.europapress.es/economia/noticia- corbacho-afirma-reducir-coste-despido-no-serviria-nada-frenar-paro-20100125181546.html (Accessed 3 June 2017). Corbacho culpa a la crisis de la necesidad de abaratar el despido [Corbacho Blames the Need to Decrease Firing Costs on the Crisis] (2010, May 28). El Diario Montañés. Retrieved from https://www.eldiariomontanes.es/v/20100528/economia/destacados/corbacho-culpa-crisis- necesidad-20100528.html (Accessed 3 June 2017). Corbacho: 'La referencia es el contrato de 33 días’ [Corbacho: ‘The 33-Day Contract Is the Reference]. (2010, June 1). El Mundo. Retrieved from http://www.elmundo.es/mundodinero/2010/06/01/economia/1275380358.html (Accessed 3 June 2017). Corbacho responde a F. Ordóñez que la crisis no tiene origen laboral [Corbacho Answers to F. Ordóñez that the Crisis Does Not Have a Labour-Related Origin]. (2009, February 2). El Confidencial. Retrieved from https://www.elconfidencial.com/mercados/2009-02-12/corbacho- responde-a-f-ordonez-que-la-crisis-no-tiene-origen-laboral_870068/ (Accessed 3 June 2017). Costamagna, Luca (2015, May 19). 'Renzi e l’amnesia sulla legge Fornero [Renzi and the Amnesia over the Fornero Law]'. Il Fatto Quotidiano. Retrieved from https://www.ilfattoquotidiano.it/2015/05/19/renzi-e-lamnesia-sulla-legge-fornero/1696812/ (Accessed 11 May 2016). De Marchis, Goffredo (2012, March 21). 'Riforma del lavoro, l'ira di Bersani "Monti ha rotto il tavolo, ora rischi per tutti” [Labour Market Reform, Bersani’s Rage: “Monti Has Broken the Table, Risking Everyone]'. La Repubblica. Retrieved from http://www.repubblica.it/politica/2012/03/21/news/lavoro_ira_bersani-31926516/ (Accessed 12 May 2016). Díez, Anabel (1994, February 28). 'Las bases del PSOE piden un giro a la izquierda del modelo propuesto por la dirección [PSOE’s Base Militants Ask for a Leftist Turn on the Policy Model Proposed by the Secretariat]'. El País. Retrieved from https://elpais.com/diario/1994/02/28/espana/762390018_850215.html (accessed 23 September 2015). A (social democratic) escape from dualisation? xci

Díez, Anabel (2009, January 23). 'La CEOE pide al PSOE abaratar el despido y reformas laborales [CEOE asks PSOE to Make Firings Cheaper and Labour Market Reforms]'. El País. Retrieved from https://elpais.com/diario/2009/01/23/economia/1232665209_850215.html (accessed 4 June 2017). Doncel, Luis (2010, April 11). España se mira en el espejo austriaco [Spain Looks at the Austrian Mirror]. El País. Retrieved from https://elpais.com/diario/2010/04/11/economia/1270936801_850215.html (accessed 3 June 2017). Eichhorst, Werner and Paul Marx (2011). 'Reforming German labour market institutions: A dual path to flexibility', Journal of European Social Policy, 21:1, 73-87. Eichhorst, Werner, Paul Marx, and Caroline Wehner (2017). 'Labor market reforms in Europe: towards more flexicure labor markets?', Journal for Labour Market Research, 51:1, 3. El Gobierno centra la reforma en los jóvenes y la contratación indefinida [The Government Focuses the Reform on the Youth and on Permanent Hirings] (2009, February 6). El País. Retrieved from https://elpais.com/diario/2010/02/06/economia/1265410805_850215.html (accessed 4 June 2017). El Gobierno cesa al director de Política Económica de Zapatero [The Government Fires Zapatero’s Head of Economic Policy]. El Periódico de Aragón. Retrieved from http://www.elperiodicodearagon.com/noticias/economia/gobierno-cesa-director-politica- economica-zapatero_593855.html (accessed 4 June 2017). Faillace, Giuseppina (2003). Il “lavoro in affitto”: dal “Pacchetto Treu” alla “Riforma Biagi” [Renting Jobs: From Treu Package to Biagi Reform]. Retrieved from the University di Pisa Thesis database (accessed 17 April 2015). FEDEA (2009). Propuesta para la reactivación laboral en España [A Proposal for Re-Activating the Spanish Labour Market]. Retrieved from https://www.fedea.net/propuesta-para-la-reactivacion- laboral-en-espana/ (accessed 4 June 2017). Fornero, Elsa (2012, April 2). “Siamo per la flessibilità buona [We Favor Good Flexibility]”. Sky News IT. Retrieved from https://tg24.sky.it/politica/2012/02/04/ministro_elsa_fornero_intervista_skytg24_lavoro_articolo_ 18.html (accessed 4 May 2015). Garea, Fernando (2009, May 13). "Zapatero lanza una ambiciosa reforma en busca de un nuevo modelo productivo [Zapatero Launches an Ambitious Reform to Find a New Production Model]". El País. Retrieved from https://elpais.com/diario/2009/05/13/espana/1242165601_850215.html (accessed 4 June 2017). Garibaldi, Pietro (2015a). "L’ Economia Politica del Jobs Act: Effetti, Rischi e sfide aperte [The Political Economy of Jobs Acts: Effects, Risks and Open Challenges] [PDF slides]". Retrieved from http://www.fondazionegiuseppepera.it/wp-content/uploads/2015/07/GARIBALDI.pdf (accessed 4 May 2015). Garibaldi, Pietro (2015b). 'Jobs act: il contratto a tutele crescenti è legge. È una buona notizia? [Jobs Act: Single Contract Is Law. Is It Good News?]. Il Fatto Quotidiano. Retrieved from https://www.ilfattoquotidiano.it/2015/02/24/jobs-act-contratto-tutele-crescenti-legge-buona- notizia/1451451/ (accessed 4 May 2015). Gómez, Manuel V. (2009a, February 5). "La patronal plantea que los despidos le cuesten menos de la mitad [Employers Organization Suggest Halving Firing Costs]". El País. Retrieved from https://elpais.com/diario/2009/02/05/madrid/1233836655_850215.html (accessed 4 June 2017). Gómez, Manuel V. (2009b, July 24). "El Gobierno carga contra la CEOE por defender el despido libre [The Government, Against CEOE for Defending Free Firings]". El País. Retrieved from https://elpais.com/diario/2009/07/24/economia/1248386402_850215.html (accessed 4 June 2017).

92 The reconfiguration of working class politics

Gómez, Manuel V. (2010, April 10). Los sindicatos rechazan la propuesta de reforma laboral del Gobierno [Unions Reject the Government’s Labour Market Reform Proposal]. El País. Retrieved from https://elpais.com/diario/2010/04/14/economia/1271196003_850215.html (accessed 4 June 2017). González, Juan J. and Miguel Caínzos (2012). 'Ciclos políticos y comportamiento electoral de jóvenes y mayores en España, 1979-2011'. Panorama Social, 15, pp. 165-179. Hancké, Bob and Martin Rhodes (2005). 'EMU and labor market institutions in Europe: The rise and fall of national social pacts'. Work and occupations, 32(2), 196-228. Jiménez, Miguel (2011, June 21). "El FMI pide a España una reforma laboral "más valiente" que rebaje el coste del despido [IMF asks Spain for a “braver” Labour Market Reform that Decreases Firing Costs]". El País. Retrieved from https://elpais.com/economia/2011/06/21/actualidad/1308641594_850215.html (accessed 3 June 2017). Griseri, Paolo (2012, January 19). "Riforma del lavoro, ecco il piano Fornero contratto unico, più soldi se lavori a termine [Labour Market Reform, This Is Fornero’s Single Contract Plan, with Higher Wages for Temporary Workers]". La Repubblica. Retrieved from http://www.repubblica.it/economia/2012/01/19/news/lavoro_ecco_piano_fornero-28395625/ (accessed 4 May 2016). Il caso Fassina [The Fassina Issue]. (2011, November 24). Il Post. Retrieved from https://www.ilpost.it/2011/11/24/il-caso-fassina/ (accessed 4 May 2016). Il Jobs act fa litigare il Pd. Malumori in casa Ncd. Cgil all’attacco: “Via ai licenziamenti di gruppo” [Jobs Act Provokes PD Infighting. Bad Mood in the NCD House. CGIL on the attack: "Leave Collective Layoffs”. (2014, December 26). La Stampa. Retrieved from http://www.lastampa.it/2014/12/26/italia/jobs-act-maggioranza-divisa-gelo-tra-pd-e-ncd-cgil- allattacco-via-ai-licenziamenti-di-gruppo-e8YHWn1xTIJIQjyjGhqJ5I/pagina.html (accessed 9 May 2016). Jobs act, Camera approva testo. Fuori dall’Aula Fi, Lega, M5s e 40 deputati Pd [Jobs Act, Parliament Approves the Law. FI, Lega, M5S and 40 PD MPs Leave the Room]. (2014, November 25). Il Fatto Quotidiano. Retrieved from https://www.ilfattoquotidiano.it/2014/11/25/jobs-act-camera- approva-testo-opposizioni-non-partecipano-voto/1232369/ (accessed 9 May 2016). Korpi, Walter (1985). 'Power resources approach vs. action and conflict: on causal and intentional explanations in the study of power'. Sociological Theory, 3:2, 31-45 La OCDE pide a Rajoy "reformas adicionales" en materia laboral [OECD Asks Rajoy for “Additional Reforms” in the Labour Market]. (2013, December 18). DW. Retrieved from https://www.dw.com/es/la-ocde-pide-a-rajoy-reformas-adicionales-en-materia-laboral/a- 17305366 (accessed 9 September 2017). Lavoro, Fornero rinvia l’incontro con le parti sociali. “Serve riforma ammortizzatori” [Labour: Fornero Restarts Meetings with Social Agents]. (2012, February 29). Il Fatto Quotidiano. Retrieved from https://www.ilfattoquotidiano.it/2012/02/29/lavoro-fornero-rinvia-lincontro-parti-sociali-serve- riforma-ammortizzatori/194684/ (accessed 12 May 2016). Fornero: sulla riforma del lavoro il Governo procederà anche senza il consenso dei partiti [Fornero: On Labour Market Reform the Government Will Proceed even without Party Consensus]. (2012, February 22). Il Sole 24 Ore. Retrieved from http://www.ilsole24ore.com/art/notizie/2012-02- 22/fornero-riforma-lavoro-senza-181414.shtml?uuid=AaxwdvvE&refresh_ce=1 (accessed 12 May 2016). Los sindicatos insisten en negociar por separado el modelo laboral 'alemán' del ‘austriaco’ [Unions Insist on a Separate Negotiation for the German and the Austrian Labour Models] (2010, April 19). El Mundo. Retrieved from http://www.elmundo.es/mundodinero/2010/04/19/economia/1271702332.html (accessed 12 September 2017). Lucidi, Federico and Alfred Kleinknecht (2009). 'Little innovation, many jobs: An econometric analysis of the Italian labour productivity crisis'. Cambridge Journal of Economics, 34:3, 525-546. A (social democratic) escape from dualisation? xciii

Mackenzie, James (2010, May 14)."Italy's Berlusconi says he was forced out by EU ‘plot’". Reuters. Retrieved from https://www.reuters.com/article/us-italy-berlusconi/italys-berlusconi-says-he- was-forced-out-by-eu-plot-idUSBREA4D0N720140514?feedName=worldNews (accessed 10 May 2016). Maltese, Curzio (2012, February 19). "Veltroni: "Basta tabù sull'articolo 18 Non lasciamo Monti alla destra” [Veltroni: Enough Taboo over Article 18. Let’s Not Leave Monti to the Right]". La Repubblica. Retrieved from http://www.repubblica.it/politica/2012/02/19/news/veltroni_articolo_18-30134125/ (accessed 12 May 2016). Mania, Roberto (2012, January 5). "Camusso: "Pronti a un nuovo patto ma sui licenziamenti non si negozia” [Camusso: “We Are Ready For an New Agreement but We Won’t Negotiate Firing Regulations”]". La Repubblica. Retrieved from http://www.repubblica.it/economia/2012/01/05/news/camusso-27617996/?ref=HREA-1 (accessed 11 May 2016). Maqueda, Antonio and Manuel V. Gómez (2017, November 24). "Bruselas sitúa a España a la cabeza de la desigualdad por renta en la UE [Brussels puts Spain up in the EU list of income inequality]". El País. Retrieved from https://elpais.com/economia/2017/11/23/actualidad/1511465471_017133.html (Accessed 25 November 2018). Mateos, Abdón (1993). El PSOE contra Franco: continuidad y renovación del socialismo español, 1953-1974. Madrid: Pablo Iglesias. Molina, Oscar and Martin Rhodes (2007). 'The political economy of adjustment in mixed market economies: A study of Spain and Italy' in Hancké, Bob, Martin Rhodes, and Mark Thatcher (eds.), Beyond Varieties of Capitalism: Conflict, Contradictions and Complementarities in the European Economy (223-252). Oxford: Oxford University Press. Moody, Barry and James Mackenzie (2011, November 7). Berlusconi to resign after parliamentary setback. Reuters. Retrieved from https://uk.reuters.com/article/uk-italy/berlusconi-faces-crunch- vote-pressure-to-quit-idUKTRE7A62D520111108 (accessed 11 May 2016). Nalbone, Daniele (2012, March 26). Bersani attacca: "Sulla riforma discussione vera in Parlamento” [Bersani Attacks: “On Labour Market Reform, the Real Debate Will Happen in the Parliament”]. Today IT. Retrieved from http://www.today.it/politica/bersani-pd-direzione-riforma-lavoro- posizione.html (accessed 11 May 2016). Noceda, Miguel A. (2009, March 29). '"Hay que subir los impuestos a los ricos” — entrevista a Cándido Méndez [“We Should Increase Taxes for Rich People” — an interview to Cándido Méndez]'. El País. Retrieved from https://elpais.com/diario/2009/03/29/economia/1238281201_850215.html (accessed 3 June 2017) Noceda, Miguel A. and Gómez, Manuel V. (2010). 'Sindicatos y empresarios apurarán hasta el lunes para pactar la reforma laboral [Unions and Employers Will Take Negotiations Until Monday to Agree on the Labour Market Reform]'. El País. Retrieved from https://elpais.com/diario/2010/05/27/economia/1274911206_850215.html (accessed 3 June 2017). OECD (2015). OECD economic surveys: Italy. Paris: Organisation for Economic Co-operation and Development Palombi, Marco (2012, January 19). "Contratto unico d’ingresso senza toccare l’articolo 18, il governo rilancia sul lavoro [Single Contract without Changing Article 18: Government Relaunches Labour]”. Il Fatto Quotidiano. Retrieved from https://www.ilfattoquotidiano.it/2012/01/19/contratto-unico-dingresso-senza-toccare-lart- governo-rilancia-riforma-lavoro/185043/ (accessed 12 May 2016). Pasquino, Giovanni and Matteo Valbruzzi (2017). "The Italian Democratic Party, its nature and its Secretary". Revista española de ciencia política, 44, 275-299.

94 The reconfiguration of working class politics

Pd, la «grana» Fassina. I liberal: «Si dimetta» Ma Bersani frena: «Questa non l'ho capita» [Pd, the «grana» Fassina. Liberals: «You resign» But Bersani brakes: «I did not understand this»]. (2011, November 23). Il Corriere della Sera. Retrieved from https://www.corriere.it/politica/11_novembre_23/pd-caso-fassina-bersani-non- capisco_e8301088-15e7-11e1-abcc-e3bae570f188.shtml (accessed 11 May 2016). Peters, B. Guy, Jon Pierre, and Desmond S. King (2005). 'The politics of path dependency: Political conflict in historical institutionalism'. The journal of politics, 67:4, 1275-1300. Picot, Georg and Alessandro Tassinari (2014). "Liberalization, dualization, or recalibration? Labor market reforms under austerity, Italy and Spain 2010-2012." Welfare State in Portugal in the Age of Austerity. Picot, Georg and Alessandro Tassinari (2017). 'All of one kind? Labour market reforms under austerity in Italy and Spain'. Socio-Economic Review, 15(2), 461-482. Pochet, Philippe and Giuseppe Fajertag (2000). 'A new era for social pacts in Europe'. Social pacts in Europe—New dynamics, 9-40. Portanova, Mario (2013, December 17). "Lavoro, il piano di Renzi per i giovani: no al precariato, ma licenziamenti più facili [Labour Market, Renzi’s Plan for the Youth: No to Preciariat, but Easier Firings]". Il Fatto Quotidiano. Retrieved from https://www.ilfattoquotidiano.it/2013/12/17/lavoro-il- piano-di-renzi-per-i-giovani-no-al-precariato-ma-licenziamenti-piu-facili/815694/ (accessed 9 May 2016). Pozzi, Sandro (2010). "El FMI apoya un contrato único con despido creciente [IMF Supports a Single Contract with Increasing Firing Costs]". El País. Retrieved from https://elpais.com/diario/2010/04/15/economia/1271282403_850215.html (accessed 23 September 2017). Renzi, Matteo (2015). “Noi siamo preoccupati non di Margaret Thatcher, siamo preoccupati di Marta [We Are Worried not about Margaret Tatcher, We Are Worried about Marta]”. Retrieved from https://www.blitzquotidiano.it/blitztv/renzi-videomessaggio-thatcher-penso-a-marta-cgil-sui- precari-dove-eravate-1975335/ (accessed 11 May 2016). Roubini, Nouriel and Stephen Mihm (2011, May 16). "Can Europe be Saved?". Slate. Retrieved from https://slate.com/business/2011/05/the-piigs-crisis-can-portugal-ireland-italy-greece-and-spain- be-saved.html (Accessed 25 November 2018). Rueda, David (2005). 'Insider-outsider politics in industrialized democracies: the challenge to social democratic parties', American Political Science Review, 99:1, 61-74. Rueda, David (2007). Social democracy inside out: Partisanship and labor market policy in advanced industrialized democracies. Oxford: Oxford University Press. Saint-Paul, Gilles (2000). The political economy of labour market institutions. Oxford: Oxford University Press. Scontro Fornero-Bersani sulla riforma del lavoro [Fornero-Bersani Standoff on Labour Market Reform]. (2012, February 23). Cinquantamila. Retrieved from http://www.cinquantamila.it/storyTellerArticolo.php?storyId=4f46a28e04b3e (accessed 9 May 2016). Salgado rechaza un contrato único con despido más barato [Salgado Rejects a Single Contract With Lower Firing Costs] (2010, April 5). Europa Press. Retrieved from http://www.europapress.es/economia/noticia-salgado-rechaza-contrato-unico-despido-mas- barato-20100405145528.html (accessed 13 June 2015). Santesmases, Antonio G. (1993). Repensar la izquierda: evolución ideológica del socialismo en la España actual. Madrid:Anthropos Editorial. Sarti, M. (2012, August 27). Il Pd censura la Fornero alle feste di partito. I renziani non ci stanno [PD Won’t Invite Fornero to the Party. Renziani Won’t Attend]. Linkiesta. Retrieved from https://www.linkiesta.it/it/article/2012/08/27/il-pd-censura-la-fornero-alle-feste-di-partito-i- renziani-non-ci-stann/8982/ (accessed 9 May 2016). A (social democratic) escape from dualisation? xcv

Solbes, P. (2013). Recuerdos: 40 años de servicio público. Madrid: Grupo Planeta. Pagnoncelli, Nando (2016, November 14). 'Più imprenditori, meno impiegati. Così si trasforma l’elettorato del Pd [More Entrepreneurs, Less Workers. This is how the PD Electorate Changes]'. Corriere della Sera. Retrieved from https://www.corriere.it/politica/14_novembre_16/piu-imprenditori-meno-impiegati-cosi-si- trasforma-l-elettorato-pd-47014e3a-6d72-11e4-a925-1745c90ecb18.shtml (accessed 9 May 2016). UGT detecta un fuerte aumento de la precariedad en el mercado de trabajo [UGT Detects a Strong Increase on Preciariousness] (1992, August 16). El País. Retrieved from https://elpais.com/diario/1992/08/16/economia/713916010_850215.html (accessed 13 June 2015). Ventura, Sofia (2018). The Italian democratic party from merger to personalism. South European Society and Politics, 23(1), 181-196. Ximénez de Sandoval, Pablo and Fernando Garea (2010, June 23). El PSOE convalida en solitario una reforma laboral provisional [PSOE Approves Alone a Provisional Labour Market Reform]. El País. Retrieved from https://elpais.com/diario/2010/06/23/espana/1277244001_850215.html (accessed 23 September 2017).

96 The reconfiguration of working class politics

Concluding remarks xcvii

Chapter 6. Concluding remarks

The last decades have witnessed deep and far-reaching changes in working class politics. Workers have been at the forefront of the public debate both as subjects of political decisions (as voters that might be leaving old, mainstream parties in search for new horizons) and as targets of broad welfare and labour market reforms that are reconfiguring the Post-war equilibrium that shaped Western political dynamics. Its chief architects, especially social democratic platforms, are under unparalleled pressure to deliver both electorally and policy-wise. These tensions interact with each other while maintaining workers at the centre of political action. The two key questions driving the present dissertation were: What is happening with working class voters of traditional parties, particularly the social democracy? And what are the underlying coalitions for and against key labour market reforms? In Chapter 1 I argued that considering both questions together (in sequence) would help to shed light on the vote-party-government-policy channel that remains at the forefront of political decision making.

6.1 Voting choices: class matters (somehow) There are three initial insights to derive from the past chapters. First, class still matters in defining individuals’ political decision. Second, it does so only to a certain extent as members of the working classes do not behave as a political bloc. Third, the definition of class is changing, and new frameworks like the division between insider (protected against the risk of unemployment) and outsider (unprotected) workers are gaining some ground without completely erasing previous divisions. Let us break these insights down through analysis. The fact that class still matters is well proven by chapter two, in which I traced the patterns of manual workers’ voting decisions within and away from social democratic parties (SDP), and towards the extremes of the . It demonstrated first that, in aggregate terms, manual workers are turning away from voting. Their relationship with the far-right is statistically significant as well, but in absolute terms, abstention is their main destination. Since the (re)distribution of working class voters is multi-directional, and since there are no perfect correlations between class belonging and voting choice in any case, it becomes evident that class only matters to a certain extent, and that additional factors are needed to explain the aforementioned patterns. Working class voters are pushed towards abstention chiefly by two factors: mistrust in existing institutions and the feeling of economic anxiety. The choice for the far-right, instead, is linked to preferences for anti-immigration preferences. The key revelation of chapter two comes, however, from measuring how much economic anxiety and anti-immigration sentiment have become more important after the Great Recession (i.e. after 2008). Chapter three considered the structural determinants on choosing populist radical right (PRR) parties in Europe as well, but the analysis departed from the insider/outsider framework, expecting to identify a positive relationship between being insider (i.e. sheltered against labour market risks thanks to a permanent contract) and favouring PRR platforms. More interestingly, adding to previous research, ours comes closer to linking the trend to economic status pressure (or its perception thereof), since the insider-far-right link is more intense for those individuals facing higher combined levels of unemployment and migrant presence. Instead, no effect could be attached to non-economic dimensions of status threat. At the same time, the effect grew in size when Bismarckian welfare states were separately considered, and correlated with lower union density, thus suggesting that contextual institutions matter to determine how much insiders want to activate exclusionary political mechanisms, probably driven by a rough personal calculation of how much they stood to lose. Adding to this conclusion came the third and last finding of the chapter: insiders belonging to the working class and to mid- and low-skilled segments (and therefore facing a higher degree of labour market vulnerability) are even more overrepresented in PRR parties, suggesting that neither the more classical occupation-based division, nor the status-based distinction between insiders and outsiders, should be discarded to fully understand the changing face of working class politics.

98 The reconfiguration of working class politics

6.2 Policy choices: class vs job position A good opportunity to observe this changing face emerges when moving the analysis from politics to policies, as is done in Chapter 4. Focusing on the same insider/outsider framework and its original proposition of insiders being more likely to favour employment protection legislation (EPL) and less likely to favour other forms of protection against the risk of unemployment (unemployment benefits - NEB, active labour market policies - ALMP), one wonders whether outsiders were actually more in favour of flexicurity arrangements (high NEB/ALMP, low EPL), or whether skill/occupation matters more than job status. My (admittedly preliminary) analysis is, to my knowledge, the first one to systematically feature all three key dimensions defining flexicurity (EPL, NEB, ALMP) for a wide range of Continental European countries since Rueda (2005), with the novel feature of pitching, at the same time, main explanatory frameworks against each other. Results show that the employed/unemployed divide (akin to the insider/outsider classification) and occupational class differences à la Oesch convey similar explanatory power. As predicted by the insider/outsider theoretical framework, the unemployed are more likely to prefer EPL reductions, ALMP/NEB increases, and as a consequence they prioritise flexicurity over dualisation. However, they do not seem to prefer liberalisation over full-scale protection (keeping EPL, NEB and ALMP high). Simultaneously, employers and professionals do display more intense preferences than mid- and low-skilled (manual) workers for any form of flexicurity or liberalisation. These findings only complicate the dilemma faced by SDP regarding labour market policies. Chapter 5 illustrates the possible strategies. Italy and Spain are countries that have several things in common: first, they share a past of labour market dualisation (making thus perfect case studies for insider/outsider perspectives). Second, dualisation may make things harder for centre-left parties. As a matter of fact, both the Spanish PSOE and the Italian PD host contending factions with regards to labour market regulation: a liberal-leaning, economist section that favours reforms co-exists along a more protectionist wing linked to unions (UGT in Spain, CGIL in Italy). Third, in both countries the 2010-2015 EMU crisis was particularly acute, with intense calls from within and from abroad to reform their labour markets in order to favour productivity increases in the long term. These calls were fairly aligned with demands made from the liberal-leaning wings of SDP elites, while met resistance from the unionist factions. Fourth, both the PSOE and the PD enjoyed considerable time in power during the worst years of the EMU crisis, thus making them responsible for reforms or the lack thereof. All these coinciding factors make the Italy-Spain case a perfect concluding comparison, not the least because the outcome in each country was different: The Italian PD pursued a full-scale de-dualising reform while the Spanish PSOE made a timid attempt at changing regulations. In Chapter four it was shown that EPL reductions were least favoured by (low-skilled) insiders. Chapter five discusses the possibility of these voters being at the core of the PSOE base, but not the PD base, which in turn gave more space for socio-liberals in Italy to pursue and enact full-scale reforms. It could be said that while the PD adopted a more liberal approach likely coupled with the composition of its constituency, the PSOE preferred to maintain a relatively protectionist position in order to not damage its own standing.

6.3 Cross-chapter potential connections: paths for future research The present dissertation considers both voting and policy aspects of individuals’ political preferences, but it does so in an admittedly compartmentalised manner. A joint consideration of the insights summarised above does, however, illuminate some interesting areas for further research. First and foremost, do policy preferences regarding labour market reforms (both before and after its enactment) affect workers’ voting decisions? If, as shown by chapter three, insiders are more likely to choose PRR parties (particularly when they feel like there is a threat to their material status), it would make sense to expect reforms or proposals that favour labour market flexibilisation to move them towards the far-right. Or, at least, to anti-establishment parties that advocate exclusionary protectionist measures. Italy itself might be an interesting case study in this regard, given the recent rise of an anti- establishment+PRR coalition in its government, just following Matteo Renzi’s de-dualising reform effort. Second, do new (particularly far-right) parties respond in any way to the incorporation of certain segments of workers? Afonso and Rennwald (2018) have already suggested that a left-authoritarian platform that advocates welfare chauvinist measures could become a major opportunity for far-right Concluding remarks xcix parties. There are, however, some limitations to it, that were highlighted already by Herbert Kitschelt’s pioneering work (Kitschelt 2007): how easy would it be to pair these “exclusionary pro-redistribution” positions with the current far-right’s appeal among small owners (see chapter two)? A possible way out could be given by an anti-establishment discourse based on mistrust towards political institutions, one that would suggest that further efforts by the state towards native workers could be done if existing resources were better administered. Chapter two highlights that mistrust towards institutions seems to drive more workers towards abstention than towards anywhere else, casting doubt on the far-right’s capacity to use that as an electoral leverage. Third, the potentially intricated link between the consequences of the Great Recession and workers’ preferences over policies and party platforms is only partially explored, both within the present work and beyond it. In several European countries, particularly in those most greatly affected by markets’ lack of confidence, the second phase of the recession contributed to provoke a series of policy decisions. Chapter five takes a close look at the Italian and Spanish case, while chapter two finds that the intensity of determinants of abstention and far-right vote among workers have intensified since 2008. Putting both insights together might indicate a promising route for further research: has the crisis itself made these factors more prominent, or are governments’ response to the downturn and external pressure that has in turn provoked a more intense dynamic? Fourth, if neither occupational class nor job status offer a degree of explanatory power and consistency when it comes to defining both policy and party preferences comparable to what they used to do, it might be worth further reconsidering the lines we use to divide and classify workers, as it has been recently done by some authors (e.g. Hausermann and Schwander 2013).

6.4 Beyond research Reconsiderations do not stop within the academy. A take-home message for social democratic parties (and for every party for that matter) when looking to build a platform to appeal workers seems to be that choice is inevitable. For Przeworski (1986), this choice was between revolution and moderation. After several decades of electoral bonanza, now the choice has returned in a new form that seems to be less dramatic, but more complex, operating at several levels. Here, I have analysed some of these angles, and all of them point to an irreversibly divided electorate, where classical occupational and class breaches are mixed with both structural and ideological factors, which adds itself to previous research pointing in the same direction. More specifically, the research presented here seems to make it hard for social democracy to keep a culturally liberal, pro-liberal democracy set of positions while appealing to the working class as a whole. Though, at the same time, it also looks like those who might aspire to substitute social democracy as the workers’ party will face their own version of the insider/outsider dilemma. The dilemma seems to be in play for center-left parties in places like Italy and Spain, although it has been shown that it is as related to internal elite divisions as much as it might be to the parties’ electoral base. Moreover, there seems to be a lack of widely popular alternatives to produce labour market reforms that compel wide sectors of the population, which adds to the problems of reuniting the working class under a single political umbrella. A final question remains, however: Is the current course of events a structural trend, or are we just witnessing a transitional period towards a new policy platform-party-class realignment? From an analytical standpoint the question is a tricky one. Everything seems transitional if looked at from far enough. As a matter of fact, the present work has focused on the demand side of the problem. However, taking a supply side perspective, from the point of view of those directly involved in the political process, the question could become a different query asking what, if anything, could be done to change the current equilibrium.

100 The reconfiguration of working class politics

References Afonso, Alexandre, and Linne Rennwald (2018). "Social Class and the Changing Welfare State Agenda of Radical Right Parties in Europe". Welfare Democracies and Party Politics: Explaining Electoral Dynamics in Times of Changing Welfare Capitalism, 171.

Häusermann, Silja and Hannah Schwander (2013). "Who is in and who is out? A risk-based conceptualization of insiders and outsiders". Journal of European Social Policy, 23(3), 248-269.

Kitschelt, Herbert (2007). "Growth and persistence of the radical right in postindustrial democracies: Advances and challenges in comparative research". West European Politics, 30(5), 1176-1206.

Przeworski, Adam (1986). Capitalism and social democracy. Cambridge University Press.

Appendix

A. Where have all the working class voters gone?

Table A.1. Descriptive statistics

Variables n Mean Min Max Categories Age 272,439 2.911 0 5 0 = 18-24; 1 = 25-34; 2 = 35-44; 3 = 45-54; 4 = 55-65; 5 = 65+ Gender 276811 1.520 1 2 1 = Male; 2= Female Education 275595 3.133 1 5 1 = ISCED 0-1; 2 = ISCED 2; 3 = ISCED 3; 4 = ISCED 4; 5 = ISCED 5-6 Social Class 268204 5.166 1 8 Based on Oesch (2006) social class scheme. Born in country 276595 1.086 1 2 1 = Yes; 2 = No Redistribution 272415 0.287 0 1 “Government should reduce income differences” | Scale: 0 = Agree strongly; 1 = Disagree strongly Trust institutions 274509 0.39 0 1 Composite index, with 0=no trust and 1=complete trust. Factors in and includes a standardized version of the three following variables: - Trust in political parties - Trust in politicians - Trust in the national parliament

Anti-Immigration 244373 2.869 0 1 Based on Polavieja (2016), with 0=for immigration and 1=against immigration. Factors in and includes the following variables: - Allow many/few immigrants of same race/ethnic group as majority - Allow many/few immigrants of different race/ethnic group from majority - Allow many/few immigrants from poorer countries outside Europe - Immigration bad or good for country's economy - Country's cultural life undermined or enriched by immigrants - Immigrants make country worse or better place to live Scale: 1 = Pro-immigration; 5 = Anti-immigration Feeling over income (economic anxiety) 271974 0.331 0 1 “Feeling about household’s income nowadays” | Scale: 0 = Living comfortably; 1 = Very difficult on present income

Table A.2. Far right, far left and social democratic parties included in the dataset

Note: based on Chapel Hill expert surveys (CHES 2016) with some modifications. Switzerland: SVP is reclassified as 'right' from 'agr' Spain: PA is considered as 'spec' instead of 'soc'; NA-BAI is classified as 'spec' (it had no coding in CHES). Portugal: PNR is classified as 'right'. Table A.3. Oesch's class schema applied to all countries in the sample

Self-empl prof & Small business Technical (semi) Production worker (Associate) manager Clerk Socio-cult (semi) Service worker large business owner owner prof. prof. AT 203 892 489 1,768 910 1,685 1,131 2,160 BE 298 1,261 963 2,454 1,883 1,292 1,539 2,089 BG 60 465 450 2,780 845 408 660 1,479 CH 422 1,383 1,055 1,991 2,147 1,261 1,751 2,184 CZ 190 1,107 825 4,149 1,461 1,309 1,197 2,401 DE 616 1,713 1,640 4,563 2,975 1,766 2,526 3,340 DK 244 743 646 1,917 1,494 823 1,559 2,028 EE 207 719 720 3,831 1,911 591 1,295 2,183 ES 263 1,965 541 3,257 1,254 983 922 2,973 FI 272 1,855 1,275 3,363 1,921 1,009 1,805 3,090 FR 158 906 1,006 1,956 1,715 1,090 1,297 1,911 GB 366 1,694 804 2,606 2,383 1,912 1,780 3,820 GR 223 2,420 172 1,468 447 826 493 1,488 HU 96 685 469 3,641 943 926 811 1,944 IE 363 1,931 702 2,996 1,987 1,653 1,678 3,758 IT 168 924 200 1,144 423 582 489 742 LT 38 380 275 2,355 805 312 778 1,486 LU 68 245 140 544 395 306 277 526 NL 346 1,230 696 1,772 3,018 1,598 2,199 2,489 NO 174 886 1,036 1,973 2,035 831 1,908 2,891 PL 160 1,968 673 3,450 1,980 755 859 1,934 PT 216 1,708 435 3,697 708 1,235 827 3,098 SE 367 1,191 1,053 2,224 1,974 1,124 1,880 3,011 SI 101 756 662 2,675 1,187 856 840 1,439 SK 99 506 369 2,484 923 709 777 1,291

5,718 29,533 17,296 65,058 37,724 25,842 31,278 55,755 Note: classification based on Oesch (2006), simplified form

Table A.4. Logistic regressions - determinants of vote choice (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) far left socdem greens agricultural issue VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

Gender (1=male; 2=female) -0.179*** -0.251*** -0.0258 -0.0565** 0.315*** 0.187*** -0.242 0.150*** -0.242*** -0.121 (0.0541) (0.0510) (0.0254) (0.0288) (0.0569) (0.0628) (0.164) (0.0552) (0.0523) (0.136) Age (five groups) 0.00570 0.0553 0.132*** 0.180*** -0.230*** -0.133*** 0.171*** 0.124*** 0.0130 -0.137** (0.0530) (0.0603) (0.0182) (0.0158) (0.0361) (0.0241) (0.0173) (0.0285) (0.0750) (0.0579) Born in country (1=yes; 2=no) -0.934*** -0.792*** -0.486** -0.344** -0.464*** -0.907*** -1.682*** -1.176*** -2.086*** -1.049*** (0.241) (0.182) (0.197) (0.162) (0.0538) (0.119) (0.358) (0.331) (0.404) (0.237) Final Oesch class position (8 classes; 1 = large business owners and managers)

Final Oesch class position - 8 classes = 2, Small business owners -0.202*** -0.0809 -0.152 -0.128 -0.800*** -0.579*** 0.949*** 1.038*** -1.112*** 0.0150 (0.0552) (0.171) (0.104) (0.104) (0.261) (0.152) (0.234) (0.355) (0.240) (0.376) Final Oesch class position - 8 classes = 3, Technical (semi-) professionals 0.230* 0.410* 0.612*** 0.516*** -0.538*** -0.482*** -0.0458 0.274 -0.167 -0.132 (0.122) (0.229) (0.0510) (0.102) (0.197) (0.0433) (0.232) (0.343) (0.445) (0.252) Final Oesch class position - 8 classes = 4, Production workers 0.268*** 0.347** 0.353*** 0.357*** -1.533*** -1.264*** 0.146 0.134 -1.040*** -0.149 (0.0753) (0.170) (0.128) (0.120) (0.389) (0.0657) (0.136) (0.406) (0.139) (0.360) Final Oesch class position - 8 classes = 5, (Associate) managers 0.154 0.297 0.393*** 0.490*** -0.717*** -0.565*** -0.142 -0.0643 -0.449*** -0.384 (0.137) (0.188) (0.0458) (0.128) (0.217) (0.0667) (0.155) (0.401) (0.0553) (0.315) Final Oesch class position - 8 classes = 6, Clerks 0.243** 0.178 0.304*** 0.532*** -0.943*** -0.760*** 0.0274 -0.275 -0.592*** 0.0204 (0.115) (0.259) (0.0666) (0.118) (0.250) (0.105) (0.213) (0.500) (0.142) (0.439) Final Oesch class position - 8 classes = 7, Socio-cultural (semi-) professionals 0.675*** 0.778*** 0.543*** 0.609*** -0.0185 -0.00395 0.0998 -0.00357 -0.461** 0.240 (0.119) (0.229) (0.0805) (0.141) (0.166) (0.0890) (0.228) (0.522) (0.179) (0.248) Final Oesch class position - 8 classes = 8, Service workers 0.203* 0.376* 0.360*** 0.383*** -1.107*** -1.135*** 0.145** 0.0892 -0.990*** -0.294 (0.119) (0.206) (0.116) (0.116) (0.311) (0.101) (0.0612) (0.440) (0.125) (0.343) Country (1=AT)

Country = 2, BE -0.597*** 1.256*** -0.204*** -0.337*** 0.129*** -0.269*** (0.0402) (0.0932) (0.0197) (0.0424) (0.0391) (0.0393) Country = 3, CH 0.786*** -1.075*** -0.549*** -1.112*** -0.576*** -0.259*** (0.0373) (0.0988) (0.0242) (0.0419) (0.0338) (0.0407) Country = 4, CZ 3.771*** 2.483*** -0.154*** -0.376*** -1.589*** -2.869*** -4.255*** -1.440*** (0.0463) (0.0881) (0.0133) (0.0508) (0.0517) (0.0338) (0.0380) (0.0897) Country = 5, DE 3.236*** 2.163*** 0.445*** -0.0959** -0.392*** -0.363*** 1.936*** (0.0496) (0.0939) (0.0204) (0.0427) (0.0396) (0.0394) (0.269) Country = 24, BG 0.331*** -0.597*** -2.752*** -5.245*** (0.0886) (0.0971) (0.109) (0.0783) Country = 6, EE -0.731*** 0.137*** -1.765*** -2.102*** -0.272*** -2.916*** (0.0581) (0.0477) (0.0654) (0.0354) (0.0407) (0.106) Country = 7, ES 3.269*** 2.716*** 0.577*** -0.0529 1.873*** 3.734*** (0.0388) (0.0978) (0.0351) (0.0454) (0.0792) (0.270) Country = 8, FI 3.485*** 2.997*** -0.204*** -0.824*** -0.361*** -0.362*** 1.107*** 5.325*** -2.825*** 2.769*** (0.0478) (0.0980) (0.0224) (0.0490) (0.0436) (0.0430) (0.0208) (0.136) (0.0947) (0.268) Country = 9, FR 3.558*** 2.007*** 0.167** -0.200*** -0.849*** -0.787*** -1.546*** 1.761*** (0.0830) (0.0899) (0.0738) (0.0436) (0.0868) (0.0388) (0.0220) (0.138) Country = 10, GB 0.247*** -0.740*** 0.578*** 0.0617 -2.225*** -1.997*** 0.207** 2.265*** (0.0413) (0.0967) (0.0215) (0.0427) (0.0393) (0.0403) (0.0844) (0.264) Country = 11, HU 0.731*** -0.798*** 0.494*** -0.860*** -1.333*** -0.182 (0.0664) (0.104) (0.0469) (0.0550) (0.0377) (0.124) Country = 12, IE 3.434*** 2.847*** -1.135*** -1.270*** -1.377*** -2.249*** (0.0353) (0.0882) (0.0217) (0.0425) (0.0391) (0.0367) Country = 15, IT 3.544*** 1.481*** -0.372*** -0.276*** -1.877*** 4.964*** (0.0235) (0.0562) (0.0353) (0.0356) (0.0343) (0.317) Country = 17, NL 3.956*** 2.782*** -0.147*** -0.604*** -0.731*** -1.180*** -1.277*** 3.205*** (0.0450) (0.0910) (0.0133) (0.0445) (0.0377) (0.0403) (0.0891) (0.267) Country = 18, NO 4.403*** 2.379*** 0.181*** 0.230*** -0.172*** 4.255*** (0.0408) (0.0842) (0.0207) (0.0418) (0.00990) (0.144) Country = 23, SI 0.787*** 0.329*** 0.726*** -2.850*** -4.949*** 1.942*** 4.199*** (0.0980) (0.0263) (0.0449) (0.0433) (0.0359) (0.0818) (0.267) Country = 20, PT 3.778*** 2.695*** 0.312*** -0.145*** -5.264*** -3.576*** (0.0479) (0.105) (0.0391) (0.0552) (0.0556) (0.0384) Country = 22, SE 3.784*** 2.219*** 0.527*** 0.0699 -0.777*** -0.373*** 4.400*** 2.960*** (0.0413) (0.0959) (0.0205) (0.0426) (0.0381) (0.0433) (0.134) (0.265) Country = 16, LT -1.225*** -0.662*** -4.469*** 4.275*** 4.703*** (0.104) (0.0548) (0.0388) (0.129) (0.259) Country = 19, PL -1.195*** -0.392*** -1.883*** -0.318*** 3.519*** -2.596*** (0.0908) (0.0283) (0.0497) (0.0229) (0.175) (0.0656) Country = 25, DK 2.549*** 1.939*** 0.191*** -0.293*** -0.0596 -0.0666 (0.0489) (0.182) (0.0181) (0.0724) (0.0384) (0.0667) Country = 26, GR 4.125*** 2.965*** 0.406*** 0.139 -2.577*** -1.954*** (0.0377) (0.238) (0.0284) (0.128) (0.0480) (0.116) Country = 27, LU 1.646*** -0.694*** -0.693*** (0.0584) (0.0607) (0.0384) Country = 28, SK 1.704*** 0.181*** 0.206** (0.0573) (0.0571) (0.0930) ESS round (1=2002)

ESS round = 2 0.181*** -0.0272 -0.0631 -0.426* -0.206 (0.0563) (0.0815) (0.0675) (0.231) (0.372) ESS round = 3 0.154* -0.187 -0.0335 0.123*** 0.253*** (0.0861) (0.139) (0.123) (0.0418) (0.0848) ESS round = 4 0.237** -0.201 0.176 -0.765* -0.806*** (0.121) (0.143) (0.145) (0.395) (0.252) ESS round (5=2010)

ESS round = 6 0.233 0.0903 -0.270 0.0239 0.755** (0.163) (0.109) (0.195) (0.627) (0.327) ESS round = 7 0.177 -0.113 -0.190** 0.000444 0.278 (0.122) (0.0984) (0.0936) (0.561) (0.416) ESS round = 8 0.316 -0.0719 -0.119 0.0161 0.778** (0.331) (0.143) (0.163) (0.495) (0.371) Constant -5.729*** -4.747*** -1.427*** -1.645*** -1.011*** -0.313** -1.215*** -6.803*** -2.147*** -5.987*** (0.252) (0.226) (0.291) (0.286) (0.368) (0.153) (0.455) (0.316) (0.598) (0.338)

Observations 116,895 115,865 135,101 129,433 108,187 107,246 40,669 44,809 52,385 55,653 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.4. (cont) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) christian liberals dem conservatives far right abstention VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

- Gender (1=male; 2=female) -0.0396 -0.0755 -0.188*** -0.0164 -0.101*** 0.0836*** -0.272*** -0.226*** 0.156*** 0.164*** (0.0479) (0.0519) (0.0687) (0.0509) (0.0210) (0.0301) (0.0622) (0.0794) (0.0444) (0.0500) Age (five groups) 0.0237 0.00133 0.175*** 0.159*** 0.0896* 0.125** 0.0886** 0.0952*** -0.222*** -0.218*** (0.0216) (0.0262) (0.0130) (0.0153) (0.0496) (0.0515) (0.0380) (0.0349) (0.0386) (0.0348) Born in country (1=yes; 2=no) -0.558*** -0.694*** -0.610*** -0.627*** -0.996*** -0.804*** -1.037*** -0.927*** 0.397*** 0.356*** (0.134) (0.0454) (0.130) (0.107) (0.161) (0.102) (0.176) (0.208) (0.110) (0.100) Final Oesch class position (8 classes; 1 = large business owners and managers)

Final Oesch class position - 8 classes = 2, Small business owners -0.358** -0.503*** -0.0425 0.00295 -0.401** -0.385** 0.797*** 0.618*** 0.598*** 0.596*** (0.149) (0.0774) (0.104) (0.0627) (0.166) (0.165) (0.165) (0.176) (0.0753) (0.0574) Final Oesch class position - 8 classes = 3, Technical (semi-) professionals -0.364*** -0.291 -0.404*** -0.148 -0.488*** -0.553*** 0.489** 0.561** 0.0489 0.234*** (0.104) (0.234) (0.114) (0.0925) (0.0803) (0.123) (0.204) (0.235) (0.0671) (0.0538) Final Oesch class position - 8 classes = 4, Production workers -0.946*** -1.083*** -0.566*** -0.361*** -1.127*** -1.103*** 0.811*** 0.736*** 0.972*** 0.994*** (0.129) (0.150) (0.0867) (0.0593) (0.118) (0.151) (0.229) (0.147) (0.114) (0.0657) Final Oesch class position - 8 classes = 5, (Associate) managers -0.279** -0.265 -0.139** -0.0312 -0.304*** -0.288*** 0.570*** 0.326 0.146* 0.0349 (0.141) (0.176) (0.0560) (0.0317) (0.0752) (0.0719) (0.179) (0.249) (0.0780) (0.0586) Final Oesch class position - 8 classes = 6, Clerks -0.534*** -0.503** -0.218*** -0.213*** -0.638*** -0.692*** 0.799*** 0.425** 0.426*** 0.407*** (0.0995) (0.244) (0.0606) (0.0586) (0.117) (0.120) (0.179) (0.209) (0.0646) (0.0616) Final Oesch class position - 8 classes = 7, Socio-cultural (semi-) professionals -0.182 -0.261 -0.395*** -0.248 -0.680*** -0.660*** 0.324 -0.0769 -0.0228 0.0291 (0.133) (0.289) (0.139) (0.208) (0.0792) (0.145) (0.279) (0.376) (0.0644) (0.0599) Final Oesch class position - 8 classes = 8, Service workers -0.889*** -0.894*** -0.447*** -0.407*** -0.942*** -0.928*** 0.761*** 0.688*** 0.798*** 0.828*** (0.127) (0.217) (0.0537) (0.0824) (0.101) (0.0990) (0.181) (0.168) (0.0965) (0.0878) Country (1=AT)

Country = 2, BE 4.326*** 1.454*** -0.414*** -0.287*** -3.184*** 1.713*** 0.720*** -1.172*** -0.463*** -0.686*** (0.0340) (0.126) (0.0197) (0.0791) (0.0427) (0.128) (0.0585) (0.0708) (0.00809) (0.0325) Country = 3, CH 3.228*** 0.681*** -1.724*** -1.046*** -3.489*** -0.638*** 1.271*** 0.304*** 1.132*** 0.648*** (0.0297) (0.126) (0.0260) (0.0815) (0.0480) (0.129) (0.0747) (0.0788) (0.0209) (0.0330) Country = 4, CZ 0.489*** 1.140*** -1.776*** -2.067*** 0.267*** 1.782*** -3.293*** -2.345*** 1.631*** 1.060*** (0.0425) (0.125) (0.0218) (0.0752) (0.0688) (0.130) (0.0701) (0.0713) (0.0145) (0.0333) Country = 5, DE 2.271*** -0.306** -0.0213 0.648*** -2.330*** -1.951*** 0.421*** -0.0370 (0.0296) (0.126) (0.0184) (0.0782) (0.0625) (0.0706) (0.0130) (0.0320) Country = 24, BG 4.050*** 0.254 -2.741*** -0.863*** 2.367*** -0.120 -1.133*** 1.097*** 0.399*** (0.100) (0.254) (0.0499) (0.0514) (0.0190) (0.145) (0.179) (0.0448) (0.0756) Country = 6, EE 3.325*** 1.581*** -5.439*** -4.806*** 0.673*** 1.196*** -5.775*** 1.278*** 0.531*** (0.0478) (0.129) (0.0329) (0.0719) (0.0153) (0.130) (0.0655) (0.0391) (0.0359) Country = 7, ES -1.075*** -0.0763 -3.733*** -4.431*** 0.575*** 2.248*** 0.353*** -0.0568 (0.0423) (0.124) (0.0329) (0.0815) (0.0509) (0.136) (0.0107) (0.0365) Country = 8, FI 2.375*** -0.138 -2.723*** -2.312*** 0.0219 1.690*** -1.418*** 0.230*** 0.504*** 0.0536 (0.0327) (0.134) (0.0226) (0.0759) (0.0595) (0.134) (0.0577) (0.0759) (0.00955) (0.0329) Country = 9, FR 0.363*** -1.673*** 0.439*** 1.922*** -0.466*** -0.429*** 0.648*** 0.607*** (0.0876) (0.130) (0.0632) (0.134) (0.173) (0.0753) (0.0254) (0.0298) Country = 10, GB 3.657*** 1.155*** -3.244*** 0.951*** -3.361*** -1.190*** 1.083*** 0.583*** (0.0306) (0.127) (0.0484) (0.136) (0.0596) (0.0743) (0.0113) (0.0304) Country = 11, HU 2.151*** -2.986*** 1.086*** 2.782*** -2.790*** -0.374*** 0.529*** 0.489*** (0.0528) (0.141) (0.0323) (0.128) (0.0606) (0.0731) (0.0219) (0.0379) Country = 12, IE 1.531*** -0.467*** 0.612*** 1.098*** 1.642*** 0.605*** 0.281*** (0.0330) (0.0249) (0.0792) (0.0498) (0.141) (0.00968) (0.0359) Country = 15, IT 3.171*** -0.436*** -3.217*** 0.679*** 0.740*** -0.440** -1.065*** -0.467*** 0.138*** (0.0464) (0.0727) (0.0299) (0.200) (0.258) (0.189) (0.0788) (0.0204) (0.0389) - Country = 17, NL 2.472*** 0.858*** 0.0508*** -0.353*** -0.188*** 1.759*** 0.484*** -0.189** 0.371*** 0.0798*** (0.0236) (0.130) (0.0171) (0.0743) (0.0448) (0.128) (0.0754) (0.0782) (0.0153) (0.0274) Country = 18, NO 2.351*** 0.0196 -1.591*** -1.634*** 0.0980*** 2.037*** 1.262*** 0.153* 0.208*** -0.405*** (0.0294) (0.122) (0.0185) (0.0659) (0.0361) (0.133) (0.0576) (0.0793) (0.0127) (0.0317) Country = 23, SI 4.142*** -0.169 -1.513*** -1.693*** -0.388*** -2.684*** 0.816*** 0.616*** (0.0360) (0.129) (0.0201) (0.0739) (0.0446) (0.0635) (0.0104) (0.0345) Country = 20, PT 4.310*** 1.367*** -5.690*** -2.182*** 1.015*** -4.911*** -3.979*** 0.833*** 0.511*** (0.0474) (0.127) (0.0688) (0.0720) (0.160) (0.0418) (0.0757) (0.0242) (0.0418) Country = 22, SE 3.382*** 0.565*** -2.097*** -1.009*** 0.340*** 2.262*** -0.763*** -0.106*** -1.049*** (0.0289) (0.130) (0.0192) (0.0699) (0.0480) (0.136) (0.0769) (0.0112) (0.0333) Country = 16, LT 1.609*** -2.115*** 1.214*** -1.720*** 1.188*** (0.139) (0.0704) (0.127) (0.0689) (0.0369) Country = 19, PL 1.586*** -0.763*** -3.076*** -1.512*** 0.222*** 2.061*** 1.497*** 1.125*** 1.137*** 0.678*** (0.0397) (0.125) (0.0278) (0.0880) (0.0600) (0.129) (0.0309) (0.0740) (0.0113) (0.0316) Country = 25, DK 4.802*** 2.078*** -3.245*** -3.403*** -0.739*** 0.411*** 0.599*** 0.199* -0.666*** -1.018*** (0.0320) (0.222) (0.0180) (0.170) (0.0530) (0.0207) (0.0598) (0.112) (0.00883) (0.0537) Country = 26, GR -3.641*** 0.594*** -1.239*** -1.196*** -0.391*** 0.0230 (0.279) (0.101) (0.0699) (0.214) (0.0111) (0.0608) Country = 27, LU 3.117*** -0.229*** -1.924*** 0.614*** (0.0733) (0.0322) (0.159) (0.0381) Country = 28, SK 1.615*** 0.271 -0.482*** 0.0599 -0.0919 -1.880*** 0.810*** 0.487*** (0.0541) (0.248) (0.0348) (0.165) (0.0883) (0.181) (0.0285) (0.0680) ESS round (1=2002)

ESS round = 2 -0.131*** -0.251*** 0.152** 0.0852 0.0994 (0.0340) (0.0400) (0.0698) (0.284) (0.0661) ESS round = 3 -0.165 -0.248*** 0.264 0.840* 0.107 (0.125) (0.0696) (0.215) (0.438) (0.0861) ESS round = 4 -0.324** -0.163* 0.469 0.524 0.0144 (0.164) (0.0873) (0.309) (0.436) (0.0775) ESS round (5=2010)

ESS round = 6 -0.109 0.00251 0.0171 -0.0513 0.0119 (0.182) (0.0706) (0.119) (0.244) (0.0800) ESS round = 7 -0.195 0.0734 -0.117 0.245 0.136*** (0.207) (0.0619) (0.0710) (0.313) (0.0461) ESS round = 8 -0.238 0.278 -0.0285 0.616*** -0.0521 (0.408) (0.289) (0.477) (0.192) (0.0796) Constant -4.460*** -1.723*** -0.117 -1.220*** -0.371 -2.039*** -3.052*** -2.064*** -2.788*** -2.288*** (0.128) (0.247) (0.0905) (0.178) (0.297) (0.252) (0.580) (0.497) (0.294) (0.211)

Observations 129,645 121,398 107,268 102,000 115,398 110,770 108,039 115,454 135,101 129,433 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.5. Marginal probabilities - voting choices of class groups (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) far left socdem greens agricultural s.issue VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

1, Large business owners and managers 0.0260*** 0.0251*** 0.188*** 0.139*** 0.0762*** 0.0850*** 0.0309*** 0.0175*** 0.0170*** 0.0260*** (0.00205) (0.00461) (0.0119) (0.0121) (0.0149) (0.00429) (0.00418) (0.00651) (0.00170) (0.00769) 2, Small business owners 0.0214*** 0.0233*** 0.167*** 0.125*** 0.0368*** 0.0504*** 0.0725*** 0.0452*** 0.00578*** 0.0264*** (0.00101) (0.00119) (0.00671) (0.00722) (0.00290) (0.00587) (0.00571) (0.00311) (0.000806) (0.00158) 3, Technical (semi-) professionals 0.0324*** 0.0371*** 0.295*** 0.210*** 0.0469*** 0.0551*** 0.0296*** 0.0226*** 0.0145*** 0.0231*** (0.00177) (0.00241) (0.0159) (0.00687) (0.00395) (0.00189) (0.00673) (0.00159) (0.00483) (0.00290) 4, Production workers 0.0336*** 0.0349*** 0.246*** 0.186*** 0.0183*** 0.0264*** 0.0354*** 0.0198*** 0.00620*** 0.0227*** (0.00205) (0.00236) (0.00951) (0.00756) (0.00313) (0.00175) (0.00182) (0.00114) (0.00137) (0.00103) 5, (Associate) managers 0.0301*** 0.0333*** 0.253*** 0.206*** 0.0398*** 0.0510*** 0.0271*** 0.0164*** 0.0110*** 0.0182*** (0.00257) (0.00129) (0.0121) (0.0105) (0.00140) (0.00219) (0.00183) (0.00209) (0.000605) (0.00134) 6, Clerks 0.0328*** 0.0298*** 0.237*** 0.213*** 0.0321*** 0.0426*** 0.0317*** 0.0135*** 0.00960*** 0.0265*** (0.00230) (0.00261) (0.00986) (0.0100) (0.00152) (0.00444) (0.00399) (0.00146) (0.000495) (0.00288) 7, Socio-cultural (semi-) professionals 0.0492*** 0.0521*** 0.282*** 0.225*** 0.0750*** 0.0847*** 0.0339*** 0.0174*** 0.0109*** 0.0324*** (0.00287) (0.00310) (0.0115) (0.0120) (0.00485) (0.00291) (0.00276) (0.00224) (0.000898) (0.00363) 8, Service workers 0.0316*** 0.0359*** 0.247*** 0.190*** 0.0275*** 0.0298*** 0.0354*** 0.0190*** 0.00651*** 0.0199*** (0.00186) (0.00104) (0.00711) (0.00556) (0.00247) (0.00150) (0.00349) (0.00159) (0.000130) (0.000809)

Observations 116,895 115,865 135,101 129,433 108,187 107,246 40,669 44,809 52,385 55,653 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.5. (cont.) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) liberals chrisdem conservative far right abstention VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

1, Large business owners and managers 0.0879*** 0.0801*** 0.157*** 0.148*** 0.236*** 0.273*** 0.0198*** 0.0378*** 0.120*** 0.130*** (0.00678) (0.0122) (0.00652) (0.00691) (0.0141) (0.0187) (0.00325) (0.00611) (0.00760) (0.00591) 2, Small business owners 0.0650*** 0.0514*** 0.153*** 0.149*** 0.179*** 0.208*** 0.0405*** 0.0649*** 0.194*** 0.210*** (0.00649) (0.00675) (0.00931) (0.00979) (0.0132) (0.0192) (0.00221) (0.00200) (0.00665) (0.00583) 3, Technical (semi-) professionals 0.0646*** 0.0622*** 0.117*** 0.133*** 0.168*** 0.183*** 0.0309*** 0.0619*** 0.125*** 0.158*** (0.00310) (0.00437) (0.00772) (0.00391) (0.00814) (0.00639) (0.00294) (0.00418) (0.00822) (0.00633) 4, Production workers 0.0385*** 0.0301*** 0.103*** 0.113*** 0.101*** 0.117*** 0.0410*** 0.0718*** 0.255*** 0.279*** (0.00200) (0.00204) (0.00288) (0.00126) (0.00332) (0.00620) (0.00384) (0.00599) (0.00817) (0.00645) 5, (Associate) managers 0.0695*** 0.0636*** 0.142*** 0.145*** 0.192*** 0.224*** 0.0332*** 0.0505*** 0.136*** 0.134*** (0.00535) (0.00380) (0.00265) (0.00602) (0.00664) (0.00869) (0.00259) (0.00483) (0.0123) (0.00755) 6, Clerks 0.0557*** 0.0514*** 0.134*** 0.127*** 0.150*** 0.164*** 0.0406*** 0.0550*** 0.170*** 0.181*** (0.00165) (0.00339) (0.00567) (0.00194) (0.00697) (0.00757) (0.00153) (0.00275) (0.00341) (0.00358) 7, Socio-cultural (semi-) professionals 0.0755*** 0.0638*** 0.118*** 0.124*** 0.146*** 0.169*** 0.0266*** 0.0352*** 0.118*** 0.134*** (0.00816) (0.00713) (0.0102) (0.0138) (0.00971) (0.0115) (0.00537) (0.00707) (0.00356) (0.00433) 8, Service workers 0.0406*** 0.0359*** 0.113*** 0.109*** 0.118*** 0.136*** 0.0393*** 0.0689*** 0.225*** 0.249*** (0.00180) (0.00156) (0.00231) (0.00249) (0.00336) (0.00324) (0.00185) (0.00445) (0.00589) (0.00753)

Observations 129,645 121,398 107,268 102,000 115,398 110,770 108,039 115,454 135,101 129,433 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.6. Logistic regressions - determinants of vote choice depending on parliamentary presence/not of far right party

(1) (2) (3) (4) (5) (6)

no far right far left social dem abstention VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

Gender (1=male; 2=female) -0.270*** -0.257*** 0.00428 -0.0689 0.227*** 0.224*** (0.0469) (0.0706) (0.0111) (0.0579) (0.0624) (0.0706) Age (five groups) 0.0598 0.0591 0.0885*** 0.180*** -0.177*** -0.186*** (0.0946) (0.116) (0.0157) (0.0202) (0.0157) (0.0132) Born in country (1=yes; 2=no) -1.194*** -0.910*** -1.000*** -0.740*** 0.504*** 0.301** (0.378) (0.335) (0.152) (0.0648) (0.0926) (0.127) Final Oesch class position (8 classes; 1 = large business owners and managers)

Final Oesch class position - 8 classes = 2, Small business owners -0.119 -0.230 -0.153 -0.112 0.721*** 0.638*** (0.0765) (0.157) (0.118) (0.115) (0.0661) (0.0818) Final Oesch class position - 8 classes = 3, Technical (semi-) professionals 0.406* 0.343 0.681*** 0.538*** -0.0150 0.216*** (0.239) (0.334) (0.0527) (0.0541) (0.0826) (0.0817)

Final Oesch class position - 8 classes = 4, Production workers 0.354** 0.239 0.328*** 0.368*** 1.116*** 1.084*** (0.138) (0.162) (0.101) (0.100) (0.154) (0.105)

Final Oesch class position - 8 classes = 5, (Associate) managers 0.455*** 0.122 0.406*** 0.484*** -0.0282 0.00982 (0.159) (0.194) (0.0308) (0.143) (0.0953) (0.103)

Final Oesch class position - 8 classes = 6, Clerks 0.253 0.117 0.284*** 0.673*** 0.516*** 0.529*** (0.222) (0.359) (0.0414) (0.0596) (0.0565) (0.0480) Final Oesch class position - 8 classes = 7, Socio-cultural (semi-) professionals 0.839*** 0.559* 0.488*** 0.459*** 0.0315 0.186*** (0.0973) (0.313) (0.0464) (0.145) (0.0570) (0.0228)

Final Oesch class position - 8 classes = 8, Service workers 0.335** 0.245 0.277** 0.374*** 0.937*** 0.955*** (0.138) (0.244) (0.119) (0.0747) (0.144) (0.0915) Country (1=AT)

Country = 2, BE

Country = 3, CH

Country = 4, CZ

Country = 5, DE -0.554*** -0.315*** 0.601*** 0.315*** -1.243*** -1.085*** (0.0218) (0.0202) (0.00857) (0.0413) (0.0123) (0.0124) Country = 24, BG

Country = 6, EE -0.549*** 0.575*** -0.426*** -0.525*** (0.0289) (0.0164) (0.0532) (0.0301) Country = 7, ES -0.487*** 0.236*** 0.718*** 0.346*** -1.320*** -1.105*** (0.00987) (0.0303) (0.0220) (0.00617) (0.0246) (0.0179) Country = 8, FI

Country = 9, FR

Country = 10, GB

Country = 11, HU -3.038*** 0.616*** -1.160*** (0.0410) (0.0199) (0.0391) Country = 12, IE -0.325*** 0.373*** -0.982*** -0.856*** -1.053*** -0.770*** (0.00981) (0.00778) (0.00606) (0.0197) (0.0134) (0.0222) Country = 15, IT

Country = 17, NL

Country = 18, NO

Country = 23, SI

Country = 20, PT 0.0209 0.204*** 0.465*** 0.243*** -0.866*** -0.568*** (0.0258) (0.0484) (0.0240) (0.0239) (0.0358) (0.0260) Country = 22, SE

Country = 16, LT -3.700*** -0.266*** 0.110*** (0.0359) (0.00502) (0.0200) Country = 19, PL

Country = 25, DK

Country = 26, GR 0.112** 0.520*** -2.194*** (0.0500) (0.0218) (0.0381) Country = 27, LU -2.089*** -0.416*** -1.022*** (0.107) (0.0571) (0.0619) Country = 28, SK

ESS round (1=2002)

ESS round = 2 0.183*** -0.0123 0.146 (0.0682) (0.157) (0.146) ESS round = 3 0.174*** -0.0687 0.247 (0.0597) (0.129) (0.164) ESS round = 4 0.222* -0.0884 0.0472 (0.120) (0.173) (0.155) ESS round (5=2010)

ESS round = 6 0.0670 -0.141 0.113 (0.212) (0.203) (0.122) ESS round = 7 0.201 -0.139 0.0986 (0.164) (0.230) (0.0878) ESS round = 8 0.577 -0.204 -0.106 (0.506) (0.300) (0.113) Constant -1.841*** -2.077*** -0.990*** -1.510*** -1.607*** -1.433*** (0.487) (0.252) (0.0318) (0.206) (0.221) (0.192)

Observations 44,668 42,125 49,231 48,877 49,231 48,877 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.6. (cont)

(7) (8) (9) (10) (11) (12) far right far left social dem abstention VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

Gender (1=male; 2=female) -0.0873 -0.236*** -0.0455 -0.0448 0.109** 0.135** (0.0761) (0.0869) (0.0435) (0.0290) (0.0441) (0.0615) Age (five groups) -0.0477 0.0511* 0.164*** 0.179*** -0.247*** -0.234*** (0.0340) (0.0293) (0.0133) (0.0219) (0.0533) (0.0467) Born in country (1=yes; 2=no) -0.679*** -0.645*** -0.141 -0.128 0.315* 0.386*** (0.224) (0.175) (0.100) (0.131) (0.164) (0.144) Final Oesch class position (8 classes; 1 = large business owners and managers)

Final Oesch class position - 8 classes = 2, Small business owners -0.306*** 0.163 -0.137 -0.126 0.508*** 0.575*** (0.0823) (0.210) (0.175) (0.173) (0.110) (0.0821) Final Oesch class position - 8 classes = 3, Technical (semi-) professionals 0.0232 0.547** 0.573*** 0.515*** 0.0615 0.234*** (0.105) (0.213) (0.0797) (0.183) (0.102) (0.0724)

Final Oesch class position - 8 classes = 4, Production workers 0.156 0.538** 0.405* 0.365* 0.858*** 0.951*** (0.123) (0.274) (0.240) (0.206) (0.131) (0.0783)

Final Oesch class position - 8 classes = 5, (Associate) managers -0.174 0.549*** 0.407*** 0.507** 0.186** 0.0363 (0.141) (0.179) (0.0990) (0.204) (0.0899) (0.0744)

Final Oesch class position - 8 classes = 6, Clerks 0.178 0.297 0.335*** 0.457** 0.361*** 0.354*** (0.151) (0.319) (0.127) (0.187) (0.0778) (0.0821) Final Oesch class position - 8 classes = 7, Socio-cultural (semi-) professionals 0.492*** 1.081*** 0.600*** 0.700*** -0.0697 -0.0464 (0.150) (0.141) (0.139) (0.223) (0.0893) (0.0746)

Final Oesch class position - 8 classes = 8, Service workers 0.0336 0.599*** 0.444** 0.406** 0.699*** 0.764*** (0.175) (0.197) (0.189) (0.205) (0.0956) (0.115) Country (1=AT)

Country = 2, BE -0.546*** 1.097*** -0.193*** -0.348*** -0.462*** -0.673*** (0.0625) (0.0671) (0.0403) (0.0488) (0.00862) (0.0442) Country = 3, CH 0.825*** -1.244*** -0.570*** -1.148*** 1.140*** 0.657*** (0.0582) (0.0632) (0.0360) (0.0439) (0.0307) (0.0473) Country = 4, CZ

Country = 5, DE

Country = 24, BG 0.409** -0.590*** 1.146*** 0.452*** (0.178) (0.0884) (0.0276) (0.0917) Country = 6, EE

Country = 7, ES

Country = 8, FI 3.549*** 2.852*** -0.182*** -0.819*** 0.505*** 0.0749* (0.0662) (0.0712) (0.0449) (0.0580) (0.0116) (0.0415) Country = 9, FR 3.665*** 1.859*** 0.237 -0.206*** 0.677*** 0.625*** (0.171) (0.0665) (0.149) (0.0505) (0.0123) (0.0383) Country = 10, GB 0.299*** -0.892*** 0.581*** 0.0553 1.087*** 0.603*** (0.0625) (0.0672) (0.0418) (0.0434) (0.0151) (0.0385) Country = 11, HU -0.942*** -0.846*** 0.517*** (0.0684) (0.0494) (0.0452) Country = 12, IE

Country = 15, IT 3.531*** 1.383*** -0.407*** -0.332*** -0.491*** 0.162*** (0.0598) (0.0880) (0.0590) (0.0341) (0.0101) (0.0502) Country = 17, NL 4.023*** 2.626*** -0.135*** -0.608*** 0.360*** 0.0987*** (0.0637) (0.0721) (0.0261) (0.0577) (0.0161) (0.0340) Country = 18, NO 4.466*** 2.223*** 0.197*** 0.215*** 0.204*** -0.393*** (0.0639) (0.0738) (0.0412) (0.0525) (0.0126) (0.0442) Country = 23, SI 0.638*** 0.339*** 0.721*** 0.823*** 0.639*** (0.0614) (0.0545) (0.0379) (0.0141) (0.0430) Country = 20, PT

Country = 22, SE 3.841*** 2.068*** 0.528*** 0.0576 -0.104*** -1.034*** (0.0594) (0.0698) (0.0394) (0.0497) (0.0127) (0.0449) Country = 16, LT

Country = 19, PL -1.344*** -0.364*** -1.872*** 1.139*** 0.694*** (0.0725) (0.0539) (0.0534) (0.0151) (0.0416) Country = 25, DK 2.617*** 1.662*** 0.204*** -0.283*** -0.667*** -0.992*** (0.0590) (0.0957) (0.0380) (0.0841) (0.00886) (0.0719) Country = 26, GR 4.522*** 2.729*** 0.685*** 0.211*** -0.169*** 0.0391 (0.225) (0.157) (0.163) (0.0591) (0.0523) (0.0786) Country = 27, LU

Country = 28, SK 1.767*** 0.229** 0.214** 0.830*** 0.534*** (0.110) (0.105) (0.0832) (0.0128) (0.0851) ESS round (1=2002)

ESS round = 2 0.183* -0.0446 0.0693* (0.0998) (0.0822) (0.0411) ESS round = 3 0.0614 -0.314 0.00889 (0.193) (0.244) (0.0337) ESS round = 4 0.123 -0.331 -0.0330 (0.279) (0.245) (0.0646) ESS round (5=2010)

ESS round = 6 0.362** 0.250*** -0.0359 (0.167) (0.0535) (0.101) ESS round = 7 0.150 -0.0953** 0.156*** (0.202) (0.0461) (0.0464) ESS round = 8 -0.0888 0.0303 -0.0253 (0.143) (0.0972) (0.0987) Constant -5.825*** -4.891*** -1.841*** -1.975*** -2.454*** -2.216*** (0.234) (0.365) (0.325) (0.319) (0.317) (0.286)

Observations 72,227 73,740 85,870 80,556 85,870 80,556 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.7. Marginal probabilities - determinants of vote choice depending on parliamentary presence/not of far right party

(1) (2) (3) (4) (5) (6) no far right far left social dem abstention VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

1, Large business owners and managers 0.0257*** 0.0449*** 0.220*** 0.154*** 0.0986*** 0.104*** (0.00343) (0.00955) (0.0117) (0.00251) (0.00748) (0.00392) 2, Small business owners 0.0229*** 0.0360*** 0.196*** 0.140*** 0.180*** 0.177*** (0.00178) (0.00290) (0.0107) (0.0133) (0.0165) (0.0152) 3, Technical (semi-) professionals 0.0380*** 0.0619*** 0.354*** 0.236*** 0.0973*** 0.125*** (0.00424) (0.00637) (0.0185) (0.00977) (0.0119) (0.00855) 4, Production workers 0.0361*** 0.0562*** 0.280*** 0.207*** 0.243*** 0.248*** (0.000904) (0.00491) (0.00709) (0.0157) (0.0124) (0.0131) 5, (Associate) managers 0.0398*** 0.0503*** 0.296*** 0.226*** 0.0961*** 0.105*** (0.00141) (0.00166) (0.0156) (0.0236) (0.0132) (0.0111) 6, Clerks 0.0328*** 0.0501*** 0.272*** 0.260*** 0.153*** 0.162*** (0.00276) (0.00675) (0.00660) (0.0116) (0.00520) (0.00798) 7, Socio-cultural (semi-) professionals 0.0571*** 0.0755*** 0.313*** 0.222*** 0.101*** 0.122*** (0.00332) (0.00710) (0.00700) (0.0239) (0.00616) (0.00376) 8, Service workers 0.0355*** 0.0565*** 0.270*** 0.208*** 0.213*** 0.225*** (0.000975) (0.00164) (0.00976) (0.0124) (0.0117) (0.00857)

Observations 44,668 42,125 49,231 48,877 49,231 48,877 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.7. (cont)

(7) (8) (9) (10) (11) (12) far right far left social dem abstention VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

1, Large business owners and managers 0.0274*** 0.0137*** 0.164*** 0.129*** 0.136*** 0.145*** (0.00240) (0.00258) (0.0198) (0.0204) (0.0108) (0.00890) 2, Small business owners 0.0205*** 0.0160*** 0.147*** 0.116*** 0.203*** 0.228*** (0.00112) (0.00104) (0.00896) (0.00883) (0.00432) (0.00368) 3, Technical (semi-) professionals 0.0280*** 0.0232*** 0.255*** 0.196*** 0.143*** 0.176*** (0.00103) (0.000943) (0.0164) (0.00895) (0.00497) (0.00653) 4, Production workers 0.0317*** 0.0230*** 0.225*** 0.174*** 0.261*** 0.296*** (0.00408) (0.00216) (0.0169) (0.00776) (0.00792) (0.00747) 5, (Associate) managers 0.0232*** 0.0232*** 0.225*** 0.194*** 0.159*** 0.150*** (0.00265) (0.00178) (0.0154) (0.00939) (0.0105) (0.00797) 6, Clerks 0.0324*** 0.0183*** 0.214*** 0.187*** 0.182*** 0.193*** (0.00240) (0.00259) (0.0140) (0.00385) (0.00510) (0.00401) 7, Socio-cultural (semi-) professionals 0.0433*** 0.0382*** 0.259*** 0.225*** 0.129*** 0.140*** (0.00353) (0.00268) (0.0178) (0.0123) (0.00389) (0.00652) 8, Service workers 0.0283*** 0.0244*** 0.232*** 0.180*** 0.233*** 0.261*** (0.00312) (0.00132) (0.00717) (0.00661) (0.00283) (0.00967)

Observations 72,227 73,740 85,870 80,556 85,870 80,556 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.8. Logistic regressions - voting for the social democrats, for the far right, abstaining

(1) (2) (3) (4) (5) (6) abstention social dem far right VARIABLES 2002-2008 2010-2016 2002-2008 2010-2016 2002-2008 2010-2016

Gender (1=male; 2=female) 0.0195 0.0510 -0.0859** -0.0216 -0.345*** -0.273*** (0.0186) (0.0449) (0.0428) (0.0521) (0.0713) (0.0867) Education (educ; 1 = no formal education) educ = 2 (low) -0.0198 -0.0498 -0.121** -0.00671 0.103 0.316*** (0.0902) (0.134) (0.0541) (0.0803) (0.203) (0.104) educ = 3 (med-low) -0.303*** -0.217** -0.0106 0.0814 0.209 0.545*** (0.0894) (0.0923) (0.0447) (0.175) (0.151) (0.139) educ = 4 (med-high) -0.459** -0.516*** 0.00412 0.0172 0.172 0.499*** (0.194) (0.0764) (0.159) (0.266) (0.173) (0.0976) educ = 5 (high) -0.435*** -0.527*** 0.0641 -0.198 0.0262 0.313 (0.0405) (0.121) (0.0864) (0.295) (0.260) (0.284) Age (five groups) -0.300*** -0.243*** 0.275*** 0.258*** 0.112* 0.112** (0.0637) (0.0408) (0.0363) (0.0298) (0.0604) (0.0477) Born in country (1=yes; 2=no) 0.233 0.373* -0.165 -0.233 -1.244*** -0.791*** (0.175) (0.208) (0.164) (0.164) (0.211) (0.162) Trust in political institutions (0=no trust; 1=full trust) -1.632*** -1.194*** 0.944** 0.489 -0.235 -0.385 (0.121) (0.172) (0.367) (0.301) (0.363) (0.432) Views on immigration (0=strongly for; 1=strongly against) 0.516*** 0.305 -0.441** -1.086*** 1.400* 2.605*** (0.166) (0.200) (0.195) (0.132) (0.769) (0.642) Difficulties on present income (0=none at all; 1=very difficult) 0.645*** 0.710*** 0.0709 -0.0784 -0.179 -0.131 (0.122) (0.109) (0.103) (0.230) (0.191) (0.142) Preferences on government-led redistribution (0=strongly for; 1=strongly against) 0.235*** 0.418*** -0.408*** -0.239*** -0.287*** -0.801*** (0.0807) (0.0601) (0.0719) (0.0910) (0.103) (0.124) Country (1=AT)

Country = 2, BE -0.616*** -0.999*** -0.225*** -0.0730** 0.939*** -0.742*** (0.0252) (0.0515) (0.0539) (0.0322) (0.0595) (0.126) Country = 3, CH 1.461*** 0.635*** -1.376*** -1.684*** 1.298*** 0.499** (0.0726) (0.0684) (0.0766) (0.0686) (0.122) (0.206) Country = 8, FI 0.722*** 0.0959*** -0.262*** -0.789*** -1.000*** 0.576*** (0.0455) (0.0211) (0.0600) (0.0492) (0.130) (0.177) Country = 9, FR 0.454*** 0.277*** -0.178 -0.413*** -0.382*** -0.188** (0.0389) (0.0457) (0.163) (0.0340) (0.123) (0.0941) Country = 10, GB 0.995*** 0.467*** 0.492*** 0.00540 -3.317*** -1.060*** (0.0340) (0.0569) (0.0478) (0.0291) (0.0740) (0.104) Country = 11, HU 0.0711*** -0.970*** -0.631*** (0.0204) (0.0206) (0.110) Country = 15, IT -0.593*** -0.329*** -0.654*** -0.469*** 0.197 -1.368*** (0.0667) (0.0678) (0.129) (0.119) (0.127) (0.0674) Country = 17, NL 0.555*** 0.163*** -0.253*** -0.891*** 0.636*** 0.484*** (0.0197) (0.0459) (0.0604) (0.0495) (0.0638) (0.187) Country = 18, NO 0.503*** -0.161*** -0.0180 0.0494 1.542*** 0.815*** (0.0282) (0.0582) (0.0409) (0.0952) (0.0884) (0.225) Country = 19, PL 0.989*** 0.304*** -0.747*** -2.203*** 1.469*** 1.227*** (0.0477) (0.0598) (0.0668) (0.0728) (0.121) (0.152) Country = 22, SE 0.0121 -0.952*** 0.619*** 0.192** 0.312 (0.0470) (0.0374) (0.0568) (0.0747) (0.239) Country = 23, SI 0.788*** 0.371*** -0.0470 0.494*** -0.381*** -2.681*** (0.0343) (0.0555) (0.0376) (0.0334) (0.0262) (0.0879) Country = 24, BG 0.591*** -0.275*** 0.198 -0.868*** -0.0359 -1.039*** (0.0960) (0.0827) (0.186) (0.0784) (0.230) (0.259) Country = 25, DK -0.215*** -0.979*** 0.119* -0.371*** 0.951*** 0.658*** (0.0169) (0.0701) (0.0697) (0.112) (0.102) (0.193) Country = 26, GR -0.659*** -0.577*** 0.404*** 0.145 -0.965*** -1.571*** (0.0525) (0.0608) (0.134) (0.142) (0.0601) (0.271) Country = 28, SK 0.665*** 0.128*** 0.131 0.294*** -0.229 -2.046*** (0.0609) (0.0490) (0.0885) (0.0327) (0.157) (0.257) ESS round (5=2010)

ESS round = 6 (2012) 0.0860 0.319*** 0.0721 (0.0970) (0.0861) (0.297) ESS round = 7 (2014) 0.143*** -0.00424 0.297 (0.0447) (0.0560) (0.379) ESS round = 8 (2016) 0.0553 -0.0242 0.515* (0.0673) (0.102) (0.278) ESS round (1=2002)

ESS round = 2 (2004) 0.00273 -0.0715 -0.0516 (0.0743) (0.128) (0.294) ESS round = 3 (2006) 0.0208 -0.380 0.848* (0.0372) (0.238) (0.438) ESS round = 4 (2008) 0.0549 -0.355 0.394 (0.0436) (0.216) (0.378) Constant -1.156*** -1.009*** -1.397*** -1.183*** -2.596*** -3.084*** (0.198) (0.184) (0.441) (0.163) (0.476) (0.412)

Observations 29,181 28,872 29,181 28,872 26,954 28,872 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.9. Marginal probabilities - voting for the social democrats, for the far right, abstaining; depending on trust on institutions (mistrust), pro-immigration sentiment (migration) and difficulties on present income (income), all of them from lowest (0) to highest (1)

abstention social dem

2002-2008 2010-2016 2002-2008 2010-2016

Mistrust Migration Income Mistrust Migration Income Mistrust Migration Income Mistrust Migration Income

0 0.361*** 0.214*** 0.217*** 0.355*** 0.251*** 0.230*** 0.184*** 0.277*** 0.233*** 0.154*** 0.265*** 0.180*** (0.00746) (0.0125) (0.00781) (0.0119) (0.0182) (0.00732) (0.0192) (0.0185) (0.00578) (0.0131) (0.0119) (0.0110)

0.1 0.328*** 0.222*** 0.227*** 0.330*** 0.256*** 0.242*** 0.197*** 0.269*** 0.235*** 0.161*** 0.246*** 0.179*** (0.00493) (0.0103) (0.00598) (0.00779) (0.0150) (0.00565) (0.0149) (0.0147) (0.00413) (0.00972) (0.00919) (0.00785)

0.2 0.296*** 0.230*** 0.237*** 0.306*** 0.262*** 0.255*** 0.211*** 0.262*** 0.236*** 0.167*** 0.228*** 0.178*** (0.00256) (0.00803) (0.00402) (0.00390) (0.0116) (0.00386) (0.0100) (0.0111) (0.00247) (0.00612) (0.00668) (0.00472)

0.3 0.266*** 0.238*** 0.248*** 0.283*** 0.267*** 0.268*** 0.226*** 0.254*** 0.237*** 0.173*** 0.211*** 0.177*** (0.000427) (0.00562) (0.00194) (0.000300) (0.00816) (0.00194) (0.00464) (0.00752) (0.000802) (0.00231) (0.00437) (0.00162)

0.4 0.238*** 0.247*** 0.259*** 0.261*** 0.273*** 0.281*** 0.242*** 0.247*** 0.238*** 0.180*** 0.195*** 0.176*** (0.00152) (0.00310) (0.000268) (0.00300) (0.00462) (0.000152) (0.00128) (0.00409) (0.000878) (0.00173) (0.00227) (0.00146)

0.5 0.212*** 0.256*** 0.270*** 0.240*** 0.278*** 0.294*** 0.258*** 0.239*** 0.239*** 0.186*** 0.180*** 0.175*** (0.00315) (0.000469) (0.00258) (0.00595) (0.001000) (0.00226) (0.00770) (0.000784) (0.00257) (0.00598) (0.000501) (0.00451)

0.6 0.187*** 0.265*** 0.282*** 0.220*** 0.284*** 0.308*** 0.275*** 0.232*** 0.240*** 0.193*** 0.166*** 0.174*** (0.00450) (0.00227) (0.00501) (0.00854) (0.00271) (0.00451) (0.0146) (0.00241) (0.00427) (0.0104) (0.00142) (0.00754)

0.7 0.165*** 0.274*** 0.294*** 0.201*** 0.290*** 0.322*** 0.292*** 0.225*** 0.242*** 0.201*** 0.152*** 0.173*** (0.00558) (0.00511) (0.00754) (0.0108) (0.00651) (0.00687) (0.0219) (0.00547) (0.00598) (0.0151) (0.00288) (0.0105)

0.8 0.145*** 0.283*** 0.306*** 0.183*** 0.296*** 0.337*** 0.310*** 0.218*** 0.243*** 0.208*** 0.139*** 0.172*** (0.00638) (0.00805) (0.0102) (0.0126) (0.0104) (0.00931) (0.0297) (0.00840) (0.00770) (0.0200) (0.00416) (0.0135)

0.9 0.127*** 0.292*** 0.318*** 0.167*** 0.301*** 0.352*** 0.329*** 0.211*** 0.244*** 0.215*** 0.128*** 0.171*** (0.00695) (0.0111) (0.0129) (0.0141) (0.0143) (0.0118) (0.0378) (0.0112) (0.00943) (0.0252) (0.00525) (0.0165)

1 0.111*** 0.302*** 0.330*** 0.152*** 0.307*** 0.367*** 0.348*** 0.205*** 0.245*** 0.223*** 0.117*** 0.170*** (0.00729) (0.0142) (0.0157) (0.0153) (0.0184) (0.0144) (0.0462) (0.0139) (0.0112) (0.0305) (0.00616) (0.0194)

Observations 29,181 29,181 29,181 28,872 28,872 28,872 29,181 29,181 29,181 28,872 28,872 28,872 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table A.9. (cont)

far right

2002-2008 2010-2016

Mistrust Migration Income Mistrust Migration Income

0 0.0691*** 0.0351*** 0.0689*** 0.102*** 0.0260*** 0.0962*** (0.00668) (0.0128) (0.00446) (0.0104) (0.00897) (0.00408)

0.1 0.0677*** 0.0398*** 0.0679*** 0.0986*** 0.0332*** 0.0952*** (0.00448) (0.0116) (0.00331) (0.00677) (0.00932) (0.00298)

0.2 0.0664*** 0.0452*** 0.0669*** 0.0956*** 0.0422*** 0.0941*** (0.00236) (0.00982) (0.00218) (0.00329) (0.00920) (0.00189)

0.3 0.0651*** 0.0511*** 0.0659*** 0.0926*** 0.0534*** 0.0931*** (0.000300) (0.00749) (0.00109) (8.02e-05) (0.00838) (0.000821)

0.4 0.0638*** 0.0577*** 0.0649*** 0.0897*** 0.0672*** 0.0922*** (0.00169) (0.00446) (3.30e-05) (0.00316) (0.00662) (0.000231)

0.5 0.0625*** 0.0650*** 0.0639*** 0.0869*** 0.0839*** 0.0912*** (0.00362) (0.000679) (0.00102) (0.00614) (0.00363) (0.00126)

0.6 0.0613*** 0.0731*** 0.0629*** 0.0842*** 0.104*** 0.0902*** (0.00547) (0.00397) (0.00203) (0.00896) (0.000950) (0.00228)

0.7 0.0601*** 0.0821*** 0.0620*** 0.0815*** 0.128*** 0.0892*** (0.00727) (0.00952) (0.00302) (0.0116) (0.00709) (0.00327)

0.8 0.0589*** 0.0918*** 0.0610*** 0.0790*** 0.156*** 0.0883*** (0.00899) (0.0161) (0.00398) (0.0141) (0.0152) (0.00425)

0.9 0.0577*** 0.102*** 0.0601*** 0.0765*** 0.188*** 0.0873*** (0.0107) (0.0236) (0.00491) (0.0165) (0.0252) (0.00521)

1 0.0565*** 0.114*** 0.0592*** 0.0740*** 0.224*** 0.0864*** (0.0123) (0.0322) (0.00582) (0.0187) (0.0370) (0.00615)

Observations 26,954 26,954 26,954 28,872 28,872 28,872 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

B. Dualism, Welfare State and the Populist Radical Right

Table B.1. Key descriptive statistics Variables N Mean Min Max Categories / Question

Age 140789 2.898 0 5 0 = 18-24; 1 = 25-34; 2 = 35-44; 3 = 45-54

4 = 55-65; 5 = 65+ Gender 143174 1.577 1 2 1 = Male; 2= Female Education 142517 3.267 1 5 1 = ISCED 0-1; 2 = ISCED 2; 3 = ISCED 3; 4 = ISCED 4; 5 = ISCED 5-6 Outsider 117235 0.539 0 2 0 = Insider; 1 = Outsider; 2 = Upscale Social Class 139767 5.260 1 8 Based on Oesch (2006) social Class scheme. Political Trust 142444 0.455 0 1 Normalised factor loadings of trust towards 1) Politicians; 2) Parliament; 3) Political Parties. 0 = No trust; 1 = Complete trust. Anti- 133663 0.448 0 1 Based on Polavieja (2016). Including: immigration - Allow many/few immigrants of same race/ethnic group as majority - Allow many/few immigrants of different race/ethnic group from majority - Allow many/few immigrants from poorer countries outside Europe - Immigration bad or good for Country's economy - Country's cultural life undermined or enriched by immigrants - Immigrants make Country worse or better place to live 0 = Pro-immigration; 1 = Anti-immigration Redistribution 141383 0.333 0 1 “Government should reduce income differences” Likert Scale: 0 = Agree str.; 1 = Disagree str.

Table B.2. PRR parties by country included in dataset

Austria FPO 567 BZO 42 TeamStronach 23 Belgium VB 537 FN 57 Other 16 Switzerland SVP/UDC 2008 LdT 13 Germany Republikaner 58 NPD 27 AfD 149 Denmark DF 633 Finland PS 722 France FN 596 MPF 73 MNR 32 United Kingdom UKIP 95 BNP 8 Netherlands Pim Fortunyn 385 PVV 828 Norway FrP 1350 Sweden SD 225

Note: Criteria for selection of PRR parties based on van Kessel (2015) and Mudde (2007).

Table B.3. Multinomial logistic regression: determinants of vote choice

(1) (2) (3) (4) (5) (6) (7) VARIABLES other mainstream left mainstream right radical right radical left green abstention

Age group (base category = 18-24)

Age group (25-34) -0.0389 0.112 -0.0357 -0.139 -0.346 -0.0926 (0.273) (0.163) (0.0850) (0.0959) (0.249) (0.0687) Age group (35-44) -0.180 0.119* -0.0667 -0.241*** -0.357*** -0.646*** (0.248) (0.0664) (0.121) (0.0806) (0.128) (0.115) Age group (45-54) -0.146 0.195*** -0.123 0.0455 -0.410*** -1.031*** (0.299) (0.0437) (0.0840) (0.0834) (0.0960) (0.133) Age group (55-64) -0.294 0.0596 -0.423*** -0.129 -0.860*** -1.581*** (0.253) (0.0801) (0.119) (0.135) (0.130) (0.239) Age group (65+) -0.575** -0.233*** -0.753*** -0.898*** -1.796*** -2.234*** (0.228) (0.0669) (0.118) (0.175) (0.109) (0.318) Education (base category = primary)

Education (lower-secondary) -0.372** -0.254** 0.0494 0.109 -0.208 -0.234* (0.159) (0.116) (0.0886) (0.202) (0.279) (0.131) Education (upper-secondary) -0.451** -0.471*** -0.139** -0.0641 0.0405 -0.694*** (0.202) (0.0859) (0.0569) (0.159) (0.236) (0.0940) Education (post-secondary) -0.664*** -0.504*** -0.565*** -0.0973 0.439** -1.081*** (0.160) (0.109) (0.187) (0.254) (0.212) (0.110) Education (tertiary) -0.563*** -0.543*** -1.003*** 0.119 0.576*** -1.258*** (0.172) (0.111) (0.193) (0.260) (0.206) (0.123) Gender (base = male) -0.126 0.0521 -0.228*** -0.134 0.264*** 0.0911 (0.0971) (0.0331) (0.0462) (0.0955) (0.0589) (0.0804) Class (base category = production worker)

Class (self-employed) -0.464*** -0.917*** -1.219*** -0.887*** 0.270** -1.149*** (0.0869) (0.129) (0.213) (0.0684) (0.129) (0.0750) Class (small business owner) -0.137 -0.993*** -0.830*** -1.037*** -0.0443 -0.811*** (0.240) (0.173) (0.133) (0.0483) (0.137) (0.0490) Class (technical pro) -0.284* -0.230 -0.579*** -0.545*** 0.238 -0.920*** (0.173) (0.173) (0.144) (0.123) (0.161) (0.0451) Class (manager) -0.379*** -0.452*** -0.913*** -0.883*** -0.0222 -1.051*** (0.101) (0.143) (0.145) (0.0509) (0.135) (0.0832) Class (clerk) -0.194*** -0.357*** -0.755*** -0.496*** 0.0152 -0.765*** (0.0597) (0.107) (0.131) (0.0677) (0.0493) (0.0716) Class (socio-cultural) 0.0782 -0.0112 -0.967*** 0.0265 0.699*** -0.758*** (0.128) (0.105) (0.129) (0.0593) (0.148) (0.0750) Class (service) -0.00841 -0.0418 -0.282* -0.128** 0.228*** -0.173*** (0.0477) (0.0442) (0.148) (0.0550) (0.0686) (0.0158) Status (base = outsider) -0.0914 0.0300 0.0415 -0.198*** -0.206*** -0.247*** (0.110) (0.0454) (0.0391) (0.0745) (0.0624) (0.0277) Country (base category = Austria)

Country (Belgium) 21.70*** -0.290*** -0.120** 18.01*** -0.579*** -1.375*** (1.049) (0.0159) (0.0479) (1.046) (0.0236) (0.0320) Country (Switzerland) 20.48*** 0.138*** 2.105*** -0.688 0.201*** 1.254*** (1.048) (0.0280) (0.0570) (1.484) (0.0486) (0.0455) Country (Germany) 18.36*** -0.0266* -1.735*** 20.50*** -0.0995*** -0.370*** (1.048) (0.0148) (0.0251) (1.047) (0.0214) (0.0331) Country (Denmark) 22.02*** -0.0283 0.440*** 21.52*** -21.91*** -1.294*** (1.049) (0.0231) (0.115) (1.046) (1.047) (0.0544) Country (Finland) 23.05*** -0.159*** 0.347*** 20.91*** 0.00836 0.145*** (1.048) (0.0156) (0.0430) (1.046) (0.0307) (0.0460) Country (France) 0.398 0.431*** 0.355*** 20.41*** -0.634*** 0.519*** (1.484) (0.0324) (0.0380) (1.045) (0.0300) (0.0335) Country (UK) 20.10*** 0.0132 -2.009*** -0.181 -2.607*** -0.0115 (1.048) (0.0202) (0.0494) (1.480) (0.0460) (0.0314) Country (Netherlands) 20.81*** -0.497*** 0.470*** 20.96*** -1.044*** -0.652*** (1.048) (0.0273) (0.0709) (1.047) (0.0373) (0.0323) Country (Norway) 21.33*** 0.165*** 0.872*** 21.16*** -3.349*** -0.517*** (1.048) (0.0248) (0.0507) (1.047) (0.0361) (0.0470) Country (Sweden) 21.13*** -0.0196 -1.248*** 20.68*** -0.834*** -1.320*** (1.048) (0.0140) (0.0381) (1.046) (0.0251) (0.0363) Year (base category = 2002)

Year (2004) -0.0145 0.0898*** 0.0205 0.120** 0.0459 0.319*** (0.0818) (0.0251) (0.450) (0.0589) (0.0520) (0.113) Year (2006) -0.377 -0.0668 0.320 0.204 -0.0361 0.237 (0.429) (0.0745) (0.386) (0.174) (0.110) (0.174) Year (2008) -0.596 -0.144* -0.258 0.434*** -0.0761 0.203 (0.429) (0.0844) (0.313) (0.152) (0.120) (0.137) Year (2010) -0.224 -0.420*** 0.0402 0.247 0.0717 0.208 (0.537) (0.123) (0.107) (0.228) (0.171) (0.190) Year (2012) -0.285 -0.239** 0.323 0.307* -0.00723 0.166 (0.481) (0.119) (0.415) (0.166) (0.0835) (0.123) Year (2014) 0.456 -0.277*** 0.965*** 0.399** 0.0860 0.524*** (0.473) (0.0852) (0.252) (0.191) (0.197) (0.139) Year (2016) 0.542 -0.145 1.467*** 0.468** 0.140 0.398* (0.348) (0.0971) (0.385) (0.207) (0.129) (0.203) Constant -22.03*** 0.467*** -0.914*** -21.97*** -0.772*** 1.870*** (1.072) (0.108) (0.238) (1.074) (0.241) (0.273)

Observations 88,519 88,519 88,519 88,519 88,519 88,519 88,519 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.3. (cont.)

(8) (9) (10) (11) (12) (13) (14) VARIABLES other mains_left mains_right rad_right rad_left green abs

Age group (base category = 18-24)

Age group (25-34) -0.0516 0.110 -0.175 -0.227** -0.320 -0.213** (0.257) (0.178) (0.121) (0.104) (0.239) (0.0848) Age group (35-44) -0.193 0.134 -0.213 -0.317*** -0.292*** -0.773*** (0.235) (0.0879) (0.165) (0.0801) (0.0963) (0.0922) Age group (45-54) -0.185 0.208*** -0.329*** -0.0664 -0.385*** -1.187*** (0.277) (0.0677) (0.0990) (0.104) (0.0764) (0.109) Age group (55-64) -0.323 0.0668 -0.625*** -0.270* -0.785*** -1.737*** (0.248) (0.0435) (0.123) (0.145) (0.0821) (0.225) Age group (65+) -0.588** -0.193*** -0.945*** -0.955*** -1.623*** -2.350*** (0.232) (0.0689) (0.129) (0.205) (0.0642) (0.310) Education (base category = primary)

Education (lower-secondary) -0.334** -0.277** 0.205*** 0.0167 -0.118 -0.204 (0.142) (0.111) (0.0697) (0.228) (0.202) (0.140) Education (upper-secondary) -0.393** -0.505*** 0.0846 -0.122 0.0575 -0.656*** (0.187) (0.0818) (0.0693) (0.190) (0.180) (0.101) Education (post-secondary) -0.606*** -0.566*** -0.188 -0.155 0.349** -0.981*** (0.151) (0.109) (0.146) (0.314) (0.161) (0.110) Education (tertiary) -0.444** -0.633*** -0.411** 0.132 0.414** -1.067*** (0.183) (0.104) (0.161) (0.351) (0.170) (0.121) Gender (base = male) -0.164 0.0261 -0.340*** -0.184* 0.242*** 0.00884 (0.103) (0.0273) (0.0709) (0.0996) (0.0411) (0.0703) Class (base category = production worker)

Class (self-employed) -0.333*** -0.860*** -0.856*** -0.778*** 0.132 -0.905*** (0.0720) (0.143) (0.126) (0.0820) (0.113) (0.0408) Class (small business owner) -0.0714 -0.951*** -0.596*** -0.904*** -0.120 -0.684*** (0.242) (0.187) (0.122) (0.0620) (0.117) (0.0819) Class (technical pro) -0.231 -0.194 -0.292* -0.367*** 0.151 -0.744*** (0.149) (0.158) (0.165) (0.137) (0.125) (0.0250) Class (manager) -0.294*** -0.418*** -0.570*** -0.745*** -0.126 -0.863*** (0.0805) (0.138) (0.127) (0.0716) (0.128) (0.0663) Class (clerk) -0.146** -0.387*** -0.448*** -0.442*** -0.0931* -0.640*** (0.0673) (0.110) (0.164) (0.0893) (0.0523) (0.0551) Class (service) 0.0882 -0.117 -0.616*** -0.0482 0.436*** -0.613*** (0.125) (0.0899) (0.129) (0.0792) (0.139) (0.0526) Class (service) 0.00232 -0.0904* -0.151 -0.132*** 0.125* -0.114*** (0.0569) (0.0538) (0.187) (0.0502) (0.0699) (0.0280) Status (base = outsider) -0.0598 0.0596 0.105** -0.126* -0.131** -0.220*** (0.112) (0.0470) (0.0490) (0.0695) (0.0650) (0.0223) Country (base category = Austria)

Country (Belgium) 21.73*** -0.293*** 0.0659 17.90*** -0.540*** -1.320*** (1.050) (0.0218) (0.0701) (1.047) (0.0210) (0.0263) Country (Switzerland) 20.58*** 0.0906*** 2.793*** -0.259 0.113* 1.572*** (1.051) (0.0277) (0.127) (1.486) (0.0645) (0.0479) Country (Germany) 18.37*** -0.0208 -1.480*** 20.31*** -0.210*** -0.386*** (1.049) (0.0352) (0.0530) (1.046) (0.0257) (0.0256) Country (Denmark) 22.28*** 0.253*** 1.151*** 22.21*** -21.46*** -0.772*** (1.056) (0.0395) (0.176) (1.052) (1.053) (0.0780) Country (Finland) 23.09*** -0.245*** 0.792*** 20.85*** -0.00133 0.379*** (1.052) (0.0227) (0.0832) (1.050) (0.0355) (0.0548) Country (France) 0.400 0.356*** 0.355*** 19.98*** -0.745*** 0.410*** (1.488) (0.0381) (0.0608) (1.045) (0.0539) (0.0426) Country (UK) 20.11*** 0.149*** -1.933*** 0.0380 -2.510*** -0.00105 (1.048) (0.0306) (0.0711) (1.478) (0.0530) (0.0298) Country (Netherlands) 20.95*** -0.351*** 1.031*** 21.29*** -0.828*** -0.287*** (1.052) (0.0257) (0.119) (1.050) (0.0613) (0.0318) Country (Norway) 21.43*** 0.185*** 1.472*** 21.31*** -3.370*** -0.205*** (1.050) (0.0310) (0.113) (1.049) (0.0633) (0.0620) Country (Sweden) 21.16*** -0.243*** -0.475*** 20.47*** -1.288*** -1.158*** (1.051) (0.0459) (0.0808) (1.054) (0.0528) (0.0414) Year (base category = 2002)

Year (2004) -0.0621 0.0871*** -0.134 -0.0237 0.0774 0.207** (0.0918) (0.0277) (0.451) (0.0374) (0.0944) (0.0980) Year (2006) -0.417 -0.0981 0.217 0.0302 -0.0590 0.126 (0.417) (0.0730) (0.372) (0.158) (0.0878) (0.165) Year (2008) -0.682 -0.252** -0.305 0.136 -0.290* 0.0824 (0.414) (0.110) (0.279) (0.166) (0.152) (0.132) Year (2010) -0.296 -0.543*** -0.0232 -0.0952 -0.112 0.0680 (0.543) (0.128) (0.0941) (0.195) (0.201) (0.200) Year (2012) -0.372 -0.381** 0.292 -0.0229 -0.278* 0.0665 (0.479) (0.165) (0.396) (0.219) (0.151) (0.141) Year (2014) 0.356 -0.431*** 0.958*** 0.0405 -0.209 0.426*** (0.468) (0.118) (0.273) (0.191) (0.277) (0.157) Year (2016) 0.462 -0.360*** 1.548*** 0.0951 -0.194 0.332 (0.347) (0.122) (0.396) (0.181) (0.148) (0.204) Anti-immigration -0.117 -1.688*** 3.236*** -2.576*** -3.982*** -0.0604 (0.414) (0.302) (0.167) (0.362) (0.170) (0.259) Anti-redistribution -0.937*** -1.782*** -0.674*** -3.239*** -1.809*** -1.026*** (0.303) (0.175) (0.173) (0.247) (0.178) (0.0647) Trust pol institutions -0.624 -0.207 -1.591*** -2.396*** -1.028*** -2.188*** (0.540) (0.201) (0.389) (0.466) (0.139) (0.328) Constant -21.42*** 2.023*** -2.293*** -18.50*** 2.118*** 3.124*** (1.212) (0.201) (0.283) (1.091) (0.108) (0.205)

Observations 83,352 83,352 83,352 83,352 83,352 83,352 83,352 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.4. Marginal effects of insiders vs outsiders by party family

(1) (2) VARIABLES Status (insider) Status (insider)

Mainstream left 0.0254*** 0.0257*** (0.00764) (0.00758) Mainstream right 0.0219*** 0.0130*** (0.00437) (0.00446) Populist radical right 0.00479*** 0.00607*** (0.00175) (0.00188) Radical left -0.00453 -0.00277 (0.00316) (0.00287) Greens -0.00972*** -0.00602* (0.00371) (0.00342) Abstention -0.0371*** -0.0355*** (0.00582) (0.00517) Observations 88,519 83,352 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.5. Logistic regression - determinants of PRR vote (incl. immigration*unemployment rate/position interaction)

(1) VARIABLES PRR

Age group (base category = 18-24)

Age group (25-34) 0.246** (0.107) Age group (35-44) 0.382*** (0.113) Age group (45-54) 0.434*** (0.134) Age group (55-64) 0.294** (0.132) Age group (65+) 0.211 (0.147) Gender (base = male) -0.435*** (0.0654) imm*unemp (immigration rate*unemployment rate) 0.00302*** (0.000878) Status (base = outsider) 0.110 (0.0846) outsider#imm*unemp -0.00278*** (0.000898) Anti-immigration for cultural reasons -0.215*** (0.0246) Anti-immigration for economic reasons -0.145*** (0.0124) Important to get respect from others -0.0127 (0.0248) Country (base category = Austria)

Country (Belgium) 0.238*** (0.0619) Country (Switzerland) 1.986*** (0.0503) Country (Germany) -1.335*** (0.0505) Country (Denmark) 0.856*** (0.117) Country (Finland) 0.840*** (0.0580) Country (France) 0.0501 (0.0803) Country (UK) -1.803*** (0.0610) Country (Netherlands) 1.237*** (0.0500) Country (Norway) 1.440*** (0.0478) Country (Sweden) -0.262*** (0.0351) Year (base category = 2002)

Year (2004) -0.217 (0.461) Year (2006) 0.103 (0.358) Year (2008) -0.187 (0.328) Year (2010) 0.162 (0.176) Year (2012) 0.462 (0.404) Year (2014) 0.787*** (0.290) Year (2016) 1.456*** (0.484) Constant -2.119*** (0.279) Observations 102,265 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.6. Marginal effects of insiders vs outsiders of PRR voters by immigration*unemployment rate/position interaction

Insider Outsider imm*unemp = 0 0.0324*** 0.0324*** -0.00194 -0.0029 imm*unemp = 25 0.0346*** 0.0326*** -0.00141 -0.00227 imm*unemp = 50 0.0369*** 0.0328*** -0.000832 -0.00165 imm*unemp = 75 0.0393*** 0.0329*** -0.000319 -0.00105 imm*unemp = 100 0.0419*** 0.0331*** -0.000737 -0.000603 imm*unemp = 125 0.0446*** 0.0333*** -0.00155 -0.000714 imm*unemp = 150 0.0475*** 0.0335*** -0.00248 -0.00125 imm*unemp = 0.0505*** 0.0336*** -0.00351 -0.00189 imm*unemp = 200 0.0537*** 0.0338*** -0.00464 -0.00256 imm*unemp = 225 0.0571*** 0.0340*** -0.00589 -0.00324 imm*unemp = 250 0.0606*** 0.0342***

-0.00724 -0.00394

Observations 102,265

Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.7. Logistic regression - determinants of PRR vote (incl. EPL/position joint variable)

VARIABLES PRR

Age group (base category = 18-24)

Age group (25-34) 0.262** (0.119) Age group (35-44) 0.424*** (0.148) Age group (45-54) 0.469*** (0.165) Age group (55-64) 0.360* (0.193) Age group (65+) 0.240* (0.133) Gender (base = male) -0.458*** (0.0685) epl_insider (level of employment protection conditional on being insider) 0.0286* (0.0162) Anti-immigration for cultural reasons -0.215*** (0.0217) Anti-immigration for economic reasons -0.142*** (0.0126) Important to get respect from others -0.0175 (0.0210) Country (base category = Austria)

Country (Belgium) 0.169*** (0.0519) Country (Switzerland) 1.985*** (0.0583) Country (Germany) -1.349*** (0.0418) Country (Denmark) 0.749*** (0.122) Country (Finland) 0.664*** (0.0473) Country (France) 0.145** (0.0583) Country (UK) -1.916*** (0.0625) Country (Netherlands) 1.073*** (0.0581) Country (Norway) 1.285*** (0.0428) Country (Sweden) -0.352*** (0.0492) Year (base category = 2002)

Year (2004) -0.145 (0.294) Year (2006) 0.0155 (0.242) Year (2008) -0.251 (0.271) Year (2010) 0.0771 (0.130) Year (2012) 0.424 (0.294) Year (2014) 0.690*** (0.265) Year (2016) 1.324*** (0.412) Constant -1.853*** (0.205)

Observations 131,347 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.8. Marginal effects on PRR voting by EPL/position joint variable, from 0 (outsider) to 2.95 (insider in the country with highest EPL)

Insider

0 0.0378*** (0.000734) 0.983 0.0388*** (0.000214) 1.966 0.0397*** (0.000331) 2.950 0.0407*** (0.000899)

Observations 131,347 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Table B.9. Logistic regression - determinants of PRR vote (incl. respect/ position interaction)

(1) VARIABLES PRR

Age group (base category = 18-24)

Age group (25-34) 0.239** (0.104) Age group (35-44) 0.374*** (0.111) Age group (45-54) 0.426*** (0.133) Age group (55-64) 0.286** (0.133) Age group (65+) 0.218 (0.145) Gender (base = male) -0.437*** (0.0657) Important to get respect from others -0.00919 (0.0264) Status (base = outsider) -0.111 (0.108) Outsider#importance_respect -0.00935 (0.0252) Immigr*unemp 0.00228*** (0.000831) Anti-immigration for cultural reasons -0.215*** (0.0246) Anti-immigration for economic reasons -0.145*** (0.0124) Country (base category = Austria)

Country (Belgium) 0.226*** (0.0611) Country (Switzerland) 1.987*** (0.0495) Country (Germany) -1.336*** (0.0491) Country (Denmark) 0.862*** (0.116) Country (Finland) 0.843*** (0.0609) Country (France) 0.0561 (0.0797) Country (UK) -1.802*** (0.0580) Country (Netherlands) 1.246*** (0.0501) Country (Norway) 1.435*** (0.0477) Country (Sweden) -0.266*** (0.0352) Year (base category = 2002)

Year (2004) -0.215 (0.461) Year (2006) 0.104 (0.359) Year (2008) -0.190 (0.330) Year (2010) 0.162 (0.178) Year (2012) 0.461 (0.405) Year (2014) 0.787*** (0.292) Year (2016) 1.461*** (0.483) Constant -2.063*** (0.292)

Observations 102,265

Table B.10. Marginal effects of insiders vs outsiders of PRR voters by closeness to the sentence “it is important to get respect from others” (1 = very much; 7 = not at all)

Insider Outsider

1 0.0408*** 0.0349*** -0.00312 -0.00374 2 0.0405*** 0.0344*** -0.00222 -0.00287 3 0.0402*** 0.0338*** -0.00135 -0.00208 4 0.0399*** 0.0333*** -0.000618 -0.00141 5 0.0395*** 0.0328*** -0.000719 -0.0011 6 0.0392*** 0.0322*** -0.00147 -0.00134 7 0.0389*** 0.0317*** -0.0023 -0.00191 Observations 102,265 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.11. Logistic regression - determinants of PRR vote (incl. class/position interaction)

(1) (2) VARIABLES PRR PRR

Age group (base category = 18-24)

Age group (25-34) 0.340*** 0.254* (0.0966) (0.136) Age group (35-44) 0.447*** 0.341*** (0.0875) (0.131) Age group (45-54) 0.521*** 0.378** (0.128) (0.156) Age group (55-64) 0.415*** 0.269 (0.159) (0.166) Age group (65+) 0.415*** 0.238 (0.133) (0.148) Education (base category = primary)

Education (lower-secondary) 0.231*** 0.364*** (0.0438) (0.0348) Education (upper-secondary) 0.292*** 0.497*** (0.0829) (0.0844) Education (post-secondary) -0.0609 0.331*** (0.163) (0.102) Education (tertiary) -0.493*** 0.0907 (0.142) (0.108) Gender (base = male) -0.301*** -0.387*** (0.0572) (0.0815) Class (base category = production worker)

Class (self-employed) -0.0346 0.209 (0.298) (0.301) Class (small business owner) -0.0303 0.211 (0.212) (0.208) Class (technical pro) -0.0154 0.165 (0.254) (0.304) Class (manager) 0.113 0.374** (0.185) (0.171) Class (clerk) -0.173 0.0605 (0.196) (0.223) Class (socio-cultural) -0.396* -0.124 (0.204) (0.181) Class (service) -0.0270 0.117 (0.309) (0.344) Status (base = outsider) 0.373** 0.398*** (0.146) (0.126) Class (self-employed)#Status (insider) -0.772* -0.768* (0.429) (0.406) Class (small business owner)#Status (insider) -0.304 -0.428* (0.209) (0.229) Class (technical pro)#Status (insider) -0.265 -0.236 (0.180) (0.218) Class (manager)#Status (insider) -0.640*** -0.641*** (0.191) (0.136) Class (clerk)#Status (insider) -0.236* -0.178 (0.123) (0.114) Class (socio-cultural)#Status (insider) -0.495*** -0.388** (0.180) (0.163) Class (service)#Status (insider) -0.179 -0.165 (0.230) (0.221) Anti-immigration 3.976*** (0.237) Anti-redistribution 0.208 (0.155) Trust pol institutions -0.570* (0.340) Country (base category = Austria)

Country (Belgium) 0.338*** 0.504*** (0.0294) (0.0445) Country (Switzerland) 1.634*** 2.280*** (0.0460) (0.101) Country (Germany) -1.394*** -1.090*** (0.0162) (0.0241) Country (Denmark) 0.704*** 1.090*** (0.102) (0.160) Country (Finland) 0.294*** 0.642*** (0.0180) (0.0467) Country (France) 0.235*** 0.302*** (0.0437) (0.0398) Country (UK) -1.578*** -1.584*** (0.0306) (0.0516) Country (Netherlands) 1.028*** 1.408*** (0.0498) (0.0986) Country (Norway) 1.228*** 1.728*** (0.0390) (0.0922) Country (Sweden) -0.730*** 0.0707 (0.0226) (0.0616) Year (base category = 2002)

Year (2004) -0.0911 -0.201 (0.451) (0.463) Year (2006) 0.208 0.129 (0.379) (0.372) Year (2008) -0.156 -0.105 (0.334) (0.318) Year (2010) 0.223 0.239 (0.166) (0.161) Year (2012) 0.503 0.567 (0.433) (0.426) Year (2014) 0.855*** 0.896*** (0.282) (0.300) Year (2016) 1.458*** 1.658*** (0.452) (0.475) Constant -3.939*** -6.430*** (0.384) (0.524)

Observations 108,429 101,044 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.12. Marginal effects of insiders vs outsiders of PRR voters by social class

(1) (2) VARIABLES Status (insider) Status (insider)

Class (self-employed) -0.0109 -0.0104 (0.0100) (0.0113) Class (small business 0.00226 -0.000955 owner) (0.00480) (0.00600) Class (technical pro) 0.00367 0.00535 (0.00463) (0.00473) Class (production worker) 0.0142*** 0.0127*** (0.00462) (0.00326) Class (manager) -0.00867** -0.00805*** (0.00386) (0.00308) Class (clerk) 0.00412 0.00687* (0.00308) (0.00364) Class (socio-cultural) -0.00274 0.000246 (0.00221) (0.00299) Class (service) 0.00676** 0.00762** (0.00334) (0.00345)

Observations 108,429 101,044 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.13. Logistic regression - determinants of PRR vote (incl. skill/position interaction)

VARIABLES PRR

Age group (base category = 18-24)

Age group (25-34) 0.204* (0.112) Age group (35-44) 0.274** (0.108) Age group (45-54) 0.310** (0.133) Age group (55-64) 0.178 (0.143) Age group (65+) 0.0783 (0.167) Gender (base = male) -0.428*** (0.0686) Skill level (base = low) 0.0445 (0.115) Status (base = outsider) 0.269*** (0.0354) Skill (low)#Status (outsider) -0.362*** (0.0984) Anti-immigration 4.064*** (0.232) Anti-redistribution 0.175 (0.169) Trust pol institutions -0.587* (0.340) Country (base category = Austria)

Country (Belgium) 0.402*** (0.0481) Country (Switzerland) 2.251*** (0.104) Country (Germany) -1.114*** (0.0290) Country (Denmark) 1.030*** (0.163) Country (Finland) 0.551*** (0.0612) Country (France) 0.241*** (0.0464) Country (UK) -1.754*** (0.0474) Country (Netherlands) 1.321*** (0.0940) Country (Norway) 1.672*** (0.0952) Country (Sweden) 0.0254 (0.0628) Year (base category = 2002)

Year (2004) -0.221 (0.463) Year (2006) 0.107 (0.369) Year (2008) -0.127 (0.321) Year (2010) 0.229 (0.170) Year (2012) 0.567 (0.428) Year (2014) 0.894*** (0.303) Year (2016) 1.646*** (0.490) Constant -5.858*** (0.345)

Observations 101,405 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.14. Marginal effects of insiders vs outsiders of PRR voters by skill level

VARIABLES Status (insider)

High-skilled -0.00273 (0.00257) Low-skilled 0.00874*** (0.00106)

Observations 101,405 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.

Table B.15. Regression - determinants on PRR vote (incl. interaction between job position and country) (1) (2) VARIABLES PRR PRR

Age group (base category = 18-24)

Age group (25-34) 0.0171** 0.0111* (0.00627) (0.00563) Age group (35-44) 0.0227*** 0.0157** (0.00607) (0.00529) Age group (45-54) 0.0206*** 0.0124** (0.00530) (0.00489) Age group (55-64) 0.0220** 0.0123 (0.00792) (0.00687) Age group (65+) 0.0266** 0.0139 (0.0118) (0.00994) Education (base category = primary)

Education (lower-secondary) 0.0181*** 0.0214*** (0.00493) (0.00561) Education (upper-secondary) 0.0203** 0.0270** (0.00816) (0.00934) Education (post-secondary) 0.0107 0.0246* (0.0107) (0.0120) Education (tertiary) -0.0145* 0.00983 (0.00663) (0.00616) Gender (base = male) -0.0160*** -0.0184*** (0.00409) (0.00379) Class (base = production worker)

Class (self-employed) -0.0309*** -0.0170** (0.00891) (0.00746) Class (small business owner) -0.0173 -0.00994 (0.0116) (0.0108) Class (technical pro) -0.0242*** -0.0146* (0.00753) (0.00705) Class (manager) -0.0309** -0.0185* (0.0100) (0.00846) Class (clerk) -0.0230*** -0.0129** (0.00611) (0.00523) Class (socio-cultural) -0.0405*** -0.0222*** (0.00864) (0.00663) Class (service) -0.0173** -0.0107 (0.00643) (0.00599) Status (base = outsider) 0.00526** -0.00204 (0.00170) (0.00261) Country (base category = Austria)

Country (Belgium) 0.0169*** 0.0107** (0.00294) (0.00357) Country (Switzerland) 0.118*** 0.134*** (0.00141) (0.00462) Country (Germany) -0.0286*** -0.0263*** (0.00184) (0.00187) Country (Denmark) 0.0519*** 0.0593*** (0.00332) (0.00385) Country (Finland) 0.0172*** 0.0235*** (0.00234) (0.00398) Country (France) 0.00857* -0.00106 (0.00406) (0.00385) Country (UK) -0.0236*** -0.0370*** (0.00284) (0.00330) Country (Switzerland) 0.0576*** 0.0601*** (0.00319) (0.00471) Country (Norway) 0.0949*** 0.102*** (0.00176) (0.00377) Country (Sweden) -0.0147*** 0.0164* (0.00202) (0.00793) Status (base = insider)#Country (Austria) 0 0 (0) (0) Status (base = insider)#Country (Belgium) 0.00306 0.0131** (0.00288) (0.00516) Status (base = insider)#Country (Switzerland) 0.0127*** 0.0226*** (0.000858) (0.00253) Status (base = insider)#Country (Germany) -0.00257** 0.00513 (0.00102) (0.00310) Status (base = insider)#Country (Denmark) -0.0226*** -0.0151*** (0.00235) (0.00399) Status (base = insider)#Country (Finland) -0.00263 0.00417 (0.00168) (0.00299) Status (base = insider)#Country (France) 0.00838*** 0.0178*** (0.00198) (0.00416) Status (base = insider)#Country (UK) -0.00209 0.00561 (0.00189) (0.00405) Status (base = insider)#Country (Netherlands) 0.00927*** 0.0168*** (0.00172) (0.00332) Status (base = insider)#Country (Norway) -0.0157*** -0.00634 (0.00202) (0.00415) Status (base = insider)#Country (Sweden) -0.00459* 0.00255 (0.00221) (0.00363) Anti-immigration 0.192*** (0.0438) Anti-redistribution 0.0186 (0.0111) Trust pol institutions -0.0477** (0.0198) Year (base category = 2002)

Year (2004) -0.000519 -0.00400 (0.0134) (0.0140) Year (2006) 0.00562 0.00262 (0.0102) (0.0101) Year (2008) 0.00437 0.00688 (0.0125) (0.0126) Year (2010) 0.0146 0.0156 (0.00986) (0.0106) Year (2012) 0.0174 0.0216 (0.0163) (0.0158) Year (2014) 0.0324** 0.0362** (0.0142) (0.0146) Year (2016) 0.0899** 0.0987** (0.0393) (0.0403) Constant 0.0187 -0.0656 (0.0170) (0.0377)

Observations 108,429 101,044 R-squared 0.062 0.090 Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table B.16. Marginal effects of insiders vs outsiders of PRR voters by country (1) (2) VARIABLES Status (insider) Status (insider)

Country (Austria) 0.00526** -0.00204 (0.00170) (0.00261) Country (Belgium) 0.00832** 0.0111** (0.00337) (0.00352) Country (Switzerland) 0.0179*** 0.0206*** (0.00211) (0.00188) Country (Germany) 0.00269 0.00309 (0.00206) (0.00187) Country (Denmark) -0.0174*** -0.0172*** (0.00259) (0.00227) Country (Finland) 0.00263 0.00214 (0.00183) (0.00141) Country (France) 0.0136*** 0.0157*** (0.00249) (0.00268) Country (UK) 0.00318 0.00357 (0.00258) (0.00258) Country 0.0145*** 0.0148*** (Netherlands) (0.00225) (0.00186) Country (Norway) -0.0105*** -0.00838*** (0.00258) (0.00258) Country (Sweden) 0.000675 0.000510 (0.00252) (0.00204)

Observations 108,429 101,044 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 C. Who wants flexicurity?

Table C.1. Descriptive statistics of the Eurobarometer dataset

N Mean Min Max Categories Age 16,682 40.980 18 90+ Discrete year by year except for “90+” Gender 16,682 1.46 1 2 1 = male; 2= female

*1 large employers and higher-grade managers *2 self-employed rofessionals *3 small owners *4 farmer & fisherman *5 employed professionals and technical experts Occupation 16,527 1 10 *6 lower-grade managers *7 clerks, sellers, transporters *8 general service workers *9 skilled manual *10 unskilled manual Job status 16,682 1 2 1=employed | 2=unemployed “more flexible contracts” improves employability: 1=totally agree Preferences on epl 15,961 2.12 1 4 | 2=tend to agree | 3=tend to disagree | 4=totally disagree “increasing income difference between working and not working Preferences on neb 15,821 2.19 1 4 people” is ____ to foster employment creation: 1=very effective | 2=fairly effecitve | 3=not very effective | 4=not at all effective “regular training improves” employability: 1=totally agree | Preferences on almp 16,489 1.64 1 4 2=tend to agree | 3=tend to disagree | 4=totally disagree

Table C.2. Distribution of key independent variables of the Eurobarometer dataset

N Freq

Occupation Large employers & higher grade managers 884 5.35 Self-employed professionals 591 3.58 Small owners 1059 6.41 Farmer & fishermen 198 1.20 Employed professionals and technical experts 370 2.24 Middle and lower-level management 2523 15.27 Clerks, sellers, transporters 3702 22.40 General service workers 2490 15.07 Skilled manual workers 3302 19.98 Unskilled manual workers 1408 8.52

Job status Employed 14,485 86.83 Unemployed 2,197 13.17

Table C.3. Logistic regression - strongly preferring EPL cuts (flexcon), NEB cuts (neb) or ALMP increases (almp)

1 2 3 flexcon neb almp

Class = 1 (large employer & high manager– base category)

Class = 2 (self-employed professional) 0.111 -0.132 0.0670 (0.171) (0.192) (0.165)

Class = 3 (small owner) 0.0253 0.325** -0.198 (0.148) (0.157) (0.147)

Class = 4 (farmers & fishermen) 0.234 0.217 -0.433** (0.221) (0.233) (0.217)

Class = 5 (employed professional and technicians) -0.255 -0.245 0.0320 (0.192) (0.214) (0.177)

Class = 6 (middle and low management) -0.452*** -0.158 0.0275 (0.129) (0.138) (0.124)

Class = 7 (clerks, sellers, transporters) -0.416*** -0.150 -0.269** (0.122) (0.131) (0.118)

Class = 8 (general service workers) -0.210* 0.0624 -0.175 (0.126) (0.135) (0.122)

Class = 9 (skilled manual workers) -0.317** 0.0930 -0.575*** (0.124) (0.133) (0.120)

Class = 10 (unskilled manual workers) -0.260* -0.0740 -0.734*** (0.144) (0.154) (0.138)

Gender (1 = male) 0.149*** 0.0611 -0.0525 (0.0539) (0.0568) (0.0500)

- - Age exact 5.25e-05 0.000786 0.000883 (0.00229) (0.00240) (0.00206)

Country = 1 (France – base category)

Country = 2, Belgium 0.0750 0.867*** 0.555*** (0.0995) (0.100) (0.0913)

Country = 3, The Netherlands -0.0897 0.394*** 1.104*** (0.102) (0.104) (0.0925)

Country = 4, Germany (West) 0.159 0.652*** 1.237*** (0.101) (0.104) (0.0950)

Country = 5, Italy -0.211** -0.0687 -0.384*** (0.103) (0.109) (0.0960)

Country = 6, Luxembourg 0.194 0.117 0.937*** (0.122) (0.134) (0.112)

Country = 7, Denmark 0.585*** 0.214** 1.709*** (0.0955) (0.107) (0.0976)

Country = 11, Greece 0.204** 0.340*** 0.616*** (0.102) (0.109) (0.0959)

Country = 12, Spain 0.0579 0.0178 0.775*** (0.0975) (0.107) (0.0904)

Country = 13, Portugal -0.521*** -0.251** 0.0993 (0.108) (0.114) (0.0939)

Country = 14, Germany (East) 0.185 1.094*** 0.944*** (0.129) (0.128) (0.121)

Country = 16, Finland 0.152 -0.459*** 0.758*** (0.100) (0.122) (0.0933)

Country = 17, Sweden -0.0821 -0.450*** 1.239*** (0.103) (0.123) (0.0948)

Country = 18, Austria 0.211** 0.661*** 0.420*** (0.0950) (0.0986) (0.0881)

Country = 19, Cyprus (Republic) 1.344*** 1.071*** 1.704*** (0.116) (0.120) (0.123)

Country = 25, Malta 0.788*** 0.799*** 1.064*** (0.142) (0.145) (0.140)

Year = 2006 (base category)

Year = 2009 -0.299** -0.511*** -0.437*** (0.127) (0.128) (0.108)

Year = 2011 -0.137 -0.140 -0.274** (0.126) (0.127) (0.107)

Position = 1 (employed – base category)

Position = 2 (unemployed) 0.348*** -0.353*** -0.174** (0.0778) (0.0905) (0.0752)

Constant -0.802*** -1.043*** 0.129 (0.220) (0.231) (0.201)

Observations 16,527 16,527 16,527 Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Table C.4. Marginal probabilities - strongly preferring EPL cuts (flexcon), NEB cuts (neb) or ALMP increases (almp)

(1) (2) (3) flexcon neb almp

Class = 1 (large employer & high manager) 0.325*** 0.254*** 0.539*** (0.0233) (0.0211) (0.0241) Class = 2 (self-employed professional) 0.350*** 0.230*** 0.554*** (0.0298) (0.0262) (0.0287) Class = 3 (small owner) 0.331*** 0.317*** 0.494*** (0.0221) (0.0217) (0.0231) Class = 4 (farmers & fishermen) 0.378*** 0.295*** 0.440*** (0.0448) (0.0405) (0.0430) Class = 5 (employed professional and technicians) 0.272*** 0.211*** 0.546*** (0.0314) (0.0291) (0.0326) Class = 6 (middle and low management) 0.236*** 0.226*** 0.545*** (0.0129) (0.0126) (0.0150) Class = 7 (clerks, sellers, transporters) 0.242*** 0.227*** 0.477*** (0.0104) (0.0101) (0.0116) Class = 8 (general service workers) 0.281*** 0.265*** 0.499*** (0.0131) (0.0129) (0.0137) Class = 9 (skilled manual workers) 0.261*** 0.271*** 0.408*** (0.0117) (0.0122) (0.0126) Class = 10 (unskilled manual workers) 0.271*** 0.240*** 0.374*** (0.0179) (0.0174) (0.0187)

Position = 1 (employed – base category) 0.261*** 0.257*** 0.481*** (0.00536) (0.00532) (0.00577)

Position = 2 (unemployed) 0.332*** 0.198*** 0.442*** (0.0157) (0.0131) (0.0156)

Observations 16,527 16,527 16,527 Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Table C.5. Logistic regression - preference for flexicurity over dualization (dualiz_flexic) and for liberalization over protection (protec_lib)

1 2 dualiz_flexic protec_lib

Class = 1 (large employer & high manager– base category)

Class = 2 (self-employed professional) -0.0621 0.166 (0.303) (0.205)

Class = 3 (small owner) -0.400 0.139 (0.261) (0.184)

Class = 4 (farmers & fishermen) -0.0346 0.0530 (0.425) (0.287)

Class = 5 (employed professional and technicians) -0.501* 0.00821 (0.283) (0.268)

Class = 6 (middle and low management) -0.267 -0.0864 (0.224) (0.156)

Class = 7 (clerks, sellers, transporters) -0.456** -0.287* (0.217) (0.151)

Class = 8 (general service workers) -0.448** -0.0441 (0.220) (0.160)

Class = 9 (skilled manual workers) -0.779*** -0.295* (0.222) (0.159)

Class = 10 (unskilled manual workers) -1.042*** -0.135 (0.243) (0.181)

Position = 1 (employed – base category)

Position = 2 (unemployed) 0.587*** -0.116 (0.140) (0.114)

Gender (1 = male) 0.208** 0.0762 (0.0991) (0.0648)

Age exact -0.00425 0.00318 (0.00419) (0.00270)

Country = 1 (France – base category) Country = 2, Belgium 0.0133 0.471*** (0.209) (0.116)

Country = 3, The Netherlands -0.0525 0.666*** (0.207) (0.119)

Country = 4, Germany (West) 0.623*** 0.812*** (0.187) (0.126)

Country = 5, Italy -0.00976 -0.497*** (0.196) (0.125)

Country = 6, Luxembourg 0.529*** 0.816*** (0.204) (0.165)

Country = 7, Denmark 1.023*** 0.856*** (0.157) (0.112)

Country = 11, Greece 0.0739 0.439*** (0.213) (0.121)

Country = 12, Spain 0.603*** 0.382*** (0.175) (0.135)

Country = 13, Portugal -0.287 -0.218* (0.203) (0.119)

Country = 14, Germany (East) 0.551* 0.992*** (0.319) (0.146)

Country = 16, Finland 0.608*** 0.192 (0.160) (0.127)

Country = 17, Sweden 0.631*** 0.404*** (0.155) (0.118)

Country = 18, Austria 0.275 0.390*** (0.200) (0.112)

Country = 19, Cyprus (Republic) 0.940*** 1.712*** (0.216) (0.136)

Country = 25, Malta 0.751** 1.131*** (0.364) (0.192)

Year = 2006 (base category)

Year = 2009 0.0368 -0.516*** (0.226) (0.139)

Year = 2011 0.125 -0.215 (0.225) (0.139)

Constant cut1 -2.270*** -3.088*** (0.388) (0.266)

Constant cut2 -0.630 -0.906*** (0.389) (0.259)

Constant cut3 0.867** 0.360 (0.390) (0.259)

Constant cut4 2.483*** 1.718*** (0.394) (0.263)

Observations 3,259 7,514 Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Table C.6. Marginal probabilities - preference for flexicurity over dualization (dualiz_flexic) and for liberalization over protection (protec_lib), where 1=strong preference for dualiz/protec and 5=strong preference for flexic/lib

Dualiz_flexic Protec_lib

1 2 3 4 5 1 2 3 4 5

Class = 1 (large employer & high manager) 0.0574*** 0.177*** 0.330*** 0.297*** 0.140*** 0.0431*** 0.229*** 0.279*** 0.266*** 0.183***

-0.0108 -0.0239 -0.0162 -0.0254 -0.0229 -0.00634 -0.0208 -0.00967 -0.0141 -0.0203

Class = 2 (self-employed professional) 0.0609*** 0.184*** 0.333*** 0.289*** 0.133*** 0.0368*** 0.205*** 0.271*** 0.279*** 0.208***

-0.0141 -0.0302 -0.0179 -0.031 -0.0283 -0.0061 -0.0222 -0.0118 -0.0143 -0.025

Class = 3 (small owner) 0.0830*** 0.227*** 0.344*** 0.246*** 0.0989*** 0.0378*** 0.209*** 0.272*** 0.277*** 0.204***

-0.0148 -0.0255 -0.0127 -0.0257 -0.0171 -0.00528 -0.0179 -0.0103 -0.0123 -0.02

Class = 4 (farmers & fishermen) 0.0593*** 0.181*** 0.332*** 0.293*** 0.136*** 0.0410*** 0.221*** 0.277*** 0.270*** 0.191***

-0.0213 -0.0469 -0.0251 -0.0464 -0.0451 -0.0103 -0.0374 -0.0146 -0.0229 -0.0382

Class = 5 (employed professional and technicians) 0.0909*** 0.241*** 0.345*** 0.233*** 0.0905*** 0.0428*** 0.228*** 0.279*** 0.266*** 0.184***

-0.0189 -0.0303 -0.0126 -0.0296 -0.0186 -0.00999 -0.0347 -0.013 -0.022 -0.034

Class = 6 (middle and low management) 0.0735*** 0.210*** 0.342*** 0.264*** 0.111*** 0.0467*** 0.242*** 0.283*** 0.258*** 0.171***

-0.0101 -0.0186 -0.0129 -0.0189 -0.0149 -0.00488 -0.0132 -0.00802 -0.0102 -0.0116

Class = 7 (clerks, sellers, transporters) 0.0874*** 0.235*** 0.345*** 0.239*** 0.0941*** 0.0564*** 0.273*** 0.288*** 0.238*** 0.145***

-0.011 -0.0169 -0.0126 -0.0171 -0.0112 -0.00497 -0.0118 -0.00794 -0.00905 -0.0091

Class = 8 (general service workers) 0.0867*** 0.234*** 0.345*** 0.240*** 0.0948*** 0.0449*** 0.236*** 0.281*** 0.262*** 0.177***

-0.0118 -0.018 -0.0125 -0.0183 -0.0116 -0.0048 -0.0134 -0.00819 -0.0102 -0.0126

Class = 9 (skilled manual workers) 0.116*** 0.277*** 0.338*** 0.198*** 0.0703*** 0.0568*** 0.274*** 0.288*** 0.237*** 0.144***

-0.0149 -0.0188 -0.0131 -0.0169 -0.00987 -0.006 -0.0137 -0.00792 -0.0103 -0.0106

Class = 10 (unskilled manual workers) 0.146*** 0.309*** 0.324*** 0.166*** 0.0551*** 0.0489*** 0.249*** 0.284*** 0.253*** 0.164***

-0.0207 -0.0217 -0.0159 -0.0187 -0.00919 -0.00653 -0.0184 -0.00841 -0.0129 -0.0161

Position = 1 (employed – base category) 0.0982*** 0.250*** 0.342*** 0.224*** 0.0852*** 0.0490*** 0.248*** 0.282*** 0.254*** 0.167***

-0.00831 -0.0116 -0.0125 -0.0108 -0.00736 -0.00383 -0.00757 -0.00771 -0.00752 -0.00656

Position = 2 (unemployed) 0.0575*** 0.175*** 0.327*** 0.298*** 0.142*** 0.0547*** 0.266*** 0.285*** 0.242*** 0.152***

-0.00828 -0.017 -0.0142 -0.0192 -0.0177 -0.0069 -0.0173 -0.00803 -0.0123 -0.0144

Observations 3,259 3,259 3,259 3,259 3,259 7,514 7,514 7,514 7,514 7,514

Robust standard errors in parenthesis *** p<0.01, ** p<0.05, * p<0.1

Table C.7. Logistic regressions - preference for neb restriction to those who turn down jobs for which they are overqualified (ubedu); and preference for almp increases in exchange of neb (eduunmp); where 1=strong preference for maintaining neb and 4=strong preference for cuts/almp increase

(1) (2) ubedu eduunmp

Class = 1 (large employer & high manager– base category)

Class = 2 (self-employed professional) -0.383 0.0652 (0.369) (0.166) Class = 3 (small owner) 0.0819 -0.138 (0.162) (0.101) Class = 4 (farmers & fishermen) -0.575 -0.0914 (0.956) (0.246) Class = 5 (employed professional and technicians) -0.505*** -0.189** (0.160) (0.0964) Class = 6 (middle and low management) -0.00965 -0.0188 (0.180) (0.103) Class = 7 (clerks, sellers, transporters) 0.173 -0.281*** (0.158) (0.0944) Class = 8 (general service workers) 0.171 -0.363*** (0.176) (0.107) Class = 9 (skilled manual workers) 0.360** -0.350*** (0.174) (0.102) Class = 10 (unskilled manual workers) 0.202 -0.339*** (0.279) (0.127)

Position = 1 (employed – base category)

Position = 2 (unemployed) -0.532*** -0.381*** (0.155) (0.0805)

Gender 0.216** -0.100** (0.0912) (0.0495)

Age exact -0.00154 -0.00832*** (0.00357) (0.00191)

Country = 1, Austria (base outcome)

Country = 2, Belgium 0.194 0.692*** (0.143) (0.0959)

Country = 3, Switzerland 0.323** 0.319*** (0.138) (0.0989)

Country = 5, Germany -0.0719 0.188** (0.126) (0.0907)

Country = 7, Spain 0.857*** -0.0807 (0.184) (0.0964)

Country = 8, Finland 0.626*** 0.274*** (0.147) (0.0912)

Country = 9, France -0.259* 0.432*** (0.146) (0.102)

Country = 15, Italy 1.641*** 1.075*** (0.143) (0.0975)

Country = 17, Netherlands 0.493*** 0.993*** (0.151) (0.102)

Country = 18, Norway 1.220*** 1.129*** (0.162) (0.0974)

Country = 20, Portugal 0.302 0.235* (0.271) (0.124)

Country = 22, Sweden 0.787*** 0.333*** (0.167) (0.107)

Constant cut1 -0.642** -3.273*** (0.269) (0.165)

Constant cut2 0.946*** -1.042*** (0.270) (0.157)

Constant cut3 1.925*** 1.878*** (0.274) (0.160)

Observations 3,162 12,509 Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Table C.8. Marginal probabilities- preference for neb restriction to those who turn down jobs for which they are overqualified (ubedu); and preference for almp increases in exchange of neb (eduunmp); where 1=strong preference for maintaining neb and 4=strong preference for cuts/almp increase

ubedu eduunmp

1 2 3 4 1 2 3 4

Class = 1 (large employer & high manager – base category) 0.242*** 0.335*** 0.191*** 0.232*** 0.0446*** 0.250*** 0.580*** 0.126***

-0.0222 -0.0126 -0.0112 -0.0212 -0.00384 -0.0128 -0.00917 -0.00885

Class = 2 (self-employed professional) 0.313*** 0.345*** 0.166*** 0.176*** 0.0419*** 0.240*** 0.585*** 0.133***

-0.0698 -0.0117 -0.025 -0.0469 -0.00614 -0.0232 -0.0135 -0.0169

Class = 3 (small owner) 0.228*** 0.331*** 0.195*** 0.246*** 0.0508*** 0.272*** 0.566*** 0.112***

-0.0195 -0.0122 -0.0108 -0.0192 -0.00393 -0.0114 -0.00931 -0.00698

Class = 4 (farmers & fishermen) 0.353* 0.344*** 0.153** 0.151 0.0486*** 0.264*** 0.571*** 0.116***

-0.199 -0.0215 -0.0686 -0.113 -0.0109 -0.0378 -0.0255 -0.0236

Class = 5 (employed professional and technicians) 0.338*** 0.345*** 0.158*** 0.160*** 0.0533*** 0.280*** 0.559*** 0.107***

-0.0227 -0.0117 -0.0102 -0.0144 -0.0038 -0.0102 -0.00876 -0.00613

Class = 6 (middle and low management) 0.243*** 0.336*** 0.190*** 0.231*** 0.0454*** 0.253*** 0.578*** 0.124***

-0.024 -0.0125 -0.0116 -0.0221 -0.00364 -0.0114 -0.00855 -0.00796 Class = 7 (clerks, sellers, transporters) 0.213*** 0.326*** 0.200*** 0.262*** 0.0581*** 0.295*** 0.548*** 0.0986***

-0.0171 -0.0123 -0.0106 -0.0181 -0.00383 -0.00954 -0.00857 -0.00537

Class = 8 (general service workers) 0.214*** 0.326*** 0.200*** 0.261*** 0.0627*** 0.309*** 0.536*** 0.0916***

-0.021 -0.0132 -0.0111 -0.0227 -0.00487 -0.0128 -0.0113 -0.00652

Class = 9 (skilled manual workers) 0.185*** 0.312*** 0.208*** 0.295*** 0.0620*** 0.307*** 0.538*** 0.0927***

-0.0191 -0.0141 -0.011 -0.024 -0.00455 -0.0118 -0.0105 -0.0061

Class = 10 (unskilled manual workers) 0.209*** 0.324*** 0.201*** 0.267*** 0.0614*** 0.305*** 0.540*** 0.0936***

-0.0396 -0.0195 -0.0152 -0.0441 -0.00619 -0.0172 -0.0147 -0.00883

Position = 1 (employed – base category) 0.232*** 0.327*** 0.193*** 0.249*** 0.0520*** 0.276*** 0.562*** 0.110***

-0.0104 -0.0109 -0.00939 -0.00952 -0.00255 -0.00541 -0.00597 -0.00368

Position = 2 (unemployed) 0.330*** 0.340*** 0.161*** 0.170*** 0.0741*** 0.339*** 0.509*** 0.0781***

-0.0302 -0.0115 -0.0119 -0.0198 -0.00602 -0.0132 -0.0129 -0.00599

Observations 3,823 3,823 3,823 3,823 14,949 14,949 14,949 14,949

Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1