Parties, Gender Quotas and Candidate Selection in France (French Politics, Society and Culture)

Parties, Gender Quotas and Candidate Selection in France Rainbow Murray

French Politics, Society and Culture Series General Editor: Robert Elgie, Paddy Moriarty Professor of Government and International Studies, Dublin City University France has always fascinated outside observers. Now, the country is undergoing a period of profound transformation. France is faced with a rapidly changing international and European environment and it is having to rethink some of its most basic social, political and economic orthodoxies. As elsewhere, there is pressure to conform. And yet, while France is responding in ways that are no doubt familiar to people in other European countries, it is also managing to maintain elements of its long-standing distinctiveness. Overall, it remains a place that is not exactly comme les autres. This new series examines all aspects of French politics, society and culture. In so doing it focuses on the changing nature of the French system as well as the established patterns of political, social and cultural life. Contributors to the series are encouraged to present new and innovative arguments so that the informed reader can learn and understand more about one of the most beguiling and compelling of all European countries.

Titles include: Gill Allwood and Khursheed Wadia GENDER AND POLICY IN FRANCE Sylvain Brouard, Andrew M. Appleton, Amy G. Mazur (editors) THE FRENCH FIFTH REPUBLIC AT FIFTY Beyond Stereotypes June Burnham POLITICIANS, BUREAUCRATS AND LEADERSHIP IN ORGANIZATIONS Lessons from Regional Planning in France Jean K. Chalaby THE DE GAULLE PRESIDENCY AND THE MEDIA Statism and Public Communications Pepper D. Culpepper, Bruno Palier and Peter A. Hall (editors) CHANGING FRANCE The Politics that Markets Make Gordon D. Cumming FRENCH NGOs IN THE GLOBAL ERA France’s International Development Role David Drake FRENCH INTELLECTUALS AND POLITICS FROM THE DREYFUS AFFAIR TO THE OCCUPATION David Drake INTELLECTUALS AND POLITICS IN POST-WAR FRANCE Graeme Hayes ENVIRONMENTAL PROTEST AND THE STATE IN FRANCE

David J. Howarth THE FRENCH ROAD TO EUROPEAN MONETARY UNION Andrew Knapp PARTIES AND THE PARTY SYSTEM IN FRANCE A Disconnected Democracy? Michael S. Lewis-Beck (editor) THE FRENCH VOTER Before and After the 2002 Elections John Loughlin SUBNATIONAL GOVERNMENT The French Experience Mairi Maclean and Joseph Szarka FRANCE ON THE WORLD STAGE Mairi Maclean, Charles Harvey and Jon Press BUSINESS ELITES AND CORPORATE GOVERNANCE IN FRANCE AND THE UK Susan Milner and Nick Parsons (editors) REINVENTING FRANCE State and Society in the Twenty-First Century Rainbow Murray PARTIES, GENDER QUOTAS AND CANDIDATE SELECTION IN FRANCE Gino G. Raymond THE FRENCH COMMUNIST PARTY DURING THE FIFTH REPUBLIC A Crisis of Leadership and Ideology Paul Smith THE SENATE OF THE FIFTH FRENCH REPUBLIC Francesca Vassallo FRANCE, SOCIAL CAPITAL AND POLITICAL ACTIVISM Sarah Waters SOCIAL MOVEMENTS IN FRANCE Towards a New Citizenship Reuben Y. Wong THE EUROPEANIZATION OF FRENCH FOREIGN POLICY France and the EU in East Asia

French Politics, Society and Culture Series Standing Order ISBN 978-0–333–80440–7 hardcover Series Standing Order ISBN 978-0–333–80441–4 paperback (outside North America only) You can receive future titles in this series as they are published by placing a standing order. Please contact your bookseller or, in case of difficulty, write to us at the address below with your name and address, the title of the series and the ISBN quoted above. Customer Services Department, Macmillan Distribution Ltd, Houndmills, Basingstoke, Hampshire RG21 6XS, England

Parties, Gender Quotas and Candidate Selection in France Rainbow Murray Lecturer in Politics, Queen Mary University of London, UK

© Rainbow Murray 2010 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6–10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The author has asserted her right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2010 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries ISBN 978-0-230-24253-1

hardback

This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Murray, Rainbow, 1979– Parties, gender quotas, and candidate selection in France / Rainbow Murray. p. cm. – (French politics, society, and culture series) Includes bibliographical references and index. ISBN 978-0-230-24253-1 (hardback : alk. paper) 1. Nominations for office–France. 2. Political parties–France. 3. Women–Political activity–France. 4. France–Politics and government–2007– I. Title. JN2997.M87 2010 324.244’015–dc22 2009048544 10 19

9 18

8 17

7 16

6 15

5 14

4 13

3 12

2 11

1 10

Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne

For Robin, Vicky and my family, with gratitude and love

This page intentionally left blank

Contents

List of Tables and Illustrations

ix

Acknowledgements

x

Chapter 1

Introduction Why gender quotas? Why parties? Why France? Parties as rational actors From multiple models to one Overview of the book

1 2 4 8 14 16 21

Chapter 2

French Parties and Parity in Context Introduction The French party system Party motivations in ratifying parity

26 26 27 38

Chapter 3

Candidate Selection, Supply and Demand Party candidate selection procedures The significance of centralisation Selection procedures in French parties Does centralisation of candidate selection assist parity? The ‘supply and demand’ model in France

46 47 47 49

The Electoral Competition Approach Introduction Overview of the electoral competition approach Electoral competition in France How electoral competition affects candidate selection Evaluating the performance of types of candidate

75 75 76 81

Chapter 4

Chapter 5

The Institutional Approach Introduction Examining the ‘institutional’ approach France’s institutions and their impact on candidate selection vii

56 57

85 89 104 104 106 109

viii Contents

Chapter 6

Chapter 7

The Ideological Approach Introduction Examining the ideological approach ‘Familles spirituelles’: Comparing ideologies and attitudes to quotas Can ideology explain the implementation of parity? Applying the ideological approach to French parties

125 125 126 131

Conclusion and the ‘Party Priorities Model’ Introduction Overview of the key themes Models of party behaviour The party priorities model Overall conclusions

148 148 148 153 157 163

134 137

Appendix: Interviews

167

Notes

169

Bibliography

175

Index

186

List of Tables and Illustrations Tables 2.1 3.1 3.2 4.1 4.2 4.3 4.4 4.5 5.1 5.2 6.1

Overview of key party characteristics Overview of party selection procedures Percentage of women in parties and politics Vulnerability of seats in French elections Measuring candidate impact on party performance Success rates of incumbents and inheritors Female incumbents and inheritors Sex and safety of seat Electoral systems and parity in France The impact of parity Ideology and implementation of parity

38 55 61 82 92 94 95 98 112 118 135

Figures and illustrations 5.1 Effects of electoral systems on women’s representation The Party Priorities Model

ix

110 159

Acknowledgements Elements of Chapter 4 were published in Party Politics (2008), vol. 14(5): 539–54. Elements of Chapter 7 were published in Parliamentary Affairs (2007), vol. 60(4): 568–84. Many people have contributed to this book in various ways. While it is not possible to produce an exhaustive list, I would like to acknowledge certain people without whom the book would not have been possible. This book emanates from my research work at Birkbeck, University of London. I would like to thank my supervisors, Joni Lovenduski and Rosie Campbell, for their wisdom, guidance and support throughout my study. Early inspiration and encouragement came from Jill Lovecy and Mariette Sineau. Robert Elgie has contributed in many ways to my work, not only through being the editor of this book series, but also as an examiner, journal editor and supportive colleague. My other examiner, Amy Mazur, has also been a pillar of support and a nurturing influence in the profession. Together, their constructive criticism and encouragement was crucial in the development of this work into a book. Vital and much-appreciated financial support came from the ESRC, who sponsored my studies through a 1+3 studentship. I am immensely grateful for this award. Many people have given feedback on various aspects of my work and some have also supported my work through conferences, publications and guidance on sources. In particular I would like to thank Tim Bale, Marc Calvini-Lefebvre, Sarah Childs, Ben Clift, Jocelyn Evans, Florence Haegel, Jennifer von Heerde, Peter John, Jerry Johnson, Hussein Kassim, Andrew Knapp, Mona Lena Krook, Katherine Opello, Judith Squires, Aurélia Troupel and Alan Ware. Many people played an invaluable role in helping me with my fieldwork. I would like to thank every person who kindly gave up their precious time to be interviewed for this thesis. In addition to the politicians listed in Appendix B, several academics generously gave me their time, including Claudie Baudino, Laure Bereni, Julien Fretel, Eléonore Lépinard, Janine Mossuz-Lavau and Lea Sgier. Further, many of the interviews x

Acknowledgements xi

with politicians would not have taken place without the assistance of friends and colleagues who helped set them up. In addition to those mentioned above, special thanks must go to Emmanuelle de Champs, Emilie Mazzacurati, Gwenaëlle Perrier, Julie Savary, and an extra big thank you to Réjane Sénac-Slawinski for unlocking several important doors for me. Good peers and colleagues are a godsend, and I have been fortunate to work with some wonderful people. My thanks go to colleagues at Birkbeck and latterly at Queen Mary, University of London, for their support throughout the preparation of this book. I am indebted to Alison Howson, Liz Blackmore, and Shirley Tan at Palgrave for their patience, efficiency and hard work in helping this book take shape. I am also grateful to the anonymous reviewer for their helpful insights. As is customary with such things, I have left thanking my family and friends till last. However, it is probably to them that I owe the greatest debt of all. I am fortunate enough to have more people to thank than I can list here, but want to make a few special mentions. Firstly, my parents for their unwavering support through every crisis, constant encouragement, their time, their love, and for providing me with the many essentials that enabled me to complete my research. Mum, Dad, thank you – I hope this brings you nachas! Secondly, my dear friend Vicky, who has been by my side from beginning to end, cheering me on, making many sacrifices for me and my work, and providing every kind of support imaginable with great patience and love. Finally, Robin Thomas Pettitt, who as a colleague has given me academic inspiration and guidance, and as a partner has given me unconditional love and support. Tak for at være min solskin. Without these special people, my world would be a darker place and this book would be non-existent. Receiving the help and encouragement of so many wonderful people has been a privilege and a blessing. Thank you all.

This page intentionally left blank

1 Introduction

This book looks at how political parties have responded to the demands of gender quotas, using France as a case study. Gender quotas are a rapidly growing phenomenon and they affect politics in a number of ways, from shaping the composition of legislatures to influencing the legitimacy of the political system. Parties are the central actors at the heart of this process – they are the ones who introduce quotas, either on a piecemeal voluntary basis or through enacting legislation, and they are then the actors charged with quota implementation. Quotas have a huge impact on parties, influencing their recruitment and selection practices, their electoral prospects and their image. However, despite the critical influence of parties on the introduction and implementation of quotas, and of quotas on the practices of parties, the interaction between the two has gone largely unstudied. Scholars of parties have not taken adequate account of the impact of gender quotas on party behaviour, while scholars of quotas have focused on the normative arguments and practical consequences of quotas rather than considering the central role of parties. This book fills this gap in knowledge by illustrating how quotas affect parties, and using theories of party behaviour to explain whether, why and to what extent parties will implement gender quotas. The case study of France is particularly pertinent, as France was the first country in the world to introduce a ‘parity’ quota – that is, a quota set at 50 per cent. Despite the global interest triggered by France’s attempt to become the world’s first parity democracy, the reality did not live up to the hype. Parity produced very variable results from one party and one election to another. In addition, although the ‘parity’ law is compulsory, it is enforced in different ways for different elections. The introduction of a compulsory quota for all elections means that France illustrates the complexity of quota implementation across a 1

2 Parties, Gender Quotas and Candidate Selection in France

variety of electoral systems and levels of government. Meanwhile, France’s varied party system allows parity’s impact on a number of different parties to be observed. This makes France a fertile testing ground for exploring the causes and consequences of these variations, and what we can learn about party responses to gender quotas as a result. This book builds on the insights provided by parity in France and brings together knowledge from the literatures on political parties and gender and politics in order to provide a comprehensive account of the interaction between quotas and parties. The first section of this chapter will explore the rationale behind the book in greater detail. Firstly, it will briefly explain the significance of gender quotas to democracy. The chapter will then explain how parties are a critical actor in the study of gender quotas, looking at the twoway relationship between quotas and parties. The specific case study of France will then be explored in more detail, before introducing the framework used for analysing the mutual relationship between quotas and parties. Parties are considered as rational actors who incorporate the demands of gender quotas into broader organisational, ideological and electoral imperatives. These different imperatives are explored using gendered approaches to candidate selection practices along with three mainstream theories of party organisation, namely the electoral competition approach, the institutional approach and the ideological approach. An overview of the core arguments of each of these approaches is offered. A central argument made is that each approach is useful but incomplete when considered in isolation from the others, resulting in the creation of an overarching model which is the first to provide a comprehensive illustration of how parties incorporate gender quotas into their broader strategies. The chapter concludes with an overview of the book, illustrating how the core ideas are organised and tested.

Why gender quotas? Women’s under-representation in politics is a global phenomenon, with a range of negative repercussions. These include an incomplete democratic process, where women’s viewpoints are overlooked; an injustice against the majority of the world’s population; an inefficient use of resources, with women’s talents going to waste; and a symbolic sense of exclusion for women voters and citizens. Attempts to redress women’s under-representation, such as positive action and a gradual evolution of societal attitudes towards women, have contributed to the

Introduction 3

gradual rise in women’s political presence around the world, but the rate of change produced by such measures is frustratingly slow. Quotas are widely advocated as a more rapid and effective means of increasing women’s presence in politics. Gender quotas have grown rapidly in popularity over the past 20 years, and there are now at least 100 countries around the world using some form of gender quota (Quota Project; Krook 2009). Quotas are increasingly seen as a ‘fast track’ solution to women’s under-representation, compared to the slower, incremental model of societal change (Dahlerup 2006; Dahlerup & Freidenvall 2005). In countries where attitudes and/or institutions are slow to adapt to a feminisation of politics, quotas may be the fastest means of increasing women’s numerical representation. However, it is acknowledged that not all quotas have the same impact. There are many variations of gender quota – for example, some are compulsory and state-wide, whilst others are introduced voluntarily by political parties. France is one of 15 countries to have a constitutional quota for its national parliament, and one of 45 countries to have national quota legislation (Quota Project). Some quotas are reinforced with penalties for non-compliance, whilst others are symbolic and reliant on parties’ will to implement change. Some quotas reserve an exact number of seats for women, while others focus on the proportion of women candidates, with or without a measure to ensure that women are placed in winnable constituencies or in winnable positions on party lists. Finally, quotas may be set at any level up to and including 50 per cent. The rapid influx of quotas around the world is starting to transform politics in a number of ways. The proportion of women in parliaments of countries where a quota is used has usually risen, sometimes dramatically. Parties using quotas have had to adapt their candidate selection practices in order to recruit and promote higher proportions of women. Men have sometimes found themselves forced to step aside in order to permit the political renewal of quotas. Symbolically, the concept of women politicians has been normalised to a greater extent. However, quotas are certainly not a panacea to the problem of women’s under-representation. They have courted a lot of controversy, both from opponents of women’s representation (who consider quotas to be unmeritocratic, discriminatory against men and a perversion of the democratic process) and feminists (who consider quotas to be essentialist and patronising, and fear that election through a quota will undermine a woman’s authority and credibility as a politician). In addition, quotas have not always been successful. Sometimes they have been introduced as a symbolic gesture, with insufficient measures to

4 Parties, Gender Quotas and Candidate Selection in France

ensure their effectiveness; for example, Brazil has a national quota law requiring 30 per cent women candidates, yet its parliament still contains only 8.8 per cent women (Quota Project). In many instances, the target proportion of women has not been met, and even more rarely has it been exceeded. Quotas are therefore interesting for a number of reasons. Firstly, their potential transformative effects for democracy are considerable. Secondly, their impact is variable and their success is not guaranteed. Thirdly, their relative newness means that research on quotas is a rapidly evolving field of study, with rich potential for further exploration of the impact of quotas and explanation for their variable success.

Why parties? One of the major actors concerned with gender quotas is political parties. They are responsible, either individually or collectively, for the introduction of quotas. They are the actors charged with quotas’ implementation, and it is parties who determine whether or not quotas will succeed. In most democracies, political parties are responsible for political recruitment and candidate selection. In many respects, then, parties are pivotal actors in explaining women’s under-representation and the effectiveness of quotas in increasing women’s presence. Parties are critical to understanding quotas, and quotas are also an increasingly important aspect of the study of parties, as more and more parties around the world choose or are forced to adopt quotas. Quotas affect many aspects of a party’s life, from its candidate selection practices (which also link to its relationship with its members) through to the composition of the parliamentary party. In the instance of compulsory quotas such as those in France, there are also financial repercussions. The effects of quotas on parties (and vice versa) may be profound, but are heavily under-theorised. Studies of gender quotas have concentrated primarily on quotas’ effects on other aspects of the political process, such as individual candidates, legislatures and voters, rather than parties. Meanwhile, the study of political parties remains largely gender-blind and has not taken sufficient account of the radical upheaval of political personnel generated by gender quotas. Let us look at each of these areas in more detail.

Introduction 5

There is a rapidly expanding literature looking at women in politics, and at quotas in particular. Mansbridge (1999) and Phillips (1991, 1995) are amongst many authors that make convincing theoretical arguments to justify the use of gender quotas in order to increase women’s representation. Other works, such as those by Childs (2004), Childs and Withey (2006), Dahlerup (1988), Dolan (1997), Franceschet and Piscopo (2008), Lovenduski (2005), Lovenduski and Norris (1996), Mateo Diaz (2002, 2005) and Meier (2000), have explored the impact that women politicians can make once elected, offering a more empirically-based argument in favour of women’s presence in politics. A recent edited volume by Dahlerup (2006) provides an overview of gender quotas from a variety of perspectives, including theoretical arguments for and against quotas, and detailed case studies of quota implementation in a range of countries. Krook has looked both at theoretical aspects of quotas and at comparative case studies around the world (Krook 2004, 2006, 2007a, 2008, 2009). Despite this growing literature on quotas, there are surprisingly few works that look directly at the role of political parties in women’s representation. Baer laments that ‘the concept of political party in women and politics research is missing where one would most expect it – in studies of recruitment and public office holding’ (1993, p. 562). She argues that the costs of this ‘myopia’ include a misunderstanding of parties by scholars on women in politics, a hostility towards women by scholars of political parties, and a continued failure to understand why parties select more men than women. Since then, a few authors such as Kittilson (née Caul) (1999, 2001, 2006), Lovenduski and Norris (1993) and Opello (2006) have written explicitly about the relationship between political parties and women candidates, and their influential works are referred to in greater detail throughout this book, but the majority of writings focus on whether women should be present in politics, the extent to which they are present, institutional and sociological variables that influence the likelihood of their being present, and what they do once they do succeed in getting elected. While all of these are essential questions that need to be answered, it is important not to forget the instrumental role that political parties play in the recruitment, retention and promotion of women politicians. However, turning to the literature on political parties for a clearer understanding of how they have responded to demands for feminisation provides only limited insights, as most works on political parties suffer from gender blindness. Many of the classic works on parties make no reference at all to women (such as Blondel 1978;

6 Parties, Gender Quotas and Candidate Selection in France

Downs 1957; Michels 1915; Sartori 1976; von Beyme 1985), and those that do tend to make only scant reference to women without integrating gender more comprehensively into their work (such as Epstein 1967; Katz & Mair 1995; and Webb et al 2002). However, I argue that no work on political parties can be complete unless it takes questions of gender into account. Gender has become an increasingly important variable for parties to take into consideration over recent decades, even prior to the introduction of quotas. Since the granting of female suffrage, parties have needed to be mindful of the needs of women voters. Well-documented studies of the ‘gender gap’ have demonstrated that women do not always vote in the same way as men, and shifts in the gender gap from right to left demonstrate that women’s votes are susceptible to influence (Achin & Lévêque 2006; Campbell 2006; Inglehart & Norris 1999; Mossuz-Lavau 2007; Norris 1999; Sineau 2006, 2007). Parties have learned that they cannot ignore female voters; for example, Mitterrand’s victory in the presidential elections of 1981 was attributed to his ability to appeal to female voters, who had previously been more right-leaning than their male counterparts (Helft-Malz & Lévy 2000, p. 38). Indeed, one of the reasons for introducing gender quotas has been as an attempt by parties to woo women voters (and appear progressive). These quotas have had a substantial influence on candidate selection procedures, and no account of political parties is complete without explicit acknowledgement of the changes to party policy and practices that feminisation has introduced. Although this book is the first to use classic approaches to party behaviour to explain quota implementation, it follows in the tradition of a small but distinguished set of works seeking to feminise the literature on political parties. Ware is one of the few authors to provide a more inclusive account of gender in his own work on political parties (1996). A work that actually dedicates itself to discussing the relationship between women and political parties is Lovenduski and Norris (1993). In the final chapter, Norris explores factors affecting parties’ relationships with women, including the institutional environment as well as the impact of different types of internal party organisation. She provides an initial suggestion that low internal democracy permits (but does not oblige) a party’s leadership to introduce more women candidates. This idea is explored in greater detail in Chapter 3 of this book. She also begins to develop ideas of supply and demand, and of parties as gatekeepers, that are explored further in Norris (1997) and Norris and Lovenduski (1995); these ideas are also tested and developed here in Chapter 3. Finally, she considers the role of ideology in parties’

Introduction 7

willingness to recruit women, an idea that is developed in works by Kittilson (née Caul) (1999, 2001, 2006) and Opello (2006) and examined in detail in Chapter 6. Much of the existing theoretical and empirical literature on the relationship between women and parties takes inspiration from this important volume. Within more specialist works on candidate selection, the landmark volume by Gallagher and Marsh (1988) considers gender along with a large number of other factors, and provides some interesting insights, although many of these are drawn from the literature on gender and politics. For example, they note that women tend to fare better under proportional representation systems due to the need to balance tickets, rather than under single member plurality systems where ‘selectors might feel, accurately or otherwise, [that women] could be electoral liabilities’ (1988, p. 260). However, they add that ‘while a PR electoral system obviously makes the selection of women more likely…the active role of the selection process in determining how many women enter Parliament should not be overlooked’ (ibid, p. 268). They also acknowledge the role of quotas in facilitating women’s entry, and the role of incumbents in impeding this. Norris and Lovenduski (1995) develop an important theoretical framework including the ‘supply and demand’ model, a more advanced and sophisticated version of the ideas referred to above by Norris (1993). They also seek to model the stages at which women face barriers in the recruitment process in order to explain how a large pool of potential candidates is reduced to a small pool of elected women. Some of these theories are applied in a more comparative context in Norris (1997). Together, these works offer a number of revealing insights into the gendered nature of the recruitment process that form foundations on which this book builds. However, none of these works relates specifically to the interaction of parties and gender quotas. Opello (2006) is an exception to this rule, and her work is considered in greater detail below. Overall, there is much to be learned from existing studies of parties and of gender quotas, but there is a dearth of theorisation as to how the two interact. This book aims to draw on the strengths of the literature described above, and to end the false dichotomy between research on gender and research on political parties. Indeed, it is a central tenet of this book that neither area of research can be complete unless it incorporates elements of the other. As parties are often responsible for introducing quotas as well as being the actors charged with their implementation, understanding their motivations for supporting (and, in some cases, resisting) quotas needs to be incorporated into any

8 Parties, Gender Quotas and Candidate Selection in France

study of gender quotas. Similarly, as quotas become more widespread, no study of political parties can be complete unless it takes into account the multiple impacts that quotas have on political parties. These include the integration of quotas into parties’ candidate selection practices, and the wider repercussions for other aspects of their organisation, such as renewal, internal democracy, unity, electoral appeal and financial stability.

Why France? France was the first country in the world to introduce a compulsory 50 per cent (‘parity’) gender quota for all parties in almost all elections. As discussed below, this has made France a fascinating case-study for scholars and practitioners alike. France may be a pioneer in terms of ambition, but its track record is rather more mixed when it comes to the successful implementation of parity. The variations of parity’s implementation from one party to another and from one type of election to another add to France’s intrigue. In addition, French parties have imposed large fines upon themselves by introducing a law which they could or would not implement. For all these reasons, France is an ideal case study of the interaction between parties and quotas in different types of election. The French parity law rose under unusual circumstances, considered in greater detail in Chapter 2. The driving motivation for a ‘parity’ quota, set at 50 per cent, was to overcome the previous legal rejection of gender quotas on the grounds that they ran contrary to the constitutional principle of universal citizenship. A quota which reflected the natural distribution of the sexes within society was based as much on an attempt to make quotas acceptable within the context of French universalist thought as it was on more principled grounds of equality and fairness (although it is on these latter arguments that other countries considering a 50 per cent quota have focused). While certain elements of ‘parity’ were specific to France, its origins were also located in an international context (Krook 2007b), with discourses on parity originating from the European Union, and France comparing unfavourably to her European neighbours in terms of women’s representation (prior to the law’s passage in 2000, France had fewer women in the lower chamber of Parliament than any other member state of the European Union except Greece1). The ‘parity’2 law specifies that French parties must field an equal number of men and women candidates to all elections. Since its intro-

Introduction 9

duction, the law has had a very variable effect, depending both on the party and the type of election concerned. At its most successful, the law induced near-parity in the composition of local councils, with 47.5 per cent women elected in 2001 and 48.5 per cent women elected in 2008 where the law applied (Fabre, Le Monde 27 March 2001a; Zimmerman 2008a). However, the law has been less successful at other levels, and its biggest failure has been in the legislative elections, where the proportion of women elected to the National Assembly rose from 10.9 per cent in 1997 to just 12.3 per cent in 2002, followed by a moderate improvement to 18.5 per cent in 2007. This variable performance between elections is interesting in itself, and the legislative elections are of particular interest for a number of reasons. Firstly, the overall percentage of women candidates tends to exceed greatly the percentage of women elected (in 2002, there were 38.9 per cent women candidates compared to 12.3 per cent women elected; in 2007, there were 41.6 per cent women candidates for 18.5 per cent women elected (Zimmerman 2003a, p. 10; Zimmerman 2007b, p. 4)). Secondly, there was huge variation between the different political parties, with parity being respected by some parties even though they were opposed to it in principle (such as the far-right Front National (FN)), while some parties, including the two main parties of the right (the UMP and the UDF) fielded fewer than 20 per cent women in 2002, and well below their target of 30 per cent women in 2007. Thirdly, the parties that did not respect parity lost millions of euros as a financial penalty for breaching the law. This raises two core questions: why parity had a variable impact from one election and from one party to the next; and why parties would consider it preferable to lose a large sum of money rather than to select more women candidates. Both of these questions are considered throughout the book. There are also several institutional factors which make France a particularly rich case-study for examining parties’ roles in quota implementation. Firstly, the ‘parity’ legislation is unusual in that it is constitutionalised and enforced upon all parties. This means that every single party in France is subject to its constraints, regardless of their beliefs and circumstances. Secondly, France has a variety of electoral systems depending on the election in question, and as a consequence, parity is applied in a range of different ways. This allows for a fruitful exploration of the efficiency of the law under different circumstances. Thirdly, the relatively large party system enables comparisons to be made between different types of party within the same country. Unlike other countries, where gender quotas have been voluntary and

10 Parties, Gender Quotas and Candidate Selection in France

piecemeal, France provides the opportunity to compare the impact of the ‘parity’ law across all parties and virtually all elections.3 Given that a focus on a single-country case study inevitably results in a small n, I deploy the ‘comparable cases’ approach of Lijphart (Collier 1993; Lijphart 1971, 1975). He argues that in small n studies, it is useful to focus on cases that are ‘similar in a large number of important characteristics (variables) which one wants to treat as constants, but dissimilar as far as those variables are concerned which one wants to relate to each other’ (Lijphart 1971, p. 687). By controlling here for shared institutional variables such as the electoral system, it is possible to focus on the variables that explain why different parties will react differently to the same situation – in this case, why some parties are more likely than others to implement the ‘parity’ law. Looking at different parties within the same system is the most useful methodology as it is their different behaviour under similar circumstances that is of theoretical interest. France also permits comparisons across different elections, allowing an assessment of whether and how parties adjust their strategy according to the status of the election in question and the electoral system in play. Understanding what effect different independent variables (the pragmatic and ideological concerns of parties and the institutional environment) have on the dependent variable (the successful implementation of a quota) is illuminating both within and beyond France. Within France, a much clearer understanding of how parity works may be gained, with the benefit of understanding why the law does not work perfectly and how it might be improved. Broader comparative insights into the benefits and obstacles of compulsory quotas may also be obtained, providing an idea for those researching or contemplating the introduction of compulsory quotas outside France as to how effective they might be under different circumstances. France is central to discussions of gender quotas (see, for example, Krook 2009, forthcoming; Lovenduski 2005; Stevens 2007), both for the audacity of its 50 per cent target and its inability to come close to reaching this target. Similarly, countries such as Spain and Italy have drawn open inspiration from the French example when debating quotas in their own country. Scholars have even come in from other disciplines such as economics in an attempt to explain the seemingly curious case of a country where parties introduced such an ambitious quota without appearing to have genuine commitment to gender equality (Fréchette et al 20084; Pemstein and Bernhard 2005). However, the existing literature on parity in France has been fairly silent on the

Introduction 11

role of political parties. This literature has focused on two distinct directions over the past 15 years – the case for parity, followed by analysis of its implementation. Both these trends are considered below, as they are illuminating for the arguments made within this book. Nonetheless, the lack of attention paid to parties within these works means that a central explanatory variable is missing from the analysis. Works written prior to the introduction of the parity law focused on philosophical and theoretical assessments of parity, with only limited assessment of the practical applicability of such a law. Some works (for example, Gaspard 1997; Gaspard et al 1992; Halimi 1994; MossuzLavau 1998) were unambiguous or even prescriptive in their support for parity, and were intended to persuade others of the need to introduce parity. The principal arguments put forward included the need for Parliament to reflect the gendered composition of society (although, interestingly, other notions of descriptive representation such as race were explicitly rejected in order to distance claims for parity from the criticisms that it was divisive and would split society into communities – a notion which runs counter to republican universalism). Parity was perceived as the best means for achieving the equal representation of men and women, firstly because of the constitutional rejection of gender quotas,5 and secondly because any quota below 50 per cent was perceived as inadequate and as having the effect of a ceiling. It was argued that only a demand for parity in principle would lead to the achievement of parity in practice. Moreover, they claimed that such a measure was necessary due to the invisible bias of the state towards men, with the seemingly gender-neutral ideal of the universal citizen concealing the underlying disadvantages to women that restricted their access to the political arena (Halimi 1994, pp. 14–18). Women’s inferior socio-economic status, the burdens of domestic life and women’s reluctance to put themselves forward as candidates were considered by some to be the primary causes of women’s underrepresentation (Halimi 1994; Roudy 1995; Sineau 1988). For others, France’s political institutions, such as the majoritarian electoral system, the ability to cumulate electoral offices, and the centralisation of power were seen as the most important explanatory variables (Guigou 1997, p. 85). Although the role of parties was not considered a central argument, a couple of party-related factors were highlighted, such as the difficulty posed to women by the preference for (male) incumbents, and the fact that women formed a minority of party members and an even smaller proportion of party officials. Parties were generally portrayed as anti-women, with the argument that men resisted women’s

12 Parties, Gender Quotas and Candidate Selection in France

entry into politics because ‘giving a place to a woman means taking it from a man’ (Dewavrin in Halimi 1994, p. 132). Sineau further argued that parties’ interest in women was limited to women voters and did not extend to women candidates (Sineau in Gaspard 1997, p. 113). In a similar vein, the Socialist minister Elisabeth Guigou was quick to condemn the sexism that was rife within her party and within the political arena in general (Guigou 1997). This sexism was simultaneously seen as an explanation for why women found it difficult to succeed in politics, and for why women were turned off politics in the first place. However, only Janine Mossuz-Lavau went as far as to claim that parties were the primary reason for women’s under-representation, arguing that they were ‘male cliques operating in closed circles, reproducing identical copies of themselves, and not being prepared to take a seat away from a man in order to give it to a woman. All accounts by female politicians heavily emphasise the fact that parties mainly agree to field women as candidates when the constituencies in question are deemed to be already lost’ (Mossuz-Lavau 1998, pp. 24–5). It was therefore argued that it was necessary to force parties to field more women, because otherwise women’s representation would not increase (Gaspard 1997, p. 32). In addition to the feminists advocating parity, a vocal minority of feminists added their voices to the wider philosophical objections to the introduction to parity. Authors such as Badinter, Ozouf and Pisier argued that parity was patronising, ineffective (because it focused on the symptom rather than the cause), and risked essentialising women and defining political candidates by their descriptive attributes rather than their beliefs or other qualities (Amar 1999). This final argument was linked to the concept of the universal citizen, which they argued was threatened by any attempt to gender citizenship. Despite their feminism, their arguments echoed those used in other countries to oppose women’s representation (Dahlerup 2006, pp. 297–9). The debate focused on philosophical questions, with pragmatic issues being limited to the recognition of the practical impossibility of pure descriptive representation. Parties did not play a central role in the debate. Following the introduction of the parity law, the focus of parity scholars shifted away from the abstract pros and cons of parity and moved towards a practical assessment of parity’s performance. Key amongst these works is a series of reports issued by the Observatoire de la Parité, the official government body appointed to monitor parity (Genisson 2002; Zimmerman 2003a, 2003b, 2004, 2005a, 2005b, 2007a, 2007b, 2008a, 2008b, 2009). As the initial successes of the law in local

Introduction 13

politics gave way to disappointment at the national level, the analysis of parity began to discuss the weaknesses in the law. In addition to the detailed discussion of results and case studies (for example Latté 2002; Sineau 2004; Troupel 2002, 2006), works began to evaluate how, and why, parity had not delivered all that it had promised (for example Achin et al 2007; Bird 2004; Sénac-Slawinski 2007, 2008). Even after this shift, the literature still focused primarily on women. There has been little discussion of the motivations for political parties in deciding whether and how to implement the law. Rather, parties have been demonised as being anti-women, without explicitly theorising whether this is in fact the case, and if so, why. Interviews conducted with parties by the Observatoire de la Parité (Zimmerman 2003a) provided some useful insights, but were presented factually without the benefit of a more critical lens. Pionchon and Derville (2004) and Zimmerman (2003a) both give some consideration to the obstacle posed to parity by incumbency. Karen Bird was one of the first to consider party motivations explicitly when implementing parity: …the major parties – those who are real contenders to form the government – find themselves struggling to balance three basic political incentives: to preserve the party’s existing electoral strongholds, and conquer new ones; to present a maximum number of female candidates and therefore minimise the financial penalty; and to avoid alienating (male) incumbents and hopefuls who could run as dissidents against the party’s official candidate, potentially splitting party support and increasing the probability of victory for opposing parties. (Bird 2004, p. 249) The only volume to date which combines an in-depth study of parties with a study of French parity is Opello (2006). Opello provides detailed and valuable insights into the role of women within the PS and the UMP, using a historical approach to provide a strong background to the introduction and initial implementation of parity. Like other works on parity, however, her book focuses more on the role of women within parties than on the perspectives of parties themselves, and the in-depth study of two parties limits the opportunities for more general comparison. A pattern emerges from the existing literature of parties struggling to juggle a number of competing priorities, some of which favour the implementation of parity, and some of which render it seemingly impossible. Some of these priorities are mapped out and tentatively

14 Parties, Gender Quotas and Candidate Selection in France

explored. Nevertheless, the emphasis remains heavily on the perspectives of women, and their attributes, struggles and experiences. The discussion of parties’ perspectives remains tentative and undertheorised. In particular, little attempt is made at measuring how a compulsory quota such as parity interacts with other elements of party organisation and ideology. Furthermore, the literature does not account adequately for the differences between different political parties. It is frequently noted that parties of the left tend to be more favourable to parity than parties of the right (as reflected by the number of candidates fielded by different parties in 2002 and 2007), but this is a wholly insufficient explanation of variation between political parties. Nor do arguments concerning incumbents suffice to explain this variation. Moreover, the interplay between different explanatory variables, and the way that different parties prioritise between them, is not considered. This book completes the picture by focusing on the perspectives of parties to explain why gender quotas are both important and difficult for parties to implement, using the full range of major parties in France.

Parties as rational actors In order to explain how parties respond to the demands of gender quotas, this book operates on the premise that parties approach quotas as rational actors whose behaviour is driven by their own needs and preferences. That parties seek to behave rationally does not necessarily mean that they succeed in doing so; their rationality may be bounded by incomplete information, inaccurate perceptions or poor choices between conflicting options. However, it is argued that parties will still seek to maximise their utility to the best of their ability, and that seemingly irrational choices will stem from bounded rationality rather than irrational behaviour – parties cannot control what they do not know, but will act as rationally as possible within the boundaries of what they do know. It is also possible that parties will make what appear to be suboptimal choices in order to satisfy an underlying strategy – a concept described by Tsebelis (1991) as ‘nested games’. For example, the decision of right-wing parties to support the passage of the parity law may appear to be a sub-optimal decision as their failure to implement it resulted in heavy financial losses. However, another way of viewing the situation is that they supported an electorally-popular piece of legislation that placed greater restraint on candidate selection for rival parties (for ideological and/or financial reasons) than it did on

Introduction 15

them, resulting in their electoral victory and consequent financial gains that exceeded any loss from failure to implement parity. When the seemingly suboptimal behaviour is placed within the wider context of party preferences and strategy, it becomes a more obviously rational decision. Locating individual decisions within the underlying gameplan of each party is one of the goals of this book. Moreover, rational choices for each party do not necessarily amount to the same thing, as for one party it will be rational to seek to maximise electoral appeal while for another party it will be rational to remain faithful to the party’s core support. Given that objectives may vary from one party to the next, and may never be either transparent or achieved, it is easy to argue that any claims of rationality are unrealistic and, in any case, impossible to prove. But even when party motivations are not clear to the observer, or do not appear to be logical, it is unsafe to assume that they are irrational. A party that makes a costly decision may still have made the best decision available to them, if they deemed an alternative strategy to be even more costly. For example, a party that chooses to forfeit a large portion of its state funding in order to field more male than female candidates might still be acting rationally if fielding more female candidates would impose an alternative cost to the party that was deemed to be greater. Likewise, a party which chooses to pursue a radical ideology that alienates it from most of the electorate will be acting rationally if ideological purity is more important to the party than electoral success. If rationality is seen as the process of knowing one’s goals, being able to rank them in order of preference and then taking the optimum path for the fulfilment of those goals, then knowing the goals – and, crucially, their preference ranking, which may vary from one party to the next – is a vital step towards knowing what path a rational party should take. This book assumes that parties will make the optimal decisions possible within the constraints of exogenous circumstances. This leads us to assume that each party’s decision on how many female candidates to field was the optimal decision for that particular party, based on the knowledge and resources available to the party at the time. From this we can infer that parties who do not implement parity have good reasons for not doing so, and that the choices they make are indicative of their preferences within the given circumstances. This then allows for the modelling of party preferences, and for an analysis of the extent to which the parity law succeeds in overriding other preferences. Illustrating how quotas interact with other party preferences is critical

16 Parties, Gender Quotas and Candidate Selection in France

to understanding whether, why and how parties will implement quotas.

From multiple models to one This book pulls together disparate theories present in the existing literature on gender and politics and on political parties to provide a more unified theory of the way that political parties respond to gender quotas. Although there is much to be learned from each of the existing elements, none is sufficient in itself to provide a holistic explanation of whether, how and why parties will implement quotas. This book is the first to bring these elements together into a more comprehensive model that accounts both for the priorities of parties and the particular demands (normative, electoral and organisational) presented by gender quotas. The book begins in the next chapter by providing the necessary contextual information about the French party system and the parity law required to understand how this case study illustrates the broader themes explored in the book. The subsequent chapters then draw on theories from the literature on gender and politics and on political parties. Although the third chapter primarily builds on work by feminist scholars, with subsequent chapters based on mainstream theories of party organisation, there is constant interlinking between the two fields. Indeed, there is much that is complementary and overlapping in these works, and this book makes more explicit how different ideas can come together to explain parties’ response to gender quotas. For example, Chapter 5, which is based on institutional approaches to party behaviour, draws on works demonstrating how some electoral systems are more favourable to women’s representation than others. Chapter 6 brings together ideas from the literature on party ideologies with work demonstrating how left-wing ideologies tend to be more supportive of women’s claims for inclusion. The symbiotic relationship between research on parties and research on gender has long existed; the purpose of this book is to make the connection more explicit and unify these ideas into a more holistic framework. The key ideas drawn from the gender and politics literature relate to candidate selection practices, considering firstly how a party’s internal organisation might affect its ability to implement a gender quota, and secondly how gender influences the availability and qualifications of aspirant candidates as well as their perceived desirability in the eyes of parties.

Introduction 17

The book then draws upon three key theoretical approaches from the literature on parties to demonstrate how useful they are in explaining party responses to gender quotas, namely the ‘electoral competition’ approach, the ‘institutional’ approach, and the ‘ideological’ approach. Each of these approaches provides some degree of insight as to why parties select the candidates that they do. Various tests are deployed, using French parties as a case study, to measure the utility of each theoretical approach in explaining parties’ choice of electoral candidates. It is a central hypothesis of this book that the reasons why parties do not select more female candidates for legislative elections are currently over-simplified and are in fact numerous, complex and diverse. Studying competing theories of party organisation and behaviour is a way of reaching a better understanding of candidate selection. However, given the complexity of party behaviour, the argument is made throughout the book that although each of these theories offers a partial explanation of party candidate selection, none of these theories is complete and adequate in itself. Each theory helps explain party priorities through a particular lens, but not all parties will have the same priorities. Although most parties will share similar preoccupations, they may order them differently. For some parties the pursuit of electoral office may be paramount; others will be more motivated by ideology, whilst others will be motivated by practical imperatives such as finances. If we assume that parties are rational actors who will evaluate candidates in accordance with their priorities, then the difference in priorities from one party to the next is the key to explaining variable outcomes. Using this concept of variable party priorities, I argue that it is possible to synthesise the approaches to party behaviour and create an overarching model that accounts for the variable impact of the parity law depending on the party and the election in question. The individual approaches, as well as the prospects for a more comprehensive model, are considered in more detail below. The ‘competition’ approach, as defined by the likes of Downs, Kirchheimer and Sartori, suggests that parties are motivated above all else by electoral success. This implies that parties would select the candidates most likely to win, irrespective of other factors. Given parties’ preference for incumbents (usually male), this theory is examined in two ways. Firstly, a profile is compiled of the types of candidate that parties are most likely to support, and secondly, a longitudinal study of electoral performance is used to determine whether these types of candidate (for example male, incumbent) do in fact perform better at the polls than their counterparts. A point of interest is to see whether

18 Parties, Gender Quotas and Candidate Selection in France

parties have an accurate perception of voter behaviour, or whether they are guilty of imputed discrimination. If parties seek electoral success and select the candidates that are the most winnable, then they are acting rationally. However, this rationality may be bounded if parties are ignorant of the electorate’s preferences, or if they are guilty of ‘bolstering’ by persisting in their belief that certain types of candidate are best, regardless of any evidence to the contrary (Ward 2002, p. 81). The ‘institutional’ approach is defined here as arguing that the choices that parties face when selecting candidates are framed by the institutional environment. Some institutionalists, such as Panebianco or von Beyme, place a greater emphasis on the institutions in place when parties were created, while others, such as Duverger, believe that parties adapt to the changing institutional environment. This book supports the latter argument and focuses primarily on contemporary institutions to see whether and how they structure party behaviour and whether parties are able to adapt to them. Particular reference is made to the ‘rational choice’ stream of new institutionalism (Peters 1999) to assess how parties attempt to maximise their utility within the particular institutional environment at any given point. This approach is tested by examining the impact of various institutions, such as the electoral systems used for different elections, on party candidate selections. The detail of the law on parity affects the way in which the law is implemented, hence the way the law is framed is considered as an institutional variable. Decisions on whether or not to implement parity can then be explained within the context of the choice-set made available to each party by the institutional environment. Some parties may have more options to choose from than others. This approach explains how the institutional environment can influence party behaviour. For example, institutional factors can explain why the same parties will select far more female candidates for some elections than for others, and can also explain why some parties behave in ways that cannot be predicted by their ideology (such as the anti-feminist Front National, who respected the parity law). Although its explanatory power is incomplete, it provides a valuable complement to the other two approaches. The ‘ideological’ approach is drawn from the ‘sociological’ approach of Lipset and Rokkan, which argues that parties’ behaviour is dictated by their social origins and ideology. In this context, the approach represents the idea that parties’ willingness to implement the ‘parity’ law can be linked both to whether or not they agree with the aim of

Introduction 19

increasing women’s presence in politics, and whether or not they believe that the ‘parity’ law is the appropriate means to achieve this aim. The ideological approach can also be linked to the electoral competition approach in two regards. Firstly, parties that are based on a strong ideology are more likely to be influenced by their ideology and less likely to compromise their ideals in the pursuit of electoral office, whereas catch-all parties are less likely to be based on a rigorous and inflexible ideology (Kirchheimer 1966; von Beyme 1985). Secondly, a party that espouses egalitarian ideals might be more likely to believe that its electorate would not discriminate between a male and a female candidate than a party that does not support gender equality (either through opposition or indifference). The approach is tested through consideration of parties’ ideological foundations and the stances that they have taken regarding the ‘parity’ law. This approach reinforces the notion that ideology may play a role over and above other motivations, thus explaining why some parties are willing to make greater sacrifices than others in order to implement parity. This is more likely to be the case when parties have ideological support for parity; a party lacking ideological support may be more motivated by other factors. In addition, support for parity depends on agreement with not one but two positions, namely agreement both with the goal of equal representation for both sexes in politics and the use of state intervention as the means to achieve this goal. For all of the theories in question, the parties are studied comparatively, with the primary focus being on French parties but with some comparisons made to other parties belonging to the same party families. As Knapp (2004) argues, French parties do not map perfectly onto von Beyme’s model of familles spirituelles (a framework of nine party ideologies into which most European parties can be placed, facilitating international comparison). Nonetheless, nearly every party family is represented within France, and six different types of party are considered in this book (the communist PCF, the socialist PS, the ecologist Greens (les Verts), the centre-right UDF, the Gaullist/ conservative UMP and, to a lesser extent, the far-right FN). The book also draws on both quantitative and qualitative methods as appropriate, rejecting the dichotomy between these methodologies in favour of King et al’s assertion that ‘most research does not fit clearly into one category or the other. The best often combines features of each’ (1994, p. 5). In this book, a question raised by the electoral competition approach is whether women’s poor electoral performance is a result of voter or party discrimination. To answer this question, a large

20 Parties, Gender Quotas and Candidate Selection in France

scale quantitative study of election results is used to evaluate whether women incur an electoral penalty or whether they are simply placed in the toughest seats. This study draws on statistical analysis of data from five elections, and the results are reported in Chapter 4. By contrast, an understanding of the intricate workings of candidate selection is much more suited to a qualitative approach of elite interviews. I interviewed party officials responsible for candidate selection (invariably men), as well as senior women within each party, to determine which selection processes are used and how fair the different actors within the process believe that process to be. The different viewpoints expressed from the two sides of the selection process provide insights into the strengths and limitations of each party’s selection process, as well as controlling for bias by ensuring that a range of viewpoints were considered. The insights obtained from these interviews are explored throughout the book and they contribute towards building theories of the selection process that are sensitive to parties’ perspectives. Thus, through a combination of the most appropriate qualitative and quantitative approaches, this volume is able to address its research questions in a holistic and systematic way. When the different methods and theories are brought together, a pattern emerges of each of the three main theoretical models’ strengths and weaknesses which is then deployed to provide an evaluation of the three approaches. A core argument made throughout the book is that each of these approaches provides useful but insufficient insights into the variable implementation of parity. In order to provide a more complete account of the gendered nature of candidate selection and of the differences between parties, I argue that it is necessary to combine these insights into a synthesised or ‘hybrid’ model. The principle on which this model is based is that parties face a number of conflicting choices when selecting electoral candidates, and their overarching motivations influence how they will decide amongst these choices. The institutional environment frames the range of choices available, while a party’s ideology, stability and its electoral ambitions will determine how it prioritises certain courses of action over others. The three theoretical approaches that are explored in the following chapters, along with insights from the literature on gender and politics, provide the building blocks upon which the concluding theoretical model is based. This model accounts both for how parties in France have responded to parity and how parties more broadly might incorporate a gender quota into their candidate selection processes.

Introduction 21

Overview of the book The final section of this chapter provides an overview of the remainder of the book, demonstrating how the themes discussed above will be organised and explored. Chapter 2 introduces the French party system and the background to the parity law. This chapter provides the contextual information on which the rest of the book builds. The key traits and history of each of the main parties are introduced, providing insights into the ideology, strength, stability and tactics of each party. Parties are also placed within the comparative framework of ‘party families’ in order to facilitate comparisons between France and other countries (von Beyme 1985). The second part of the chapter explores the origins of the parity law and its development into legislation, with a particular focus on the attitudes of parties towards the law. The motivations for each party in supporting the initial legislation are indicative both of the way in which party traits fed into policy formation, and the ways in which this influenced subsequent implementation of parity. The findings suggest that some parties were genuine in their support for parity, while others were more pragmatic in their outlook. This pragmatism explains why the same party might choose to support the passage of a law that it did not intend to implement, and the argument is made that this seemingly inconsistent behaviour might actually be consistent with the party’s underlying goals. For example, if a party seeks to win votes, it may support the introduction of parity as a popular policy, but may not then implement the law if it is concerned that doing so would harm its electoral prospects. This concept is explored throughout the book as a measure of whether parties’ behaviour is driven by conviction or pragmatism. The question is also raised of whether the loopholes in the legislation were a deliberate strategy by parties to give the impression of supporting women’s representation whilst allowing themselves the freedom to maintain the status quo. Chapter 3 considers candidate selection practices and draws on theories from the literature on gender and politics to explain how gender interacts with candidate selection. The first part of the chapter builds on the argument made by Lovenduski (1993) and Norris (1997) that the degree of centralisation of a party’s candidate selection processes may affect a party’s ability to implement a gender quota. The hypothesis is put forward that centralisation is a necessary but not sufficient condition of successful quota implementation. Parties are more easily able to ensure compliance with parity across the country if the process is centralised, but centralisation only facilitates rather than guarantees

22 Parties, Gender Quotas and Candidate Selection in France

effective implementation, as the will of a party to implement the quota is also paramount. This hypothesis is tested by exploring the candidate selection practices of each of the main parties and comparing them to the party’s track record for parity implementation. The chapter then moves on to consider another major theory of gendered candidate selection, namely the ‘supply and demand’ model (Norris & Lovenduski 1995). If the shortage of women is a supply-side problem – that is, due to an insufficient number of suitable women candidates – then parties’ hands will be tied. Conversely, if it is a demand-side problem – that is, good women candidates come forward but parties reject them – then greater understanding needs to be found of why parties do not select women candidates. The chapter argues that supply and demand explanations may actually be interlinked, if parties are unconsciously prioritising qualities that are more likely to be present in a male candidate. This would lead parties to believe that the problem was one of supply even if it were triggered in part by demand. This theory is tested through an examination, based on interviews with party election officials, of what it is that parties are looking for in a candidate. These qualities are then evaluated to see whether they are gendered by examining the factors that shape a good candidate (such as education levels, career choices, prior political experience and availability) through a gendered lens. The findings suggest that, although women as a group are still disadvantaged relative to men in a range of areas that contribute to political careers, supply-side explanations alone are not sufficient to explain women’s under-representation in politics. The first of the three mainstream approaches to party behaviour is explored in Chapter 4. The chapter introduces the party competition approach and explains why parties might seek to prioritise electoral success when selecting electoral candidates. The rationality of parties is also considered, both in terms of whether parties are rational to prioritise the candidates that are the most electable, and whether it is possible for parties to make rational decisions concerning candidates’ electoral appeal. The chapter then proceeds to test how useful this approach is for explaining parties’ choice of electoral candidate. One of the core hypotheses of the chapter is that parties perceive certain types of candidate as being more electable than others, and therefore prioritise these types of candidate. This hypothesis is firstly tested through profiling the types of candidates that are preferred by parties, drawing on a variety of information including elite interviews. A quantitative analysis of candidate performance over time is then provided, with an

Introduction 23

emphasis on measuring any differences between the success rates of incumbents versus non-incumbents, and men versus women (given that incumbency is the main reason cited by parties for not replacing male candidates with women). By separating seat effects from candidate effects, a neutral measure of candidate performance can be obtained that can then be compared against party perceptions of how certain types of candidates will be received by the electorate. Comparing party perceptions of the electorate’s preferences to actual electoral outcomes illustrates whether or not parties are rationally targeting their electorate and fielding the candidates with the greatest chance of winning. An analysis is then provided of whether parties are willing and able to base their selection of candidates on a rational estimation of electoral performance, and hence of whether the competition approach is useful in explaining parties’ candidate selection preferences. The second approach to be explored is the institutional approach, and this is the subject of Chapter 5. The chapter considers the influence of the institutional environment on the choices of candidate that parties make. The institutions considered include the electoral system, the party system, the relative importance of the election, the detail of the parity law, and the relative size, strength and wealth of the party. One of the central areas of interest is the role of institutions in shaping and framing the range of choices available to political parties. A related and equally important theme is the ability of parties to adapt to the changing institutional environment. The institutional environment varies for each type of election, and this is a powerful explanatory variable in demonstrating why parity has a varying impact from one level of election to another. Certain types of party are more constrained by institutional factors than others, and the ability of a party to adapt to and even profit from the institutional environment is highly significant in determining whether or not a party will implement parity. If a party does not wish to implement the ‘parity’ law, institutional factors will be critical in determining how successful a party is at avoiding implementation. The approach is insightful in demonstrating how party choices are framed, especially where institutional constraints effectively leave parties with very few options to choose from. However, the approach is less helpful in explaining how and why parties that are faced with a range of options will discriminate between these options, and the institutional approach cannot in itself account for why different parties faced with the same choices will select different outcomes.

24 Parties, Gender Quotas and Candidate Selection in France

This leads to the ideological approach, explored in Chapter 6. The core question here is how much parties’ ideology can explain their attitude towards candidate selection. This produces a number of related questions to consider. Firstly, it is important to determine whether or not a party will be favourable to the principle of the ‘parity’ law. Factors influencing party attitudes to ‘parity’ include their attitudes to gender equality and also to the desirability of state intervention in enforcing equality. Secondly, the strength of a party’s beliefs is almost as important as the beliefs themselves. A party that is weakly in favour of parity is hypothesised to be less willing to make sacrifices in order to implement the ‘parity’ law than a party that is strongly and vocally committed to enforcing gender equality. How strongly a party feels about parity is likely to be linked to how large a role ideology plays in the party’s organisation, with some parties being heavily programmatic while others are more willing to dilute their ideology in order to maximise their electoral appeal. Where parties have weak ideological support for parity, this may still motivate them to implement parity, but only where it does not interfere with other imperatives such as vote maximisation. Where parties do not have ideological support for parity, ideology may cease to be significant, as parties may implement parity anyway for reasons explored in the preceding two chapters. These theories are tested against French parties, and the findings are used to evaluate the significance of ideology in determining whether or not parties will incorporate gender parity into their candidate selection procedures. Although each of the three main theories explored above provides essential insights into the reasons why parties select certain electoral candidates over others, no one approach on its own is sufficient to account fully for all the variation in the implementation of parity. Institutional variables provide a framework of options for parties to choose from, but do not account sufficiently for how parties choose between these options and why different parties favour different options. The electoral and ideological approaches go some way towards providing an explanation of party priorities, but their explanatory power is incomplete when each is considered in isolation from the other two approaches. Therefore Chapter 7 concludes the book by considering how all the different approaches interact. This chapter evaluates what has been learned so far and then reviews the relative strengths and weaknesses of the approaches explored in the preceding chapters. To provide a more complete answer to the central research questions, a new theoretical approach called the ‘party priorities

Introduction 25

model’ is introduced. This model uses the insights from the other approaches to provide an overall account of how different parties will react to the parity law under different circumstances. The model demonstrates the options available to parties and accounts for how different parties will prioritise certain options over others. The difference between parties is explained in terms of their overall motivations and priorities, and these priorities are then used to predict the path that each party will take when navigating through the competing choices available. The range of choices combined with the order in which they are prioritised will be determinant of the degree to which the ‘parity’ law will be implemented. The theoretical foundations and the functionality of the model are explored before applying the model to French parties to see how accurately the model can account for each party’s behaviour. The model is then evaluated in a number of ways. Firstly, its explanatory power is measured. Secondly, the utility of the model for analysing the application of other types of quota in other scenarios is considered. Finally, the book concludes with a discussion of its contribution to studies of parties and gender quotas, and the implications of this research for future policy developments.

2 French Parties and Parity in Context

Introduction The French ‘parity’ law has a specific history and context which underpin the issues explored in the remainder of the book relating to the law’s implementation. Similarly, the actors involved in the law’s passage and implementation – France’s political parties – are diverse in a number of ways that are relevant to the analysis of gender parity. This diversity influenced both the initial passage of the ‘parity’ law and the willingness and ability of each party subsequently to meet the requirements of parity. This chapter begins by providing an overview of the French party system and the six main parties within the system which are studied in this book. The history, current situation and ideological positioning of each party, along with the dynamics of the French party system, all help shape each party’s available options, thus illuminating the analysis provided in subsequent chapters. It is also necessary to understand the circumstances under which parity arose, given both the particular national context of parity and the broader international context of gender quotas. The pressure under which parties found themselves to introduce a law which they did not all support is important in explaining both why the law was passed and why it contained so many loopholes. The seemingly contradictory legislation which resulted, and the subsequent disregard shown for the law by some parties, can be better understood when considering party motivations. Although the actions of parties appear inconsistent, with varying levels of enthusiasm for gender parity, their underlying motivations reveal a more consistent, pragmatic approach. 26

French Parties and Parity in Context 27

The French party system France is often referred to as having a distinctive party system known as ‘bipolar quadrillism’. At its peak, the party system formed two distinct blocks of two parties each – the PCF and PS on the left, and the UDF and Gaullists on the right. The ‘bipolar’ characteristic is a product of the electoral system, with elections being held over two rounds, encouraging a proliferation of parties in the first round and a left-right stand-off in the second round. With no party large enough to dominate on the left or the right, co-operation between the two major parties on either side produced the infamous quadrille. Over time, this description has become increasingly inaccurate as new parties have entered the party system. The two major new players are the Greens (les Verts), who originally defined themselves as being outside the left-right spectrum but eventually joined the block on the left; and the FN, who have stolen votes from both left and right and continue to be a disruptive influence to the bipolar system. Moreover, the unity of each block has waxed and waned over the years. The left won the election of 1997 by uniting its various parties under the umbrella of the ‘gauche plurielle’ (plural left), while the parties of the right appeared divided. However, by 2002 these roles had reversed and a new, unified party of the right emerged in the form of the UMP, while the growing divisions on the left led to the collapse of the gauche plurielle. The UMP combined the Gaullist RPR party with smaller rightwing parties, as well as absorbing a significant portion of the UDF. The disintegration of the UDF was compounded in 2007, when most of the party shifted its loyalties to the UMP and changed its name to the Nouveau Centre. Those who remained loyal to the party’s leader, François Bayrou, regrouped under the name MoDem. The waxing and waning of party fortunes is another feature of the French party system. In the period 1978–2002, no party succeeded in winning a second term of office. Instead, the French electorate changed their mind at each election, leading to alternance – the replacement of each successive government with the opposition, resulting in each party enjoying a period of relative success followed by a period of relative decline followed by a revival in fortune, and so on. A third distinctive feature of the French party system is the changing nature and number of parties present. Although the number of parties capable of winning seats in Parliament is fairly limited, a large number of parties exists at the margins thanks to the two-round electoral system which allows the French electorate to vote with their heart in

28 Parties, Gender Quotas and Candidate Selection in France

the first round before voting tactically in the second round. These parties may not appear significant, but their ability to eat into the vote of larger parties has occasionally led to their co-option into electoral coalitions. Moreover, most French parties reinvent themselves on a fairly regular basis, occasionally through a change of name but more often through entering and leaving loose coalitions or through changes of direction under a new leadership. Indeed, leadership is the fourth notable aspect of the French party system, resulting from the presidentialised nature of French politics. Parties are used as vehicles for presidential candidates, and this has a dramatic influence on the organisation of parties as well as on the tactics deployed by key political players. Presidential politics was at the heart of the final disintegration of the UDF. Whereas the UDF and UMP had previously been allies who seldom pitted candidates against each other in parliamentary elections, the warm relations between the two parties cooled in the 2002–7 parliament as the UDF leader, François Bayrou, withdrew his support for the UMP government and increasingly positioned himself as a rival political force. Following a strong showing by Bayrou in the 2007 presidential elections, the UMP presidential candidate, Nicolas Sarkozy, threatened to pit UMP challengers against UDF incumbents in the forthcoming legislative elections unless they declared their loyalty to Sarkozy rather than Bayrou. This led to the split in the party between Bayrou loyalists (MoDem) and Sarkozy converts (Nouveau Centre). This is a prime illustration of the power of presidentialism to change the party system. In addition, new parties are occasionally formed by ambitious or deluded politicians who have become disillusioned with the party they were in and seek to lead their own party, as was the case, for example, with Jean-Pierre Chevènement’s Mouvement des Citoyens (Citizens’ Movement). Parties are defined by their leaders and even those with no realistic hope of getting their candidate elected will put forward a presidential candidate. The consequence of this emphasis on présidentiables (potential candidates for the presidency) was never more evident than in the presidential elections of 2002, when a record number of candidates led to a splitting of the vote on the left and the shock elimination of Lionel Jospin, who had been widely tipped to win, in the first round, and the progression to the second round of the far-right leader, Jean-Marie Le Pen. France’s party system is therefore characterised by change and uncertainty, by division, unification, reinvention and alternance. The bipolar quadrillism has been replaced by bipolar multipartism, and even this latter term is questionable. The electoral system and the

French Parties and Parity in Context 29

presidentialisation of French politics continue to be dominant influences in the shaping of the party system. The section below explores the characteristics of the six most influential parties in the French party system to demonstrate the history and nature of each party, as well as the challenges that they currently face. The PCF (Communist Party) The PCF is an ideological party in nature, and gender equality, along with more general ideals about equality, is a feature of this ideology. Their current leader, Marie-George Buffet, is female, and their official documents use a feminist writing-style.1 Their national leadership bodies and committees are all comprised of 50 per cent men and 50 per cent women (www.pfc.fr). However, women still comprise a minority of the party membership and of those holding elected office in local, regional and national politics. The party’s origins lie with the international movement of Communist parties, although unlike many of the Communist parties within Western Europe, the PCF achieved mainstream political importance, especially following the role it played in the Résistance during Nazi occupation. In its heyday, it commanded 28.6 per cent of the vote (1946), making it the largest party in France, and it continued to receive around 20 per cent of the vote until the end of the 1970s (Wright 1989, p. 240). Since then, the PCF has faced a period of steady decline, and has suffered from the strengthening of the Socialist Party (and, more recently, the rise of environmental parties such as the Greens), the fall of communism in eastern Europe, and the loss of its natural constituency of working-class blue-collar workers and ex-revolutionaries as the latter die out and French society becomes increasingly white-collar and middle-class. In the 2002 legislative elections, it received just 4.7 per cent of the vote, and this dwindled even further to 4.3 per cent in 2007 (Bell 2003, p. 30; Ministère de l’intérieur). This humiliation was compounded by Buffet’s weak performance in the 2007 presidential elections, where she obtained a mere 1.9 per cent of votes and was eclipsed by the far-left candidate Olivier Besancenot, who more than doubled her score with 4.1 per cent (Ministère de l’intérieur). One consequence of the PCF’s decline is that it has had to contemplate a change of strategy. In many respects, the party has diluted its ideology in order to attempt to survive in the French electoral market, meaning that it can no longer be considered a purely ideological party

30 Parties, Gender Quotas and Candidate Selection in France

that places ideology over all other considerations. In addition to making such changes as embracing democracy and abandoning its ties with the Soviet Union, the party has also acknowledged that it can no longer hope to have any significant influence if it operates in isolation from mainstream politics. The party has therefore co-operated with other left-wing parties, and notably the Socialist Party. Although the PS has emerged as the dominant partner within this coalition, the PCF has had little choice but to persevere with left-wing alliances as its only route to power. This strategy has provided the party with a route into government that would otherwise have been impossible given the low parliamentary presence of the party (currently 15 seats in Parliament, although this figure was temporarily higher following the ‘vague rose’ (surge of support for the left) in 1997, when they won 36 seats). Their participation in a coalition government – albeit a coalition of parties that were all on the left – forced them to make some compromises which made some of their members uncomfortable, but did at least afford them two ministries plus a junior minister (Knapp 2004, p. 114). However, they are now left in a position where they are dependent both on their alliance with the PS, and on their existing incumbents, for survival (ibid, p. 115). The PS (Socialist Party) The PS have been pioneers in terms of women’s representation. They have had at least some degree of internal gender quota since 1974, were the leading party behind the parity reform, have been the only party to nominate a female Prime Minister (the short-serving Edith Cresson, who held the office from 1991–2 under François Mitterrand’s presidency), and in 2006 nominated a woman – Ségolène Royal – as their presidential candidate.2 Yet their commitment to women’s representation, and to women’s rights more generally, has been somewhat unreliable. Yvette Roudy complained of the underhand tactics used to minimise gender quotas in the late 1970s (Roudy 1995, pp. 15–22) while Sineau and Jenson (1995) lamented the failure of Mitterrand to deliver on his promises to women. Mitterrand’s initial enthusiasm for women’s rights may have been influenced at least in part by his attempts to woo the female electorate, and they steadily dropped off the agenda as his presidency progressed. It was not until Lionel Jospin took over the party in 1995 that the PS regained its feminist edge. Moreover, the PS do not have a good track record of implementing their own quotas; in 1988 and 1993, they had a lower percentage of female deputies than the RPR, despite the fact that the RPR had no

French Parties and Parity in Context 31

gender quotas and was not a vocal advocate of women’s representation. Even when Jospin promised that 30 per cent of electoral districts would be reserved for women in 1997, this still only translated into 16 per cent of seats, although even this was a substantial improvement on the three seats (5.6 per cent) won by women in 1993. The PS claims to be a party driven by ideas rather than by vote maximisation (Opello 2006, p. 39). Despite this, the party’s electoral pragmatism has seen its ideas shift towards the centre over the past 25 years. After its first two years in power from 1981 to 1983 led to looming economic disaster, the party was forced to reappraise its position on matters such as nationalisation and capitalism. It gradually shifted from a traditional socialist party to a party of the centre left. During the 1997–2002 parliament, the Jospin government: …sold off more state property, in money terms, than all of its rightwing predecessors put together. It saw in the euro, and with it the stability pact perpetuating the Maastricht convergence criteria. And it used riot police to dislodge activists for the unemployed from benefit offices. (Knapp 2004, p. 115) The PS might not have abandoned their socialist roots altogether, but they have shown that they can and will make ideological compromises in order to be a credible party of government. Shifts towards the centre of the political spectrum have not been the PS’s only electoral strategy. The party has also been successful in creating the ‘plural left’ coalition, whereby they used the combined strength of the left-wing vote to get themselves back into power, whilst still being the dominant partner in all coalition negotiations. In 1997, they managed to translate a 50 per cent share of the left-wing vote into 70.4 per cent of government posts (19 positions out of 27; Hanley 2003, p. 82). However, the party has long been riddled by internal division and factionalism, and this division was evident in 2002, when the party imploded following its disastrous results in first the presidential and then the legislative elections. The party also failed to unite successfully around Royal’s 2007 presidential campaign, contributing to her defeat. The bitterly contested leadership election of 2008 gave a very narrow victory to Martine Aubry over Royal, resulting in a party with a female leader but ongoing rivalries. The party is also struggling to maintain pre-eminence amongst the left-wing electorate, who are facing an everexpanding number of parties from which to choose. Indeed, it was the

32 Parties, Gender Quotas and Candidate Selection in France

plurality of left-wing presidential candidates in the first round of voting in 2002, and the split of the left-wing vote between these candidates, that prevented the Socialist candidate from qualifying to the second round of voting, in which he might otherwise have triumphed. The party is therefore faced with several pressing needs at the moment if it wishes to return to power, not least of which are the need to balance the factions within the party, and the need to unify the leftwing vote. The Greens The Greens are the most successful of France’s environmentalist parties, due in part to their decision to participate in the ‘plural left’ coalition. They are based on a strong ideology which is centred on the environment but also includes liberal and post-materialist values, one of which is gender equality. The Greens have a strong record of commitment to feminist values, and have been the most successful of the major parties in initiating and implementing gender quotas. The Greens have been committed to the principle of parity since 1988, and have largely achieved this within the internal party organisation, although they have had more trouble achieving a parity of electoral candidates, due in part to the lower proportion of female members (Allwood & Wadia 2000, p. 70). In the 1997 and 2002 elections, onethird of their parliamentary delegation was female, which made them the most feminised party in the National Assembly, even if their absolute numbers were very low. The gain of an extra male deputy in 2007 pushed down the proportion of women to 25 per cent, slightly lower than the PS. Although the Greens are founded on a clear and distinct ideology, even they have demonstrated willingness to compromise and temper their idealism with electoral pragmatism (Spoon 2007; Villalba and Vieillard-Coffre 2003). Their initial motto that declared that they were ‘neither left nor right but in front!’ was soon abandoned as they realised that their best prospects for electoral bargaining lay with the left. The Greens enjoyed an increasing vote share throughout the 1980s and 1990s, but the SMP electoral system made it hard for them to translate their votes into seats. Their breakthrough came through negotiations with the PS, which needed the Green vote to help it regain power, and these negotiations resulted in a few seats being ‘reserved’ for a Green candidate, in exchange for which the Greens would support left-wing candidates in the constituencies not reserved for their party. The Greens were rewarded with six seats in the 1997

French Parties and Parity in Context 33

Parliament and the most coveted prize of all, the Ministry for the Environment. However, their weak presence in a coalition dominated by the PS came at a high price; they were not able to enact most of the promises that they had made in their manifesto, and in 2002 they were left wondering whether the benefits of a limited role in the government were sufficient to justify the cost of ideological compromise and disillusionment of the membership, whose high hopes of a Green environment minister did not translate into significant environmentalist policy outcomes. The UDF (non-Gaullist right) The UDF were, nominally at least, a party of the centre, although their alignments and alliances placed them firmly on the right (Knapp 2004, p. 200). In fact, for most of their existence, the UDF were scarcely a party at all, but rather a loose confederation of centrist movements, bringing together ‘liberal, Christian democrat and radical currents’ (Sauger 2003, p. 107). As a result, the party’s ideology was somewhat muddled between the different elements of the party, although gender equality never featured strongly in the party’s rhetoric, and one of their deputies, Christine Boutin, was the only person to vote against the parity legislation (Fabre, Le Monde 27 January 2000). Moreover, the UDF faced major upheaval over the past decade. In 1998, one of the major parties within the confederation, Démocratie Libérale (DL), withdrew from the UDF, taking with them nearly a third of the party’s members and deputies. An even bigger crisis faced the UDF in 2001–2002 when the UMP party was formed (see below for greater discussion of the UMP), and much of the UDF was subsumed into this larger party. The UDF, who have traditionally formed governing alliances with the Gaullists, now found themselves a much weakened minority party. In 1988, they actually won slightly more seats than the Gaullist RPR (133 seats to 130); in 1993, following a huge electoral swing to the right, they won 216 seats and were by a considerable margin the second largest party in the National Assembly. However, by 2002, the loss of much of the party to DL and the UMP left the UDF floundering with only 29 seats. With the UMP possessing a large absolute majority of seats in Parliament, the UDF were no longer needed for coalition formation, and found themselves lost on the sidelines. They initially supported the government, and were rewarded with ministerial office for their vice-president, Gilles de Robien. However, following the government’s defeat over the referendum on the European constitution in 2005, and mindful of the fact that it was less

34 Parties, Gender Quotas and Candidate Selection in France

than two years until the next presidential and parliamentary elections, the UDF’s leader, François Bayrou, took the decision to withdraw their support for the UMP. When de Robien decided to remain in the government, he was stripped of his party functions within the UDF (Montvalon and Roger, Le Monde 16 June 2005). The consequent rivalry between Bayrou and Sarkozy in 2007, as detailed above, led to the final disintegration of the party into the Nouveau Centre and MoDem. In electoral terms, the UDF in 2002 and its offshoots in 2007 could be considered parties in crisis, with defections to other parties by both members and deputies contributing to the UDF’s demise. The UMP’s enhanced electoral capabilities meant the UMP was able to compete in many seats which had previously been contested only by the UDF, and this proved very costly for the UDF in 2002. Many of the seats it did win were not contested by the UMP, which was why these incumbents were willing to defect rather than risk losing their seats in 2007. The party’s final split happened so close to the 2007 elections that it is still possible to analyse the UDF’s candidate selection processes for 2007. Indeed, despite the recent split of the UDF, much of the analysis of this book focuses on the behaviour of the UDF prior to this split, and hence examines the UDF as a single party. UMP (Right-wing umbrella party) The UMP is distinctive from the other parties discussed here for a number of reasons. Firstly, the party has only been in existence since 2002, and its long-term future remains uncertain (although the party has grown steadily in strength since its inception). Secondly, the UMP has replaced the UDF as the umbrella party of the right; although the remnants of the Gaullist RPR party are the largest and most influential component of the UMP, they now sit alongside a large section of the former UDF, DL and smaller groupings on the right. The party is still establishing its identity with the French public, and Sarkozy is steering the party in different directions to his predecessor, Jacques Chirac. The UMP was formed, in part, as a response to the particular political situation that emerged in 2002, and the institutions in France that created this situation. Prior to this year, France had a president that was elected every seven years, and a legislature that was elected every five years (or earlier at the president’s request). It was therefore possible to have periods during which the president was not from the same party as the party which had a majority in the National Assembly (known as cohabitation). Indeed, the electorate frequently change their

French Parties and Parity in Context 35

mind in France, and have thrice voted for the opposition when a parliamentary election fell part way through a presidential term. Two of these periods of cohabitation directly affected Jacques Chirac; once as Prime Minister under Mitterrand from 1986–8, and once as President, cohabiting with the Socialist PM Lionel Jospin, from 1997–2002. Chirac also turned down the opportunity of being Prime Minister in 1993 due to his reluctance to serve under Mitterrand for an additional period of cohabitation due to his previous negative experience of deadlock. The year 2002 presented a natural opportunity to address this situation, as the presidential and legislative elections happened to fall in the same year, and five long years of cohabitation had made the major parties and the electorate keen to end the deadlock once and for all. Two political measures were taken to reduce the possibility of any further cohabitation; firstly, the presidential term was reduced from seven years to five so that the electoral cycles would continue to be in harmony, thus maximising the opportunity for the electorate to vote the same party into the Elysée and Matignon3 before the public mood shifted. Secondly, the electoral cycle was reversed so that the presidential elections would precede the legislative elections; it was assumed, with good reason, that whichever candidate was victorious in the presidential elections would see their party triumph at the legislative elections two months later. It was within this context that the UMP was formed. Seeking re-election and eager to maximise his chances, Jacques Chirac aimed to build a broader coalition to match the ‘plural left’ coalition of his main rival, Lionel Jospin. Alliances were cultivated over the two year period preceding the elections with key figures in the UDF and Démocratie Libérale (Knapp 2004, p. 230). Once Chirac had secured his win in the presidential elections, his party, the RPR, spearheaded the alliance that took on the name Union for a Presidential Majority, or UMP. The aim of the alliance, and the message they sent out to voters, was simple: provide the president with a parliamentary majority to avoid another period of cohabitation. The campaign was straightforward and very effective. The Union consolidated its form in November 2002 and changed its name to a Union for a Popular Movement, thus retaining the UMP initials but shedding the temporary name that it had acquired for the 2002 elections. The stunning victory of the UMP in 2002, and the party’s re-election in 2007 – breaking 25 years of alternance – conceal underlying problems with the coalition. The idea of a unified party of the right had long been popular amongst elements of the right-wing electorate that

36 Parties, Gender Quotas and Candidate Selection in France

wished to fend off threats from the left and from the far-right (Knapp 2004, p. 257). Yet one of the reasons why the idea of the Union was so popular was because its component parts were so weak. In particular, the RPR was fragile and had lost its position of dominance over the right. The party had been tainted by corruption scandals and was not performing well in local elections compared to other right-wing parties (ibid, p. 90). Its role as the dominant partner in the UMP is greater than any role it might have hoped for if it had continued as a separate party, and it is uncertain whether the RPR will be able to sustain its privileged position within the UMP. The party’s future depends on the ability of its various factions to continue co-operating with each other – something that has often proven problematic for French parties in the past. The FN (Front National) The FN shot onto the political scene in the early 1980s and quickly established itself as a political force that could not be ignored. Its breakthroughs came courtesy of proportional representation, first in local elections and then in the temporary use of PR in the 1986 legislative elections, where the party won 35 seats (Borella 1990, p. 208). Since then, the SMP electoral system has frustrated the FN’s attempts to gain parliamentary representation, with the party only gaining one seat in 1988 and 1997, and no seats at all in 1993 and 2002. Nonetheless, it consistently receives 9–20 per cent of the national vote, and its presence has unsettled the other parties in the party system, especially when it has managed to qualify for the second round of voting. In some cases, this has led to triangulaires between the left, the right and the FN. The mainstream right has chosen to dissociate itself from the FN, even though they have lost some of their voters to the FN. The FN has also attracted voters away from the PCF, particularly within deprived working-class areas. The far-right ideology of the FN is particularly unfavourable to feminism, and the FN was open in its disapproval of parity and in its belief that a woman’s place is in the home. It even proposed the policy of a ‘maternal salary’ that would pay women to stay at home with their children. Ironically, however, it tends to benefit from fielding female candidates as, rightly or wrongly, women are perceived as less extreme and have helped to soften the hard-line image of the party. Nonetheless, the FN has consistently performed far less well amongst female voters than male voters (Helft-Malz and Lévy 2000, pp. 46–7; Sineau 2008). The party has therefore found the implementation of parity to

French Parties and Parity in Context 37

be both undesirable and difficult to achieve (Ivaldi 2003, p. 144; Zimmerman 2003b). The FN has enjoyed mixed fortunes over recent years. Internal divisions resulted in a major upheaval in 1999 when a faction splintered off to form a rival party, the Mouvement National Républicain (MNR), under the leadership of Bruno Mégret. The short-term effect of this schism was costly for the FN, although their electoral fortunes recovered significantly in 2002 with the surprise success of the FN’s charismatic leader, Jean-Marie Le Pen, in the presidential elections. Although Le Pen was resoundingly beaten in the second round of the elections, his mere presence served both to place the FN firmly back on the electoral map, and to assert the FN as the major party of the far-right. The influence of the FN in framing the terms of debate that dominated the presidential elections, such as crime and insecurity, also highlighted the party’s continued significance. Since this high point, the FN have entered a new period of turmoil. Their poor performance in the 2002 legislative elections yielded no seats, and the electorate were still haunted by Le Pen’s 2002 success in the 2007 presidential elections. Le Pen’s vote share in 2007 plummeted to 10.4 per cent, and the party again failed to win seats in the subsequent legislative elections. Le Pen is now at the end of his political career and looks set to be succeeded by his daughter Marine. Paradoxically, the least feminist of all parties may benefit from the more moderate image provided by Marine Le Pen, perhaps as a result of her gender. Given the FN’s limited electoral impact in legislative elections, the FN is studied in less close detail than the other parties in this book. Nonetheless, the FN is still worthy of some consideration, not least as an example of a party which opposed parity on ideological grounds and yet felt compelled to respect the law for more pragmatic reasons. In addition, the centrifugal impact of the FN on the French party system ensures that it qualifies as a party with ‘blackmail potential’ under Sartori’s criteria of relevance (Sartori 1976). As such, the FN cannot be ignored in any discussion of the party system. The key characteristics of the parties are summed up in Table 2.1. The variations between the different parties provide an initial indication as to how party motivations towards parity might differ. For example, some parties are more ideological than others, and some are in a more vulnerable position than others. Indeed, a dominant theme for all parties is one of survival in a party system which is constantly evolving. In order to make a more informed assessment of how these variations may influence attitudes towards parity, it is first necessary to

38 Parties, Gender Quotas and Candidate Selection in France Table 2.1

Overview of key party characteristics

Party

Year founded1

Ave. seats in NA2

Ave. % women in NA2

Membership3

PCF PS Greens UDF UMP FN

1920 1971 1984 1978 2002 1972

25.2 180 4.34 100.65 239.26 0.4

11.1% 14.8% 30.8%4 4.2%5 9.7%6 50%

133,767 147,275 8,810 40,000 164,500 60,000

Political persuasion Left/far left Centre-left Green/left Centre/right Moderate right Far right

Sources: Haegel 2005; Le Monde; Knapp 2004 1 This refers to the year that the party was founded in its current incarnation (excluding the recent collapse of the UDF). Some parties, such as the PS, have existed in earlier incarnations under different names, while the UMP is an umbrella party formed of older parties such as the RPR, which was founded in 1976 and has Gaullist roots that go even further back. 2 Average number of seats held in the National Assembly over the last five elections (1988, 1993, 1997, 2002 and 2007), and the average percentage of these seats held by women. 3 Membership figures are for the early 2000s, and are based on the claimed figures, although these tend to be significant exaggerations. The figure for the Greens is an estimation, which makes it seem relatively low compared to the claimed figures (there is no claimed figure available). The figures are taken from Knapp 2003, p. 355. There is no reliable contemporary data on the percentage of women within each party’s membership, although women tend to form a minority of between 30 and 40 per cent of members in most parties. 4 The Greens have only had seats in Parliament since 1997. 5 The figures for 1988–2002 are for the UDF; the 2007 uses the combined totals for Nouveau Centre and MoDem. These figures conceal a sharp decline from a height of 216 seats in 1993 to just 18 seats in 2007, as a result of the disintegration of the UDF. 6 The UMP has only existed since 2002; the figure for elections prior to this date is for the RPR.

explore the origins of parity and the way in which initial party responses to the parity movement may have shaped the legislation itself and the subsequent attempts by each party to implement the legislation. The issues facing parties at the time of parity’s conception, and their consequences for the framing of the law, are considered below. These ideas are then revisited in Chapters 4, 5 and 6, which consider the various challenges and priorities for parties when attempting to implement the law.

Party motivations in ratifying parity The full history of how parity came into being is detailed and complex, and is well documented elsewhere in the literature (Bereni 2004,

French Parties and Parity in Context 39

2007a, 2007b; Haase-Dubosc 1999; Helft-Malz & Levy 2000; Krook 2005, 2009; Lépinard 2004; Lovecy 2000; Mossuz-Lavau 1998; Scott 2005; Tremblay 2002). What is of interest here is not a complete chronological account of events, but the core background information in order to understand the motivations of parties in introducing and then ratifying the parity legislation. I hypothesise that these motivations will be a useful tool in predicting parties’ motivations in the subsequent implementation of the ‘parity’ law. For example, a party that was primarily driven by ideological support for parity might be more inclined to implement the law than a party that sought only to appease the female electorate. As will be demonstrated, the choice of gender parity, as opposed to a less constrictive gender quota, was primarily the result of France’s constitution and its universalist tradition. The placing of parity on the political agenda came mostly from pressure groups external to the party system, although ideological support for parity from within the governing left-wing coalition led to parity being placed on the legislative agenda. Its actual passage, however, was above all due to the desire of all parties to woo the female electorate. This combination of ideology and electoral pragmatism, within a framework dictated by France’s political institutions, is as present in parity’s foundations as in its application. Institutional obligations The French republican tradition is based on the theoretical model of the universal citizen. For this reason, the French Constitutional Council shot down an attempt by the socialist government to introduce gender quotas in 1982, arguing that a gendered concept of citizenship was incompatible with the constitution. As a result, gender quotas have been an option for parties internally but have not been possible at a national, state-enforced level. Given France’s low level of female representation compared to her neighbours,4 campaigns for a mechanism to introduce larger numbers of women into politics persisted, but had to change tack. ‘Parity’ as an idea actually originated outside political parties, with its initial formulation taking place at the EU level and then being popularised by Gaspard et al (1992). The idea subsequently gathered momentum within the feminist movement and eventually took hold in parties, as will be discussed later in this chapter. What is of note here is that the concept of ‘parity’ was very much a product of France’s political traditions and of the 1982 ruling. Once it became evident that a gender quota of the kind seen in

40 Parties, Gender Quotas and Candidate Selection in France

Belgium or Argentina would not be possible for France, feminists had to become more creative in designing a form of quota that would be compatible with the constitution. With parity, they argued, the universal citizen model remained intact, because parity was merely reinforcing the existing gender balance amongst French citizens. Although theoretically weak, and highly disputed by proponents of universalism, this argument was sufficient to get gender quotas back on the agenda, albeit repackaged under the name of ‘parity’. This point gains added significance when it is borne in mind that, under different circumstances, it is highly unlikely that parties would have opted for a quota set as high as 50 per cent. Indeed, the PS (the main proponents of parity) had previously fought to keep quotas as low as possible. For example, the party initially had a quota for women as low as 10 per cent, and attempts to raise the quota to 20 per cent in 1977 were thwarted at the last minute so that the quota was only raised to 15 per cent (Roudy 1995, pp. 20–1). A similar practice was seen in 1982 when the original proposal for gender quotas asked for a minimum of 30 per cent women, and this was then revised down to 25 per cent and expressed as a maximum of 75 per cent of either sex (Mossuz-Lavau 1998, p. 31). By 1997, the PS was using a gender quota of 30 per cent, but this was still well short of parity and even then they struggled to implement it.5 Even these limited measures surpassed those of other parties, who had not implemented any electoral quotas at an internal level at this point (Quota Project; Allwood & Wadia 2000, ch. 2). It is therefore highly improbable that any party would have suggested parity if a gender quota with a lower minimum percentage of women had been a feasible alternative. Parity’s form can thus be attributed to institutional circumstances independent of party preferences. Its arrival on the political agenda, by contrast, was motivated by electoral competition, as the next section will demonstrate. Electoral obligations The defining moment for placing parity on the political agenda was the presidential elections of 1995. Jacques Chirac, representing the RPR, had once described quotas as ‘reverse sexism’, and had publicly stated only a year before the election that he was opposed to the idea of a parity law and to gender quotas (Allwood & Wadia 2000, p. 71; Remy & Mandonnet, L’Express 11 February 1999). However, he was threatened not only by his left-wing opponent, Lionel Jospin, but also by the presence of a second strong candidate on the right, Edouard

French Parties and Parity in Context 41

Balladur, the then prime minister. All three candidates in this tight three-horse race sought to maximise their vote. Over the 50 years since the granting of women’s suffrage in 1944, the gender gap in voting had reversed, with women now being more likely than men to vote for parties of the left (Allwood & Wadia 2000, p. 124). All candidates knew that the female vote had been instrumental in granting Mitterrand two terms and ending the domination of the Right since the start of the Fifth Republic; Jospin sought to capitalise on this trend, and his opponents sought to reverse it. Moreover, the debate on parity was gathering momentum, and each candidate was forced to take a stand on the matter. Unwilling to alienate the female electorate, each candidate came out in favour of increasing women’s representation. Jospin, who himself had once been ‘vehemently against the principle’ of parity, declared his wish to ‘boldly advance towards parity’, suggesting that ‘the number of women elected should be one of the criteria determining the level of public finance for political parties’ (Mossuz-Lavau 1998, p. 53; Roudy 1995, p. 18). For his part, Chirac did something of a U-turn; whilst still expressing certain reservations about parity, he declared that ‘we need to set clear objectives concerning women’s representation. For me, this can only mean parity’ (Mossuz-Lavau 1998 pp. 53–4). Like Jospin, he approved linking state finance of parties to their level of parity, and also promised to create an Observatoire de la Parité if elected. By contrast, Balladur made the less ambitious, but more concrete, offer of modifying the constitution to install a 30 per cent gender quota for elections using proportional representation. Only the far-right FN leader Jean-Marie Le Pen openly rejected the principle of parity, stating that it was incompatible with universalism. The attempts of the other political leaders to pander to the female electorate led two commentators to observe that their sudden interest in parity was ‘for some by conviction, for others by opportunism’ (Aulagnon & Bezat, Le Monde 10 March 1995). Although Chirac’s commitment to parity was very questionable, he did make good on his promise to install an Observatoire de la Parité; however, in its early days the Observatoire was severely under-resourced and intended to be nothing more than symbolic.6 Meanwhile, Jospin’s commitment to parity was growing under the influence of his second wife, the feminist philosopher Sylviane Agacinski. Capitalising on the bad press received by the RPR for dismissing eight of the 12 female ministers in Juppé’s government within the first six months of office,7 and the growing popularity of parity in the opinion polls,8 the PS made a manifesto pledge to introduce parity in the legislative elections of

42 Parties, Gender Quotas and Candidate Selection in France

1997. When the PS won the election, Chirac found his hands tied. Having previously done a U-turn to support parity in his presidential campaign, he could not now afford to thwart the government’s attempts to introduce parity. Both leaders therefore chose to ride the wave of popularity that parity enjoyed in the late 1990s, and help parity become law. By this point, it would be fair to suggest that Jospin was ideologically committed to parity, whilst Chirac was still more concerned with electoral expediency. Ideological battles Surprisingly, some of the key ideological battles surrounding parity did not take place between parties but within them. On the left, there was broad ideological support for parity within the Greens, who had already inscribed the principle of parity into their statutes, but less agreement within the PS. Although the principle of bringing more women into politics had unanimous support, members of the party were divided as to whether or not parity was the best way to achieve this goal. Robert Badinter, a PS senator who, like Jospin, was married to a prominent feminist philosopher, did not share Jospin’s enthusiasm for parity (and nor, for that matter, did his wife!).9 In an impassioned speech on the floor of the Senate he argued that parity risked essentialising women and dividing society into communities. In so doing he added his voice to a small but vocal anti-parity movement within the left, headed by numerous prominent women (L’Express 11 February 1999; Pisier, Le Monde 11 February 1999). Things were hardly plain-sailing for the right either. The overall sentiment towards parity was, at best, one of unease, and at worst, one of outright hostility. One conservative senator asked ‘do they [women] even really like politics?’ (Bacque, Le Monde 05 March 1999). Meanwhile, the (female) president of the RPR, Michèle Alliot-Marie, objected to parity on the grounds that ‘I find it insulting towards women and I have calculated the difficulties in implementing it’ (Saux, Le Monde 25 November 2001). Nonetheless, Jacques Chirac maintained his public commitment to supporting parity, and coerced his reluctant party into following suit. His control over his party was tested when the right-wing Senate made several attempts to shoot parity down by rewording the text so as to render it ineffective.10 He was helped in this endeavour by Christian Poncelet, the president of the Senate and a member of the RPR, who provided crucial support in helping the passage of parity through the Senate. He was also aided by several high-profile female politicians from the right who gave their whole-

French Parties and Parity in Context 43

hearted support to parity, such as Simone Veil and Roselyne Bachelot. For these women, support for parity came from a passionate ideological attachment to the idea. For the vast majority of their party, however, grudging support for parity was founded in far more pragmatic motivations. They saw in parity an opportunity to distance themselves from the claims of sexism that had followed the ‘Juppette’ scandal in 1995. The backlash that this had caused, combined with the RPR’s poor performance in the 1997 election, raised the fear that the right had alienated the female electorate. To alienate them further by taking a public stand against parity, which undoubtedly would have played into the hands of the left, was a risk that parties of the right were not prepared to take. The long-term implications of parity were considered less harmful than the immediate ramifications of obstructing parity’s passage. There was also an additional factor that would have helped parties of all sides to placate those members opposing parity on the grounds that it intervened with the candidate selection process, rather than on more ideological grounds. This was the fact that the law was fairly malleable for the crucial elections to the National Assembly, and did not apply to elections to the powerful departmental assemblies. The exact details of the law and its impact on candidate selection will be discussed further in Chapter 5. Of note here is that the weaknesses of the law may well have been intentional, and may have been an integral part of the design of a law that was intended more to please the electorate than to revolutionise the composition of France’s political elites. Indeed, this would place parity within a wider tradition of symbolic womenfriendly policies in France, such as equal employment policies, as well as within a wider international trend for gender quotas which may promise more than they deliver (Matland 2006; Mazur 1995). Parity almost fits the mould of symbolic reform whereby ‘decision makers, more interested in image making than problem solving, design policy statements with no teeth … destined to have little real impact’ (Mazur 1995, p. 2). The law does do more than the bare minimum required for symbolic reform, especially in second-order elections, but the loopholes left in the policy to minimise its impact in first-order elections are revealing of the compromises needed to secure the law’s passage. It is also interesting that the word ‘parity’ was deliberately left out of the actual wording of the law. Once the law had been weakened down enough to be tolerable to its opponents, it was passed with quasi-unanimity.11 As the above sections have shown, there is no unified pattern that maps out party motivations in implementing parity. This is principally

44 Parties, Gender Quotas and Candidate Selection in France

because parity had supporters and opponents across the political spectrum, and different actors within the same party had different attitudes towards parity. The influence of the 1982 Constitutional Council ruling in shaping later versions of parity is clear (Krook 2005). The dynamic between ideological and electoral motivations, however, is rather more muddled. Parity did not fit into a clear left-right ideological spectrum, and garnered support and opposition from parties of both sides. Overall, ideology appeared to have a stronger impact on parties of the left, both in support for and opposition to parity, whilst parties of the right seemed to have been motivated primarily by electoral considerations both in their initial opposition to, and eventual support of, the parity law. Although parity found supporters and detractors on both sides, left-wing arguments tended to focus on ideology (with debates centred on the philosophical aspects of parity, and disagreement focusing on the means rather than the end). By contrast, right-wing arguments were more concentrated on electoral motivations, with detractors resenting the restriction on the free choice of candidates that parity would entail, while its supporters emphasised the need to regain the female electorate. (Again, this is consistent with international trends for gender quota debates (Dahlerup 2006).) This is not to say that ideology did not play a part in influencing the decisions of parties of the right, nor that electoral competition was not a factor in determining the behaviour of parties of the left. Evidently, these factors work in tandem and cannot be artificially separated. However, the priority accorded to different factors seems to have varied, with ideological arguments carrying more weight for parties of the left and pragmatic arguments being of greater importance to parties of the right. Although some actors, especially from parties of the left, were acting out of an ideological commitment to parity, the majority of actors, and particularly those on the right, were acting primarily out of electoral interests. It was believed that the electoral advantage of voting for parity exceeded the costs that parity would entail, especially as efforts were made to ensure that these costs would be as low as possible. Whether or not these priorities carry through to the implementation of parity is a theme that will be revisited in greater detail in subsequent chapters, as competing theories of party motivations are examined. This chapter has illustrated the core background information underpinning the remainder of the book. The parties featured in the book, and the law that they are charged with implementing, can now be situated within their appropriate contexts. Before moving on to studying competing explanations for why some parties are better able to imple-

French Parties and Parity in Context 45

ment parity than others, it is first necessary to examine the question of candidate selection. The process by which parties select candidates, along with the criteria candidates are expected to meet (and whether or not these are gendered), are essential for understanding how parity interacts with the wider process of political recruitment. This is the theme of the next chapter.

3 Candidate Selection, Supply and Demand

The primary impact of the parity legislation on political parties is at the level of candidate selection. In order to meet the requirements of the law, parties need to select more women candidates and fewer men candidates. In order to understand parties’ willingness and ability to meet this requirement, it is important to understand both the process of selection and the type of candidate selected. The process of selection is significant for understanding who within a party is responsible for implementing parity, and the ease with which this process can be coordinated and controlled. Examining which candidates are sought after by parties is revealing as it indicates the extent to which women are able to meet the demands of parties, and hence whether or not there is a sufficient supply of women candidates to fulfil the requirements of parity. In this sense, this chapter will build on theories of supply and demand in the candidate selection process, by illustrating that lack of supply is partly a function of subtly gendered demand (Norris & Lovenduski 1995). The first section of this chapter explores the hypothesis that the degree of centralisation of the selection process affects how easily a party can implement parity. A more centralised process should facilitate the implementation of parity, although it is not a sufficient condition in itself. This requires a study of the formalised process of candidate selection, namely, how candidates are selected rather than which candidates are selected. The second section then examines which traits parties seek in electoral candidates, and whether these are as gender-neutral as they first appear. Although none of the requirements for being a candidate is obviously gendered, the overall profile required by parties is more likely to be found in a man than a woman. This is illustrated by a study of women’s roles in the social, political, economic and domestic 46

Candidate Selection, Supply and Demand 47

spheres. There is certainly some evidence to support the claim that the supply of women candidates may be lower than that of men. However, supply is still sufficient to meet the requirements of parity, implying that the continuing shortage of women in French politics is primarily a problem of demand.

Party candidate selection procedures Pippa Norris identifies four levels of analysis in the study of political recruitment: the political system, the recruitment process, the supply of candidates and the demands of gatekeepers (Norris 1997, p. 1). The political system is considered, within the conceptual framework of this book, to be part of the institutional environment within which parties operate, and as such is discussed in Chapter 5. The supply and demand of candidates is considered in detail in the second part of this chapter. First, we are concerned with the recruitment process, that is to say, the technical process of selecting candidates. This section considers the hypothesis that parties with a more centralised recruitment process will find it easier to implement parity, as they avoid the collective action problem that a decentralised process would entail. This hypothesis is put to the test by considering the recruitment process within each party. The means by which parties select their candidates are not uniform, but vary considerably from one party to another. Some parties select their candidates via a centralised bureaucracy, while some parties allow their local members full autonomy in selecting candidates. Most parties opt for a mixed system which allows local members a certain degree of input whilst still allowing for some centralised control of the process. The first part of this section will explore the theory behind the hypothesis in more detail, and explain how the varying degrees of centralisation might make parity more or less easy to realise. A more detailed study will then be conducted of the individual recruitment processes of the five main parties studied in this book. An analysis will then be offered of how each process enables or cripples attempts to implement parity. It will be argued that a degree of centralised control over candidate selection is a necessary, but not a sufficient, factor in implementing parity.

The significance of centralisation The degree of centralisation of a party’s recruitment process can be important in understanding how easily a party will implement parity

48 Parties, Gender Quotas and Candidate Selection in France

in legislative elections (Kittilson 2006; Matland & Studlar 1996; Norris 1993). The reason for this is simple; given that the electoral system uses single member districts, parity requires an equal proportion of men and women at the aggregate level. This requires national co-ordination in order to ensure that the male and female candidates put forward by different constituencies balance each other out. This is a common issue for countries using single member districts, and similar problems have been witnessed in countries such as the UK. Parties can regulate this problem in a number of ways. They can select all candidates via the national party bureaucracy, thus ensuring that the list of candidates is gender balanced; they can over-rule local choices of candidate in order to boost the number of women selected; they can reserve certain constituencies for women in advance so that local members will only be able to select from a pool of female candidates;1 or they can preapprove lists of suitable candidates for members to vote upon, and thus make sure that there are sufficient women within the candidate pool.2 Each of these methods, in ascending order of internal democracy, is deployed by one or more parties in France. By allowing the national party bureaucracy to have some role in the candidate selection process, parties minimise the risk of a collective action problem, whereby each constituency selects a man in the hope that another constituency somewhere else will select a woman to balance them out. A highly decentralised process may therefore be disadvantageous to the successful implementation of parity. A certain degree of centralisation is necessary to ensure that constituencies do not all rely on somewhere else to select the female candidates for the party. A certain degree of authority is also required to enforce parity if too many constituencies insist on selecting a man. Without this authority, the collective action problem would again arise, with each constituency that selected a man believing that they should be allowed to keep their candidate, and that the ‘sacrifice’ of switching a man for a woman should be made elsewhere. Although a degree of centralised control over the recruitment process is arguably a necessary condition for the achievement of parity (given the unlikelihood, under the present political climate, of constituencies spontaneously selecting an equal number of men and women), it is not in itself a sufficient condition. Although centralised control assists parties in implementing parity, it does not oblige them to do so. Nor does it always ease tensions in those constituencies that are expected to select female candidates. As discussed later in this chapter, many parties will wish to select more men than women, which inevitably leads to conflict if some constituencies are

Candidate Selection, Supply and Demand 49

allowed to field their candidate of choice while others are not. Even when the power to oversee the selection process is formally attributed to the national party bureaucracy, this power does not always confer the necessary goodwill or diplomacy required to persuade a disgruntled constituency to pass over its preferred candidate in favour of a woman. This is particularly the case in parties that do not benefit from strong ideological support for parity, as ideological commitment may be a mitigating factor in soothing the sting of fielding a candidate imposed from above.

Selection procedures in French parties The selection procedure for each of the five main parties in France will now be examined in greater detail. In each case, the key element of interest is the interplay of power between the local members and the national party bureaucracy in determining who has the first and last say on the selection of candidates. PCF The selection process within the PCF commences with a call for candidates across the party membership (although sympathisers who are not members might also be considered). As a general rule, aspirant candidates are put forward by local party sections3 or by the departmental4 executive committee. Individual candidatures are also possible but are unlikely to be successful, as the party has a cultural preference for candidates who are put forward by others. Nominated aspirants are assembled on a list that is circulated to the party membership within that constituency, and they are then invited to participate in a debate. Following this debate, a secret ballot takes place. All paid-up members resident within the constituency are entitled to vote. Once each constituency has selected their candidate, the departmental committee reviews the candidatures in an attempt to ensure parity. The aim is that each department should respect parity, thus automatically regulating the problem at the national level. If for any reason this is not possible, regional committees try to intervene to ensure that parity is respected. This can result in an ongoing dialogue between the various strata of party organisation to try and ensure parity, but their elections official claims that in the end ‘we get there, more or less’.5 Indeed, the PCF selected a total of 43.95 per cent women for the legislative elections of 2002 and 46.5 per cent in 2007, so they did come close to parity. However, the fact that most of these

50 Parties, Gender Quotas and Candidate Selection in France

women were in unwinnable seats6 suggests that the negotiations at departmental and regional level probably consisted of encouraging the constituencies with the least chance of winning to alter their choice of candidate, as these constituencies had the least to lose. Consequently, the chosen method of regulating parity was reasonably effective in terms of percentage of female candidatures, but was still of rather limited benefit to women. The party also acknowledged the primacy of its local members; intervention from higher party authorities consists of ‘dialogues’ rather than coercion, and the final choice of candidate remains in the hands of individual constituencies. This was the reason cited for the failure of the party to achieve perfect parity. Greens The Greens have several issues to take into consideration when selecting their candidates due to the relatively small size of the party. They are dependent on their alliance with the PS to obtain any seats at all. This alliance ensures that, in certain constituencies agreed prior to the election, either the PS or the Greens will not field a candidate so that their votes may be transferred to the other party. Given the difference in size between the parties, there are more constituencies reserved for the PS than for the Greens. However, this alliance is vital for the Greens as, without it, they would not win any seats. Their selection process therefore begins with negotiations with the PS to see how many constituencies will be reserved for the Greens, and how many other constituencies they will be allowed to field candidates in (if neither party has a realistic chance of winning the election, both parties will field candidates so as to maximise the number of votes received and hence increase the portion of state finance determined by national vote share). A second headache for the Greens is the attempt to ensure an even balance within the small number of candidates that they succeed in getting elected. They do not only have parity to contend with, but also a harmonious balance of the different factions present within the party. Unlike a party such as the UMP, the Greens only tend to win a small handful of seats, making it all the more vital that the select few who get elected are as representative as possible of the distribution of factions within the party. Bearing these restrictions in mind, the Greens invented a system whereby each constituency designates four candidates (comprising two men and two women), nominated by the electoral committee within that constituency. If the party does not have enough suitable candi-

Candidate Selection, Supply and Demand 51

dates to complete the list, potential candidates from other constituencies within the department will be invited to join the list. This enables each constituency to put forward a gender-balanced list of candidates and also provides an outlet for men or women who do not make the shortlist, as an opportunity for them may arise in a neighbouring constituency. This list of four candidates is then submitted to a vote by the members within that constituency. Once each constituency has selected a candidate, their choice is put forward to the national selection committee for validation. This committee tends to validate the local choice except in one of two circumstances; firstly, they regulate the most winnable seats to ensure an even distribution of party factions, as discussed above; and secondly, they regulate seats at a regional level to ensure that parity is respected. To remedy the problem cited by some constituencies of a lack of suitable female candidates, a national group of female candidates, nicknamed ‘Amazones’, was established. These women are prepared to relocate themselves to wherever they are needed in order to ensure that enough constituencies have female candidates for parity to be respected. (One Amazone went all the way to Martinique to represent her party!). With the exception of this group of women, however, it is very rare for the national party to impose a candidate that has not originated from the constituency in question. If a constituency is not satisfied with the request of the national party to change their candidate, they have a right of appeal, although the final say rests with the national selection committee. This has occasionally led to disputes, with threats of resignation from disillusioned activists. It is of note that, although the Greens work very hard to ensure that a parity of candidatures is respected, their primary concern within their few winnable seats is ensuring an even distribution of factions rather than an even gender balance. This explains why there was one woman out of three candidates elected in 2002, and one woman out of four in 2007. It is also important to note that the Greens place a high premium on their incumbents, as re-selecting an incumbent strengthens their bargaining position with the PS when trying to ensure that the constituency remains reserved for the Greens. PS The recruitment process within the PS commences with a call for members to put themselves forwards as candidates. Very occasionally (in fewer than 1 per cent of cases), a national figure may also be invited

52 Parties, Gender Quotas and Candidate Selection in France

to stand within the constituency. All aspirant candidates within a constituency, including the incumbent and/or national personality where applicable, are then voted upon by the party members within that constituency. The choice made by the local members is then put forward to the national committee for validation. As with the Greens, the local choice is usually validated, but may not be if the national committee disagrees with the choice of a candidate for a particular reason. The PS also resembles the Greens in that, due to their alliance, there are certain seats in which the PS does not field a candidate.7 This can be difficult for prominent would-be candidates within these constituencies, but they recognise that this alliance is important for the greater good of the party, as it allows the PS to expand its electorate in its target seats and hence maximise its chances of winning. The PS has an internationally-inspired solution for regulating the question of parity. As most constituencies, when left to their own devices, do not select women, the PS has begun to reserve constituencies for women, via the use of all-women shortlists. The number of reserved constituencies grows from one election to the next and is the party’s main tool in achieving parity. It should be noted that these constituencies rarely figure among the more winnable seats available, as these tend to be held by incumbents. However, the PS debated the idea of reserving their ‘swing seats’, namely those won in 1997 and lost in 2002, for women in the 2007 elections. These seats benefit from having no sitting incumbent (although the ousted incumbent will often wish to try and regain their seat at the next election), and were the party’s prime target seats for 2007. The party’s women’s section called for all swing seats that were lost at the previous election to be reserved for women at the subsequent election. Neither the Women’s Secretary nor the Elections Secretary was optimistic when interviewed about the chances of such a measure becoming compulsory, and with good reason. Instead, it was made a target, with the opportunity of fielding a male candidate where this was deemed to be the best choice. This flexibility allowed a greater degree of negotiation between the national party and the local constituency, by permitting the constituency to plea its case for a male candidate. UDF The UDF used a recruitment process that accorded far more control to the national executive committee than was the case in parties of the left. There was a national selection committee, drawn from members of the national party organisation, which oversaw the selection of

Candidate Selection, Supply and Demand 53

candidates to all elections. For legislative elections, potential candidates could be put forward for consideration to the national selection committee in one of several ways. Firstly, the president of the departmental party organisation could propose candidates. Secondly, the committee could send a team out into local constituencies to seek out talent and identify potential electoral candidates. Thirdly, some aspirant candidates could put themselves forward on their own initiative. Finally, people holding prominent local electoral office could also be considered for national office. Once all aspirant candidates had been chosen, the national selection committee then made its selection of candidates and took them to the national executive committee for approval. From time to time the executive committee would not agree with the choices made by the selection committee and would override their selections. So the selection committee did the majority of the groundwork, but it was the national executive committee who had the final say. What is notable in all this is the virtual absence of involvement from the local party membership. In very rare cases, where a constituency had two very strong candidates and the executive committee was undecided, the local membership of that constituency might be consulted for their opinion, but even in these isolated cases the opinions of the members were not binding, and merely served to help the executive committee make their choice. This lack of internal democracy was partly due to the loose-based and fragile nature of the UDF, which had only existed as a single party since 1998, was partially subsumed by the creation of the UMP in 2002 and then divided back into separate parties in 2007. The official reason offered, however, was that ‘the appreciation of the political value of a candidate should be made outside of any pressure from party activists’.8 The choice of candidate should be carefully considered and based on the qualities of the individual, with the candidate selected being the one most likely to perform well in the constituency rather than the one best able to mobilise support amongst activists in the event of a members’ vote. One factor that certainly did influence the choice of candidate was incumbency. Although it was not official party policy to reselect incumbents automatically, the party only deselected incumbents in one of two cases: firstly, if the incumbent had demonstrated signs of political weakness during their term that suggested that they might lose their seat, such as the loss of a local mandate during their parliamentary term; and secondly, if the candidate had grown too old but did not volunteer to retire. In the latter case, the party was more concerned with ensuring the

54 Parties, Gender Quotas and Candidate Selection in France

long-term health of the party within the constituency, and if this was best ensured by introducing a younger candidate to maintain the party’s presence, then the party could force the incumbent to stand down. In the former case, the party had to tread very carefully. For example, a lesson was learned in the senatorial selections when a candidate who had lost a local office was deselected in favour of a local mayor. The ousted incumbent stood as an independent candidate against the official party candidate and won the election. This experience served to reinforce the party’s reluctance to deselect incumbents. By contrast, parity played a relatively small part in the process of candidate selection. The party took the step of introducing gender parity for the internal party hierarchy, taking effect from 2006, which meant that an increased proportion of the selection and executive committees were female. There was rather less interest in respecting parity at the level of candidacies, however. Where there were two strong candidates, one male, one female, then the existence of the parity law might have persuaded the UDF to prioritise the female candidate. Nonetheless, where candidates were not evenly matched, the candidate deemed to be the strongest was chosen, regardless of sex. Respecting parity did not appear to be a matter of great concern for the party. Where it was possible to select more women (usually in unwinnable seats that no-one else wanted) then they would happily do so, but they openly acknowledged that they would not force the issue. Their elections officer was also confident that the question of parity was one that would resolve itself over time, as more women rose through the ranks thanks to the introduction of parity at a more local level.9 The party’s primary concern was choosing the electoral candidates that it perceived to be most likely to win, and this concern appeared to over-ride all other factors – internal democracy, appeasing members, appeasing incumbents, and parity. This might explain why, in 2007, not a single woman was elected from either of the parties emanating from the UDF (even though the proportion of women candidates rose as a result of the party’s sudden split and hence the urgent need to recruit extra candidates by the two new component parties). UMP The recruitment process of the UMP has changed from the procedure used by the RPR, under the guidance of the then UMP president Nicolas Sarkozy. The RPR had used a national selection committee that operated on an old boys’ network and did not have a strong appreciation of the needs of local constituencies.10 Subsequent reform has led

Candidate Selection, Supply and Demand 55

to party members being able to select candidates within their constituencies. These candidates are then scrutinised and validated by a national selection committee. This in turn reports to the Political Bureau, which is the main corpus of the party. Although the Political Bureau officially approves the process, it is too large and unwieldy and meets too infrequently to have any significant input into the process, meaning that ultimate decisions almost always lie with the national selection committee. It is of note that this selection committee comprises only three women members out of a total membership of 24, and both the chair of this committee and the national secretary for elections are men (www.u-m-p.org). Table 3.1 provides an overview of the key characteristics of candidate selection procedures within each party to facilitate comparison. Table 3.1

Overview of party selection procedures

Party

Who selects aspirants for list?

Members select candidate from list?

National selection committee (NSC) overrides choice of candidate?

Who has final say?

PCF

Local party organisations

Yes

No; may ask for change but cannot force it

Membership

Greens

Local party organisations

Yes

Yes, to ensure parity National party or balance of factions

PS

Individuals invited to come forward

Yes

Yes, if not satisfied with choice of candidate

National party

UDF

Mixture of local, national and individual

No; NSC decides

National executive committee may override selection committee

National executive committee

UMP

Party members

Yes

National selection committee oversees process, then validated by Political Bureau

NSC/Political Bureau

FN

National party

No; NSC decides

NSC is the highest authority, and includes the party president, general secretary and vicepresidents

NSC

56 Parties, Gender Quotas and Candidate Selection in France

Does centralisation of candidate selection assist parity? The above table does not provide clear support for the hypothesis that parties with centralised control over candidate selection are better able to implement parity. Of the parties under study, the only party to allow members the final say over the selection of candidates (the PCF) selected one of the highest levels of female candidates in the 2002 and 2007 elections (43.95 per cent and 46.5 per cent women respectively). By contrast, two of the parties with the highest degree of centralised control, the UDF and the RPR (who were responsible for the majority of the candidate selections for the UMP in 200211), were the same parties who fielded the lowest proportion of female candidates (in 2002, both parties fielded fewer than 20 per cent women, and in 2007 these proportions had only risen to 26.5 per cent for the UMP and 27.9 per cent for the Nouveau Centre). However, the hypothesis was that centralised control over the selection process would be a necessary, but not sufficient, condition for parity. Clearly, the latter is true – centralised control will not benefit women if parties do not use this control to impose parity upon local constituencies. What is less clear is whether centralised control is, in fact, a necessary condition. The Greens and the FN are good examples of how, when used effectively, centralised control of selection can be used to enforce the implementation of parity. Centralised control can therefore be considered as an enabling tool in parity’s implementation. Yet for it to be a necessary condition assumes that, left to their own devices, local party organisations will not be capable of selecting women in sufficient numbers. There is considerable evidence, however, to support such a hypothesis. All three parties of the left spoke of the need to intervene in order to ensure that more women were selected by local constituencies. In the case of the PCF, this intervention took the form of a ‘dialogue’, whereas the Greens actively intervened to force certain constituencies to change their choice of candidate from a man to a woman until the requirement of parity was met. If parity occurred organically, such measures would not be necessary. Moreover, the use of all-women shortlists by the PS was precisely because of the ‘difficulty of ensuring that women are selected by local activists’.12 Implementation of parity, whether it be at a local or centralised level, requires a degree of political will – parity must be perceived as a higher priority than other factors influencing the selection process. The reasons whether or not parity achieves this priority status are considered throughout this book. What is particularly important here is that, even where the will to

Candidate Selection, Supply and Demand 57

implement parity exists, its implementation is likely to entail certain sacrifices, such as passing over a prominent (male) mayor as a potential candidate in favour of a less prominent female candidate. Where such sacrifices are required, they are more easily made at the aggregate level, as the demand for sacrifices at the individual level is likely to produce a collective action problem. Each constituency risks ‘free-riding’ by selecting a man in the hope that another constituency will select a woman. By having a nationalised scrutiny of the process, the party can oversee which candidates are selected where, and determine which constituencies are best placed to field a woman. In the case of the PS, this decision is made in advance of selections to eliminate the problem of free-riding. For the Greens, the decision is made a posteriori. The PCF have acknowledged that the high degree of internal party democracy, with party members having the final say on candidate selection, has proved a barrier to the successful implementation of parity, as their members select more men than women and they cannot override this choice.13 Thus the theory suggests and the practice confirms that a degree of centralised control over the recruitment process is currently a necessary, though not sufficient, condition for the successful implementation of parity. One of the reasons why centralisation is significant is because of the cross-party tendency to select more male candidates than female candidates. As the next section demonstrates, this suggests that parties have either a supply problem or a demand problem concerning female candidates; in other words, either there are not enough good female candidates available, or parties are passing over good female candidates in favour of men.

The ‘supply and demand’ model in France This second section of the chapter will deploy a variant of the ‘supply and demand’ model proposed by Norris and Lovenduski (1995) to see whether it can provide an explanation for why parties do not select more women. The model suggests that there are two explanations for why parties do not select more female candidates. The first, ‘demandside’ explanation suggests that parties do not select more women due to discrimination. This discrimination can be direct, whereby their choices are actively motivated by the selectors’ attitudes towards women, or imputed, whereby selectors themselves have no objections to a female candidate but believe that another group, such as the electorate, would be likely to discriminate against a woman. The second,

58 Parties, Gender Quotas and Candidate Selection in France

‘supply-side’ explanation suggests that parties’ hands are tied because there simply are not enough qualified women available from which to choose. This implies that parties are not choosing men out of discrimination but out of necessity due to a lack of an alternative. This model is useful conceptually for distinguishing between the areas that parties control and those that they do not. If parties have good quality female candidates available, then their choice of male rather than female candidates is exactly that – a choice. If the women are not available, then the element of choice is removed and the shortage of women becomes a contextual factor that affects party choices rather than resulting from them. It will be suggested here that the two explanations – supply-side and demand-side – may in fact interact in a way not included in the original model. Political parties will select their candidates on the basis of certain criteria: a candidate’s political experience, their involvement in the party, their personal qualities. In some instances, these criteria may be invisibly gendered, because they are qualities that are more prevalent in men than in women, and may be the result of an unconscious valorisation of masculine or male-dominated qualities rather than a deliberate decision to favour men. In these circumstances, parties may genuinely believe that the problem is supply-side, because there are more men than women that meet their criteria for a good candidate. However, if these criteria are subtly gendered then the problem may also be partially demand-side, because parties are seeking a candidate portfolio that discriminates in favour of men. This hypothesis shall be tested here in a number of ways. Firstly, the various stages of the process of going from being a citizen to being a deputy are considered to see where and why the proportion of women decreases (women comprise 51.4 per cent of the population and 18.5 per cent of deputies (l’INSEE 2004; National Assembly)). According to the Norris and Lovenduski model, this process forms a ‘ladder of recruitment’, going from the electorate to party membership, then to the pool of applicants, then to being a candidate, and finally to becoming a deputy (1995, p. 16). Unfortunately, there is no French data available for the pool of applicants. Even if an estimation were to be made of how many people competed for each party in each constituency to be selected as the party candidate, it would not be known how many more people sought, unsuccessfully, to be placed on the list of aspirant candidates. So this chapter will look at the proportion of women voters, party members, candidates and deputies, by party.

Candidate Selection, Supply and Demand 59

A second test of the model will be to look at the criteria cited by party election officers as being sought after in an ideal candidate, to see whether these criteria are gender-neutral or more likely to favour one sex or the other. No party cited any factors that constituted direct discrimination (although this does not mean that it does not occur), and all parties also denied that the electorate would be more likely to vote for a man than a woman, suggesting that imputed discrimination was not a causal factor either.14 However, if the criteria for selection were subtly gendered, then this may have resulted in indirect discrimination. Finally, a study will be offered of women’s involvement in wider French society, to identify whether there is likely to be an adequate supply of women possessing qualities that might render them suitable for electoral office. Of course, it is not one individual quality or another that makes a person a good candidate, but rather their sum. Nevertheless, the wider the potential candidate pool, the higher the likelihood that strong candidates might emerge from the pool. Given that the National Assembly has 555 seats on mainland France, each party would require, at most, 278 strong female candidates, out of c.18,500,000 women eligible.15 This section considers why it might be so hard to find enough suitable women from such a large potential candidate pool. From citizen to deputy The process of going from being a citizen, eligible to stand for election, to actually being elected is not straightforward. There are numerous routes into the National Assembly; most, though not all, involve a degree of party activism, and it is virtually impossible to enter Parliament without affiliation to a political party (in 2002, for example, there were no independent candidates elected and all but nine deputies belonged to one of the five main parties examined in this book). Both the PCF and the UDF said that they would consider selecting candidates who were party sympathisers rather than party members. As this is rather difficult to measure, the non-party membership route to parliament will be considered under the third test of the model, namely the section looking at potential supply of candidates. This section shall focus on the traditional route through party membership. As stated above, the number of women ‘eligible’ for Parliament is some 18.5 million. The number of women who are members of the main political parties is, of course, considerably lower. Party membership in France is generally low at around 1.1 per cent (Knapp 2004, p. 354), with women less likely to be members of parties than men,

60 Parties, Gender Quotas and Candidate Selection in France

and ‘the number of women decreas[ing] as one climbs the party hierarchy’ (Allwood & Wadia 2000, p. 53). Women’s lower presence in parties is a long-term trend (Duverger 1995), with the male domination of parties possibly acting as a deterrent to women members and hence becoming a self-perpetuating vicious circle. Once women have joined a party, they can then rise through the ranks in a number of ways. One route is to seek a position of authority within the party. Another is to be well-connected within the party. Many women in the past have relied on the support of a powerful man to help promote them into the limelight (Appleton & Mazur 1993, p. 110), with the associated dependence that this entails. Sometimes this man may be a relative.16 Overall, however, ‘these relationships escape the writ of formal party rules, and … tend to discriminate against women’ (1993, p. 111). On a related note, sometimes women with appropriate educational qualifications (normally a degree from the prestigious postgraduate Ecole nationale d’administration (ENA)) may get promoted into the government without having previously sought electoral office, as was the case for Elizabeth Guigou (PS), for example. Women may also break into a career as a deputy by first being a suppléante (substitute deputy).17 However, the most common route into a parliamentary career is to build up a local political support base, for example through the acquisition of a local mandate. It has often been observed in French politics that the higher the position within politics, the scarcer the women (Helft-Malz & Lévy 2000; Pionchon & Derville 2004, p. 28). It has already been noted that women form the majority of the population yet a small minority of deputies. Table 3.2 provides a fuller account of how, as the political stakes get higher, the proportion of women gets lower. For all parties, the percentage of women members lies somewhere between 30 per cent and 40 per cent, thus forming a significant minority. This figure does not indicate causality (there are many reasons why parties might have fewer female than male members, such as women’s reduced availability, an off-putting male image of parties, or women’s preference for other forms of civil engagement), but the figure does demonstrate that, from the outset, parties have more men available than women. One could therefore make the argument that, if parties select candidates for electoral office from within their membership pool, a non-discriminatory approach would produce gender percentages of elected officials that approximate their gender percentages for party members.18 Table 3.2 demonstrates clearly that this is not the case. Party executives are the one area where the percentages of women

Table 3.2

Percentage of women in parties and politics Percentage of women by party and function1 Members

PCF Greens PS UDF6 UMP7 FN

37–40% 33% 34.45% 30% 36.7% 29%

Party Local executives office2

50% 35% 33%4 28% 22.2% 23.3%

41.6% 45.1% 37.1% 32.6% 31.9% 31.0%

Mayors2

9.5% 0% 5.9% 7.2% 6.4% 50%2

Conseils Généraux3

14.8% 8.3% 12.3% 11.8% 7.6% n/a3

Legislative elections 2002

Legislative elections 2007

Candidates

Deputies

Candidates Deputies

43.95% 48.91% 36.13% 19.68% 19.93% 48.4%

18.2% 33.3% 16.2% 6.9% 10.7% n/a8

46.5% 50.4% 45.52%5 27.9%6 26.5% 48.8%

16.7% 25% 26% 0%6 14.2% n/a8

Sources: Zimmerman (2003a, 2003b, 2008a), Genisson (2002); National Assembly; Ministère de l’intérieur 1 Member data comes from party interviews (source: Zimmerman 2003b) but, as data on membership is notoriously unreliable, this should be treated with caution. Data for party executives is also from 2003. The data for legislative candidates and deputies are from 2002 and 2007. All other data are from the 2001 elections, as the subsequent local elections (in 2008) took place after the legislative elections. 2 The data for ‘local office’ refers to municipal elections. The data for ‘mayors’ refers to mayors of municipalities having 3500 or more inhabitants. See Chapter 5 for a fuller discussion of the impact of parity in local elections. The figure for the women mayors in the FN is distorted by the low number of mayors elected; the proportion of women FN mayors in villages of fewer than 3500 inhabitants is a more modest 14.3%. 3 These are the departmental governing bodies, and one of the most powerful bodies in local politics. Data was not available for the FN in this category, presumably due to their insufficient presence. 4 The figure for the higher profile position of National Secretary is 40 per cent. 5 This is the combined figure for the PS and their electoral allies such as the PRG. The figure for the PS alone was higher. 6 The 2007 figures are for Nouveau Centre. 7 All data pre-2002 is for the RPR; data from 2002 onwards is for the UMP. 8 No deputies of either sex were elected for the FN in 2002 or 2007.

61

10.1057/9780230275294 - Parties, Gender Quotas and Candidate Selection in France, Rainbow Murray

62 Parties, Gender Quotas and Candidate Selection in France

are broadly in proportion to those of women members, except for the PCF, who observe parity for their internal organisation, and the UMP, who fall slightly short (36.7 per cent female members, 22.2 per cent women in the party executive).19 Once it comes to elected positions, the differences become starker. Women are fairly well represented in local (municipal) office, largely thanks to the effective implementation of parity in larger districts (although the lack of parity in smaller districts pulls the figures down from near the 50 per cent mark).20 However, women are not reaching the more powerful executive positions within local government in the same proportions. In particular, women are a very small minority in the most politically valuable office of mayor. As mayors are the most visible, prominent and powerful of local officials, the office of mayor is a useful resource when seeking to be selected as a legislative candidate. It is not surprising that, within the current composition of the National Assembly, more than half of deputies (339) are current or former mayors (National Assembly). Only 34 of these 339 (10 per cent) are women, and even this represents an increase from 8.1 per cent in 2001. Another useful office to hold when seeking selection is that of a conseiller général (departmental councillor). Elections to conseils généraux, like legislative elections, are held using a two-ballot single-member plurality system. The emphasis is therefore very much on the individual, and the individual of choice is more often than not a man. By failing to get elected to the conseils généraux, women are missing out on another foot on the ladder to a legislative career. Looking at deputies demonstrates that in the 2002 parliament, only 9.4 per cent of the 191 deputies who were also conseillers généraux were women. In 2009, 129 deputies also held departmental office, 17 per cent of whom were women, which is closer than the 2002 figure to the overall proportion of women in parliament. However, as with local politics, women are scarcer at the executive level, and this carries through to Parliament: of the 19 deputies who are currently presidents of a conseil général, only one is a woman.21 It is also of note that elections to the conseils généraux are one of the few elections where parity has not applied, and hence where the numbers of women have not risen significantly. This calls into question the validity of the claim made by the UDF elections official that women would naturally start to rise through the ranks and that no special measures were required to help boost their numbers, that it was simply a matter of time. If the numbers of women deputies are dependent on the (very slow) rises in the numbers of female

Candidate Selection, Supply and Demand 63

mayors and conseillères générales, it could be a very long time indeed before parity is achieved. An extra way to get into Parliament is through being a suppléant; in the 2002 Parliament, 4.2 per cent of deputies were suppléants, and a further 9.5 per cent of deputies had served as suppléants in the past; the figures for 2007 are 2.4 per cent and 7.6 per cent respectively. Hence this can be a useful route into the National Assembly. In 2007, women comprised nearly 38 per cent of suppléants, with several cases of parties trying to balance the ticket by offsetting a male candidate with a female suppléante. However, this potential opportunity for women has actually been exploited in a negative way by some parties. For example, Françoise de Panafieu (UMP) dryly noted that her party would offer women this ‘second-rate’ position as a way of giving women an obligation without giving them any real power.22 This impression was compounded by the fact that the deputies most likely to be selected for governmental office were also likely to have male suppléants. In the first government formed after the 2002 election, only four of the 20 male ministers were replaced with female suppléantes, and each of the four female ministers was replaced with a man. Similarly, in the first government formed after the 2007 election, all five women nominated to the government were replaced with a man while only three of the 12 men nominated were replaced with women (resulting in a decrease in women’s presence in parliament). This suggests that when the position of suppléant carries with it a genuine possibility of power, it is as elusive as any other route into Parliament. Finally, there is the selection of candidates for legislative elections themselves. The difference between the parties is starkest in terms of the proportion of women that they have selected for elections to the National Assembly. While parties of the left, and the PCF and Greens in particular, have a substantially higher proportion of women candidates than members, the reverse is true for the parties of the right. For all parties, the proportion of female candidates is far higher than the proportion of women actually elected, for reasons that are discussed in detail in Chapter 4. What is particularly interesting to note is that, with the sole exception of the UDF, the proportion of female deputies is higher for all parties than the proportion of women mayors or conseillères générales. What can be concluded from all of this? The first conclusion that can be drawn is that the first hurdle to achieving parity takes place long before candidates are selected, in that women consistently comprise a lower proportion of party members than men. However, this

64 Parties, Gender Quotas and Candidate Selection in France

difference in itself is not enough to account for the lower levels of women in high political office. This suggests that women encounter further barriers in the journey from party membership to national office. Key amongst these barriers is the fact that women are not penetrating the higher ranks of local office. Despite their significant influx into local politics in 2001, women are currently more predominant in the lower ranked positions that are less useful when seeking selection for national office. This cannot wholly be attributed to the fact that women have only arrived in local politics very recently, as even prior to the application of parity in 2001, women still comprised 21.7 per cent of local councillors (Zimmerman 2003b, p. 4). This therefore suggests that women face a barrier between the entry to local politics and the access to more prestigious, powerful local positions. This then has a detrimental effect on women seeking national office. Indeed, the fact that there are more female deputies (18.5 per cent) than mayors (9.5 per cent) or conseillères générales (13.1 per cent)23 suggests that women may be less likely than men to use this traditional route into politics, and more likely to use other routes described above, such as prominence within a party hierarchy, support from a powerful male politician, or success from outside a political party, for example in an NGO. Overall, women’s reduced involvement in political parties is compounded by their restricted access to the local positions that form a springboard for national office, meaning that the battle for a national seat is often lost well before the selection process even commences. This suggests that supply and demand factors interact to the detriment of women’s representation. Selection criteria As discussed above, there is more than one route to Parliament. This section considers what political parties look for in an ideal candidate, and whether these criteria are gender-neutral, or whether one sex might be more likely to meet the required profile than the other. If the former, then this would suggest that parties are faced with a problem of supply (or that they are not being fully honest when discussing their selection criteria, which is also a possibility when conducting elite interviews). If the latter, then parties may be discriminating against women even if they are not aware that they are doing so, resulting in a demand-side explanation. Although different parties cited different criteria in answer to the question ‘what do you look for in an ideal candidate?’, there were some common themes.24 Being well-known was the most commonly

Candidate Selection, Supply and Demand 65

sought characteristic, with parties of the left and the FN also seeking someone who was active within the party. Other qualities sought included having a first mandate, ideally won through an election fought under a majoritarian electoral system (such as an incumbent, or a conseiller général) (UDF); being well adapted to the local constituency (PS, PCF and FN); being able to deal with national issues (Greens); being able to draw together people from different political persuasions (PCF); being motivated (UDF) and hard-working (PS); being eloquent (Greens); being supported by a high-profile personality within the party (UDF); loyalty to the party and its policies (FN); and possessing leadership skills (Greens). Obviously, not all these qualities would carry equal weight, although one might assume that parties would ideally be seeking a candidate possessing a combination of these qualities. Let us consider whether each of these qualities might be deemed gender-neutral or not. The requirement of being ‘well-known’ is somewhat vague and can be satisfied in a number of ways. Holding a prominent local office such as mayor is a good way of being well-known within the constituency, with certain caveats – for example, being the mayor of a large town is clearly more advantageous in this respect than being the mayor of a tiny village. This placed women at a disadvantage in selections for the 2002 and 2007 legislative elections, for while women were 11.2 per cent of mayors in the smallest towns, they were only 6.7 per cent of mayors in towns with a population above 3500 inhabitants. In the largest towns (pop. 30,000+), there was a total of 20 female mayors and 232 male mayors (Zimmerman 2003b, p. 38). Even now, women only comprise 13.8 per cent of mayors (www.observatoire-parite.gouv.fr). So it is unlikely that, if it is a mayor that parties are seeking, women will be the first people to spring to mind. There are other ways of achieving prominence, however. Prominence within a party hierarchy may not be of much benefit when facing the electorate, but might well carry weight during the candidate selection process. Aspirant candidates might also be able to establish themselves through other, less overtly political acts within the community, such as campaigning for a local hospital or being a well-known business leader. This requirement is therefore only gendered if it is associated with a particular type of electoral mandate, as was the case with the UDF. Being active within the party is something that favours women less than men, as women comprise less than 50 per cent of party members and party executive members. On the basis of this criterion, one might

66 Parties, Gender Quotas and Candidate Selection in France

expect to see women in proportions closer to 30 per cent or 40 per cent rather than 50 per cent. However, the requirement of being welladapted to the local constituency is not necessarily gender specific and should not serve as a reason to discriminate against women. The requirements of being able to deal with national issues, being motivated and hard-working and being eloquent are all attributes that are based on personal qualities more than political experience. However, if discrimination were taking place, it is possible that a party selector might perceive men in general as being more likely to possess these qualities than women. Unfortunately, it is not possible to tell whether or not this is the case. What is clear, however, is that these requirements are not automatically gendered; any possible discrimination in this regard would be on the basis of prejudice rather than fact. Being supported by a high-profile personality within the party is a quality that is potentially gendered. Although receiving the backing of a ‘personality’ within the party has enabled certain women to break through barriers and accede to Parliament, it has also enabled many men to impose themselves in constituencies where a woman might otherwise have been chosen. A party that encourages this strategy is also encouraging a system of patronage that often results in an ‘old boys’ network’ that works to the disadvantage of women. Party loyalty is another quality which is based more on the individual than on gender. As men form a higher proportion of party members, it might be easier to question the loyalty of women who have been attached to a party for less time or who have had less time to devote to party campaigns. On the other hand, a study of women MPs in the British parliament revealed that they were more likely than men to toe the party line and less likely to rebel (Childs 2004; Cowley & Childs 2003). Finally, the quality of ‘leadership skills’ is one that parties are likely to judge on previous demonstrations of leadership. Although this quality is not automatically gendered, it might favour men over women. For example, as will be demonstrated in the next section, women are significantly less likely than men to hold the most senior positions in either business or politics, and will therefore have had fewer opportunities to demonstrate their leadership ability. Overall, none of the qualities listed by parties as being desirable in a candidate is directly gendered. However, several of them may be less favourable to women than to men. Women are less likely than men to hold prominent local electoral office that would help them be ‘well known’, and are particularly scarce in the valuable positions of

Candidate Selection, Supply and Demand 67

conseiller général and mayor of a large town. They are also less likely than men to have been active within their party, as there are fewer women than men who are members of political parties. Women have reduced access to leadership positions, both in politics and within wider society, thus rendering the requirement for ‘leadership skills’ unfavourable to women. The potential to be steam-rollered in by a big name within the party is also more likely to work against women than for them. The remaining requirements are more individualised and due to personal attributes, and are not gendered in themselves, although their subjective nature leaves scope for discrimination, especially where party selectors are using stereotypes to help make speedy judgements about people they do not know well (Norris & Lovenduski 1995, p. 14 & p. 107). As none of the requirements listed was overtly gendered, it is unlikely that parties are selecting men over women due to a deliberate, calculated preference for men (although if parties were to do this, they would be unlikely to admit to this during an interview; but let us give them the benefit of the doubt). However, as several of the requirements were subtly gendered, that is to say, more likely to favour a man than a woman, it is possible that what appear to be supplyside problems (not enough women meeting the requirements of a good candidate) may actually be linked to demand-side problems (parties are seeking qualities that are more prevalent in men). Given that the qualities that are gendered tend to be qualities requiring experience that women are less likely to possess, it is worth reflecting on why it is that women are less experienced. A partial explanation has already been provided above, in that women are less likely to accede to the political positions that act as a springboard to parliamentary office. This still leaves the wider questions of why women are less likely to get involved in the first place, and whether women lack the resources that would help them work their way to the top of the political ladder. The third test of this model seeks to answer these questions. Social and economic factors One factor that was frequently mentioned in interviews with parties was that women tended to be less available than men and have less time to dedicate to politics. This is symbolic of a wider pattern of social inequality that may serve as a barrier to women seeking a career in politics. This section will explore whether the biggest explanations of women’s absence from parliament are actually located beyond

68 Parties, Gender Quotas and Candidate Selection in France

political parties and are due to women’s place in wider society. If this were the case, then the key reason for the shortage of women would be due to supply-side problems. The variables to be considered are education, employment, domestic obligations and forms of political participation other than party membership. Education According to Allwood and Wadia (2000), ‘studies revealed that political interest and voting increased with the level of education attained by women’ (p. 141). It was therefore believed that education would be a crucial factor in enabling women to participate more actively in politics. In this regard, women’s current educational status initially seems very promising. From an early age, French girls are outperforming boys in the classroom, with teenage boys being significantly more likely than girls to have to repeat a year at school, while girls are more likely than boys to be a year ahead of their studies (l’INSEE 2004, p. 57). However, there is a marked difference in the career orientations that boys and girls choose, with girls tending to go for arts subjects and training in areas like office and secretarial skills, while boys are far more likely to favour subjects such as electronics. Girls also comprise 82.9 per cent of students sitting the literature-oriented baccalauréat, but only 11.6 per cent of those taking the engineering option, despite the fact that girls form the majority (55.4 per cent) of baccalauréat students due to their superior performance at school (ibid, p. 61). Women are also more likely to go to university than men, comprising an ever-higher proportion of students (56.9 per cent in 2007–8, according to l’INSEE 2008). However, the picture is not all rosy for women. The differences in career orientation that develop at school carry on into adulthood and help to confine women within certain sectors, as the next section demonstrates. For example, women comprise 73.4 per cent of arts students, but only 45.5 per cent of science students, despite their overall numerical superiority (l’INSEE 2008). Women are also less likely to access the most elite universities, with the ENA being a prime example of this. The ENA is a breeding ground for political elites, and many of France’s politicians and top civil servants have passed through its doors. Yet women comprised only 27 per cent of new entrants in 2003, and even this figure is an improvement on some previous years25 (www.ena.fr). Overall, the indications are that women are at least as able and motivated as men, but may be choosing (or being directed into) gendered career paths from an early age.

Candidate Selection, Supply and Demand 69

Employment One of the most prominent indicators of women’s participation in society is their activity level, and it has been argued that women need to participate in the labour market in order to develop the social and economic capacity to participate actively in politics (Allwood & Wadia 2000, pp. 141–3). France appears to do well in this regard, having one of the highest rates in Europe for women’s economic participation, with women comprising 47.2 per cent of the workforce (l’INSEE 2008).26 However, this figure again requires qualification. Although many women work, 30.2 per cent of them work part-time (compared to 5.7 per cent of men), and women occupy 82 per cent of all part-time jobs (ibid). There are also certain sectors of the economy that are very gendered, following on from the gender gap in education noted in the previous section. Women are very over-represented in certain, less prestigious, sectors (administrative, service-based and so on) and under-represented in others (such as finance, science and technology). This goes some way towards explaining the prevalent pay gap in France, with an average hourly pay gap between men and women of about 19 per cent (ibid, pp. 94–5). Significantly, the pay gap increases as the job’s status rises, meaning that women in the more highly qualified, better remunerated jobs that might lead towards political office are earning substantially less money than their male colleagues (l’INSEE 2008).27 Women are also confronted by a glass ceiling, with barely 5 per cent of France’s largest companies being run by women (Belot & Rocco, Le Monde 08 March 2001; l’INSEE 2008). Women’s lower rate of pay, combined with their higher rate of unemployment (female unemployment is consistently 2–3 percentage points higher than male unemployment) are both symptomatic of the idea that it is still men who are the primary bread-winners in a family, even though women-headed single-parent families now comprise 15.4 per cent of households in France (l’INSEE 2004, p. 29). It is this mentality that pervades women’s careers, in both the economic and political spheres, and remains an important explanation of how women have not advanced further. Domestic obligations In keeping with the image of men as the primary bread-winners, women are seen as the primary carers in French society (as in most other countries), and this is reflected in the division of domestic labour. Where women stay at home, they spend an average of 6 hours and 45 minutes per day on domestic chores and childcare, compared

70 Parties, Gender Quotas and Candidate Selection in France

to 1 hour and 52 minutes by their male partner (l’INSEE 2004, p. 119). Even where women work full time, they still spend nearly four hours a day on chores compared to only 2 hours 14 minutes a day for men. Consequently, women on average work shorter hours at their jobs than men, resulting in less dedication to their careers and lower prospects for advancement, and even then women enjoy nearly an hour less of free time per day than men. In both of these respects, women are placed at a disadvantage compared to men as they are less available to pursue their economic and political careers. This was reflected in the comments made by political parties, who lamented the fact that women struggled to attend party meetings and events, and therefore did not gain the required profile to make them suitable for consideration as parliamentary candidates. Female deputies also felt that the ‘double day’ of professional and domestic life took its toll on women, physically and psychologically. The head of the Greens’ women’s committee commented that ‘women with children feel guilty if they do not spend enough time with them. There is a social pressure that is both exterior and internalised. The timing of meetings is also incompatible with a family life; this needs to change’.28 Her socialist counterpart concurred, remarking that building a political career is time-consuming and ‘because of the division of domestic duties, women are often less available than men’, thus damaging their political careers.29 The burden of childcare responsibilities means that women tend to start their political careers later in life after having a family, leaving them less time to climb the political ladder (Dewavrin 1994, p. 131; Sineau 2001, pp. 215–16). Alternative forms of political participation Given that women are less involved in political parties than men, despite their higher levels of education, it is possible that they are preferring to express their political interests in other forums. This may provide them with an opportunity to gain the political skills, networks and social capital to advance in politics from outside the conventional party route. Norris et al (2004) identify four dimensions of political participation: voting, campaign-oriented, cause-oriented and civicoriented activism. Unfortunately there is not sufficient data available to consider each of these dimensions here, but a brief examination of voting and civic-oriented activism will be offered. As with many other countries in Europe, France has witnessed a gender gap in voting, with women traditionally voting to the right of men (Helft-Malz & Lévy 2000, pp. 37–8; Inglehart & Norris 1999). Like

Candidate Selection, Supply and Demand 71

many of her neighbours, France has seen this trend gradually reverse over recent years, and women now vote slightly to the left of men. Perhaps of most interest in terms of their participation, women have become increasingly likely to use their vote, with the increased rate of abstention that was witnessed amongst women in the years after their receiving the vote gradually disappearing to the point that both sexes now vote in equal proportions. By contrast, women are less likely than men to participate in civicoriented activism. The overall rate of participation is 45 per cent for men compared to 40 per cent for women, although there are also notable gender differences in the types of participation (l’INSEE 2004, pp. 122–3; l’INSEE 2008). For example, men are significantly more likely than women to be involved in a sporting association or an environmental lobby, while women are more likely than men to be involved in a religious group, residents’ association or humanitarian group. However, these latter groups are at least as likely, if not more so, to confer political skills and networks to their members than a sports club. Nonetheless, in the most politically useful positions of chair, treasurer or leader, men are more prevalent (30 per cent compared to 20 per cent) while women are more likely to be a beneficiary or simple participant (60 per cent compared to 50 per cent men) (l’INSEE 2004). The proportion of women presiding over an association has risen to 31 per cent, but women are still more likely to be the secretary (57 per cent) (l’INSEE 2008). These results suggest that women do show the same level of interest as men when it comes to voting, but are less likely to get involved in civic-oriented activism, perhaps due to the greater constraints on their free time detailed in the previous section. This may also explain why women are less prominent in the more time-consuming positions of responsibility. However, their lower levels of involvement, particularly in the positions more likely to be of benefit to a political career, suggest that women are once again less present than men in an area of value to legislative recruitment. This reinforces the supply-side explanation of political recruitment. Evaluating the supply-demand model This section has looked at three different tests of the supply-demand model. The first test, which examined the process of becoming a candidate, demonstrated that women’s political careers face a string of hurdles along the path to national election, which go some way to explaining why women are less likely than men to reach the end of this path.

72 Parties, Gender Quotas and Candidate Selection in France

Many of these hurdles affect women more than men, such as the fact that women are not gaining access to the most politically useful offices in local politics, and the fact that many women are being offered a position as a suppléante rather than a candidate (and, more often than not, as a suppléante to someone who is unlikely to be nominated to the government). Women face an additional barrier in that, where they are selected, they are disproportionately fielded in unwinnable seats. This test therefore suggests an interaction of supply and demand explanations. Women are in shorter supply than men in the traditional political recruiting grounds of mayors and conseillers généraux. However, their presence in the party hierarchy and in local politics suggests that their absence from these springboard positions may be a demand-side problem. This theory is reaffirmed by the fact that women are disproportionately placed in unwinnable seats, suggesting that even when parties do recruit women, they are unwilling to hand over any real power. These demand-side explanations at lower levels of politics then cut off the supply of women at the more elite levels, meaning that although there is indeed a shortage of the type of candidate that parties are seeking, it actually appears to be a consequence of party attitudes towards women lower down the ranks. The second test looked at the qualities sought by party selectors to see whether they were neutral or gendered. It demonstrated that none of these qualities was overtly gendered, but that several of the qualities sought after were more prevalent in men than in women, meaning that the selection process eliminated more women than men from the candidate pool. This again demonstrates an interaction of supply and demand, as the demand is not for men per se, but for characteristics more often found in men, thus resulting in a lower supply of potential female candidates. It is unlikely that parties are aware of this interaction, and probably see the problem purely as one of supply. Finally, the third test looked at women’s wider involvement in society to see whether this could explain their lower levels of participation within political parties. Women are actually better educated on the whole than men, and should therefore benefit politically in this regard, but their attainments are tempered by their tendency to be within ‘feminised’ fields that are valued less by society, and by the fact that they are less likely to be accepted into the most prestigious universities such as the ENA. Women are also being held back in the professional sphere, with a high activity rate being confounded by a concentration within part-time jobs and lower paid sectors of the economy that both contribute to a significant gap in earnings between

Candidate Selection, Supply and Demand 73

men and women. In terms of alternative forms of political participation, women are now as likely to vote as men, but still less likely to participate in civic-oriented activism. Their lower attainments in the work force and their lower involvement in both political parties and civic organisations appear to stem from the crucial factor of the division of domestic labour, which falls primarily on the shoulders of women. As long as women are doing more than their fair share of housework and childcare, they will be prevented from putting in the necessary time and commitment to build a political career, whether this be in a political party or an association. In this regard, therefore, women’s lower levels of participation can be attributed to a supply-side problem that is independent of political parties. Overall, it is clear that parties are not entirely responsible for the shortage of women in politics, but they do appear to compound an existing problem. Supply-side explanations account for why women are less likely than men to join a political party and to take on positions of responsibility, but they do not account for why women’s political careers appear to be stifled at the local level, nor for why women are placed in second-rate positions (unwinnable seats and/or suppléantes) at the national level. It is also true that, even if women are present in lower numbers than men within political parties, they are still present in sufficient numbers to provide a candidate pool that should, at the very least, permit their election to parliament in proportion to their activism in political parties. This is confirmed by women from both the PS and UMP, who claimed that women were available as long as there were genuine opportunities for them, and that there were more women seeking to be deputies than there were seats available.30 However, the valorisation of male attributes and the prioritisation of men for the key positions in local and national office both suggest that demand-side explanations compound the supply-side explanations to result in the lower number of women who succeed in being elected to the National Assembly. Conclusion Understanding how candidate selection works is central to any study of a gender quota. The first section of this chapter discussed the recruitment process within each party, and demonstrated the importance of having some degree of central control over this process. Centralisation assists in the implementation of parity by viewing parity from an aggregate perspective, and allowing parties to control for collective action problems. However, a centralised recruitment process does not,

74 Parties, Gender Quotas and Candidate Selection in France

in itself, guarantee that parity will be implemented, as it is only effective when accompanied by the political will to enforce parity from the top downwards. Rather, centralisation is an enabling tool that will help parties who wish to take advantage of it, such as the Greens. The second section discussed the ‘supply and demand’ model of political recruitment to see whether the shortage of female deputies was due to discrimination by parties, a shortage of women, or a combination of the two. The findings suggest that the career structure within parties does not work in favour of women, with women frequently falling at hurdles in the earlier stages of a political career that prevent them from reaching the ultimate goal of a seat in parliament. Moreover, although parties do not appear to deliberately discriminate against women, they prioritise qualities that are more prevalent in men when selecting electoral candidates. Both of these findings demonstrate an interaction of supply and demand explanations; there is an inadequate supply of women meeting the criteria for electoral candidates, partly because their political careers are being stymied lower down the ranks, and partly because the selection process is subtly gendered. There is also a wider supply-side explanation of women’s low presence due to their inferior economic status, lower levels of civic engagement and their reduced availability for political activities. These all stem from an unequal division of domestic labour, leaving women less time for other activities. Overall, supply-side explanations do account for a significant degree of women’s lower levels in politics, but these are compounded by the behaviour of political parties, resulting in a female presence in politics that starts low and gets smaller as the stakes rise. This chapter has examined parties’ recruitment processes and the candidate pool that is available to parties. It is now possible to advance to theories of party organisation and how these interact with the factors we have examined here to determine the choice of candidates that each party will make at legislative elections. Such is the purpose of the next three chapters.

4 The Electoral Competition Approach

Introduction This chapter is the first of three that each explores one of the main theoretical approaches drawn from the literature on political parties. In each chapter, the key features of the relevant approach are summarised and then applied to the French case study to see how well each can explain parties’ choice of election candidates. The first of these approaches and the topic of this chapter is the electoral competition approach. Ware neatly summarises this approach as follows: ‘parties are conceived as self-interested actors responding to the logic of the situation in which they find themselves – a logic that is dictated by the need to compete for votes’ (Ware 1996, p. 9). In this chapter, the party competition approach is used to consider whether electoral competition between parties is the prime motivation in explaining party behaviour. Electoral competition theory argues that parties are organised around the central goal of winning as many votes as possible, and will modify their behaviour in whatever way necessary in order to achieve this goal. For example, the Downsian model refers to the malleability of party ideology in order to maximise electoral appeal (Downs 1957). The approach requires further specification for the purpose of this chapter, where it is applied specifically to the area of candidate selection. If the approach is correct in claiming that parties’ prime concern is winning as many votes as possible, then it follows that parties’ top priority in selecting candidates would be to select the candidates most likely to win. This chapter explores the reasoning behind this approach, before examining its practical application. The desire to win votes is a central goal for most political parties. That parties wish to win votes may therefore be taken as a given. The 75

76 Parties, Gender Quotas and Candidate Selection in France

distinctiveness of this approach lies in its suggestion that this desire to win votes will be parties’ first priority and will override all others. This implies that parties will be prepared to put aside other objectives, such as ideological coherence or appeasement of their membership, in favour of winning votes. If parties are to be perceived as rational actors for whom maximising their electoral appeal is the ultimate goal, then other priorities may be ordered and structured in terms of how they help to fulfil this ultimate goal. Thus, if adopting a certain policy assists the party in winning votes then they will adopt this policy, but for the purpose of enhancing their electoral performance rather than out of other motivations. When applied to the question of candidate selection, parties will prioritise the candidates that are most likely to win, irrespective of other factors. Whether or not parties choose to respect parity would, in these terms, be determined by whether or not respecting parity would be likely to affect their electoral performance. This chapter will begin with a discussion of the competition approach, mapping out its origins and its relevance to the questions addressed in this book. The approach will then be evaluated from a number of angles. Firstly, a more detailed consideration will be offered of why French parties might place a high emphasis on winning votes. Secondly, an evaluation will be provided of why parties might consider certain types of candidate to be more electable than others. This evaluation will then be put to the test by seeing which types of candidate parties actually prioritise, and whether these candidates are, in fact, the most electable. If parties are consistently selecting the most electable candidates, then the electoral competition approach could be considered a useful explanation of parties’ choices of candidates. If parties are not prioritising candidates according to how electable they are, then either parties are not acting solely out of electoral motives, or they do not have an accurate perception of the type of candidate that will appeal to their electorates.

Overview of the electoral competition approach Advocates of the electoral competition approach argue that parties exist in order to win elections, and party policy is responsive to what voters want rather than what the party believes is best. Parties are not seen as programmatic or ideological, but rather as power-seeking and vote-maximising.1 They will be prepared to modify their behaviour in whatever way necessary in order to achieve their goal of winning as many votes as possible. This section will explore the origins and argu-

The Electoral Competition Approach 77

ments of this approach, and demonstrate how it can be applied to the study of candidate selection. Some of the earliest work on parties as electoral machines was by Schumpeter (1942), whose ‘innovation was to render irrelevant all motivations of political elites with the exception of their shared appetite for power’ (Bartolini 2002, p. 86). This led to the landmark text on electoral competition theory by Downs in his work entitled ‘An Economic Theory of Democracy’ (1957). Downs assumed that parties were rational, self-interested actors and argued that parties would rationally choose the most efficient path towards the achievement of their goal which, he claimed, was electoral office: Our model is based on the assumption that every government seeks to maximise political support. We further assume that the government exists in a democratic society where periodic elections are held, that its primary goal is re-election, and that election is the goal of those parties now out of power. (Downs 1957, p. 11) The term ‘primary goal’ indicates that election will be prioritised over all other factors. Downs provides a model based on Hotelling’s economic analysis which claims that parties can be located, in terms of ideology, on a linear left-right scale, and the desire to broaden their electoral appeal to the widest possible audience will result in parties’ converging on the centre. Ideology is important to parties only to the extent that its successful adaptation will win them votes. Downs explains that: We assume that [parties] act solely in order to attain the income, prestige and power which comes from being in office. Thus politicians in our model never seek office as a means of carrying out particular policies; their only goal is to reap the rewards of holding office per se.2 They treat parties purely as a means to the attainment of their private ends, which they reach only by being elected. (ibid, p. 28) This echoes Duverger’s description of parties as ‘purely electoral organisations, of which the prime function consisted in ensuring the success of their candidates: the election was the end and the party the means’ (Duverger 1964, p. 366). While Duverger then goes on to contrast these electorally-motivated parties with parties whose primary purpose was the dissemination of ideologically coherent policies,

78 Parties, Gender Quotas and Candidate Selection in France

Downs sees ideology more as a means for parties to distinguish themselves from their competitors and hence capture a specific electoral market. Parties thus become vote-maximisers: ‘each party seeks to receive more votes than any other’, and hence, ‘the party…manipulates its policies and actions in whatever way it believes will gain it the most votes’ (Downs 1957, p. 31). Sartori develops this theory and makes some important modifications, including providing a better account of spatial competition in a multi-party system as opposed to the Downsian model that worked best in a two-party system (Sartori 1976). The theory was also expanded by Kirchheimer to incorporate the notion of ‘catch-all parties’, namely parties that seek to appeal to the widest possible audience (Kirchheimer 1966). Catch-all parties are typified by a flexible and inclusive programme that adapts to the electoral market and seeks as broad an appeal as possible. These parties will aim to follow the public mood rather than leading it, and typically have low levels of party membership and broad socio-economic bases, in contrast to Duverger’s membership-based ‘mass parties’. The concept of catch-all parties has become widely adopted in the literature (although frequently criticised, the concept endures), and in France, the centre-right UMP party and centre-left PS party can both be argued to be ‘catch-all’ parties (Haegel 2005; Knapp 2002). This is perhaps slightly more the case with the UMP; the PS is torn between its ideological foundations and the practical imperatives of office-seeking, and is more likely to co-operate with than to co-opt rival parties. Nonetheless, its willingness to make a number of compromises over the years (in areas such as privatisation) in order to maximise its electoral appeal indicate its shift towards catch-allism. When applied to the introduction of parity as a policy, the electoral competition theory suggests that parties would only introduce such a law if they believed that doing so would win them votes, rather than out of a deeper ideological belief in the need for gender parity. This would certainly explain why parties’ enthusiasm for introducing the law exceeded their enthusiasm for implementing it; if they believed that voters would approve of parity in principle, but would be indifferent to or unaware of its application in practice, then the law might be entirely manufactured to broaden the electoral appeal of the parties behind it, with no deeper purpose. Moreover, the electoral motivations of each party to implement parity might be influenced by the level of competition surrounding the policy. Parties of the left might be able to afford to miss the benchmark of parity provided they still fielded more

The Electoral Competition Approach 79

women than parties of the right. Meanwhile, parties of the right might have felt that they could not gain from parity as an issue even if they were to respect it faithfully, as it was an issue that was ‘owned’ by the left. According to Budge et al (1987), certain issue areas will always play to the advantage of one party over another, regardless of the discourse surrounding the issue. For example, the British Conservative party is seen as the stronger party on the issue of law and order, and it is therefore pointless and even disadvantageous for the Labour party to broach the topic, as ‘even mentioning the topic runs the danger of bringing it into prominence and thus benefiting their rivals. Rather than arguing about it, the best strategy is not to mention it at all – or if some reference must be made in view of public concern, to deal with it cursorily’ (Budge et al 1987, p. 25). The consequence is that ‘parties talk past each other, glossing over areas which might favour their rivals while emphasising those on which they feel they have an advantage’ (ibid, p. 24). Considering that left-wing parties are generally seen as stronger on issues of equality, and that the momentum for parity came from the left (both the law itself and earlier attempts at legal and party quotas were initiated by left-wing parties), right-wing parties may have felt that their best option was to skate over the issue, and focus on policies that would be electorally rewarding for them. The literature thus paints a picture of parties competing in a linear spatial model for a maximum vote share, with ideological programmes playing second fiddle to vote maximisation.3 The approach does not make any specific reference to electoral candidates; it is assumed that voters identify with parties on the basis of policies rather than people. However, if the policies are of secondary importance and the key issue is getting elected, it follows that parties will seek to select the candidates that are most likely to win. One of the arguments that Downs puts forward is that rational voters will only have limited information about parties, and the cost of obtaining information often exceeds the benefit. It then follows that most voters will know very little about the candidates that they are voting for, and may focus only on the characteristics that are most salient to them. Where no other information is available, voters may only notice the candidate’s party label and their most obvious physical characteristics such as their sex and, in some cases, race. Hence candidate sex becomes important by virtue of its visibility; in the absence of superior knowledge about a candidate, voters will make a decision on the basis of the information available. They may also make assumptions about the candidate based on the little information that they do know. For example, Norris and

80 Parties, Gender Quotas and Candidate Selection in France

Lovenduski argue that pre-judgements may be used as an ‘information shortcut’, with certain characteristics being used ‘as a proxy measure of abilities and character’ (1995, p. 107). Yet the visibility of candidate sex does not necessarily make it significant. If voters do not discriminate in any way between men and women candidates then sex will be of no consequence. This raises the important question of whether or not voters evaluate candidates any differently on the basis of their sex. If one sex were regarded more favourably by voters than the other, then parties would be rational to prioritise candidates of that particular sex, in accordance with the notion that parties will follow the electorate’s preferences in order to raise their vote share. Such behaviour assumes that parties will know whether or not candidate sex will affect voter choice, and if so, in what way and to what extent.4 A second, and related, point is that a change of candidate would represent an increased information cost to voters. A candidate that had represented a constituency in the past would be better known to voters and they would therefore be able to judge such a candidate on more than just the obvious traits described above. Duverger argues that the more a candidate is known within a constituency, the more the election will be about the personality of that candidate, whereas when the candidate is not known personally to voters, they will vote on the basis of party label (1964, p. 357). Where this personal vote for the candidate is likely to be positive, it may represent an electoral boost over and above the vote for the party. Replacing a known candidate with a newcomer might cost the party this electoral boost and, depending on the marginality of the constituency, this might cost the party the seat. Regardless of how talented and qualified a new candidate might be, these talents and qualities will be of no consequence if voters do not become aware of them in time for an election. As political renewal is a necessary side-effect of parity, parties would need to replace many wellknown male candidates with relatively unknown female candidates. If local notoriety, normally achieved through current or previous occupation of the seat in question, is influential to party success, then electorally-minded parties would be rational to favour established candidates over newcomers.5 In order to assess the utility of the electoral competition approach, it is necessary to see whether there is any difference in the electoral performance of men compared to women, and of established candidates compared to newcomers. This will be addressed in detail later in the chapter. First, it is necessary to explain why French parties might be so

The Electoral Competition Approach 81

concerned with winning votes. As the next section demonstrates, the peculiarities of the French political system exacerbate the emphasis on votes, both for obtaining electoral office and for wider political purposes.

Electoral competition in France In order for the electoral competition approach to be applicable, French parties would need to organise their behaviour around the central priority of winning elections. This section considers why winning votes is of such importance to French parties, and how this might impact on candidate selection. Why every vote matters For most political parties, winning votes is a means to an end, with winning seats being the ultimate goal. In France, maximising the number of votes received fulfils not one but several core objectives of a political party. Firstly, there is the obvious target of winning seats, and hence gaining representation within a legislature (and, for larger parties, there is the additional goal of winning enough seats to control the legislature). For some parties, particularly those on the left, maximising their vote share serves a secondary purpose – to improve their position when negotiating for an electoral deal with a coalition partner. Thirdly, every vote received in a legislative election translates into a portion of a party’s state funding. Parties therefore have multiple reasons for wishing to receive as many votes as possible. Each of these reasons is explored in more detail below. Winning seats The wish to win seats is the most self-evident reason for wishing to win as many votes as possible. France has long had a reputation for electoral volatility, and from 1978 to 2002 each newly-elected government was voted out again at the next election. As a consequence, France has a large number of swing seats where each vote is at a premium (on average, 36.2 per cent of seats in the National Assembly swung between the left and the right at each election between 1993 and 2002, with a massive 41.8 per cent of seats changing hands in the ‘vague rose’ of 1997).6,7 This level is very high compared to the UK, where even in 1997, when the general election brought about a change of government, the percentage of seats to change hands was 29.7 per cent, and this figure has been much lower for other recent elections.8 As

82 Parties, Gender Quotas and Candidate Selection in France

Table 4.1 demonstrates, a large proportion of seats at any given election is held with quite a narrow majority. Bartolini argues that the availability of seats – that is to say, the openness of a seat to more than one party, typical of a marginal seat – increases the level of electoral competition (2002, p. 93). Furthermore, in the literature on candidate effects, findings suggest that individual candidates may have a small effect on the percentage of the vote received by their party (Cain et al 1984; Coates 1995; Norris et al 1992; Norris & Lovenduski 1995). In safe seats, this ‘personal vote’ effect will be insignificant, but the more marginal the seat, the greater the emphasis on choosing the candidate that will win, rather than lose, those all-important extra few votes. Thus the high electoral volatility in France increases the number of constituencies where every Table 4.1

Vulnerability of seats in French elections Percentage of seats affected per election

Swing required1

1993

1997

2002

2007

2012

<1% <3% <5%

15% 38.7% 58.2%

10.3% 31.8% 51.3%

12.8% 36.6% 57.2%

11.6% 30.2% 44.4%

9.7% 26.0% 47.1%

Average swing required Mean Median

5.04% 4.16%

5.65% 4.75%

5.26% 4.11%

6.80% 5.83%

6.16% 5.25%

Actual swing at election (national mean)2 Mean swing

8.48%

–7.91%

4.91%

–1.48%

n/a

Number of seats actually overturned (out of 577) Swing seats 1

220

241

165

733

n/a

The swing required to overturn the seat at the next legislative election (for example, the figure for 1993 is based on the election results for 1988, and shows the swing required to overturn the seat at the 1993 election). These figures are based on results for the second round of voting and are cumulative; for example, the figure for seats requiring less than 5 per cent swing includes the seats requiring less than 3 per cent and 1 per cent. 2 The ‘actual swing’ for the 1993 column is based on the national mean swing that occurred in the 1993 election. It should be borne in mind that this is a mean and does not take into account significant local variations. A positive figure indicates a swing from left to right; a negative figure indicates a swing from right to left. 3 The 2007 election was the exception to the rule, as the incumbent right-wing government was re-elected. Of the 73 seats which changed hands, 61 were overturned by left-wing parties, but 12 were actually gained by right-wing parties.

The Electoral Competition Approach 83

vote counts, and where the right choice of candidate is of particular importance. Negotiating with coalition partners The particular nature of France’s two-ballot electoral system, combined with the large number of parties in her party system, means that French parties are prone to forging electoral alliances. Voters have a potentially wide range of candidates in the first round but the second round tends to be a left-right stand off. If there are too many candidates in a constituency from the same side of the political spectrum, for example the left, there is the risk of splitting the vote and hence of no left-wing party gaining the required percentage of votes to enable qualification to the second round. This prompts parties to form coalitions with parties close to them on the ideological spectrum in order to concentrate the vote in each constituency behind a single candidate and, in so doing, maximise that candidate’s chances of qualification to the second round. For parties of the left, this was exemplified by the ‘gauche plurielle’ umbrella of the late 1990s, whereby parties of the left agreed to co-operate and translate their vote share into seat share. In other words, constituencies with large left-wing electorates were divided amongst the main parties of the left in proportion to the national share of vote that each party received. This tactic allowed each party to benefit from receiving the votes of the entire left-wing electorate within the constituencies that were reserved for them, thus avoiding internal division and strengthening their position against parties of the right.9 The electoral strength of each party was also important in influencing how many government positions each party within the coalition could hope to claim when the left came into power in 1997. These arrangements were crucial in enabling certain smaller parties, such as the Greens, to win seats, but they also increased the emphasis on vote share (Spoon 2007). Given that the initial balance of power within the coalition was influenced by vote share rather than seat share, parties had to be careful to win as many votes as possible in order to maximise their power within the coalition, and hence the number of seats and government positions allocated to them. This meant that votes became significant in all seats and not just target seats, thus making national electoral performance an instrumental factor in determining each party’s power and status. Although France’s right-wing parties have not formed a coalition equivalent to the ‘gauche plurielle’, they have formed other types of electoral alliance, the most recent and obvious example of which is the

84 Parties, Gender Quotas and Candidate Selection in France

formation of the UMP. This umbrella party, created in 2002 to help ensure a right-wing victory in the legislative elections, was formed from a number of smaller right-wing parties. Each party’s prior electoral strength was influential in determining the new balances of power within the UMP. Although the creation of the UMP was an isolated event, it is not uncommon for parties on the French right to merge, split and form loose coalitions (such as the UDF from 1978–1998). For this reason, even the smaller parties with few seats will wish to maximise their vote share in order to have the best possible bargaining position in the event of the formation of an electoral coalition. The fractured, fluid and bi-polarised nature of France’s party system thus leaves many parties fighting both for seats and for power relative to the parties close to them in the ideological spectrum. Electoral performance matters even in seats that cannot be won. By increasing their vote share in elections, and hence their electoral muscle, parties increase their influence and power. They also strengthen their bargaining position when discussing programmatic demands, thus helping them to realise their more ideological agendas. Winning as many votes as possible is therefore an essential element of party strategy. The spoils of victory In France, political parties are financed by the state, and the amount of money that each party receives is in direct proportion to their electoral performance. Parties receive two distinct portions of state finance; the first is related to how many votes they receive, and the second, to how many seats they win. All parties, large and small, will have a direct financial incentive to win as many votes as possible in every seat in order to assure the party’s financial health. However, the second portion of state finance is the more lucrative, and thus places a particular premium on winning seats. Each vote is worth 1.67 euros per year, while a seat is worth 44,492 euros per year.10 Parties will therefore wish to do everything they can to maximise their electoral performance, particularly in constituencies where they stand a chance of winning, because their future resources depend on their current success. Rather than being at the mercy of their membership, parties are dependent on the electorate for the bulk of their funding. Placing an emphasis on electoral performance thus allows parties to maximise the resources available to them. The more votes they receive, the more power, political and financial capital they have at their disposal. Every vote and every seat counts towards helping parties achieve

The Electoral Competition Approach 85

their goals. As a result, it is worth parties’ while to ensure that they select the candidate that is the most electable. How they seek to achieve this goal is the subject of the next section.

How electoral competition affects candidate selection As we have seen above, parties have good reason to seek to win as many seats and as many votes as possible, and hence to seek the candidate most likely to perform well at the polls. If there is such a thing as a personal vote – namely, a proportion of the vote that is influenced directly by the candidate rather than the party – then this will be of particular value to French parties. Although research on candidate effects is limited and inconclusive, there is enough evidence to suggest that a personal vote is possible, and may be of some value, especially in marginal constituencies (Norris et al 1992; Norris & Lovenduski 1995). If parties are fuelled by the desire to win elections, then they will seek to prioritise candidates who are likely to add to the party vote, and avoid candidates who might be perceived as having a negative effect. From a rational choice perspective, if parties are rational actors seeking to maximise their electoral performance, then it follows that it is rational for parties to select the candidate most likely to perform well at the polls. This assumes that parties will have a perfect knowledge of their candidates and of their electorate, and will be able to forecast with accuracy which candidate would be the most valuable. In practice, this is not possible, and parties operate under a version of bounded rationality, in that their choices are constrained by the limits of the information available. Parties are therefore likely to be influenced by generalisations and types of candidate to assist them in making their choices. They will look for candidates with demonstrable experience, but also for candidates who fit a recognisable and reassuring profile. Incumbents The single biggest generalisation that all parties have made is that incumbents are preferable to non-incumbents. This preference for incumbents has numerous foundations that extend beyond a simple electoral imperative. For example, incumbents may be popular within the local party, and will benefit from the networks accrued during their term(s) in office. They will also have benefited from the privileges of patronage at their disposal to create a pool of influential people who are dependent on or indebted to them (Pionchon & Derville 2004,

86 Parties, Gender Quotas and Candidate Selection in France

p. 76). They will have acquired a reputation in the constituency, especially if they have served more than one term. Their high profile combined with their previous electoral success will strengthen their position and credibility as a candidate. As discussed earlier in this chapter, their familiarity to the electorate increases the probability of receiving a personal vote. Furthermore, parties will be unwilling to offend a loyal deputy by showing them the door. We have also seen in Chapter 3 that reselecting incumbents is a useful way for parties to defend a constituency for themselves against claims from other coalition partners. Finally, if a party deselects an incumbent against their will, there is the risk that the incumbent will stand anyway and defeat the party nominee. So there is the double issue of seeing the incumbent as the stronger candidate, and wishing to avoid the potentially harmful consequences of deselection. For all these reasons, parties have been unapologetic for their practice of favouring incumbents at the expense of introducing new, women candidates to meet the requirement of parity. The PS explained in 2001 that it was trying to increase the numbers of women in constituencies without incumbents, and to replace incumbents who stood down with women (although the latter was not properly observed in practice). However, the right of incumbents to stand again if they so chose was protected (Fabre and Garin, Le Monde 13 November 2001), and their subsequent reselection was described as a ‘formality’.11 Meanwhile, the UDF suggested that a proportional representation system would be required to get round the problem of incumbency (Zimmerman 2003b), and their elections officer cited incumbency as the number one priority when selecting candidates.12 Michel Laurent (PCF) claimed that in the single-member district legislative elections, the automatic reproduction of incumbents was inevitable.13 In all cases, the question of incumbency was presented as an insurmountable problem. There was an underlying assumption that it was always better to keep an incumbent than to replace him or her (usually him) with a new candidate. Former incumbents In addition to incumbents, there is a phenomenon specific to France of former incumbents who routinely seek to regain their seats. In some political systems, a politician who had lost their seat might be deemed a poor choice of candidate. Not so France. Given the French tendency, prior to 2007, of alternance (switching between governments of the right and left at each election), it is par for the course for certain polit-

The Electoral Competition Approach 87

icians to have a term in office, be voted out and then seek to regain their seat when the political tide shifts back in their favour. Indeed, many former deputies are reselected automatically as their party’s candidate in their constituency. Although the Women’s Secretary of the PS hoped to reverse this trend by asking that these swing seats be reserved for women, she admitted that it was the least likely of her demands to be met, given the strong tradition of according preference to the former incumbent.14 So strong is this tradition that women’s movements in other parties have not even dared ask for such a measure. Local politicians A third preference amongst parties, especially for candidates who have not been elected to Parliament before, is for candidates who already possess a local electoral mandate, as confirmation that they are electable and capable of holding office. High profile positions such as mayor are particularly valued. For example, all but 23 (3.9 per cent) of the deputies elected in 2007 had already won an election at local level prior to their election to the National Assembly. More than half of deputies are, or have been, mayors. If a candidate is not an incumbent, parties need an alternative method of assessing the candidate’s suitability for national office, and their performance in local office is a convenient starting point. Perpetuating the status quo Fourthly, in constituencies where the incumbent is retiring, parties often seek to recruit a new candidate that will in some way resemble the outgoing deputy (Norris & Lovenduski 1995, p. 127). Parties sometimes have an attitude of ‘if it isn’t broken, don’t fix it’ – if the previous candidate in that area has been popular with the electorate, then a candidate with a similar profile might prove likewise popular. This practice is often encouraged by the outgoing incumbent, who will have nurtured a candidate as the preferred successor. More often than not, this ‘successor’ will be male, and will benefit from the endorsement of the outgoing candidate. This was the case in 2002 with the PS, for example, where several male candidates who were retiring ensured that their preferred (male) successor took their place, despite the party policy to give vacant seats to women (Noblecourt, Le Monde 16 December 2001). Favouring men Parties may not overtly seek to select male candidates, but the above factors work in tandem to the advantage of men. The vast majority of

88 Parties, Gender Quotas and Candidate Selection in France

both incumbents and former incumbents are male. As we have seen in the previous chapter, there are far more men than women holding the most valuable local mandates. Men are more likely than women to be groomed as the favoured successor to an outgoing incumbent, and are more likely to be seen as having the required profile due to their resemblance to the predominant, masculine model of a politician. Indeed, ‘masculine political behaviour is seen as the norm and women are seen as the deviation’ (Halimi 1994, p. 118). Pionchon and Derville (2004, p. 77) also argue that parties believe men to be better suited to politics than women, and see female candidates as more ‘risky’. This concept of women as risky, less qualified or less well suited to politics has led parties to perceive replacing (usually male) incumbents with female candidates as too costly in electoral terms. Even where there is no incumbent, other male candidates may be perceived as more likely to win than a woman, by virtue of their greater political experience, their better networks or their more familiar profile. A rational party seeking to maximise electoral performance might decide that the political cost of respecting parity is greater than the financial cost of fielding more men than women. If it is believed that a man is more likely to win than a woman, the financial cost of not winning the seat might also exceed the financial penalty for not implementing parity, due to the money tied to winning seats. The UMP summed up this position when a spokesperson for the party, Pierre Bédier, told L’Express that ‘a man who wins is less costly than a woman who loses’ (Bitton, L’Express 16 August 2001). This sentiment was echoed by M. de Saint Quentin, party financial director and a member of its selection committee, who claimed that ‘it is more financially valuable to have male candidates who win than female candidates who lose’ (Fabre, Le Monde 10 May 2002). The message of these statements is clear: a party is better off disregarding parity and fielding a man rather than a woman, despite the fine that this will incur, because the man is more likely to win the election. This view still had currency when parties were selecting candidates for the 2007 elections; for example, in 2006, a party official from the UMP headquarters claimed that ‘we would probably have lost the elections [in 2002] if we had fielded unknown women in half the constituencies. We still prefer to pay fines than lose elections!’ (www.lefigaro.fr, 06 January 2006). Although other parties have been less explicit in explaining why they do not select more women, their reluctance to do so almost speaks for itself. For example, the justification used by both main parties in prioritising men in 2002 was the fact that they needed to play safe after the sur-

The Electoral Competition Approach 89

prise progression of the far-right FN candidate Jean-Marie Le Pen into the second round of the presidential elections two months earlier (Seelow, Le Monde 13 July 2002). ‘Playing safe’ can be interpreted as prioritising incumbents, who are already familiar to the electorate, and/or as avoiding the ‘risk’ of fielding women. A clear pattern thus emerges of favouring incumbents over nonincumbents, and favouring men over women. Whether or not these choices are rational, and hence whether or not they support the electoral competition approach, is evaluated in the remainder of this chapter.

Evaluating the performance of types of candidate We have seen that parties prioritise certain types of candidate over others, and notably male incumbents. However, do parties have any evidence to support their belief that men in general, and incumbents in particular, are a better election prospect than women? At a first glance, the answer would be yes. Women are less likely than men to win their elections. This is visibly manifested in the fact that the proportion of women candidates has always exceeded the proportion of women elected. However, this simplistic evidence has been contested by women’s groups, and most notably by the Observatoire de la Parité. In a study of the June 2002 legislative elections, they observed that women were fielded in constituencies that were already known to be difficult on the basis of their party’s performance in the April 2002 presidential elections (Zimmerman 2003a). Meanwhile, studies conducted elsewhere imply that women, for the most part, do not receive an electoral penalty, while incumbents receive only a small electoral boost that is unlikely to make a difference except in the most marginal of constituencies (Norris et al, 1992). However, without a systematic study of electoral performance of candidates in France over more than one election, it is very difficult to prove (or, indeed, disprove) conclusively whether parties are right to be wary of women candidates, or whether women’s poor electoral performance is due solely to the seats in which they are placed. The real test of voter preferences lies not in their stated preferences but in whether, all other things being equal, they discriminate against women candidates when voting. To answer this question, I conducted a longitudinal study of candidate performance in France. If the results demonstrate that incumbents do indeed enjoy an electoral bonus, and/or that women receive an electoral penalty, then the behaviour of parties may be deemed to be

90 Parties, Gender Quotas and Candidate Selection in France

electorally motivated. If not, then alternative explanations need to be considered. Separating seat effects from candidate effects The key difficulty in measuring candidate success is isolating the impact of a candidate from other factors affecting the election result. One obvious influence on a seat’s electoral outcome is the political environment on a wider scale. A way of controlling for this is to measure overall swing across the country and compare the swing within a seat against the national average. However, this is of limited benefit because, as in the US or UK, France has strong regional differences of political affiliation. So it is far more accurate to compare the swing within a seat to the regional, rather than national, average. Even this does not tell the full story. Each seat is also affected by its own history. The safety of the seat is a very important factor, and can vary considerably even within seats in the same region. Thus it is necessary not only to make comparisons across space, but also across time. By comparing a seat against its own history, after controlling for swing, it is easier to identify the effects that changes in candidates may have from one election to the next. This study will look at election results spanning the last five general elections in France – those of 2007, 2002, 1997, 1993 and 1988.15 Changes in party fortunes within each seat over time will be crossreferenced with changes in the candidates presented in that seat to see whether candidate effects have any impact on party performance. The significance of candidate sex and incumbency status will first be monitored by seeing whether a change in these affects the election result. The performance of incumbents will then be compared to other types of candidate. Finally, the safety of seats in which women are fielded will be closely examined to see if this offers an alternative explanation for women’s lower rate of electoral success. Measuring sex and incumbency effects In order to separate seat effects and candidate effects, a measure was created to compare a candidate’s performance against the result their party16 could expect to achieve, based on the party’s previous history within that seat and its performance in the rest of that region in the same election.17 Any difference in party performance from one election to the next, after controlling for swing and for the party’s prior performance in that seat, might arguably be attributed to candidate effects. The two candidate effects considered to be of most theoretical interest

The Electoral Competition Approach 91

are sex (to see whether men enjoy an electoral advantage over women) and incumbency (to see whether incumbents benefit from standing against newcomers). The first test to separate seat effects from candidate effects is a regression model, reported in Table 4.2. The dependent variable models change in party performance from the previous election (t–1) to the election in question (t0) after controlling for swing and seat effects.18 A decrease in the dependent variable indicates a better performance than expected for the left; an increase indicates a strong performance for the right. The model uses two sets of independent variables to control for changes in the sex and incumbency status of candidates for both parties from t–1 to t0. Changes in candidate sex are measured through a series of dummy variables which control the different possible combinations of sex and party over both elections. For the sake of brevity, left-wing candidates are denoted in the model by the letter L, and right-wing candidates are denoted by an R. A change from a man candidate at t–1 to a woman candidate at t0 is denoted as m–w, and vice versa as w–m. For example, if the right-wing candidate was a man in both elections while the left-wing candidate was a man at t–1 and a woman at t0, this would be labelled in the model as ‘R same; L m–w’. As these dummy variables were mutually exclusive, and as the n for the dummy indicating no change in candidate sex between the two elections was larger than the combined n for the other candidate sex variables, each regression was run twice, once using the variable indicating whether or not there was any change in the sex of either candidate, and once using the variables to illustrate the different kinds of change that did take place. Not every combination occurred in every election, and combinations were excluded if the n was lower than five. Incumbency is also controlled for, using a variable to measure different kinds of incumbency status.19 This variable was used for the candidates of both parties, and was used for both time periods in order to control for any change. The most striking finding is that none of the variables measuring change in candidate sex is significant in any of the models. Once safety of seat and incumbency are controlled for, sex carries no additional impact. Incumbency does appear to have some significance, although its impact varies from one election to another. As the majority of incumbents and former incumbents are men, while the majority of women are new candidates, it could be argued that there may be a sex effect that is masked by controlling for incumbency. However, a closer examination of the findings makes such a conclusion implausible. In the 1997 election, there was a particularly high proportion of

92

Table 4.2

Measuring candidate impact on party performance, 1997, 2002 and 200720

Independent variables

1997

Candidate sex

No change R same; L m–w R same; L w–m R m–w; L same R & L m–w

Incumbency

a

R w–m; L same R status t0 L status t0 R status t–1 L status t–1 Adjusted R2

2002 b

β

Sig

0.211

.834

–0.456 .134 .844 .032** –0.116 .742 –0.323 .276 0.028

2007

a

β

Sig

0.005 –2.121 0.294

.997 .372 0.898

b

β

Sig

0.721

0.225

–0.425 .167 –0.063 .802 0.856 .032** .029 .902 –0.107 .763 –0.384 .077* –0.322 .280 0.117 0.640 0.023 0.005

a

β

Sig

–0.813 –0.515 –0.806 2.315

.472 .741 .300 .329

–0.849 .510 –0.007 .979 0.109 .677 –0.394 .074* 0.131 .606 –0.003

b

β

Sig

–0.714

.292

β

0.100 .936 0.356 1.900

Sig

.918 .513 .753 .541

2.791 .107 1.182 .000*** 1.227 .000*** -0.700 .022** –0.748 .020** –0.125 .656 –0.072 .801 –0.275 .300 –0.247 .368 0.127 0.122

Source: election results provided by www.assemblee-nationale.fr and www.lemonde.fr *significant at 10% (p<0.1); ** significant at 5% (p<0.05); *** significant at 1% (p<0.01)

10.1057/9780230275294 - Parties, Gender Quotas and Candidate Selection in France, Rainbow Murray

The Electoral Competition Approach 93

right-wing incumbents following the landslide victory of the right in 1993. Far from enjoying an advantage, these incumbents lost their seats in large numbers as the electoral tide shifted back to the left. Meanwhile, although incumbency status does appear to have been significant for left-wing candidates in 1997, a higher level of experience is actually associated with a better than expected performance for the right. This may be explained by the fact that the only seats to be held by the left in 1993 were those where their baseline support was strongest. For this reason, swing to the left in these seats was likely to be lower than the swing experienced in the seats which returned to the left in 1997. The PS came close to fulfilling its target of 30 per cent women candidates in 1997, resulting in a significant rise in the number of new women left-wing candidates. If an increase in the ranking of prior experience results in a better score for the right, it follows that the new candidates – many of whom were women – were those who performed most strongly for the left in 1997. In the 2002 election, the incumbency status of candidates is completely insignificant. The control variable for incumbency status in the previous election (1997) indicates weak significance for right-wing candidates, where incumbents faced a larger swing to the left. This is consistent with the findings for 1997, where many right-wing incumbents lost their seats. Finally, in 2007, incumbency becomes highly significant for both parties. This is intuitive, given that in 2007, unlike all other elections included in this study, the incumbent government was re-elected, with a relatively low turnover of seats. The very low R2 for all the regressions, and especially those prior to 2007, suggest that sex and incumbency status are both poor predictors of electoral outcome. However, given the varying significance of incumbency status, it is worth looking at the power of incumbency in greater detail. Incumbents versus ‘inheritors’ In addition to incumbents, there are several other categories of candidate who stand in French elections. These include ‘former incumbents’, as discussed above. There are also ‘inheritors’, defined by Norris and Lovenduski as ‘candidates selected for an open seat previously held by their own party, where the previous MP retired’ (1995, p. 24). Moreover, some seats will be contested by the same candidate at each election, even if the candidate is not successful – indeed, 11 seats have had the same runner-up candidate for at least four elections in a row. The combined effect of incumbents, former incumbents seeking to

94 Parties, Gender Quotas and Candidate Selection in France

regain their seats, inheritors (many of whom are the ‘favoured sons’ of the outgoing deputy) and ‘serial losers’ means that even in difficult seats, opportunities for would-be candidates are hard to come by. Although there is clear evidence of parties’ willingness to recycle their election candidates, there is far less evidence to suggest that the electorate like to see the same names on the ballot paper year after year. As Table 4.3 demonstrates, inheritors (who, in this chapter, are further defined as candidates with no previous electoral history in that seat21) may perform as well as, or even better than, present or former incumbents. Over the course of the past four elections, regardless of party swing, the overall success rate of inheritors has been steadily increasing, although the opportunities accorded to them have diminished overall due to the increasing numbers of incumbents standing for re-election. This is despite the fact that incumbency, in itself, is no guarantee of electoral success, as the figures for 1997 more than amply demonstrate – barely one in two incumbents managed to hold on to their seat. In that year, incumbents fared particularly badly while an unusually high proportion of former incumbents were successful in regaining their seats. This is in accordance with the significant test results for incumbency in 1997 found in Table 4.2, and is related to the sharp electoral swing to the right in 1993. Many right-wing candidates exceeded their expectations and won seats in former left-wing strongholds, giving the right a staggering majority of 401 seats. When many of these newly-won seats were overturned in 1997, many uprooted left-wing politicians returned to their former seats. Therefore, while the relationship between incumbency and electoral performance was statistically Table 4.3

Success rates of incumbents and inheritors Incumbents

1993 1997 2002 2007 1

Former Incumbents

Inheritors

Total standing1

Success rate2

Total standing1

Success rate2

Total Success standing1 rate2

75% 80.9% 84.2% 84.1%

65.4% 57.6% 71.4% 87.4%

n/a3 14.9% 13.5% 13.2%

n/a3 82.6% 59.0% 30.3%

27.6% 17.5% 13.5% 15.3%

47.2% 62.3% 64.8% 81.8%

The percentage of seats in which a candidate of this group was standing.22 The proportion of candidates within this group who were elected. 3 Reliable data was not available for elections prior to 1988, therefore it is not possible to determine whether candidates in 1993 were seeking to regain a seat won before 1988. 2

The Electoral Competition Approach 95

significant in 1997, it was clearly a party effect caused by the rapid surge and decline in support for the right in the early to mid-1990s, rather than a candidate effect. Similarly, the success of incumbents in 2007 is unlikely to be an indicator of a sudden enthusiasm from the electorate for incumbent candidates. A much more plausible explanation for the significance of incumbency in 2007 lies in the new electoral calendar introduced prior to the 2002 elections, resulting in a harmonisation of presidential and legislative electoral terms, with the presidential elections occurring shortly before the legislative elections. Prior to this harmonisation, it was possible to have a president of one party and a government of the other party, resulting in a scenario of ‘cohabitation’ which was unpopular with voters due to its associations with political stalemate and lack of accountability. The harmonisation was designed to increase the likelihood of giving the president a parliamentary majority.23 For this reason, the best predictor of the 2002 and 2007 legislative election results was the presidential elections that preceded them. Had Jospin qualified to the second round of the 2002 elections instead of Le Pen, he might well have secured a victory for himself and then his party. In a similar vein, the success of Sarkozy in the 2007 election secured a follow-on victory for the UMP despite the unpopularity of the outgoing government. If anything, then, these results only serve to underline that it is party rather than incumbency effects that dictate a candidate’s electoral fortune. One final point of note in the discussion of incumbency is the fact that women, in addition to forming a very small proportion of present and former incumbents, consistently performed less well than their Table 4.4

Female incumbents and inheritors Incumbents

1993 1997 2002 2007

Former Incumbents

Inheritors

Percent women1

Success rate (%)2

Percent women1

Success rate (%)2

Percent women1

Success rate (%)2

5.3 6.2 11.3 14.0

65.2 48.3 58.2 88.2

n/a3 4.7 (4) 5.1 (4) 17.1 (13)

n/a3 75 (82.9) 50 (59.4) 38.5 (28.6)

8.2 13.9 27.6 28.4

46.2 50 51.7 72

(23) (29) (54) (68)

(65.4) (58.2) (73) (87.3)

(13) (14) (29) (25)

(47.2) (64.4) (69.7) (75)

All figures are percentages. 1 The percentage of total candidates in this category who were female. The figures in brackets are the actual n that this percentage represents. 2 The success rate of female candidates. The figures in brackets represent the success rate for male candidates. 3 Reliable data was not available for elections prior to 1988, therefore it is not possible to determine whether candidates in 1993 were seeking to regain a seat lost in 1988.

96 Parties, Gender Quotas and Candidate Selection in France

male counterparts prior to 2007, as Table 4.4 demonstrates. In fact, although the number of women standing in each category has risen with each election, their success rate in proportion to men steadily decreased. In 1993 there was only a slight difference between the performance of male and female incumbents and inheritors. By 2002, the gender gap in electoral performance had become marked in all categories. By contrast, in 2007 women actually outperformed men both as incumbents and former incumbents. There are several, linked, explanations for this trend. Given that the results above imply that the electorate do not penalise female candidates, this is an unlikely explanation for their poor results, and this also does not explain their improved performance in 2007. More probable is that women tend to be placed in difficult seats. As incumbents, they are frequently defending difficult seats that were only won at the last election due to the extent of the political swing in their party’s favour, and hence that are held with slim majorities that are likely to be overturned at the next election if the electorate swing back again.24 Moreover, although the number of women fielded as inheritors has increased quite dramatically – due in part to the policy of the Socialist Party of replacing retiring incumbents with women in 1997, and in part to the demands of parity in 2002 – it would seem that parties have been rather less generous to women than to men in handing out safe seats. If women are more likely than men to hold seats won with small margins, then a relatively low swing would be enough to cost them the seat. This might explain why women incumbents fared better in 2007, when the low swing meant that even those in volatile constituencies were able to hold onto their seats. This also explains why the success rate for former incumbents of both sexes was much lower than usual in 2007, although it is less clear why women were more successful than men within this category. Although there is some existing research to support the hypothesis that it is the safety of seat rather than the candidate’s sex that causes women to perform badly (Murray 2004; Southwell & Smith 2007; Zimmerman 2003a), there has not yet been a systematic study definitely proving whether or not women are disproportionately placed in difficult seats. The final test of this chapter aims to address this problem. Candidate sex and safety of seat In order to make a fair appraisal of safety of seat, I used the measure of the swing required to overturn the seat at the next election. Using the

The Electoral Competition Approach 97

typical swing as a barometer (with the average swing per election falling between 4.9 per cent and 8.5 per cent25), I then grouped seats into five categories of safety according to the swing required to overturn the seat (in brackets): very safe (> 15 per cent), fairly safe (8–15 per cent), ‘could swing’ (5–8 per cent), vulnerable (2–5 per cent) and marginal (< 2 per cent). These categories were then compared against the sex and party of candidates for the 1993, 1997, 2002 and 2007 elections. The results are reported in Table 4.5. Table 4.5 is very interesting for a number of reasons. For each election, the gender distribution is significant for half of the seats, but not for the other half. It is not a coincidence that this significance oscillates to the same rhythm as electoral swing. The incumbent party (‘defender’) will try to defend as many of its seats as possible while the party in opposition (‘challenger’) can rely on at least some swing back in its favour at the next election and will try to regain as many seats as it can. Thus, in 1993 and 2002, the Left were defending and the Right were challenging; in 1997 and 2007 the reverse was true. For the party that is defending, the most desirable seats are the safest ones as these are the ones least likely to be overturned by a swing to the opposition. Conversely, for the party that is attacking, the best seats are the ones that are marginal, as these are the seats that are most easily won. Regardless of the party in question, whichever side is defending will put its female candidates on the front line, in the most vulnerable seats, whilst conserving a disproportionate number of the safest seats for its male candidates. For example, in 1993, men representing the Left had a monopoly over the safest seats while a much higher proportion of women than men were fielded in marginal constituencies. When the large differences in the number of male and female candidates are taken into consideration, the paucity of women in the safest seats becomes even more stark; for example, although 4 per cent of female right-wing candidates defended safe seats in 1997 compared to 2 per cent of men, this actually equates to six men and only one woman in the safest seats, with 78 out of the 84 ‘fairly safe’ right-wing seats that year also being defended by men. Meanwhile, the side which is attacking will do the reverse – the male candidates are disproportionately placed in the most winnable seats, while the majority of women are relegated to seats in which they stand little or no chance of election. For example, in 2002, women on the Right were more than four times more likely to contest a seat that was considered safe or fairly safe for the Left than they were to contest a marginal seat with a good chance of victory, whereas the majority of

98

Table 4.5

Sex and safety of seat 1993

1997

L

R

L

M

F

R

M

F

L

M

F

M

R F

M

L F

M

R F

M

F

DEFENDER

F2

2007

Very safe Fairly safe Could swing Vulnerable Marginal

8 52 35 63 37

(4) (25) (16) (30) (17)

0 4 6 2 8

(0) 1 (1) (2) 7 (7) (3) 10 (10) (1) 29 (30) (4) 40 (42)

0 1 3 3 2

(0) (1) (3) (3) (2)

0 0 7 19 21

(0) (0) (13) (36) (40)

0 0 1 3 2

(0) (0) (2) (6) (4)

6 78 51 71 35

(2) (29) (19) (27) (13)

1 6 4 6 7

(0) (2) (2) (2) (3)

5 62 39 68 43

(2) (23) (14) (25) (16)

0 13 11 11 20

(0) (5) (4) (4) (7)

4 23 37 64 50

(2) (12) (19) (33) (26)

1 2 2 7 5

(1) (1) (1) (4) (3)

5 26 29 33 36

(3) (15) (17) (19) (21)

3 5 9 12 12

(2) (3) (5) (7) (7)

24 92 60 59 53

(7) 4 (1) (28) 10 (3) (18) 8 (2) (18) 10 (3) (16) 9 (3)

CHALLENGER

M1

2002

Very safe Fairly safe Could swing Vulnerable Marginal

11 28 36 8 52

(11) (29) (38) (4) (25)

2 4 6 0 4

(2) 31 (15) (4) 63 (30) (6) 42 (20) (0) 1 (1) (2) 7 (7)

4 1 3 0 1

(2) (0) (1) (0) (1)

44 67 39 0 0

(17) (25) (15) (0) (0)

11 10 3 0 0

(4) (4) (1) (0) (0)

6 21 20 6 78

(11) (40) (38) (2) (29)

2 1 3 1 6

(4) (2) (6) (0) (2)

19 36 36 5 62

(10) (18) (18) (2) (23)

20 35 19 0 13

(10) (18) (10) (0) (5)

38 61 57 4 23

(14) (22) (21) (2) (12)

12 18 6 1 2

(4) (7) (2) (1) (1)

14 40 41 38 42

(4) (12) (12) (12) (13)

14 62 27 31 20

(4) (19) (8) (9) (6)

2 12 18 24 23

(1) (7) (11) (14) (14)

6 19 20 21 25

(4) (11) (12) (12) (15)

1 Male candidate 2 Female candidate Figures represent the actual n. Figures in brackets are percentages within the columns for defenders and for challengers, rounded to the nearest whole number. Percentages may not add up to exactly 100 per cent due to rounding. Pearson chi-square results for this table are as follows: 1993 Areas shaded in grey: p=0.042 Unshaded areas: No significance 1997 Areas shaded in grey: p=0.035 Unshaded areas: No significance 2002 Areas shaded in grey: p=0.042 Unshaded areas: No significance 2007 Areas shaded in grey: p=0.031 Unshaded areas: No significance

10.1057/9780230275294 - Parties, Gender Quotas and Candidate Selection in France, Rainbow Murray

The Electoral Competition Approach 99

men stood in the constituencies that were easiest to win back from the Left. Conversely, parties are less markedly divided on gender in the seats which are not likely to swing, and the results for these seats (which are unshaded in the table) are not significant. It would seem that parties are more prepared to place women in target seats which are theoretically winnable (that is, only requiring a fairly small swing in their party’s favour) if they know that a swing against their party is far more probable. The significance of these findings is confirmed by statistical tests, and also by the fact that the trend was as strong in 2007 as in 1993, despite the increasing numbers of women candidates. It is also of note that the trend applies equally to parties of the left and of the right. Therefore, although parties are starting to make token gestures towards increasing women’s representation, these results prove that women are consistently placed in the most difficult seats. Either parties are doing this in the (misguided) belief that they are minimising their electoral losses by placing women candidates in the seats that matter the least, or parties are not acting purely out of electoral motivations. Either way, the trends observed in Table 4.5 cannot be explained by the logic of electoral competition alone. Overview of findings This study provides considerable evidence to suggest that it is parties, and not the electorate, who discriminate against female candidates and prevent them from being elected in equal proportion to their candidate numbers. After controlling for both geographical and temporal electoral trends, there was no evidence that changes in candidate sex had any impact on electoral performance. Moreover, there was only scant evidence to suggest that incumbency had an impact on election results. The only elections which appeared to be affected by incumbency were 1997, in which incumbency actually appeared to have a negative effect, and 2007. In both cases this significance was in fact a party effect, linked to the large swing towards and then against the Right in the 1990s, and to the victory of Sarkozy in the 2007 presidential elections. The beneficial effects of incumbency were brought further into question when it was revealed that inheritors are performing increasingly well at the polls – indeed, in 1997, they out-performed incumbents. It is not yet clear whether this was an exceptional result, or whether there might be a trend for incumbents to perform better under

100 Parties, Gender Quotas and Candidate Selection in France

right-wing victories, and for former incumbents and inheritors to perform better under left-wing victories. If 1997 was an exceptional result then incumbents do appear to enjoy a certain, slight advantage over newcomers, in which case parties are rational in according incumbents their privileged status. If the trend continues for incumbency effects to vary according to which party wins the election, then this reinforces the suggestion that the effect is primarily one of party rather than incumbency. The question regarding the electoral weight of incumbency therefore remains partly unresolved. The question regarding women’s electoral performance, on the other hand, is far more conclusive. In addition to refuting the suggestion that women perform less well at the polls, this study also provides strong evidence to suggest that women are, in fact, placed in the most difficult seats, and further, that this is an integral part of party electoral strategy. This leads to the conclusion that parties may not be objective in seeking the candidates that are the most electable, and/or that they may be selecting candidates for motives other than anticipated performance at the polls. Either way, fears of the electoral cost of political renewal and feminisation appear to be unfounded. Conclusion The electoral competition approach assumes that party behaviour is dictated by the goal of winning as many votes as possible. For this approach to provide an accurate account of candidate selection, parties would need to prioritise selecting the candidate that they believed to be the most electable. This chapter has demonstrated that parties might have considerable motivations for choosing this approach to candidate selection, given the significance of each and every vote received to the achievement of party goals. Votes help increase parties’ wealth and power, which in turn help them achieve parties’ programmatic demands. France also has a large number of constituencies that are won or lost on a relatively small margin, thus further increasing the importance of selecting the right candidate. The approach was tested by firstly considering what type of candidate parties tend to select, and then determining whether parties are making their choices based on an accurate and deliberate strategy to maximise their electorate. In terms of which candidates parties selected, all parties openly acknowledged their preference for incumbents over non-incumbents, and demonstrated an extended preference for candidates with other types of electoral experience, including

The Electoral Competition Approach 101

former incumbents. A more subtle preference for male candidates was also demonstrated, although only occasionally expressed explicitly. More often it was hidden through a preference for qualities that men were more likely to possess, or expressed through suggestions that selecting women amounted to an electoral ‘risk’. Parties did appear to be acting out of the conviction that they were choosing the best ‘men’ for the job, but there was little empirical evidence to ascertain whether men in general, and incumbents in particular, benefited from any kind of electoral advantage. This study aimed to remedy this gap in knowledge by separating candidate effects from seat effects over five elections to determine what effect, if any, the sex and incumbency status of a candidate had on their electoral performance. This produced two significant findings. Firstly, there was no demonstrable evidence that sex had any impact whatsoever on candidate performance, and inconclusive evidence as to whether incumbency enhanced electoral performance. Secondly, there was considerable evidence to suggest that parties of both the left and right were deliberately and consistently fielding women in the most unwinnable seats, whatever their overall electoral strategy. Taken together, these findings refute the argument that parties are selecting the most electable candidates, and suggest that it is actually parties and not the electorate who are discriminating against women. What remains unclear is whether or not parties are aware of the fact that they may not necessarily be selecting the most winnable candidates. Given the lack of research in this area prior to this study, it is possible that parties genuinely believe that the candidates they select are those most likely to win, and that their strategy of supporting incumbents and marginalising women is necessary for electoral success. If this is the case, then this could support the hypothesis of bounded rationality, namely that parties are making rational decisions within the limits of the information available, even if perfect knowledge of a situation might demonstrate their choices to be suboptimal. However, given the lack of concrete evidence to suggest that their preferred type of candidate does perform better at the polls, this casts a doubt on such a conclusion. If parties do not have an accurate understanding of their electorates, then something else must be informing their choices. This could either be a rational preference for certain types of candidate based on factors other than electoral performance, or it could be an irrational preference for certain types of candidate based on a projection of parties’ attitudes onto the electorate. Let us consider each of these in more detail.

102 Parties, Gender Quotas and Candidate Selection in France

Given the lack of evidence to suggest that men and incumbents bring an electoral bonus to parties, this cannot be considered a rational explanation for party choice. However, this does not automatically suggest that parties are irrational actors. Rather, it suggests that parties may be selecting candidates for reasons other than their electoral performance. For example, as we have seen in this chapter, there are other motivations for selecting certain types of candidate, such as a sense of obligation towards incumbents, or a wish to prevent the risk of deselected incumbents’ standing against the party nominee. The preference for men over women may also be linked to the problems of supply at elite level that were discussed in Chapter 3. If this were the case, then parties may have reasoned and rational motivations for prioritising men over women and incumbents over inheritors. This would suggest that party competition is not, after all, the primary motivation in determining party behaviour, at least within the domain of candidate selection. Alternative explanations are considered in more detail in the following chapters. A further rational possibility would be if parties were electorally motivated, but saw the policy effect of parity to be more electorally significant than the candidate effect. In other words, if parties felt that voters might be attracted to the party by the idea of parity, this might provide a rational electoral motivation for parties to field more women candidates. For this to be true, two conditions would need to be met. Firstly, parties would have to assume that the vote for a party would be dictated by the party’s policies rather than by their candidates, meaning that they would be rewarded rather than penalised for fielding women candidates. Secondly, any positive electoral effects incurred through implementing parity would have to exceed any possible negative effects. It is unlikely that parties would be able to measure these positive and negative effects with any degree of certainty. However, the idea of ‘issue ownership’ suggests that this strategy would be less likely to be adopted by a party of the right (Budge et al 1987). It is more likely to explain attempts by parties of the left to improve subsequent attempts at implementing parity. On the other hand, parties may indeed be seeking to field the most electable candidates available, and simply misinterpreting the electorate by projecting their own preferences onto those of their voters, rather than making a rational evaluation of the situation. This would effectively amount to imputed discrimination, whereby the party suggests that it is the electorate that would not support a female candidate, in order to justify selecting more men than women. Party selectors

The Electoral Competition Approach 103

might also ignore any evidence that countered their ingrained belief that men and/or incumbents were more electable than women and/or newcomers. This theory is supported by Etzioni and by Ward, who explains that ‘[t]he norms that the individual adheres to and the affective orientations they have may prevent feasible options being considered and relevant information being obtained, as well as biasing decision-making away from what is instrumentally rational’ (Ward 2002, p. 80). The problem is that it is not clear which of the above assessments is the correct conclusion. It is without a doubt that parties take into consideration a variety of factors when selecting candidates, but less clear whether or not electoral performance is their overriding priority. It is also difficult to determine whether or not they are genuinely convinced that incumbents and men perform better than women at the polls, or whether this is a convenient excuse used to justify a biased selection process. In order to solve the puzzle, it is necessary to delve deeper into party motivations, commencing with institutional explanations of candidate selection, the subject of Chapter 5.

5 The Institutional Approach

Introduction The second of the three approaches to party behaviour explored in this book is the institutional approach. This approach uses political institutions and the political environment to explain party choices. In the context of compulsory quotas such as parity, this approach can be used to explore how the quota legislation itself, along with other aspects of the political system, can influence and define the choices available to parties. In so doing, it is possible to illustrate the choice set available to parties. Some parties may find themselves with a range of options, in which case the institutional approach will need to be used in conjunction with other approaches to explain how parties prioritise between the available choices. But in certain circumstances, institutional constraints may be so great that parties may find themselves with no choice at all when it comes to quota implementation. In these cases, it can be argued that the external political environment is of greater significance than other predictors of party behaviour such as electoral or ideological motivations, as the choice of which course of action to take is effectively taken out of a party’s hands. This chapter will explore the institutional approach before looking at various institutions to illustrate how each will affect the choice set available to parties when deciding whether or not to implement parity. The institutional approach is predicated on the belief that party motivations and behaviour are fundamentally influenced by the external institutional environment. Party choices are shaped and defined by the political environment in which they are located. Parties cannot be considered as free agents operating on a blank canvas, but rather as players within a game. As such, the motivations of each player can 104

The Institutional Approach 105

only be understood within the context of the rules of the game. In the game of politics, these ‘rules’ are the collective impact of the institutional framework. Every aspect of French political life will help to define the parameters within which parties operate, and to limit the choices that are available to parties. Institutionalists believe that it is the institutional environment, more than anything else, which determines party behaviour. This is used to explain the difference between political systems in different countries. In keeping with the overarching themes of the book, this chapter uses a rational choice version of institutionalism to demonstrate how parties seek to maximise their utility within the changing institutional environment. If this approach is a good predictor of party behaviour, then it should be possible to identify ways in which institutions shape party motivations and control the choices available to parties. It should also be possible to observe changes in party behaviour in response to changes in the institutional environment. If parties are attempting to play the game of politics to their best advantage, then they should be sensitive to any changes in the rules of the game and adapt their game-play accordingly. For example, the same parties might adopt different strategies under different electoral systems. Similarly, parties might change their strategies in response to the introduction of the parity law, in a way that was moulded by the nature of the law and its contribution to the wider institutional setting. Institutions may also have a variable impact on different parties depending on each party’s resources and goals. In some instances, institutions might be able to explain variation in parity’s implementation from one party to the next; in other instances, they might be able to provide a partial explanation that is rendered more complete when considered in conjunction with other factors, which is the theme of Chapter 7. The aim of this chapter is to explore the institutional approach in greater detail, and then to test the approach by seeing how useful it is for explaining party behaviour in France. Although there is debate as to the boundaries of what can be considered an ‘institution’, with new institutionalists using a somewhat broader definition of institutions than their predecessors, this chapter shall focus on political institutions that have a direct bearing on candidate selection, such as the electoral system, the party system, the relative importance of the office at stake, and the resources at the disposal of the various parties. The exact detail of the parity law will also be considered to see how this interacts with France’s other political institutions to influence party

106 Parties, Gender Quotas and Candidate Selection in France

behaviour. The utility of the institutional approach as an explanatory tool will then be assessed to see whether party behaviour is, indeed, the inevitable consequence of external factors rather than being due to the needs and particularities of individual parties.

Examining the ‘institutional’ approach The central argument of institutionalism is that institutions matter; that their effects and influence on political behaviour must be taken into consideration. As a theory of political science, institutionalism had its first wave in the early twentieth century, and began to lose popularity in the behavioural revolution of the 1960s (Peters 1999, p. 11). However, March and Olson (1984) led the attempts to ‘bring the state back in’ using an approach known as ‘new institutionalism’. New institutionalist theory in political science is not a homogeneous school of thought, however. Hall and Taylor (1996) identify three stands of ‘new institutionalism’, namely historical, sociological and rational choice institutionalism, each of which has a different definition of what constitutes an institution, and a different conception of how institutions interact with and influence political life. Historical institutionalism argues that an institution’s past will define its current procedures and effects, with traditions, norms, values and precedent resulting in path-dependence and limited ability to deviate from what went before. Sociological institutionalism incorporates not just the more formal rules of the game but also social elements, symbols and morals, with institutions being seen as interacting with and guiding human actions (Krook 2003). There is merit to both of these approaches. For example, it could be argued from a historical institutionalist perspective that the prevalence of male incumbents is an obstacle to parity, as it is difficult to unseat these men once they have become established in their positions of power. As the historical approach is often used to explain institutional resistance to change, this may also be of some utility in explaining why parity has been of limited effectiveness. Path dependence might explain why changes to the composition of political elites are so hard to achieve. A limitation with this approach is that it struggles to explain how change does occur, and tends to attribute this to ‘unintended consequences’ of policies (Hall & Taylor 1996; Kenny 2007; Krook 2003). This approach, while intuitively plausible, risks describing more than it explains, with continuity being the default and any deviation from this being an accident. Such an approach would limit the analysis of political parties as rational actors seeking to adapt to a changing institutional environment.

The Institutional Approach 107

Similarly, sociological institutionalism is of potential interest. For example, the subordinate position of women in French society that was explored in Chapter 3 could be accounted for in a sociological institutionalist approach, with cultural attitudes towards women and their role within the family and society being a barrier to their inclusion in public life. This approach might explain why deeply embedded beliefs, values and attitudes towards women might affect the judgement of party selectorates. This also overlaps with party ideology, which is explored in greater detail in the next chapter. As a central approach of this book is to consider parties as rational actors in order to explain whether and how they have incorporated parity into their candidate selection procedures, rational choice institutionalism is the branch of institutionalist theory that is most compatible with the methodological approach deployed throughout the book, so it is on this strand of new institutionalism that I shall concentrate here. If the premise of institutionalism is that institutions affect political life, and the premise of rational choice theory is that individual actors seek to maximise their utility, it follows that a rational choice institutionalist approach will consist of rational actors trying to maximise their utility within a framework that it is dictated by institutions. In effect, institutions create a bounded version of rationality that defines and limits the range of choices available to individual actors. Within these limits, individuals may still be rational and self-interested. Institutions are perceived as important due to their role in establishing the parameters within which rationality may occur. As Peters puts it, …institutions are conceptualised as collections of rules and incentives that establish the conditions for bounded rationality, and therefore establish a ‘political space’ within which many interdependent political actors can function. Thus, in these models, the individual politician is expected to manoeuvre to maximise personal utilities, but his or her options are inherently constrained because they are operating within the rule set of one or more institutions. (1999, p. 44) In essence, therefore, institutions constitute the rules of the game in which political actors are the players. Indeed, North describes institutions as ‘the rules of the game in a society’ (North 1990, p. 3). Actors accept these rules as a necessary part of the game, and part of this acceptance is an acknowledgement that the other players in the game

108 Parties, Gender Quotas and Candidate Selection in France

will be constrained by the same rules. This means that all players will experience the costs and benefits of adherence to the rules, and the behaviour of other players will be easier to predict. Although not all rules will benefit each player equally, the assumption is that the collective set of rules will balance each other out in the course of the game. So how does this theory relate to explanations of party behaviour and candidate selection? The actors to be analysed here are political parties. Although most theories of rational choice institutionalism focus on individuals rather than organisations, this study focuses on the final outcomes of party decision-making processes, therefore parties will be perceived here as individual actors navigating their way through an institutional framework and seeking to maximise their utility within the range of options that this framework makes available to them. This leads to the equally crucial definition of an institution. There is no definitive account of what constitutes a political institution, and some accounts are quite restrictive (for example, considering only specific and embodied entities such as parliaments) while others are so broad and all-encompassing as to be almost meaningless. One useful and dynamic account of what is an institution is offered by Shepsle: ‘An institution is a script that names the actors, their respective behavioural repertoires (or strategies), the sequence in which the actors choose from them, the information they possess when they make their selections, and the outcome resulting from the combination of actor choices’ (2006, p. 24, original emphasis). Although this definition is somewhat broad and vague, it is also very useful in identifying the analytical benefit of the study of institutions; their intrinsic interest here is not what they are or even what function they serve, but rather, how they affect the rules and the actors of the game under study. In this particular case, the game is candidate selection and the actors are political parties. The institutions to be considered in this study should therefore be those that have a direct bearing on how parties select their candidates, and the remainder of this chapter will explore these in greater detail. The study will include how these institutions limit the options available to parties and inform and shape their preferences, thus affecting the choices that parties go on to make when deciding whether or not to implement parity. Once an analysis has been offered of how institutions affect the actors (parties) and their strategies, consideration will be given in the following chapter to the sequence in which actors make their choice, and how this affects outcomes.

The Institutional Approach 109

France’s institutions and their impact on candidate selection This section builds on the theory outlined in the previous section and demonstrates how rational choice institutionalism can be used to explain the impact of various aspects of France’s political system on the choice of candidate of French parties. In each case, the effects of the institution will be considered in some detail, mapping out how they affect the rule set available to French parties, and how they influence the choices that parties then make. The institutions that meet the necessary criteria of being directly influential on parties’ choice of candidates are the electoral system, the ‘parity’ legislation, the party system, the relative size, wealth and power of French parties, and the importance of the office at stake. Each of these will be considered and analysed in greater detail below. In addition, it should be noted that parties themselves can be considered institutions. This is not considered directly here as we are concerned with the impact of other institutions on parties and party behaviour; but elements of parties as institutions are considered elsewhere, for example the role of party organisation and internal democracy (Chapter 3) and party ideology (Chapter 6). The electoral system Electoral systems have a central role in shaping party behaviour, and are acknowledged by Bartolini and Mair as being one of the key institutional factors affecting party competition and electoral volatility (1990, p. 152). Duverger (1951), Lijphart (1994) and Sartori (1976) are amongst a range of authors who have argued that electoral systems shape party systems and patterns of electoral competition; for example, ‘Duverger’s law’ argues that majoritarian electoral systems lead to twoparty systems (1951). Electoral systems are also particularly influential in structuring party choices when selecting candidates. There has been much research indicating that proportional electoral systems are more favourable to women than majoritarian electoral systems (Dahlerup 2006; Matland 2002; Norris & Lovenduski 1995; Rule 1987; Rule & Zimmerman 1994). This is because different electoral systems offer different choices and opportunities to parties in several ways. Firstly, majoritarian electoral systems place a greater emphasis on the individual candidate, thus increasing the importance of choosing the ‘right’ candidate, whereas in proportional list systems people are more likely to vote on the basis of the party rather than the candidate. As explored in

110 Parties, Gender Quotas and Candidate Selection in France

Chapter 4, this emphasis on choosing the ideal candidate can work to the detriment of women. Secondly, it is easier to introduce new candidates onto party lists without removing the incumbent (although these new candidates might not be placed in winnable positions). This makes it easier for women to get their first foot in the door – a base on which to build. Thirdly, it is more conspicuous on party lists if there is a stark gender imbalance, meaning that parties might be more likely to wish to ‘balance the ticket’ by including female candidates. This is particularly the case where ‘contagion’ takes effect – namely, if one party strives to present a gender-balanced list, this might induce other parties to follow suit (Matland 1998, p. 74; Nelson & Chowdhury 1994). For all these reasons, proportional electoral systems tend to be associated with higher levels of female presence than majoritarian electoral systems. Figure 5.1 illustrates this effect by comparing the average percentage of women elected in 24 developed countries, some with proportional and some with majoritarian electoral systems. The speed and extent to which women’s representation has increased over the Figure 5.1

Effects of electoral systems on women’s representation

30.00

Average % women in Parliament

25.00

20.00 Majoritarian systems Proportional systems 15.00

10.00

5.00

0.00 1945

1950

1960

1970 1980 Year

1990

1998

2009

Sources: Data for 1945–1998 taken from Matland (2002); data for 2009 taken from Interparliamentary Union, 14 August 2009.

The Institutional Approach 111

past 60 years is much more pronounced in those countries with proportional electoral systems. This trend is ostensibly mirrored in France, which uses different electoral systems for different levels of election. In elections conducted under a majoritarian system, including the legislative elections, and smaller districts for senatorial and municipal elections, women have traditionally fared less well than under elections governed by proportional representation. The ‘parity’ law is also applied in different ways depending on the type of electoral system used and the size of the constituency, and the introduction of parity appears only to have increased the gulf between the different types of electoral system. A summary of the different electoral systems used and the appropriate ‘parity’ legislation is provided in Table 5.1. Several effects can be noted from studying this table. The elections with the lowest number of women are all majoritarian (departmental elections, legislative elections and the districts of senatorial elections using a majoritarian electoral system). Conversely, elections using a proportional system come much closer to achieving gender parity. This is true in two senses. Firstly, even before the parity legislation was introduced, women were clearly faring better under proportional systems. Secondly, lists have been far more favourable to the implementation of parity, and hence it can be observed that parity has had a far more dramatic effect in elections using proportional representation than in elections using a majoritarian electoral system. Perhaps the strongest evidence to support this theory is provided by the Senatorial elections, where all factors were equivalent other than the electoral system used, with larger districts (four seats or more)1 being elected under a proportional system, and smaller districts being elected under a majoritarian system. The pre-parity difference between the two groups is negligible in groups A and B, with a somewhat larger difference in Group C. It is after the introduction of the parity legislation that the real difference becomes most evident. Under a proportional electoral system, subject to the requirement of parity, the number of female senators rose from 6.8 per cent to 27 per cent in the first cohort of senators to be elected post-parity, and then from 14.8 per cent to 34.9 per cent in the second cohort. The third group rose from no women whatsoever (1998) to 27.5 per cent in 2008. By stark contrast, in the districts using a majoritarian electoral system, and hence where parity did not apply, all cohorts stagnated, with the very low levels of women in these districts showing no increase in 2001 and 2004, and much smaller increases in 2008 than in the districts using proportional representation.

112

Table 5.1

Electoral systems and parity in France

Election

Electoral system

Size of district1

‘Parity’ legislation

% women pre-parity (year of election)

% women post-parity (year of election)

Current % women (year of election)

Local (council) elections1

Two-ballot ‘winner takes all’ list system

< 3500 inhabitants

None

21% (1995)

30.1% (2001)

32.2% (2008)

Local (council) elections1

Two-ballot list system; 50% to winner, 50% proportional

> 3500 inhabitants

List not recognised unless three in every six candidates are female (one in two from 2007)6

25.7% (1995)

47.5% (2001)

48.5% (2008)

Departmental elections2

Two-ballot majoritarian

n/a

None

Group A4: 6.3% (1995) Group B: 8.3% (1998)

Group A: 9.8% (2001) Group B: 10.9% (2004)

12.3% (2008)

Regional elections

Proportional by list

n/a

List not recognised unless three in every six candidates are female6

27.5% (1998)

47.6% (2004)

47.6% (2004)

European elections

Proportional by list

n/a

List not recognised unless one in every two candidates is female

40.2% (1999)

43.6% (2004)

44.4% (2009)

10.1057/9780230275294 - Parties, Gender Quotas and Candidate Selection in France, Rainbow Murray

Table 5.1

Electoral systems and parity in France – continued

Election

Electoral system

Size of district1

‘Parity’ legislation

% women pre-parity (year of election)

% women post-parity (year of election)

Senatorial elections1

Proportional by list; elected by electoral college

4 seats or more3

List not recognised unless one in every two candidates is female

Group B5: 6.8% (1992) Group C: 14.8% (1995) Group A: 0% (1998)

Group B: 27% (2001) Group C: 34.9% (2004) Group A: 27.5% (2008)

Senatorial elections1

Majoritarian system; 3 seats or elected by electoral fewer3 college

None

Group B: 7.1% (1992) Group C: 4.8% (1995) Group A: 3.3% (1998)

Group B: 7.1% (2001) Group C: 4.4% (2004) Group A: 9.5% (2008)

Legislative elections

Two-ballot majoritarian system

Financial penalty for parties not respecting parity

10.9% (1997)

12.3% (2002)

n/a

Current % women (year of election)

15.8% women elected in total in Group A (2008), taking the total women in the Senate to 21.9%

18.5% (2007)

10.1057/9780230275294 - Parties, Gender Quotas and Candidate Selection in France, Rainbow Murray

113

Sources: Stevens 1996, p. 169; Le Monde; Bird 2003; Bird 2004; Zimmerman 2004, 2005a, 2008a, b, 2009; l’INSEE 2008 1 For some elections, a different electoral system will be deployed depending on the number of seats being elected within a district (senatorial elections) or the size of the population within that district (local elections). 2 Départements are administrative districts broadly equivalent to British counties. There are 96 départements in mainland France. Departmental councils are known as ‘conseils généraux’. 3 The rules concerning the minimum district size for proportional lists have changed twice in recent years; in 2000, the minimal district size was reduced from five to three, although in 2003 it was raised back to four. 4 Conseils généraux are elected in two groups, with each group being elected for six years and the groups being elected alternately, three years apart. 5 The Senate is elected in three groups, with each group being elected for nine years and the groups being elected alternately, three years apart (the 2007 elections were delayed for a year to avoid election overload following the presidential and legislative elections earlier in 2007). 6 The law was changed in 2007 from the three in six rule to zipping.

114 Parties, Gender Quotas and Candidate Selection in France

Given that the same parties behaved in quite different ways depending on the electoral system in place, this invites several conclusions. Firstly, it supports the theory that it is the institution – in this case, the electoral system – that is influencing party behaviour. Secondly, it confirms that parties will adapt their strategies according to the rules set by the electoral system in place. Finally, this also supports the theory that women fare better under proportional electoral systems than majoritarian systems, especially when a quota is in place. This finding was confirmed in interviews with parties, who stressed that it was harder to field women in majoritarian elections. The PS, for example, argued that under majoritarian electoral systems, they ‘select the candidate that is most likely to win within that constituency’, whereas under proportional systems ‘the role of political parties is multiplied’.2 When the electorate are voting for a party list, the choice of candidates is more flexible; when the election is based on a single member district, the choice of candidate is all-important, and Chapters 3 and 4 have demonstrated that this scenario disadvantages women. However, a closer look at the results reveals that the conclusion that women fare better under proportional electoral systems is overly simplistic, as it does not account for several anomalies. For example, in 2001, women fared better in the local election districts where parity did not apply than in the senatorial election districts that were governed by parity. Similarly, they fared far better than women in the legislative elections, which were also governed by parity but using a majoritarian electoral system. Indeed, despite falling under the remit of parity, the legislative elections showed a smaller increase in women in 2002 than the departmental elections, which used the same electoral system but were not covered by the ‘parity’ law. Moreover, the elections that were held using a ‘zipping’ rule – namely, that every other candidate should be female – should have resulted in at least as many, if not more, women being elected than those governed by the more lax rule of three women in every six candidates.3 However, this was not the case. The 2001 local and 2004 regional elections, held under the three-in-six rule, both elected more women than the European and senatorial elections that used zipping. Finally, the electoral system for legislative elections was temporarily changed in the 1986 election to one of proportionality, yet this had virtually no bearing on the number of women elected (5.9 per cent, compared to 5.3 per cent in 1981 and 5.7 per cent in 1988) (Allwood & Wadia 2000, p. 28).4

The Institutional Approach 115

It is clear, therefore, that while electoral systems play an important role in determining party behaviour – as manifested by the different approaches taken by parties depending on the electoral system in use – they are not a sufficient explanation of party candidate choice, and must be considered in conjunction with other political institutions. Institutional theory, after all, considers the collective impact of the different institutions (rules of the game), and each rule can only be understood fully when considered within the wider context of the game. In this particular case, we have already seen that the type of electoral system interacts with the specific parity legislation to produce a double effect. To understand this double effect fully, let us take a closer look at the precise rules of parity. The ‘parity’ legislation The rules governing parity are considered here as an institutional explanation in themselves. Unlike most gender quotas, parity is constitutionalised and is a legal requirement, and as such can be considered an external constraint on party candidate selection. Parity is a major factor in determining the rule-set within which parties operate when selecting their candidates, and its varying forms account for much of the difference between the levels of parity achieved from one party to the next and from one election to the next. As can be noted in Table 5.1, there are three main types of regulation governing parity: zipping, which is used for European elections, some senatorial elections and local elections in 2008; the three-in-six measure, which was used for local elections in 2001 and regional elections; and financial penalties, which are used for legislative elections. The first two measures are reasonably self-explanatory; the list is not accepted unless it meets the one-in-two or three-in-six gender parity requirements. Unsurprisingly, all parties therefore had no choice but to present lists that complied with the parity regulations, and reservations of being unable to find enough female candidates proved to be unfounded (Bird 2004). However, as previously stated, more women were elected under the three-in-six rule than under the one-in-two rule. This result is slightly counter-intuitive, but can be explained when the choices available and their impact on the preferences of parties are taken into consideration. Under the three-in-six rule, there is some degree of flexibility as to where men and women should be placed on the list, allowing parties to accommodate incumbents whilst still respecting the requirements of parity. By contrast, the one-in-two rule is somewhat stricter, especially

116 Parties, Gender Quotas and Candidate Selection in France

as it applies to elections with smaller district magnitude, meaning that anyone placed third or lower on the list is unlikely to be elected. In constituencies where parties had two male incumbents, they were forced to place one of these incumbents third on the list behind the obligatory woman. Some incumbents did not accept this third placing as they knew that it would amount to losing their seat. Instead, they either defected to another party (often at the expense of the woman who might otherwise have headed that party list) or stood as independent candidates. These dissident candidates were often elected in lieu of the women placed second on their original party list. For example, in La Manche in 2001, two male RPR candidates stood on separate lists, leaving the woman in second place on one of the lists effectively relegated to third ranking. Similar behaviour was seen by leftwing parties; for example, in Loire Atlantique, one of the two male incumbents rejected the third place offered to him on the PS list and instead defected to the Greens to head their list, thus undermining both the woman in second place on the PS list, and the woman who might otherwise have headed the list for the Greens. As a consequence, the elections held under the one-in-two rule fell some way short of parity in terms of seats, despite the fact that all parties respected the rule in terms of candidatures.5 The third type of regulation, financial penalties, is also an important factor in determining how parties implemented parity. The rules for legislative elections are quite particular. As French parties are financed by the state in proportion to their electoral performance, the penalty amounts to a withdrawal of state funds rather than a fine. As discussed in Chapter 4, French parties receive two distinct portions of state funding; the first in proportion to the number of votes they receive, and the second in proportion to the number of deputies that they get elected. The parity penalty applies to the first portion of funding. If the disparity between the number of men and women exceeds 2 per cent – that is, 49 per cent of one sex and 51 per cent of the other – the party will have their funding cut by half the percentage of the gender gap. For example, if a party fields 55 per cent men and 45 per cent women, the gender gap will be 10 per cent, resulting in a 5 per cent loss of funding.6 The significance of these regulations lies in the fact that smaller parties who do not win seats depend on the first portion of state funding, while larger parties are more concerned with the greater financial rewards attached to the number of seats won. This means that parties’ strategy will be influenced by whether or not they expect to

The Institutional Approach 117

win seats. Smaller parties cannot afford to lose any of their precious state funding by failing to respect parity. This explains why the FN was so careful to respect parity, despite being the party most ideologically opposed to the law. Clearly, in this case, the institutional effect of parity was more powerful than the ideological opposition to it (the FN leader, Le Pen, was also the only candidate who openly opposed the principle of parity when he stood in the 1995 presidential elections). The other parties without many seats were also careful to minimise their financial losses. For example, in 2002 the main five parties faced an average penalty of 16.69 per cent, whereas smaller parties only sacrificed an average of 5.19 per cent of their precious funds.7 By contrast, larger parties can offset any losses incurred through a breach of parity as long as they win enough seats. If parties believe that they are more likely to win seats with male candidates, the financial penalty will not be an effective deterrent from fielding a majority of male candidates. For example, despite losing nearly 4 million euros per year from the first portion of its funding, the UMP was the richest party following the 2002 legislative elections, with an annual state subsidy approaching 30 million euros (Bordenave, Le Monde 22 June 2002; Zimmerman 2003a). Similarly, in 2007, the UMP had approximately 50 per cent more funding than the PS despite having barely half the proportion of women candidates. This also helps to explain why the legislative elections generated so few female deputies, despite the relatively high number of female candidates: not only did the law not stipulate whether or not women should be placed in winnable seats, but the degree to which parties respected parity also appeared to be inversely proportional to the number of seats that a party expected to win. The perfect parity displayed by the smaller parties who could not afford the financial penalty had no impact on the final composition of the legislature. Conversely, the most seats were won by the parties with low proportions of female candidates. Table 5.2 demonstrates the proportion of female candidates fielded and elected, and the penalties incurred by various parties. It should be noted that in addition to the gap between large parties and small parties, there is also a noticeable gap between the parties of the left (the PS, PCF and Greens) and the parties of the right (the UMP and UDF). This suggests that institutional explanations may not be sufficient – ideology is still a significant variable. Overall, the difference between parties can largely be attributed to the combination of the electoral system and the nature of the parity law, although ideology does still play a part (see Chapter 6 for a more

118

Table 5.2 Party

UMP PS2 UDF3 MoDem PCF Greens FN

The impact of parity Number of seats won

Number of candidates fielded

% women candidates

Penalty as % of funding1

Penalty in euros

2002

2007

2002

2007

2002

2007

2002

2007

2002

2007

365 149 29 – 22 3 0

313 186 22 3 15 4 0

577 535 226 – 491 460 570

548 510 104 529 384 524 557

19.6% 34.6% 19.9% – 43.6% 48.9% 48.4%

26.6% 46.5% 27.9% 36.9% 48.2% 50.2% 48.8%

30.4% 15.4% 30.1% – 6.4% 1.1% 1.6%

23.4% 3.5% 22.1% 13.1% 1.8% 0% 1.2%

3,936,742 1,519,397 624,726 – 118,385 19,110 67,779

4,127,0414 516,806 n/a4 449,548 64,107 0 19,713

Sources: Zimmerman (2003a), pp. 44–5; National Assembly; OPFH. 1 This refers to the first portion of funding, namely the portion based on number of votes received. 2 This figure includes minor parties who affiliated themselves to the Parti Socialiste. 3 The 2007 data for the UDF is for Nouveau Centre. 4 The 2007 figure is for the UMP and Nouveau Centre combined.

10.1057/9780230275294 - Parties, Gender Quotas and Candidate Selection in France, Rainbow Murray

The Institutional Approach 119

detailed discussion). These rules have a differential impact on different parties and frame their choices and preferences accordingly, with smaller parties more motivated to implement parity than larger ones. The stark contrasts between parties of different sizes and resources deserve further consideration. The nature of the parity law goes some way towards explaining these contrasts, but some of the institutional effects detailed above only become apparent when considered in conjunction with the nature of the French party system. The party system Politics is not a game of solitaire; winning in politics is not just about working the rules of the game to your advantage. Another key factor to take into account is who the other players are, and what their relative strengths are. The number of parties is important, as is the potential for collaboration and interaction between these parties. As Sartori argues, ‘the number of parties immediately indicates, albeit roughly, an important feature of the political system: the extent to which political power is fragmented or non-fragmented, dispersed or concentrated. Likewise, simply by knowing how many parties there are, we are alerted to the number of possible ‘interaction streams’ that are involved’ (1976, p. 120). Duverger supports this point of view, arguing that the nature of a party system is determined both by the number of effective parties and by the alliances between them (1972, p. 20). A larger party system, numerically speaking, creates greater ideological diversity between parties, increases competition at a range of points on the ideological spectrum rather than simply at the centre, and increases the potential for political parties to interact and negotiate with each other in order to strengthen their position. As demonstrated below, these factors are all significant for French parties when selecting candidates. France has a party system with a relatively large number of parties; there are five main parties with coalition potential, and many smaller parties who have success at lower levels of government. Competition for votes is tight, and the two-ballot electoral system enables many smaller parties to stand in the first round of voting. This can have a detrimental effect on large parties, especially if the smaller parties are close to the larger one on the ideological spectrum, as the splitting of the vote in the first round might prevent any party from that part of the spectrum from qualifying to the second round. In order to maximise their electoral chances, several parties of the left (and, to a lesser extent, certain parties of the right) work together as a loose coalition, with pre-election agreements that certain constituencies will be

120 Parties, Gender Quotas and Candidate Selection in France

reserved for certain parties in order to concentrate the left– or rightwing vote.8 The division of the mainstream parties of left and right into two electoral blocks – the ‘bipolar multipartism’ referred to in Chapter 2 – is a side-effect of the electoral system and can be detrimental to parity in three ways. Firstly, inter-party negotiations reinforce the emphasis on incumbents, as a party is much more likely to be successful in claiming a winnable seat for itself during negotiations if it is refielding the incumbent. As most incumbents are men, this can have a negative effect on parity. Secondly, as a result of these negotiations, parties will have a clear idea in advance of which seats they are most likely to win, which gives them the option of whether or not to respect parity within these priority seats. Competition for places in these seats will be intense, meaning that the stakes are high and women risk getting overlooked in favour of men. Thirdly, these parties tend not to field a candidate in every seat, as some seats are ‘reserved’ for their coalition partners. This can increase the competition for the seats that are available, which can work to the detriment of women. We have also noted that the rules of parity have a varying impact depending on how many seats a party expects to win. An extra factor to bear in mind is that smaller parties do not just field more women because it is financially expedient to do so, but also because they do not face the same imperatives for male candidates that larger parties face. This is primarily because smaller parties do not have the problem of incumbents. As we saw in Chapter 4, incumbents have been a significant barrier to the implementation of parity because replacing an elected male with an elected female effectively means deselecting an incumbent. The natural rate of political renewal (for example from retirements) is insufficient to meet the demands of parity. There have also been protests in parties at the idea of giving all newly available seats to women, as young men are thus deprived of the opportunity of a political career (for example, a petition was circulated in the PS by men who considered themselves to be a ‘sacrificed generation’ (Mandraud, Le Monde 05 February 2006)). It is therefore much harder for parties with a larger number of incumbents to implement parity than it is for smaller parties who have few or no incumbents. On a related note, the stakes are lower for smaller parties because competition to be a candidate is less fierce when the party has no chance of winning the seat. If a party is struggling to find enough candidates to fill each seat, it is far more likely to embrace female candidatures than a party which already has several people fighting to represent each seat. This principle of allowing more women in when

The Institutional Approach 121

the stakes are lower can be expanded to explain party behaviour across the range of elections; the higher the stakes, the rarer women become. One exception to this rule is if a party is in decline. A party with a dwindling number of seats, as is the case for the PCF, will find it harder to implement parity within the remaining seats because there will be a growing pool of male former incumbents that have lost their seats competing for a decreasing number of winnable seats, resulting in additional difficulties for the party in implementing parity.9 The party system therefore has significant effects on party choices and outcomes. French parties have to incorporate negotiations with coalition partners into their electoral strategy and their choice of electoral candidate. The freedom available to parties to select the candidate of their choice is also affected by the party system. Larger parties may enjoy a wider pool of aspirant candidates from which to choose but may also be limited by the need to reselect incumbents. Smaller parties without incumbents do not have this obligation and are less likely to suffer electoral consequences for fielding an unknown female candidate. Thus, the number of parties, their size and power and their interactions with each other are all significant to predicting the likelihood of women’s selection as candidates. Parties adapt their strategies according to the situation they are in. One final factor is of note here. We have seen that parties are more likely to struggle to field women when the seat is winnable, because there will be greater pressure from male candidates. This fits into a wider pattern of parties being less likely to field women when the stakes are high, an area which is considered in greater detail below. The importance of the election If parties are rational and self-interested, then their willingness to play by or circumvent the rules depends on both the ease of defection and the cost of compliance with these rules. For each election, compliance with parity incurs a cost if parties are prevented from fielding their preferred candidate. However, defection will also entail a cost as parties are penalised for not respecting parity. Where parties place a high priority on selecting one candidate over another, the temptation to defect from the obligation of parity will be higher, as the costs of compliance (not fielding the preferred candidate) begin to outweigh the costs of defection (penalty for non-compliance). So parties will evaluate how rational it is to comply with parity based not only on the costs, but

122 Parties, Gender Quotas and Candidate Selection in France

also the benefits of defection, and these rise with the status of the office at stake. Parity has obliged parties to cede a considerable amount of political space and power at all levels to women which was previously held by men. It is striking, however, how resistance to parity increases depending upon the importance of the election. For example, within local politics, power is concentrated in two areas: firstly, the conseils généraux (departmental councils); and secondly, municipal executive office, and mayoral office in particular. Unsurprisingly, as discussed in Chapter 3, these are the areas of local politics that have become the least feminised. The conseils généraux, broadly exempt from parity regulations, remain the most masculine of all France’s political institutions.10 Meanwhile the office of mayor is predominantly held by men, and the percentage of female mayors, although feeble, is still at its highest in the smallest towns (14.2 per cent) (OPFH). It is no coincidence that female mayors are more prevalent in the towns where the role is the least significant. Furthermore, research by Bird after the 2001 local elections demonstrated that women become rarer as the value of the executive office becomes higher. For example, in towns with over 3500 inhabitants, women comprised 47.5 per cent of council members, but only 36.5 per cent of members of council executives. For the more coveted position of deputy-mayor (premier adjoint), this figure fell to 26.3 per cent, while for the top job of mayor, women comprised a mere 6.7 per cent (Bird 2004). This is despite the fact that, for the basic position of councillor, parties exceeded the minimum requirement of placing women fourth, fifth and sixth on lists (which would have resulted in a total of 43 per cent women, rather than the actual figure of 47.5 per cent) (Fabre, Le Monde 27 March 2001b). The recent change to the law, obliging parity in local executives as well as councils, has had a mitigating but incomplete impact, with the proportion of women mayors rising to 9.6 per cent, and deputy-mayors rising to 36.8 per cent in 2008 (OPFH). It appears that parties do not mind making a little bit of space for women in the lower ranks, as long as it is still men who hold the real positions of power. The same principle can be seen for other elections. A relatively high proportion of women were elected in both the regional and European elections, reflecting the lower status of these elections (the European elections in particular are often viewed as ‘second order’ elections; former deputies will sometimes seek a European mandate to keep themselves busy whilst waiting to regain their seat in the National Assembly, but European office is not seen as a useful springboard for national office).

The Institutional Approach 123

By contrast, resistance to parity was much higher in the senatorial elections, resulting in the dissident candidatures described above. For the most coveted prize of all, a seat in the National Assembly, the combined impact of the highest stakes and the least strict rules produced the poorest performance for parity. In a cost-benefit analysis, the most powerful political positions will motivate parties to pay greater costs to ensure the election of their preferred candidate, because the potential benefits of so doing are sufficiently high. In the case of the legislative elections, the cost of defection from parity is relatively low, making it all the more probable that parties will choose to defect rather than comply with parity for these elections. The perceived value of the different political offices is therefore an important explanatory variable in predicting parties’ response to a law such as parity. Indeed, a similar pattern can be observed in a number of other countries, where the percentage of women in local office exceeds the percentage of women in the national parliament. Examples include Australia, Belgium, Canada, the Czech Republic, Hungary, the UK and the USA (Stevens 2007, p. 79). Conclusion Institutional explanations of party behaviour have a lot to offer in the French case. There are many aspects of candidate choice that have a demonstrable link to institutional variables. For example, there is clear evidence that, overall, women fare better under proportional rather than majoritarian electoral systems. This effect is compounded when used in conjunction with the parity legislation. The ‘parity’ law itself affects parties in different ways depending on their size and wealth. For larger parties, parity is harder to implement but easier to circumvent, while smaller parties do not face the impediment of male incumbents but are more likely to be constrained by the financial penalty. A pattern thus emerges of different parties being affected by institutions in different ways, depending on their resources and on their motivations. The interplay of institutional effects makes some scenarios more advantageous than others, and parties will tend to play the hand that they are dealt to their best advantage. Where they have more to gain by implementing parity than by not doing so – for example, where the law is very restrictive, and/or the stakes are low – then they will respect parity. However, where the law has little or no impact, and/or the political office at stake is of great value, parties will attempt to circumvent parity in pursuit of other goals. As different parties are affected by institutions in different ways depending on their

124 Parties, Gender Quotas and Candidate Selection in France

resources, it is possible to use the institutional approach to account for the variation between different parties, rather than citing this variation as evidence that the approach is flawed. Institutional explanations also account for the variable impact of parity over different elections. Although there is no doubt that the institutional approach is of use in demonstrating how the range of options available to parties is framed and evaluated, it cannot explain everything. There are some aspects of party behaviour that cannot be accounted for by the institutional approach alone. For example, the PS in 2002 was a much larger party than the UDF, and expected to win far more seats. The above theory would therefore suggest that the UDF, with fewer incumbents and less income from the second portion of state finance, would be more likely to respect parity than the PS. However, this was clearly not the case. Rather, it is indicative of a wider trend for parties of the left to be more open to parity than parties of the right, suggesting that party ideology is also an important factor. This is the theme of the next chapter. Without taking into account the effects of party ideology, the institutional approach remains an important, but incomplete, account of party behaviour.

6 The Ideological Approach

Introduction So far this book has looked in some detail at why compulsory quotas may be problematic for political parties, using the example of parity in France. The book has demonstrated that parity places institutional constraints on parties that limit their ability to select their preferred candidates. Parties are torn between the obligation to respect parity and the desire to keep incumbents happy, retain support amongst their membership, negotiate with coalition partners and, most crucially, win elections. While the claim that women candidates lose votes does not appear to be substantiated, there is certainly no solid evidence to suggest that fielding more women poses an electoral advantage to parties. For all these reasons, parties’ unwillingness to comply with the parity law appears understandable. What is harder to comprehend is why parties would choose to introduce or implement parity under any conditions other than absolute obligation. Explaining this is the purpose of this chapter. This chapter explores the role of party ideology as a motivating factor for supporting parity. It argues that parties serve a purpose beyond basic existence and office-seeking; that they are programmatic in nature and that they have fundamental beliefs about how the world should be. Regardless of necessities and ambitions (such as survival and election), all parties will retain at least some ideals and offer their vision of the world to the electorate as their way of distinguishing themselves from other parties. This world vision can include ideas about how society should be organised, how resources should be distributed, and the role of the state and the citizen. Gender relations can form a part of party ideology, especially if framed within a wider discourse of equality. Where party 125

126 Parties, Gender Quotas and Candidate Selection in France

ideology does incorporate a desire for gender equality, parties may be more likely to support parity. The greater the importance attached to equality within party ideology, the greater the chance that the party will prioritise parity and make sacrifices to ensure its effective implementation. Ideology may therefore offer the key to why some parties implement parity despite the compelling arguments against doing so. To test this theory, this chapter will first examine the literature on party ideology and the ideological approach, and discuss whether ideology really does matter or not, and if so, how. Parties will be considered comparatively using a framework that identifies their key ideological traits. The existing literature on the role of party ideology in determining levels of women’s representation will also form a crucial part of this discussion. I will then consider how these theories can be applied to the case-study of French parity to see whether, and under what conditions, ideology might be used to explain party choices when selecting candidates. Variables such as level of support for gender equality will be compared against the main French parties to see whether they are, in fact, useful predictors of party behaviour. The utility of the ideological approach, especially when considered in conjunction with the other approaches explored in this book, will then be evaluated.

Examining the ideological approach The electoral competition approach, which was studied in Chapter 4, suggests that parties are concerned above all with winning elections. Downs (1957) argues that ideology is malleable and that parties are willing to shift their ideological stance and relocate themselves towards the median voter in order to maximise their vote share. This behaviour is particularly associated with Kirchheimer’s ‘catch-all parties’ thesis (1966). However, not all parties can be categorised as ‘catch-all parties’ (indeed, most cannot), and even those that can will still have distinctive programmes that are based on a deeper underlying philosophy about how society should be and how the party could change and affect this if elected. All parties put forward programmes and manifestos on which to be judged by the electorate, and voters are not just selecting individuals to lead them but are also choosing between competing policies, programmes, visions of society. The roots of these competing visions are ideological and sociological and explain why loyalty to a particular party goes beyond appreciation of an individual leader or candidate. It is a party’s ideas which define it, which give it its raison

The Ideological Approach 127

d’être and which allow voters to identify with it. Indeed, von Beyme claims that ‘over the longer term only parties based on an ideology have succeeded in establishing themselves in Europe’ (1985, p. 29). Ideology also offers voters a more fundamental indication of what a party stands for than does an electoral manifesto, for example by giving an indication of how a party might adapt to a new issue or unexpected event. Marks et al (2002) argue that ‘basic social divisions give rise to ideological commitments that condition the response of political parties to new issues’ (p. 592). If we accept that all parties have some degree of ideological foundation, we then need to establish some further parameters. Parties need to have some kind of distinctiveness from one another, to offer competing visions of society that offer some form of choice to the electorate between one programme and another. We might therefore expect policy differences between parties and different stances on more abstract ideals and principles such as equality. A further question lies in the extent to which ideology will triumph over other considerations. We have already seen how ‘catch-all’ parties perceive ideology to be fairly malleable. If the Downsian spatial analysis is used, with parties being located on a political spectrum from left to right, then parties are perceived to converge towards the centre in order to appeal to the maximum range of voters. However, actual practice confirms that this is rarely the case for any party and certainly not the case for all parties. This suggests that for some parties, maximising their vote share will be a higher priority than for others, with some parties more concerned with ideological purity even if their refusal to broaden their appeal results in the limited development of a party. A party that seeks to be a big tent will invariably have to encompass a wide range of political views within it and will offer a less clearly defined message to the electorate, allowing a wider range of people to find something within the party that they identify with. The classic example of ‘big tent’ parties are the Democratic and Republican parties in the USA, sometimes referred to as ‘empty vessels’ due to the fact that their programmatic nature is submerged by the diversity of their beliefs and membership (Katz & Kolodny 1994). (Although even this classic view is contested, for example by Gerring, who argues that ‘there are plausible grounds for regarding the American parties as ideologically motivated’ (1998, p. 257), by Grofman et al (1999) who argue that the Democrats are considerably more catch-all than the Republicans and that this is to the Democrats’ cost, and by Ware (1996, p. 62).) Big tent parties are more likely to be found near the centre of the ideological spectrum,

128 Parties, Gender Quotas and Candidate Selection in France

even if they can still be distinguished from one another with some parties leaning more to the left and others more to the right.1 By contrast, other parties will be willing to be located further away from the political centre. Communist, green, agrarian and nationalist parties often fall into this second category. These parties are more likely to offer a narrow and consistent ideological stance which remains rigid and is the defining reason for the party’s existence. By limiting the appeal of their policies to a smaller sector of society, these parties are unable to achieve the size or electoral success of ‘big-tent’ parties. However, it is easier for them to maintain a distinctive identity and for them to remain true to the ideals that underlie their existence. Using Strøm’s typology, the ‘big-tent’ parties might be considered to be office-seeking parties, while the smaller parties that focus on policy objectives could be considered policy-seeking parties (Strøm 1990). Epstein argues that ideology is also more closely linked to massmembership parties. He claims that ‘in order for a party to attract regular dues-payers…it is useful if not essential to have a program to which loyalty can be given over a period of time, between elections as well as during campaigns’ (Epstein 1967, p. 261). Party members, he argues, will be loyal to the ideology behind a party and will resist attempts to change that ideology. By contrast, ‘a non-membership party, even if with a well-developed program…would have little trouble changing or discarding it when expedient for electoral purposes’ (ibid). While this latter description might be in accordance with Kirchheimer’s model of catch-all parties, many parties do fall into the programmatic category, and for these parties ‘belief in a cause is the substitute for material interest’ (ibid). This would support the notion that parties might be willing to make sacrifices, such as the loss of electoral support, in order to remain true to their programmatic beliefs. Party ideology may also be perceived in terms of parties’ sociological origins. Parties may be seen as the organised representation of particular interests. For example, the origins of many political parties are classbased, with conservative parties representing traditional elite interests whilst communist and socialist parties were developed to represent the interests of the proletariat. Partisan identification can be conceived of in terms of cleavages (Lipset and Rokkan 1967). This theory was developed by Klaus von Beyme (1985), who organises parties into families based on their origins and representation of interests, and this is explored in greater detail below. A party that has the representation of particular interests as its raison d’être will be keen to ensure that its policies are in keeping with these interests. Although these interests may

The Ideological Approach 129

not be directly ideological in nature, certain premises often underlie them that influence the party’s programmatic aims. For example, a socialist party that seeks a more equal distribution of wealth might support egalitarian policies more generally, while a party that was more focused on the individual might reject such policies. It appears, then, that the importance of ideology will vary from one party to another. For some parties, ideology is a general principle that allows them to distinguish themselves from their nearest rivals when contesting elections. For others it is more fundamental, and programmatic purity will override pragmatic concerns. However, it is misleading to assume that any party, even the most electorally motivated, will have no concern for ideology. Ideology does not just serve to differentiate parties from one another and appeal to different sections of the electorate; it is also what inspires the actors within parties and motivates them to pursue the practice of politics. Ideology is fundamental to every political actor. As Blondel puts it, ‘ideological struggle continues to be an essential part of political conflict’ (1978, p. 120). The key difference between parties is not whether they have ideological foundations or not, but rather the extent to which they are prepared to adapt and compromise that ideology in the pursuit of power. If considered from a rational perspective, the extent to which parties will be rational in allowing ideology to dictate their behaviour depends on the fundamental desires of a party. A party that is based on an alternative vision of society in which environmental concerns are key to all policy-making might be rational to pursue policies that are influenced by this vision, such as raising taxes on fossil fuels, even if these policies are unlikely to appeal to most of the electorate. The party might not be seeking to govern single-handedly, but rather to influence the policy agenda through garnering support for their ideas, demonstrating the popularity of these ideas amongst certain sectors of the electorate, and possibly gaining sufficient electoral strength to have coalition potential. Under such circumstances, it would be irrational for this party to abandon the ideals that underpin its existence and opt for watereddown policies that are less in keeping with its ideology. On the other hand, a party which feels that the first step towards effecting change in society is obtaining the power to govern will be more rational in adapting its policies to the electoral market in order to win office. Such a party might argue that it is better to abandon 50 per cent of its ideals in order to be sufficiently appealing to the electorate to win office and implement the remaining 50 per cent of its programme than it is to remain entirely true to its ideals, lose elections and implement 0 per

130 Parties, Gender Quotas and Candidate Selection in France

cent of its programme. From this perspective, a certain malleability of principles is not only rational but also demonstrates how electoral motivations might still be influenced by an underlying desire to achieve programmatic aims. One further thing to bear in mind when considering the extent to which party ideology is malleable is consideration of how far along the political spectrum a party is capable of travelling (assuming, of course, that one accepts that parties are indeed located within the spatial concept of a ‘spectrum’). Parties may be able to relocate themselves further from or closer to the centre, but there is a limited capacity for travel within the spectrum, and the larger the party system, the greater the limitations on parties to move from their established position on the spectrum. It is difficult for one party to ‘leapfrog’ another without risking alienating their core voters. For example, a left-wing party that moves so far right that it leapfrogs a more moderate left-wing party might find its core supporters abandoning it for the party that it leapfrogged, which has now become the party closest to their ideals, or – as Downs (1957) acknowledged – its supporters may simply abstain. Moreover, where there is ‘crowding’ of a limited ideological space, there may not be anywhere else for a party to move, as the space on either side of the spectrum to the party in question might already be occupied by well-established parties (Gerring 1998, p. 32). Budge et al further argue, as we saw in Chapter 4, that certain parties will have ‘ownership’ of certain policy issues and that parties are therefore wise to focus on the issues that are positively associated with their party. Any attempt by a party to take a position on an issue that ‘belongs’ to another party will only help to further the interests of the second party rather than broadening the appeal of the first party. Budge et al argue that parties are therefore limited in their ability to move towards the median voter on more than a restricted number of policies, leaving parties locked in to their established ideological territory. ‘In spatial terms this implies that certain areas…are open only to one party – certain policy areas “belong” to it, as do the votes of electors found there. In other words we are dealing with “partyreserved space”, with some areas open to only one party, rather than a pure policy space of unlimited free movement for parties, such as that postulated by Downs and Hotelling’ (Budge et al 1987, p. 27). This argument is furthered by Klingemann et al (1994), who contend that parties build up policy positions over time that result in ‘essential and indelible associations with particular issues and policies’ (p. 24).

The Ideological Approach 131

A picture thus emerges of a situation whereby ideology does matter and does inform the policy direction that parties will take. Parties will be willing and able to adapt and modify their policy positions to varying degrees, depending on a range of factors including how entrenched their ideology is, whether ideological purity is of greater or lesser importance to them than electoral success, whether they are trying to speak primarily to their core membership or to maximise their appeal to a broad electoral base, and the extent to which issues and positions on an ideological spectrum are captured by themselves or their competitors. If a party has strong ideological roots and is primarily concerned with honouring these roots and appeasing its membership, then ideology will play a key role in determining policy positions. If a party is trying to broaden its appeal to as wide a section of the electorate as possible then the catch-all thesis of Kirchheimer and the malleability of ideology for electoral purposes suggested by Downs become more likely explanations of party behaviour. Even in the latter case, ideology cannot be dismissed entirely. Most parties, even those that are primarily electoral-professional, can be classified within a broad ideological grouping or ‘family’ that links their ideas to those of other parties in different systems. The next section explores this concept in greater detail.

‘Familles spirituelles’: Comparing ideologies and attitudes to quotas Klaus von Beyme adopts a historical sociological approach to the study of parties and ideology and argues that it is possible to categorise parties from different countries across the world into one of nine broad ideological families, or ‘familles spirituelles’ (1985). His is the most robust and comprehensive of several attempts to group parties according to their ideology (see also Blondel 1978; Lipset & Rokkan 1967; Mair & Mudde 1998; Rokkan 1970). Von Beyme acknowledges that his model is still not broad enough to include every single party (and is perhaps even less comprehensive than von Beyme claims, as Ware points out (1996, p. 23)), but the model is still useful in highlighting similarities and ideological convergences between seemingly different parties. A detailed definition of each of the familles is not required here. Rather, the aim is to identify whether France’s parties can be located within von Beyme’s model and, if so, what insights into these parties’ ideologies can be obtained. The comparative literature on gender and politics will be used to see if similar parties in other countries have proven more or less willing to support gender quotas.

132 Parties, Gender Quotas and Candidate Selection in France

The nine party categories deployed in von Beyme’s schema are Liberal and Radical; Conservative; Socialist and Social Democrat; Christian Democrat; Communist; Agrarian; Regional and Ethnic; Right-wing Extremist; and Ecology. Knapp argues that ‘party divisions in France have never corresponded exactly to von Beyme’s familles spirituelles, and the fit is becoming less and less perfect’ (2004, p. 15), but the familles are still the most useful tool for placing French parties in a comparative ideological context, and some French parties do correspond fairly well to von Beyme’s model. The PS belong to the socialist family, as the PCF belong to the communist family. The FN are right-wing extremist and the Greens fit in the ecology family. The UMP and UDF are rather harder to categorise within this model. The UDF has contained an unusual hybrid of liberal and radical elements with Christian democrats – given the anticlericalism of the former and the Catholicism of the latter, this is a union that cannot be accounted for within von Beyme’s typology. The UMP are the inheritors of France’s Gaullist traditions, which also defy classification, although since the 1970s the Gaullists have developed into a party that best fits the conservative party family model. Von Beyme considers the communists to be the most ideological of all party families. Their strongest emphasis is on class and the redistribution of capital and the means of production. They favour nationalisation and a strong central state. Social equality is perceived primarily in class terms. Some of these themes are also present in the socialist family, who believe in strong state intervention in the economy and public ownership. They also seek to enhance social equality through the most realistic means available, such as the social and welfare state (1985, p. 70). Socialist parties are less revolutionary than communist parties, being attached to the concepts of democracy, universal suffrage and an acceptance of the legitimacy of the political system. Some parties have softened their stance over time, to the extent of becoming ‘catch-all’ parties (1985, p. 75). Von Beyme categorises ecology movements as being difficult to place on the left-right scale due to their diverse social origins and their mixture of ideologies. At the time that he wrote this, ecology movements had only recently emerged. Writing more than ten years later, Ware (1996) claims that ecology movements are a form of postmaterialist New Left party family. Kittilson (2006) similarly classifies green parties as having ‘New Left’ values such as permissive social values and a concern for the environment.

The Ideological Approach 133

The ideologies of right-wing extremist parties are not well specified by von Beyme, who provides instead a rather descriptive account of the history of these parties across Europe. Commonalities between these parties tend to include ultra-nationalist doctrines, xenophobia, an authoritarian approach to law and order, and hostility towards immigration. Economically their policies are much harder to compare. Similarly, von Beyme claims that it is very difficult to compare the programmatic principles of Conservative parties from one country to another. Recent trends include advocating a reduction in state involvement in the economy, privatisation, traditional social values and nostalgia for the past. Finally, the definitions of liberal-radical and Christian democrat parties are of rather limited utility when studying French parties as they fail to account for the specificities of the UDF. So what can all this tell us about French parties and their attitudes towards parity? Placing parties within internationally recognisable ideological groupings facilitates comparison with similar parties in other countries and makes it easier to identify international trends concerning party attitudes towards women. Works from the gender and politics literature are particularly instructive here. Lovenduski and Norris’s comparative study (1993) identifies certain party traits that are associated with attitudes towards quotas. Lovenduski argues that ‘parties of the left have traditionally been more willing than parties of the centre and the right to make agreements to nominate women and they also appear to be more able to deliver on such agreements’ (1993, p. 13). She does go on to qualify this claim by arguing that ‘ideology is a less reliable indicator of party support for women’s representation than it once was’ (ibid). At the end of the same volume, Norris concurs, and introduces an additional variable, namely the willingness to use state intervention to achieve equality goals. She claims that: [S]ocial democratic and Green parties are far more likely to believe intervention in the recruitment process is necessary and appropriate, hence positive discrimination is justified to bring about short-term change. Parties of the right and centre are more likely to rely upon rhetorical strategies, and possibly affirmative action, in the belief that women should be encouraged to stand, and party members should be encouraged to select them, but the recruitment process has to involve ‘fair’ and open competition. … Just as parties of the right tend to favour a minimal role for government in the

134 Parties, Gender Quotas and Candidate Selection in France

free market economy, so they lean towards non-intervention or regulation of the candidate recruitment process. (Norris 1993, p. 320) Norris supports this argument with evidence of the use of various interventionist mechanisms to raise women’s presence in politics in France, Britain, Norway, the Netherlands and Germany. This initiative from left-wing parties may sometimes lead to a ‘contagion’ effect whereby parties of the right follow suit, but usually out of a wish to appear modern and competitive rather than out of conviction (Caul 2001; Matland 2006; Matland and Studlar 1996). Norris’s findings are confirmed by Miki Caul Kittilson. She claims that ‘left-wing parties may be more likely to support women’s candidacies than right parties because left parties espouse egalitarian ideologies’ (Caul 1999, p. 81). She further argues that New Left parties such as Green parties, with their increased emphasis on minority rights and social equality, may be even more favourable to women’s representation than Old Left parties such as socialist parties (ibid, p. 82; Kittilson 2006, p. 46). This is confirmed by the findings of her comparative study which indicate that left-wing parties are more favourable to women’s representation than right-wing parties, with environmental and communist parties having the highest percentages of women, and ultra-right parties having the lowest (Caul 1999, p. 85). A trend thus emerges from the comparative literature whereby leftwing parties tend to be more favourable both to gender equality and to state intervention, both of which are essential for support of a constitutionalised quota like parity. By contrast, right-wing parties are more likely to have traditional, conservative views about women and the family and are also individualist and state reductionist in their approach and therefore less likely to support measures such as compulsory quotas. The next section explores these ideas further and develops them into testable hypotheses that are then applied to the case study of French parties.

Can ideology explain the implementation of parity? There is some evidence to suggest that parties will have differing attitudes towards gender quotas, driven by their underlying ideologies. This section explores whether these attitudes can be used to predict and/or explain the extent to which different parties implement the parity law. Given the costs associated with implementing parity, it is

The Ideological Approach 135

possible that ideology might be able to explain why parties would be motivated to introduce and implement parity in spite of these costs. This might explain why parties on the left, including the PS, PCF and Greens, have shown a greater inclination towards parity than have the equivalent parties on the right. However, there is no perfect correlation between ideological support for gender equality and the implementation of parity, as Table 6.1 reveals. Table 6.1 demonstrates a general trend towards parties of the left being more favourable to gender equality than parties of the right, but it also reveals three immediate contradictions to the theory that this will result in parity implementation. Firstly, the PS fielded significantly fewer women than the other two parties on the left in 2002. Secondly, they also fielded fewer women than the FN, despite the fact that the FN was the party most openly opposed to parity. Thirdly, women candidates from the PCF were particularly likely to be placed in seats where they did not win. The causes of these contradictions are explored in greater detail below. The implication here is that ideology, whilst of some explanatory power, cannot be considered alone when explaining levels of implementation of parity. To what extent, then, and under what conditions might ideology be a useful predictor of parity implementation? I hypothesise that for Table 6.1

Ideology and implementation of parity

Party

Left

Right 1

PCF Greens PS UDF UMP FN

Believes Believes % women in gender in state candidates equality intervention to achieve equality outcomes 2002 2007

2002

2007

Yes Yes Yes Neutral1 Neutral1 No

18.2% 33.3%2 16.8% 6.9% 10.4% n/a4

16.7% 25% 26% 0% 14.2% n/a4

Yes Yes Yes No No No

43.95% 48.91% 34.58% 19.91% 19.58% 48.42%

48.2% 50.2% 46.5% 27.9%3 26.6% 48.8%

% women elected

The UDF are neutral in that they are neither strongly for nor against gender equality. The UMP are also neutral in that there are elements within the party that strongly advocate gender equality, elements that oppose it, with most of the party falling somewhere in-between. 2 As the Greens only won three seats, this was the closest that they could achieve to parity. 3 The 2007 data is for Nouveau Centre. 4 The FN did not win any seats in 2002 or 2007.

136 Parties, Gender Quotas and Candidate Selection in France

parties to be motivated by their ideology to implement parity, three conditions need to be met. The first two of these three conditions have already been touched upon in the previous section; parties need to have ideological support for gender equality, and they also need to agree that a compulsory measure introduced via the state is the best means of achieving the goal of gender equality. If parties do not support gender equality in principle, then they will not be motivated by their ideology (although perhaps by other factors, as will be considered below) to implement parity. Likewise, parties might agree with gender equality in principle, but disagree that enforced parity is the appropriate means to go about achieving equal numbers of women politicians in practice. The debates surrounding parity, explored in Chapter 1, illuminate this mentality. Many of the opponents of parity were feminists who were all in favour of seeing more women in politics, but they did not agree with the use of a compulsory quota to achieve this aim. Beliefs in universalism, small government, or that an artificial measure would be patronising have all contributed to a rejection of parity for reasons other than a corresponding rejection of gender equality. This second condition is as important as the first in predicting ideological support for parity. The third condition is somewhat more complex. Parties may be motivated by both of the first two conditions to implement parity, but still not regard it as a sufficiently high priority to override other factors. The first two conditions relate to ideological direction; this third factor relates more to the strength of ideological attachment. The first section of this chapter demonstrated that ideology matters more to some parties than to others. Parties that have strong ideological foundations are more likely to let their programmatic beliefs dictate their policies. This is true of both the PCF and the Greens. By contrast, the PS falls much more closely into the ‘catch-all’ party type described by Kirchheimer, whereby ideology becomes secondary to electoral goals. If a party has weak ideological support for a policy, this may not be sufficient to override more pressing imperatives such as the need to appease incumbents or to win over voters.2 So the third condition is that the degree of ideological support for parity should be sufficient to over-ride other concerns when selecting electoral candidates. To summarise, I argue that ideology will be most effective as a predictor of the positive implementation of parity if parties: a. support the principle of women’s representation; b. support the practice of parity as a means of achieving it; and c. are sufficiently motivated by this ideological support to be willing to bear any costs incurred by prioritis-

The Ideological Approach 137

ing parity over other imperatives when selecting electoral candidates. The validity of these conditions when applied to French parties will be explored in the next section. Firstly, an obvious caveat presents itself: what if parties do not meet these conditions? This needs to be considered in two parts: parties that do not agree with parity, and parties that do agree with it but do not consider it a priority. If a party does not meet the first two conditions – those concerning the direction of ideological support – then they are unlikely to support parity. In this respect, ideological indifference, and especially ideological opposition, to parity may be negative predictors of parity’s implementation. A party that does not approve of parity in principle may be reluctant to implement it in practice. Yet this fails to explain the anomalous behaviour of the FN. Chapter 5 demonstrated that the FN is compelled to support parity for financial reasons, as it is dependent on the portion of state finance that is targeted by the sanction for non-compliance with parity. Thus the FN does implement parity but for reasons other than ideological belief. This implies that while ideological support for parity may be a useful positive predictor of parity’s implementation, ideological opposition to parity may not be an effective negative predictor. Parties that do not agree with parity might implement it anyway but for other reasons. In practice, therefore, ideological explanations are limited to cases where parties are motivated by their support for parity. If a party does not meet the third condition – concerning the strength of ideological support – then their behaviour becomes harder to predict. A party might not be sufficiently motivated by their beliefs to implement parity at any cost, but might still be willing to tolerate a higher cost in parity’s implementation than a party that had no support for parity. Ideology may still be of significance even if it is not the key determinant of party behaviour. This is obviously rather subjective and difficult to measure. A more sophisticated analysis of the interplay of ideology with other variables is offered in the next chapter. At this stage, it is useful to see how well the theories mapped out above explain actual party behaviour in France.

Applying the ideological approach to French parties This section draws on the previous two sections to see how well the ideological approach can explain the extent to which different parties choose to implement parity. The ideological directions of the six main

138 Parties, Gender Quotas and Candidate Selection in France

parties studied in this book are explored in greater depth and tested against the theories outlined above. The PCF The Communist Party has strong ideological foundations. The principles of communism place the party towards the extreme left of the ideological spectrum and this, combined with its reliance on its membership, has made it a poor candidate for a move towards electoralprofessionalism. That is not to say that the party is entirely committed to ideals at the expense of pragmatism. Indeed, the PCF has had to adapt its core beliefs in several ways in order to survive. Firstly, the party slowly abandoned its international orientations and its allegiance to the Soviet Union (Knapp & Wright 2006, p. 180). Secondly, the party has had to cope with a sharp decline in its core electorate due to sociological changes in France. The shift from an agrarian and industrial economy to a service economy, with a corresponding decline in blue-collar workers and a rise in white-collar workers, has robbed the PCF of its core constituency. The PCF routinely commanded around 25 per cent of the vote in the post-war period, and this figure has dwindled over the years to just 3.37 per cent for its presidential candidate in 2002 (Bell 2003, p. 30). Faced with a spiralling decline in popularity and a shift of direction towards domestic politics, the PCF was forced to contemplate abandoning its anti-system status and collaborating with the PS. Once unthinkable to both parties, a coalition gradually emerged, with varying levels of support for the PS from the PCF. Having abandoned its revolutionary principles and recognised the legitimacy of the state, the PCF was able to enter government in coalition with the PS, allowing both parties the combined electoral clout to defeat the Right for the first time in 1981. The PCF now directs its ideas more towards defending workers’ rights and public services. Although its ideological origins are still present in its current doctrine, most of its defining policies have been modernised and adapted, demonstrating that the party is not so tied to its programmatic beliefs as to be incapable of survival. In terms of gender equality, the PCF espouses egalitarian ideals but many of these focus on class issues rather than gender. It contains within its membership a number of working-class men that have not been particularly interested in or even favourable towards questions of equality for women. Nonetheless, the overall doctrine has been broadly favourable to women and the party has a comparatively good track record for fielding women candidates, even prior to the

The Ideological Approach 139

introduction of parity, as Muguette Jacquaint was keen to emphasise.3 Mme Jacquaint added that the party has shown a strong commitment to gender equality, both in terms of parity and ‘in all other areas as well’. The party currently has a woman leader (Marie-George Buffet). Parity fits in well with the party’s egalitarian ideals. It also accords well with the party’s emphasis on achieving equality through state intervention. For both these reasons, the PCF is a good example of a party that shows ideological support for parity. The strength of this ideological support is harder to measure. The party is committed first and foremost to class issues and is also fighting a losing battle against decline, and for both these reasons gender equality is not top of the party’s list of priorities. The party’s attempts to implement parity have been hampered by this decline and also by the party’s insistence on internal democracy and according a high level of autonomy to its membership, as outlined in Chapter 3.4 The belief in respecting the candidate choices of local cells wherever possible has superseded the desire to achieve a parity of candidates, which explains why the party selected 43.58 per cent women candidates in 2002 rather than the 50 per cent target, despite having few incumbents.5 Furthermore, the party’s declining vote share has reinforced the emphasis on the incumbents it does have because, as explained in the previous chapter, it is through re-fielding an incumbent that the party has the strongest claim to seats within a left-wing electoral coalition. As most of these incumbents are male, this would explain why there was such a gulf between the percentage of women selected and elected in both 2002 and 2007. The overall picture for the PCF, then, is of a party that has strong ideological roots that support gender equality without prioritising it, and the party’s seemingly terminal decline has further reduced the emphasis on parity as the party deals with more urgent imperatives. The Greens The Greens are perhaps the best example of a party that meets all three of the conditions for supporting parity. The Greens are a small party that is very programmatic in nature, built on a strong New Left philosophy of environmental protection and social justice. Gender equality accords well with the party’s vision of a fair and just society and also corresponds well to the party’s idealist, post-modern demographic. According to their women’s officer, parity is part of their culture.6 They have long had a woman leader, Dominique Voynet. The party has proven itself willing to embrace gender equality and to use quotas as a means

140 Parties, Gender Quotas and Candidate Selection in France

of achieving this aim. For example, at their annual conference in 2005 they agreed a party policy to respect parity both in practice and principle, that is to say, both in terms of candidates and seats. One reason why the Greens are happy to promote parity is that they have fewer constraints than some of their rivals; for example, their relatively new and small status means that they have few incumbents. This corresponds with findings in the comparative literature; the ‘newness’ of a party is a factor identified by Caul (1999) as being positively associated with quota implementation. The Greens are also more concerned with policy influence and remaining true to their ideals than they are with seeking power. The main way that they have achieved power is by appealing to their core electorate and then forming alliances with larger parties such as the PS, rather than trying to be a large party themselves. This strategy enabled them, despite their relatively weak electoral muscle, to have their leader, Voynet, be appointed to the cherished post of Environment Secretary in the gauche plurielle government of 1997. When a survey of their members was conducted in 1998, 70 per cent felt that their representatives had been influential in the area of the environment while nearly as many (61 per cent) cited gender parity as an area of successful influence (Villalba and Vieillard-Coffre 2003, p. 69). This demonstrates both their commitment to parity and their ability to influence the policy areas that matter to them without having to adopt a more electoralist approach. As such, they are the perfect example of a party that fulfils the conditions for supporting parity through ideological motivation, and this is reflected through their being the closest of the parties under study to achieving parity (see Table 6.1). This concurs with the findings in the literature that New Left and especially Green parties are the most supportive of gender quotas. The PS The Socialist Party is one that has seen a progressive weakening of its ideology in order to be more electorally appealing. Under the Fifth Republic, the party had its first taste of power in 1981. Two years of implementing Socialist economic policies proved disastrous and forced the party to backtrack and take a more moderate stance towards the economy. Further key reforms came in 1991 when the party abandoned its attacks on capitalism and its emphasis on class struggle and replaced these with humanist republican values such as equality and justice (Opello 2006, p. 44). The party has the strongest electoralprofessional tendencies of any of France’s left-wing parties, and is

The Ideological Approach 141

certainly the most mainstream, both in terms of ideas and votes. That is not to say that the party has discarded all its ideology; it still remains programmatic and distinctive from its competitors on both sides of the political spectrum. The PS has been fairly supportive of gender equality in its doctrine and its policies. Despite claims that women-friendly policies were largely a ploy by Mitterrand to woo the female electorate, Opello argues that the party’s underlying ideology is the key determinant of the party’s women-friendly policies. Using interview data of PS officials, she found that: When asked why the party adopted a gender-based quota in 1974, a majority of respondents cited the party’s programmatic beliefs. They believed that women’s advancement was part of a larger goal to advance society in general, and to achieve sexual equality in particular, and that the quota was one way this could be accomplished. Similarly, most respondents stated that Jospin implemented a 30 per cent quota during the 1997 legislative elections because of the party’s programmatic beliefs. In their view, the quota was a way for Jospin to fulfil the party’s goals of modernizing and democratizing society as well as achieving sexual equality in politics. (Opello 2006, p. 50) The PS appears to fulfil both of the first two conditions – it supports gender equality in principle, and like its fellow parties of the left, it believes in state intervention as an appropriate means to ensure equality outcomes. The strength of these convictions is the party’s Achilles heel, however. The party is more diverse in its composition and beliefs than the PCF or Greens, with numerous factions encompassing a broad range of beliefs and attitudes. Some of these actors have proven to be more favourable to women’s interests than others, with enduring sectors that have resisted quotas on a number of grounds, ranging from universalist principles to beliefs that it is not appropriate to replace one injustice with another. The party has also seen self-interested resistance from male activists who are not sufficiently persuaded by ideational arguments to be willing to cede their power to women. For example, disgruntled younger men within the party circulated a petition expressing their fears of being a ‘sacrificed generation’, cast aside in favour of women candidates (Mandraud, Le Monde 05 February 2006). Laurence Rossignol, until recently the National Secretary for Women’s Rights and Parity within the PS, complained that the party ‘still belongs to a

142 Parties, Gender Quotas and Candidate Selection in France

world where men hold power and are determined to preserve it for themselves’ (Mandraud, Le Monde 10 July 2005). She further argued that ‘we don’t believe that laws and regulations are sufficient. The PS also has a role of educating the public, which it is no longer doing. It is no longer engaging in the ideological battle’ (ibid). This weak ideological support for parity is compounded by the variables explored in the preceding chapters that have hampered the PS’s efforts to implement parity, such as electoral motivations and an emphasis on incumbents. Together, these factors explain why, despite being the initiators of the parity law and demonstrating support for both the ends and the means of parity, the PS has not been entirely successful in implementing the law. Its failure to prioritise parity over other imperatives means that it only fulfils two out of the three conditions required for ideological motivations to be successful. Nonetheless, the fact that the party does still support parity in principle should not be discounted. As Opello has demonstrated, its ideas have been influential in putting parity onto the agenda, and if the party has yet to achieve parity, it has still made significant progress and continues to make efforts to increase women’s representation at a faster rate than natural political renewal. This suggests that, whilst not being a high priority for the PS, parity is still taken seriously due to ideological support, and this has induced the PS to make greater efforts to implement parity than would otherwise be the case. The UDF The UDF was not renowned for having a strong ideological identity or coherence. The party was formed in 1978 as an umbrella organisation of centre-right parties representing the Non-Gaullist Moderate Right or NGMR (Knapp 2004) and comprised several different ideological strands within it, including a Christian Democrat strand (the CDS) and a liberal-radical strand (the PR). The diluted ideology of the different strands explains, firstly, how the party could integrate the Catholic CDS with the anti-clerical PR, and secondly, how the party could contemplate flirtations with different actors across the political spectrum as pragmatism dictated. The UDF mainly formed alliances with the Gaullists under their various guises, sharing government and agreeing not to field candidates against each other in certain elections. The UDF was also tempted at points by alliances with the left and, since 1998, with the far-right. This latter alliance, with certain UDF regional presidents choosing to maintain their power through an alliance with the FN, was one of the causal factors of a split in the UDF in 1998 and

The Ideological Approach 143

the withdrawal from the federation of the radicals, who renamed themselves Démocratie Libérale (DL), leaving behind a party dominated by the CDS (Knapp 2004; Sauger 2003). A further scission took place in 2002 when a large proportion of the party decided to throw their lot in with the UMP, leaving behind a small parliamentary group that retained the title of the UDF. Having lost some of its more rightwing elements, the UDF became a party oriented more towards the centre, and it progressively withdrew its support of the UMP government during the 2002–7 parliament. The final division of the party in 2007 into pragmatists who were willing to support the government (Nouveau Centre), and centrists who wished to establish a separate identity (MoDem), was in keeping with the party’s muddled history and identity. Within this lack of a coherent ideology lay a mixed pattern of support for women. The centrist government that preceded the UDF, during the presidency of Valéry Giscard d’Estaing, was notable for legalising abortion. In many respects, however, the UDF has been marked by its disinterest in women’s issues, and its liberal elements opposed measures seen as positive discrimination, preferring a more individualist approach. This was typified by their attitude towards parity, which they considered unnecessary. Their elections officer, Eric Azière, felt that the issue should not be forced.7 This was also reflected in the fact that the sole deputy to vote against the parity law in the National Assembly was Christine Boutin, a (woman) UDF deputy from the conservative wing of the party.8 The UDF could best be categorised as a party lacking in strong ideological coherence, that was rather ambivalent towards the question of women’s representation and also rather hesitant about the use of compulsion to enforce a quota. Given the party’s much reduced size and status following the schisms of 1998 and 2002, the party was more interested in ensuring its survival, conserving what few seats it had remaining and trying to rebuild its lost fortunes than it was with respecting parity. Its failure to meet the conditions for ideological support of parity was reflected in the fact that women fared badly within the UDF in 2002. The party later introduced parity into its internal hierarchy, which helped raise the levels of women selected through having a more gender-balanced selection committee, but the party’s chief priority was conserving its incumbents. This was most starkly evident in 2007, when not one of the party’s women candidates was elected.9

144 Parties, Gender Quotas and Candidate Selection in France

The UMP The UMP has only existed in its current incarnation since 2002. It is an umbrella party that was formed by Jacques Chirac to help him secure a parliamentary majority following his re-election as president in 2002. Its chief component is Chirac’s old RPR party although it also contains several smaller right-wing parties, including numerous elements that previously belonged to the UDF. As such it is something of an ideological melting pot, ranging from those fairly close to the political centre to those of a strongly right-wing persuasion. The one thing that the party seems to have lost, in terms of partisan identification, is its Gaullist roots; according to Knapp, ‘in 2002, the whole of France’s moderate Right was non-Gaullist’ (2004, p. 233). The RPR progressively shifted from a Gaullist stance to a more generic conservative position that was not accompanied by a strong ideological drive. Indeed, ‘the RPR had failed to develop distinctive and coherent policies on issues as diverse as Europe, culture, the role of women, the place of young people in society, the integration of immigrants, industrial relations, local and regional identities, and assistance to the poorest in society’ (Knapp 2004, pp. 242–3). This is of interest to women’s representation because Gaullism had differed from other ideologies associated with the political right in that it was fairly accommodating of state intervention to achieve social goals. However, the Gaullists were also very socially conservative towards women, viewing them primarily as wives and mothers and making ‘women’s rights subordinate to family policy’ (Opello 2006, p. 53). By contrast, the RPR and the UMP have had a more welcoming approach to women, but this has been accompanied by a stronger emphasis on individualism and a rejection of artificial measures to increase women’s representation. Thus the party is now in a situation where the goal of parity (women’s presence in politics) is broadly acceptable even if not actively encouraged, but the means (state enforcement) is unwelcome. Interviews conducted with RPR officials by Katherine Opello revealed that the RPR did not support quotas and parity before they became popular with the electorate in the mid-1990s because ‘positive action measures were not to “our way of thinking”.’ They further claimed that ‘women should obtain their place in politics by their merit and not by quotas and “other rules”’ (Opello 2006, p. 55). Given the ambivalence shown by the UMP towards women’s presence in politics and the general distaste for parity (indeed, it was due to pressure from dissenting RPR senators that the word ‘parity’ does not feature in the law and that the law only promises to ‘favour’ gender

The Ideological Approach 145

equality rather than to ‘guarantee’ it), it is unsurprising that parity is not a priority for the party. This has started to change slightly under the direction of Nicolas Sarkozy, who is in favour of positive discrimination and has supported measures to increase women’s role in the party, especially as the French Right have traditionally built their policies around their leadership. The UMP improved their performance in the 2007 election (although they remain one of the least feminised parties), and the numerous high-profile women appointed to government posts since 2007 suggests a further feminisation of the party. Despite these top-down improvements – which have caused some resentment amongst men who were overlooked for government posts – change at the party’s grassroots is likely to come rather more slowly. The FN The FN are a party with a strong programmatic identity, although their most salient feature is not their ideology but their leadership in the form of Jean-Marie Le Pen. The party has a number of policies associated with far-right nationalist parties, such as strong lines on immigration and law and order. For example, the party has favoured the repatriation of immigrants, and is strongly in favour of reinstating the death penalty. The party also has very traditional attitudes towards women, with policies including the offer of a maternal salary as a financial incentive for women to stay at home and raise children rather than being in paid employment. It is perhaps unsurprising, then, that the FN was opposed to parity. What is more surprising is that the FN is not adverse to the inclusion of women politicians amongst its ranks, seeing them as a means of softening the party’s hardline image and making them more appealing to the electorate. One of the leading figures within the party and its probable future leader is Le Pen’s daughter, Marine. The party has always been more popular with the male electorate though, with men being twice as likely as women to vote Frontiste. Prior to the introduction of parity, the FN only fielded 12 per cent women in the 1997 elections (Camus 1998, p. 15), suggesting that the explanation for the respect of parity in 2002 and 2007 lies very much as a pragmatic response to the institutionalist arguments laid out in Chapter 5 rather than an ideological acceptance of parity. Indeed, despite Le Pen’s claims prior to the 2002 legislative elections that women softened the party’s image (along with a claim that the FN would win just as many votes if it fielded a donkey – suggesting that people were voting for the party and its leader rather than the individual candidate), he went on to say after the elections that the

146 Parties, Gender Quotas and Candidate Selection in France

poor result was due in part to the obligation to respect parity, arguing that this had given the FN ‘a handicap of notoriety that those who created the law nonetheless managed to avoid’ (Chombeau, Le Monde 11 June 2002). The FN is therefore the best example of a party that fails to meet the criteria of ideological support for parity, and yet the FN implemented parity anyway due to its financial dependence on the first portion of state funding, confirming the theory that ideology is not useful as a negative predictor of parity’s implementation. Overview Overall, the case study of French parties corresponds fairly well to the theories mapped out early in this chapter. Only the Greens meet all three conditions for parity’s implementation (agreement with gender equality and with state intervention to achieve this, as well as considering this a priority), and they are the party that has come the closest to achieving parity and that has placed the most emphasis on this. The Socialists and the Communists both meet the first two conditions of direction of ideological support, but the strength of this support is weak, especially in the PS, who are more concerned with other factors. The extent to which this support is still influential on the overall outcome is explored in greater detail in the next chapter. Meanwhile, the UDF and UMP are both parties that are ambivalent towards gender equality whilst disagreeing with artificial measures to promote women candidates, and this is reflected in their low implementation of parity. Finally, the FN are ideologically opposed to parity despite implementing it, demonstrating that ideological explanations are ineffective in explaining parity implementation in the absence of ideological support. These findings are all in keeping with international patterns identified in the literature on gender and politics; the New Left party (the Greens) has strongest ideological support for parity, followed by the Communists and then the Socialists, while parties on the Right are less in favour of parity and the far-right party shows the least support. Conclusion This chapter has considered the role that party ideology might play in motivating parties to implement parity. The literature suggests that ideology can still be a strong indicator of party policies and behaviour despite trends amongst certain parties to prioritise electoral success over ideological purity. Even the most electoral-professional parties will still have certain ideological origins that feed into their attitudes and

The Ideological Approach 147

ideas. Many parties in France are programmatic in nature and their ideas are the main determinants of their policies. Although French parties do not map perfectly onto the model of familles spirituelles, it can be seen that they bear many similarities to their European counterparts, where leftist values have been strongly associated with approval of gender quotas. These values encompass an emphasis on equality and social justice, including gender equality, and an acceptance of the need for positive discrimination and state intervention in order to achieve equality in practice. These attitudes were developed into a theoretical model that was then measured against the values of French parties to test the usefulness of the ideological approach in explaining parity implementation. The overall picture that emerges is one where the ideological approach is of rather limited utility in explaining whether or not parties will implement parity. The approach is robust only when strict conditions are met, and in the cases of most French parties these conditions are not met. Ideology can only be useful as a strong positive predictor of parity implementation provided parties meet the three conditions of agreeing with the principle of parity (its goal), the practice of parity (using enforcement to achieve this goal) and considering parity a sufficient priority to over-ride other factors. Only one party, the Greens, meets all three conditions. In the absence of the first two criteria, ideology does not provide an adequate explanation of parity implementation. This is the case for the three parties of the right (the UDF, UMP and FN) examined in this book. In the case of meeting the first two criteria but not the third, ideology may provide a limited explanation of party behaviour. Parties that have weak support for parity may incorporate this into their electoral strategy when other factors are not dominant. The PS and PCF both come into this category. Overall, the main benefit of the ideological approach lies in explaining why parties might be willing to make sacrifices in order to implement parity, but it is of no utility when parties do not have ideological support and has only very limited explanatory power when considered in isolation from other approaches. Exploring the interplay of ideological motivations with other factors is necessary to appreciate fully its role in explaining party behaviour, and this is the subject of the next and final chapter.

7 Conclusion and the ‘Party Priorities Model’

Introduction This book has considered the various ways in which parties respond to a compulsory quota, using parity in France as a case study. This final, concluding chapter brings together the arguments made throughout the book and offers a new model for understanding the implementation of compulsory quotas. We have seen that each of the theoretical models explored in Chapters 4, 5 and 6 provides a partial but incomplete explanation of party behaviour, and they are most useful when considered collectively rather than independently of each other. The most useful insights into party choices when selecting candidates can be found in the interactions between the different approaches. To advance this argument, I develop a new theoretical model that I call the ‘party priorities model’. This model incorporates the insights of the three other approaches explored in this book and demonstrates how different parties will choose different paths of decision-making dependent on their overall priorities and goals. After laying out the theory behind the model I put it to the test by considering how well it explains the behaviour of the six French parties studied here. This model then leads to a broader discussion of the role of parties in quota implementation, and the potential insights of the model for a wider range of studies.

Overview of the key themes In this section I revisit the key themes developed throughout the book and bring them together to offer some answers and conclusions. These themes include lessons from the literature; the motivations of parties 148

Conclusion and the ‘Party Priorities Model’ 149

when introducing parity compared to their motivations in implementing it; the question of centralisation of candidate selection; and supply and demand factors when selecting candidates. Lessons from the literature Chapter 1 provided a summary of the various bodies of writing that are relevant to the study of sex and candidate selection in France. This includes works on political parties and candidate selection, the broad literature on gender and politics and the more specialist literature on parity within France. Many insights can be found from all of these areas of research, but one problem identified was that there was rather limited overlap between existing works on these different areas, even though they need to be considered collectively in order to address the question of gendered candidate selection in France. There are a few exceptions to this rule. For example, Lovenduski and Norris (1993) explicitly examine the relationship between gender and political parties, and lay some important theoretical foundations that are pursued in later works by both these authors. Opello (2006) goes one further and considers in detail the role of French parties in the development of parity. These works are very useful and have been highly influential in the development of the ideas explored here. Other works remain more isolationist and, whilst providing very useful insights, are of limited benefit to the research questions explored here except when examined more holistically and in conjunction with each other. For example, the literature on French parties has made rich contributions to the study of these parties in this book, but takes very little account of questions of gender. Meanwhile works on parity have provided a number of insights into the origins and implementation of the law on which this book has built, but pay little attention to the role of political parties as the key actors in parity’s implementation. At most, they have lamented parties’ failure to produce more women deputies without theorising in any depth why this is the case. My argument is that parity has become a significant variable that French parties must take into account when selecting their election candidates. Parity has had wide repercussions for parties in a number of areas, including candidate selection but also internal organisation, party finance, promotion within the party, the party’s image and so on. It is no longer possible to offer a credible study of French parties that does not take the impact of parity into account. Similarly, the fundamental role of political parties as the implementers of parity and also, in their legislative function, as the only actors capable of

150 Parties, Gender Quotas and Candidate Selection in France

reforming the law, means that their behaviour (and, just as importantly, an understanding of the motivations underlying this behaviour) should be central to any attempt to explain the creation and implementation of parity. This book serves to bridge the gap between studies of French parties and studies of the parity law, arguing that the two are inextricably linked. The book draws on and feeds into insights taken from the wider literature on political parties and on gender and politics. All parties are gendered, and the growing uptake and significance of gender quotas throughout the world has drawn attention to this fact and highlighted the growing need to explain the role of gender in parties, and the way in which parties are more or less able to incorporate gender quotas into wider candidate selection practices. Party motivations from quota creation to implementation Chapter 2 raised the question of whether or not party attitudes towards implementing parity were consistent with their attitudes towards the introduction of the law in the first place. Party attitudes were considered at the time when parity was first conceived, debated and then introduced as a law. The subsequent chapters have offered more detailed insights into the motivations and attitudes of party when implementing the law. Have the two been consistent? Was the level of enthusiasm for the law displayed by parties at its inception a good predictor of how parties would go on to implement the law? I argue here that it was to a certain extent, but with some reservations. There is at least some evidence to suggest that parties have been consistent in their attitudes towards parity. Indeed, it is perfectly reasonable to expect that they would be, although inconsistencies might reveal that parties were acting tactically rather than out of conviction. For the most part, parties’ enthusiasm for the creation of parity can be correlated with their willingness to implement the law. The Greens, who prided themselves on their role in getting parity on the agenda,1 have been the most enthusiastic at implementing the law. The Socialists, who had a strong wing within the party pushing the law onto the statute books but also some internal dissenters, have been moderately enthusiastic about respecting parity. By contrast, the RPR barely concealed their disagreement with parity when it was first introduced and this has been reflected in the limited efforts by the UMP to select more women. Similarly, the UDF was less than warm in its attitude to parity, with the only deputy to vote against the law belonging to the UDF and the party then being the poorest performer amongst the major parties in the 2002 and 2007 elections. This reflects the argument in Chapter 6

Conclusion and the ‘Party Priorities Model’ 151

that parties with ideological support for parity will be the most likely to be willing to make sacrifices in order to implement parity. As noted in Chapter 6, however, attitudes towards parity are not the only explanation for whether or not a party will implement the law; the oft-cited example is the FN, who were the only party that openly opposed parity at the time of its passage but still went on to implement it for financial reasons. This demonstrates how the parties who supported parity were, to a certain extent, able to force the hand of recalcitrant parties and make them follow suit. A more complex explanation is required for the RPR/UMP, who voted for the law and then showed little willingness to implement it. It was demonstrated in Chapter 2 that the RPR were generally unenthusiastic about parity, with Jacques Chirac playing a key role in steering the party towards acceptance of the law, and the RPR-led Senate being a stumbling block that resulted in a watering down of the legislation. Nonetheless, the party’s deputies in the National Assembly were unanimous in voting for the law. Their change in heart between passage and implementation may best be explained by electoral motivations. The party may have seen an electoral advantage in supporting the law, as failure to do so might alienate the female electorate and would go against public opinion, which was shown to be in favour of the law. Opello supports this argument, claiming that ‘RPR officials had electoral incentives to favour gender-based measures when they did’ (2006, p. 86). By contrast, Chapter 4 demonstrated that the UMP might have electoral incentives to field incumbents and high-profile male candidates at the expense of women. The party might thus be acting consistently in prioritising whichever course of action it perceived to be in its best electoral interests, as a rational vote-maximising party, even if this meant inconsistency in its attitude towards parity. Centralisation of the selection process Chapter 3 argued that centralisation of the candidate selection process is a necessary but not sufficient condition of parity implementation. This confirms the argument put forward by Lovenduski that ‘weak or decentralised party organisation means that party centres are less able to implement policies to promote women because they have low levels of control over their local branches and constituency organisations’ (1993, p. 13). Where French parties have the will to implement parity, as in the case of the PCF, this is limited by their inability to impose this will effectively in a top-down manner. Where parties are organised centrally but do not have the will to implement parity, as in the case of

152 Parties, Gender Quotas and Candidate Selection in France

the UDF, they do not use their central authority to impose parity. Thus centralisation is a necessary condition as without a central organising and coercing force, the collective action problem of getting the local constituencies to select enough women means that it is almost impossible to achieve parity. However, it is not a sufficient condition as centralisation is only useful to parity implementation when it is deployed to that end. Supply and demand explanations Chapter 3 also considered whether the shortage of women selected and elected in legislative elections was due to supply or demand factors, using the model advanced by Norris and Lovenduski (1995). Supplyside factors would indicate a deficiency of well-qualified women with the profile necessary to stand for national office. There was some evidence to support this, with women being under-represented in the positions of power such as mayor that traditionally act as springboards for national office in French politics. Women were also hampered by socio-economic factors, for example by being less prevalent in the professions that favour a parliamentary career, being less likely to attend the elite ENA which traditionally serves as a breeding ground for politicians, and being more likely than men to be held back by domestic obligations. Although these explanations do account to some extent for the reason why there appear to be more men than women with the required profile for national office, there was also evidence that the definition of what qualities a deputy should have were subtly gendered. Parties valorised certain attributes in their ideal candidate that did not discriminate directly in favour of men but nonetheless these attributes, such as holding local executive office or being high-profile, are more prevalent in men than in women. The factors that prevent women from building the required profile mean that the qualities sought after by parties favour male candidates. The bias against women in powerful local positions is another way in which parties may discriminate indirectly against women when selecting national candidates. There is also evidence that, even where well-qualified women candidates are available, they are passed over in favour of men or relegated to the subordinate position of suppléant. Where parties take on a promising candidate who does not have the traditional career trajectory through local office, this is more often than not a man, meaning that women do not benefit even when parties are willing to broaden their selection criteria. This all suggests that difficulties with finding sufficient women candidates may not, in fact, be a clear-cut problem of

Conclusion and the ‘Party Priorities Model’ 153

either supply or demand but rather a complex interaction of the two. Even when parties seek to recruit women, they may not find women who meet their criteria because their criteria are subtly gendered in a way that parties may not appreciate. Negative perceptions of women as politicians combined with socio-economic disadvantages serve to stifle the political careers of women from the stage of joining a political party all the way through to the attainment of national office. This is reflected, for example, in the party attitudes noted in Chapter 4 whereby parties tend to perceive male candidates, especially if they are incumbents, as more electable than women, even where there is no solid evidence to support this claim. Furthermore, problems of supply are more easily overcome when demand is high. For example, the Greens have managed to find sufficient women candidates to achieve parity even though men outnumber women within the party by two to one, thus overcoming the genuine potential for supply-side arguments. The UMP have a higher proportion of women members than the Greens and a much higher membership pool to select from, yet they have selected far fewer women than the Greens, suggesting that their failure to field enough women candidates is more a problem of demand than supply. The themes explored above all interact with the core theoretical models explored in Chapters 4 to 6. The next section of this chapter offers a more direct examination of the merits and limitations of each of the three central approaches to party behaviour before considering a synthesised model that enhances their explanatory power.

Models of party behaviour This section will consider each of the three main approaches to party behaviour (the electoral competition approach, the institutional approach and the ideological approach) in turn. The overall insights provided by each approach will be identified along with the questions that cannot be answered when each approach is examined independently of the others. An attempt will then be made to answer these questions by developing an interactive ‘party priorities model’ which incorporates the insights of all three approaches and accounts for how different parties will address the question of parity and candidate selection in different ways dependent on their overarching priorities. The electoral competition approach The electoral competition approach is based on the premise that parties are vote-maximisers who will select the electoral candidates that they

154 Parties, Gender Quotas and Candidate Selection in France

believe to be most likely to win. Parties have argued that their electoral interests are best served by fielding male candidates, especially where they are incumbents, rather than replacing established and/or familiar faces with unknown women. Chapter 4 demonstrated that male candidates do not actually incur any electoral advantage, although there was more mixed evidence to support the argument that incumbents perform better at the polls than non-incumbents. Given the additional electoral costs of deselecting an incumbent, such as loss of a local support-base and the risk of the unseated incumbent standing as a dissident, parties prefer to ‘play safe’ and prioritise male candidates. This approach has significant utility in explaining why certain parties will not implement parity where they perceive that it is not in their best electoral interests to do so. The approach has two important limitations, however. The first is that not all parties fit Kirchheimer’s ‘catch-all’ thesis (1966) and some parties are motivated more by other factors, such as ideological purity, than by winning the highest possible vote share. These parties will be more influenced by other priorities and their behaviour is not well explained by the electoral competition approach. Secondly, this approach is most powerful in its explanation of party support for candidates that are incumbents. There is a weaker electoral argument for prioritising men in seats where there is no incumbent, especially where the seat is unwinnable. Chapter 4 demonstrates that, for reasons of party finance and coalition bargaining power, every vote counts in France, even where votes do not translate into seats. This goes some way towards explaining why parties are still concerned with maximising their electoral appeal even in seats where they are not constrained by incumbents. Nonetheless, the electoral arguments for prioritising male candidates in these seats are weak. Nor can the approach account for the fact that in 2002, the PS had more incumbents than the UMP yet still succeeded in fielding significantly more women than the UMP, despite the fact that both parties can be labelled ‘catch-all’. Overall, the electoral competition approach is insightful in explaining why rational parties might not implement parity. However, it does not explain why some parties still do choose to implement parity, nor does it explain divergences between parties with similar electoral interests. The approach is also of reduced explanatory power in constituencies where parties are not reselecting an incumbent. The institutional approach The institutional approach, outlined in Chapter 5, is useful for demonstrating the rules of the game that frame the options available to polit-

Conclusion and the ‘Party Priorities Model’ 155

ical parties. The institutional environment imposes various and varying limitations on parties and dictates the choice-set that is available to each party when selecting its candidates. The institutions considered here include the electoral system, the party system, the importance of the office at stake, the parity legislation and the resources at the disposal of the different parties. Chapter 5 demonstrates that each of these can be influential in determining what options are available to parties, and their combined effect leaves each party with a particular range of possible options when selecting candidates. For some parties, such as the FN, this range is very restrictive, with the party’s financial dependence on the first portion of state finance effectively leaving it with no choice other than to implement parity or face financial disaster. Where this is the case, the institutional approach is a very powerful predictor of the course of action that a party will take. For other parties, the range of options is more varied, and here the explanatory power of the institutional approach becomes more limited. Although the approach helps to identify which options are available, it does not account for how parties will choose between these different options, nor can it explain why parties faced with a similar choice-set will reach different outcomes. The optimal choice from within a fixed choice-set is not the same for all parties. Overall, the institutional approach is useful in two ways. Firstly, it demonstrates which options are available to parties, and illustrates how some parties have a wider choice-set than others. Secondly, where the choice-set is very limited, the institutional approach is the best explanation of party behaviour. However, the approach does not explain how parties with wider choice-sets will discriminate between the available choices, and for this reason it does not have sufficient explanatory power when considered alone. The ideological approach The ideological approach is based on the principle that all parties are programmatic in nature to at least a certain extent and can be identified on the basis of an underlying ideology that is at the root of all their policy choices. This allows voters to distinguish between parties and select the one which best reflects their own ideological leanings. It also allows voters to anticipate the stances that parties might take on policy areas not covered at election time. Party behaviour and choices can be explained and predicted on the basis of their overall ideological outlook. On this premise, it should be possible to identify whether or not parties will seek to implement parity based on their ideological positions. Two aspects of ideology are considered significant; firstly,

156 Parties, Gender Quotas and Candidate Selection in France

the party’s attitude towards gender equality, and secondly, its attitude towards state intervention as the appropriate tool for achieving gender equality. Chapter 6 argues that only when parties are in favour of both these aspects can it be deemed to have ideological support for parity. Even in the event of positive ideological identification with the ends and means of parity, a further obstacle is identified, in that parties need to identify fairly strongly with these goals in order to prioritise them over other factors. In most parties this is not the case. Where ideological support for parity is positive and strong, the ideological approach has powerful explanatory power. Where support is positive but weak, the approach is more limited as other goals in the candidate selection process may be accorded higher priority. Nonetheless, the approach may explain why parties with ideological support are willing to tolerate higher costs in order to implement parity than those without ideological support. This is largely reflected in the behaviour of French parties, with parties whose values accord with parity (the parties of the left) being more likely to implement it than those that do no support it (the parties of the right), and the differences between the parties of the left can be explained by the importance accorded to parity and to ideology more generally within the party. Thus, the Greens, to whom ideology is of central importance, prioritised parity more highly than the PS, who are more oriented towards maximising their electoral appeal, but both parties were more successful in implementing parity than those parties that did not support parity. The one exception to this rule was the FN, who opposed parity but still implemented it for pragmatic reasons. Herein lies the main weakness of the ideological approach; it works best as a positive predictor of parity, but is less effective in explaining the behaviour of parties that do not agree with parity in principle as they may still implement it in practice for other reasons. Overall, the ideological approach is a powerful explanatory variable for parties that have both positive and strong ideological support for parity. The approach is still useful but less effective in explaining the behaviour of parties that support parity but only weakly, due to its interaction with other variables. The approach is least useful for explaining the behaviour of parties who do not have ideological support for parity. These parties will not be motivated by ideology to implement parity and will be more likely to see the costs of parity as outweighing the benefits, but may still be motivated to implement parity for reasons that cannot be explained by the ideological approach at all.

Conclusion and the ‘Party Priorities Model’ 157

The three approaches – an evaluation Each of the three mainstream theoretical approaches explored in this book has provided some useful insights into the competing choices and priorities that parties face when selecting their electoral candidates. Institutional variables provide a framework of options for parties to choose from, but do not account sufficiently for how parties choose between these options and why different parties favour different options. Meanwhile, the electoral and ideological approaches go some way towards providing an explanation of party priorities, but their explanatory power is incomplete when each is considered in isolation from the other two approaches. The electoral competition approach demonstrates why parties might seek to avoid deselecting incumbents in order to field more women, but does not explain why, for example, the UMP fielded fewer women than the PS in 2002 even though it had fewer incumbents. The ideological approach demonstrates why parties might be prepared to implement parity even if doing so goes against their other interests, but does not consider how ideology interacts with other priorities, nor does it account sufficiently for the behaviour of parties that are not ideologically motivated to implement parity. Each approach provides a partial explanation that is necessary but not sufficient to understand the choices made by parties when deciding whether or not to implement parity. A picture begins to emerge where each approach has the most explanatory power in extreme cases where parties fit into a particular category. The electoral competition approach works best for parties that have many male incumbents and place a high emphasis on electoral performance. The institutional approach is most powerful when explaining the behaviour of parties whose choice-set is so limited by institutional variables that they are directed towards a particular course of action. The ideological approach is best at explaining the behaviour of parties with strong positive ideological support for parity. In most instances, however, parties do not fit any of these models perfectly, and in most cases the behaviour of each party can be explained at least in part by more than one approach. A holistic picture of party behaviour can only be obtained when all three approaches are considered together. Synthesising the approaches into a more powerful explanatory model is the purpose of the next section.

The party priorities model This section introduces the party priorities model, which draws on the strengths of the above approaches and uses them to predict how

158 Parties, Gender Quotas and Candidate Selection in France

different parties will react to parity dependent on their circumstances. The resulting hybrid model provides a more complete explanation of party behaviour that takes into account the different needs and priorities of each party when selecting candidates and deciding whether or not to implement parity. The model is based on the premise that there are three overarching priorities that parties face when selecting electoral candidates, namely electoral motivations, party motivations, and ideological motivations. Electoral and ideological motivations correspond largely to the approaches explored in Chapters 4 and 6 respectively; party motivations are based on the impact that institutional factors have on shaping party choices, as discussed in Chapter 5. The model incorporates the core decisions that a party faces when deciding whether or not to implement parity, and the priorities indicate the order in which parties will make these core decisions, depending on how important each priority is to the party in question. Each party will begin with their top priority (see below for a discussion of what different parties might prioritise), and then make choices depending on whether or not parity is compatible with its needs within this category. In each of the three categories, two types of scenario are ultimately possible: either a conclusive decision is made, based on the necessity or impossibility of implementing parity; or the effect of parity is not considered significant enough to be decisive, in which case parties will move on to evaluating parity within their second area of priority. Parties will continue to move through the options until they have reached a decisive result. In the event that they do not come to a decisive result after exploring the options in all three categories, it is assumed that their position is neutral. In this scenario, they would implement parity in order to avoid an unnecessary financial cost. The level of decision may also vary. For some stages of the model, the outcome may have a blanket effect for the entire selection process; for other stages, the decision must be made on a seat-by-seat basis. The two outcomes in the model whereby parity is an imperative are both blanket decisions; they would need to be implemented as an overarching strategy for candidate selection. By contrast, decisions on whether or not a female candidate would be as electable as her male alternative, or on whether choosing one candidate over another might create dissent within the party, are both dependent on the individual circumstances within each constituency, and thus must be made on a piecemeal basis. For parties that have an overarching imperative to implement parity, the model might still be used to decide in which seats parity should be implemented.

Conclusion and the ‘Party Priorities Model’ 159

The Party Priorities Model Political Parties

Ei

Pi

Eiia

Eiiia No parity

Eiib

Eiiib Ii

Ii Pi

Piia

Pi

Piib

Piiia No parity

Ii

Parity*

Piiib Ii

Ei

Iiia

Iiiia Parity *

Iiib

Iiiib Ei

Ei

Pi

Pi

Key Ei Eii

Electoral motivations Would a female candidate be less electable than the male candidate that would otherwise be fielded (for example, the incumbent)? (a yes; b no) Eiii Does the electoral cost of implementing parity exceed the financial cost of not implementing it? (a yes; b no) Pi Party motivations Pii Can the party bear the financial penalty for not implementing parity? (a yes; b no) Piii Does the cost to the party (such as through party unity, appeasing internal factions, satisfying the party faithful, risking defections and so on) of enforcing parity exceed the financial cost of not implementing it? (a yes; b no) Ii Ideological motivations Iii Is the party pro-parity? (a yes; b no) Iiii Does ideology over-ride all other factors? (a yes; b no) * Decision has a blanket effect across all seats

Two additional things are of note. Firstly, for parties with weak but positive levels of support for parity, ideology may not be sufficient to have an overarching effect and may be overridden in seats where other variables have a strong influence, but may be decisive in seats where all the options in the other two categories have been played out and have not led to a decisive conclusion. Overall, degree and direction of ideological support for parity might influence the level of cost that parties are prepared to tolerate in order to implement parity; a higher level of cost might be supported by a party that agreed with parity in principle than a party that did not. Secondly, it should be noted that a verdict of ‘no parity’ does not equate to a party fielding no women at all. Rather, it means that in those specific seats where this is the outcome, parties will not introduce a female candidate at the expense of a male one. Given that the ‘no parity’ verdict is only applicable on a seat-by-seat basis rather than across the board, and given the expectation that at least some seats will

160 Parties, Gender Quotas and Candidate Selection in France

result in a neutral conclusion that allows parity, especially in those seats already being represented by women, it can be assumed that all parties will field at least some women. Indeed, it is assumed that in seats where the favoured candidate is a woman, the model will not be applied as there is no conflict of interest. The extent to which parity is achieved is dependent on how often a ‘no parity’ verdict is reached. How does this model work when applied to political parties? Firstly, I argue that different types of party will follow the model in different ways. Secondly, I illustrate how close the model comes to explaining the actual behaviour of French parties. When considering how different parties will progress through the model, I categorise parties into three main types based on their dominant characteristics. Parties often possess more than one of the traits contained in the categories below, and this is reflected in their secondary decisions when following the model, but the categories below indicate the dominant features of each party and hence the initial direction of their trajectory through the model. Type 1: Electoral-professional parties Electoral-professional parties are electorally motivated above all else, so will begin with Ei. Their progress through E will be determined on a seat-by-seat basis. For seats that do not provide a decisive outcome, parties will then proceed to Pi. It is assumed that electoral-professional parties will be large enough to tolerate the financial penalty for ignoring parity, and that they will therefore proceed to Pii. In all seats where the electoral and/or party cost of implementing parity is high enough to be decisive, parity will not be implemented. In seats that do not incur a significant cost in either E or P, parity will be implemented. How high a cost the party is willing to tolerate may also be determined by whether or not the party has ideological support for parity. Type 2: Vulnerable parties Parties may be vulnerable in one or both of two ways. Firstly, they may be financially vulnerable due to their small size. Parties in this category will be expected to follow a simple path of Pi → Piib (parity). Secondly, parties may suffer from weak internal unity. In this scenario, they would also be expected to start with Pi, and might be more likely than other parties to conclude with Piiia (no parity). Type 3: Parties with strong ideological foundations All parties that are defined primarily by a clear ideological stance will begin with Ii. Those that are in favour of parity should reach a

Conclusion and the ‘Party Priorities Model’ 161

decisive outcome by selecting Iiiia (parity); those that are not in favour will proceed to other categories in accordance with their secondary priorities. How would the model apply to French political parties? Let us consider the cases of each of the parties studied in turn: PCF: As an ideological party with a long history and some electoral presence, the Communist Party (PCF) could be classed under type 3, so would begin with Ii (ideology). The PCF is pro-parity but not to the extent that this over-rides all other factors. This suggests that they would therefore proceed to other categories, but prioritise parity in all scenarios where the cost of parity was not excessively high. Although the PCF is fairly stable, it is a party in decline, which means that there are usually more incumbents and former incumbents than there are safe seats available. This results in high competition for winnable seats, meaning that party motivations (P) would come into play in these seats, with Piiia (no parity) being the probable outcome. For the remainder of their seats, implementing parity should be relatively low-cost and desirable. PS: The PS is primarily an electoral-professional party, and as such would be classed as a type 1 party despite its ideological support for parity. The PS would therefore be expected to begin with Ei and work its way through E and then P. Like the Communists, the PS would be expected to implement parity wherever possible. However, more seats would be expected to reach the decisive, negative outcomes of Eiiia and Piiia (no parity) due to the greater priority accorded to electoral success, the higher number of incumbents and the internal factionalism within the party. Greens: The Greens are the best example of a party that follows the type 3 party trajectory. As a relatively young, ideological party with few incumbents, they would be expected to begin with Ii. As strong supporters of parity, they are then expected to proceed to Iiia (parity) and enforce this across the board. Where problems are encountered due to individual seats that have prioritised Eiiia or Piiia outcomes (for example due to the presence of a male incumbent), the Greens would be expected to use other seats to compensate so that parity overall is achieved. UDF: The UDF were vulnerable due to their changing status and the loss of a considerable portion of their party to the UMP. As such, they

162 Parties, Gender Quotas and Candidate Selection in France

would be classed as a type 2 party, and be expected to begin with Pi. Being a cadre party with sufficient financial resources, they would be expected to skip Pii and focus their decisions on Piii and then on E. Given their ideological indifference to parity, this would not be a motivating factor for them. It would therefore be quite easy for the party and electoral costs of implementing parity to exceed the financial penalty for ignoring it, and with no ideological support for parity to mitigate these costs, a high number of Piiia and Eiiia (no parity) scenarios would be expected. UMP: Like the PS, the UMP is an electoral-professional party that would be classed as type 1. However, it differs from the PS in two regards. Firstly, it does not have ideological support for parity, so might be more likely than the PS to consider the cost of implementing parity to outweigh the benefits. Secondly, as a recently-formed umbrella party, it is arguably more vulnerable to P factors than the PS. Its trajectory would therefore be expected to be fairly similar to that of the PS, but with a higher number of Piiia and Eiiia (no parity) outcomes. FN: The FN is a classic example of the first kind of type 2 party, namely a party that is vulnerable due to its small size and hence its financial dependence on the first portion of state funding. Hence it would be expected to proceed directly to Piib (parity). The practice of parity broadly confirms the theory (this is perhaps to be expected as the model is fairly inductive in approach; I argue that this does not deduct from its explanatory power). The PCF has come close to achieving a parity of candidates, especially in 2007, but it is not prepared to make as much room for women in the seats that matter. The PS has come less close than the PCF to parity, and problems of incumbency combined with internal divisions have made parity harder for this party to achieve. Nonetheless, the party’s ideological support for parity has been reflected in the party’s ongoing efforts to increase the proportions both of women candidates and of women elected, leading to improved results in 2007 compared to 2002. The Greens and the FN are, as predicted, the two parties closest to achieving parity, although for very different reasons. Meanwhile, the UDF has struggled to achieve parity, especially in terms of seats, as has the UMP. Indeed, even the low proportion of women elected by the UMP in 2002 exceeded expectations; many of these women were in difficult seats that would not have been won had the UMP not enjoyed such a

Conclusion and the ‘Party Priorities Model’ 163

landslide victory following the controversial presidential elections two months earlier. The UMP did perform slightly better in 2007, perhaps due to electoral pressure and pressure from certain actors within the party such as Nicolas Sarkozy and Marie-Jo Zimmerman,2 but even then, the party aimed to select 30 per cent women candidates rather than the parity requirement of 50 per cent, and still undershot its target. Although it is not possible to prove with certainty that the model is a perfect predictor of party behaviour, both the outcomes of parity and the reasons for achieving those outcomes appear congruent with the theory. Overall, the model demonstrates a number of things. Firstly, it shows how the different theories explored throughout this book have most explanatory power when one considers the interactions between them. This model illuminates these interactions. Secondly, the model demonstrates how parties prioritise between the different options available to them based on their central goals and motivations. Knowledge of a party’s key characteristics makes their behaviour easier to understand and predict. Thirdly, the model explains how the ordering of priorities translates into the order in which decisions are taken. Lower level priorities will only come into play in the event that higher order priorities do not produce a decisive outcome. This explains how and when different priorities will be significant and allows for the complexities of and differences between parties. In so doing, it provides a more comprehensive account of the decision-making processes that underlie parties’ choices when selecting electoral candidates. This section has brought together the different theoretical approaches of the book and demonstrated how they can be combined into a more powerful explanatory model. Having offered a more complete theoretical account of parity implementation, it is now possible to draw some overall conclusions.

Overall conclusions This section brings together the main arguments made throughout the book to provide more comprehensive conclusions to the book’s central research questions and illustrate how this work has contributed to knowledge in this area. The first big conclusion is that parties are the key actors in quota implementation and understanding their perspectives is vital to knowing the conditions under which a quota will be implemented. A study of candidate quotas simply cannot be complete without considering

164 Parties, Gender Quotas and Candidate Selection in France

the motivations of those charged with their implementation. Incorporating party perspectives is a significant addition to the specific literature on parity in France and the more general literature on candidate quotas and their implementation. Secondly, the decisions concerning whether or not to implement parity are complex and cannot be explained fully with any of the existing theories of party behaviour. The three main theoretical approaches in the literature on political parties serve many purposes and are useful in many ways, and they have provided numerous insights into party motivations when selecting candidates. But none of these approaches is sufficient in itself to account for all the variation between parties or to explain all of the motivations underlying party choices. Hence, the literature on political parties can be strengthened by using a more holistic approach to party organisation that synthesises the strengths of the three models explored. This leads to the third conclusion, which is that the key to understanding whether and when parties will implement parity lies in locating parity within the wider context of party beliefs, priorities and opportunities. This is illustrated in the party priorities model. The way in which parties order and prioritise their central goals is both variable from one party to the next, and predictive of the way in which parties will order and prioritise the options available to them when selecting electoral candidates. Accounting for the variation between parties at the more fundamental level of party goals is the best way to account for subsequent variation between parties at secondary levels such as policy formulation and implementation; in this case, the parity law. This supports the rational actor approach to political parties, indicating that much of their behaviour can be understood through attempts – albeit sometimes bounded by institutions – to maximise their utility on the basis of their individual goals and motivations. On a related note, the fourth conclusion is that the party priorities model has wider applicability. Firstly, it could be used to explain the implementation of other kinds of quota within France. For example, French parties have begun to show concern for other kinds of descriptive representation such as the proportionate presence of ethnic minorities, and this might one day be expressed in the form of a quota. Secondly, the model could be adapted to be applicable to any other scenario in which parties have to evaluate competing costs when selecting one type of candidate over another. For example, parties in other countries would use a similar process to decide whether or not to implement candidate quotas, even though the associated costs would

Conclusion and the ‘Party Priorities Model’ 165

be different to those specified within this particular application of the model. The weight of the variables would vary depending on the circumstances under which the model was applied, leading to an altered probability of different outcomes. Nonetheless, and crucially, the order in which decisions were taken should remain unaltered. It should therefore be possible to estimate, given sufficient knowledge of the costs associated with different outcomes and the principle characteristics of the parties involved, how these parties would react to a measure such as parity. Thirdly, the model might also have wider applicability for the literature on political parties as a rational choice model of party behaviour. By illustrating how party priorities can be used to predict paths of decision-making, the model could be used to assess how party priorities would translate into policy outcomes. There is much potential for the future theoretical development and empirical testing of this model under different scenarios. Finally, this book has demonstrated the limitations of the parity law and indicated why the law’s success has remained so limited. Parties without ideological support for the law will only implement it under two circumstances; either if the law becomes so constraining that they are forced to do so, or if public attitudes towards the law lead them to believe that it is in their electoral interests to do so. Parties with ideological support for the law would be more motivated to change the law to make it more effective; firstly because this would please their own voters if it accords with their ideology, and secondly because if the parties are likely to implement the law despite the cost of doing so then they would be likely to wish to impose the same cost on their rivals to avoid being placed at a disadvantage. Parties without ideological support for the law are unlikely to alter it to make it more constraining unless there is clear electoral pressure for them to do so. Electoral pressure did have some impact prior to the 2007 legislative elections, when Ségolène Royal was polling strongly in the presidential elections and her success was attributed partly to the political renewal that she embodied as a woman candidate. Perhaps as a result, the parity law was modified in early 2007 to reinforce its role in local executives, and to increase the penalty for non-implementation in the legislative elections by 50 per cent. As with the original law, a loophole remained – the law will not come into effect until 2012. Whether parties will find the higher penalties sufficiently costly to persuade them to implement the law completely remains to be seen, and whether a potential rise in women candidates will have any corresponding impact on the proportion of women elected is also uncertain. What is

166 Parties, Gender Quotas and Candidate Selection in France

clear is that the only way to ensure that rational parties will implement parity is if a cost-benefit analysis of implementation versus defection for each and every party shows defection (non-implementation of parity) to be the more costly option. Knowing how parties perceive the costs and benefits of parity dependent on their overall goals and priorities is the first step; persuading the parties to modify the law so that all parties are motivated to implement it properly is the next big step, and essential if France is ever to achieve a parity democracy.

Appendix: Interviews

The book draws on elite interviews conducted in Paris, April–July 2005. The aim of the interviews was to learn more about the intentions, methods and outcomes of candidate selection by French political parties. Two interviews from each party were sought: one with a party official responsible for elections and/or candidate selection, and one with a woman politician within the party, in order to ensure that both sides of the selection process were given due consideration. Where the opportunity to conduct additional interviews arose it was taken rather than wasted, even when this resulted in an uneven balance of interviews across the parties. The parties selected were those with the highest levels of parliamentary presence, namely the UMP, PS, UDF, PCF and the Greens (in addition, the FN were consulted via email). All interviews were successfully obtained except for an interview with a woman deputy from the UDF, and an interview with a party electoral official from the UMP. Given that the period of fieldwork coincided with the referendum on the European constitution, I am particularly grateful to those who gave their time to be interviewed for this book, all of whom spoke on the record. The full list of interviewees is as follows:

UMP • Marie-Jo Zimmerman, UMP deputy, head of the Observatoire de la Parité, and President of the Parliamentary Committee on Women’s Rights and Gender Equality. Former Women’s Officer for the UMP. Interviewed on 22 June 05 • Françoise de Panafieu, UMP deputy and mayor of an arrondissement in Paris. She was the UMP’s candidate for Mayor of Paris in the 2008 local elections. It is also of note that in 2002 she failed to get selected for her preferred constituency (the one for which she was also mayor), so stood against the official UMP candidate and won the seat. Interviewed on 08 July 2005

PS • Catherine Génisson, PS deputy, former head of the Observatoire de la Parité during the period when parity was introduced, member of the Parliamentary Committee on Women’s Rights and Gender Equality, and then (in 2005) National Secretary for Women within the PS. Interviewed on 06 July 2005 • Bruno le Roux, PS deputy, then National Secretary for Elections. Interviewed on 19 July 2005 • Catherine Picard, former PS deputy who won a seat in 1997 and then lost it in 2002. She stood again in 2007 but did not succeed in regaining her seat. Interviewed on 02 June 2005 167

168 Appendix: Interviews

UDF • Eric Azière, National Secretary in charge of elections. Interviewed on 30 June 2005

PCF • Michel Laurent, party election official. Interviewed on 08 November 2004 • Muguette Jacquaint, one of the few women PCF deputies. Interviewed on 13 July 2005

Greens • Jean Desessard, Senator, Party election officer until 2004 (including selections for the 2002 legislative elections). Interviewed on 29 June 2005 • Emilie Mazzacurati, then Women’s Officer for the Greens. Interviewed on 31 May 2005

Notes Chapter 1

Introduction

1 www.ipu.org, situation as of 1 January 1999. 2 The law is commonly known as the ‘parity’ law, but the word ‘parity’ was deliberately excluded from the final wording of the legislation. See Chapter 2 for further discussion. 3 Parity does not yet apply to the elections of the conseils généraux (powerful regional bodies) and to certain constituencies in local and Senatorial elections. See Chapter 5 for further details. 4 I disagree profoundly with the findings of Fréchette et al (see Murray et al 2009), but the point here is that parity has drawn in scholars from other disciplines. 5 This is discussed in greater detail in Chapter 2.

Chapter 2

French Parties and Parity in Context

1 Standard French grammar refers to mixed groups of people in the masculine plural. Feminists have recently begun to challenge this by offering a gender-neutral term that indicates the possibility that individuals within the group are female. For example, a standard reference to members of the National Assembly would be députés; a feminist version of the same word would be député-e-s, with the additional e representing female deputies. The gender-neutral term is not a grammatical requirement and is not widely used, and can therefore be seen as a conscious effort to demonstrate gender inclusiveness. 2 They were not the first party to have a woman presidential candidate, but were the first party to do so where the candidate stood a credible chance of qualifying to the second round. 3 The Elysée is the Presidential palace; Matignon is the official residence of the serving Prime Minister. 4 In 2001, prior to parity’s first application for the legislative or senatorial elections, France had the lowest number of women in Parliament (both chambers combined) than any other EU member state, with just 9.6 per cent women in the National Assembly and 5.9 per cent women in the Senate, leading to a combined total of 8.3 per cent. The figure for the National Assembly was lower than the level of women elected in 1997 (10.9 per cent) due to the replacement of female deputies with male suppléants. (Source: www.db-decision.de (website no longer available)). 5 In total, the PS fielded 27.8 per cent women in 1997 (www.observatoireparite.gouv.fr) although women only comprised 16 per cent of winning candidates (based on data collected from www.assemblee-nationale.fr). 6 Roselyne Bachelot and Gisèle Halimi, founder members of the Observatoire, speaking at the Conference to mark ten years of the Observatoire de la Parité. 169

170 Notes 7 This was known as the ‘Juppette’ scandal, referring to the slang name given to these women ministers – a patronising play on Prime Minister Juppé’s name that translates as ‘short skirts’. 8 For example, in a survey published in L’Express on 6 June 1996, 77 per cent of those surveyed said that they were in favour of modifying the constitution to introduce the principle of parity (Mossuz-Lavau 1998, p. 26). In 1997, a survey of voters revealed that 38 per cent of the electorate would prefer a female deputy, 25 per cent would prefer a male deputy, with 37 per cent undecided (Helft-Malz & Lévy 2000, pp. 71–2). 9 Elizabeth Badinter has published numerous articles providing a philosophical objection to parity. See, for example, Amar 1999, pp. 18–22 & 40–8. 10 To the constitutional amendment ‘Equality of access for women and men to electoral mandates and elective office will be promoted by statute’, the Senate tried to add the clause ‘whilst respecting the choice of representatives and the right to stand as a candidate’. This was over-ruled by the National Assembly on the grounds that it ‘negates the principle’ of parity (Elizabeth Guigou, quoted in Le Monde 05 March 1999). However, the Senate was successful in ensuring that the word ‘promoted’ was used rather than ‘determined’ or ‘guaranteed’, thus weakening its effectiveness. 11 Only one deputy, Christine Boutin of the UDF, voted against (Le Monde Interactif 26 January 2000).

Chapter 3

Candidate Selection, Supply and Demand

1 A practice similar to the all-women shortlists used by the British Labour party. 2 A practice similar to the British Conservative Party’s ‘A-list’ strategy. 3 The PCF is divided into different levels of organisation, and the smallest of these is the section. There are usually several sections per constituency. 4 Départements are administrative districts broadly equivalent to a British county. There are 96 départements in mainland France. 5 Michel Laurent, interviewed on 08 November 2004. 6 In 2002, the women elected as a percentage of total communist candidates elected was 18.2 per cent; the total percentage of women fielded in the seats that the PCF had won in 1997 was even lower, at 11.4 per cent. In 2007, the percentage of women communists elected fell to 16.7 per cent, despite the rise in the percentage of candidates. 7 Being the largest party of the left, and the initiator of the ‘gauche plurielle’ coalition of the late 1990s, the PS has alliances with numerous smaller parties of the left. 8 Eric Azière, party elections and federations officer, interviewed on 30 June 2005. 9 Eric Azière, interviewed on 30 June 2005. 10 Opinion of Françoise de Panafieu, deputy for the UMP and formerly for the RPR, interviewed on 08 July 2005. 11 The RPR was one of the founding members of the UMP (see Chapter 2 for a fuller discussion). Given the two-month time span between the UMP’s creation and the 2002 legislative elections, the candidates fielded under the

Notes 171

12 13 14

15

16

17

18

19

20 21 22 23 24 25

26 27

UMP banner were, for the most part, selected by the UMP’s component parties prior to the UMP’s creation. Bruno le Roux, National Secretary for Elections, PS. Interviewed on 19 July 2005. Interview and correspondence with Michel Laurent, National Executive member responsible for federations and regions, PCF. Although Michel Laurent (PCF) did admit that party members, when selecting candidates, might employ imputed discrimination in the belief that the electorate would be less likely to vote for a woman. Moreover, the responses given in interviews may have been subject to a social desirability bias, with respondents unlikely to be willing to acknowledge discrimination even if it did exist. This figure is based on population data from l’INSEE (2004), using women from the age range of 25–74. It is unlikely that parties would consider women younger or older than this age range, and this age range reflects the range currently present in the National Assembly. For example, Martine Aubry, current PS leader, prominent mayor and former minister and deputy, is the daughter of Jacques Delors, former minister and one of the founders of the European Union. Deputies are elected together with a substitute who replaces them in the event of their death or resignation. Given the fact that parliamentary office is incompatible with government office, all deputies who are promoted to the government have to hand their seat over to their suppléants in order to avoid a by-election. This ignores the fact that parties do not manage to attract an equal number of female members, and puts to one side the important question of why this is so, and how it could be remedied. Data on party executives for the UDF is not available following their proposal to implement a 50 per cent quota within the internal party hierarchy, which would suggest an increase to 50 per cent. A detailed discussion of the effects of parity at different elections and its implementation in the local elections of 2001 is offered in Chapter 5. A further 17 deputies – all male – have been presidents of a conseil général in the past. Interview from 08 July 2005. Figures for 2008 (source: www.observatoire-parite.gouv.fr). Please note that it was not possible to obtain an interview with an election official from the UMP, hence there is limited data available for the UMP. In 2001, women formed 33.7 per cent of candidates for entry, but only 20.8 per cent of those admitted, suggesting that something within the entry system discriminates against women (www.insee.fr, 08 March 2002), although data for 2007 suggests that this problem is being resolved. It should also be borne in mind that past performance is even more important than current performance for determining the current availability of suitable electoral candidates, as their education would normally precede their pursuit of political office by a number of years. Within the European Union, only Sweden, Finland and Denmark have a higher rate of female employment. For example, the pay-gap for managers is 24 per cent.

172 Notes 28 Emilie Mazzacurati, head of the Greens’ women’s committee, interviewed on 31 May 2005. 29 Catherine Genisson, PS National Women’s Secretary and former head of the Observatoire de la Parité, interviewed on 06 July 2005. 30 Catherine Genisson (PS) and Françoise de Panafieu (UMP).

Chapter 4

The Electoral Competition Approach

1 See Chapter 6 for an alternative perspective and a discussion of parties’ ideological motivations. 2 Original emphasis. 3 However, for some parties, ideology is still perceived to be of more importance than a transitory means to an electoral end. This is discussed further in Chapter 6. 4 This is unlikely, as voter preferences can only be measured reliably by studying actual voting behaviour, as voters are not always honest about their voting intentions when surveyed – a phenomenon clearly illustrated, for example, by significant under-reporting of intentions to vote for the National Front. 5 The French are so used to the concept of an electoral advantage going to the incumbent that they have a common phrase for it: ‘prime au sortant’, meaning ‘incumbent’s bonus’. 6 Own calculation, based on electoral data from www.assemblee-nationale.fr and www.lemonde.fr. These figures include the gain and loss of one seat by the FN. 7 Vague rose is the French colloquial term used to describe the Socialist sweep to victory in 1997. 8 Data for 1997 taken from the British Parliamentary Constituency database, available from www.pippanorris.com, and refers to seats in England and Wales (data for other seats was not available). The total percentage of seats to change hands for the whole of the UK was 4.1 per cent in 2001 and 9.6 per cent in 2005 (based on results taken from www.bbc.co.uk). 9 It should be noted that the gauche plurielle was a series of bilateral negotiations between the PS and each smaller party rather than a collective negotiation between all left-wing parties (Knapp 2004, p. 71). 10 ‘Financement des partis’, document available on www.observatoire-parite. gouv.fr, accessed 03 November 08. 11 Interview with Bruno le Roux, National Secretary for Elections, PS, 19 July 2005. 12 Interview with Eric Azière, party election officer, UDF, 30 June 2005. 13 Interview with Michel Laurent, party election official, 08 November 2004. 14 Interview with Catherine Genisson, then National Secretary for Women, PS, 06 July 2005. The PS did make some gestures towards reserving some of these seats for women in 2007, although numerous exceptions were permitted. 15 It is not possible to make meaningful comparisons beyond 1988 because a proportional electoral system was temporarily deployed for the 1986 general election, and elections prior to that date operated under different boundaries and with fewer seats.

Notes 173 16 France operates under a multi-party system; however, all electable parties, with the exception of the National Front (FN), tend to fall into one of two broad electoral coalitions representing the Left and the Right. To avoid unnecessary complexity I have divided parties for the purpose of this analysis into the three categories of Left, Right and FN. 17 This measure was achieved by measuring the swing from the first election to the second election, and then deducting the swing in each seat from the average swing for that region to produce a ‘deviation’ score. This deviation score was then deducted from the ‘deviation’ score of the preceding election to produce a dependent variable that accounted for party performance across space (the region) and time (the previous election). 18 In order to create the dependent variable (using the method described above), it is necessary to use data from two previous elections; for this reason, results can only be generated for 1997, 2002 and 2007. 19 The four different measures were incumbent (coded as 4), former incumbent (coded as 3), former candidate (anyone who had previously qualified to the second round in that seat but without winning, coded as 2) and new candidate (coded as 1). 20 The data for this and all subsequent tables was assembled using election data provided on www.assemblee-national.fr and www.lemonde.fr 21 To the extent of my knowledge. The data is based on the second round candidates at the time of the legislative elections in 1988 onwards. Some omissions may be possible due to candidates who stood in that seat prior to this date, who stood in by-elections, and/or who previously stood in that seat but did not qualify to the second round. These candidates may also have served as a suppléant (reserve candidate) to a sitting deputy. 22 In principle, the total proportion of incumbents plus inheritors should not exceed 100 per cent. In practice, due to the fact that the figures are calculated on a left-right basis rather than an individual basis, and that there are some seats in which both the winner and runner-up are of the same political coalition, the figures are slightly elevated. These figures do, however, include candidates in elections for which there was no second round. 23 Indeed, the UMP party was initially formed as an electoral coalition of right-wing parties to provide a parliamentary majority for Chirac, and the initials originally stood for ‘Union for a Presidential Majority’. 24 The success rates for former incumbents are rather arbitrary as the number of women concerned was so low (four in both 1997 and 2001). 25 This is the average swing for the 1988–2002 elections. The 2007 elections experienced an unusually low swing to the left of 1.48 per cent.

Chapter 5

The Institutional Approach

1 As explained in footnote 3 of the table, the minimum district size for proportional representation has changed twice in recent years. In 2000 it was expanded from three seats to five, but then reduced back to four in 2003. 2 Interview with Bruno le Roux, National Secretary for Elections, PS, 19 July 2005.

174 Notes 3 Under ‘zipping’, women are guaranteed at least second, fourth and sixth place on the list, if not better; under the three in six rule, women may be placed fourth, fifth and sixth on the list. 4 It should be noted that this temporary change in the electoral system was for party political reasons and not out of any wish to raise the proportion of female deputies. Rather, the PS, faced with certain electoral defeat, introduced proportional representation in order to minimise their electoral losses. The RPR reverted the electoral system back to its original, majoritarian form as one of their first acts when they came into office in 1986. 5 This practice was slightly less prevalent in the 2008 senatorial elections, but only due to the weakened position of the UMP, who could not afford to take the risk of splitting their vote. (Thanks to Paul Smith, University of Nottingham, for providing this information.) 6 A law was passed in January 2007, raising the penalty by 50 per cent, but it will not come into effect until the 2012 elections. 7 The ‘five main parties’ refers to the UMP, UDF, PS, PCF and Greens. The ‘smaller parties’ refers to all parties who fielded more than 100 candidates but won fewer than three seats. 8 See Chapter 4 for a more detailed discussion of electoral coalitions. 9 Interview with Michel Laurent, PCF election official, 08 November 2004. 10 The modification of the parity law in 2007 obliges candidates in cantonal (departmental) elections to have a suppléant of the opposite sex, but this has had only a minimal impact.

Chapter 6

The Ideological Approach

1 In the American case, both parties are actually right of the ‘centre’ as defined by European party systems, although the Republicans are clearly further to the right than the Democrats. The key thing is that the centre ground of American politics, the ideological space where the median voter lies, tend to fall in the grey area between the two parties. 2 This is explored in greater detail in Chapter 7, which explores how different party priorities interact and shape party decisions. 3 Interview with Muguette Jacquaint, PCF deputy, 13 July 2005. 4 Interview with Michel Laurent, PCF elections official, 08 November 2004. 5 See Chapters 4 and 5 for an explanation of why incumbents pose a barrier to the implementation of parity. 6 Interview with Emilie Mazzacurati, Paris, 31 May 2005. 7 Interview with Eric Azière, UDF elections officer, 30 June 2005. 8 Boutin has since joined the UMP and is now a government minister. 9 This takes into account the party’s divide; no women were elected under the Nouveau Centre or MoDem banners.

Chapter 7

Conclusion and the ‘Party Priorities Model’

1 See Chapter 6. 2 Director of the Observatoire de la Parité and leader of the parliamentary committee on women’s rights.

Bibliography Newspapers, newspaper articles and websites L’Express (French periodical) Le Monde (French daily newspaper) Le Monde Diplomatique (French periodical) Le Figaro (French daily newspaper) Aulagnon, M & J-M Bezat (10 March 1995), ‘L’idée de “quotas” de femmes dans la vie politique refait surface’, Le Monde. Bacque, R (05 March 1999), ‘La droite sénatoriale accepte la parité approuvée par Jacques Chirac’, Le Monde. Belot, L & A M Rocco (08 March 2001), ‘Les états-majors des entreprises s’ouvrent timidement aux femmes’, Le Monde. Bitton, A (16 August 2001), ‘La parité de A à Z’, L’Express. Bordenave, Y (22 June 2002), ‘Le “parti du président” sera la formation politique la mieux dotée, avec près de 30 millions d’euros par an’, Le Monde. Chombeau, C (11 June 2002), ‘Le Front national est en mesure de maintenir ses candidats dans seulement 37 circonscriptions’, Le Monde. Fabre, C (27 January 2000), ‘Les députés ont adopté le projet de loi instituant la parité des hommes et des femmes aux élections’, Le Monde. Fabre, C (27 March 2001a), ‘Les “glaneuses” du canton font une maigre récolte’, Le Monde. Fabre, C (27 March 2001b), ‘Quatre femmes administrent des villes de plus de 100 000 habitants’, Le Monde. Fabre, C (10 May 2002), ‘Pour les élections législatives, la droite a relégué la parité au second plan’, Le Monde. Fabre, C & C Garin (13 November 2001), ‘La laborieuse mise en oeuvre de la parité pour 2002’, Le Monde. Mandraud, I (05 February 2006), ‘Le malaise des jeunes hommes blancs du Parti socialiste’, Le Monde. Mandraud, I (10 July 2005), ‘Les “jeunes hommes pressés” ont fait régresser la parité dans le parti’, Le Monde. Montvalon, J B & P Roger (16 June 2005), ‘François Bayrou obtient la suspension de Gilles de Robien de la direction de l’UDF’, Le Monde. Noblecourt, M (16 December 2001), ‘Les candidates socialistes tentent de s’octroyer des circonscriptions’, Le Monde. Pisier, E (11 February 1999), ‘Contre l’enfermement des sexes’, Le Monde. Remy, J & E Mandonnet (11 February 1999), ‘Oui à l’égalité, non à la parité’, L’Express. Saux, Jean-Louis (25 November 2001), ‘Trois Questions À…Michèle AlliotMarie’, Le Monde. Seelow, S (13 July 2002), ‘L’UMP, l’UDF et le PS mauvais élèves de la parité’, Le Monde. 175

176 Bibliography www.assemblée-des-femmes.com National Assembly website: www.assemblée-nationale.fr www.bbc.co.uk www.db-decision.de www.elections2002.sciences-po.fr www.ena.fr www.idea.int www.insee.fr www.ipu.org www.lefigaro.fr www.lemonde.fr www.lexpress.fr www.lesverts.fr www.liberation.fr www.nouvelsobs.com OPFH – Observatoire de la Parité entre Femmes et Hommes: www.observatoireparite.gouv.fr www.parti-socialiste.fr www.pcf.fr www.pippanorris.com www.premier-minstre.gouv.fr www.senat.fr www.udf.org www.u-m-p.org

All other references Achin, C & M Paoletti (2002), ‘Le “salto” du stigmate. Genre et construction des listes aux municipales de 2001’, Politix 15(60): 33–54. Achin, C & S Lévêque (2006), Femmes en Politique (Paris: La Découverte). Achin, C et alii (2007), Sexes, Genre et Politique (Paris: Economica). Allwood, G and K Wadia (2000), Women and Politics in France 1958–2000 (London: Routledge). Amar, M (ed.) (1999), Le piège de la parité: arguments pour un débat (Hachette). Appleton, A & A Mazur (1993), ‘Transformation or Modernisation: The Rhetoric of Gender and Party Politics in France’, in J Lovenduski and P Norris (eds), Gender and Party Politics (London: Sage Publications). Baer, D L (1993), ‘Political Parties: The Missing Variable in Women and Politics Research’, Political Research Quarterly 46(3): 547–76. Bartolini, S (2002), ‘Electoral and Party Competition: Analytic Dimensions and Empirical Problems’, in R Gunther et al (eds), Political Parties: Old Concepts and New Challenges (Oxford: Oxford University Press). Bartolini, S & P Mair (1990), Identity, Competition, and Electoral Availability: The Stabilisation of European Electorates 1885–1985 (Cambridge: Cambridge University Press). Bataille, P & F Gaspard (1999), Comment les femmes changent la politique et pourquoi les hommes resistent (Paris: La Découverte).

Bibliography 177 Baudino, C (2003), ‘Parity Reform in France: Promises and Pitfalls’, Review of Policy Research 20(3): 385–400. Bell, D (2003), ‘The French Communist Party: From Revolution to Reform’, in J Evans, The French Party System (Manchester: Manchester University Press), pp. 29–41. Bereni, L (2004), ‘Le mouvement français pour la parité et l’Europe’, in S Jacquot & C Woll (eds), Les usages de l’Europe: acteurs et transformations européennes (Paris: L’Harmattan). Bereni, L (2006), ‘Lutter dans ou dehors du parti? L’évolution des stratégies des féministes du Parti socialiste (1971–1997)’, Politix 20(78): 187–209. Bereni, L (2007a), ‘French Feminists Renegotiate Republican Universalism: The Gender Parity Campaign’, French Politics 5: 191–209. Bereni, L (2007b), ‘De la cause à la loi. Les mobilisations pour la parité politique en France (1992–2000)’, Doctoral Thesis (l’Université Paris I Panthéon-Sorbonne). Bereni, L & E Lépinard (2004), ‘Les strategies de légitimation de la parité en France’, Revue Française de Science Politque 54(1): 71–98. Bird, K (2003), ‘Who are the Women? Where are the Women? And What Difference Can They Make? Effects of Gender Parity in French Municipal Elections’, French Politics 1(1): 5–38. Bird, K (2004), ‘The Effects of Gender Parity in Elections: The French Case’, in J Gaffney (ed.), The French Presidential and Legislative Elections of 2002 (Ashgate: Aldershot), pp. 238–59. Bird, K & M Dubesset (2003), ‘La parité en politique: espoirs et désillusions’, Regards sur l’actualité 289: 65–71. Blondel, J (1978), Political Parties (London: Wildwood House). Borella, F (1990), Les partis politiques dans la France d’aujourd’hui (5th edition) (Paris: Le Seuil). Budge, I, D Robertson & D Hearl (1987), Ideology, Strategy and Party Change: Spatial Analyses of Post-War Election Programmes in 19 Democracies (Cambridge: CUP). Cain, B, J Ferejohn & M Fiorina (1984), ‘The Constituency Service Basis of the Personal Vote for U.S. Representatives and British MPs’, American Political Science Review 78: 110–25. Campbell, R (2006), Gender and Voting Behaviour in Britain (Colchester: ECPR Press). Camus, J (1997), Le Front national. Histoire et analyses (Paris: Olivier Laurens). Camus, J (1998), Les extrémismes en Europe. Etat des lieux (Paris: CERA/Editions de l’Aube). Caul, M (1999), ‘Women’s Representation in Parliament: The Role of Political Parties’, Party Politics 5(1): 79–98. Caul, M (2001), ‘Political Parties and the Adoption of Candidate Quotas: A Cross-National Analysis’, Journal of Politics 63(4): 1214–29. Childs, S (2004), New Labour’s Women MPs: Women Representing Women (London: Routledge). Childs, S & Krook, M L (2006), ‘Gender and Politics: The State of the Art’, Politics 26(1): 18–28. Childs, S & J Withey (2006), ‘The Substantive Representation of Women: Reducing the VAT on Sanitary Products in the UK’, Parliamentary Affairs 59(1): 10–23.

178 Bibliography Chagnollaud, D (ed.) (1993), La vie politique en France (Paris: Editions du Seuil). Coates, D (1995), ‘Measuring the “Personal Vote” of Members of Congress’, Public Choice 85(3–4): 227–48. Cole, A, P Le Galès & J Levy (eds) (2005), Developments in French Politics 3 (Basingstoke: Palgrave). Collier, D (1993), ‘The Comparative Method’, in A Finifter (ed.), Political Science: The State of the Discipline II (Washington: APSA). Cowley, P & S Childs (2003), ‘Too Spineless to Rebel? New Labour’s Women MPs’, British Journal of Political Science 33: 345–65. Dahlerup, D (1988), ‘From a Small to a Large Minority: Women in Scandinavian Politics’, Scandinavian Political Studies 11(4): 257–98. Dahlerup, D (ed.) (2006), Women, Quotas and Politics (Oxford: Routledge). Dahlerup, D & L Freidenvall (2005), ‘Quotas as a “Fast Track” to Equal Representation for Women’, International Feminist Journal of Politics 7(1): 26–48. Dauphin, S & J Praud (2002), ‘Debating and Implementing Gender Parity in French Politics’, Modern & Contemporary France 10(1): 5–11. Dewavrin, N (1994), ‘Un comportement specifique des femmes en politique’, in G Halimi (ed.), Femmes: Moitié de la terre, moitié du pouvoir (Paris: Editions Gallimard), pp. 129–36. Dolan, J (1997), ‘Support for Women’s Interests in the 103rd Congress: The Distinct Impact of Congressional Women’, Women and Politics 18(4): 81–94. Downs, A (1957), An Economic Theory of Democracy (New York: Harper and Row). Duverger, M (1951) Les Partis Politiques (Paris: Seuil) (English edition (1964): Political Parties: Their Organisation and Activity in the Modern State (Cambridge: Cambridge University Press)). Duverger, M (1955), The Political Role of Women (Paris: UNESCO). Duverger, M (1972), Party Politics and Pressure Groups: A Comparative Introduction (London: Thomas Nelson and Sons). Elgie, R & Griggs, S (2000), French Politics: Debates and Controversies (London: Routledge). Elgie, R (ed.) (2000), The Changing French Political System (London: Frank Cass). Elgie, R (2003), Political Institutions in Contemporary France (Oxford: Oxford University Press). Elgood, Vinter & Williams (2002), Man Enough for the Job? A Study of Parliamentary Candidates (Research Discussion Series, Equal Opportunities Commission). Epstein, L (1967), Political Parties in Western Democracies (London: Pall Mall). Evans, J (ed.) (2003), The French Party System (Manchester: Manchester University Press). Fabre, C (ed.) (2001), Les femmes et la politique: du droit de vote à la parité (Paris: Le Monde & Librio). Flood, C & L Bell (eds) (1997), Political Ideologies in Contemporary France (London: Pinter). Franceschet, S & J Piscopo (2008), ‘Gender Quotas and Women’s Substantive Representation: Lessons from Argentina’, Politics and Gender 4(3): 393–426. Frears, J (1991), Parties and Voters in France (London: Hurst). Fréchette, G, F Maniquet & M Morelli (2008), ‘Incumbents’ Interests and Gender Quotas’, American Journal of Political Science 52(4): 891–907.

Bibliography 179 Freedman, J (2004), ‘Increasing Women’s Political Representation: The Limits of Constitutional Reform’, West European Politics 27(1): 104–23. Gallagher, M & M Marsh (1988), Candidate Selection in Comparative Perspective: The Secret Garden of Politics (London: Sage). Gaspard, F, C Servan-Schreiber & A Le Gall (1992), Au pouvoir citoyennes! Liberté, égalité, parité (Paris: Le Seuil). Gaspard, F (ed.) (1997) Les Femmes dans la prise de décision en France et en Europe (Paris: Editions l’Harmattan). Genisson, C (2002), ‘La parité entre les hommes et les femmes: une avancée décisive pour la démocratie’, report for l’Observatoire de la Parité (Documentation française). Gerring, J (1998), Party Ideologies in America, 1828–1996 (Cambridge: Cambridge University Press). Grofman, B, S Merrill, T L Brunell & W Koetzle (1999), ‘The Potential Electoral Disadvantages of a Catch-All Party: Ideological Variance among Republicans and Democrats in the 50 US States’, Party Politics 5(2): 199–210. Guigou, E (1997), Etre femme en politique (Paris: Plon). Gunther, R, J Ramon-Montero & J Linz (2002), Political Parties: Old Concepts and New Challenges (Oxford: Oxford University Press). Guyomarch, A, P A Hall, J Hayward & H Machin (eds) (2001), Developments in French Politics 2 (Hampshire: Palgrave). Haase-Dubosc, D (1999), ‘Sexual Difference and Politics in France Today’, Feminist Studies 25(1): 183–210. Haegel, F (2004), ‘The Transformation of the French Right – Institutional Imperatives and Organisational Changes’, French Politics 2(2): 185–202. Haegel, F (2005), ‘Parties and Organization’, in A Cole, P Le Galès & J Lévy (eds), Development in French Politics 3 (Basingstoke: Palgrave), pp. 18–34. Halimi, G (ed.) (1994), Femmes: Moitié de la terre, moitié du pouvoir (Paris: Editions Gallimard). Hall, P, J Hayward & H Machin (eds) (1994), Developments in French Politics (London: Macmillan Press Ltd). Hall, P & R Taylor (1996), ‘Political Science and the Three New Institutionalisms’, Political Studies, 44: 936–57. Hanley, D (2003), ‘Managing the Plural Left: Implications for the Party System’, in J Evans, The French Party System (Manchester: Manchester University Press), pp. 76–90. Helft-Malz, V & P Lévy (2000), Les femmes et la vie politique française (Paris: Que sais-je? Presses universitaires de France). Henry, N (1995), ‘Gender Parity in French Politics’, The Political Quarterly 66(3): 177–80. Hofferbert, R (1998), Parties and Democracy: Party Structure and Party Performance in Old and New Democracies (Oxford: Blackwell). Huckabee, D (2003), Re-election Rates of Incumbents (New York: Novinka Books). Inglehart, I & P Norris (1999), ‘The Developmental Theory of the Gender Gap: Women’s and Men’s Voting Behaviour in Global Perspective’, The International Political Science Review 21(4): 441–63. Ivaldi, G (2003), ‘The FN Split: Party System Change and Electoral Prospects’, in J Evans, The French Party System (Manchester: Manchester University Press), pp. 137–54. L’INSEE (2004), Femmes et hommes: regards sur la parité (Paris: Statistique publique).

180 Bibliography L’INSEE (2008), Femmes et hommes: regards sur la parité (Paris: Statistique publique). Katz, R & P Mair (1994), How Parties Organise: Change qnd Adaptation in Party Organisations in Western Democracies (London: Sage). Katz, R & P Mair (1995), ‘Changing Models of Party Organisation and Party Democracy: The Emergence of the Cartel Party’, Party Politics 1(1): 5–28. Katz R & R Kolodny (1994), ‘Party Organization as an Empty Vessel: Parties in American Politics’, in R Katz & P Mair, How Parties Organise: Change and Adaptation in Party Organisations in Western Democracies (London: Sage). Kauppi, N (1999), ‘Power or Subjection?: French Women Politicians in the European Parliament’, The European Journal of Women’s Studies 6(3): 329–40. Kenney, S (1996), ‘New Research on Gendered Political Institutions’, Political Research Quarterly 49(2): 445–66. Kenny, M (2007), ‘Gender, Institutions and Power: A Critical Review’, Politics 27(2): 91–100. King, G, R Keohane & S Verba (1994), Designing Social Inquiry (Princeton: Princeton University Press). Kirchheimer, O (1966), ‘The Transformation of West European Party Systems’, in LaPalombara & Weiner (eds), Political Parties and Political Development (Princeton University Press). Kittilson, M (2006), Challenging Parties, Changing Parliaments (Ohio State University Press). Klingemann, H, R I Hofferbert & I Budge (1994), Parties, Policies and Democracy (Boulder, Colorado: Westview Press). Knapp, A (2002), ‘France: Never a Golden Age’, in P Webb, D Farrell, and I Holliday (eds), Political Parties in Advanced Industrial Democracies (Oxford: Oxford University Press), pp. 107–50. Knapp, A (2003), ‘From the Gaullist Movement to the President’s Party’, in J Evans (ed.), The French Party System (Manchester University Press), pp. 121–36. Knapp, A (2004), Parties and the Party System in France: A Disconnected Democracy? (Hampshire: Palgrave Macmillan). Knapp, A & V Wright (2006), The Government and Politics of France (5th Edition) (London: Routledge). Krook, M L (2003), ‘Not All Quotas Are Created Equal: Trajectories of Reform to Increase Women’s Political Representation’ (paper presented at the European Consortium of Political Research Joint Sessions, Edinburgh, March–April 2003). Krook, M L (2004), ‘Gender Quotas as a Global Phenomenon: Actors and Strategies in Quota Adoption’, European Political Science 3(3): 59–65. Krook, M L (2005), ‘Embedded Reform: Legal Setbacks and Candidate Gender Quotas in France and the United Kingdom’ (paper presented at the Political Studies Association Women and Politics Conference, University of Bristol, 19 February 2005). Krook, M L (2006), ‘Reforming Representation: The Diffusion of Candidate Gender Quotas Worldwide’, Politics & Gender 2(3): 303–27. Krook, M L (2007a), ‘Candidate Gender Quotas: A Framework for Analysis’, European Journal of Political Research 46(3): 367–94.

Bibliography 181 Krook, M L (2007b), ‘National Solution or Model from Abroad? Analyzing International Influences on the Parity Movement in France’, French Politics 5(1): 3–19. Krook, M L (2008), ‘Quotas Laws for Women in Politics: Implications for Feminist Practice’, Social Politics 15(3): 345–68. Krook, M L (2009), Quotas for Women in Politics: Gender and Candidate Selection Reform Worldwide (New York: Oxford University Press). Krook, M L (forthcoming), ‘Beyond Supply and Demand – A FeministInstitutionalist Theory of Candidate Selection’, Political Research Quarterly. Latté, S (2002), ‘Cuisine et dépendance. Les logiques pratiques du recrutement politique’, Politix 15(60): 55–80. Le Bras-Chopard, A & J Mossuz-Lavau (eds) (1997), Les femmes et la politique (Paris: L’Harmattan). Lépinard, E (2004), ‘L’égalité introuvable: Stratégies de légitimation et mise en oeuvre de la parité politique en France’, Doctoral Thesis (Paris: EHESS). Lijphart, A (1971), ‘Comparative Politics and the Comparative Method’, American Political Science Review 64(3): 682–93. Lijphart, A (1975), ‘The Comparative-Cases Strategy in Comparative Research’, Comparative Political Studies 8(2): 158–77. Lijphart, A (1994), Electoral Systems and Party Systems. A Study of Twenty-seven Democracies, 1945–1990 (Oxford: Oxford University Press). Lipset, S M & S Rokkan (eds) (1967), Party Systems and Voter Alignments: CrossNational Perspectives (New York: Free Press). Lovecy, J (2000), ‘“Citoyennes à part entière?” The Constitutionalisation of Gendered Citizenship in France and the Parity Reforms of 1999–2000’, Government and Opposition 35(4): 439–62. Lovenduski, J (1993), ‘Introduction: The Dynamics of Gender and Party’, in J Lovenduski and P Norris (eds), Gender and Party Politics (Thousand Oaks: Sage), pp. 1–15. Lovenduski, J & P Norris (eds) (1993), Gender and Party Politics (London: Sage Publications). Lovenduski, J & P Norris (eds) (1996), Women in Politics (Oxford: Oxford University Press). Lovenduski, J (2005), Feminizing Politics (Cambridge: Polity). Lowndes, V (2002), ‘Institutionalism’, in D Marsh and G Stoker (eds), Theory and Methods in Political Science, 2nd edition (Basingstoke: Palgrave Macmillan). Lundell, K (2004), ‘Determinants of Candidate Selection: The Degree of Centralisation in Comparative Perspective’, Party Politics 10: 25–47. Mair, P and C Mudde (1998), ‘The Party Family and Its Study’, The Annual Review of Political Science 1: 211–99. Mansbridge, J (1999), ‘Should Blacks Represent Blacks and Women Represent Women?: A Contingent “Yes”’, The Journal of Politics 61(3): 628–54. March, J & J Olsen (1984), ‘The New Institutionalism: Organizational Factors in Political Life’, American Political Science Review 78(3): 734–49. Marks, G, C J Wilson & L Ray (2002), ‘National Political Parties and European Integration’, American Journal of Political Science 46(3): 585–94. Mateo Diaz, M (2002), ‘Do Quotas Matter? Positive Actions in the Belgian Parliament’, Res Publica 44(1): 49–72.

182 Bibliography Mateo Diaz, M (2005), Representing Women? Female Legislators in West European Parliaments (Essex: ECPR press). Matland, R (1998), ‘Enhancing Women’s Political Participation: Legislative Recruitment and Electoral Systems’, in Azza Karam (ed.), Women in Parliament: Beyond Numbers, Institute for Democracy and Electoral Assistance, Stockholm, Sweden. Matland, R (2002), ‘Enhancing Women’s Political Participation: Legislative Recruitment and Electoral Systems’, in A Karam (ed.), Women in Parliament: Beyond Numbers (2nd edition) (International IDEA, available online: http://www.idea.int/publications/wip/index.cfm). Matland, R (2006), ‘Quotas: Frequency and Effectiveness’, in Dahlerup, D (ed.), Women, Quotas and Politics (Oxford: Routledge), pp. 275–92. Matland, R & D Studlar (1996), ‘The Contagion of Women Candidates in Single Member District and Proportional Representation Systems: Canada and Norway’, Journal of Politics 58(3): 707–33. Mazur, A (1995), ‘Strong State and Symbolic Reform: The Ministère des Droits de la Femme in France’, in D McBride Stetson & A Mazur (eds), Comparative State Feminism (London: Sage Publications). McAllister, I & D Studlar (2002), ‘Electoral Systems and Women’s Representation: A Long-Term Perspective’, Representation 3(1): 3–14. Meier, P (2000), ‘The Evidence of Being Present: Guarantees of Representation and the Belgian Example’, Acta Politica 35(1): 64–85. Michels, R (1915), Political Parties: A Sociological Study of the Oligarchical Tendencies of Modern Democracy (London: Jarrold and Sons). Mossuz-Lavau, J & M Sineau (1983), Enquête sur les femmes et la politique en France (Paris: Presses Universitaires de France). Mossuz-Lavau, J (1998), Femmes/Hommes pour la parité (Paris: Presses de Sciences-Po). Mossuz-Lavau, J (2007), ‘Les femmes’, in P Perrineau (ed.), Atlas électoral: qui vote quoi, où, comment? (Paris: Presses de Sciences Po), pp. 75–8. Murray, R (2004), ‘Why Didn’t Parity Work? A Closer Examination of the 2002 Election Results’, French Politics 2(3): 347–62. Murray, R, M L Krook & K Opello (2009), ‘Elite Bias, Not Voter Bias: Gender Quotas and Candidate Performance in France’ (paper presented at the European Conference on Politics and Gender, Belfast, January 2009). Nelson, B & N Chowdhury (eds) (1994), Women and Politics Worldwide (London: Yale University Press). North, D (1990), Institutions, Institutional Change and Economic Performance (Cambridge: Cambridge University Press). Norris, P (1993), ‘Conclusions’ in J Lovenduski & P Norris (eds), Gender and Party Politics (London: Sage Publications). Norris, P (ed.) (1997), Passages to Power: Legislative Recruitment in Advanced Democracies (Cambridge: Cambridge University Press). Norris, P (1999), ‘Gender: A Gender-Generation Gap?’, in P Norris (ed.), Critical Elections: British Parties and Voters in Long-Term Perspective (London: Sage). Norris, P & J Lovenduski (1995), Political Recruitment: Gender, Race and Class in the British Parliament (Cambridge: Cambridge University Press). Norris, P, J Lovenduski & R Campbell (2004), Gender and Political Participation, report for the Electoral Commission. Norris, P, E Vallance & J Lovenduski (1992), ‘Do Candidates Make a Difference? Gender, Race, Ideology and Incumbency’, Parliamentary Affairs 45(4): 496–517.

Bibliography 183 Opello, K (2006), Gender Quotas, Parity Reform and Political Parties in France (Oxford: Lexington Books). Panebianco, A (1988), Political Parties: Organisation and Power (Cambridge: Cambridge University Press). Pemstein, D & W T Bernard (2005), ‘The Strategic Placement of Women Candidates in French Legislative Elections’ (paper presented at the annual meeting of the Midwest Political Science Association, Chicago, April 2005). Peters, B Guy (1999), Institutional Theory in Political Science (London: Continuum). Phillips, A (1991), Engendering Democracy (Cambridge: Polity Press). Phillips, A (1995), The Politics of Presence (Oxford: Oxford University Press). Pintat, C (1997), ‘Les Femmes dans les Parlements et dans les partis politiques en Europe et en Amerique du nord’, in C Fauré (ed.), Encyclopédie politique et historique des femmes (Paris: Presses Universitaires de France). Pionchon, S & G Derville (2004), Les femmes et la politique (Grenoble: Presses Universitaires de Grenoble). Quota Project, www.quotaproject.org Rokkan, S (1970), Citizens, Elections, Parties: Approaches to the Comparative Studies of the Processes of Development (New York: David Mackay Company). Roudy, Y (1995), Mais de quoi ont-ils peur? Un vent de misogynie souffle sur la politique (Paris: A. Michel). Roux, N (ed.) (2004), Rien sans elles: de la parité en politique (Nantes: l’Atlante). Rule, W (1987), ‘Electoral Systems, Contextual Factors and Women’s Opportunity for Election to Parliament in Twenty-Three Democracies’, Western Political Quarterly 34: 477–98. Rule, W & J Zimmerman (1994), Electoral Systems in Comparative Perspective: Their Impact on Women and Minorities (Connecticut: Greenwood Press). Russell, M, F Mackay & L McAllister (2002), ‘Women’s Representation in the Scottish Parliament and the National Assembly for Wales: Party Dynamics for Achieving Critical Mass’, Journal of Legislative Studies 8(2): 49–76. Sartori, G (1976), Parties and Party Systems: A Framework for Analysis (Cambridge: Cambridge University Press). Sauger, N (2003), ‘The UDF in the 1990s: The Break-up of a Party Confederation’, in J Evans, The French Party System (Manchester: Manchester University Press), pp. 107–20. Schumpeter, J (1942), Capitalism, Socialism, and Democracy (New York: Harper and Brothers). Scott, J W (2005), Parité: Sexual Difference and the Crisis of French Universalism (Chicago: Chicago University Press). Seltzer, N & V Leighton (1997), Sex as a Political Variable: Women as Candidates and Voters in US Elections (London: Lynne Rienner). Sénac-Slawinski, R (2007), L’ordre sexué. La perception des inégalités femmeshommes (Paris: Presses Universitaires de France). Sénac-Slawinski, R (2008), La Parité (Paris: Presses Universitaires de France). Shepherd-Robinson, L & J Lovenduski (2002), Women and Candidate Selection in British Political Parties (London: Equal Opportunities Commission). Shepsle, K (2006), ‘Rational Choice Institutionalism’, in R Rhodes, S Binder & B Rockmam (eds), The Oxford Handbook of Political Institutions (Oxford: Oxford University Press), pp. 23–38. Sineau, M (1988), Les femmes en politique (Paris: Editions Economica).

184 Bibliography Sineau, M & J Jenson (1995), Mitterrand et les Françaises: un rendez-vous manqué (Paris: Presse de la Fondation Nationale des Sciences Politiques). Sineau, M (2001), Profession: Femme Politique. Sexe et Pouvoir sous la Cinquième Republique (Paris: Presses de Sciences Po). Sineau, M (2004), ‘L’impossible féminisation de l’Assemblée nationale, sanctuaire du pouvoir républicain’, in N Roux (ed.), De la Parité en politique (Nantes: L’Atalante), pp. 59–73. Sineau, M (2006), ‘Paradoxes of the Gender Gap in France’, French Politics, Culture and Society, 24(1): 40–58. Sineau, M (2007), ‘Effets de genre, effets de génération? Le vote hommes/ femmes à l’élection présidentielle’, Revue française de science politique 57(3): 351–67. Sineau, M (2008), La force du nombre: femmes et démocratie présidentielle (Clamency: l’Aube). Southwell, P L & C P Smith (2007), ‘Equality of Recruitment: Gender Parity in French National Assembly Elections’, Social Science Journal 44(1): 83–90. Spoon, J J (2007), ‘Evolution of New Parties: From Electoral Outsiders to Downsian Players – Evidence from the French Greens’, French Politics 5(2): 121–43. Squires, J & M Wickham-Jones (2001), Women in Parliament: A Comparative Analysis (Research Discussion Series, Equal Opportunities Commission). Stevens, A (1996), The Government and Politics of France (second edition) (Hampshire: Macmillan). Stevens, A (2007), Women, Power and Politics (Basingstoke: Palgrave Macmillan). Strøm, K (1990), ‘A Behavioural Theory of Competitive Political Parties’, American Journal of Political Science 34(2): 565–98. Studlar, D & S Welch (1987), ‘Understanding the Iron Law of Andrarchy: Effects of Candidate Gender on Voting in Scotland’, Comparative Political Studies 20(2): 174–91. Tremblay, M (2002), ‘Les élites parlementaires françaises et la parité: sur l’evolution d’une idée’, Modern & Contemporary France 10(1): 41–57. Troupel, A (2002), ‘Pistes et matériaux: disparités dans la parité. Les stratégies de contournement de la parité dans le département des Alpes-Maritimes’, Politix 15(60): 147–68. Troupel, A (2006), ‘Disparités dans la parité: Les effets de la loi du 6 juin 2000 sur la féminisation du personnel politique local et national’, Doctoral Thesis (Université de Nice-Sophia-Antipolis). Tsebelis, G (1991), Nested Games: Rational Choice in Comparative Politics (California: University of California Press). Villalba, B & S Vieillard-Coffre (2003), ‘The Greens: From Idealism to Pragmatism (1984–2002)’, in J Evans, The French Party System (Manchester: Manchester University Press), pp. 56–75. Von Beyme, K (1985), Political Parties in Western Democracies (Hants: Gower). Ward, H (2002), ‘Rational Choice’, in D Marsh and G Stoker (eds), Theory and Methods in Political Science, 2nd edition (Basingstoke: Palgrave Macmillan). Ware, A (1996), Political Parties and Party Systems (Oxford: Oxford University Press). Ware, A (ed.) (1987), Political Parties: Electoral Change and Structural Response (Oxford: Basil Blackwell). Ware, A (2003), ‘Book Reviews’, Party Politics 9(4): 523–5. Webb, P, D Farrell & I Halliday (eds) (2002), Political Parties in Advanced Industrial Democracies (Oxford: Oxford University Press).

Bibliography 185 Welch, S & D T Studlar (1996), ‘The Opportunity Structure for Women’s Candidacies and Electability in Britain and the United States’, Political Research Quarterly 49(4): 861–74. Wright, V (1989), The Government and Politics of France, 3rd edition (London: Routledge). Ysmal, C (1989), Les partis politiques sous la Cinquième République (Paris: Montchretien). Zimmerman, M (2003a), ‘Pourquoi la parité en politique reste-elle un enjeu pour la démocratie française?’, report for l’Observatoire de la Parité (Documentation française). Zimmerman, M (2003b), ‘Elections à venir: faire vivre la parité’, report for l’Observatoire de la Parité (Documentation française). Zimmerman, M (2004), ‘Parité et modes de scrutin: conforter l’égalité’, report for l’Observatoire de la Parité (Documentation française). Zimmerman, M (2005a), ‘Effets directs et indirects de la loi du 6 juin 2000: un bilan contrasté’, report for l’Observatoire de la Parité (Documentation française). Zimmerman, M (2005b), ‘Cinq ans après la loi: parité…mais presque’, minutes of a conference organised by l’Observatoire de la Parité, available here: http:// www.observatoire-parite.gouv.fr/travaux/publications.htm Zimmerman, M (2007a), ‘Les clés de la parité’, booklet by l’Observatoire de la Parité, available here: http://www.observatoire-parite.gouv.fr/portail/minilivre_20070326. htm Zimmerman, M (2007b), ‘Elections législatives 2007: résultats des épreuves paritaires’, research note by l’Observatoire de la Parité, available here: http://www.observatoireparite.gouv.fr/travaux/doc16/legislatives_2007.pdf Zimmerman, M (2008a), ‘Elections municipales et cantonales 2008: les partis politiques résistent encore à la parité’, report for l’Observatoire de la Parité, available here: http://www.observatoire-parite.gouv.fr/espace_presse/dossiers_de_presse/ OPFH_elections_032008.pdf Zimmerman, M (2008b), ‘Elections sénatoriales: seulement 15,8% de sénatrices élues en 2008’, research note by l’Observatoire de la Parité, available here: http:// www.observatoire-parite.gouv.fr/espace_presse/communiques/pdf/OPFH_NS_ Senatoriales2008.pdf Zimmerman, M (2009), ‘Elections européennes 2009: la parité confortée’, report by l’Observatoire de la Parité, available here: http://www.observatoire-parite.gouv.fr/ espace_presse/communiques/pdf/OPFH_RPE_2009.pdf

Index alternance, 27, 35, 86 bipolar multipartism, 28, 120 bipolar quadrillism, 27

institutional motivations, 39–40, 44 104–9, 123–4, 154–5 issue ownership, 79, 130 local politics, 61–5, 87, 112, 122

candidate selection centralisation, 46–57, 73–4, 151–2 criteria, 64–7 catch-all parties, 78, 126–8, 131–2, 154 cohabitation, 34–5 Communist Party, see PCF domestic barriers, 69–70 Downs, 75, 77–9, 126–7, 130–1 education, 68 electoral motivations, 40–2, 44, 75–89, 100–3, 129–30, 153–4 electoral systems, 7, 27–8, 109–15 employment, 69 familles spirituelles, see party families FN (Front National), 9, 36–7, 55, 145–6 former incumbents, 86–7, 94–6, 121 see also incumbents ‘gauche plurielle’, 27, 83, 140 Gaullists, see UMP gender gap in voting, 41, 70–1 gender quotas arguments for and against, 3–4 constitutional ban, 8, 39–40 types of, 3 Greens, 32–3, 50–1, 55, 139–40, 161 ideological motivations, 42–4, 125–31, 134–7, 146–7, 155–6 incumbents, 51–4, 85–6, 88–96, 99–102, 120–1, 139, 154 see also former incumbents

MoDem (Mouvement Démocrat, Democratic Movement), 27–8, 34, 143 National Front, see FN (Front National) New Centre, see Nouveau Centre Non-Gaullist Right, see UDF Nouveau Centre (New Centre), 28, 34, 143 parity arguments for and against, 11–12 details, 115–19 financial penalties for nonimplementation, 9, 115–18 impact, 9 origins, 8, 39–44 Parti communiste française, see PCF Parti socialiste, see PS parties priorities, 14–15, 20, 44, 123, 136, 148, 158–63 women, 11–12 see also individual parties party families, 131–4 see also von Beyme party system, 27, 38, 119–21 PCF (Communist Party), 29–30, 49–50, 55, 128–9, 161 pragmatism, 43, 129 presidentialism, 28 proportional representation, see electoral systems PS (Socialist Party), 30–2, 51–2, 55, 124, 140–2, 161 186

Index 187 rational choice institutionalism, 105–9 rational choice theory, 8–14, 85, 106–7, 129–30 single member plurality, see electoral systems suppléants, 63 supply and demand model, 22, 57–9, 67, 71–4, 152–3 Senate, 111–13, 151 Socialist Party, see PS

UDF (Union for French Democracy), 9, 27–8, 33–4, 52–5, 124, 142–3, 161–2 UMP (Union for a Popular Movement; Union for a Presidential Majority), 9, 34–6, 54–5, 137, 144–5, 151, 162 Verts, see Greens Von Beyme, 127–8, 131–3 see also party families