Voting in Quebec Municipal Elections: A Tale of Two Cities 9781487540081

This volume offers an in-depth look at municipal voting behaviour in Montreal and Quebec City, two of Canada’s most impo

136 85 3MB

English Pages 266 [265] Year 2022

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Voting in Quebec Municipal Elections: A Tale of Two Cities
 9781487540081

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

VOTING IN QUEBEC MUNICIPAL ELECTIONS A Tale of Two Cities Edited by Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

While Quebec is well known for its provincial-level party politics and thriving nationalism, voting behaviour and electoral campaigning at the municipal level have failed to gain much attention to date. Voting in Quebec Municipal Elections seeks to transform the state of municipal elections research in Quebec through a systematic study of the 2017 Montreal and Quebec City elections. Drawing upon data from the Canadian Municipal Election Study, the authors demonstrate not only the importance of Quebec municipal politics, but the many ways that municipal elections research can inform our broader understanding of voting behaviour in the province. This volume considers the features particular to the Quebec local context, such as the importance of language and nationalism, the effects of local party labels for down-ballot races, and the role of ideology. Voting in Quebec Municipal Elections represents the largest-ever collection of work on local elections in the province’s history, making a significant contribution to our understanding of the municipal voter in Quebec. éric bélanger is a professor in the Department of Political Science at McGill University. cameron d. anderson is an associate professor in the Department of Political Science at Western University. r. michael mcgregor is an associate professor in the Department of Politics and Public Administration at Ryerson University.

This page intentionally left blank 

Voting in Quebec Municipal Elections A Tale of Two Cities

EDITED BY ÉRIC BÉLANGER, CAMERON D. ANDERSON, AND R. MICHAEL MCGREGOR

UNIVERSITY OF TORONTO PRESS Toronto Buffalo London

©     University of Toronto Press 2022 Toronto Buffalo London utorontopress.com Printed in the U.S.A. ISBN 978-1-4875-4006-7 (cloth) ISBN 978-1-4875-4007-4 (paper)

ISBN 978-1-4875-4009-8 (EPUB) ISBN 978-1-4875-4008-1 (PDF)

Library and Archives Canada Cataloguing in Publication Title: Voting in Quebec municipal elections : a tale of two cities / edited by Éric   Bélanger, Cameron D. Anderson, and R. Michael McGregor. Names: Bélanger, Éric, editor. | Anderson, Cameron D. (Cameron David),   1973–, editor. | McGregor, R. Michael, editor. Description: Includes bibliographical references and index. Identifiers: Canadiana (print) 20210351411 | Canadiana (ebook) 20210351527 |   ISBN 9781487540067 (cloth) | ISBN 9781487540074 (paper) |   ISBN 9781487540098 (EPUB) | ISBN 9781487540081 (PDF) Subjects: LCSH: Local elections – Québec (Province) – Montréal. |   LCSH: Local elections – Québec (Province) – Québec. | LCSH: Voting – Québec   (Province) – Montréal. | LCSH: Voting – Québec (Province) – Québec. |   LCSH: Montréal (Québec) – Politics and government – 21st century. |   LCSH: Québec (Québec) – Politics and government – 21st century. Classification: LCC JS1721.Q8 V68 2022 | DDC 324.9714 – dc23

We wish to acknowledge the land on which the University of Toronto Press operates. This land is the traditional territory of the Wendat, the Anishnaabeg, the Haudenosaunee, the Métis, and the Mississaugas of the Credit First Nation. This book has been published with the help of a grant from the Federation for the Humanities and Social Sciences, through the Awards to Scholarly Publications Program, using funds provided by the Social Sciences and Humanities Research Council of Canada. University of Toronto Press acknowledges the financial support of the Government of Canada, the Canada Council for the Arts, and the Ontario Arts Council, an agency of the Government of Ontario, for its publishing activities.

Contents

List of Figures  vii List of Tables  ix Acknowledgments  xi  1  Introduction  3 éric bélanger, cameron d. anderson, and r. michael mcgregor   2 Gender Affinity Voting and the 2017 Municipal Elections in Quebec City and Montreal  35 scott pruysers   3 Language and (Not) Voting: 2017 Municipal Elections in Quebec 56 cameron d. anderson and laura b. stephenson   4 Nationalism and Municipal Elections in Quebec  78 jean-françois daoust   5 Does the Left–Right Axis Matter in Municipal Elections?  94 sandra breux, jérôme couture, and anne mévellec   6 Understanding Municipal Partisanship  116 éric bélanger and laura b. stephenson   7 Who Relies on Party Labels in Municipal Elections?  135 cameron d. anderson and r. michael mcgregor

vi Contents

  8 Economic Voting in the 2017 Montreal and Quebec City Municipal Elections 157 éric bélanger and charles tessier   9 Setting the Record Straight: Identifying the Correlates of Split Ticket Voting in Montreal and Quebec City  180 charles tessier and r. michael mcgregor 10 Do Municipal Campaigns Enlighten Municipal Voters? Evidence from Montreal and Quebec City  202 randy besco and j. scott matthews Appendix: Canadian Municipal Election Study Codebook  225 Contributors  247 Index  249

Figures

1.1 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.1 4.2 4.3 4.4 6.1 7.1 7.2 7.3

Competing party systems since the 2001 municipal elections  10 FSA index of linguistic dissimilarity in Montreal  66 Mayoral turnout in Montreal by FSA diversity  67 Council turnout in Montreal by FSA diversity  68 Borough mayoral turnout in Montreal by FSA diversity  68 Mayoral turnout in Montreal by diversity perceptions  69 Council turnout in Montreal by diversity perceptions  69 Borough mayoral turnout in Montreal by diversity perceptions 70 Mayoral turnout in Montreal for anglophones vs francophones by FSA diversity  71 Mayoral turnout in Montreal for anglophones vs francophones by diversity perceptions  72 Vote choice determinants in Montreal  86 The heuristic of Coderre’s political affiliation  87 Vote choice determinants in Quebec City  88 The heuristic of Gosselin’s political affiliation  89 Mean ideological self-placement of partisans  123 “Don’t know” responses and knowledge (predicted margins)  151 “Don’t know” responses and attention (predicted margins)  151 “Don’t know” responses and partisanship (predicted margins) 152

This page intentionally left blank 

Tables

1.1 1.2 1.3 1.4 2.1 2.2 2.3

City council seats obtained in the 2017 Quebec City election  18 City council seats obtained in the 2017 Montreal election  19 Socio-demographic characteristics and voter turnout  23 Socio-demographic characteristics and vote choice  26 Voter sex and vote choice in Montreal  46 Voter sex and vote choice in Quebec City  46 Binary logistic regression results (voting for a woman in Montreal – Plante vs Coderre)  47 2.4 Binary logistic regression results (voting for a woman in Quebec City – Guérette vs other)  48 A2.1 Binary logistic regression results (vote choice in Quebec City – Gosselin vs other)  51 A2.2 Binary logistic regression results (vote choice in Quebec City – Laubeaume vs other)  51 3.1 Reported turnout by language group by office, Montreal  65 3.2 Mean voluntary group membership by language, Montreal  66 A3.1 Predicting perceptions of linguistic diversity  74 4.1 Descriptive statistics for Montreal and Quebec City  81 4.2 Municipal candidates and political parties  84 4.3 Logistic regression models: Coderre = 1; Plante = 0  85 A4.1 Multinominal logit regression models: reference = Labeaume 91 5.1 Montreal: Descriptive analysis  102 5.2 Quebec City: Descriptive analysis  103 5.3 Logistic regression Montreal vote  105 5.4 Results for Quebec City  107 6.1 Creation date of the main municipal parties competing in Montreal and Quebec City in the 2017 elections  120 6.2 Distribution of partisanship across levels of government  121

x Tables

6.3 Strength of partisanship across cities and levels of government 122 6.4 Strength of partisanship by political party  122 6.5 Impact of ideology and partisanship on vote choice in Montreal 126 6.6 Impact of ideology and partisanship on vote choice in Quebec City 126 6.7 Mean candidate evaluations by partisanship in Montreal  127 6.8 Mean candidate evaluations by partisanship in Quebec City 127 6.9 Partisanship and city council vote choice in Montreal  129 6.10 Partisanship and city council vote choice in Quebec City  129 6.11 Partisanship and borough mayor vote choice in Montreal  131 7.1 Rates of “don’t know” responses and the availability of party labels 145 7.2 Determinants of “don’t know” responses  149 8.1 Perceived direction of the economy in Montreal and Quebec City 164 8.2 Variation in the economy in Montreal and Quebec City between 2016 and 2017  164 8.3 Satisfaction with incumbent in Montreal and Quebec City  165 8.4 The economy and incumbent support in Montreal  166 8.5 The economy and incumbent support in Quebec City  167 8.6 Model-based predicted probability of voting for the incumbent, economic perceptions model  168 8.7 Pooled analysis of economic voting  169 8.8 Economic policies and incumbent support in Montreal  170 8.9 Economic policies and incumbent support in Quebec City  171 8.10 Model-based predicted probability of voting for the incumbent, policies model  172 8.11 The economy/policies at the provincial and federal levels and municipal incumbent support in Montreal  172 8.12 The economy, provincial and federal party ID, and incumbent support in Montreal  173 9.1 Split ticket voting by city and position  192 9.2 Split ticket voting rates by mayoral vote choice  192 9.3 The correlates of split ticket voting  194 10.1 Economic opinionation and accuracy  208 10.2 Candidate placements opinionation and accuracy  212 10.3 Priming effects, conventional tests  215 10.4 Priming effects, two-wave tests  218

Acknowledgments

One of the main reasons this volume exists is the existence of the Canadian Municipal Election Study (CMES). The opinion data collected through that project form the backbone of the collection. That project could not have been completed had it not been for a generous research grant received from the Social Sciences and Humanities Research Council of Canada (Insight Grant, number 435-2017-0993). A major step toward the development of the current volume was the holding of a one-day conference at McGill University in June of 2018. That event allowed the volume’s contributors to meet in person, present early versions of their respective chapters, and receive useful feedback on them. Warm thanks to Andrew Sancton for having delivered the keynote address on that occasion. We also thank Alexandre Blanchet, Adrienne Davidson, Chris Erl, Eric Guntermann, Jack Lucas, Aaron Moore, Mireille Paquet, Erin Tolley, and Florence Vallée-Dubois for their participation in this event, either as additional presenters or commentators. The organization of that conference was again made possible by SSHRC (this time in the form of a Connection Grant, number 611-20170539). Several units within McGill also contributed financially to the organizing of this conference: the Centre for the Study of Democratic Citizenship (CSDC), the Centre for Interdisciplinary Research on Montreal (CIRM), the McGill Institute for the Study of Canada (MISC), the Quebec Studies Program, the Department of Political Science, and the Faculty of Arts (through the Dean’s Development Fund). The event also benefited from the financial support of Ryerson University’s Dean of Arts, Ryerson’s Department of Politics and Public Administration, and Western University’s Associate Dean of Research in the Faculty of Social Science. We acknowledge the precious help with conference logistics that

xii Acknowledgments

we received from Wim Wolfs, Sara Vissers, Esther Armaignac, Stéphan Gervais, and Giovanni Giuga. Finally, we are extremely thankful for the meticulous and gracious work of this volume’s reviewers, and of our editor at the University of Toronto Press, Daniel Quinlan.

VOTING IN QUEBEC MUNICIPAL ELECTIONS A Tale of Two Cities

This page intentionally left blank 

1 Introduction éric bélanger , cameron d . anderson , and r . michael m c gregor

Municipal elections have consequences. Local governments in Quebec and Canada are responsible for a wide range of policies and services that directly affect voters’ daily lives. Local politicians make decisions about services and infrastructure that citizens often take for granted, including water and sewage, garbage removal, roads, fire and police ­services, land use, social services, and public transit. As the order of government “closest” to the voter, municipal politicians are also often seen as the most responsive to public demands (Olson, 1969; Ostrom et al., 1961; Fischel, 2001). On top of this, the overwhelming majority of elections in this country are fought at the municipal level. Municipalities a­ ccount for 99.6% of all Canadian governments (Quesnel, 2007; S ­ancton, 2015) and 95.8% of all politicians in Canada (Federation of Canadian ­Municipalities, 2013). Despite the importance and ubiquity of municipal elections, we know shockingly little about these contests, and our understanding of voting behaviour in local elections is notably weak. To date, Toronto (­McGregor et al., 2016) and Vancouver (Cutler & Matthews, 2005) are the only ­Canadian municipalities to have been the focus of individual-level election studies. By contrast, comparable studies have been conducted for every federal election since 1965 (see Kanji et al., 2012), and provincial election studies are increasingly common (see the Making Electoral Democracy Work project [Blais, 2010] and the Comparative Provincial Elections Project [Wesley, 2015]). Election studies have yet to be conducted in any city in Quebec – a noteworthy omission given the unique nature of politics and municipal government in that province (Bherer & Breux, 2011). We cannot properly understand voters in Quebec without understanding their behaviour toward all orders of government. In this edited volume, we seek to fundamentally transform the state of municipal elections research in Quebec by undertaking a systematic

4  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

study of the 2017 Montreal and Quebec City elections. Drawing upon data from the Canadian Municipal Election Study, we aim to demonstrate not only the importance of Quebec municipal politics but also the many ways in which municipal elections research can inform our broader understanding of voting behaviour in the province. Quebec’s municipal institutions are complex and in many ways unique in C ­ anada. Yet we still know little about various factors the affect them, such as party systems and multi-tier voting, or how language and nationalism impact electors. As Mévellec (2011: 289) bluntly put it not too long ago: “­L’analyse électorale municipale est une discipline qui reste largement à inventer au Québec.” The project at hand goes a long way toward filling this significant gap in the existing academic literature. The Study of Local Political Behaviour in Canada and Quebec Canadian scholars have often bemoaned the fact that, compared to the national and provincial levels, municipal elections and voters have received very little academic attention (Kushner et al., 1997; Cutler  & ­Matthews, 2005; McGregor et al., 2016). This aligns with broader claims that the field of urban politics more generally is a “black hole” (­Eidelman and Taylor 2010: 305). And Canada is hardly alone in this; American researchers in particular have said much the same about their own literature on local political behaviour (Oliver & Ha, 2007; Marschall et al., 2011). That said, the Canadian literature has recently been growing substantially in this regard. In this section, we discuss the state of the ­literature in Canada as a whole, but with special emphasis on the province of Quebec, and on Montreal and Quebec City in particular. Early studies of Canadian local elections suffered from a lack of individual-level (survey) data and tended to focus on candidates rather than electors. In particular, they focused on two topics: the characteristics of successful candidates (incumbency status, spending levels, etc.) and the role of gender in local electoral results (Kushner et al., 1997, Gidengil & Vengroff, 1997; Kushner et al., 2001). Gender has been an ­especially popular topic in the literature on local elections, largely ­because of claims that “local government has traditionally been considered more accessible to women” (Gidengil & Vengroff, 1997: 514). Such claims are based on arguments that local politics may be friendlier to women because elected positions tend to be part-time and do not involve relocating (the assumption being that women would look upon these aspects more favourably than men). The non-partisan nature of local politics, and the relative lack of barriers to entry, is also thought

Introduction 5

to encourage women to participate. More recent research, however, has found that women remain underrepresented at the local level (Tolley, 2011). Not surprisingly, gender and local candidacy continues to be an important topic in the literature on local elections (Breux et al., 2019; Davidson et al., 2020). Like women, visible minorities have also long been politically marginalized, and their political candidacies have also received more attention recently at the local level (Bird et al., 2016; Spicer et al., 2017; McGregor et al., 2017). Though the role of race in local elections has yet to receive the consideration that gender has, there is clear evidence that visible minorities face different and often greater challenges than do women at the local level. For instance, there is evidence that racial affinity effects are stronger than gender affinity effects (McGregor et al., 2017) and that female candidates are much more likely to win elections than are visible minorities (Spicer et al., 2017; Lucas et al. 2021). Though local candidates and candidacies continue to receive attention from scholars (Breux et al., 2019; Davidson et al., 2020; Lucas, 2021), the balance of research in the field has shifted recently toward the study of voters. The increasing availability of survey data means that studies of how individual electors view and behave toward local politics can be undertaken. ­Vancouver was the first city where a detailed analysis of the correlates of vote choice was carried out (Cutler & Matthews, 2005), and the vote choices of T ­ orontonians have received a great deal of attention (­Schatten et al., 2015; Bird et al., 2016; McGregor et al., 2016; Silver et al., 2020; Kiss et al., 2020; M ­ cGregor et al., 2021). Our understanding of the socio-demographic and attitudinal drivers of local vote choices is steadily advancing. The increased availability of survey data also means that the correlates of l­ocal voter turnout are becoming better understood (Breux et al., 2017; McGregor, 2018; McGregor & Lucas, 2019). This latter trend is important because local turnout rates tend to be significantly lower than those seen in provincial and federal elections, and also because the factors that drive local turnout are not always the same as those we find when examining participation patterns at the other levels of government. In brief, scholarship in the field of Canadian local elections is growing quickly in both breadth and depth. Even so, the vast majority of ­Canadian cities have yet to be subjected to the in-depth analysis required to begin to satisfactorily explain local voting behaviour and election outcomes. While our understanding of local elections has increased dramatically as of late, there is much left to be learned about elections in most Canadian cities.

6  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

In many ways, the study of municipal elections in Quebec is relatively advanced when compared to other parts of the country. Still, municipal voting behaviour research on Quebec has remained limited due to the quality and quantity of available data. Past work explaining local election outcomes has been based largely on qualitative approaches (e.g., Breux and Bherer, 2011; Chiasson & Mévellec, 2014). Quantitative work has largely been at the aggregate level (e.g., Bourassa, 1971; Drouilly, 1991, 2010; Couture et al., 2014), though it is difficult to draw conclusions from aggregate data about the attitudes and behaviour of individual electors (Robinson, 1950; Kramer, 1983). The Canadian Municipal Election Study thus enables us to make significant strides in this neglected area of research. So far, scholars of Quebec municipal elections have focused most of their attention on low levels of political participation (e.g., Collin, 2011; Dostie-Goulet et al., 2012; Couture et al., 2014), the socio-demographic profiles of candidates (e.g., Simard, 2004; Tremblay & Mévellec, 2013), the incumbency effect (e.g., Collin, 2011), and the nature and character of municipal party politics (e.g., Quesnel & Belley, 1991; Belley, 1992; Belley & Lavigne, 2008; Mévellec & Tremblay, 2013; Mévellec, 2014). The research presented in this volume both widens the number of questions examined and deepens our understanding of municipal voting ­behaviour in Quebec. In particular, several unique (or nearly unique) features of municipal politics in Quebec are explored by our contributors, including local political parties, language, nationalism, and the presence of multi-level structures (council and borough elections) within city institutions. The first factor receiving attention is the presence of political parties. In Canada, most municipal elections are non-partisan. But in Quebec, well-established parties contest elections against quasi-parties built around mayoral candidates, but also against non-partisan candidates (see ­Mévellec & Tremblay, 2013; Belley et al., 2014). At the provincial and federal levels, parties structure political competition, and partisanship (the attachment some electors feel toward a specific party) strongly affects vote choice. Does this hold at the municipal level when parties are present, and does the type of party matter? Chapters 6, 7, and 8 in this volume show that the presence of municipal parties can indeed affect Quebec voters’ calculus. Second, how does a city’s linguistic composition affect municipal elections? Montreal has large anglophone and allophone populations, whereas Quebec City is overwhelmingly francophone. Past research has found that in largely bilingual Ottawa, wards with a high percentage of French-speakers tend to elect French councillors (Cardinal & Mévellec,

Introduction 7

2012). Does this phenomenon hold in other environments like M ­ ontreal and Quebec City? Until now, only aggregate-level studies have examined the linguistic cleavage in Montreal election outcomes (Drouilly, 1991, 2010). We might expect linguistic heterogeneity to affect voters’ ­decisions to turn out and to choose one candidate over others and that these effects overlap to some extent with nationalist attitudes. Drouilly (2010: 412) hints strongly at this possibility but lacks the appropriate individual-level data to test it rigorously. The impact of voters’ opinions about ­Quebec’s place within Canada on their behaviour in municipal elections is a third Quebec-specific factor that remains under-studied. Chapters 3 and 4 examine the phenomena of language and nationalism in greater depth. A fourth factor of note in these elections is the presence of multi-level structures. The literature on multi-level governance in Canada is extensive (see Anderson, 2008; Young & Leuprecht, 2004; Young & Horak, 2012). To this point, however, it has focused exclusively on federal–provincial/ territorial–municipal relationships. This overlooks the multi-level structure within some municipal governments, many of which have sub- or supra-city bodies of elected representatives. Many cities in Quebec have borough councils, which include members who sit on city councils as well as those who sit on borough councils only. Scholars have debated how local governments should be structured (Kushner & Siegel, 2003; Sancton, 2000; Quesnel, 2001, 2005), but meanwhile, we know almost nothing about vote choice in sub- or supra-city elections. Are electors aware of these multi-level features of local government, and how do such structures affect behaviour? Also, municipalities are the only level of government in Canada in which simultaneous elections are held (council, mayoral, borough, and, in some provinces, school board). How do ­Quebec voters reason about and behave toward municipal elections at other levels than the mayoral race at the “top” of the ticket? These ­research questions are taken up in chapters 7 and 9. We have identified four distinct features of Quebec municipal elections that are studied in this volume. However, local elections in general are novel in many other ways that are worthy of examination. For instance, female candidates in municipal elections face a significant number of barriers to success (Tolley, 2011; Bird et al., 2016). But do they benefit from disproportional levels of support among female voters? Chapter 2 investigates whether gender affinity effects were present in the 2017 Quebec local elections. Also, municipal politics has long been described as almost apolitical, in the sense that ideological considerations are generally thought to be absent from political parties’ campaigns and voters’ decisions, owing to the administrative nature of municipal issues

8  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

and governance (Bherer, 2006; Belley & Lavigne, 2008). Yet the winner of the 2017 Montreal municipal elections, Projet Montréal, clearly embraces a left-leaning ideology and agenda. Is the apparent emergence of an ideological cleavage in Montreal mirrored in Quebec City? Do the 2017 elections signal the end of “apolitisme” at the local level, at least in Montreal? Chapter 5 tackles these questions in depth. Also worthy of attention are economic conditions, which vary substantially between cities. We know that local tax policy affects incumbents’ success rates (Vermeir  & ­Heyndels, 2006), but do differences in local economies have similar effects? Chapter 8 assesses the degree of economic voting in the Montreal and Quebec City local elections. Finally, do municipal electoral campaigns have an effect on voters? Do they help them d ­ ecide which candidates to support? Given the low-information environment of ­municipal elections (Cutler & Matthews, 2005; Matson & Fine, 2006), these are important questions to ask, and chapter 10 takes up this task. All of these factors (gender, ideology, the economy, and campaign e­ ffects) have been well studied at the federal and provincial levels in ­Canada but have yet to be considered vis-à-vis Quebec local elections. The main goal of this edited volume, then, is to boost significantly the level of attention devoted to municipal elections research in Quebec. In bringing together twelve experts in municipal politics and voting behaviour from institutions in Quebec and other parts of Canada, we believe that this project will increase awareness among scholars of the value and importance of the study of local political behaviour in general, and the Quebec case in particular. Though this volume focuses on two cities in Quebec, it represents a significant advance in our more general understanding of municipal electoral behaviour in Canada. The Canadian Municipal Election Study The data employed throughout this volume are from the Canadian ­Municipal Election Study (CMES). The study includes survey data from eligible voters in eight Canadian municipal elections held in 2017 and 2018. We focus here on two cases: Montreal and Quebec City. The CMES surveys are largely modelled on comparable federal and provincial ­election studies such as the Canadian Election Study (note that the complete pre- and post-election questionnaires are provided in the appendix at the end of this book). Surveys were conducted online and included many questions similar to those used for studies of other levels of government, but also included some questions specific to municipal politics in Montreal and Quebec City. The survey consisted of two waves: the pre-election questionnaire was administered in the weeks immediately

Introduction 9

preceding Election Day (5 November 2017) and achieved about 1,900 completes in each city.1 The post-election survey was fielded in the weeks after the municipal elections.2 In Quebec City, 1,313 respondents completed the post-election survey; in Montreal, 957. The CMES was one of the largest election studies ever conducted in Canada, and to the best of our knowledge, it is the largest ever individual-level study of local elections in the world (containing data from over 14,000 respondents). The survey methodology employed by the CMES aimed to combine the strengths of telephone and online survey methods. For reasons of ­ efficiency and cost, all questionnaires were administered online. Respondents were recruited in two ways. Most (roughly 75%) were ­ ­recruited via random digit dialling (RDD). Potential interviewees who passed a screener question and agreed to participate were then provided with a link to complete our survey online. This approach was employed to ensure that our sampling of municipal populations was as random, and thus as representative, as possible. An additional pool of respondents from an existing online panel was used to supplement the sample recruited via RDD (given that RDD recruitment presents a strain in terms of both cost and time).3 Forum Research Incorporated was tasked with recruiting respondents and administering the surveys. Local Political Parties in Quebec As we underscored earlier, the local political scene in several Quebec cities is characterized by the presence of political parties. This is the case in ­Montreal and Quebec City, the two municipalities examined in this volume. Because the reader may not be familiar with the parties competing in these two cities, we next provide an overview of the history of municipal political parties since 2001 in Montreal and Quebec City, along with some brief details about their organization and level of activity. The local party systems in place in both Montreal and Quebec City at the time of the 2017 elections were developed relatively recently. Figure 1.1 illustrates the evolution of these party systems in the early twenty-first century. Since the 2001 amalgamations, the competition in Montreal had pitted two parties against each other: Vision Montréal, founded in 1994 by former mayor Pierre Bourque, and Union Montréal, founded in 2001 by Gérald Tremblay to support his candidacy in the

1 The pre-election questionnaire was fielded from 20 October to 4 November. 2 The post-election questionnaire was fielded from 6 to 29 November. 3 Panel recruitment was based on appropriate quotas for age and gender.

10  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor Figure 1.1.  Competing party systems since the 2001 municipal elections a) Montreal Party System Vision Montréal Union Montréal Projet Montréal Équipe Denis Coderre Coalition Montréal Vrai changement pour Montréal 2001

2004

2007

2010

2013

2016

2004

2007

2010

2013

2016

b) Quebec City Party System Action civique de Québec Renouveau municipal de Québec Vision Québec Défi vert de Québec Équipe Labeaume Démocratie Québec Alliance citoyenne de Québec Québec 21 Option Capitale-Nationale 2001

first post-amalgamation elections. The 2001 and 2005 elections both saw Tremblay and ­Bourque running against each other, with the former winning both times. In the 2005 elections, a third notable candidate entered the fray, Richard Bergeron of the newly formed left-leaning alternative Projet Montréal. Tremblay and Bergeron ran again in the 2009 elections, together with Vision Montréal’s new leader, Louise Harel. Tremblay won again, only to resign three years later under the burden of allegations of illegal party financing involving the Mafia. This corruption crisis led

Introduction 11

to the collapse and dissolution of Union Montréal; together with the early withdrawal of Louise Harel from the race, this paved the way for an almost complete reconfiguration of Montreal’s party competition prior to the 2013 elections. Indeed, the 2013 and 2017 elections both saw three new political parties competing with Projet Montréal: Équipe Denis Coderre,4 Coalition Montréal (led in 2013 by Marcel Côté), and Vrai Changement pour Montréal (led in 2013 by Mélanie Joly). In Quebec City, the first two post-amalgamation elections saw a confrontation between candidates from the Renouveau Municipal de Québec (RMQ) on one hand and Andrée P. Boucher (former mayor of Sainte-Foy) on the other. Boucher initially ran under the banner of the Action Civique de Québec in 2001 before running as an independent candidate in the 2005 elections. L’Allier won the 2001 elections as the RMQ candidate but did not seek re-election at the end of his mandate (he had in fact been mayor of Quebec City since 1989). Boucher managed to win the 2005 elections; however, she passed away suddenly two years later. Her death led to a mayoral by-election, held in December 2007. That is when Régis Labeaume emerged as an independent candidate and became the new mayor of Quebec City, with 59% of the vote. In 2008 he founded his own party, Équipe Labeaume; during the mayoral election of 2009, he increased his vote share to 80%. Anticipating Labeaume’s landslide, the RMQ chose not to run a mayoral candidate (the party would dissolve in 2010). The incumbent mayor ended up facing one noteworthy independent candidate (controversial radio host Jeff Fillion) as well as a candidate from a new pro-environment party (Défi Vert de Québec). The 2013 ­elections saw a candidate from a brand-new party, Démocratie Québec, facing Labeaume, but the latter was easily re-elected. Three other new parties joined the mayoral race in the 2017 elections: Québec 21, O ­ ption Capitale-Nationale,5 and the Alliance Citoyenne de Québec (in the 2013 elections, the latter had presented district candidates but no mayoral candidate). Local political parties do not operate with the same scale of internal activity as other, upper-level party formations. In terms of membership, the number of members often remains undisclosed, which suggests that local parties do not have many of them. Among Montreal and Quebec

4 Note that following the defeat and departure of Denis Coderre, the party changed its name to Ensemble Montréal. 5 That party has since changed its name to Transition Québec.

12  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

City parties that do reveal their membership numbers, we find that ­Projet Montréal may be an exception; it had more than 5,000 members at the time of the 2017 elections (Gerbet & Schué, 2020). Québec 21 and Option Capitale-Nationale have had between 200 and 300 members, ­according to 2019 numbers (Porter, 2019; Ricard-Châtelain, 2020). As for Équipe Labeaume, the party does not have a formal membership. In local party life, fundraising remains modest.6 In 2017, none of the parties in Quebec City had revenues above $100,000. The most affluent party that year was Équipe Labeaume with $85,300 in funding, about twice as much as its two closest competitors ($44,200 for Démocratie Québec and $32,428 for Québec 21). In Montreal, the two main p ­ arties raised quite a bit more than that. Équipe Denis Coderre declared $441,285 in funding in 2017, with Projet Montréal not too far behind at $372,273. Fundraising events are only one aspect of municipal party ­activity between elections. Other activities include rallies, conferences, and meetings ­between citizens and elected party officials. That said, ­comparing levels of party activity online provides an idea of the unevenness found across local parties in terms of political engagement between elections.7 ­Indeed, while all parties have a presence online, some of them are much more active than others. In Montreal, the two main parties, Équipe Denis Coderre and Projet Montréal, make extensive (almost daily) use of social media platforms like Facebook and Twitter. In Quebec City, only Québec 21 displays a similar level of online activity; the other parties are relatively quiet in comparison. Another particularity of municipal political parties in Quebec is that they are not affiliated with any upper-level parties. Indeed, no organizational links exist between local party formations and those competing in provincial or federal elections. This is in part due to the ­aforementioned “apolitical” character of municipal party politics. The only traces of crosslevel party links are related to party membership and to candidacies. First, a candidate or elected municipal official may be a member in good standing of an upper-level party, although this partisan affiliation is rarely discussed openly on the municipal scene (Breux, 2019: 78). P ­ erhaps an extreme example of this situation is Denis Coderre, who at one time had been federal minister in a Liberal government. Coderre’s federal past

6 Party funding numbers come from the reports filed with the DGEQ, which can be accessed online: https://www.electionsquebec.qc.ca/francais/municipal/financement -et-depenses-electorales/sommaire-rapports-financiers.php. 7 We rely here on data collected by Christopher Erl (McGill PhD candidate), whom we thank for granting us access to it.

Introduction 13

was never really brought forward during electoral campaigns, nor was it during his tenure as mayor; even so, electors made that cross-level connection when thinking of him, as chapter 4’s analysis reveals. S ­ econd, while the few ideologically proximal parties may help one another, they do so only very informally. For instance, the members of a party at one level may lend a hand to the campaign of an ideologically close party at another level, or even run as candidates for it. This was seen with Projet Montréal, the provincial Québec Solidaire party, and the federal New Democratic Party, which all share the same progressive base (Jetté, 2019; Lévesque, 2019). The 2017 Elections in Focus All local election campaigns ran simultaneously across Quebec from 22 September to 5 November 2017. We focus in this volume on the two largest municipalities in the province, Montreal and Quebec City. The election in Montreal included two major mayoral candidates. The first was incumbent mayor Denis Coderre at the head of a party bearing his name, Équipe Denis Coderre. The second was Valérie Plante, the newly selected leader of Projet Montréal. A third candidate, Jean Fortier of C ­ oalition Montréal, dropped out of the mayoral race in mid-campaign (on 18 October) and called for his supporters to rally behind Plante.8 Five other independent candidates also competed in the Montreal mayoral campaign, but they collectively received less than 2% of the vote.9 In the eyes of Montrealers, the two main candidates seemed to represent significantly different political projects and ideologies. The CMES asked respondents to locate each candidate on a general left/right ideological scale, with 0 meaning left, 10 meaning right, and 5 representing the centre. Valérie Plante (with a mean placement of 4.27 on the scale) was clearly to the left of Denis Coderre (mean of 5.95).10 Montreal voters clearly had two ideologically distinct mayoral candidates to choose between. In Quebec City, three main candidates took part in the mayoral race. Régis Labeaume, under the banner of Équipe Labeaume, sought a fourth consecutive mandate as mayor. He faced several challengers, the

8 Fortier's decision to abandon the race was made too late in the campaign for his name to be removed from the ballots. He ended up receiving 1.26% of the vote on Election Day. 9 They were Bernard Gurberg, Fabrice Ntompa Ilunga, Tyler Lemco, Philippe Tessier, and Gilbert Thibodeau. 10 This difference is significant at p < 0.01, N = 1,411.

14  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

two most prominent being Jean-François Gosselin (the candidate for Québec 21) and Anne Guérette (the Démocratie Québec candidate). Three other candidates took part in the mayoral race but garnered only 2.5% of the popular vote combined.11 As was the case in Montreal, CMES data suggest that the three main candidates in Quebec City were associated with very different positions on the left/right ideological spectrum. Quebecers located Labeaume toward the centre on the 0 to 10 scale (mean placement of 5.59) and the two other main candidates on either ideological side of the incumbent. Guérette was perceived as left-leaning (mean of 3.87) and Gosselin as right-leaning (mean of 6.98).12 In total, there are 65 elected officials on Montreal’s City Council: 46 city councillors, 18 borough mayors, and the mayor of Montreal (who also serves as mayor of the Ville-Marie borough). In all, 35 incumbent city councillors ran for re-election in 2017, and among the 18 borough mayors, 16 were on the ballot.13 In other words, nearly 80% of incumbents appeared on the ballot in 2017 (76% of incumbent city councillors and 89% of incumbent borough mayors). The portrait for Quebec City council is easier to summarize. First, there are no borough mayors in Quebec City. Second, there are 22 representatives on the City C ­ ouncil – the mayor of Quebec City and 21 city councillors – effectively making the Quebec City council one third the size of the Montreal’s. What the two cities do have in common, however, is that the vast majority of i­ ncumbents ran for re-election in 2017. Of the 21 incumbent councillors in Quebec City, 15 were also on the ballot in 2017. The Campaigns In Montreal, a public opinion poll published at the beginning of the campaign suggested that Denis Coderre enjoyed a substantial 15-point lead over Valérie Plante, confirming another poll result from earlier in the year.14 The mayor started the campaign with confidence, mostly putting forward his record of achievements (especially in terms of boosting Montreal’s economy) over the course of his tenure. Plante, who needed

11 They were Nicolas Lavigne-Lefebvre (Option Capitale-Nationale), Daniel Brisson (­Alliance citoyenne de Québec), and Claude Gagnon (independent). 12 Differences all significant at p < 0.01, N = 1,309. 13 In addition, there are 38 borough councillors in Montreal who need to be elected. However, because these councillors do not sit on City Council, they were not ­included in the CMES study and will not be considered in this volume. 14 Mainstreet poll conducted 20–22 September and Léger poll conducted 14–19 June, respectively. Source: Qc125 website (http://blog.qc125.com/p/montreal-2017.html).

Introduction 15

to make herself better-known to voters, spent the early weeks of the campaign making a number of attention-grabbing pledges. The most noteworthy of these was her promise on 10 October to build a brand-new metro line called the Pink Line. This pledge was pivotal in positioning herself as the “mobility” candidate. Her intent was to distinguish herself clearly from Coderre and remind electors of the incumbent mayor’s lacklustre record on public transport in Montreal, including the notoriously jammed metro lines at peak rush hours and the high number of orange cones signalling roadwork throughout the city. Coderre, for his part, supported the idea (already floating for several years) of extending the existing Blue Line to Anjou in the east of the Island of Montreal. The two candidates argued over how to compensate small businesses for roadwork hassles, with Coderre pleading for some level of centralization in terms of urban development and criticizing some of the boroughs’ decisions – most notably pointing at the Plateau, whose mayor (Luc ­Ferrandez) was affiliated with Projet Montréal. While public transit and the economy proved to be important issues in this campaign and came to define the candidates and their platforms, there was another factor that hampered Coderre’s campaign – the growing impression that he was not being transparent enough to the public. The issue that perhaps hurt him the most in that regard was the Formula E. The incumbent mayor had made much of the coming to town of this electric car-racing event in July 2017. However, in mid-October, Plante publicly raised concerns over the actual economic success of this costly event, citing rumours of a massive number of unsold tickets. Coderre ­refused to disclose the number of Formula E tickets sold. On 1 ­November, the event’s organizers finally revealed that as many as 20,000 tickets had been distributed free of charge and that the remaining 25,000 had been either sold to the public or made available to the event’s partners and sponsors. In reply, Coderre simply said he regretted not having disclosed these numbers himself sooner. On 23 October, Radio-Canada had made public several statements from journalists and public officials to the effect that Coderre had been trying to control information on a regular basis since his election as mayor in 2013. All of this only fuelled Plante’s rhetoric and gave the impression that incumbent Coderre had things to hide and was not trustworthy. Whether or not these transparency issues had any effect on voters, by the end of October 2017 polls had Plante two percentage points ahead of Coderre.15

15 CROP poll conducted 19–24 October. Source: Qc125 website.

16  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

Regarding the Quebec City municipal campaign, it is fair to say that its outcome was less surprising. A poll published in early September 2017 suggested that Régis Labeaume’s lead was insurmountable. It showed him with 49% of support, far ahead of Jean-François Gosselin (14%) and Anne Guérette (10%).16 By the end of the campaign, that lead had increased slightly: according to a late October poll, 51% of Quebec City voters intended to vote for him.17 Gosselin and Guérette had both also ­increased their support (to 22% and 14%, respectively), but not enough to threaten the incumbent. It seems that Labeaume’s central campaign message – Quebec City residents must re-elect him to preserve and strengthen the city’s political weight vis-à-vis higher-level governments – resonated strongly with voters. However, the same poll suggested that Gosselin’s party, Québec 21, was closing in on Équipe Labeaume in a few boroughs, most notably in Beauport and Charlesbourg. The campaign in Quebec City was dominated by three issues. The most important one was transportation and mobility in and around the city. It was debated significantly in all three of the televised debates, and it ­allowed the candidates to demarcate themselves clearly from one another. To address traffic congestion problems, Guérette proposed an 11 kilometer tramway across the city; Labeaume supported the development of a fast bus service. For his part, Gosselin took the car drivers’ side and strongly endorsed the building of a third road link across the St. ­Lawrence River (be it a bridge or a tunnel) to simplify automobile travel between the eastern part of the city and the south shore’s suburbs. Labeaume’s and Guérette’s main argument against the third link idea was that it might never be built because responsibility for an infrastructure project of this sort fell to the provincial and federal governments, not to the municipality. A second important issue in this campaign was the city’s economy. In particular, the Labeaume administration’s past decisions to increase property and commercial taxes were criticized by Québec 21 as creating a burden on taxpayers and businesses and ultimately hampering ­Quebec City’s economic growth. The tax issue and the third link proposal allowed G ­ osselin and his party to align themselves with the right-leaning provincial party C ­ oalition Avenir Québec (CAQ). The third salient campaign issue was quality of life. Gosselin emphasized park maintenance as key in this r­ espect. Labeaume pledged to develop more green space, more landscaping along the river, and more bicycle paths. Guérette proposed giving more authority to borough

16 Léger poll conducted 1–4 September. Source: Journal de Québec. 17 Léger poll conducted 25–26 October. Source: Journal de Québec.

Introduction 17

councils in this policy area and declared that she was in favour of subsidizing electric bikes. The Results On 5 November, Montrealers elected Valérie Plante as their first female mayor, with 51.42% of the vote. Denis Coderre had failed to secure a second term, though he received a significant share of electoral support, with 45.66% voting for him. Such was not the fate of Régis Labeaume in Quebec City, who was chosen as mayor for a fourth consecutive time with 55.27% of the vote. His closest rival ended up being Jean-François Gosselin, who received 27.63%. Anne Guérette closed the Quebec City mayoral race with 14.64% of support. The results of these two municipal election campaigns were quite different in a few noteworthy ways. First, the incumbent mayoral candidate won in Quebec City but lost in Montreal. Incumbent mayors usually ­succeed when seeking re-election (Collin 2011; Krebs 1998; Kushner et al. 1997; Mévellec 2011; Moore et al. 2017). In this sense, Labeaume was the norm, Coderre the exception. Relatedly, there was a change in partisan control of the City Council in Montreal but not in Quebec City, where 93% of ­incumbent city councillors were re-elected (14 out of 15). The only incumbent who was defeated in Quebec City was ­Marie-France ­Trudel of Équipe Labeaume, who was beaten by the Québec 21 candidate. Nevertheless, Équipe Labeaume was able to keep control of the council with 17 seats (see Table 1.1). Québec 21 was able to elect two councillors, both in the borough of Beauport in the eastern part of the city (near where the proposed third road link across the river would likely be built). Démocratie Québec managed to win only one seat. The story in M ­ ontreal is very different. The rate of incumbent ­re-election was lower than in ­Quebec City: 80% of incumbent city councillors (28 out of 35) and 75% of incumbent borough mayors (12 out of 16) were re-elected there. Most of the incumbents who lost had been running for Équipe Denis Coderre. Three high-profile borough mayors who had decided to run under the banner of Coderre’s party lost their re-election bids: Réal Ménard and Russell Copeman (formerly with C ­ oalition ­Montréal) and Anie Samson (who had left Vision Montréal to run for Coderre in 2013).18 Overall, Projet Montréal gained control of the City Council with 18 Similarly, three well-known incumbent city councillors who decided to run for Coderre's party failed to be re-elected: Richard Bergeron (formerly with Projet Montréal), Elsie Lefebvre (formerly with Coalition Montréal), and Lorraine Pagé (formerly with Vrai Changement pour Montréal).

18  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor Table 1.1.  City council seats obtained in the 2017 Quebec City election City councillors Équipe Labeaume Québec 21 Démocratie Québec Option Capitale-Nationale Alliance citoyenne de Québec Independents

17 2 1 0 0 1

Total

21

Source: Quebec government, Ministère des Affaires municipales et de l’Habitation.

33 elected officials (a gain of 14 seats from 2013). Équipe Denis Coderre saw 25 of its candidates elected. Marvin Rotrand of ­Coalition Montréal kept his seat; two borough-based parties, Équipe Anjou and Équipe Barbe (in LaSalle), hold the City Council’s five r­ emaining seats (see Table 1.2). Second, there was much more of a landslide in Quebec City than in Montreal. Labeaume was re-elected by a comfortable 28 points over his closest competitor. While Plante’s victory in Montreal was historic, her victory over Coderre was much narrower (less than six percentage points). Some of this difference may be attributable to the different number of viable mayoral candidates in each city, in the sense that the anti-incumbent vote was split between two main challengers in Quebec City while in Montreal it was essentially a two-way race. Still, Labeaume’s victory is impressive, especially given that he had been mayor for ten consecutive years by that point. The race in Montreal had a much closer outcome, which was a surprise, since most pre-campaign polls had predicted that Coderre would be re-elected. A third way that these two outcomes differ relates to electoral participation rates. In Quebec City, turnout in 2017 was 50.7%, while in Montreal it was just 42.5%. This gap in turnout between the two cities is similar to what has been observed in the recent past. For example, in the 2013 municipal elections, turnout was 54.9% in Quebec City while it was 43.3% in ­Montreal. Electoral participation in both cities declined in 2017 compared to 2013 but remained much higher in Quebec City than in Montreal. That said, the decline in turnout between the two elections was slightly smaller in Montreal, perhaps owing to the closer mayoral race in that city. Some Preliminary Survey Findings Before providing an overview of this volume, we discuss some key features of the elections in Montreal and Quebec City using CMES survey data.

Introduction 19 Table 1.2.  City council seats obtained in the 2017 Montreal election City councillors

Borough mayors

Total

Projet Montréal Équipe Denis Coderre pour Montréal Équipe Barbe Team Équipe Anjou Coalition Montréal Vrai changement pour Montréal Équipe Dauphin Lachine Plateau sans frontières Independents

23 19 2 1 1 0 0 0 0

10 6 1 1 0 0 0 0 0

33 25 3 2 1 0 0 0 0

Total

46

18

64

Source: Quebec government, Ministère des Affaires municipales et de l’Habitation.

First, we consider why turnout was higher in Quebec City than in M ­ ontreal and briefly examine the socio-demographic correlates of turnout in the two cities. We then attempt to explain the different fates of the two incumbent mayoral candidates. Finally, we run standard vote choice models to identify the correlates of support for the major mayoral candidates. These results provide important context for the chapters that follow.19 Explaining Turnout As noted earlier, one noteworthy way in which the elections in Montreal and Quebec City differed from each other was turnout. CMES data point to a number of factors that help account for the gap between the two cities.20 One was the competitiveness of the races: specifically, the mayoral election in Montreal was much more competitive than in Q ­ uebec City – Valérie Plante won by less than 6 percentage points, whereas Régis Labeaume won by almost 28 over his closest rival, Jean-François Gosselin. Such a finding is inconsistent with a well-established literature that predicts that competitiveness should increase voter turnout (Pattie & Johnston, 2001; Arnold, 2018). CMES data confirm that electors themselves perceived the race in Montreal as much closer than in Q ­ uebec City. Respondents were asked during the campaign to rate the candidates’ chances of winning, on a scale from 0 to 100. In Quebec City, Labeaume was a clear favourite, given an average of 74.9% chance of 19 All results are weighted for age and gender. 20 We recognize that, for some electors, turnout decisions may be influenced by council races. Nevertheless, for the purpose of our overview here, we focus on mayoral races.

20  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

winning.21 In Montreal, Coderre was perceived as a slight favourite, seen as having a 59.9% chance of victory, compared to 57.7% for Plante.22 Though Coderre held a slight edge in the minds of voters, this race was viewed as much more competitive than the one in Quebec City.23 The significant gap in turnout rates may thus be surprising in light of the actual and perceived competitiveness of the mayoral races. Competitiveness, however, is not the only factor that determines voter turnout. CMES data point to three differences between electors in ­Montreal and Quebec City that may account for the sizable turnout gap. Residents of Quebec City are more interested in municipal politics than Montrealers; they perceive municipal government as having a greater impact on their lives and feel a much higher sense of duty to vote in municipal elections. These three factors are all widely believed to be positively associated with turnout (Blais, 2000, 2006). In terms of political interest, residents of Quebec City had a slight edge over their counterparts in Montreal. When asked to rate their level of interest in municipal politics on a scale from 0 to 10, Quebeckers gave an average rating of 6.43. In Montreal, this value was 6.03 (this difference is significant at p < 0.01).24 While the magnitude of this ­effect may be somewhat difficult to contextualize, the direction is clear – ­Montrealers are less interested in local politics than are residents of the provincial capital. And it is not the case that Montrealers are simply uninterested in politics more generally. When the same “interest” question was asked regarding provincial and federal politics, no significant differences were found ­between residents of the cities with respect to interest in provincial politics (average scores here were 7.23 in Quebec City versus 7.07 in Montreal). Montrealers are actually much more interested in federal ­politics – the average level of interest in Montreal was 7.24, but only 6.81 in ­Quebec City (this difference is significant, at p < 0.01). So

21 N = 1,729. 22 This difference is significant at p < 0.01, N = 1,698. 23 The observant reader will note that the competitiveness values for Coderre and Plante add up to greater than 100 (as one might expect, given the nature of these questions). Such a finding is common in election study data, however, as at the aggregate, electors seem unwilling or unable to answer these types of questions such that values sum to 100%. The same pattern exists in Quebec City, where Labeaume is given a 74.9% chance of winning, while Gosselin is assigned a score of 36.7% (again, these values add up to well over 100%). For our purposes, the important pattern here is that Coderre and Plante are seen as very competitive, but Labeaume is seen as having a substantial advantage. 24 N = 1,867 for Montreal and 1,890 for Quebec City.

Introduction 21

there is something about local politics that either piques the interest of Quebecers or loses the interest of Montrealers (or both). Perhaps relatedly, Quebecers are much more likely to hold the opinion that municipal government has a significant bearing on their lives: 47.8% of respondents in Montreal believed that their local government had “a lot” of impact on their quality of life,25 while 64.9% of respondents in Quebec City thought so – a significant gap of more than 18 points (p < 0.01).26 It is no surprise that turnout is higher among those who feel that government has a great impact on their lives. A third explanation for the turnout differential between the two cities has to do with attitudes toward the act of voting. CMES data reveal that residents of the provincial capital are much more likely than their counterparts in Montreal to view voting in local elections as a duty rather than a choice. There is plenty of literature that points to the duty/choice distinction as key to turnout decisions (Loewen & Dawes, 2012; Blais & Achen, 2019). In this regard, 66.9% of Montrealers but 82.7% of Quebec City residents reported that voting in a local election is a duty – a significant difference (at p < 0.01).27 Given that Quebecers are much more likely to view voting as a duty, it is unsurprising that they turned out in relatively high numbers.28 CMES data thus show that Montrealers and Quebecers have very different attitudes toward local politics. Plain interest, the perception that local government is important, and viewing the act of voting as one’s democratic duty were enough to generate a relatively high turnout in Quebec City, despite the very uncompetitive mayoral race.

25 36.5% of respondents answered that municipal government had “some” impact, while the remaining 15.7% of respondents said that there was “not very much” or “no” impact on their lives. The corresponding values in Quebec City are 25.5%, 7.5% and 1.1%, respectively. N = 1,818 in Montreal and 1,863 in Quebec City. 26 Montrealers (57.4%) are also less likely to think that the provincial government has “a lot” of impact than are Quebecers (68.7%) (p < 0.01). The two cities are statistically indistinguishable with respect to attitudes toward the impact of the federal government: 48.8% of Montrealers and 51.8% of Quebecers believe that the federal government has “a lot” of impact. 27 N = 1,819 in Montreal and 1,880 in Quebec City. 28 In general, Quebecers are more likely than Montrealers to view voting as a duty rather than a choice. 76.5% of respondents in Montreal and 88.6% of Quebecers view voting in provincial elections as a duty (a gap of 12.3 points). The corresponding values for the federal level are 79.8% and 88.1% (a difference of 8.3 points) – note that differences for both provincial and federal estimates are significant at p < 0.01). The estimated gap between the cities is nevertheless greatest at the municipal level (15.8 points).

22  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

Finally, there is value in briefly considering the individual-level correlates of turnout in our two cases. Table 1.3 shows the results of two logistic regression models (entries show marginal effects), where the outcome variable is coded as 1 for voters and 0 for abstainers. We include a series of standard socio-demographic indicators as explanatory factors. These factors (all of which are coded as dummies) include age, gender, education, immigrant and visible minority status, mother tongue (francophone serves as the base), and household income.29 All independent variables are coded as dummies. Table 1.3 reveals many noteworthy findings. First, only one factor, income, is significant at the 95% level in both cities – electors with household incomes above the median are relatively likely to vote. Only one variable (immigrant status) is insignificant in both cities. All remaining factors are significant in one city only. In Montreal, men, as well as electors with a university education, are relatively likely to turn out. In Quebec City, older electors, whites, and allophones are comparatively likely to vote. However, these differences between cities should not be overblown. For most factors, the direction of effects is the same, even if a variable is significant in only one city. For only two factors do the cities differ from each other in a statistically significant way30: visible minorities are more likely to vote in Montreal, and allophones are more likely to vote in Quebec City. Though an explanation for these differences lies outside of the purview of this chapter, they do raise the question of whether these results help account for overall differences in turnout between the two cities. In this instance, we believe it highly unlikely that these differences are sufficient to explain the overall difference in turnout. ­Montreal has higher numbers of both immigrants and allophones, a pattern that should, at least to a certain extent, mean that these factors cancel each other out. Moreover, the r-squared values of the models are quite small, which means that socio-demographic factors explain only a small fraction of variation in voter turnout in either city. Still, while these results do little to help us explain aggregate-level differences in turnout between the cities, they nevertheless provide further evidence that the two municipalities differ in important ways and that the voting populations in Montreal and Quebec City are unique.

29 Education is coded as 1 for university degree and 0 for less than that. Visible minorities are those who identify themselves as having a background other than European. Income is coded as above/below the median. 30 This is according to a model where each socio-demographic factor is interacted with city. Results not shown but available upon request from the authors.

Introduction 23 Table 1.3.  Socio-demographic characteristics and voter turnout Montreal

Quebec City

Over 50 Female University educated Immigrant Visible minority Anglophone Allophone High income

0.06 (0.03) –0.10 (0.03)** 0.06 (0.03)* –0.03 (0.05) 0.12 (0.06) –0.01 (0.05) –0.07 (0.06) 0.08 (0.03)*

0.08 (0.02)** –0.04 (0.02) 0.03 (0.02) –0.01 (0.05) –0.17 (0.07)* –0.03 (0.07) 0.21 (0.10)* 0.08 (0.02)**

N Pseudo R2

851 0.0440

1,196 0.0549

Note: Entries report marginal effects and standard errors (in parentheses). * p < 0.05; ** p < 0.01

Explaining Incumbent Performance Besides very different rates of turnout, another significant difference between the elections in Montreal and Quebec City was the fate of the incumbent mayors and their parties. Régis Labeaume, first elected as mayor of Quebec City in 2007, held on to office with relative ease, and his party maintained its dominance on City Council, winning 86% of the seats. In Montreal, one-term mayor Denis Coderre was ousted by Valérie Plante, and many members of Coderre’s party at the council and borough levels were defeated by candidates affiliated with Plante’s party, Projet Montréal. CMES data point to a number of factors that help explain these vastly different outcomes. The first indicator is a standard measure of retrospective evaluations of mayoral performance. Respondents were asked how satisfied they were with the performance of the incumbent mayors. Levels of satisfaction were very much higher in Quebec City than in Montreal. As many as 63.9% of CMES respondents in Quebec City were satisfied with Labeaume’s performance: 18.0% were very satisfied and 45.9% somewhat satisfied. By contrast, evaluations of Coderre were much less rosy: 54.3% of Montrealers were satisfied with his performance (12.9% “very” and 41.4% “somewhat”). This “satisfaction” gap of nearly 10 points goes a long way toward explaining the different fates of the incumbent mayors. Still, given that a majority of respondents were satisfied with Coderre’s performance, his loss may seem surprising. CMES data suggest that his loss was the result of attitudes not only toward the mayor but also toward his opponent. Montrealers were satisfied with Coderre’s performance – they

24  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

simply liked Plante more than him. CMES respondents were asked to rate the two candidates on a standard 0–100 “feeling thermometer” scale, where a high value indicates a positive evaluation. Coderre received an average score of 50.1; the corresponding value for Plante was nearly 15 points higher, at 64.9.31 In Quebec City, Régis Labeaume suffered from no such deficit. His average feeling thermometer value was 57.3. In contrast, his closest rival, Jean-François Gosselin, received an average score of just 44.6, meaning that Labeaume had nearly a 13-point advantage.32 Put simply, voters were more satisfied with Labeaume than they were with Coderre, and in addition, Labeaume was more popular than his closet rival. Finally, besides being less popular than his challenger, Coderre was at a disadvantage because many more voters simply ruled him out ­altogether. CMES respondents were asked if there was a mayoral candidate for whom they absolutely would not vote. In Quebec City, just over one in five (20.6%) respondents identified Labeaume as someone they would never vote for (the remaining respondents either identified other candidates they would never vote for, or did not rule any candidate out). In contrast, nearly one third (33.2%) of Montrealers i­ ndicated that they would absolutely not vote for Coderre.33 Montreal’s incumbent was therefore at a significant disadvantage, relative to his counterpart in Quebec City: Coderre had a much smaller share of the voting population to ­potentially appeal to throughout the campaign than did Labeaume. The literature indicates that incumbents usually hold a significant advantage in local elections. Why was Coderre an exception to this ­ norm? The CMES data reveal that, even though most voters were relatively s­ atisfied with the performance of Montreal’s mayor, they simply did not like him as much as they did his opponent. At the same time, a relatively large share of residents ruled out the possibility of supporting the mayor. Such conditions make it difficult even for an incumbent to win. Explaining Vote Choice The final pieces of CMES data we present in this introductory chapter are the results of analyses of the socio-demographic correlates of mayoral vote choice. Survey respondents reported their vote choice in the post-election questionnaire, and these responses can be combined with responses to a variety of other questions to paint a comprehensive picture of the supporters of each of the major mayoral candidates. In 31 N = 1679. Difference significant at p < 0.01. 32 N = 1678. Difference significant at p < 0.01. 33 N = 1885 in Montreal and 1892 in Quebec City.

Introduction 25

this chapter, we model the relationship between vote choice and sociodemographic indicators. The chapters that follow include a variety of analyses of other types of indicators. Election studies at other levels of government in Canada (and elsewhere) often seek to explain vote choices through reference to the “funnel of causality” or a “bloc recursive model” (Miller & Shanks, 1996) (in Canada, see Blais et al., 2002; Gidengil et al., 2012; Fournier et al., 2013). The funnel of causality provides a conceptual framework for understanding how many factors can theoretically influence vote choice. The model incorporates (from most distal to most proximate) socio-demographics, values and beliefs, party identification, economic evaluations, issue opinions, and leader evaluations. It does not assume that every voter will draw on each of these factors when making a vote choice. Rather, the value of this approach is that it reflects an ordering of factors that may influence voters when they make their electoral choices (Gidengil et al., 2012). We believe that the conceptual framework provided by the funnel of causality model readily applies to elections at the local level. These electoral contests are inherently similar to those at other levels of government – voters can develop broadly ideological positions about the role and place of local government, can position themselves on competing visions of policy issue importance and direction, can develop identification with local politicians and challengers, and can develop perceptions of records of performance while in office. All of these factors (and more) existing at the federal and provincial levels in Canada also exist at the local level, especially in the larger cities that are the focus of this study. So we think it appropriate to draw on this framework to structure analyses that seek to understand the electoral choices of voters in M ­ ontreal and Quebec City. In our analysis, we consider only the socio-demographic correlates of vote choice (other factors will be considered in later chapters). We do so by way of a series of logistic regression models, where vote choice serves as the dependent variable. In Montreal, the outcome variable in the model compares Plante voters to Coderre supporters. In Quebec City, where there were three mayoral candidates of note, we run a model for each candidate, comparing his or her supporters to those who voted for the other two candidates. We thus have one model for Montreal and three for Quebec City.34 Table 1.4 shows the results of the socio-demographic analysis. Explanatory variables are the same as those included in Table 1.3 (the turnout

34 We omit respondents who supported minor candidates. The major candidates received over 97% of the vote in both cities.

26  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor Table 1.4.  Socio-demographic characteristics and vote choice Montreal

Quebec City

Plante vs Coderre

Gosselin vs others

Guérette vs others

Over 50 Female University educated Immigrant Visible minority Anglophone Allophone High income

–0.12 (0.04)** –0.05 (0.04) 0.01 (0.04)

–0.16 (0.03)** –0.10 (0.03)** –0.17 (0.03)**

–0.10 (0.07) –0.09 (0.09) 0.09 (0.07) 0.02 (0.08) –0.16 (0.04)**

–0.19 (0.12) –0.36 (0.23) –0.04 (0.14) 0.22 (0.19) 0.07 (0.03)*

0.13 (0.07) –0.15 (0.16) –0.03 (0.13) –0.07 (0.14) –0.05 (0.03)

–0.02 (0.10) 0.47 (0.22)* 0.07 (0.15) –0.11 (0.17) –0.02 (0.03)

N Pseudo R2

638 0.0375

1,024 0.0803

1,024 0.0750

1,024 0.0123

0.12 (0.03)** 0.05 (0.03) 0.16 (0.03)**

Labeaume vs others 0.05 (0.03) 0.06 (0.03) 0.02 (0.03)

Note: Entries report marginal effects and standard errors (in parentheses). * p < 0.05; ** p < 0.01

models). The results presented are marginal effects and can be interpreted as the effect of each characteristic on voting for the candidate of focus, as compared to the other candidates. In Montreal, Table 1.4 reveals only two socio-demographic factors related to vote choice: age and income. Young voters and lower-income Montrealers were attracted to the new mayor. These findings are not surprising, given that, as noted earlier, Plante was widely viewed as a leftleaning candidate and as representing a younger generation of politicians on Montreal’s municipal scene. Of note here is the null finding for the “female” variable: men and women were no different from each other when it came to mayoral vote choice. Plante is the first female mayor in the city’s history, and Montreal is now the largest city in North America with a female mayor. Likely because there were three major candidates in Quebec City, the results for that city are somewhat more complicated. Gosselin, the candidate furthest to the right on the political spectrum, did very poorly among older voters, women, those with a university education, and lowincome voters. Anne Guérette, the candidate of the left, performed well among older voters and those with higher education. The relationship between age and vote choice is therefore unusual in this instance. In general, older voters tend to support more conservative candidates, yet the ­opposite appears to be true in Quebec City. In fact, the sociodemographic make-up of Jean-François Gosselin’s supporters loosely

Introduction 27

mirrors the infamous phenomenon of Quebec City’s “young angry males” (Langlois, 2004; Turbide et al., 2008). Only one variable in the Labeaume model is statistically significant – the mayor was heavily favoured by visible minorities. The observant reader will notice, however, that the R-squared value in the Labeaume model is but a small fraction of those observed for the other candidates. For the incumbent mayor, therefore, socio-demographics do very little to explain his success in the 2017 election. One additional observation from Table 1.4 is worth mentioning. Language is not correlated with vote choice in either Quebec City or Montreal. These null findings are noteworthy, given the centrality of language to politics in the province and the large anglophone population in Montreal. What Is to Come To this point, the reader has been offered a brief overview of the literature on local political behaviour in Quebec and Canada, our case for the value in a book-length study of the 2017 elections in Montreal and ­Quebec City, an overview of the CMES, background on the candidates and the elections themselves, and some relevant data on the elections. This information provides important contextual information for the chapters that follow. The following chapters are arranged in an order that corresponds to the previously discussed funnel of causality (Campbell et al., 1960). Early chapters focus on those factors that are considered distal to vote choice and comparatively stable across time. In chapter 2, Pruysers tests for the presence of gender affinity effects in the two mayoral elections. Both cities had serious female mayoral candidates. He finds that, on its own, gender has little effect on mayoral vote choice, but that affinity effects were moderated by electors’ attitudes toward feminism. Anderson and Stephenson focus on the impact of language on voter turnout. More specifically, in chapter 3, they consider the effect of linguistic diversity on participation rates, finding mixed evidence that such a relationship exists. The next chapters consider factors that are assumed to also be relatively long-term but that are attitudinal rather than socio-demographic. Daoust, in chapter 4, explores the role of nationalism in local elections. One might expect the nationalist/federalist dimension to have little influence on local politics, given that municipal governments have little or no policy power over issues related to either separatism or federal/ provincial dynamics. Even so, Daoust finds evidence that nationalist attitudes significantly affected vote choice in both of our two cities. In

28  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor

chapter 5, Breux and colleagues examine the relationship between traditional left/right ideology and vote choice. Contrary to claims often made that municipal politics are non-ideological, the authors conclude that ideology did indeed play a major role in vote choice in both cities. Bélanger and Stephenson, in chapter 6, assess the presence and influence of local partisanship in these elections. They make the case that, despite the somewhat transient nature of local political parties, the concept of partisanship (developed in the context of national elections) travels well to the local level in the province of Quebec. Chapter 7 further inspects the effects of parties and party labels on local politics in our two cities. Here, Anderson and McGregor report the results of a survey experiment that show that party labels have a significant effect on the ability of electors to state a vote preference. H ­ owever, these effects are weaker among highly knowledgeable electors and stronger among partisans. The final chapters focus on short-term factors of the sort that are most proximate to vote choice in the funnel of causality. In chapter 8, Bélanger and Tessier consider the relationship between retrospective economic evaluations and vote choice. One might assume that mayors have relatively little power over the state of the economy; however, the authors conclude that economic evaluations significantly influenced vote choice in Montreal (though not in Quebec City). In chapter 9, Tessier and McGregor investigate the phenomenon of straight-ticket voting (or voting the same party for multiple positions in a concurrent election). After estimating rates of straight-ticket voting in the two cities, they explore the correlates of such behaviour. In the tenth and final chapter, Besco and Matthews examine the influence of campaigns on local voters. They find no evidence that the municipal campaigns in either Montreal or Quebec City had an “enlightening” effect on voters (i.e., the campaigns did not improve knowledge levels among the electorate). Municipal elections in Canada are substantially different from those held at the provincial and federal levels. As is the case with respect to many political factors, Quebec is unique within Canada. The presence of unique party systems, unusual government structures, and the ­nationalist dimension, coupled with the comparative lack of a­ cademic interest in local political behaviour, makes a thorough investigation of elections in this context invaluable for our understanding of ­Canadians’ voting behaviour. To date, however, our understanding of municipal elections in the province are largely shrouded in darkness. By telling the tales of the elections in these two cities, we hope to begin to cast light on these important and fascinating elections and the people who participate in them.

Introduction 29 REFERENCES Anderson, C. (2008). Economic voting, multilevel governance, and information in Canada. Canadian Journal of Political Science, 41(2), 329–354. https://doi.org /10.1017/S0008423908080414 Arnold, F. (2018). Turnout and closeness: Evidence from 60 years of Bavarian mayoral elections. Scandinavian Journal of Economics, 120(2), 624–653. https:// doi.org/10.1111/sjoe.12241 Belley, S. (1992). Les partis politiques municipaux et les élections de 1986 à Montréal et de 1989 à Québec. Politique, 21, 5–35. https://doi.org/10.7202 /040711ar Belley, S., & Lavigne, M.-A. (2008). Apolitisme, partis politiques et prégnance des institutions: le cas de l’élection municipale de 2005 à Québec. Recherches Sociographiques, 49(1), 47–68. https://doi.org/10.7202/018193ar Belley, S., Lavigne, M.-A., & Quesnel, L. (2014). L’élection municipale de 2013 à Québec: la légitimation par les urnes d’un nouveau leadership local. Canadian Journal of Urban Research, 23(2), 100–122. Bherer, L. (2006). Le cheminement du projet de conseils de quartier à Québec (1965–2006): un outil pour contrer l’apolitisme municipal? Politique et Sociétés, 25(1): 31–56. https://doi.org/10.7202/013514ar Bherer, L., & Breux, S. (2011). Démocratie locale et élections: prémices d’une comparaison. In S. Breux & L. Bherer (Eds.), Les élections municipales au Québec: enjeux et perspectives (pp. 1–32). Presses de l’Université Laval. Bird, K., Jackson, S., McGregor, M., Moore, A., & Stephenson, L. B. (2016). Sex (and ethnicity) in the city: Affinity voting in the 2014 Toronto mayoral election. Canadian Journal of Political Science, 49(2), 359–383. https://doi.org /10.1017/S0008423916000536 Blais, A. (2000). To vote or not to vote. University of Pittsburgh Press. Blais, A. (2006). What affects voter turnout? Annual Review of Political Science, 9(1), 11–25. https://doi.org/10.1146/annurev.polisci.9.070204.105121 Blais, A. (2010). Making electoral democracy work. Electoral Studies, 29(1), 169–170. https://doi.org/10.1016/j.electstud.2009.05.006 Blais, A., & Achen, C. H. (2019). Civic duty and voter turnout. Political Behavior, 41(2), 473–497. https://link.springer.com/article/10.1007/s11109-018-9459-3 Blais, A., Gidengil, E., Nadeau, R., & Nevitte, N. (2002). Anatomy of a Liberal victory: Making sense of the vote in the 2000 Canadian election. Broadview Press. Bourassa, G. (1971). Les relations ethniques dans la vie politique montréalaise. Documents of the Royal Commission on Bilingualism and Biculturalism. Ottawa: Information Canada. Breux, S. (2019). Offre électorale, enjeux et électeurs: les élections municipales au Québec. In J.-P. Brady (Ed.), Introduction à la vie politique municipale québécoise (pp. 63–92). Presses de l’Université Laval.

30  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor Breux, S., & Bherer, L. (2011). Les élections municipales au Québec: enjeux et perspectives. Presses de l’Université Laval. Breux, S., Couture, J., & Koop, R. (2017). Turnout in local elections: Evidence from Canadian cities, 2004–2014. Canadian Journal of Political Science, 50(3): 1–24. https://doi.org/10.1017/S000842391700018X Breux, S., Couture, J., & Koop, R. (2019). Influences on the number and gender of candidates in Canadian local elections. Canadian Journal of Political Science, 52(1): 163–181. https://doi.org/10.1017/S0008423918000483 Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter. Wiley. Cardinal, L., & Mévellec, A. (2012). La représentation politique des francophones d’Ottawa: la situation des élus francophones au conseil municipal, 2000–2010. Francophonies d’Amérique, 34, 81–103. https://doi.org /10.7202/1023782ar Chiasson, G., & Mévellec, A. (2014). The 2013 Quebec municipal elections: What is specific to Quebec?” Canadian Journal of Urban Research, 23(2), 1–17. Collin, J.-P. (2011). Quel avenir pour la démocratie municipale québécoise? In S. Breux & L. Bherer (Eds.), Les élections municipales au Québec: enjeux et perspectives (pp. 311–351). Presses de l’Université Laval. Couture, J., Breux, S., & Bherer, L. (2014). Analyse écologique des déterminants de la participation électorale municipale au Québec. Canadian Journal of Political Science, 47(4), 787–812. https://doi.org/10.1017/S0008423914001152 Cutler, F., & Matthews, S. (2005). The challenge of municipal voting: Vancouver 2002. Canadian Journal of Political Science, 38(2), 359–382. https://doi.org /10.1017/S0008423905040151 Davidson, A., McGregor, R. M, & Siemiatycki, M. (2020). Gender, race, and political ambition: The case of Ontario school board elections. Canadian Journal of Political Science, 53(2), 461–475. https://doi.org/10.1017 /S0008423919001057 Dostie-Goulet, E., Blais, A., Fournier, P., & Gidengil, E. (2012). L’abstention sélective, ou pourquoi certains jeunes qui votent au fédéral boudent les élections municipales. Canadian Journal of Political Science, 45(4), 909–927. https://doi.org/10.1017/S0008423912001084 Drouilly, P. (1991). L’élection municipale de 1990 à Montréal. In D. Monière (Ed.), L’Année politique au Québec 1990–1991. Éditions Québec/Amérique. Drouilly, P. (2010). Montréal 2009: Curieuse élection. In M. Fahmy (Ed.), L’État du Québec 2010 (pp. 406–412). Boréal. Eidelman, G., & Taylor, Z. (2010). Canadian urban politics: Another “black hole”? Journal of Urban Affairs, 32(3), 305–320. https://doi.org/10.1111 /j.1467-9906.2010.00507.x Federation of Canadian Municipalities (FCM). (2013). 2013 – municipal statistics: Elected officials gender statistics. http://www.fcm.ca/Documents /reports/Women/2013_municipal_statistics_elected_official_gender_EN.pdf

Introduction 31 Fischel, W. A. (2001). The homevoter hypothesis: How home values influence local government taxation, school finance, and land-use policies. Harvard University Press. Gerbet, T., & Schué, R. (2020). Projet Montréal a perdu plus de la moitié de ses membres depuis 2017. Ici.Radio-Canada.ca, 30 October 2020. Gidengil, E., Nevitte, N., Blais, J. E., & and Fournier, P. (2012). Dominance and decline. University of Toronto Press. Gidengil, E., & Vengroff, R. (1997). Representational gains of Canadian women or token growth? The case of Quebec’s municipal politics. Canadian Journal of Political Science, 30(3), 513–537. https://doi.org/10.1017/S0008423900015997 Jetté, É. (2019). “Je suis tatouée QS sur le front,” dit Nima Machouf qui veut représenterle NPD aux prochaines élections. Le Journalde Montréal, 15 March 2019. Kanji, M., Bilodeau, A., & Scotto, T. (2012). The Canadian election studies: Assessing four decades of influence. UBC Press. Kiss, S., Perrella, A., & Spicer, Z. 2020. Right-wing populism in a metropolis: Personal financial stress, conservative attitudesm and Rob Ford’s Toronto. Journal of Urban Affairs, 42(7), 1028–1046. https://doi.org/10.1080/07352166 .2019.1657021 Kramer, G. H. (1983). The ecological fallacy revisited: Aggregate- versus individual-level findings on economics and elections, and sociotropic voting. American Political Science Review, 77(1), 92–111. https://doi.org/10.2307 /1956013 Krebs, T. B. (1998). The determinants of candidates’ vote share and the advantage of incumbency in city council elections. American Journal of Political Science, 42(3), 921–935. https://doi.org/10.2307/2991735 Kushner, J., & Siegel, D. (2003). Effect of municipal amalgamations in Ontario on political representation and accessibility. Canadian Journal of Political Science, 36(5), 1035–1051. https://doi.org/10.1017/S0008423903778950 Kushner, J., Siegel, D., & Stanwick, H. (1997). Ontario municipal elections: Voting trends and determinants of electoral success in a Canadian province. Canadian Journal of Political Science, 30(3), 539–553. https://doi.org/10.1017 /S0008423900016000 Langlois, S. (2004). Jeunes hommes en colère à Québec: malaise de classe et de génération. In M. Venne (Ed.), L’annuaire du Québec 2005 (pp. 92–94). Fides. Lévesque, F. (2016). Le NPD mise sur une élue de Projet Montréal. La Presse+, 12 August 2019. Loewen, P. J., & Dawes, C. T. (2012). The heritability of duty and voter turnout. Political Psychology, 33(3), 363–373. https://doi.org/10.2307/23260396 Lucas, J., Merrill, R., Blidook, K., Breux, S., Conrad, L., Eidelman, G., Koop, R., Marciano, D., Taylor, Z., & Vallette, S. (2021). Women’s municipal electoral performance: An introduction to the Canadian municipal elections database. Canadian Journal of Political Science, 54(1), 125–133. https://doi.org/10.1017 /S000842392000102X

32  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor Lucas, J. (2021). The size and sources of municipal incumbency advantage in Canada. Urban Affairs Review, 57(2), 373–401. https://doi.org/10.1177 /1078087419879234 Marschall, M., Shah, P., & Ruhil, A. (2011). The study of local elections; editors’ introduction: A looking glass into the future. PS: Political Science and Politics, 44(1), 97–100. https://doi.org/10.2307/40984492 Matson, M., & Fine, T. S. (2006). Gender, ethnicity, and ballot information: Ballot cues in low-information elections. State Politics and Policy Quarterly, 6(1), 49–72. https://doi.org/10.1177/153244000600600103 McGregor, M. (2018). Voters who abstain: Explaining abstention and ballot roll-off in the 2014 Toronto municipal election. Urban Affairs Review, 54, 1081–1106. https://doi.org/10.1177/1078087416688960 McGregor, M., & Lucas, J. (2019). Who has school spirit? Explaining voter participation in school board elections. Canadian Journal of Political Science, 52(4), 923–936. https://doi.org/10.1017/S0008423919000088 McGregor, R. M., Moore, A., Jackson, S., Bird, K., & Stephenson, L. (2017). Why so few women and minorities in local politics? Incumbency and affinity voting in low information elections. Representation, 53(2), 135–152. https:// doi.org/10.1080/00344893.2017.1354909 McGregor, M., Moore, A., & Stephenson, L. (2016). Political attitudes and behaviour in a non-partisan environment: Toronto 2014. Canadian Journal of Political Science, 49(2), 311–333. https://doi.org/10.1017/S0008423916000573 McGregor, R. M., Moore, A., & Stephenson, L. (2021). Electing a mega-mayor: Toronto 2014. University of Toronto Press. Mévellec, A. (2011). Les élections municipales de 2009 dans les villes moyennes du Québec: entre changement et reconduction. In S. Breux & L. Bherer (Eds.), Les élections municipales au Québec: enjeux et perspectives (pp. 289–310). Presses de l’Université Laval. Mévellec, A. (2014). Did the 2013 municipal elections destabilize municipal politics in Quebec? Canadian Journal of Urban Research, 23(2), 18–37. Mévellec, A., & Tremblay, M. (2013). Les partis politiques municipaux: la “westminstérisation” des villes du Québec? Recherches Sociographiques, 54(2), 325–347. https://doi.org/10.7202/1018284ar Moore, A., McGregor, R. M., & Stephenson, L. B. (2017). Paying attention and the incumbency effect: Voting behavior in the 2014 Toronto municipal election. International Political Science Review, 38(1), 85–98. https://doi.org /10.1177%2F0192512115616268 Oliver, J., & Ha, S. (2007). Vote choice in suburban elections. American Political Science Review, 101(3), 393–408. https://doi.org/10.1017/S0003055407070323 Olson, M. (1969). The principle of “fiscal equivalence”: The division of responsibilities among different levels of government. American Economic Review, 59(2), 479–487.

Introduction 33 Ostrom, V., Tiebout, C., & Warren, R. (1961). The organization of government in metropolitan areas: A theoretical inquiry. American Political Science Review, 55(4), 831–842. https://doi.org/10.1017/S0003055400125973 Pattie, C., & Johnston, R. (2001). A low turnout landslide: Abstention at the British general election of 1997. Political Studies, 49(2), 286–305. https://doi .org/10.1111%2F1467-9248.00314 Porter, I. (2019). Jean-François Gosselin cherche à sauver son parti politique. Le Devoir, 9 April 2019. Quesnel, L. (2001). Municipal reorganisation in Quebec. Canadian Journal of Regional Science, 21(1), 115–134. Quesnel, L. (2005). La restructuration des agglomérations métropolitaines au Québec. In R. Hudon and J.-P. Augustin (Eds.), Villes, régions, universités, 221–240. Presses de l’Université Laval. Quesnel, L., with Hamel, S. (2007). Your guide to municipal institutions in Canada. Federation of Canadian Municipalities, International Centre for Municipal Development. Quesnel, L., & Belley, S. (1991). Partis politiques municipaux: une étude de sociologie électorale. Agence d’ARC. Ricard-Châtelain, B. (2020). Option capitale-nationale devient Transition Québec. Le Soleil, 25 May 2020. Robinson, W. S. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15(3), 351–357. https://doi.org/10.1093/ije /dyn357 Sancton, A. (2000). Merger mania: The assault on local government. Price-Patterson. Sancton, A. (2015). Canadian local government: An urban perspective (2nd ed.). Oxford University Press. Schatten, W., Ruderman, N., Zhao, Z, & Nevitte, N. (2015). The 2014 Toronto municipal election. Paper presented at the 2015 meeting of the Canadian Political Science Association, Ottawa. Silver, D., Taylor, Z., & Calderon-Figueroa, F. (2020). Populism in the city: The case of Ford Nation. International Journal of Politics, Culture, and Society, 33(4), 1–21. https://link.springer.com/article/10.1007/s10767-018-9310-1 Simard, C. (2004). Qui nous gouverne au municipal: reproduction ou renouvellement? Politique et Sociétés, 23(2), 135–158. https://doi.org/10.7202 /010887ar Spicer, Z., McGregor, R. M., & Alcantara, C. (2017). Political opportunity structures and the representation of women and visible minorities in municipal elections. Electoral Studies, 48, 10–18. https://doi.org/10.1016 /j.electstud.2017.01.002 Tolley, E. (2011). Do women “do better” in municipal politics? Electoral representation across three levels of government. Canadian Journal of Political Science, 44(3), 573–594. https://doi.org/10.1017/S0008423911000503

34  Éric Bélanger, Cameron D. Anderson, and R. Michael McGregor Tremblay, M, & Mévellec, A. (2013). Truly more accessible to women than the legislature? Women in municipal politics. In L. Trimble, J. Arscott, & M. Tremblay (Eds.), Stalled: The representation of women in Canadian governments (pp. 19–35). UBC Press. Turbide, O., Vincent, D., & Laforest, M. (2008). Les “X” à Québec: la construction discursive d’un groupe exclusif. Recherches sociographiques, 49(1): 87–112. https://doi.org/10.7202/018195ar Vermeir, J., & Heyndels, B. (2006). Tax policy and yardstick voting in Flemish municipal elections. Applied Economics, 38(19), 2285–2298. https://doi.org /10.1080/00036840500427536 Wesley, J. (2015). Big worlds: Politics and elections in the Canadian provinces and territories. University of Toronto Press. Young, R., & Horak, M. (2012). Sites of governance: Multilevel governance and policy making in Canada’s big cities. Montreal & Kingston: McGill–Queen’s University Press. Young, R., & Leuprecht, C. (2004). Canada: The state of the federation 2004: Municipal–federal–provincial relations in Canada. Montreal & Kingston: McGill– Queen’s University Press.

2 Gender Affinity Voting and the 2017 Municipal Elections in Quebec City and Montreal scott pruysers

It should come as no surprise that there is mounting evidence that women face considerable barriers in electoral politics. Recent research suggests that women are less interested in a seeking a political career than similarly qualified men (Lawless & Fox, 2005; Preece & Stoddard, 2015; Pruysers & Blais, 2017, 2018a, 2018b), that women are less likely to be recruited into politics as candidates by political party officials than men (Butler & Preece, 2016; Cross & Young, 2013; Pruysers & Cross, 2016), that women are more likely to be nominated in unwinnable or long-shot ridings (i.e., “sacrificial lambs”; see Thomas & Bodet, 2013), that women may face a motherhood penalty (Thomas & Bittner, 2017), and that women candidates face gender stereotypes (Aalberg & ­Jenssen, 2007; Huddy & Terkildsen, 1993; ­Sanbonmatsu & Dolan, 2009; ­Sanbonmatsu, 2006) and voter bias (­Anzia & Berry, 2011; Fulton, 2012; ­Pearson & McGee, 2013; but see also Sevi et al., 2018).1 Furthermore, the ­Canadian case suggests that women do not perform better at lower levels of government (Tolley, 2011), which dispels the myth of the “municipal advantage.”2 However, women candidates are often expected to have an advantage when it comes to women voters. As Dolan (2008: 79) writes, “one rather common assumption people make is that women voters are the most likely source of support for female candidates, a phenomenon referred to as the ‘gender affinity effect.’”

1 While Sevi and colleagues (2018) find that women candidates in Canada do not receive fewer votes than their male counterparts, recent literature has convincingly demonstrated that similar vote shares for men and women is not a gender-neutral process. To perform as well as men, women candidates have to be more qualified and experienced (see, for example, Bauer, 2020). 2 Women, for instance, account for less than 20% of mayors in Canada and less than 30% of city councillors. The percentage of city councillors who are women is on par with that of women Members of Parliament (26%).

36  Scott Pruysers

While there are a variety of reasons why we might expect women to behave in this manner, including politicized gender consciousness (Goodyear-Grant & Croskill, 2011), the desire to increase the number of women elected to democratic institutions (Sanbonmatsu, 2002), and a reliance on readily available heuristic cues such as candidate gender (King & Matland, 2003), the basis of the gender affinity hypothesis is rather simple: all other things equal, women are expected to vote disproportionately in favour of women candidates. As discussed below, however, evidence of a gender affinity effect in both real-world and experimental elections has been mixed. Moreover, while studies in the United States have found some evidence of gender affinity voting, research on the ­Canadian case has tended to find very little support for a gender affinity effect in actual (as opposed to experimental) elections (see, for instance, Goodyear-Grant & Croskill, 2011; McGregor et al., 2017). Furthermore, when affinity effects have been identified, they have tended to be moderated by other factors such as race (see, for instance, Bird et al., 2016). This chapter contributes to the debate by testing for the existence of gender affinity effects (and potential moderating factors) in two mayoral elections: the 2017 contests in Quebec City and Montreal. In both cities, a relatively popular woman was running for mayor: Valérie Plante who won 51% of the vote in Montreal, and Anne Guérette who won 15% in Quebec City. Using CMES data from these two cities, the results reveal a number of important findings regarding gender affinity effects but also the role of gender in municipal elections more generally. O ­ verall, the results provide little evidence of a direct gender affinity effect. In ­neither of the two cities did women vote overwhelmingly for either Plante or Guérette. However, the data also reveal that gender did matter for some voters, depending on the attitudes held by those individuals. In particular, the interaction between voter sex and attitudes about gender had a strong relationship with vote choice. In both cities, women who held favourable views of feminists, for instance, were more likely to vote for the lone woman mayoral candidate. Also in both cities, voters who thought council should reflect the make-up of society (i.e., 50% women) were more likely to cast their ballot for a woman. As an aside, data from ­Quebec City also suggest that while women did not necessarily vote in favour of the woman in the race (Guérette), they did tend to vote against a particular candidate (Gosselin). The results therefore demonstrate that gender matters in Canadian municipal elections, although perhaps not in the direct and straightforward ways in which the gender affinity hypothesis would predict. This adds to mounting evidence that affinity effects are likely to be moderated by other factors – in this case, by political attitudes.

Gender Affinity Voting  37

Gender Affinity Voting Why Gender Affinity? Why are women expected to be a particularly strong base of support for female candidates? The simplest explanation has to do with an in-group/ out-group effect: women, preferring same-sex representation, vote for women candidates because of their shared sex (Sanbonmatsu, 2002). Voting for female candidates, therefore, may be an attempt to address the underrepresentation of women and challenge the representational status quo (Dolan, 2008). Given that women continue to be considerably underrepresented in national legislatures around the world (InterParliamentary Union, 2018), women’s desire to actively challenge the status quo is not unreasonable. Social identity and gender consciousness arguments push this explanation further. A social identity explanation focuses on feelings of group membership and solidarity. As Goodyear-Grant and Croskill (2011: 226) explain, gender identity, or gender consciousness, “has a cognitive component whereby group members understand women’s economic, social, and political disadvantage, and see it as the result of systemic discrimination, not individual failure; gender consciousness also has an affective component whereby group members resent the inequality and acquire feelings of connectedness with the group and its interests as a response.” This shared identity and positive affect draws women voters to support women candidates. In some instances, this identity may be strong enough to encourage women voters to “switch” parties in order to vote for a woman candidate. Another explanation centres on the (uneven) distribution of political resources such as knowledge and campaign attention. Given that women tend to score lower on tests of political knowledge than men (see Deli Carpini & Keeter, 1996; Pruysers & Blais, 2014), women are expected to make greater use of heuristic shortcuts, such as demographic similarity, when casting their ballots – though there is compelling evidence that the gap may not be as large as typically portrayed (Mondak & Anderson, 2004; Stolle & Gidengil, 2010).3 That said, the knowledge gap is particularly

3 Mondak and Anderson (2004), for instance, reveal that men are more likely to guess on knowledge tests than are women, a fact that accounts for some 50% of the g ­ ender gap. Similarly, Stolle and Gidengil (2010: 93) demonstrate that at least part of the gender gap in knowledge is due to “gender-biased measures that underestimate ­women's political knowledge.”

38  Scott Pruysers

relevant in low-information municipal elections where policy and other campaign information may be difficult to obtain but where cues such as candidate gender are readily available for even low-attention voters. As King and Matland (2003: 596) write, “voters are cognitive misers who, when faced with the need to evaluate a candidate, are very unlikely to invest significant resources in gathering comprehensive information.” Individuals with low levels of political knowledge, when voting in an ­already low-information election, are expected to rely on shared demographic traits such as gender to assist in their voting decisions. Evidence of Gender Affinity? The phenomenon of affinity voting has been studied extensively over the past three decades in a variety of settings (Brians, 2005; Dolan, 2008; Goodyear-Grant 2010; King & Matland, 2003; Matson & Fine, 2006; ­McDermott, 1998; Plutzer & Zipp, 1996; Sanbonmatsu, 2002; Smith & Fox, 2001; Zipp & Plutzer, 1985). Literature from the United States, the case that has been studied most extensively, has been somewhat mixed regarding evidence in support of a gender affinity effect. While a number of studies have demonstrated that women are indeed more supportive of female candidates (e.g., Brians, 2005), others have found such an effect does not exist (e.g. King & Matland, 2003). Still other studies have concluded that affinity effects are often mediated or conditioned by forces beyond shared sex (e.g., Plutzer & Zipp, 1996). In other words, there does not appear to be any clear consensus regarding gender affinity voting. If the evidence from the American case is mixed, what is the evidence from Canada? While there is a considerable body of research that ­explores the various correlates of vote choice in Canadian federal elections (see, for example, Blais et al., 2002, 2003; Bittner, 2011; Kanji et al., 2012; Gidengil, 1992; Gidengil et al., 2012; see also Bélanger and Tessier, this volume; Breux et al., this volume), less attention has been devoted to the question of gender affinity.4 Studies that have examined affinity ­voting, however, have typically found little support for the phenomenon in Canada. For example, Goodyear-Grant (2010) studied the 2004 and 2006 Canadian federal elections and found no evidence of a gender ­affinity effect. In fact, she reports that in both elections, men were more likely than women to vote for female candidates. Likewise, Cutler

4 While most Canadian studies have found little support for a gender affinity effect, Cutler (2002) finds that Canadian voters are more likely to support a federal political party that has a leader with the same sex as the voter.

Gender Affinity Voting  39

and Matthews (2005) examined voter behaviour in the 2002 Vancouver ­municipal election and demonstrated that women were no more likely to support the female mayoral candidate (Jennifer Clarke) than were the men in their sample. These general findings are echoed by McGregor and colleagues (2017), who find no evidence of gender affinity at the ward level in Toronto municipal elections. Although the evidence from Canada seems rather clear, two recent studies are worth reviewing in greater depth. While the two studies analyse different scales (municipal vs federal) and different actors (mayoral candidates vs local constituency candidates), their results are similar in a number of ways. In particular, neither study finds evidence of a direct or unmediated gender affinity effect, and both conclude that when affinity effects are found, they are conditioned by other factors. That is, both studies highlight the need to explore interactions between voter sex and other factors (race, attitudes, ideology, etc.) in the search for a better understanding of gender affinity effects. Bird and colleagues (2016) explored gender and ethnic affinity voting in the 2014 Toronto mayoral election. Using data from the Toronto Election Study, they found no evidence of a traditional gender affinity effect. Indeed, in none of their final models was gender a significant predictor of support for Chow once attitudes and ideology were introduced. Thus, gender did not seem to be an important factor. The authors did find, however, that the interaction between race and gender was a significant predictor of Chow support: After differentiating between voters on the basis of race, TES data suggest the relationship between gender and vote choice depends upon the race of voters. We observe null results for the gender variable when it is not interacted with race because the impact of voter gender on candidate choice runs in opposite directions for white and visible minority voters. Controlling for egalitarian attitudes, ideology and partisanship, white women were less likely to support Chow than were white men, while visible minority women (Chinese or otherwise) offer more support to Chow than did their male counterparts. (375)

Overall, Bird and colleagues (2016) found much more evidence of an ethnic affinity effect than of a gender affinity effect.5 Their research does,

5 For other recent research regarding ethnic affinity effects, see Besco (2015), Bird (2016), and Goodyear-Grant and Tolley (2017). Note that much of this research is based on online experiments, which may explain why these studies find affinity effects whereas other studies, typically using election data, do not.

40  Scott Pruysers

however, point to important ways in which gender intersects with other factors such as race to become a significant predictor of vote choice. Exploring possible gender affinity affects in the 2000 and 2004 ­Canadian federal elections, Goodyear-Grant and Croskill (2011) used the Canadian Election Study to explore whether women were more likely to support women candidates at the constituency level. The authors found “very little evidence of gender affinity effects between women voters and women candidates,” noting that during the 2004 election, men were actually more likely to support women candidates (243). The same study examined whether voter flexibility conditions voting for women candidates – that is, whether women whose partisan ties are less strong, and who are therefore able to more freely choose among local candidates, are more likely to support women candidates. Here too, however, the authors found no evidence of a gender affinity effect: “Even when women are relatively less encumbered by strong partisan loyalties – a situation in which candidate considerations could come to the fore in the vote calculus – women voters are still no more likely than male voters to vote for women” (246). Goodyear-Grant and Croskill did find some (modest) evidence of a gender affinity effect in their analysis of whether gender affinity might be conditioned by political attitudes. They found that in the 2004 election, women who were in favour of greater women’s representation were more likely to vote for women candidates, although this was not the case in the 2000 election. Given this, the interaction between voter sex and political attitudes might be an important predictor of affinity voting. The evidence, though limited, suggests that gender affinity voting is not at all prevalent in Canada, and that when it does exist, it tends to be conditioned by other factors such as race or political attitudes. There is limited research on this topic, however, so further investigation is warranted. Moreover, in light of the findings of Bird and colleagues (2016) and Goodyear-Grant and Croskill (2011), it is important to consider whether certain kinds of women are more likely to support female candidates. Given that the elections being examined in this chapter do not allow us to consider the interaction between gender and race, one goal here is to unpack the interaction between voter sex and political attitudes on the one hand, and the interaction between voter sex and political resources on the other. In terms of attitudes, I consider beliefs relating to women in politics (i.e., gender representation on council) and gender more broadly (i.e., attitudes towards feminists or childrearing), and for resources, I consider political knowledge and campaign attention.

Gender Affinity Voting  41

2017 Municipal Elections in Montreal and Quebec City To explore the question of gender affinity voting, this chapter considers recent mayoral elections in Montreal and Quebec City. The 2017 mayoral election in Montreal was a two-way race between incumbent mayor Denis Coderre (Équipe Denis Coderre) and Valérie Plante (Projet Montréal). Coderre, a former Member of Parliament and federal cabinet minister, had served as mayor since 2013; Plante had served on City Council since 2013. A third candidate, Jean Fortier, was officially on the ballot, but he dropped out of the race weeks before Election Day and publicly endorsed Plante for mayor. With 51% of the vote, Plante won the election and became the first woman mayor in the city’s history. Coderre, having received 46% of the vote, announced his retirement from politics. The 2017 election in Quebec City was a three-way race between incumbent mayor Régis Labeaume (Équipe Labeaume), Jean-François ­Gosselin (Québec 21 Équipe JF Gosselin), and Anne Guérette (Démocratie Québec). All three candidates had considerable political experience. Labeaume was the long-time incumbent mayor; Gosselin was a former member of the Quebec National Assembly; Guérette had served on City Council for a decade. Labeaume secured his fourth mandate with 55% of the vote, compared to 28% for Gosselin and 15% for Guérette. Labeaume was victorious, but his base of support had diminished considerably, down from 74% in 2013, when his party won all but three City Council seats. While gender was not a central campaign theme or a focus of either election, issues related to gender did arise in various ways throughout the campaigns. Plante, for instance, began her mayoral campaign with the slogan “L’homme de la situation” (The Man for the Job). This was seen as an attempt by Plante to subvert expectations regarding gender and political leadership as well as to highlight her irreverent sense of humour. In terms of policy, both Plante and Coderre made campaign pledges to ensure gender parity on their respective executive committees (CBC, 2017a). So, both elections had a woman running for mayor. Besides that, both cities saw a slight rise in the number of women seeking seats on the City Council. In fact, 31% of the 12,924 candidates seeking election across the province of Quebec were women, up from 25% in 2005 (CBC, 2017b). Nearly 400 of those women were running for mayor in various cities. In Montreal, for instance, 43% of all candidates were women, although neither major party had achieved gender parity in their slate of candidates. Given the greater visibility of women candidates during these elections, gender may have been cued to voters, even if only subtly. In other words,

42  Scott Pruysers

voters may have been primed to consider gender, given the visibility of women candidates, even though gender was not a central campaign theme. Gender was not a prominent campaign theme; that said, both cities had a strong female contender for mayor, and this allows us to consider whether, and in what circumstances, gender was an important factor in voters’ decision-making calculus. Building on the work of Bird and colleagues (2016) and their analysis of gender and ethnic affinity voting in the 2014 Toronto municipal election, the rest of this chapter considers the following four research questions: 1 Is there evidence of gender affinity voting in Montreal and Quebec City? That is, did women voters disproportionately support women mayoral candidates (i.e., Plante in Montreal and Guérette in Quebec City)? 2 What, if anything, is the role of attitudes toward gender, and women more generally, in voting for women candidates? Are voters who hold positive views of feminists and who are concerned about the underrepresentation of women more likely to support women mayoral candidates? 3 In addition to attitudes toward gender, is there a relationship ­between political resources (such as political knowledge and campaign attention) and the propensity to engage in gender affinity voting? 4 Finally, how does voter sex interact with views about gender and with political resources such as knowledge? Data and Methods Data for this chapter have been drawn from the CMES voter surveys conducted in Montreal and Quebec City. Those surveys include 1,148 individuals who voted in the 2017 Quebec City mayoral election and 738 who voted in the Montreal mayoral election.6 The dependent variable used in the various regression models is respondent vote choice. In Montreal, where there were only two main contenders, the models use a binary logistic regression analysis where voting for a woman (Plante) = 1, and voting for a man (Coderre) = 0. In Quebec City, where there are three main mayoral contenders, I focus my analysis on the extent to which women voters supported the lone woman mayoral candidate in the race. As such, voting for a woman (Guérette) = 1, whereas voting for a man (Gosselin or Labeaume) = 0. Focusing on Guérette versus

6 Excluded from the analysis are voters of minor mayoral candidates.

Gender Affinity Voting  43

others in Quebec City allows us to consider specifically whether women voters were disproportionately supporting the woman in the race.7 The primary explanatory variable explored throughout the chapter is voter sex (female = 1; male = 0). Besides sex, the empirical analysis in the following section includes various other explanatory and control variables. Consistent with Bird and colleagues (2016), I include a number of variables tapping into attitudes toward feminism, women’s democratic representation, and women more broadly. This is operationalized in three ways. First, I ­include a 100-point feeling thermometer question regarding attitudes toward feminists. Second, I include responses to an agree/disagree question about the composition of the Municipal Council: “The gender composition of City Council should reflect the gender composition of the population. (i.e., roughly 50% of the population is female, so 50% of councillors should also be female).” Third, I include a measure of respondent “gender ideology” or “gender consciousness.” Following Goodyear-Grant and Croskill (2011), I utilize agree/disagree responses to the following statement: “Society would be better off if more women stayed home with their children” (reverse coded). Broadly speaking, these measures were meant to tap into attitudes toward the role of women in public life. As Goodyear-Grant and Croskill (2011: 235), write, “low levels of knowledge and low attention to the election are the key precursors for women voters’ use of candidate gender as a shortcut to the vote.” That being so, I include a measure of political knowledge. This is measured using a brief battery of standard political knowledge questions: identifying the name of the provincial premier, the name of the city’s mayor, the minister of finance at the federal level, and the number of City Council seats. Using these questions, a total score for each respondent is established ranging from 0 (no correct answers) to 4 (perfect). In addition to knowledge, I include a political attention variable that utilizes responses to the following question: “On a scale of 0–10, how much attention did you pay to the mayoral election campaign?” Political knowledge, campaign attention, and the three gender attitudinal variables are then each interacted with respondent sex and the interactions included in the analysis. This allows us to push the analysis further and explore whether these attitudes toward gender interact with one’s own identity to shape support for women candidates.

7 Though not central to the analysis, the appendix does include models for Gosselin and Labeaume support, and a brief summary of the results is highlighted in the text.

44  Scott Pruysers

In addition to the above explanatory variables, the analysis includes a series of standard socio-demographic controls. These include age, income, and education. Note that I do not include a variable for municipal party identification, even though Montreal and Quebec City are unique among Canadian cities insofar as they are among the few that have municipal political parties. Municipal party identification is left out, in large part, due to the high levels of personalization or leader-centric politics at the municipal level.8 It would be difficult, for instance, to disentangle feelings toward Denis Coderre and his party, Équipe Denis Coderre. Included instead of party identification is respondent ideology as captured by self-placement on the left/right scale.9 Results A brief look at open-ended text responses suggests that gender may indeed have played a role in these two municipal elections. When asked what, if anything, respondents liked about Plante, for instance, one f­ emale voter responded as follows: “It may be a silly reason, but I like that she is a female candidate in this election.” Specifically mentioning Plante’s sex as something that voters liked about her was not completely uncommon. Other voters wrote “une femme pour du changement” (a woman for change) and “une femme comme maire” (a woman as mayor) in ­response to this question. Around 7% of respondents mentioned ­Plante’s sex as one of the things they most liked about her. There is, of course, also the flipside. Some respondents included Plante’s appearance, the sound of her voice, and her laugh as things they disliked about her. This is consistent with research suggesting that women candidates and elected officials often face extra scrutiny regarding these kinds of issues (see, for instance, Bashevkin, 2009; Lawless, 2009; Trimble, 2007).10

8 For a discussion of party identification and the use of party labels in municipal elections, see Anderson and McGregor (chapter 7, this volume) and Bélanger and Stephenson (chapter 6, this volume). 9 Ideology is measured with an 11-point scale where 0 is left and 10 is right; income is measured with nine categories ranging from less than $25,000 to $200,000+; and education is measured with eleven categories ranging from no formal education to a doctorate. 10 Studying the 2004 Canadian Conservative leadership election, for example, Trimble (2007) finds that one in three news stories mentioned Belinda Stronach's appearance compared to only 2% of news stories that referenced Stephen Harper's appearance. More importantly, Stronach's appearance was the first thing mentioned about her in fully one half of these news stories – before her qualifications, policy positions, ­background, and so on.

Gender Affinity Voting  45

A similar pattern exists for Guérette: in their responses regarding what they liked about her, over 5% of voters in our sample mentioned that she was a woman. One respondent, for instance, wrote that Guérette was a “bon modèle de femme en politique” (good model of women in politics). Others simply liked that she was a woman: “simplement qu’elle est une femme candidate!” (simply that she is a female candidate!). Other respondents mentioned the general lack of diversity in electoral politics: “c’est une femme et il y peu de femmes en politique” (she is a woman and there are few women in politics”) or “j’aime qu’il y ait une femme dans le paysage politique” (I like that there is a woman in the political landscape). These responses from both Montreal and Quebec City suggest that voters were at least thinking about candidate sex when considering whom to vote for. The next question, of course, is whether any of this mattered when it came time to casting a ballot. In other words, if voters were thinking about candidate sex and gender, what does the reported vote choice tell us about a possible gender affinity effect? The analysis begins with some descriptive bivariate data regarding voter sex and vote choice in Montreal and Quebec City. Table 2.1 breaks down vote choice by men and women in Montreal. It is immediately obvious that women did not vote for Plante at significantly higher rates than men. Indeed, 63.2% of men voted for Plante compared to 62.7% of women.11 The bivariate results, therefore, provide no support for a gender affinity effect in this particular election. Table 2.2 provides the same breakdown in Quebec City. Similar to Montreal, women did not vote for Guérette at considerably higher rates than men. Indeed, there is no significant difference between the rate at which women supported Guérette over Labeaume. A 13-point gap does, however, separate women and men in their support for Gosselin (38% vs 25%). Thus, while there is very little difference between men and women in their support for Guérette and Labeaume, the real gender difference in Quebec City appears to be in support for Gosselin compared to Guérette. As revealed in the introduction to this volume, women tended to vote against Gosselin rather than for Guérette. 11 Note that the reported vote choice in the survey is not entirely reflective of the actual election results. In Montreal, Plante won 51% of the vote compared to Coderre's 46%. Our data therefore include more Plante supporters and fewer Coderre supporters than the actual results would suggest. In Quebec City, the actual results were as follows: Labeaume 55%; Gosselin 28%; Guérette 15%. The data therefore overestimate support for Guérette and underestimate support for Labeaume.

46  Scott Pruysers Table 2.1.  Voter sex and vote choice in Montreal (%) Men

Women

Coderre Plante

36.8 63.2

37.3 62.7

N

370

351

Table 2.2.  Voter sex and vote choice in Quebec City (%) Men

Women

Gosselin Guérette Labeaume

37.8 17.4 44.6

24.4 23.1 52.5

N

631

507

Moving to the multivariate analysis, the results include three separate models and report marginal effects. Model 1 includes voter sex, age, education, income, and self-placement on the left/right scale. Model 2, by contrast, includes these same variables but adds attention to the election, political knowledge, and three variables that tap into attitudes toward women (feminist feeling thermometer, attitudes toward women’s political representation, and gender ideology). Finally, Model 3 includes all of the variables in Models 1 and 2 along with a series of interactions between the attitudinal and attention variables and respondent sex. We start the analysis in Montreal. Beginning with Model 1 in Table 2.3, there is no evidence that women were more likely than men to vote for Plante. The results do suggest that older and more right-wing voters were less likely to support Plante. Model 2 reveals that those voters who wanted to see more women on council were significantly more likely to vote for Plante. Attitudes regarding gender are therefore relevant. Again, however, voter sex does not appear to be related to vote choice, though age and right-wing ideology remain significant and negative. Model 3, which adds the interaction variables, reveals an interesting finding. Here we find further evidence of a gender affinity effect. Women who agree that City Council should better reflect the gender composition of the broader society (i.e., those who want more women on council) were much more likely to support Plante. While attitudes toward gender representation on City Council had an independent main effect in Model 2, the interaction in Model 3 is both significant and substantive. Women who provide

Gender Affinity Voting  47 Table 2.3.  Binary logistic regression results (voting for a woman in Montreal – Plante vs Coderre)

Sex (female) Age Education Income Ideology Attention Political knowledge Attitudes toward feminists Underrepresentation of women Gender ideology Female X attention Female X political knowledge Female X attitudes toward feminists Female X underrepresentation Female X gender ideology 2

Pseudo R

Model 1

Model 2

Model 3

–0.004 (0.036) –0.004 (0.001)*** –0.003 (0.003) –0.000 (0.001) –0.002 (0.001)*** – – – –

–0.013 (0.037) –0.004 (0.001)*** –0.003 (0.003) –0.000 (0.001) –0.002 (0.001)*** 0.005 (0.004) –0.019 (0.027) 0.000 (0.000) 0.138 (0.049)***

–0.277 (0.227) –0.004 (0.001)*** 0.000 (0.005) –0.000 (0.000) –0.002 (0.001)** 0.024 (0.012)** –0.006 (0.039) 0.000 (0.000) –0.001 (0.064)

– – – –

–0.017 (0.060) – – –

–0.052 (0.076) –0.002 (0.018) –0.063 (0.055) 0.002 (0.001)*





0.324 (0.105)***





0.060 (0.130)

0.020

0.028

0.055

Note: Marginal effects reported. Standard errors in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01

positive assessments of feminists were also significantly more likely to support Plante. Interestingly, the notion that women might tend to support other women due to a lack of political resources (knowledge and attention) is not borne out by the data. Moving now to Quebec City, Table 2.4 shows the results for a binary logistic regression where the dependent variable is voting for the one woman in the election (Guérette). For Quebec City, this is the most relevant analysis for this chapter as it reveals the extent to which women supported the lone female mayoral candidate. Model 1 reveals that her voters were more likely to be women, older, and educated. This, of course, provides some preliminary evidence of a gender affinity effect at play. Model 2 shows that a number of attitudinal variables related to gender are significantly related to her support as well. Those who think that ­Municipal Council should be more representative, those who are supportive of feminists, and those who have a more inclusive gender ideology, are more likely to support Guérette. The main effect of gender, however, is no longer significant once the attitudinal variables are introduced. F ­ inally, Model 3 provides further evidence of attitudinal correlates: individuals

48  Scott Pruysers Table 2.4.  Binary logistic regression results (voting for a woman in Quebec City – Guérette vs other)

Sex (female) Age Education Income Ideology Attention Knowledge Attitudes toward feminists Underrepresentation of women Gender ideology Female X attention Female X knowledge Female X attitudes toward feminists Female X underrepresentation Female X gender ideology Pseudo R2

Model 1

Model 2

Model 3

0.061 (0.024)*** 0.004 (0.001)*** 0.003 (0.001)* –0.000 (0.001) –0.001 (0.001)* – – – –

0.027 (0.025) 0.002 (0.001)** 0.003 (0.001)** –0.000 (0.001) –0.001 (0.001) –0.011 (0.006)* 0.015 (0.014) 0.001 (0.000)*** 0.206 (0.032)***

–0.185 (0.161) 0.002 (0.001)** 0.002 (0.001)* –0.00 (0.001) –0.001 (0.001) –0.016 (0.008)** 0.016 (0.021) 0.000 (0.000) 0.219 (0.040)***

– – – –

0.120 (0.053)* – – –

0.021 (0.063) 0.007 (0.012) –0.013 (0.029) 0.003 (0.001)***





–0.094 (0.068)

– 0.028

– 0.094

0.155 (0.110) 0.117

Note: Marginal effects reported. Standard errors in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01

concerned with representation on City Council were significantly likely to support Guérette. This is an interesting contrast to Montreal, where it is the interaction effect, not the main effect, that is significant. Consistent with the support for Plante in Montreal, women who are supportive of feminists were more likely to vote for Guérette. As in Montreal, gender clearly played a role in the Quebec City mayoral election. While not the primary focus of the chapter, Tables A2.1 and A2.2 in this chapter’s appendix show similar results for Gosselin and Labeaume. Model 1 of Table A2.1 reveals that women were significantly less likely to vote for Gosselin. Voter sex continues to be significantly (and negatively) related to voting for Gosselin in Model 2, where the attention and attitudinal variables are included. Here we can see that voters who were ­attentive and knowledgeable were more likely to vote for Gosselin than for his competitors. Additionally, those voters who had more positive views toward feminists and a more inclusive gender ideology, as well as those who wanted to see more women on City Council, were significantly less likely to be Gosselin supporters. Both gender itself and attitudes

Gender Affinity Voting  49

t­oward gender are therefore important. In Model 3, the main effect of gender is no longer significant. However, the data reveal that a certain subset of women, those who are supportive of feminists, are significantly less likely to be Gosselin supporters. Finally, Table A2.2 reports the results for Labeaume supporters. Like Guérette voters, those who voted for Labeaume tended to be older and female (Model 1). Also, the results in Model 2 are largely consistent between Labeaume and Guérette. Here, again, we see strong evidence of attitudinal correlates, in terms of both gender representation on council and the role of women more generally. The results in Model 3 continue to confirm the importance of attitudes toward women in explaining Labeaume’s support, though it is worth noting that none of the interactions are significant. Discussion and Conclusion Exploring gender affinity voting in Montreal and Quebec City has produced a number of interesting results. In both cities, a woman ran for mayor: Valérie Plante and her party Projet Montréal won the election with 51% of the vote in Montreal, whereas Anne Guérette and her party Démocratie Québec came in third place in Quebec City with 15% of the vote. The evidence of a direct, unmediated, gender affinity effect in both cities is rather limited. The bivariate data, for example, show that women did not vote for women candidates disproportionately (Tables 2.1 and 2.2). This is consistent with the findings of recent studies of gender affinity voting in Canada (see Bird et al., 2016; Goodyear-Grant, 2010; GoodyearGrant & Croskill, 2011; McGregor et al., 2017). This study therefore adds to the literature on gender affinity in municipal elections, providing two additional cities, as well as providing further evidence against the presence of a strong and unmediated gender affinity effect. This, of course, is not to say that gender and voter sex was unimportant or that there is no evidence of gender affinity. Indeed, gender does appear to have mattered in both elections in a number of other ways. In both Montreal and Quebec City, positive attitudes toward feminists and a desire to see more women elected to political office were related to voting for a woman mayoral candidate. In Montreal, this manifests as an interaction with voter sex, as women who hold these views were significantly more likely to vote for Plante. In Quebec City, by contrast, the main effect is significant, revealing that respondents who agreed that council should have more women were more likely to vote for Guérette. A desire to see more descriptive gender representation in democratically

50  Scott Pruysers

elected bodies therefore appears to be an important part of voting for women candidates. In both cities, women who view feminists favourably were more likely to vote for a woman mayoral candidate. So we can conclude that gender and attitudes about gender were both relevant factors in each election. Though outside the scope of this chapter, it is curious that men who view feminists favourably behave differently than women with the same views. It may be that the cognitive and affective components of gender identity make voting for a woman more salient (and even urgent) for women voters compared to their male counterparts. Thus, while men may view feminists favourably and agree that council should be more diverse, they do not have the same feelings of connectedness to female candidates that women do. Views about the role of women in society were related to vote choice as well. Individuals who believed that society would be better off if women stayed at home to raise children were significantly more likely to vote for Gosselin. Likewise, women who held positive views of feminists were significantly less likely to vote for Gosselin and significantly more likely to vote for Guérette. This, of course, is an important reminder not to treat groups as completely homogeneous. Not all women voted the same, and as the data reveal, attitudes were an important moderating factor in the relationship between gender and vote choice. While there is ­limited ­evidence of a direct and unmediated gender affinity effect, gender clearly mattered in both mayoral elections. As Dolan (1998, 2001) reminds us, however, gender can be more or less important depending on the electoral context. In campaigns where gender is an important theme or where gender has been politicized, voters can be cued to pay greater attention to questions of gender, including candidate gender. Under such circumstances we may be more likely to find evidence of an unmediated gender affinity effect. Another aspect of the electoral context may be the degree of competitiveness. In relatively uncompetitive elections where there is a clear front-runner who is expected to win, the sex of challenging candidates may become a more salient factor for voters. Future studies, therefore, should consider examining gender affinity effects from a comparative perspective, examining the phenomenon during elections where gender is and is not a salient campaign issue, as well as in competitive and uncompetitive elections.

Gender Affinity Voting  51

Appendix to Chapter 2 Table A2.1.  Binary logistic regression results (vote choice in Quebec City – Gosselin vs other)

Sex (female) Age Education Income Ideology Attention Knowledge Attitudes toward feminists Underrepresentation of women Gender ideology Female X attention Female X knowledge Female X attitudes toward feminists Female X underrepresentation Female X gender ideology Pseudo R2

Model 1

Model 2

Model 3

–0.138 (0.027)*** –0.005 (0.001)*** –0.002 (0.002) –0.000 (0.001) 0.000 (0.001)

–0.059 (0.028)** –0.003 (0.001)*** –0.004 (0.003) 0.000 (0.001) 0.001 (0.001) 0.011 (0.006)* 0.029 (0.016)* –0.001 (0.000)** –0.224 (0.024)*** –0.157 (0.041)***

0.171 (0.147) –0.003 (0.001)*** –0.003 (0.003) 0.000 (0.001) 0.000 (0.001) 0.016 (0.009)* 0.048 (0.023)** –0.000 (0.000) –0.246 (0.032)*** –0.148 (0.058)** –0.007 (0.013) –0.041 (0.034) –0.002 (0.001)** 0.075 (0.058) –0.003 (0.085)

0.040

0.124

0.125

Note: Marginal effects reported. Standard errors in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01

Table A2.2.  Binary logistic regression results (vote choice in Quebec City – Labeaume vs other) Model 1 Sex (female) Age Education Income Ideology Attention Knowledge Attitudes toward feminists Underrepresentation of women Gender ideology Female X attention Female X knowledge Female X attitudes toward feminists Female X underrepresentation Female X gender ideology Pseudo R2

Model 2

0.075 (0.030)** 0.030 (0.032) 0.002 (0.001)* 0.002 (0.001) –0.001 (0.002) –0.000 (0.002) 0.000 (0.001) 0.001 (0.001) 0.001 (0.001) 0.000 (0.001) –0.001 (0.001) –0.046 (0.018)** 0.000 (0.000) 0.090 (0.033)*** 0.098 (0.052)*

0.007

0.019

Note: Marginal effects reported. Standard errors in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01

Model 3 0.054 (0.173) 0.002 (0.001) 0.000 (0.002) 0.000 (0.001) 0.000 (0.001) –0.002 (0.010) –0.065 (0.026)** 0.000 (0.000) 0.112 (0.042)*** 0.163 (0.076)** 0.001 (0.015) 0.047 (0.038) –0.010 (0.001) –0.022 (0.070) –0.084 (0.107) 0.024

52  Scott Pruysers REFERENCES Aalberg, T., & Jenssen, A. T. (2007). Gender stereotyping of political candidates: An experimental study of political communication. Nordicom Review 28(1), 17–32. http://dx.doi.org/10.1515/nor-2017-0198 Anzia, S. F., & Berry, C. R. (2011). The Jackie (and Jill) Robinson effect: Why do congresswomen outperform congressmen? American Journal of Political Science 55(3), 478–493. https://doi.org/10.1111/j.1540-5907.2011.00512.x Bashevkin, S. (2009). Women, power, politics: The hidden story of Canada’s unfinished democracy. Oxford University Press. Bauer, N. (2020). Shifting standards: How voters evaluate the qualifications of female and male candidates. Journal of Politics, 82(1), 1–12. http://dx.doi.org /10.1086/705817. Online First. Besco, R. (2015). Rainbow coalition or inter-minority conflict? Racial affinity and diverse minority voters. Canadian Journal of Political Science, 48(2), 305–328. https://dx.doi.org/10.2139/ssrn.2325159 Bird, K. (2016). Understanding the local diversity gap: Supply and demand of visible minority candidates in Ontario municipal politics” (pp. 180–220). In A. Bilodeau (Ed.), Just ordinary citizens? Towards a comparative portrait of the political immigrant. University of Toronto Press. Bird, K., Jackson, S., McGregor, M., Moore. A., & Stephenson, L. (2016). Sex (and ethnicity) in the city: Affinity voting in the 2014 Toronto mayoral election. Canadian Journal of Political Science, 49(2), 359–383. https://doi.org /10.1017/S0008423916000536 Bittner, A. (2011). Platform or personality? The role of party leaders in elections. Oxford University Press. Blais, A., Gidengil, E., Dobrzynska, A., Nevitte, N., & Nadeau, R. (2003). Does the local candidate matter? Candidate effects in the Canadian election of 2000. Canadian Journal of Political Science 36(3), 657–664. https://doi.org /10.1017/S0008423903778810 Blais, A. E., Gidengil, R. N., & Nevitte, N. (2002). Anatomy of a Liberal victory: Making sense of the vote in the 2000 Canadian election. Broadview Press. Brians, C. L. (2005). Women for women? Gender and party bias in voting for female candidates. American Politics Research, 33(3), 357–375. http://dx.doi.org /10.1177/1532673X04269415 Butler, D., & Preece, J. (2016). Recruitment and perceptions of gender bias in party leader support. Political Research Quarterly, 69(4), 842–851. http:// dx.doi.org/10.1177/1065912916668412 CBC. (2017a, 20 November). Gender parity but little diversity in Valérie Plante’s executive committee. CBC News. https://www.cbc.ca/news/canada /montreal/mayor-valérie-plante-executive-committee-1.4409738

Gender Affinity Voting  53 CBC. (2017b, 5 November). Voters cast ballots in municipal elections across Quebec. CBC News. https://www.cbc.ca/news/canada/montreal/quebec -votes-today-2017-1.4387967 Cross, W., & Young, L. (2013). Candidate recruitment in Canada: The role of political parties. In R. Koop & A. Bittner (Eds.), Parties, elections, and the future of Canadian politics (pp. 24–45). UBC Press. Cutler, F. (2002). The simplest shortcut of all: Sociodemographic characteristics and electoral choice. Journal of Politics, 64(2), 466–490. http://dx.doi.org /10.1111/1468-2508.00135 Cutler, F., & and Matthews, J. S. (2005). The challenge of municipal voting: Vancouver 2002. Canadian Journal of Political Science, 38(2), 359–382. https:// doi.org/10.1017/S0008423905040151 Delli Carpini, M. X., & Keeter, S. (1996). What Americans know about politics and why it matters. Yale University Press. Dolan, K. (1998). Voting for women in the “Year of the Woman.” American Journal of Political Science, 42(1), 272–293. https://doi.org/10.2307/2991756 Dolan, K. (2001). Electoral context, issues, and voting for women in the 1990s. Women and Politics, 23(1–2), 21–36. https://doi.org/10.1300/J014v23n01_03 Dolan, K. (2008). Is there a “gender affinity effect” in American politics? Information, affect, and candidate sex in US House elections. Political Research Quarterly, 61(1), 79–89. https://doi.org/10.1177/1065912907307518 Fulton, S. (2012). Running backwards and in high heels: The gendered quality gap and incumbent electoral success. Political Research Quarterly, 65(2), 303–314. https://doi.org/10.1177%2F1065912911401419 Gidengil, E. (1992). Canada votes: A quarter century of Canadian national election studies. Canadian Journal of Political Science, 25(2), 219–248. https:// doi.org/10.1017/S0008423900003966 Gidengil, E., Nevitte, N., Blais, A., Everitt, J., & Fournier, P. (2012). Dominance and decline: Making sense of recent Canadian elections. University of Toronto Press. Goodyear-Grant, E. (2010). Who votes for women candidates and why? In C. D. Anderson & L. B. Stephenson (Eds.), Voting behaviour in Canada. (pp 43–64). University of British Columbia Press. Goodyear-Grant, E., & Croskill, J. (2011). Gender affinity effects in vote choice in Westminster systems: Assessing “flexible” voters in Canada. Politics and Gender, 7(2): 223–250. https://doi.org/10.1017/S1743923X11000079 Goodyear-Grant, E., & Tolly, E. (2017). Voting for one’s own: Racial group identification and candidate preferences. Politics, Groups, and Identities, 7(1), 131–147. https://doi.org/10.1080/21565503.2017.1338970 Huddy, L., & Terkildsen, N. (1993). Gender stereotypes and the perception of male and female candidates. American Journal of Political Science, 37(1), 119–147. http://dx.doi.org/10.2307/2111526

54  Scott Pruysers Inter-Parliamentary Union. (2018). Women in national parliaments. http://archive .ipu.org/wmn-e/classif.htm Kanji, M., Bilodeau, A., & Scotto, T. (Eds.). (2012). The Canadian election studies: Assessing four decades of influence. UBC Press. King, D., & Matland, R. (2003). Sex and the Grand Old Party: An experimental investigation of the effect of candidate sex on support for a Republican candidate. American Politics Research, 31(6): 595–612. http://dx.doi.org/10.1177 /1532673X03255286 Lawless, J. L. (2009). Sexism and gender bias in election 2008: A more complex path for women in politics. Politics and Gender, 5(1), 70–80. Lawless, J. L., & Fox, R. L. 2005. It takes a candidate. Cambridge University Press. Matson, M., & Fine, T. S. (2006). Gender, ethnicity, and ballot information: Ballot cues in lowiInformation elections. State politics and policy quarterly, 6(1), 49–72. https://doi.org/10.1177/153244000600600103 McDermott, M. L. (1998). Race and gender cues in low-information elections. Political Research Quarterly, 51(4), 895–918. https://doi.org/10.1177%2F 106591299805100403 McGregor, R. M., Moore, A., Jackson, S., Bird, K., & Stephenson, L. (2017). Why so few women and minorities in local politics? Incumbency and affinity voting in low information elections. Representation, 53(2), 135–152. https:// doi.org/10.1080/00344893.2017.1354909 Mondak, J., & Anderson. M. (2004). The knowledge gap: A reexamination of gender-based differences in political knowledge. Journal of Politics, 66(2), 492–512. http://dx.doi.org/10.1111/j.1468-2508.2004.00161.x Pearson, K., & McGhee, E. (2013). What it takes to win: Questioning “gender neutral” outcomes in US House elections. Politics and Gender, 9(4), 439–462. https://doi.org/10.1017/S1743923X13000433 Plutzer, E., & Zipp, J. F. (1996). Identity politics, partisanship, and voting for women candidates. Public Opinion Quarterly, 60(1), 30–57. https://doi.org /10.1086/297738 Preece, J., & Stoddard, O. (2015). Does the message matter? A field experiment on political party recruitment. Journal of Experimental Political Science, 2(1), 26–35. https://doi.org/10.1017/xps.2014.17 Pruysers, S., and Blais, J. (2014). Anything women can do men can do better: An experiment examining the effects of stereotype threat on political knowledge and efficacy. Social Science Journal, 51(3), 341–349. http://dx.doi .org/10.1016/j.soscij.2014.05.005 Pruysers, S., & Blais, J. (2017). Why won’t Lola run? An experiment examining tereotype Thtreat and political ambition. Politics and Gender, 13(2), 232–252. https://doi.org/10.1017/S1743923X16000544 Pruysers, S., & Blais, J. (2018a). A little encouragement goes a (not so) long way: An experiment to increase political ambition among young women.

Gender Affinity Voting  55 Journal of Women, Politics, and Policy, 9(3), 384–395. https://doi.org/10.1080 /1554477X.2018.1475793 Pruysers, S., & Blais, J. (2018b). Narcissistic women and cash-strapped men: Who can be encouraged to consider running for office and who should do the encouraging? Political Research Quarterly, 72(2). http://dx.doi.org/10.1177 /1065912918786040 Pruysers, S., & Cross, W. (2016). Candidate selection in Canada: Local autonomy, centralization, and competing democratic norms. American Behavioral Scientist, 60(7), 781–798. http://dx.doi.org/10.1177/0002764216632820 Sanbonmatsu, K. (2002). Gender stereotypes and vote choice. American Journal of Political Science, 46(1), 20–34. https://doi.org/10.2307/3088412 Sanbonmatsu, K. (2006). Do parties know that “women win”? Party leader beliefs about women’s electoral chances. Politics and Gender, 2(4), 431–450. http://dx.doi.org/10.1017/S1743923X06060132 Sanbonmatsu, K., and Dolan, K. (2009). Do gender stereotypes transcend party? Political Research Quarterly, 62(3), 485–494. http://dx.doi.org/10.1177 /1065912908322416 Sevi, S., Arel-Bundock, V., & Blais, A. (2018). Do women get fewer votes? No. Canadian Journal of Political Science, 52(1), 201–210. https://doi.org/10.1017 /S0008423918000495 Smith, E., & Fox, R. L. (2001). The electoral fortunes of women candidates for Congress. Political Research Quarterly, 54(1), 205–221. http://dx.doi.org /10.1177/106591290105400111 Stolle, D., & Gidengil, E. (2010). What do women really know? A gendered analysis of varieties of political knowledge. Perspectives on Politics, 8(1), 93–109. https://doi.org/10.1017/S1537592709992684 Thomas, M., & Bittner, A. (Eds.). (2017). Mothers and others: The role of parenthood in politics. UBC Press. Thomas, M. & Bodet, M. A. (2013). Sacrificial lambs, women candidates, and district competitiveness in Canada. Electoral Studies, 32(1), 153–166. http:// dx.doi.org/10.1016/j.electstud.2012.12.001 Tolley, E. (2011). Do women “do better” in municipal politics? Electoral representation across three levels of government. Canadian Journal of Political Science, 44(3), 573–594. https://doi.org/10.1017/S0008423911000503 Trimble, L. (2007). Gender, political leadership, and media visibility: Globe and Mail coverage of Conservative party of Canada leadership contests. Canadian Journal of Political Science, 40(4), 969–993. Zipp, J. F., & Plutzer, E. (1985). Gender differences in voting for female candidates: Evidence from the 1982 election. Public Opinion Quarterly, 49(2), 179–197. https://doi.org/10.1086/268913

3 Language and (Not) Voting: 2017 Municipal Elections in Quebec cameron d . anderson and laura b . stephenson

A great deal of research examines how political behaviour is shaped by context. We, as individuals, do not make our choices in a vacuum; we are influenced by those around us and the circumstances in which we find ourselves. Early studies of vote choice (e.g., Lazarsfeld et al., 1948) found that our social characteristics are very important determinants of our vote choices. Others have found that our social interactions, or who we talk to about politics, have important implications for our attitudes and behaviours (Huckfeldt & Sprague, 1987; Huckfeldt, 1984; Huckfeldt et al., 1993, 1995; Zuckerman & Kroh 2006; Zuckerman et al., 1994, 2007). The types of behaviours affected include political learning, tolerance, discussion, and political activity. Knowing that social context matters, many scholars have considered whether the composition of a community is relevant. Community hetero­ geneity can take many forms relating to such things as partisan attitudes, social class, income, and ethnicity. The existing literature reveals that there is no simple answer to whether community heterogeneity has positive or negative effects on political engagement. The results suggest that both are possible, with important nuances related to the type and context of heter­ ogeneity. Indeed, while some suggest that diversity increases competition, which in turn can spur engagement (Cox & Munger, 1989; Blais, 2000), others who study social networks argue that the disagreement that comes with diversity can introduce feelings of ambivalence and social accounta­ bility concerns that can depress citizen involvement (Mutz, 2002; McClurg, 2006). The work on social capital (see, for example, Putnam, Leonardi, & Nanetti, 1993) posits a clearer directional hypothesis – that similar individu­ als coalesce, which in turn increases civic participation. However, examina­ tions of specific types of diversity, such as economic, ethnic, and linguistic, and of different measures of social capital (such as trust and political parti­ cipation), have produced a range of mixed results in the literature.

Language and (Not) Voting  57

In this chapter, we focus on understanding one type of heterogene­ ity (linguistic) and one type of behaviour (voting) in a single, diverse community: Montreal, Quebec. That city is ideal for this analysis. Lan­ guage issues are highly politicized and salient in Quebec, and linguistic diversity is especially notable in Montreal. We make use of census data on linguistic diversity and a municipal election survey to investigate the presence of mobilizing or marginalizing effects related to the linguistic heterogeneity of one’s neighbourhood. Our analysis reveals that linguis­ tic diversity matters, at the aggregate level, and that it does not have significantly different effects on different language groups. Given the importance of the language cleavage in Quebec, our results raise ques­ tions about how linguistic heterogeneity affects societies, or not. Theories of Heterogeneity and Political Participation Since our goal is to understand how the linguistic diversity of one’s neigh­ bourhood might affect voting behaviour, we ground our work in terms of three strands of research – the mobilization hypothesis, social network research, and studies of social capital. First, there is research in the field of political behaviour that considers how a heterogeneity of preferences (often expressed through party support) can mobilize individuals to become more involved in politics. This mobilization hypothesis can be traced to two different mechanisms. First, there is an expectation of com­ petition. In a fight for scarce resources, such as political power, greater heterogeneity means that groups are more evenly matched, which means there is a stronger chance of losing benefits if other groups mobilize more effectively. So in communities that are evenly divided in terms of party support, there is an incentive for people to turn out to vote when the outcome is likely to be close, because the chance of affecting the outcome is greater. There is also likely to be increased campaigning as groups escalate their efforts to mobilize members. In this way, heteroge­ neity can increase participation as campaign resources mobilize more members. This view of the mobilization hypothesis can be found in Cox and Munger (1989) and Blais (2000). The second mechanism has to do with increased information. In diverse communities there is a greater likelihood that people will be exposed to new (and potentially countervailing) information. When confronted with new ideas, people can be spurred to learn more, pay more attention, and get more involved. For example, Scheufele and colleagues (2004) contend that disagreement within one’s social net­ work increases political engagement by promoting political learning and the use of hard news media. DeSante and Perry (2016) note that

58  Cameron D. Anderson and Laura B. Stephenson

political knowledge increases, even among non-voters, when the size of the minority group increases. Moving to political activities, Harell and colleagues (2009) have found that political diversity fosters participation among young people; they hypothesize that this is related to increases in political knowledge, discussion, and interest that arise from a diverse social context. Similarly, in Belgium, Quintelier and colleagues (2012) found that diversity is related to an increase in the participation of young people. Another strand of research focuses on the potential negative effects of social networks. In this literature, some research claims that heter­ ogeneity can in fact decrease political participation, through two mech­ anisms: social desirability and ambivalence (Mutz, 2002, 2006). Social desirability affects one’s willingness to express political views in mixed company – that is, among those who do not necessarily agree. Ambiv­ alence is related to the absorption of competing information that can lead citizens to be uncertain about their political stances and, in turn, less likely to engage. Nir (2005) has demonstrated that individual-level ambivalence exerts an impact on political participation. Finally, also relevant to our study is the literature on social networks. Informed by Putnam and colleagues’ influential study of Italy (1993), a substantial line of research considers how a community’s composition can affect a host of outcomes at the aggregate and individual levels (such as economic development and citizenship behaviours; for a review of research, see Costa & Kahn, 2003). The logic behind this literature is that people tend to prefer to be with others like themselves. As Costa and Kahn (2003) note: Whether in choosing a college roommate, a residential community, or a place to pray, people tend to self-segregate. They prefer to interact with others like them because of shared interests, socialization to the same cul­ tural norms, and greater empathy toward individuals who remind them of themselves. Members of minority groups may prefer to interact with other minority members if they fear discrimination. The coordination necessary for groups to form and to become active is easier if members speak a com­ mon language. (104)

When people coalesce into communities, one specific type of social capital – bonding – is increased. There is also a bridging variant of social capital, one in which people from different groups work together and thereby reduce their differences (for a discussion of the different types and the organizations they apply to, see Coffé and Geys [2007]). For the most part, research has found that social heterogeneity decreases social

Language and (Not) Voting  59

capital (Coffé, 2009). Putnam (2007) suggests that ethnic diversity, at least in the short run, leads groups to separate from one another or “hunker down.” This is the marginalization hypothesis. The more iso­ lated groups are from one another, the less likely they are to engage in bridging behaviours. Often understood as trust among individuals, norms of reciprocity, and networks (Coffé, 2009), when social capital does exist it can increase political participation (Putnam, Leonardi, & Nanetti, 1993; Nakhaie, 2006). Voting is a social experience, which means that when neighbours and household members participate, they encourage participation among others (Buton et al., 2012). It has also been demonstrated that so­ cial pressure (or even threats of it) can increase voter turnout (Gerber et al., 2008). Heterogeneity in a society thus works against these impulses. Hill and Leighley (1999) suggest that the relationship between diversity and lower turnout in the United States is related to weaker mobilization. Anderson and Paskeviciute (2006) have found supportive evidence with data from 44 countries in terms of citizenship behaviours, both struc­ tural (political discussion and voluntary organization membership) and cognitive (political interest and interpersonal trust). Both are affected by heterogeneity, but with variation across measures. In particular, they have found that linguistic heterogeneity is related to more organiza­ tional involvement but less trust, and that these values further vary in established and weak democracies. Sandovici and Listhaug (2010) have found that voting is (negatively) affected by belonging to an ethnic or linguistic minority, but not by other citizenship behaviours. Similarly, Bellettini and colleagues (2016) found that ethnic and economic ine­ quality depressed neighbourhood turnout in the 2004 and 2009 local elections in Italy; they attribute this to weaker social cohesion. An important nuance to the idea that heterogeneity decreases social capital and therefore participation is that the density of the group is very important. Fieldhouse and Cutts (2008) found that the population density of Asians in the UK matters for turnout: density increases Asian turnout and decreases non-Asian turnout in the same community. Their work in 2010, comparing the US and the UK, similarly points to the importance of both diversity and density. They also note, however, that standard influences on turnout, such as socio-economic standing, re­ main relevant and need to be taken into account when looking at the effects of diversity. Marschall and Stolle (2004) further investigated the differential effects on majorities and minorities in communities. Look­ ing at blacks and whites, they found that social trust is created differently in the two groups: whites tend to develop less generalized trust in lowstatus neighbourhoods, whereas generalized trust increases for blacks

60  Cameron D. Anderson and Laura B. Stephenson

when they live in contexts that provide close interactions with a broad sample of people. The variation in findings in the literature, especially between groups, points to the need to look at the individual level to best understand the effects of heterogeneity. If some groups are mobilized and other margin­ alized, simply looking at aggregate turnout does not tell us much about the specific effects of heterogeneity. There is also the issue of related socio-economic factors. If a community is highly diverse but there is sig­ nificant poverty in one group, what looks like a negative effect of diver­ sity on participation could actually be an effect of resources (Fieldhouse & Cutts, 2010). Geys (2006) conducted a meta-analysis of turnout studies and found significant variation in the results for heterogeneity. While he was able to point to a finding “on balance,” by no means did the studies all agree and looking at the big picture would have missed important variation. The research discussed above, and indeed the study of the effects of heterogeneity in the greater literature, has encompassed many countries and groups. In this chapter, we add to this body of literature by explor­ ing the effects of linguistic heterogeneity in Quebec, the only predom­ inantly French-speaking province in Canada, a bilingual country, and in particular in the linguistically diverse city of Montreal. As the work reviewed indicates, there are contrasting theories and findings regard­ ing these effects in terms of both direction and mechanism. Cognizant of this disagreement, we proceed in a more exploratory fashion to con­ sider how linguistic heterogeneity may influence turnout in municipal electoral contests in Montreal. Language Diversity in Quebec Diversity is an important issue in Quebec. Most importantly, as a “distinct society” and a “nation within a united Canada” (the latter declared in a parliamentary motion by former Prime Minister Stephen Harper in 2006), Quebec has a unique imperative to be concerned about heteroge­ neity in society. To protect the Québécois culture, the provincial govern­ ment (especially since the Quiet Revolution of the 1960s) has attempted to maintain control over non-French influences. This became clear with the passing of Bill 101 in 1977, which made French the official language of the province and banned signs in any language other than French. The prevailing view is that the French culture, including especially the French language, needs to be preserved, and this has influenced opin­ ions about immigration. Turgeon and Bilodeau (2014) found support for a “linguistic insecurity hypothesis” – that is, the feeling that French is

Language and (Not) Voting  61

threatened is associated with Quebeckers preferring fewer immigrants. Similarly, a 2017 CROP poll found that 67% of Quebeckers are in favour of new immigrants adopting Canadian culture, compared to 60% of Ca­ nadians in the rest of the country. Attitudes toward Muslims are similarly divided (Hinkson & Laframboise, 2017). Given the salience of language for the French culture, the degree of linguistic heterogeneity in a community is likely to be a key factor in political participation for many of the reasons discussed above. In par­ ticular, one can imagine that social capital/cohesion is likely to be af­ fected by the presence of large groups of non-francophones, with the result that linguistic diversity is likely to negatively affect participation in Quebec. Because of the relatively small size of the anglophone and allophone communities, and the substantial bilingual population, it is doubtful that there is any real campaign activity related to mobilizing non-francophones, so the mobilization hypothesis may be questioned. The province of Quebec is strongly francophone – 79% of Quebeckers report French as their mother tongue, and 94.5% know either French or both French and English. Only 4.6% report knowing only English, and 0.9% neither English nor French (Statistics Canada, 2017). In Mon­ treal specifically, 7.1% know only English, and 1.6% neither English nor French, and 55.1% have knowledge of both official languages (Statistics Canada, 2017). However, the Montreal area has many anglophone Que­ beckers (many with very deep roots in the province). As noted earlier in the work of Fieldhouse and Cutts (2008, 2010) and Marschall and Stolle (2004), the behaviour of the minority and majority lin­ guistic groups may vary. We can expect, then, that local language density is extremely important when considering the behaviour of minority language groups. Given that there are many non-francophones in Montreal (and that anglophones have a long history there), opportunities for community or­ ganizations and social networking among anglophones or allophones (but especially anglophones) are significant. Therefore, we hypothesize there may be a mobilization effect on the voting behaviour of non-francophones in Montreal. Regarding francophone populations in particular, the litera­ ture suggests that the majority group in a community with high heteroge­ neity often has weaker participation. In Montreal, we expect to find that francophones who perceive diversity are less likely to vote. To summarize, the hypotheses we investigate are as follows: H1. Linguistic diversity will increase turnout through mobilization and competitive dynamics. H2. Linguistic diversity will decrease turnout because it tends to undermine the development of social capital.

62  Cameron D. Anderson and Laura B. Stephenson H3. Minority language group members will be more likely to vote as they perceive greater linguistic diversity. H4. Francophones will be less likely to vote as they perceive greater linguistic diversity.

Data and Methods The data we use in this chapter come from two sources: the Canadian Municipal Election Study (CMES) and the 2016 Canadian Census. All individual-level data come from the CMES. Because the CMES asked respondents to list the first three digits of their postal code, we were able to merge Canadian Census data by Forward Sortation Area (FSA) to provide objective measures of neighbourhood linguistic diversity for each respondent. Our main variable of interest in this chapter is linguistic diversity. There are benefits and disadvantages to using either objective or subjec­ tive measures of linguistic diversity, and we are in the fortunate position of having data for both kinds of measures. Using perceptions of diversity would be in keeping with McDonald and Tolbert (2012), who found that perceptions of competitiveness affected turnout behaviour more than factual information. However, people may not be aware of what contex­ tual information shapes their behaviours and beliefs, which means that objective data can be a better measure of real-life experience. Therefore, we consider both types of measures in our analyses. Our objective measure of linguistic diversity comes from Canadian Census data for each respondent’s FSA. For each valid FSA reported in our individual-level sample, we coded the proportion of individuals who are francophone, anglophone and allophone. From this, we calculated a measure of linguistic diversity using the Herfindahl Index of Dissim­ ilarity or ethnic fractionalization.1 This produced values for each FSA between 0 and a theoretical value of 1, with smaller (larger) values in­ dicating less (more) linguistic diversity.2 The index scores ranged from 0.06 to 0.70 with an FSA average of 0.49. 1 Formally, this index is written: Herfindahl Index of Dissimilarityj = 1–∑ne=1 s2ej. The fractionalization index in a given FSA measures the possibility that two randomly cho­ sen people in that FSA are from a different linguistic group. Applied to our purposes in this chapter, the calculation for each FSA of the Index of dissimilarity looks like this (where the proportion of linguistic members from each FSA is squared): 1 – (∑ (Francophone)2 + (Anglophone)2 + (Allophone)2). 2 Technically, the upper limit of the Herfindahl Index is defined as 1 – (1/n). Thus, the upper limit here equals 0.67 (and can be a little higher if the proportions do

Language and (Not) Voting  63

The CMES survey asked a relatively fine-grained subjective question about linguistic diversity in one’s neighbourhood. That question read: “With respect to the following characteristics, how diverse would you say your neighbourhood is? Linguistically (very diverse, fairly diverse, not very diverse, not at all diverse, don’t know/prefer not to answer).” We also have measures of respondent language. As we expected dif­ ferent reactions to linguistic diversity depending on the individual’s own linguistic group, we created variables for francophones, anglophones, and allophones from first responses to the question: “What is the first language you learned and still understand?” Our dependent variable is turnout. We looked at turnout for the may­ oral, City Council, and borough mayoral elections, as individuals did not have to mark all ballots. We considered turnout in each of these races because council and borough mayoral contests are much more decen­ tralized and likely to better reflect more local and diverse community interests. These behaviours were probed with a question that first asked about voting in general and included options that would allow people to indicate that they did not vote but had considered it or usually did. In this way, we provided options that would decrease the social desirabil­ ity bias related to the question (see Morin-Chassé et al., 2017, for more information about the measure). A follow-up question for those who in­ dicated they had voted asked respondents to indicate whether they had voted in the races for mayor, council, and borough mayor. The level of information about and degree of familiarity with these different offices and the competing candidates likely varied greatly. May­ oral contests are well-publicized and can structure municipal elections in Quebec because council candidates often run as part of mayoral “équi­ pes.” For example, the former mayor of Montreal led “Équipe Denis Coderre pour Montréal,” which drew council candidates from other mu­ nicipal parties as well as independent incumbents and new candidates. Much like parties at other levels of government, the ones in Quebec announce party platforms, and the mayoral candidate expects support from their party members while in office. Elections for city councillors and borough mayors are less high-profile. We included a number of control variables in our models, such as age, education, gender, income, interest in the municipal election, and not exactly add up to 1, which is the case here). The maximum value of the index is obtained only if every person present in an FSA is from a different language group. A value of 0 would indicate that everyone is from the same language group. Other stud­ ies that look at linguistic heterogeneity have used this measure, such as Anderson and Paskeviciute (2006).

64  Cameron D. Anderson and Laura B. Stephenson

feeling that voting is a duty at the municipal level, as well as the actual percentage of francophones in a respondent’s FSA.3 For our initial tests, we also looked at feeling thermometers of attitudes toward different lin­ guistic groups (francophones, anglophones, and allophones). Finally, we used a variable that was coded 1 if the respondent indicated they belonged to a voluntary organization and 0 if not as a proxy for social capital, given its prevalence in the literature as a mobilizing factor for minority groups.4 Background Results First and foremost, it is important to note some objective details: the actual turnout rate in the mayoral election was 42.5%; also, just under 76% of our sample was francophone, about 10% anglophone, and about 13% had a first language other than French or English (allophone). The second background piece of information that is relevant for our discussion is a sense of the reported turnout in the municipal races held in 2017. Table 3.1 presents turnout rates by office type and language group. In the first instance, it appears that desirability bias and selection effects are present, even with the careful wording of the survey question, as our reported rates of turnout well exceed population averages. We also find evidence of ballot roll-off dynamics, as reported turnout rates drop significantly from mayoral to council to borough mayoral races. The last substantive point to make regarding Table 3.1 is the suggestive evidence of the effects of language on turnout in these elections. There is little difference in turnout for the mayoral race by language group, but a gap does emerge in council and borough mayoral races, in which anglophones have the highest rate of turnout. Additionally, as a way of laying the groundwork for understanding the linguistic features of this case, we consider the perceptions of language

3 Coding for each of the control variables is as follows:- Gender: women = 1, men = 0; Age: born in 1928–1961 = 2, 1962–1986 = 1, 1986–1999 = 0; Education: completed university or higher = 1, all else = 0; Income: nine categories of ascending household income by $25,000 increments (e.g., $0 – 25K = 1 …); Voting at the municipal level is a duty (=1) or a choice (=0); Interest in municipal politics 0 = not interested … 10 = very interested. 4 All models in the chapter also include a dummy control variable for the method of contact. Of the 1,924 campaign completes in Montreal, 1,222 were recruited through a random-digit-dial process and 702 were recruited from Forum’s existing online panel of respondents. The dummy variable for method of contact allows us to control for unobserved heterogeneity in the characteristics of respondents from each group.

Language and (Not) Voting  65 Table 3.1.  Reported turnout by language group by office, Montreal (%) Office

All

Francophone

Anglophone

Allophone

Mayor Council Borough mayor

79.3 69.2 58.7

79.4 69.0 58.2

81.5 78.0 68.5

77.6 67.2 65.0

groups. The mean value on the 0 to 3 linguistic diversity variable is 1.82, suggesting that diversity is certainly perceived by respondents. However, moving to the neighbourhood level, the data suggest that respondents are not very good at gauging the extent of linguistic diversity: the correla­ tion coefficient between objective and subjective indicators of linguistic diversity is only 0.42.5 As perceptions may differ depending on whether someone is part of a majority or minority language group, we also exam­ ined these correlations by language. The correlation coefficient of ob­ jective and subjective indicators was greatest among francophones (just higher than the average) and lowest among anglophones. This finding suggests that the relative size of the language group (majority or minor­ ity) strongly conditions the accuracy of perceptions of linguistic diversity. Another initial piece of data to consider focuses on the presence of social capital among linguistic groups. The literature suggests that social capital has important influences on engagement, and the presence of social capital in a language group may be an important factor in under­ standing how linguistic diversity influences turnout in municipal elec­ tions in Montreal. Table 3.2 shows the distribution of social capital by language group. Between groups, there is some interesting variation. Francophones are less likely to belong to volunteer groups than both anglophones and allophones, and the rate of group membership is not significantly different between the two minority language groups. A last opening look at our data is found in Figure 3.1, where we plot the distribution of respondents across the range of FSA Index of lin­ guistic dissimilarity values. The histogram shows the percentage of all respondents at each unique value of the linguistic diversity index. While we see that values range from about 0.05 to 0.7, there are clearly far more respondents living in FSAs that are higher on that range, as the mean of 0.47 would suggest. Notwithstanding this general observation about the distribution of respondents, there are a number of peaks and valleys throughout the range. 5 A strong association between these variables would be somewhere in the 0.7–0.8 range and above.

66  Cameron D. Anderson and Laura B. Stephenson Table 3.2.  Mean voluntary group membership by language, Montreal All

Francophone +#

36.3

33.7

Anglophone

Allophone

47.2*

45.3*

Note: Cell values indicate the percentage of linguistic group who are members of a voluntary association. *p < 0.05 difference from francophones; + p < 0.05 difference from anglophones; # p < 0.05 difference from allophones

Figure 3.1.  FSA index of linguistic dissimilarity in Montreal

Percentage of Respondents

10

8

6

4

2

0 0

.1

.2

.3

.4

.5

.6

.7

Index of Linguistic Dissimilarity

To summarize, thus far we have demonstrated the different rates of turnout in the 2017 municipal electoral contests and have sketched some of the city’s linguistic and social capital dimensions. We now turn to con­ sider our first two (and contrasting) hypotheses – that linguistic diversity will positively (H1) or negatively (H2) impact turnout in Montreal. Linguistic Diversity and Turnout To analyse the individual-level impact of linguistic diversity on turnout, we ran logistic regression models for mayor, council, and borough mayor voting. We estimate a model that includes controls (female, education, age, income, duty, and interest in municipal

Language and (Not) Voting  67 Figure 3.2.  Mayoral turnout in Montreal by FSA diversity 1

Pr(Mayoral turnout)

.8

.6

.4

.2

0 0

.1

.2

.3

.4

.5

.6

.7

FSA Diversity

politics), social capital, percentage of francophones in the respond­ ent’s FSA, and our measures of linguistic diversity (first objective indicators, then separate models with subjective perceptions). We present marginal effects in Figures 3.2 to 3.4 to display the results in an interpretable format.6 Figures 3.2 to 3.4 display the effect of increasing objective linguistic di­ versity on turnout in mayoral, council, and borough mayoral elections. These figures present the point estimates with 95% confidence inter­ vals. As objective diversity at the FSA level rises, turnout in the Montreal elections increases, providing evidence in favour of the mobilization hypothesis (H1). Figures 3.5 to 3.7 display the same pattern with sub­ jective linguistic diversity – though with slightly less substantive effect.7 Taken together, these figures reveal the mobilizing effect of linguistic

6 All models involving objective data cluster standard errors on the FSA. Full regression results are available, on request, from the authors. 7 Indeed, while the direction of effect is in a positive direction (consistent with H1), the large confidence intervals associated with the point estimates suggest that the differ­ ence in predicted turnout across the range of subjective linguistic diversity perception values is not statistically significant for Figures 3.6 and 3.7.

68  Cameron D. Anderson and Laura B. Stephenson Figure 3.3.  Council turnout in Montreal by FSA diversity 1

Pr(Council turnout)

.8

.6

.4

.2

0 0

.1

.2

.3 .4 FSA Diversity

.5

.6

.7

.6

.7

Figure 3.4.  Borough mayoral turnout in Montreal by FSA diversity

Pr(Borough mayoral turnout)

.8

.6

.4

.2

0 0

.1

.2

.3

.4

FSA Diversity

.5

Language and (Not) Voting  69 Figure 3.5.  Mayoral turnout in Montreal by diversity perceptions

Pr(Mayoral turnout)

.9

.8

.7

.6 0

1 2 Perceptions of Linguistic Diversity

3

Figure 3.6.  Council turnout in Montreal by diversity perceptions .8

Pr(Council turnout)

.75

.7

.65

.6

.55 0

1 2 Perceptions of Linguistic Diversity

3

70  Cameron D. Anderson and Laura B. Stephenson Figure 3.7.  Borough mayoral turnout in Montreal by diversity perceptions .7

Pr(Borough mayoral turnout)

.65

.6

.55

.5

.45 0

1

2

3

Perceptions of Diversity

diversity in Montreal. In terms of our hypotheses, recall that we posed contrasting expectations based on the literature. As anticipated by H2, we considered the possibility that diversity might decrease turnout be­ cause heterogeneity can work against the development of social capi­ tal. However, in light of the mobilization hypothesis (H1), our results broadly suggest that diversity may spur feelings of competition that lead to greater involvement. Our remaining expectations are based in literature that suggests that groups may react to linguistic heterogeneity differently. Before looking at turnout, we first considered whether there was in fact a difference in perceptions of diversity across language groups. The OLS regression results are reported in Appendix Table A3.1. From the first model, an­ glophones are more likely to perceive linguistic diversity (compared to francophones). Francophones are significantly less likely to perceive lin­ guistic diversity than the minority language groups. However, Models 2 and 4 reveal that when controlling for actual levels of linguistic diversity at the FSA level, language group fails to have a significant effect. This latter finding is not surprising and holds even when controlling for so­ cial capital. Thus, there are no systematic differences in perceptions of heterogeneity by language group.

Language and (Not) Voting  71 Figure 3.8.  Mayoral turnout in Montreal for anglophones vs francophones by FSA diversity 1.5

Pr(Mayoral turnout)

1

.5

0

–.5 0

.1

.2

.3 .4 FSA Diversity Francophone

.5

.6

.7

Anglophone

We ran models similar to those in Figures 3.2 to 3.7 to test for unique effects of linguistic diversity on different language groups.8 Figures 3.8 and 3.9 display the effects of increasing linguistic diversity (objective and subjective) on the probability of turning out in the mayoral election. Both figures show that the effects are not significantly different for an­ glophones compared to francophones. However, even though they are insignificant across all levels (unlike in Figures 3.2 to 3.7), we do observe an interesting pattern among anglophones – the trend is decreasing turn­ out levels as perceptions of linguistic diversity increase (Figure 3.9). We only show these results for anglophones in Montreal as they are exemplary of the null results we find in all other analyses. Insignificant interaction effects were observed for all language groups, with both types of linguistic diversity data, and among all three types of elections. The results suggest that linguistic diversity (either in actuality or in the perception of it) did not impact the turnout decisions of majority and minority language groups differently in the 2017 municipal elections. As such, our individual-level results do not support the expectations for 8 All results available from authors on request.

72  Cameron D. Anderson and Laura B. Stephenson Figure 3.9.  Mayoral turnout in Montreal for anglophones vs francophones by diversity perceptions

Pr(Mayoral turnout)

1.2

1

.8

.6

.4 0

1

2

3

Perceptions of Diversity Francophone

Anglophone

majority and minority language groups that emerged from the litera­ ture. Furthermore, these null findings with individual-level data do pres­ ent a contrast to the aggregate-level turnout data, which did find turnout differences between the language groups. Discussion and Conclusion Language and linguistic diversity is a distinguishing feature of Quebec politics. In this chapter, we outlined the empirical landscape of this di­ versity in Montreal and considered the extent to which linguistic diver­ sity influenced turnout in the 2017 municipal elections. There are some important differences in reported rates of turnout by language group. At the council and mayoral levels, anglophones have higher rates of turnout than the other language groups (francophones and allophones). This comports with our expectations for how diversity and density interact. We also find that levels of social capital are significantly differentiated by language group (as francophones are least likely to be group members). However, when we investigate our main research questions at the in­ dividual level, we find mixed evidence for how linguistic diversity affects turnout, especially in light of our expectations developed from existing

Language and (Not) Voting  73

research. Perhaps this is to be expected, as these effects find support in the wide range of findings that have been realized in the broader literature. While we expected to find some indication that the effects are conditional on language group, we only find that local linguistic diversity impacts turnout for all respondents positively, regardless of lan­ guage group. What does all this mean? Generally speaking, language and percep­ tions of linguistic diversity are important considerations in Quebec politics. At the municipal level, we find that linguistic diversity affects engagement but perhaps not as much as one might think given the de­ gree of diversity in Montreal. Clearly, the results show that FSA linguis­ tic diversity is positively related to turnout decisions in these elections. However, the effects of perceptions of diversity and conditional impacts (based on majority and minority position) are much less prominent. In­ deed, the literature regarding the conditional effects among minority and majority language groups, while disagreeing on the direction of the effects, was unified in finding results. As such, we find our lack of results perplexing. We believe that further investigation is warranted to better understand whether our null findings are unique to our data, our meas­ ures, or the linguistic context in Montreal more generally.

74  Cameron D. Anderson and Laura B. Stephenson

Appendix to Chapter 3 Table A3.1.  Predicting perceptions of linguistic diversity Model 1 0.041 (0.063) 0.002 (0.065) 0.016 (0.045) 0.003 (0.016) 0.025+ (0.013) 0.273** (0.110) 0.016 (0.109) – 0.208** (0.069) –

Model 2 0.033 (0.060) 0.004 (0.065) –0.012 (0.041) –0.009 (0.016) 0.022 (0.016) 0.091 (0.110) 0.180 (0.115) – 0.170** (0.070) –1.410*** (0.160)

Model 3 0.041 (0.063) –0.003 (0.065) 0.023 (0.045) 0.005 (0.016) 0.025+ (0.014) – – –0.132+ (0.079) 0.212** (0.067) –

Model 4 0.030 (0.060) –0.004 (0.066) –0.010 (0.041) –0.008 (0.016) 0.022 (0.016) – – 0.060 (0.080) 0.180** (0.070) –1.420*** (0.160)

N

1.49*** 0.14 848

2.45*** 0.200 788

1.613*** 0.151 848

2.41*** 0.194 788

R2

0.037

0.153

0.032

0.148

Female Education Age Income Interest Anglophone Allophone Francophone Social capital % FSA  francophone Constant

+

p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001

Language and (Not) Voting  75 REFERENCES Anderson, C. J., & Paskeviciute, A. (2006). How ethnic and linguistic heterogeneity influence the prospects for civil society: A comparative study of citizenship behavior. Journal of Politics, 68(4), 783–802. http://dx.doi.org /10.1111/j.1468-2508.2006.00470.x Bellettini, G., Ceroni, C. B., & Monfardini, C. (2016). Neighbourhood heterogeneity and electoral turnout. Electoral Studies, 42, 146–156. https:// doi.org/10.1016/j.electstud.2016.02.013 Blais, A. (2000). To vote or not to vote. University of Pittsburgh Press. Buton, F., Lemercier, C., & Mariot, N. (2012). The household effect on electoral participation: A contextual analysis of voter signatures from a French polling station (1982–2007). Electoral Studies, 31(2), 434–447. https:// doi.org/10.1016/j.electstud.2011.11.010 Coffé, H. (2009). Social capital and community heterogeneity. Social Indicators Research, 91(2), 155–170. https://doi.org/10.1007/s11205-008-9275-y Coffé, H., & Geys, B. (2007). Toward an empirical characterization of bridging and bonding social capital. Nonprofit and Voluntary Sector Quarterly, 36(1), 121–139. https://doi.org/10.1177/0899764006293181 Costa, D. L., & Kahn, M. E. (2003). Civic engagement and community heterogeneity: An economist’s perspective. Perspectives on Politics, 1(1), 103–111. https://doi.org/10.1017/S1537592703000082 Cox, G. W., & Munger, M. C. (1989). Closeness, expenditures, and turnout in the 1982 US House elections. American Political Science Review, 83(1), 217–231. https://doi.org/10.2307/1956441 DeSante, C. D., & Perry, B. N. (2016). Bridging the gap: How geographic context affects political knowledge among citizen and non-citizen Latinos. American Politics Research, 44(3), 548–577. https://doi.org/10.1177 /1532673X15606967 Fieldhouse, E., & Cutts, D. (2008). Diversity, density, and turnout: The effect of neighbourhood ethno-religious composition on voter turnout in Britain. Political Geography, 27, 530–548. http://dx.doi.org/10.1016/j.polgeo.2008.04.002 Fieldhouse, E., & Cutts, D. (2010). Does diversity damage social capital? A comparative study of neighborhood diversity and social capital in the US and Britain. Canadian Journal of Political Science, 43(2), 289–318. http://dx.doi .org/10.1017/S0008423910000065 Gerber, A., Green, D., & Larimer, C. (2008). Social pressure and voter turnout: Evidence from a large-scale field experiment. American Political Science Review, 102(1), 33–48. https://doi.org/10.1017/S000305540808009X Geys, B. (2006). Explaining voter turnout: A review of aggregate-level research. Electoral Studies, 25, 637–663. http://dx.doi.org/10.1016/j.electstud.2005 .09.002

76  Cameron D. Anderson and Laura B. Stephenson Harell, A., Stolle, D., & Quintelier, E. (2009). Do diverse networks demobilize young people? European Journal on Child and Youth Research, 3(7), 86–92. Hill, K. Q., & Leighley, J. (1999). Racial diversity, voter turnout, and mobilizing institutions in the United States. American Politics Quarterly, 27(3), 275–295. https://doi.org/10.1177/1532673X99027003001 Hinkson, K., & Laframboise, K. (2017, 13 March). Canadians divided when it comes to immigration, poll suggests. CBC News. http://www.cbc.ca/news /canada/montreal/canadian-immigration-poll-quebec-1.4022664 Huckfeldt, R. R. (1984). Political loyalties and social class ties: The mechanisms of contextual influence. American Journal of Political Science, 29(2), 399–417. https://doi.org/10.2307/2110879 Huckfeldt, R., Beck, P. A., Dalton, R. J., & Levine, J. (1995). Political environments, cohesive social groups, and the communication of public opinion. American Journal of Political Science, 39(4), 1025–1054. Huckfeldt, R., Plutzer, E., & Sprague, J. (1993). Alternative contexts of political behavior: Churches, neighborhoods, and individuals. Journal of Politics, 55(2), 365–381. https://doi.org/10.2307/2132270 Huckfeldt, R., & Sprague, J. (1987). Networks in context: The social flow of political information. American Political Science Review, 81(4), 1197–1216. https://doi.org/10.2307/1962585 Lazarsfeld, P. F., Berelson, B., & Gaudet, H. (1948). The people’s choice: How the voter makes up his Mind in a presidential campaign (2nd ed.). Columbia University Press. Marschall, M. J., & Stolle, D. (2004). Race and the city: Neighbourhood context and the development of generalized trust. Political Behavior, 26(2), 125–153. https://doi.org/10.1023/B:POBE.0000035960.73204.64 McClurg, S. D. (2006). Political disagreement in context: The conditional effect of neighborhood context, disagreement and political talk on electoral participation. Political Behavior, 28, 349–366. McDonald, M. P., & Tolbert, C. J. (2012). Perceptions vs. actual exposure to electoral competition and effects on political participation. Public Opinion Quarterly, 76(3), 538–554. https://doi.org/10.1093/poq/nfs029 Morin-Chassé, A., Bol, D., Stephenson, L. B., & Labbé St-Vincent, S. (2017). How to survey about electoral turnout? The efficacy of face-saving response items in 19 different contexts. Political Science Research Methods, 5(3), 575–584. http://dx.doi.org/10.1017/psrm.2016.31 Mutz, D. C. (2002). The consequences of cross-cutting networks for political participation. American Journal of Political Science, 46(4), 838–855. http:// dx.doi.org/10.2307/3088437 Mutz, D. C. (2006). Hearing the other side: Deliberative versus participatory democracy. Cambridge University Press.

Language and (Not) Voting  77 Nakhaie, M. R. (2006). Electoral participation in municipal, provincial, and federal elections in Canada. Canadian Journal of Political Science, 39(2), 363–390. https://doi.org/10.1017/S000842390606015X Nir, L. (2005). Ambivalent social networks and their consequences for participation. International Journal of Public Opinion Research, 17(4), 422–442. http://dx.doi.org/10.1093/ijpor/edh069 Putnam, R. D. (2007). E Pluribus Unum: Diversity and community in the twentyfirst century: The 2006 Johan Skytte Prize Lecture. Scandanavian Political Studies, 30(2), 137–174. https://doi.org/10.1111/j.1467-9477.2007.00176.x Putnam, R. D., Leonardy, R., & Nanetti, R. (1993). Making democracy work: Civic traditions in modern Italy. Princeton University Press. Quintelier, E., Stolle, D., & Harell, A. (2012). Politics in peer groups: Exploring the causal relationship between network density and political participation. Political Research Quarterly, 65(4), 868–881. https://doi.org/10.1177 %2F1065912911411099 Sandovici, M., & Listhaug, O. (2010). Ethnic and linguistic minorities and political participation in Europe. International Journal of Comparative Sociology, 51(1–2), 111–136. https://doi.org/10.1177/0020715209347070 Scheufele, D. A., Nisbet, M. C., Brossard, D., & Nisbet, E. C. (2004). Social structure and citizenship: Examining the impacts of social setting, network heterogeneity, and informational variables on political participation. Political Communication, 21, 315–338. http://dx.doi.org/10.1080/10584600490481389 Statistics Canada. (2017). Focus on Geography series, 2016 Census. Statistics Canada catalogue no. 98-404-X2016001. Ottawa. Turgeon, L., & Bilodeau, A. (2014). Minority nations and attitudes toward immigration: The case of Quebec. Nations and Nationalism, 20(2), 317–336. http://dx.doi.org/10.1111/nana.12068 Zuckerman, A. S., Dasović, J., & Fitzgerald, J. (2007). Partisan families: The social logic of bounded partisanship in Germany and Britain. Cambridge University Press. Zuckerman, A. S., & Kroh, M. (2006). The social logic of bounded partisanship in Germany: A comparison of West Germans, East Germans, and immigrants. Comparative European Politics, 4(1), 65–93. http://dx.doi.org/10.1057 /palgrave.cep.6110071 Zuckerman, A. S., Valentino, N. A., & Zuckerman, E. W. (1994). A structural theory of vote choice: Social and political networks and electoral flows in Britain and the United States. Journal of Politics, 56(4), 1008–1033. https:// doi.org/10.2307/2132071

4 Nationalism and Municipal Elections in Quebec jean - françois daoust

In Canada, the province of Quebec has a partisan system largely based on political parties’ positions regarding the national question, that is, whether Quebec should be an independent state (Bélanger & Nadeau, 2009). The founding of the Parti Québécois (PQ) in 1968 polarized the debate at the provincial level; that new party became the alternative to the Quebec Liberal Party (QLP), which has always favoured keeping Quebec within the Canadian federation. In recent decades, the third most prominent party, whether it be the ADQ (1994–2011) or the Coalition Avenir Québec (since 2011), has often tried to bridge this divide by being “agnostic” or “autonomist” on the national question, albeit with limited success.1 A similar dynamic exists at the federal level: the Bloc Québécois has offered an independentist option since 1991. Not surprisingly, there is a rich literature demonstrating that attitudes toward Quebec independence are among the most important determinants of vote choice. This is true both at the provincial level (Bastien et al., 2013; Bélanger & Nadeau, 2009; Bélanger et al., 2018; Blais & Nadeau, 1984) and at the federal level (Bakvis & Macpherson, 1995; Blais et al., 2002; Daoust & Dassonneville, 2018; Fournier et al., 2013; Gagnon & Boucher, 2017). Yet we do not know whether the impact of issues related to nationalism is as salient during municipal elections as it is for other levels. Regarding the 2009 Montreal municipal election, Drouilly (2010) found that in Montreal, francophones and non-francophones were two very different communities with respect to vote choice. To explain this cleavage, he pointed out the connection between the two main opponents: on the one hand, there was Gérald Tremblay, a federalist and a former provincial Liberal minister, and on the other hand, there was Louise Harel, a 1 That is, until the Coalition Avenir Québec won the 2018 provincial election. See Bélanger et al., 2022.

Nationalism and Municipal Elections in Quebec  79

separatist and a former minister for the Parti Québécois. Drouilly hypothesized that this situation reflected the fact that the national question structures Quebec’s political life even at the municipal level. Beyond Drouilly’s observation, there are theoretical reasons to believe that nationalism may be important in municipal contests. First, the literature on the “contamination” effect is insightful in its simple form: attitudes and/or behaviours at a certain level may not be independent of attitudes and/or behaviours at another level.2 So there may be a “contamination” effect from other levels of electoral competition that are polarized around, among other things, the national question. Supporters of the Parti Québécois and Québec Solidaire (two provincial parties that view independence favourably) may want to support local candidates for reasons similar to those that explain their vote choice at the provincial level. The same rationale applies to supporters of the Bloc Québécois at the federal level. In sum, there may be spillover across electoral arenas so that, even though local governments have no real influence regarding the separatism question, voters may use the stances of local candidates as a “cue” when forming local vote preferences. Second, municipal elected representatives can and sometimes do take actions that are at least tangentially related to nationalism. For example, Denis Coderre, the incumbent mayor of Montreal (2013–17), changed the city’s flag and coat of arms in 2017 to include the iconic white pine representing the First Nations people who, prior to European settlement, had inhabited the land on which the city has since been built. In this way, both emblems now include Indigenous nations as full representatives, equal to the other four “nations” (French, Scottish, Irish, and English). Also, the municipal government has the power (and has used it) to suspend the parking rules on days when Jewish people are forbidden by their religion to move their cars to legal parking zones. In Quebec City, Mayor Régis Labeaume publicly defended the Islamic community’s right to its own cemetery. While not directly related to the national question, these issues do acknowledge rights that are so often the focus of debates over French language and culture. In this chapter, I examine whether the national question has an effect on vote choice in municipal elections in Quebec. I also consider the effects of left/right ideological cleavage on vote choice (see chapter 5 in this volume for a detailed analysis of the role of ideology). We know that both nationalist and ideological cleavages play a role in elections at 2 See Riera (2013) for an analysis of contamination effects focusing on vote choice, which is also not a stranger to the literature on shortcuts and heuristics (for example, at the local level, see Boudreau and colleagues, 2015).

80  Jean-François Daoust

other levels, though some studies have found that the national question is declining in importance among voters.3 Whether one cleavage predominates over the other seems to be an open question at all levels, and comparing the impact of nationalist attitudes to the impact of left/right ideology will be useful in the present examination of the municipal level. Hence, this chapter will provide new insights about nationalism and its impact in municipal elections, but also in relation to one of the most studied cleavages (left/right ideology). Overall, results indicate that in Montreal municipal politics, attitudes toward Quebec independence matter as much as the left/right ideological dimension in citizens’ vote choice calculus. However, this effect is driven by citizens who favour Quebec independence and who are cognizant of Denis Coderre’s affiliation to the federalist Liberal Party of Canada. In Quebec City, attitudes toward Quebec independence were also a significant predictor of vote choice, but that factor was less important than the classic left/right cleavage. I conclude that nationalism is quite important in the province’s municipal elections. My results support the idea, suggested by Drouilly (2010), that the national question is central at all levels of government in Quebec, not just the provincial and federal levels, and should be taken into account during municipal elections. Data and Indicators I use the CMES dataset for Montreal and Quebec City described earlier in this volume. The wording of all questions can be found in the appendix at the end of this book. Below, I describe the indicators used and how they are coded. Detailed descriptive statistics can be found in Table 4.1. Age: A continuous measure ranging from 18 to 92. Rescaled [0,1]. Sex: Based on self-identification. Woman = 1, Man = 0. Education: “What is the highest level of education that you have completed?” No schooling, Some elementary school, Completed elementary school, Some secondary/high school, Some technical, community college, Completed technical, community college, Some university, Bachelor’s degree, Master’s degree, professional degree or doctorate. Variable is a dummy where 1 = some university (or more), 0 = otherwise. Non-responses were coded as missing.

3 See Blais et al. (2002) for federal elections; Bélanger and Nadeau (2009) and Daoust and Jabbour (2020) for provincial elections; and, for insights on ideology at the ­municipal level in Canada, Bherer and Breux (2012) and Couture (2017).

Nationalism and Municipal Elections in Quebec  81 Table 4.1.  Descriptive statistics for Montreal and Quebec City Variables

Mean

Standard deviation

Age Sex (woman) Education French Ideology Quebec independence

0.50/0.51 0.48/0.42 0.68/0.59 0.78/0.97 0.46/0.54 0.45/0.36

0.21/0.22 0.50/0.42 0.47/0.49 0.41/0.17 0.21/0.21 0.40/0.37

Note: Values for Montreal/Quebec City are displayed. All variables range from 0 to 1. N = 612/998.

Language: “What is the first language you learned and still understand?” Francophones are coded 1, 0 otherwise.4 Ideology (left-right): “In politics people sometimes talk of left and right. Where would you place yourself on a scale from 0 to 10, where 0 means left and 10 means right?” Non-responses are coded as 5. Rescaled [0,1]. Quebec independence: “Are you very favourable, somewhat favourable, somewhat opposed, or very opposed to Quebec sovereignty, that is, Quebec is NO LONGER A PART OF CANADA?” Response choices were rescaled from 0 to 1 where 1 represents someone who is very favourable. Based on Table 4.1, we can see that there is very little variation between the two cities regarding the socio-demographic variables. It is worth noting, however, that there are slightly more women in the Montreal sample and that almost all respondents are francophones in the Quebec City sample. On the attitudinal variables, Quebec City voters are further to the right than are Montrealers, and less favourable to Quebec independence, which is consistent with the literature (Daoust, 2017). Voter’s Perceptions of Candidates’ Political Affiliation Of course, nationalism will have a greater potential impact on vote choice in municipal elections if voters are aware of candidates’ positions on the issues related to nationalism and, more specifically, whether they are associated with independentist or federalist parties. Hence, I first assess whether voters associated mayoral candidates with a sovereigntist or a federalist party at the provincial and federal levels. 4 As a robustness check, I ran an analysis separating francophones, anglophones, and “other,” and it did not change the results of regression models below.

82  Jean-François Daoust

There are a number of reasons to believe that citizens perceived a connection between the five mayoral candidates considered here (two in Montreal and three in Quebec City) and political parties at different levels. In Montreal, the incumbent Denis Coderre’s partisan association was very clear: he had always been a federal Liberal. He was first elected in the 1997 federal election, and before that, he had been a Liberal candidate in 1988, 1990 (in a by-election against Gilles Duceppe), and 1993. He was also the president of the youth federal Liberals when he was 24 years old. In sum, Denis Coderre had been publicly identified with the federal Liberals for almost 30 years. Moreover, he was known to be a clear anti-separatist. In fact, in 1995, the year of the referendum on independence, Coderre stated that he regretted that Canada lacked a law that would see to the deportation of sovereigntists who “spit on his flag” (Gagnon, 2013). All in all, it was obvious that Denis Coderre was a strong federalist and a Liberal. In the case of Valérie Plante, his opponent in the mayoral race, things were not as straightforward. Her career path had been characterized by various stereotypes associated with the left. She worked in the community and union milieu, notably helping women and young girls, and was on the board of the Broadbent Institute (a progressive, left-leaning institute). She was associated with the sovereigntist movement because of the general proximity between her municipal political party and Québec Solidaire. More recently, she had publicly celebrated Saint-Jean-Baptiste Day (the province’s “national” holiday) but not the Canada Day celebrations, which reinforced her nationalist image (Lau, 2018). Plante clearly leaned left ideologically, and there were good reasons to suspect that she was also independentist-leaning as well. In Quebec City, the incumbent Régis Labeaume was known to work well with other levels of government, regardless of the party in power (including the Conservative government in Ottawa and the Liberals at the provincial level). However, a review of his career suggests that he has a nationalist streak. He worked for a minister in the PQ government for five years in the 1980s and had publicly stated that he voted YES in both referendums on Quebec’s independence. Moreover, it would appear (from media reports) that he had close ties with Pierre-Karl Péladeau, ex-PQ leader and CEO of Québécor, with whom he worked to develop a partnership to build a hockey arena in Quebec City. More anecdotally, he publicly paid tribute to Agnès Maltais when, after 20 years representing the PQ in a Quebec City riding, she decided to retire. Labeaume seems to be more pragmatic than consistently ideological. On the nationalist front, it seems that he is sovereignist-leaning, though he rarely says so explicitly.

Nationalism and Municipal Elections in Quebec  83

His opponents, Jean-François Gosselin and Anne Guérette, were much less well-known by the populace. Gosselin had been a Member of the National Assembly (2007–8), where he was part of the right-leaning Action Démocratique du Québec. He lost his seat in 2008; then in 2012, he ran as a candidate for the Liberal Party of Quebec (he again lost). The city’s popular radio stations, which are associated with the ideological right, informally supported his mayoral candidacy (with some anchors doing so explicitly). It seems that this was largely because he favoured a third road link between Quebec City and its suburbs in the form of a new bridge across the St. Lawrence. The CAQ and the Conservatives at provincial and federal levels favoured that project as well. Finally, regarding Anne Guérette, her past positions were clearly left-leaning, but we have very little in the way of public information (at least as reported in the media) about her position on Quebec independence. But what were citizens’ perceptions of the major mayoral candidates’ partisan affiliations? The CMES survey asked respondents whether they perceived a connection between the candidates and various political parties at both the federal and provincial levels. The question used to measure these perceptions was as follows: “Which [provincial/federal] political party, if any, would you associate with [Plante/Coderre/Labeaume/Gosselin/Guérette]?” Federal and provincial results for the five mayoral candidates are shown in Table 4.2. Perhaps unsurprisingly, the percentage of “don’t know” responses is quite high for all candidates. After all, it might not be strategic for a candidate to seek to be clearly associated with a party, for doing so could drive away opponents of that party. Furthermore, a candidate in a major city like Montreal or Quebec City wants to be seen as able to work with other levels of government, whichever party is in power. The striking exception to this is in Montreal, where Denis Coderre was clearly associated with a federalist party, that is, the Liberal Party. Just over 60% of respondents associated him with both the provincial and the federal Liberals. Only 6% associated him with an independentist party at the provincial level (either the PQ or QS), and 3% at the federal level (the Bloc). This no doubt reflects the fact that he was a Liberal MP from 1997 to 2013 and a cabinet minister for several years. Valérie Plante was less well-known in the public sphere. This is illustrated by the higher proportion of people (slightly more than one third) who did not know how to associate her with other parties. In terms of federal parties, it seems that Plante was perceived as a leftist, probably an NDP supporter. In terms of provincial parties, the two main independentist parties lean left, and about 20% of respondents associated her

84  Jean-François Daoust Table 4.2.  Municipal candidates and political parties (%) Montreal Provincial ID CAQ PLQ PQ QS Other party Don’t know None Federal ID CPC Green LPC NDP Bloc Québécois Other party Don’t know None

Quebec City

Plante

Coderre

Labeaume

Gosselin

Guérette

10.2 9.6 11.8 19.6 2.9 34.7 11.3

3.9 60.2 4.4 1.4 1.2 20.1 8.8

5.4 30.9 29.2 0.7 1.5 19.7 12.5

39.8 6.2 4.8 2.6 3.4 31.9 11.4

6.8 12.6 14.3 18.9 2.1 32.3 13.1

5.1 8.1 9.5 19.9 9.0 2.0 33.6 12.9

6.5 1.0 62.1 1.8 3.1 1.0 17.6 7.1

9.0 0.3 30.5 2.0 18.5 1.9 22.3 15.6

38.5 0.7 5.1 5.8 4.2 1.7 32.1 12.0

4.2 8.3 11.9 21.1 7.3 1.7 31.5 13.9

Note: N = 612 for Montreal and 998 for Quebec City. Percentages (%) are displayed.

with Québec Solidaire and 12% with the Parti Québécois. Clearly, a large proportion of Montrealers associated her with an independentist party. Overall, it seems that voters perceived the Montreal municipal election as a campaign between a Liberal federalist (Coderre) and a nationalist/ sovereignist on the left (Plante). In Quebec City, the incumbent Régis Labeaume was, as expected, the best-known candidate (and thus had the lowest rate of “don’t know” responses). At the federal level, he was seen as a Liberal (30.5%) or a Bloquiste (18.5%); at the provincial level, voters were again split, this time between the Quebec Liberal Party and the Parti Québécois. Voters, then, seemed quite divided over Labeaume’s partisan affiliation. We see a very different picture for Gosselin: Quebecers clearly saw him as a candidate of the right, associating him with the CAQ (39.8%) at the provincial level and with the Conservative Party of Canada (38.5%) federally. Finally, Guérette seems to have been perceived as left-leaning, given that voters are most likely to associate her with Québec Solidaire (18.9%) and the NDP (21.1%). All in all, the most striking result in Quebec City is that Gosselin, the second-place finisher, was perceived as a federalist. Hence, in both cities, there was potential for issues related to nationalism to be significant determinants of vote choice. I now test this contention.

Nationalism and Municipal Elections in Quebec  85 Table 4.3.  Logistic regression models: Coderre = 1; Plante = 0 Model 1 Age Sex (woman) Education French Ideology Quebec independence Coderre as liberal Coderre as liberal x Quebec Independence Constant Observations Pseudo R2

+

0.83 (0.42) –0.01 (0.17) 0.26 (0.18) –0.19 (0.20) – – – – –0.99* (0.32) 612 0.009

Model 2 +

0.78 (0.44) 0.03 (0.18) 0.37+ (0.20) 0.27 (0.23) 1.98* (0.46) –1.29* (0.26) – – –1.82* (0.42) 612 0.085

Model 3 0.72 (0.45) 0.04 (0.18) 0.35+ (0.20) 0.23 (0.23) 1.92* (0.46) –0.20 (0.50) 0.78* (0.30) –1.42* (0.56) –2.32* (0.47) 612 0.095

Note: Logistic regression coefficients displayed. Robust standard errors in parentheses. + p < 0.10; * p < 0.05

Vote Choice and Nationalism in Montreal The two main candidates in Montreal together received over 90% of the vote,5 so I will focus on them and exclude the small number of respondents in the survey who supported other candidates. Using a binary dependent variable where a vote for Coderre equals “1” and a vote for Plante equals “0,” I run logistic regression models to analyse the determinants of Montrealers’ vote choices. Results are displayed in Table 4.3. Model 1 includes socio-demographic determinants, and shows that age is the only such factor to be related to vote choice (with older voters tending to support Coderre). Model 2 includes two underlying beliefs and values, that is, voters’ opinions on Quebec independence and their left/ right ideology. It reveals, first, that support for Quebec independence had a significant negative impact on one’s proclivity to vote for Coderre. Even though these attitudes toward independence are clearly associated with vote choice, logistic regression results are not directly interpretable. Hence, I computed the predicted probabilities of voting for Denis Coderre in Figure 4.1 (based on Model 2 in Table 4.3). I find that, ceteris paribus, someone who was “very opposed” to Quebec independence had a predicted probability of .49 of voting for Coderre, as compared to a value of .22 for someone who was “very in favour.” This maximum effect of 27 percentage points is very important. However, a more realistic estimate would be to compare someone moving from 5 In the survey, 63.2% of respondents said that they voted for Plante while 36.8% supported Coderre.

86  Jean-François Daoust

Pr(Coderre)

Figure 4.1.  Vote choice determinants in Montreal .8

.8

.7

.7

.6

.6

.5

.5

.4

.4

.3

.3

.2

.2

.1

.1

0

0

Very Opposed

Somewhat Opposed

Somewhat Favourable

Support for Quebec Independence

Very Favourable

Left

.1

.2

.3

.4

.5

.6

.7

.8

.9 Right

Ideology

somewhat opposed to somewhat favourable. In that case, the effect is still quite important, with a decrease of 9 percentage points in vote support for Coderre. These effects are important, but how do they compare with the effects of ideology? As shown in Figure 4.1, being more rightist increased voters’ proclivity to support Coderre. Going from 0 to 1 on the left/right scale increases the probability of doing so by about 40 percentage points, which is huge. However, taking one standard deviation around the mean (comparing someone moving from .3 to .7) gives an effect of about 20 percentage points. It is not impossible that the impact of the national question would vary across voters who realize that Denis Coderre is affiliated with the Liberal Party of Canada, which is strongly opposed to independence. Luckily, the dataset included questions about the perceptions of the affiliations of the candidates, as reported in Table 4.2. Coderre is a striking example of a candidate perceived to be affiliated with a political party at both the provincial and federal levels. In Model 3 of Table 4.3, I included an interaction term between the national question variable and a dichotomous variable indicating whether respondents identified Coderre as being tied to the Liberal Party of Canada. The coefficient indicates that the effect of the national question is significantly greater when respondents do have the heuristic of the affiliation. Figure 4.2 shows the interaction effect. In fact, the effect of the national question is undistinguishable from zero when respondents do not know Coderre’s affiliation. The effect

Nationalism and Municipal Elections in Quebec  87 Figure 4.2.  The heuristic of Coderre’s political affiliation

Average Marginal Effects of the National Question on Pr(Coderre)

.2

.1

0

–.1

–.2

–.3

–.4 Coderre Not Seen as Liberal

Coderre Seen as Liberal

Note: Estimations based from Model 3, Table 4.3, with 95% confidence intervals.

found in Model 2 is driven by people who perceive Coderre as a federal Liberal. Those who do are (at maximum) 32 percentage points less likely to support Coderre. All in all, nationalism seems important in the Montreal election, but it appears at first glance to be less important than ideology. However, one needs to take into account whether citizens have the cue about the candidates’ political affiliation. When voters hold positive views about Quebec independence and perceive Coderre to be tied with the federalist Liberal Party of Canada, the impact of the national question substantially increases. Vote Choice and Nationalism in Quebec City In contrast to Montreal, Quebec City had three “major” candidates. Hence, the dependent variable is categorical and represents a vote for Labeaume, Gosselin, or Guérette.6 The appropriate modelling approach to analyse a nominal dependent variable is a multinomial logistic regression. I discuss statistically significant findings from this analysis below; however, due to space considerations, the full results of these models are found in Table A4.1 of the appendix at the end of this chapter. I will

6 The shares of respondents who supported the candidates are as follows: 47.4% for Labeaume, 33.3% for Gosselin, and 19.3% for Guérette.

88  Jean-François Daoust

Pr(Vote choice)

Figure 4.3.  Vote choice determinants in Quebec City .8

.8

.7

.7

.6

.6

.5

.5

.4

.4

.3

.3

.2

.2

.1

.1

0

0

Very Opposed

Somewhat Opposed

Somewhat Favourable

Very Favourable

Support for Quebec Independence Labeaume

Gosselin

Guérette

Left

.1

.2

.3

.4

.5

.6

.7

.8

.9 Right

Ideology (Left-Right) Labeaume

Gosselin

Guérette

display the effects of the variables, comparing them to Labeaume (the reference category). Model 1 of Table A4.1 shows that education and age are significant control variables. Relative to Labeaume’s voters, supporters of Gosselin are less educated and those who voted for Guérette are older. As for Montreal, Model 2 adds the national question and ideology. Both are statistically significant. The predicted probabilities of voting for each candidate are shown in Figure 4.3 based on Model 2. All else being equal, a voter who is “very opposed” to independence has a predicted probability of .43 of voting for Labeaume; this value increases to .58 for someone who is “very favourable.” There is also a positive effect for Guérette, albeit a less substantial one – the maximum effect is 6 points. To the contrary, voters more favourable to Quebec independence are less likely to support Gosselin, and it appears to be the most important effect, with a maximum impact of 23 percentage points. That is, someone very opposed is predicted to have 40% chance of supporting him as compared to 17% for voters very favourable to independence. Overall, these maximum effects are important, and are even more important when we look at the changes in terms of the “somewhat opposed” to “somewhat favourable” attitudes. How do these effects compare to the ideological left/right cleavage’s impact? In Quebec City, the impact of ideology is very important. Regarding the left/right cleavage, not surprisingly, right-wingers are more likely to support Gosselin while leftists are more likely to vote for Guérette.

Nationalism and Municipal Elections in Quebec  89 Figure 4.4.  The heuristic of Gosselin’s political affiliation Average Marginal Effects of the National Question on Pr(Gosselin)

.2 .1 0 –.1 –.2 –.3 –.4

Gosselin Not Seen as Conservative

Gosselin Seen as Conservative

Note: estimations based from Model 3, Table A4.1, with 95% confidence intervals.

The effect for Labeaume is curvilinear, showing that he had an appeal among centrists. Focusing on one standard deviation around the mean, the effects for both challengers are about 30 percentage points, which is very important. In terms of potential heuristics, the most noteworthy case is Gosselin, who is perceived as affiliated with the Conservative Party of Canada.7 I included an interaction term between the national question variable and a dichotomous variable to indicate whether respondents identified Gosselin as being tied to the Conservative Party of Canada. The interaction is showed in Figure 4.4. The main finding is similar to the one for Montreal. That is, the ­heuristic clearly works by amplifying the effect of the national question when citizens associate Gosselin with the federalist Conservative Party of Canada. In fact, when they do not have that heuristic, the effect is negative but fails to reach conventional levels of statistical significance. However, among those who are very favourable to Quebec independence and who perceive Gosselin as tied to the Conservatives, the maximum effect is almost 40 percentage points less likely to support him. This is a huge effect. 7 He is also seen as tied with the CAQ at the provincial level, but in terms of the national question, the Conservative Party of Canada is clearly federalist while the CAQ is quite appealing for several independentists.

90  Jean-François Daoust

Discussion Quebec nationalists know that independence, if ever achieved, would come through the Quebec National Assembly, which is elected at the provincial level. But this has not prevented them from bringing their fight to the federal level, most famously through the Bloc Québécois. In this chapter I have analysed, for the first time, whether nationalism is a significant determinant of vote choice at the municipal level as well. First, the vast majority of voters had an opinion when they were asked to associate municipal candidates with provincial and federal political parties. In Montreal, voters clearly associated Denis Coderre with the federal and Quebec Liberal parties, which is not surprising considering his past career. His opponent, Valérie Plante, was associated with a party by fewer respondents, but when they did link her with a party, it tended to be with leftist and sovereigntist parties, namely the Parti Québécois and Québec Solidaire at the provincial level and the NDP at the federal one. In Quebec City, findings were mixed regarding Régis Labeaume but were quite clear for Jean-François Gosselin, who was associated with the Conservative Party of Canada. Overall, a significant part of the electorate had some sense of the partisan affiliation of the candidates. Second, attitudes toward Quebec independence were a significant determinant of vote choice in both mayoral elections. In Montreal, the maximum effect was very important (about 20 percentage points): citizens with more favourable views of Quebec independence were more likely to support Valérie Plante. In Quebec City, attitudes toward independence had a significant effect on two of the three candidates, and the magnitude of this impact was similar to what was found in Montreal. Labeaume enjoyed support among voters favourable to Quebec independence, while the opposite was the case with Gosselin, his main opponent. Third, throughout the analyses I included ideology in terms of the left/right cleavage as a benchmark in order to compare it to the impact of nationalism. Overall, it seems that ideology plays a somewhat greater role than nationalism in citizens’ vote choice calculus in municipal elections in Quebec. However, subgroups of voters who used a particular heuristic and who perceived some candidates as affiliated with clearly federalist parties were more strongly influenced by the national question. The effect, if any, was very limited when they did not have the cue, but turned out to be very important (more than a maximum effect of 30 percentage points) when the heuristic was available. Overall, the findings in this chapter indicate that taking nationalism into account in electoral studies in Quebec should not be restricted to the provincial and federal levels. Doing so deepens our understanding of citizens’ voting behaviour and electoral outcomes at the municipal level too.

Nationalism and Municipal Elections in Quebec  91

Appendix to Chapter 4 Table A4.1.  Multinominal logit regression models: reference = Labeaume Model 1 Gosselin Age Sex (woman) Education French Ideology Quebec Independence (support) Gosselin as conservative Gosselin as conservative x Quebec Independence Constant Guérette Age Sex (woman) Education French Ideology Quebec Independence (support) Gosselin as conservative Gosselin as conservative x Quebec Independence Constant Observations Pseudo R2

Model 2

Model 3

–1.32* (0.34) –0.59* (0.15) –0.57* (0.15) 0.58 (0.48)

–1.04* (0.37) –0.50* (0.17) –0.44* (0.16) 0.86 (0.52) 3.74* (0.48) –1.21* (0.27)

–0.99* (0.37) –0.37* (0.17) –0.52* (0.17) 0.75 (0.59) 3.56* (0.48) –0.43 (0.35) 1.06* (0.22) –1.98* (0.51)

0.27 (0.50)

–2.13* (0.66)

–2.38* (0.72)

1.01* (0.43) 0.13 (0.17) 0.82* (0.20) 0.21 (0.49)

1.22* (0.48) 0.14 (0.19) 0.68* (0.21) 0.11 (0.51) –4.33* (0.66) 0.20 (0.26)

1.46* (0.50) 0.25 (0.19) 0.53* (0.22) 0.02 (0.49) –3.99* (0.67) –0.32 (0.37) 0.41 (0.34) 0.65 (0.53)

–2.27* (0.58) 998 0.047

–0.36 (0.68) 998 0.185

–0.62 (0.68) 998 0.204

Note: Logit regression coefficients displayed. Robust standard errors in parentheses. + p < 0.10; * p < 0.05

92  Jean-François Daoust REFERENCES Bakvis, H., & Macpherson, L. G. (1995). Quebec block voting and the Canadian electoral system. Canadian Journal of Political Science, 28, 619–635. https://doi .org/10.1017/S000842390001934X Bastien, F., Bélanger, É., & Gélineau, F. (Eds.). (2013). Les Québécois aux urnes: les partis, les médias et les citoyens en campagne. Presses de l’Université de Montréal. Bélanger, É., Daoust, J.-F., Mahéo, V.-A., & Nadeau, R. (2022). Le nouvel électeur québécois. Presses de l’Université de Montréal. Bélanger, É., & Nadeau, R. (2009). Le comportement électoral des Québécois. Les Presses de l’Université de Montréal. Bélanger, É., Nadeau, R., Henderson, A., & Hepburn, E. (2018). The national question and electoral politics in Quebec and Scotland. McGill–Queen’s University Press. Bherer, L., & Breux, S. (2012). L’apolitisme municipal. Bulletin d’histoire politique, 21(1), 170–184. https://doi.org/10.7202/1011705ar Blais, A., Gidengil, E., Nadeau, R., & Nevitte, N. (2002). Anatomy of a Liberal victory: Making sense of the vote in the 2000 Canadian election. University of Toronto Press. Blais, A., & Nadeau, R. (1984). L’appui au Parti québécois: évolution de la clientèle de 1970 à 1981. In Jean Crête (Ed.), Comportement électoral au Québec (pp. 279–318). Gaëtan Morin. Boudreau, C., Elmendorf, C. S., & MacKenzie, S. A. (2015). Lost in space? Information shortcuts, spatial voting, and local government representation. Political Research Quarterly, 68(4), 843–855. http://dx.doi.org/10.1177 /1065912915609437 Couture, J. (2017). Mesurer les préférences budgétaires des maires et des mairesses au Québec à l’aide de Wordfish. In P.-M. Daigneault et F. Pétry (Eds.), Les idées, le discours et les pratiques politiques au prisme de l’analyse des données textuelles (pp. 309–322). Les Presses de l’Université Laval / Hermann. Daoust, J.-F. (2017). Le mystère de Québec. Politique et Sociétés, 36(2), 143–158. https://doi.org/10.7202/1040416ar Daoust, J.-F., & Dassonneville, R. (2018). Beyond nationalism and regionalism: The stability of economic voting in Canada. Canadian Journal of Political Science, 51(3), 553–571. https://doi.org/10.1017/S000842391800001X Daoust, J. F., & Jabbour, A. (2020). An extraordinary election? A longitudinal perspective of the Quebec 2018 election. French Politics, 18(3), 253–272. Drouilly, P. (2010). Une curieuse election. In M. Fahmy (Ed.), L’état du Québec 2010 (pp. 406–412). Boréal. Fournier, P., Cutler, F., Soroka, S., Stolle, D., & Bélanger, É. (2013). Riding the orange wave: Leadership, values, and issues in the 2011 Canadian election. Canadian Journal of Political Science, 46(4), 863–897. https://doi.org/10.1017 /S0008423913000875

Nationalism and Municipal Elections in Quebec  93 Gagnon, A.-G., & Boucher, F. (2017). Party politics in a distinct society: Two eras of block voting in Quebec. In A.-G. Gagnon and A. B. Tanguay (Eds.), Canadian parties in transition (4th Ed.) (pp. 277–295). University of Toronto Press. Gagnon, K. (2013, 3 September). Denis Coderre: le Rocky de la politique. La Presse. http://www.lapresse.ca/actualites/politique/201311/09/01-4708918 -denis-coderre-le-rocky-de-la-politique.php Lau, R. (2018, 3 July). Montreal mayor Valerie Plante defends decision to not walk in Canada Day Parade. https://globalnews.ca/news/4309411 /montreal-mayor-valerie-plante-defends-canada-day-parade-decision Riera, P. (2013). Voting differently across electoral arenas: Empirical implications from a decentralized democracy. International Political Science Review, 34(5), 561–581. http://dx.doi.org/10.1177/0192512112467216

5 Does the Left–Right Axis Matter in Municipal Elections? sandra breux , jérôme couture , and anne mévellec

In North America, a number of studies have attempted to explain ­municipal electoral behaviour while seeking to liberate themselves from the so-called Michigan model. In this context, several researchers have examined the role of ideologies in municipal politics, with quite ambivalent results. Political ideology is generally considered to play an important role in electoral behaviour, but with a few exceptions (Abrajano et al., 2005; Boudreau et al., 2015; Sances, 2018) it is rarely dominant (Hajnal & Trounstine, 2014). In Quebec, the traditional approach marked by the reformist movement at the municipal level appears to have led researchers to abandon this analytical approach a long time ago. For some, the reformist movements were indeed the origin point of the depoliticization of the Quebec municipal scene and the emergence of a form of apoliticism that pressed for business-type management of institutions (Bherer & Breux, 2012). The creation of municipal political parties in the wake of a 1978 bill of the Quebec government1 that allowed such parties to raise public funds just like the provincial ones (Mévellec & Tremblay, 2013) thus repoliticized the municipal scene. However, these parties had some unique features (Couture et al., 2018; Mévellec & Tremblay, 2013). The absence of ideological links with federal or provincial parties, weak internal organization, and the general absence of activism between elections all brought into question these parties’ relevance and real impact on the electorate. Moreover, the political debate in Quebec has been strongly structured around the question of national sovereignty. As a result, the Liberal/PQ model has been hardly applicable to the municipal level. The provincial

1 Loi concernant les élections de 1978 dans certaines municipalités et modifiant la Loi des cités et villes. Assemblée nationale du Québec, 23 June 1978.

Does the Left–Right Axis Matter in Municipal Elections?  95

ideological divide has found little purchase in the province’s cities in terms of individual candidates or municipal parties. The research conducted as part of the Canadian Municipal Election Study (CMES) offers an opportunity to re-examine this question. Following Daoust, who, in chapter 4 explored the role the national question played in the behaviour of municipal voters, in this chapter we examine how the ideological dimension manifests itself at the municipal level – specifically in the behaviour of voters. The actual ideological positioning of municipal political parties is debatable; that said, in the 2017 municipal elections several parties could easily be positioned on a right/left spectrum. In Quebec City, the Québec 21 party, a new player on the municipal scene, did not try to hide its right-wing values, while in Montreal, Projet Montréal was an environmentalist party that clearly positioned itself on the left of the political spectrum. Thus we will explore a new aspect of municipal electoral competition – namely, the influence of ideology on voters’ choices, a theme that has been well-researched at other levels of government (e.g., Blais et al., 2002). This chapter asks the following question: does ideology influence voters’ choices at the municipal level? To answer it, we begin by addressing the explanatory variables of individuals’ electoral choices as they relate to the presence of political parties; this includes examining sociodemographic variables as well as ideology and partisan attachment. We are mindful of the unique nature of municipal political parties in Quebec (see chapter 6). We then present our methodology, followed by our results based on the mayoral elections in Montreal and Quebec City. We will show that, contrary to what is often asserted about municipal elections, ideology played a major role in voter’ choices. We conclude by reflecting on the importance of political parties and the specific relationship between voters and those parties. Understanding the Voting Orientation at the Municipal Scale Understanding electors’ choices is a delicate operation. Indeed, as the different models of electoral participation show, there is no single variable that can explain their behaviour (Smets & van Ham, 2013). In the more specific matter of vote choice, Blais and colleagues (2002) write that “no single factor can provide an explanation for why voters voted as they did” (10). At higher levels of government, electoral sociology has highlighted the influence of several variables. Blais and colleagues (2002) have constructed an explanatory model with eight main variables, three of which are directly related to the presence of political parties:

96  Sandra Breux, Jérôme Couture, and Anne Mévellec

socio-demographic variables, ideology, and partisan attachment. These three broad categories are interesting, because, to our knowledge, they have never been systematically tested at the municipal level in Quebec. Socio-economic variables have long been used to understand electoral choices. However, Blais and colleagues (2002) note that in Canada these variables are generally viewed as having little explanatory weight, although some of them, such as gender and region, have become more predominant in recent decades (91). These authors demonstrated the importance of age and language in the federal Liberals’ 2000 success in Quebec. Unfortunately, such analyses have rarely been conducted at the municipal level. Also, most of these studies have not been based on individual-level survey data. Moreover, most of the research has tried to describe the voter instead of trying to understand how she or he votes. We know that the municipal elector tends to be a property owner, older, wealthier, and bettereducated than the average electorate (Oliver, Ha, & Callen, 2012). It is also known that age plays an important role (Nakhaie, 2006), with young people tending to abstain more from elections at this level of government (Dostie-Goulet et al., 2013). In Quebec, in a rare study of survey data, Breux and colleagues (2018) scaled down the importance assigned to the ownership variable by pointing out that first and foremost, it is information that seems to influence the act of going to the polls. Seen from that angle, being an owner is only one among many pathways to information. All of these surveys establish a profile of the voter yet do nothing to explain why people vote the way they do. The explanatory factors vary. For example, in the US context, Hajnal and Trounstine (2014, 2016) found that racial and class conflict dominated electoral behaviour. Other studies, including many Canadian ones, have emphasized location, especially the urban/suburban divide (Oliver & Ha, 2007). For example, in their analysis of the 2014 Toronto mayoral election, Bird and colleagues (2016) concluded that voting tended to be based on ethnic affinity and that the confluence of gender and race might influence vote choice. Yet at the turn of the 2000s, a number of studies – most of them focusing on higher levels of government – found that suburbanites tended to vote more to the right and urban voters more to the left (Walks, 2004). Villeneuve and colleagues compared voter choice at all three levels of government and concluded that “place of residence is one of the factors that explains political orientations, regardless of the socio-demographic attributes of individuals. In other words, the vote is partly explained by neighbourhood effects. In a very clear way, suburban neighbourhoods vote more to the right than

Does the Left–Right Axis Matter in Municipal Elections?  97

downtown neighbourhoods” (2007, 394).2 A few years earlier, Belley (2003) had likewise pointed out this difference, in his case between the city centre and the suburbs of Quebec City. Findings like this call on us to look at how values manifest themselves at the municipal level, especially since US studies indicate that although partisan identification and ideology are likely to influence vote choice, other factors do so to a greater extent (Hajnal & Trounstine, 2014). In Quebec, political ideology is often described in binary terms: “In Quebec, the cleavage between sovereigntists and federalists is so powerful that there is little room for other values to play a significant role” (Blais et al., 2002: 112). Such analyses have never been conducted at the municipal level, partially because of the lack of formal partisan ties between the municipal scene and other levels of government. Nevertheless, we know that some elected municipal politicians are affiliated with provincial or federal political parties. Mévellec and Tremblay (2016: 63) found that 58% of the women and 70% of the men in their sample of elected municipal government officials had a provincial partisan affiliation and that 35% of the women and 48% of the men had a federal one. Although these elected representatives do not display these affiliations during their municipal campaigns and such information may be unequally known by voters (as Daoust underscores in his contribution to this volume), this affiliation may nonetheless have influenced voters (Elmendorf & Schleicher, 2012a). In a similar vein, Stephenson and colleagues (2018) point out that insofar as some electors’ voting orientations are influenced by the candidate’s socio-demographic variables, partisanship and even family ties and background also play a “shortcut” role. There is a debate about the role of ideological preferences at the municipal level. Some argue that municipal issues are too technical to generate political momentum (Peterson, 1981) or simply do not induce traditional partisan cleavages (Kaufmann, 1998: 658). Others argue that all public policy choices are ideological (Elmendorf & Schleicher, 2012b), including at the municipal level (Graham, Philips, & Maslove, 1998). In this regard, the scope of powers and responsibilities of Montreal and Quebec City were expanded by the Quebec government in 2016, giving these governments greater say in economic development, land use planning, internal governance, relations with immigrants, and the regulation of commerce. Unlike provincial jurisdictions, municipal jurisdictions are not directly redistributive in nature; even so, they can be instrumental in ways that reflect liberal economic or interventionist ideals.

2 Author translation.

98  Sandra Breux, Jérôme Couture, and Anne Mévellec

This is especially the case with regard to issues such as public transit, urban planning, and recreation. Yet municipal politicians do not always express their ideals as such in their speeches and statements; they instead seem to emphasize policy implementation. With regard to independent local parties in Europe, Otjes (2018) writes that “because of [their] localist orientation, these parties are independent of the ideological left or right ... At the [same] time there is considerable programmatic and ideological heterogeneity among independent local parties” (309). In his study of the 1989 election in Quebec City, Belley highlighted the presence of different political cultures, which oscillated between popular reformism and managerial populism. Over time, however, these differences became blurred in the face of “the rise of neoliberal ideas advocating the reduction of the size and role of the state, and the concerns of a growing part of the population towards the protection of the environment and heritage in urban areas” (Belley, 1992: 6).3 The issue of municipal mergers then shifted the left/right debate toward a citycentre/suburban debate, making it more difficult for parties and candidates to position themselves, which made for a more apolitical vision of the municipal scene. Bherer and Breux (2012), however, have emphasized that being apolitical “does not do away with the ideological question. Apolitical attitudes can be found both on the right and on the left of the political spectrum. However, studies done in a European context and in some local circles have shown that the supposed apoliticism is sometimes ‘only for show.’ Would it be possible to arrive at a similar finding in Quebec?” (171).4 Moreover, Couture (2017) has found, based on a quantitative analysis of the content of budget speeches given by mayors across Quebec, that there is a gap between conservative and liberal discourses with regard to taxation issues. The topic of municipal political parties thus leads us to wonder precisely what relationship voters have with these organizations. Explanatory voting models have already shown the importance this variable holds for voters at other levels of government. These models emphasize that a voter may well make her or his choice before the campaign even starts, such is the importance of their partisan attachment. Blais and colleagues (2002) found that this concept explains Canadians’ choices quite well; that is why Clarke and colleagues (1979) developed the notion of durable versus flexible partisan attachment. While Blais and colleagues

3 Author translation. 4 Author translation.

Does the Left–Right Axis Matter in Municipal Elections?  99

postulate that flexible partisan attachment may undermine durable partisan attachment, the fact remains that the concept of flexible partisan attachment seems particularly applicable at the municipal level, because of the nature of the political parties that are competing. In fact, municipal political parties in Quebec “are, for the most part, more ephemeral political teams than highly organized partisan structures” (Bherer & Breux, 2012: 172).5 In addition, the average life of a municipal political party is only seven years, which limits the potential leverage of partisan attachment. However, these general data must be qualified in light of certain figures. In cities with 20,000 to 500,000 inhabitants, there is a certain rootedness of parties (mainly victorious) from one election to the next (Mévellec, 2014). Collin and colleagues (2009) similarly noted that among elected officials, partisan attachment was unstable. However, recent research on cities with populations between 20,000 and 500,000 tends to emphasize the permanence in partisan attachment among elected municipal officials (Chiasson et al., 2017). To our knowledge, though, no study has assessed the impact on voters of partisan defections by elected officials. Finally, most municipal political parties bear the name of their leader or the city, which makes their ideological distinctions difficult to grasp, at least for the neophyte. This tends to centre these organizations around their leader rather than around a corpus of ideas. Collin (2011) specifies that “each municipal team and each municipal party is essentially defined as a unique political experience that revolves around its candidates and local issues” (341).6 All of this suggests that partisan attachment, insofar as it exists in Quebec at the municipal level, is probably expressed differently at the municipal level than at other levels because of the particularities of the political parties on the political chessboard at this scale (on this, see also chapter 6). In addition, there are the various corruption scandals (Bherer & Breux, 2016), which at times have damaged the reputation and relevance of these formations in the eyes of the voters. These general findings, however, are not necessarily applicable to Quebec City and Montreal, which have distinctive histories on the Quebec political landscape. These two municipalities have had municipal political parties since 1960, long before the provincial legislature formally recognized them. Despite this long presence, clear ideological claims have been the exception among them rather than the rule. Moreover, since the early 2000s, there has been a trend in municipal

5 Author translation. 6 Author translation.

100  Sandra Breux, Jérôme Couture, and Anne Mévellec

politics away from the political party format and toward the valorization of independent candidates. Nonetheless, the municipal parties and some of their members still tend to come from certain provincial or federal parties (Collin, 2011). These observations concerning the question of whether ideology influences voters’ choices at the municipal level have led us to propose two hypotheses: • H1: Electors who view themselves as being ideologically on the right will vote for the candidate who is ideologically on the right. The assumption that voters who are more left-leaning will vote for the candidate on the left and that those who are more right-leaning will vote for a candidate on the right is a classic hypothesis from political science (Downs, 1957) that has been widely demonstrated (Smets & van Ham, 2013). However, several factors undermine the applicability of such a hypothesis at the municipal level, thereby confirming the relevance of carrying out a new empirical analysis. As Collin (2011) points out: “The 2000s will nevertheless stand out for having ushered in some significant and lasting changes in the municipal scene in Quebec, at least in the big cities. The rise of entrepreneurial town halls whose style and type of politics make the headlines, the return to political parties that are more civic than programmatic and militant” (346).7 Moreover, the absence of any reference to higher ideological affiliations or any left/ right debate suggests that such referents do not sway voters. • H2: The larger the ideological gap between an elector and a candidate, the lower the probability of voting for that candidate. This hypothesis is related to the idea that an elector is likely to choose a candidate who shares his or her ideological affiliation over one who does not (Boudreau et al., 2015). The effect of ideology on voting could find a mediating effect through the ideological distance between the candidates and the voters. Methodological Design and Analysis To answer our research question, we set up two specific research designs. In the case of Montreal, the design is based on a logistic regression with the goal of understanding the support given to the incumbent candidate. The dependent variable is coded as follows: 1 = Vote for Denis Coderre and 0 = Vote for Valérie Plante. We estimate three models.

7 Author translation.

Does the Left–Right Axis Matter in Municipal Elections?  101

Model A considers the effect of ideology on voting; Model B assesses the effect of ideology on voting when controlling for partisan attachment8; and Model C considers the effect of ideology when controlling for the gap between the ideology of the respondent and that of the candidates. Each of the models controls for socio-demographic variables including age, age squared, gender, education, homeownership, and mother tongue. Our research design for Quebec City is based on a multinomial logistic regression. In this case, the dependent variable is a non-dichotomous dummy variable. The reference category applies to the respondents who voted for the incumbent Régis Labeaume. Our aim was to try to understand the support for the incumbent, and the analysis compares those who voted for Gosselin versus Labeaume and those who voted for Guérette versus Labeaume. The three models and the socio-economic control variables are the same as for the city of Montreal. Montreal For the city of Montreal, the results show that 38% of respondents voted for Denis Coderre and 62% for Valérie Plante. Our analysis did not include the 54 respondents who voted for another candidate or who preferred not to reveal their vote, since they were not numerous enough. In terms of ideology, the average score of Montreal respondents is 4.61, which comprises a perfect centre-left distribution on a scale of 0 to 10, with 0 being left and 10 right. Only 19% of respondents say they do not feel close to a municipal political party in Montreal,9 which demonstrates the importance of political parties in understanding that city’s political dynamics. Indeed, that is the first original result of this research. At the same time, 21% of the respondents state that they identify fairly or very strongly with the Coderre team, and 23% state that they identify fairly or very strongly with Projet Montréal (variables Strong Coderre supporter and Strong Projet

8 In the municipalities of Montreal and Quebec City, political parties structure the electoral offer. It is thus possible to think, in keeping with the Michigan model, that this structuring of the electoral offer allows electors to make their choice before the election, based on the attachment they have to their party; and that if any relationship was identified between the ideology and the vote, it would be less pronounced if one controlled for the partisan factor. Nevertheless, the fact that most of these political formations are centred around a mayoral candidate is likely to cast this potential for partisan attachment in a different light. 9 Variable is No partisan attachment.

102  Sandra Breux, Jérôme Couture, and Anne Mévellec Table 5.1.  Montreal: Descriptive analysis Variables

Minimum

Maximum

Mean

S.D.

Vote Denis Coderre (vs Plante) Age Age2* Male French (first language) University degree Completed high school (or less) Property owner Ideology Ideology Denis Coderre Ideology Valérie Plante No partisan attachment Strong Équipe Coderre supporter** Strong Projet Montréal supporter** Ideological deviation Denis Coderre Ideological deviation Valérie Plante

0 18 324 0 0 0 0 0 0 0 0 0 0 0 0 0

1 89 7,921 1 1 1 1 1 10 10 10 1 1 1 10 10

0.38 50.0 2,724.7 0.44 0.56 0.49 0.16 0.32 4.61 6.37 3.88 0.19 0.21 0.23 2.87 1.85

– 14.89 1,482.22 – – – – – 2.18 2.22 2.06 – – – 1.85 1.83

* The effect of age on voting may be curvilinear. One method of testing this effect is to introduce age squared in the equation. ** The scores are not exactly the same as those presented in chapter 6 of this book, since non-partisan respondents are considered here.

Montréal supporter). Each of the two parties can therefore count on a fairly comparable partisan base. For the purposes of the analysis, the variables No partisan attachment, Strong partisan Coderre and Strong partisan Projet Montréal must be interpreted in relation to the 39% of respondents who report having an affiliation with a Montreal municipal political party, albeit to a weaker degree. Respondents were also asked to position candidates on the same left/ right axis (0 to 10). Here, Coderre is positioned on the right with an average of 6.37 and Plante on the left with an average of 3.88. To confirm Hypothesis 1, we expect ideology to be positively correlated with the vote. The difference in absolute value between the position of the respondent and that of the candidate is illustrated by the Ideology Deviation variable. On average, respondents have a gap of 2.87 with Coderre and 1.85 with Plante. Quebec City Some 48% of respondents voted for Régis Labeaume, 32% for JeanFrançois Gosselin, and 20% for Anne Guérette. Our analysis did not include the 56 respondents who voted for another candidate or who preferred not to reveal their vote, since they were not numerous enough.

Does the Left–Right Axis Matter in Municipal Elections?  103 Table 5.2.  Quebec City: Descriptive analysis Variables

Minimum

Vote Jean-François Gosselin 0 Vote Anne Guérette 0 Age 19 361 Age2 Male 0 French (first language) 0 University degree 0 Completed high school (or less) 0 Property owner 0 Ideology 0 Ideology Anne Guérette 0 Ideology Régis Labeaume 0 Ideology Jean-François Gosselin 0 No partisan attachment 0 Strong partisan for Démocratie Québec* 0 Strong partisan for Équipe Labeaume* 0 Strong partisan for QC21* 0 Ideology Deviation Anne Guérette 0 Ideology Deviation Régis Labeaume 0 Ideology Deviation Jean-François Gosselin 0

Maximum

Mean

S.D.

1 1 92 8,464 1 1 1 1 1 10 10 10 10 1 1 1 1 10 10 10

0.32 0.20 53.7 3,101.3 0.55 0.96 0.47 0.13 0.74 5.40 3.88 5.53 7.03 0.12 0.09 0.27 0.17 2.87 2.63 2.87

– – 14.68 1,544.53 – – – – – 2.16 2.31 2.33 2.35 – – – – 2.44 2.27 2.48

As for the ideology, the average of the respondents of Quebec City is 5.40, which means a distribution slightly to the right of centre on the scale of 0 to 10, with 0 being left and 10 right. In all, 12% of respondents say they do not feel close to a municipal political party in Quebec City, which is lower than in Montreal.10 This once again demonstrates the importance of political parties in understanding political dynamics at the municipal level in Quebec’s two largest cities. Nine per cent of respondents say they are fairly or very close to the party Démocratie Québec, 27% to the Labeaume team, and 17% to QC21 (variables Strong Démocratie Québec supporter, Strong Team Labeaume supporter, and Strong QC21 supporter, respectively). This means that the Labeaume team enjoys a larger political base than its opponents; QC21, a brand-new party, has a fairly large political base; and Démocratie Québec, the official opposition before the 2017 election, finds itself with a weak party base. For the purposes of this analysis, the variables No partisan attachment, Strong Démocratie Québec supporter, Strong Team Labeaume

10 Variable is No partisan attachment.

104  Sandra Breux, Jérôme Couture, and Anne Mévellec

supporter, and Strong QC21 supporter should be interpreted in relation to the 35% of respondents who say they are close to a municipal political party, albeit to a weaker degree or hardly at all. Respondents were also asked to position candidates on the left/right axis (0 to 10). Guérette is positioned on the far left, with an average of 3.88; Labeaume slightly to the right of centre, at 5.53; and Gosselin on the far right, with an average of 7.03. To confirm Hypothesis 1, we ­expect the ideology to be negatively correlated when comparing those who voted for Labeaume and Guérette and positively correlated when comparing those who voted for Labeaume and Gosselin. The average difference in absolute value between respondents and candidates is 2.63 for Régis Labeaume and 2.87 for Anne Guérette and Jean-François Gosselin. Results As we have done in the methodology section of this chapter, we present the results by city. Montreal Our results show that ideology partly explains the vote of the respondents in the Montreal election: the more a respondent is on the right on the axis, the more likely he or she is to vote for Coderre, which confirms Hypothesis 1. Specifically, the marginal effect (Dy/Dx) found in Model A shows that each additional point on the scale increases the probability of voting for Coderre by 6 percentage points. Part of the effect of the ideology seems to result from partisan attachment, according to Model B. Indeed, the Ideology variable remains significant, but its marginal effect (Dy/Dx) on the vote decreases by about 0.5 to 3 percentage points per additional point on the left/right scale. In addition, respondents who do not have a partisan attachment are not more likely to vote for one or the other candidate, as shown by the non-significant coefficient for the variable No partisan attachment. However, respondents who feel fairly or very close to one or the other of the political parties are more likely to vote for the party’s mayoral candidate. More specifically, supporters of the Coderre team have a 43-point higher probability of voting for Coderre, while Projet Montréal supporters have a 41-point lower probability of voting for the latter, according to the marginal effects reported in Model B of Table 5.3. Hence, the mobilization of supporters was quite comparable for both parties. The effect of ideology also seems to derive in part from the candidate’s positioning on the left/right scale, according to Model C. The Ideology

Table 5.3.  Logistic regression Montreal vote Dependent variable: 1 = Vote for Denis Coderre; 0 = Vote for Valérie Plante

A)  Ideology (left–right) Coef. (S.E.)

Constant

0.05 (1.09)

B)  Ideology mediated by partisanship

Dy/Dx

Coef. (S.E.)

Dy/Dx



–0.33 (0.56)



–0.02 0.0002 n.s. –0.10 n.s. n.s. n.s. 0.06

–0.12* (0.05) 0.001** (0.0005) –0.43* (0.21) –0.22 (0.24) 0.61* (0.25) –0.18 (0.30) 0.00006 (0.23) 0.17** (0.04) –0.36 (0.29) 1.96*** (0.27) –2.69*** (0.40)

–0.02 0.0003 –0.08 n.s. 0.12 n.s. n.s. 0.03 n.s. 0.43 –0.41

C)  Ideology mediated by the self-rated ideology of candidates Coef. (S.E.) 0.06 (1.25)

Dy/Dx –

Independent variables Age Age2 Male French (first language) University degree Completed high school (or less) Property owner Ideology No partisan attachment Strong Équipe Coderre supporter Strong Projet Montréal supporter Denis Coderre Ideology deviation Valérie Plante Ideology deviation

–0.09** (0.04) 0.001** (0.0004) –0.12 (0.17) –0.42* (0.20) 0.38 (0.20) –0.18 (0.30) 0.003 (0.09) 0.25*** (0.04)

N Log pseudo-likelihood LR Chi-square Pseudo R2

653 –396.15 59.81*** 0.07

* p < 0.1; ** p < 0.05; *** p < 0.01

–0.06 (0.05) 0.0007 (0.0004) –0.06 (0.20) –0.29 (0.23) 0.39 (0.22) –0.20 (0.35) –0.13 (0.21) 0.13* (0.06)

–0.47*** (0.06) 0.41*** (0.06) 653 –282.22 287.68*** 0.34

582 –330.24 85.85*** 0.21

n.s. n.s. n.s. n.s. n.s. n.s. n.s. 0.03

–0.10 0.08

106  Sandra Breux, Jérôme Couture, and Anne Mévellec

variable remains significant, but its marginal effect (Dy/Dx) on the vote decreases once again by about 0.5 to 3 points per additional point on the left/right scale, which confirms Hypothesis 2. The larger the ideological gap between a candidate and a respondent, the less likely the latter is to vote for that candidate. The results show that each point of difference with Coderre decreases the probability of voting for this candidate by 10 points and that each point of difference between the respondent and Plante increases the probability of voting for Coderre by 8 points. As for the socio-economic variables, the results indicate that support for Denis Coderre is found among both young people and older people, since the age variable has a negative and significant coefficient in models A and B; age squared likewise has a positive and significant coefficient in these two models. However, age is not significant in Model C. As for the other control variables, none show stable results in more than one model. Quebec City Once again, ideology partly explains how respondents voted during the election in Quebec City – indeed, more so than in Montreal. The more a respondent is on the right side of the axis, the more likely she or he is to vote for Jean-François Gosselin, and the more she or he is on the left, the higher the probability of voting for Anne Guérette, which confirms Hypothesis 1. More specifically, the marginal effect (Dy/Dx) calculated in Model A shows that each additional point on the scale increases the probability of voting for Gosselin rather than for Régis Labeaume by 10  percentage points. In the same vein, each additional point on the left/right axis decreases the probability of voting for Guérette rather than for Labeaume by 6 percentage points. A small part of the effect of ideology seems to derive from partisan attachment, according to Model B. Indeed, the Ideology variable remains highly significant, although its marginal effect (Dy/Dx) on the vote diminishes somewhat, decreasing to 9 percentage points per additional point on the left/right scale for a vote for Jean-François Gosselin. The mediation effect is larger for Guérette – namely a decrease by one half, which is 3 points per point on the scale for Guérette. The effect of mediation on ideology seems lesser in the case of a vote for Gosselin. Furthermore, the partisan effect of moderation is somewhat less pronounced in Quebec than in Montreal. Moreover, respondents who do not have a partisan attachment are not more likely to vote for either one or the other candidate, as shown by the non-significant coefficient for the variable No partisan attachment.

Table 5.4.  Results for Quebec City

A) Ideology (left–right) Guérette Reference: Vote Coef. (S.E.) Dy/Dx for Labeaume

Gosselin Coef. (S.E.)

C) Ideology mediated by the self-rated ideology of candidates

B) Ideology mediated by partisanship Guérette

Dy/Dx n.s.

Coef. (S.E.)

Dy/Dx

Gosselin Coef. (S.E.)

Dy/Dx

Guérette Coef. (S.E.)

Dy/Dx

Gosselin Coef. (S.E.)

Dy/Dx

Age

0.05 (0.04)

n.s.

0.07 (0.04)

0.05 (0.05)

n.s.

0.06 (0.05)

n.s.

0.06 (0.05)

n.s.

0.08 (0.05) n.s.

Age2

0.00 (0.00)

n.s.

–0.00 (0.00)

–0.001*

0.00 (0.00)

n.s.

–0.00 (0.00)

n.s.

0.00 (0.00)

n.s.

–0.00 (0.00) n.s.

Male

–0.18 (0.19)

n.s.

0.41 (0.16)

0.09**

–0.18 (0.21)

n.s.

0.21 (0.20)

n.s.

–0.26 (0.21)

n.s.

0.39 (0.21) n.s.

French (first language)

0.29 (0.52)

n.s.

1.00 (0.51)

0.15*

0.03 (0.56)

n.s.

0.74 (0.62)

n.s.

–0.00 (0.58)

n.s.

1.52 (0.74) 0.21***

University degree

0.61 (0.20)

0.09***

–0.12***

0.24 (0.23)

n.s.

–0.61 (0.22) –0.14** –0.33 (0.30)

n.s.

High school (or less)

–0.83 (0.42) –0.10*

–0.49 (0.17)

0.59 (0.23)

0.08**

–0.29 (0.21) n.s.

–0.93 (0.58)

n.s.

–0.04 (0.32) n.s.

–0.00 (0.26)

n.s.

–0.02 (0.25) n.s.

–0.20 (0.24)

n.s.

–0.83 (0.46)

n.s.

0.19 (0.20)

n.s.

–0.35 (0.26)

n.s.

0.19 (0.25)

n.s.

0.43 (0.04)

0.10*** –0.26 (0.06) –0.03***

0.39 (0.05)

0.09*** –0.24 (0.08) –0.04***

Property owner

–0.10 (0.21)

Ideology

–0.39 (0.05) –0.06***

n.s.

No partisan attachment









0.81 (0.27)

0.06**

0.47 (0.28)

n.s.









Strong Dém. Québec supporter









3.16 (0.41)

0.27***

0.61 (0.58)

n.s.









Strong É. Labeaume supporter









–1.78 (0.32) –0.12***

–2.07 (0.27) –0.37*** –







Strong QC21 supporter









–3.12 (2.04)







n.s.

3.08 (0.38)

0.67*** –

0.30 (0.07) 0.07***

(Continued )

Table 5.4.  Results for Quebec City (Continued)

A) Ideology (left–right) Guérette Reference: Vote for Labeaume Coef. (S.E.) Dy/Dx

B) Ideology mediated by partisanship

Gosselin

Guérette

Gosselin

C) Ideology mediated by the self-rated ideology of candidates Guérette

Coef. (S.E.)

Dy/Dx

Coef. (S.E.)

Dy/Dx

Coef. (S.E.)

Dy/Dx

Coef. (S.E.) –0.52 (0.08)

Dy/Dx

Guérette Ideology deviation Labeaume Ideology deviation Gosselin Ideology deviation

































0.52 (0.07) –0.12***

















0.18 (0.12)

N Log PLKH LR Chi-square Pseudo R2

1 051 –896.97 390.673*** 0.18

* p < 0.1; ** p < 0.05; *** p < 0.01

1 051 –635.65 913.29*** 0.42

828 –622.20 502.32*** 0.40

0.13***

n.s.

Gosselin Coef. (S.E.) 0.13 (0.12)

Dy/Dx n.s.

0.56 (0.06) –0.12***

–0.67 (0.06)

0.11***

Does the Left–Right Axis Matter in Municipal Elections?  109

The result obtained here is similar to the one presented for the city of Montreal. However, respondents who feel fairly or very close to one or the other of the political parties are more likely to vote for the party’s mayoral candidate. Specifically, supporters of Démocratie Québec have a 27-point higher probability of voting for Guérette, while supporters of Québec 21 have a 67-point higher probability of voting for Gosselin, according to the marginal effects reported in Model B of Table 5.4. Hence, the mobilization of supporters was stronger for QC21 than for Démocratie Québec. In the same vein, supporters of the Labeaume team have a probability of less than 37 points to vote for Gosselin and of less than 12 points to vote for Guérette. The most interesting effect is that the variable Strong Démocratie Québec supporter is not negatively and statistically related to a vote for Gosselin and that the variable Strong QC21 supporter is not negatively correlated with a vote for Guérette. This indicates that the vote for these two candidates is not transferable. Indeed, the number of voters for whom the choice was between Guérette and Gosselin was too low. As a result, for some voters the election was between Guérette versus Labeaume and for others it was between Gosselin and Labeaume. Part of the effect of ideology also seems to derive from the candidate’s positioning on the left/right scale, according to Model C. The Ideology variable remains significant, but its marginal effect (Dy/Dx) on the vote decreases by approximately one third, dropping to 4 percentage points per additional point on the left/right scale for Guérette and by 7 percentage points per point on the scale for Gosselin. This means that mediation by ideology is less significant in Quebec City than in Montreal (see previous section). Once again, the wider the ideological gap between a candidate and a respondent, the less likely the latter is to vote for that candidate. The results show that each point of difference with Labeaume decreases the probability of voting for this candidate by 12 points; that each point of difference with Guérette increases the probability of voting for Labeaume by 13 points; and that each point of difference with Gosselin increases the probability of voting for Labeaume by 11 points. Again, Model C shows that a vote for Gosselin or Guérette is not transferable from one to the other. Results for the socio-demographic variables show that support for Gosselin is found among men, mother-tongue francophones, and those with a vocational or technical education. Support for Guérette is mainly found among university graduates. Labeaume seems to pull votes from across the board, with coefficients inverted for a number of variables between the Guérette-versus-Labeaume models and the

110  Sandra Breux, Jérôme Couture, and Anne Mévellec

Gosselin-versus-Labeaume models. Note that Labeaume seems to have made more gains with older people and people with a high school degree or less. Discussion These different results allow us to contribute to the reflections on ­Canadian and, more specifically, Quebec municipal electoral sociology. More specifically, two important results emerge from our analysis. First, our study shows clearly that ideology played a major role in the election results in Montreal and Quebec City. In Montreal, right-wing voters were more likely to vote for Denis Coderre and left-wing voters more likely to vote for Valérie Plante. Plante’s victory can be explained, at least in part, by the fact that she was perceived as more toward the centre of the political spectrum than Coderre. Thus, her positioning was more in line with that of Montrealers. In addition, she was able to attract the vote of the most left-leaning voters, while Coderre was unable to do the same with the most right-leaning voters. In Quebec City, the situation is the same. Voters on the right tended to vote for Gosselin, those on the left for Guérette, and those in the centre – many – for Labeaume. In this sense, Labeaume’s victory can be explained by the fact that he was perceived as more of a political centrist than the other candidates. His position on the centre-right was almost perfectly aligned with that of Quebec City’s voters. This allowed him to attract voters to his right and left, something the other two candidates were unable to do. As emphasized earlier, this role of ideology tends to suggest that apoliticism is not the same as ideological neutrality. However, one can hypothesize that elections are won at the centre, which perhaps echoes a certain apolitical view of things: “Apoliticism reflects the communitarian ideal represented by the municipal level, an ideal that should not tolerate any obstacle between the mayor and his or her citizens, no conflict or debate of ideas. The municipal scale would by definition be harmonious and consensual. In this context, apoliticism alone constitutes a political project that allows the person claiming to represent it to speak on behalf of all” (Bherer & Breux, 2012, 180).11 This hypothesis challenges the definition of the centre in politics, both at the municipal level and at other levels of government. The fact remains that Plante was victorious in Montreal because she drew most left-leaning voters. Based on our research, then, it seems fair to say that the notion that municipal politics

11 Author translation.

Does the Left–Right Axis Matter in Municipal Elections?  111

are non-ideological no longer holds. Indeed, there are left and right ways of doing politics in Montreal and Quebec City, and some voters are aware of this and align their vote accordingly. Second, we have shown that part of the effect of ideology on voting relates to partisan attachment. It seems that in Montreal and Quebec City, political parties – whose relevance is often questioned in the media – were one driver of electoral behaviour. It seems that parties forge links with voters, and few voters in either city claimed to be truly independent. This is significant, because previous studies had found that municipal political parties had played a somewhat contradictory role in electoral participation. Our study clearly underscores that parties, by creating links with the electorate, help structure electoral behaviour by ensuring a “legible” offer to voters (see also chapter 6). Moreover, in Montreal both parties had a strong support base of roughly the same size. They also received very strong and comparable support from their supporters. This new information helps us understand the elections in Montreal and Quebec City. In Montreal, this result contradicts a statement by Projet Montréal, which, at its April 2018 convention, attributed its victory to the greater mobilization of its supporters (Projet Montréal, 2018: 34). In Quebec City, the effect of ideology on voting was stronger for Québec 21 than for the other parties. The same applied to the effect of the ideological “fit” and the ability to attract the vote of outsiders. Thus, Labeaume had a larger partisan base than his opponents and was more efficient than Guérette at getting it out; but he was less effective than Gosselin. However, Labeaume’s supporters were much more numerous than Gosselin’s. Nonetheless, the links between the municipal parties and the electorate remain to be explored. Some chapters in this volume set out to do so. However, the question of partisan attachment raises several methodological and theoretical challenges. The CMES asked the following question: “Do you usually think of yourself as a [...]?,” and provided names of the local parties as response options. One might wonder what this question actually covers. In other words, what does it mean to identify with a party that has existed for only a few months? Table 6.1 in chapter 6 shows the founding dates for the political parties that competed in 2017 in Quebec City and Montreal. Many have short-lived histories, and some of them had never been in power, thus preventing voters from basing their party identification on evaluations of past performance. Chapter 6 also suggests there are similarities between a new sports team and a municipal political party; a comparison of such attachments could point toward an explanation. It seems that political parties on the municipal scene try to foster personalized attachment with voters. Indeed, it is clear that besides laying

112  Sandra Breux, Jérôme Couture, and Anne Mévellec

out ideological platforms, parties try hard to personalize their electoral campaigns. Thus, one may wonder if, in the end, the partisan proximity displayed by voters is not ultimately a type of proximity with the candidates themselves. Consider here that municipal political parties often include the candidate’s name in their party name. According to some observers, these “alphabet” parties, devoid of ideology, are simply campaign organizations. But it is also conceivable that the use of candidates’ names in this way helps to foster a closer attachment to the party itself. This is implied by the practices of the already established political parties of Gatineau and Sherbrooke, which, only for the duration of the electoral campaign of 2017, added the name of the incumbent mayor to their party name (Action Gatineau – Équipe Pedneaud-Jobin; Renouveau Sherbrookois – Équipe Sévigny). More generally, Quebec municipal election campaigns take the form of a “quasi-presidential election” (Bherer & Breux, 2012, 173), so that all eyes and spotlights are turned almost exclusively on the race for mayor. For example, although Projet Montréal had been in existence since 2004, it was very much Plante’s campaign that had prominence. Thus, municipal political parties seem to resort to the strategic use of mayoral power. In this way, ideological positioning is mediated by the personality of the party leader/candidate for mayor. The municipal level provides an opportunity to question the notion of partisan attachment in a context that is less standardized by the presence of these parties than are the federal or provincial levels. Elmendorf and Schleicher (2012a) point out that “partisanship for some voters is more affective than informational. These ‘Michigan voters,’ as we have elsewhere dubbed them, tend to conform their comments and beliefs to their party identification rather than the other way around” (1854). This explanation would appear to be quite satisfactory in terms of how voters position themselves on the left-right axis and how they position the candidates, given the nature of municipal political parties and the fact that they are focused on their leader. Our analysis leaves a number of elements unresolved. With regard to socio-demographic variables, in the absence of the emergence of a clear profile, analysing ideology by place of residence would probably be an interesting avenue of research. Similarly, the other explanatory variables of electoral choice presented by Blais and colleagues (2002) also need to be examined in more detail (economic perceptions and beliefs about issues, evaluations of government performance, leader evaluations, strategic considerations). In this sense, it would be interesting for future surveys to integrate the “incumbent” variable into the models, in order to see how partisan proximity is negotiated when

Does the Left–Right Axis Matter in Municipal Elections?  113

an incumbent candidate is present, particularly when he or she is not re-elected, as was the case in Montreal. More generally, party ideology has been shown to have an effect on electoral choice. ­Therefore, the fragile presence of municipal party systems in Montreal and Quebec City calls into question the ideological positions adopted by these parties, as well as the centripetal and centrifugal forces that may ­influence them. Together these results invite us to question once again the specificity of the municipal political scene. As soon as we grasp that the municipal scene is ideological, it loses some of its specificity. This first work therefore opens up a new research project: Is there continuity between ideological choices at the municipal level and those made at other levels of government? Moreover, does the positioning on the left/right axis by the voters correspond to the values of the party? REFERENCES Abrajano, M., Nagler, J., & Alvarez, R. M. (2005). A natural experiment of race-based and issue voting: The 2001 city of Los Angeles elections. American Politics Quarterly, 58(2), 203–218. http://dx.doi.org/10.1177 /106591290505800202 Belley, S. (1992). Les partis politiques municipaux et les élections municipales de 1986 à Montréal et de 1989 à Québec. Politique, 21, 5–35. https://doi.org /10.7202/040711ar Belley, S. (2003). Forces et faiblesses des acteurs de la nouvelle scène municipale. Télescope, 10(2), 24–34. Bherer, L., & Breux, S. (2012). L’apolitisme municipal. Bulletin d’Histoire politique, 21(1), 170–184. http://dx.doi.org/10.7202/1011705ar Bherer, L., & Breux, S. (2016). Réflexions sur les conditions d’exercice de la démocratie municipale. In P. Brissette (Ed.), Corruption (pp. 75–94). Léméac. Bird, K., Jackson, S. D., McGregor, R. M., Moore, A., & Stephenson, L. (2016). Sex (and ethnicity) in the city: Affinity voting in the 2014 Toronto mayoral election. Canadian Journal of Political Science, 49(2), 359–383. https://doi.org /10.1017/S0008423916000536 Blais, A., Gidengil, E., Nadeau, R., & Nevitte, N. (2002). Anatomy of a Liberal victory: Making sense of the vote in the 2000 Canadian election. University of Toronto Press. Boudreau, C., Elmendorf, C. S., & MacKenzie, S. A. (2015). Lost in space? Information shortcuts, spatial voting, and local government representation. Political Research Quarterly, 68(4), 843–855. http://dx.doi.org/10.1177 /1065912915609437

114  Sandra Breux, Jérôme Couture, and Anne Mévellec Breux, S., Couture, J., & Gélineau, V. (2018, March). Le rôle de l’information dans la participation électorale individuelle à l’échelle municipale. Conférence sur invitation, Ministère des Affaires municipales et de l’Occupation du Territoire, Québec. Chiasson, G., Gauthier, M., & Mévellec, A. (2017). Partisanisation ou politisation des municipalités québécoises? ST 36 Le pouvoir local, entre politisation et dépolitisation. Congrès de l’Association française de science politique, Montpellier, France. Clarke, H. D., Jenson, J., Le Duc, L., & Pammett, J. H. (1979). Political choice in Canada. McGraw-Hill Ryerson. Collin, J.-P. (2011). Quel avenir pour la démocratie municipale québécoise? In S. Breux & L. Bherer (Eds.), Les élections municipales au Québec: enjeux et perspectives (pp. 311–351). Presses de l’Université Laval. Collin, J.-P., & Bherer, L. with Breux, S., & Plourde, J.-S. (2009). Le mode de scrutin proportionnel à l’échelle municipale. INRS-UCS. Couture, J. (2017). Mesurer les préférences budgétaires des maires et des mairesses au Québec à l’aide de Wordfish. In P.-M. Daigneault & F. Pétry (Eds.), Les idées, le discours et les pratiques politiques au prisme de l’analyse des données textuelles (pp. 309–322). Les Presses de l’Université Laval/Hermann. Couture, J., Breux, S., & Bherer, L. (2018). Political accountability and responsiveness: What is the role of municipal political parties? In S. Breux & J. Couture, Accountabilty and responsiveness at the municipal level: Views from Canada (pp. 23–48). McGill–Queen’s University Press. Dostie-Goulet, E., Blais, A., Fournier, P., & Gidengil, E. (2013). L’abstention sélective, ou pourquoi certains jeunes qui votent au fédéral boudent les élections municipales. Canadian Journal of Political Science, 45(4), 909–927. https://doi.org/10.1017/S0008423912001084 Downs, A. (1957). An economic theory of democracy. Harper and Row. Elmendorf, C., & Schleicher, D. (2012a). Districting for a low-information electorate. Yale Law Journal, 121(7), 1846–1886. https://digitalcommons.law .yale.edu/ylj/vol121/iss7/4 Elmendorf, C., & Schleicher, D. (2012b). Informing consent: Voter ignorance, political parties, and election law. UC Davis Legal Studies Research Paper no. 285, 1–70. https://doi.org/10.2139/ssrn.2010115 Graham, K., Philips, S. D., & Maslove, A. (1998). Urban governance in Canada. Harcourt Brace. Hajnal, Z., & Trounstine, J. (2014). What underlies urban politics? Race, class, ideology, partisanship, and the urban vote. Urban Affairs Review, 50(1), 63–99. http://dx.doi.org/10.1177/1078087413485216 Hajnal, Z., & Trounstine, J. (2016). Race and class inequalities in local politics. APSA Task Force on Racial and Class Inequalities in the Americas, 11, 1–17.

Does the Left–Right Axis Matter in Municipal Elections?  115 Kaufmann, K. M. (1998). Racial conflict and political choice: A study of mayoral voting behaviour in Los Angeles and New York. Urban Affairs Review, 33(5), 655–685. https://doi.org/10.1177%2F107808749803300503 Mévellec, A. (2014). Did the 2013 municipal elections destabilize municipal politics in Quebec? Canadian Journal of Urban Research / Revue canadienne de recherche urbaine, 23(2), 18–37. Mévellec A., & Tremblay, M. (2013). Les partis politiques municipaux: la “Westminsterisation” des villes du Québec? Recherches sociographiques, 54(2), 325–347. https://doi.org/10.7202/1018284ar Mévellec, A., & Tremblay, M. (2016). Genre et professionnalisation de la politique municipale. Presses de l’Université du Québec. Nakhaie, M. R. (2006). Electoral participation in municipal, provincial, and federal elections in Canada. Canadian Journal of Political Science, 39(2), 363–390. https://doi.org/10.1017/S000842390606015X Oliver, E. J., & Ha, S. (2007). Vote choice in suburban elections. American Political Science Review, 102(3), 393–408. https://doi.org/10.1017 /S0003055407070323 Oliver, E. J., Ha, S., & Callen, Z. (2012). Local elections and the politics of small-scale democracy. Princeton University Press. Otjes, S. (2018). Pushed by national politics or pulled by localism? Voting for independent local parties in the Netherlands. Local Government Studies, 44(3), 305–328. http://dx.doi.org/10.1080/03003930.2018.1427072 Peterson, P. E. (1981). City limits. University of Chicago Press. Sances, M. W. (2018). Ideology and vote choice in US mayoral elections: Evidence from Facebook surveys. Political Behaviour, 40(3), 737–762. https:// link.springer.com/article/10.1007/s11109-017-9420-x Smets, K., and Van Ham, C. (2013). The embarrassment of riches? A metaanalysis of individual-level research on voter turnout. Electoral Studies, 32(2), 344–359. https://doi.org/10.1016/j.electstud.2012.12.006 Stephenson, L., McGregor, M. R., and Moore, A. (2018). Sins of the brother: Partisanship and accountability in Toronto. In S. Breux & J. Couture, (Eds.), Accountability and responsiveness at the municipal level: Views from Canada, McGill–Queen’s University Press, 23–48. Villeneuve, P., Jodoin, Y., & Thériault, M. (2007). L’énigme de Québec … ou de ses banlieues: une analyse de géographie électorale. Cahiers de géographie du Québec, 51(144), 375–397. https://doi.org/10.7202/017625ar Walks, R. A. (2004). Suburbanization, the vote, and changes in federal and provincial political representation and influence between inner cities and suburbs in large Canadian urban regions, 1945–1999. Urban Affairs Review, 39(4), 411–440. https://doi.org/10.1177%2F1078087403260787

6  Understanding Municipal Partisanship éric bélanger and laura b . stephenson

In studies of political behaviour, partisanship holds a special place.1 Since the publication of The American Voter in 1960, researchers have built a large literature examining the origins, parameters, and consequences of political party affiliation. It was Campbell and his colleagues (1960) who first defined partisanship as a long-standing psychological attachment. In this view, party identification (PID) is something that develops early through socialization and subsequently shapes political attitudes and behaviours. This definition has been countered by those who argue that partisanship is more of a running tally of evaluations (Fiorina, 1981), but the effect on electoral behaviour of being a partisan is generally agreed to by all. Partisanship colours how people understand information, evaluate candidates, and make decisions. As Norpoth (1978: 38) puts it, “an enduring sense of partisanship supplies an individual with an invaluable cognitive-evaluative guide to the world of politics.” In studies of voting behaviour, partisanship is assumed to be a relevant factor. It is measured differently in studies around the world (see, for example, the discussion in Blais et al., 2001), but it is consistently identified by its role in decision-making. To the best of our knowledge, no research has examined whether the nature of PID varies with the level of government. To be fair, the very idea is nonsensical in most countries, where the same parties compete at every level of government. In Canada, by contrast, the situation varies substantially across all three levels of government. In many provinces (British Columbia, Alberta, and 1 The authors thank the faculty members of the Escuela de Política y Gobierno at the Universidad Nacional de San Martín in Buenos Aires – especially María Laura Tagina, Carlos Varetto, Rocío Annunziata, and Gabriela Delamata – for their useful comments on a previous version of this chapter that Bélanger completed while on a research stay there.

Understanding Municipal Partisanship  117

Quebec, for example), the provincial level is marked by some unique party system configurations, and in the few municipalities where parties exist, those parties are not replicated in any other forum. In this chapter, we examine partisanship at the municipal level in Quebec, which offers an interesting context for studying this question, given that the two cities we are examining in this volume are marked by electoral rivalry between political parties. Given the extensive research about party identification at the federal level of government, our purpose is very straightforward – to ascertain whether municipal partisanship should be understood in the same way. To that end, we consider four elements: intensity, ideological coherence, candidate evaluations, and vote choice. We evaluate whether there is party-specific variation by looking in depth at the parties in Montreal and Quebec City, where there are two distinct types – those that form around an individual (équipes) and those that form around an ideology. It is not obvious that both types of parties can generate the same type of long-standing psychological attachments as are formed around parties at other levels; nor is it clear that municipal politics can engender the same party loyalties as other levels of government. The Canadian Municipal Election Study (CMES) asked voters about their partisan attachments at all three levels of government. This allows us to delve into comparisons across levels, cities, and parties. Our results suggest that, to a good extent, the idea of partisanship at the municipal level seems to make sense. While we certainly find fewer partisans at the local level of politics in Quebec, those individuals who identify themselves with a municipal political party do display traits and behaviours that we ought to expect if partisanship is to be a meaningful concept in the context of municipal elections. Literature Review and Theoretical Expectations Parties have long structured political systems around the world. Parties act as “teams,” bringing together legislators who favour similar policy outcomes in a way that creates efficiencies (especially when it comes to elections and branding) and support (especially in the legislature). For voters, party labels can provide a useful shortcut for understanding previous and future activity. In Miller’s (1976) words, “party loyalty does much to shape cognitions and evaluations of political events” (29). Voters often find that they prefer the policy stances of a specific party (even if that preference was inherited from their parents), and it is not surprising that voters can develop an attachment to a particular party. Partisan identity may signal the preferences of a given voter; it may also

118  Éric Bélanger and Laura B. Stephenson

provide voters with a simple decision criterion for evaluating future candidates. Voters can ask themselves, “Are they ‘like me’ or not?” A similar information shortcut can exist with respect to policies: individuals can decide where they stand on issues based on the stances of the parties, party members, or other partisans. This can save significant time and effort as an election looms. (For a discussion of the benefits of PID for voters, see Norpoth, 1978). Dinas (2014) has further demonstrated that partisanship is reinforced by vote choice, in the sense that consonant choices strengthen one’s ties to a party. Around the world, scholars have examined whether partisanship is a useful concept when trying to understand political opinions and choices. The idea was developed in the United States and has since become a standard element in vote choice models (Bowler et al., 1994), but its universal applicability has been questioned. Thomassen and Rosema (2009) argue that the concept has little traction in the Netherlands, being even less stable than vote choice, and that its applicability in other European parliamentary systems is similarly limited. Norpoth (1978) notes that in Europe, the prevalence of social cleavages has led some to question partisanship’s value with regard to voting. In Canada, partisanship has been shown to be a strong and relatively stable predictor of vote choice (Gidengil et al., 2012); nonetheless, there have been significant debates about whether it is identical to the American variant (for a review, see Bélanger & Stephenson, 2014). Even how to measure the concept so as to address both its psychological and durable components is a matter of contention (see, for example, Blais et al., 2001), as language equivalence and the decision to include party names and explicit non-partisan response options in public opinion surveys have led many to question the appropriateness of cross-national comparisons. Some of the explanations for why partisanship may vary across parties are of great interest to us here. One point raised in the Canadian context is that federalism and variation in provincial party systems weakens partisanship because voters can have different identities at different levels of government (Clarke et al., 1979, 2019; Clarke & Stewart, 1987; Stewart & Clarke, 1998). Clarke and his colleagues (1984) also make the argument that brokerage-style parties, as are found in Canada, mitigate against the development of partisan identities because the parties have no clear stances: “When the parties do not align themselves along major divisions in society, voters need make only very limited commitments to ‘their’ party” (56). The distinctiveness of partisan options can thus make partisanship more useful. Similarly, others have argued that the format of American elections (i.e., the presence on the ballot of multiple offices) and the greater number of elections (including primaries) make PID a

Understanding Municipal Partisanship  119

more useful and reinforced identity separate from vote choice (­ Gidengil, 1992; Butler & Stokes, 1969). Bélanger and Stephenson (2010) also find variation by type of party (ideological vs brokerage), with brokerage parties having somewhat less intense and loyal partisans. A recent theoretical statement about variation in partisanship comes from Schmitt (2009). He conceptualizes party identification as a political concept rather than a social one, taking into account the style of competition (adversarial vs consensual), ideological competition, and political mobilization. He finds that the electoral-system-induced style of competition has the most impact on partisanship in a country. He also finds that the impact of partisanship on vote choice is strongly affected by the politicization of the electorate, in the sense that the strength of social cleavages reduces the effect of partisanship (echoing work mentioned earlier). These insights are relevant for evaluating municipal-level partisanship. First, the arguments made about voting behaviour reinforcing partisanship and the limited opportunities outside the United States to utilize party identification in elections are relevant on three counts. As noted earlier, municipal parties exist only in municipalities, and therefore, any party identification can only be relevant sporadically (typically every four years). Another consideration is that the municipal parties that compete in a given election year have rarely been around for a long time. In Montreal and Quebec City, all of the main competing political parties but two (Projet Montréal and Équipe Labeaume) had been created during the five years prior (see Table 6.1). Regarding parties that develop around a mayoral candidate, there is the added element that such parties will come and go with individuals, thus eliminating the usefulness of the identity across elections. Furthermore, the opportunities for vote choice to affect partisanship are limited. Second, the ideological content of the parties is relevant for whether municipal partisanship can assume the traditional partisanship functions, in that opinion-informing information shortcuts are more likely to exist when there is an ideological program that underpins partisanship. Given the variation in the types of municipal parties, it is not clear that some of the parties have such content. In the next section, we take up the question of whether partisanship, as we understand it at the federal level, exists for municipal parties in Quebec. Our analysis proceeds in several parts. First, we consider the nature of PID as expressed by voters. Here we look at frequency and strength and compare across levels as well as across parties. We might expect there to be fewer partisans at the local level than at the provincial or federal level for the reasons just explained. That said, if PID is similar in nature across levels, we should not see much cross-level difference in

120  Éric Bélanger and Laura B. Stephenson Table 6.1.  Creation date of the main municipal parties competing in Montreal and Quebec City in the 2017 elections Name of party

Date

Projet Montréal Équipe Labeaume Démocratie Québec Équipe Denis Coderre Coalition Montréal Québec 21 Option Capitale-Nationale

2004 2008 2012 2013 2013 2017 2017

partisanship strength (or intensity). Second, we look at the link between partisanship and ideology. We might expect variation in partisanship to arise from the ideological range of the parties (put differently, partisanship should distinguish citizens along ideological lines) and from the nature of the parties themselves (whether they are candidate-focused or ideological). Third, we assess the extent to which partisanship may influence how citizens evaluate candidates. As a “perceptual screen” and incentive for motivated reasoning, partisanship should lead voters to rate their party’s candidates more highly. Finally, we examine the relationship between PID and municipal vote choice. Identifying with a municipal political party should lead one to vote in favour of council candidates from that party and for borough mayors affiliated with that party as well. Data Analysis To study partisanship at the municipal level, the CMES surveys measured party identification in their pre-election waves using a fairly standard question formulation, worded as follows: “In municipal politics, do you usually think of yourself as a […].” A list of the competing municipal parties, together with the options “none of the above” and “don’t know” (both recoded as representing non-partisans), were presented to the respondent as possible answers. A follow-up question designed to sort out “strong” and “weak” identifiers was also asked (“How strongly do you identify with that party?”), with three response categories (very strongly, fairly strongly, not very strongly). These two questions – party identification and its strength – were repeated identically for provincial and federal politics in the CMES surveys, so we have at our disposal a measure of voters’ PID at each of the three levels of government, enabling a systematic cross-level comparison.

Understanding Municipal Partisanship  121 Table 6.2.  Distribution of partisanship across levels of government (%)  

Municipal

Provincial

Federal

Non-partisans Not very strong Fairly strong Very strong N

30.8 13.7 39.8 15.7 3,778

17.0 17.8 46.4 18.8 3,778

15.3 18.6 50.2 15.9 3,778

Distribution and Intensity of PID The first question to examine with the available CMES survey data is whether municipal partisanship is like partisanship at the provincial and federal levels. Table 6.2 provides the results obtained with each of these three party identification measures, both cities combined, allowing for a direct comparison of PID at all three levels of government. The table shows that there are fewer partisans at the municipal level than at the provincial and federal levels (69% identify with a party at the local level as opposed to 83% provincially and 85% federally). This is as expected, since municipal parties typically do not enjoy the longevity of provincial or federal parties and hence cannot hope to foster as much long-term attachment among voters as the latter. That said, the data indicate that despite the somewhat unique nature of local-level electoral competition, a sizable proportion of citizens do identify with a local party. Obviously, some of these municipal partisans are also partisans at other levels, but the levels are far from entirely overlapping. For instance, the survey indicates that 43% of provincial non-partisans hold municipal partisanship (46% of federal non-partisans do as well). When partisans are compared in terms of the strength of their partisanship (Table 6.3), we find no substantial difference across levels. What is more, the table shows few differences between the cities. The proportions of very strong, fairly strong, and not very strong party identifiers in Montreal and Quebec City are all within 3.1 percentage points of one another. In short, according to the data presented in these first two tables, there is little that distinguishes partisanship at the municipal level from partisanship at the provincial and federal levels, other than the fact that there are fewer local-level partisans. Of particular interest is that local-level partisans do not come exclusively from the pools of partisans found at other levels. It is also relevant to consider whether the type of political party affects municipal partisan identification. Is there a difference in the nature of party allegiances depending on the type of party? Table 6.4 presents the strength of partisanship as measured for each of the municipal parties

122  Éric Bélanger and Laura B. Stephenson Table 6.3.  Strength of partisanship across cities and levels of government (%)  

Quebec City

Montreal

Provincial

Federal

Not very strong Fairly strong Very strong

20.5 56.0 23.5

19.1 59.1 21.8

21.5 55.9 22.6

21.9 59.3 18.8

N

1,408

1,208

3,134

3,200

Table 6.4.  Strength of partisanship by political party (%)

   Not very strong Fairly strong Very strong N

Équipe Option Coalition Denis Projet Démocratie Équipe Capitale- Québec Montréal Coderre Montréal Québec Labeaume Nationale 21 21.7

21.4

16.8

34.0

18.1

27.8

17.0

55.9

57.8

60.6

49.5

58.8

43.1

55.1

22.5

20.8

22.6

16.5

23.2

29.1

27.9

70

521

618

215

785

43

365

separately. The nature of partisanship does seem to vary depending on party type, but only moderately so. Indeed, we observe slightly more intense party attachment for two of the parties that could be characterized as “ideological” (as confirmed by the data presented below in Figure 6.1 and in chapter 5), namely Projet Montréal and Québec 21. Only 17% of each of these two parties’ self-declared partisans indicate that they are not very strongly attached to their party, which constitutes the lowest proportion found for this answer category in the table. That said, there is no pattern that emerges clearly other than this one. For instance, we do not find a similarly low proportion of weak partisans for Coalition Montréal, which is the other party that could be characterized as ideological based on the data we presented earlier. And partisans of the two most leader-centred parties, Équipe Denis Coderre and Équipe Labeaume, do not stand out as being particularly weaker than other partisans. All in all, the proposition that partisanship is more intense in the case of ideological, as opposed to leader-centred, parties is only weakly supported by the data. Ideological Coherence of PID If party identification at the municipal level is meaningful, it should distinguish partisans along ideological lines as much as it does at other levels

Figure 6.1.  Mean ideological self-placement of partisans .8

0.63 0.66

Green

Bloc Québécois

Ideology

0.31 .3

Coalition Montréal

Équipe Denis Coderre

Project Montréal

Démocratie Québec

Équipe Labeaume

Option Capitale-Nationale

Québec 21

Liberal Party of Quebec

Parti Québécois

Coalition Avenir Québec

Québec Solidaire

Liberal Party of Canada

Conservative Party of Canada

NDP

0.42 0.37 0.37

Federal

0.41 0.45

Provincial Quebec City Montreal

0.54 0.56

0.43 0.4 .4

0.54 0.57 .6

0.7 0.69 .7

.5

.2

.1

0

124  Éric Bélanger and Laura B. Stephenson

of government. Figure 6.1 uses the survey respondents’ self-­placement scores on a left/right ideological scale (running between 0 and 1, where 0 is most left and 1 is most right) and compares the mean of this score among partisan groups. The distribution of these mean scores appears consistent with the municipal parties’ platforms in both cities. In Montreal, the most left partisans are those of Projet Montréal, which is to be expected given the policy orientations of this party (see chapters 1 and 5). While Coalition Montréal cannot really be characterized as a typical right-leaning party, it was founded in 2013 to support the mayoral bid of Marcel Côté, a well-known economist in Quebec; it may thus be unsurprising to see Coalition Montréal partisans locating themselves more to the right than those from the other two main Montreal parties. Partisans in Quebec City similarly differentiate themselves along ideological lines. In that municipality, the partisans who clearly stand out from the others in terms of their ideological orientation are those of Québec 21. That party indeed adopted a right-leaning policy profile during the campaign, arguing for tax cuts and car-friendly infrastructure developments (see also chapters 1 and 5). Figure 6.1 also allows a comparison with provincial- and federal-level partisanship. As can be seen, partisanship helps distinguish supporters of various political parties along the left/right spectrum at these levels, too. In provincial politics, the two most polarized groups are those identifying with Québec Solidaire (on the left) and the Coalition Avenir Québec (on the right), as should be expected (Pétry, 2013). In federal politics, a similar opposition can be observed between partisans of the Conservative Party (most to the right) and those of the New Democratic Party (most to the left), with partisans of the Green Party and the Bloc Québécois being close to NDP ones. This ideological distribution of partisan groups at the federal level is also as expected (Cochrane, 2010). With such results being consistent with what we know of partisans at the provincial and federal levels in Quebec and Canada, and with their meaningfulness seemingly reflected at the municipal level by being congruent with what we know of the municipal parties’ policy orientations, it seems safe to conclude that partisanship at the municipal level does act as a broad ideological anchor of policy preferences. The above findings are corroborated when one looks at other indicators of policy positions, such as economic conservatism, social conservatism, and issue importance. These results are too numerous to be shown in this chapter but can be easily summarized here. Projet Montréal partisans score least highly on economic and social conservatism, whereas those of Coalition Montréal score most highly (and just as with left/right ideology, partisans of Équipe Denis Coderre locate themselves

Understanding Municipal Partisanship  125

in between the other two partisan groups). In Quebec City, partisans of Québec 21 score most highly on both conservatism scales. In terms of issue importance, again the results are clear. In Montreal, the partisans who assign the most importance to the issue of public transit are those of Projet Montréal, which is consistent with the party’s campaign pledges on this issue (see chapter 1). In Quebec City, also consistent with campaign positions, we find that the partisans of Démocratie Québec are the ones most concerned about public transit while those of Québec 21 are the ones who give the issue the least importance. When we consider the issue of property taxes, partisans who attribute the most importance to the issue are those of the two most right-leaning parties, Coalition Montréal and Québec 21, while partisans who give it the least importance are those of Projet Montréal and Option Capitale-Nationale in Quebec City. As a final consideration, we examine whether partisanship is simply a proxy for ideology. If so, then it should eliminate the latter’s association with vote choice by entirely absorbing its effect, thus acting as an intervening variable. To test this, we first estimate the magnitude of the relationship between ideology and mayoral vote choice, in both Montreal and Quebec City, and then add partisanship to the vote equation. The results, presented in tables 6.5 and 6.6, are similar to those presented in chapter 5 in that they show that the addition of partisanship to the vote model does reduce the relationship between ideology and vote choice, suggesting that partisanship effectively moderates the impact of ideology on voting preferences. That said, ideology’s association with mayoral vote choice, though diminished, remains statistically significant in both cities, indicating that partisanship’s relationship to the vote is not entirely attributable to ideology. While the findings presented in this section do indicate that ideology, as a long-term attitude, shapes local partisanship to a good extent, it seems that municipal partisanship, as a concept, is defined by more than just ideology – a conclusion that further justifies its study from a variety of angles, as we do here. Candidates and PID We can now assess behavioural evidence of PID. If partisanship at the municipal level acts as PID should according to the literature – that is, as a perceptual screen that colours people’s evaluations of political figures and events – then we should find that municipal partisans are more likely to evaluate the mayoral candidate of their party favourably. To assess this question, we analyse candidate evaluation scores as measured in the pre-election wave of the survey. Tables 6.7 and 6.8 report the mean feeling thermometer score given by partisans of each of the main

126  Éric Bélanger and Laura B. Stephenson Table 6.5.  Impact of ideology and partisanship on vote choice in Montreal (logit) Vote for Coderre over Plante 2.29* (0.42)

Vote for Coderre over Plante

Ideology Party ID Équipe Denis Coderre Projet Montréal Intercept Observations

1.33** (0.54)

– – –1.66* (0.21) 676

2.07* (0.27) –2.54* (0.36) –1.08* (0.29) 676

Pseudo R2

0.043

0.349

Note: Robust standard errors in parentheses. * p