Full Participation: A Comparative Study of Compulsory Voting 9280811665, 9789280811667

Table of contents :
Full participation: A comparative study of compulsory voting
Half Title Page
Title Page
Copyright
Contents
List of tables
Preface
1. Introduction
2. A history of compulsory voting and an overview of contemporary experience
3. Normative arguments for and against compulsory voting
4. Compulsory voting and election campaigns
5. Compulsory voting and electoral turnout
6. Compulsory voting, electoral integrity and democratic legitimacy
7. Compulsory voting and political outcomes
8. Conclusion
Appendix: sources of data and variable construction
References
Index

Citation preview

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page i

Full participation

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page ii

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page iii

Full participation A comparative study of compulsory voting

S A R A H B I RC H

Manchester University Press

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page iv

Copyright © Sarah Birch 2008 The right of Sarah Birch to be identified as the author of this work has been asserted by her in accordance with the Copyright, Designs and Patents Act 1988. Published by Manchester University Press in the UK, Europe, Australia and New Zealand Oxford Road, Manchester M13 9NR, UK www.manchesteruniversitypress.co.uk British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library

ISBN

978 0 7190 7762 3 hardback

First published 2008 17 16 15 14 13 12 11 10 09 08

10 9 8 7 6 5 4 3 2 1

The publisher has no responsibility for the persistence or accuracy of URLs for any external or third-party internet websites referred to in this book, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Typeset in Photina MT by Servis Filmsetting Ltd, Stockport, Cheshire

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page v

Contents

List of tables Preface 1 2 3 4 5 6 7 8

Introduction A history of compulsory voting and an overview of contemporary experience Normative arguments for and against compulsory voting Compulsory voting and election campaigns Compulsory voting and electoral turnout Compulsory voting, electoral integrity and democratic legitimacy Compulsory voting and political outcomes Conclusion

Appendix: sources of data and variable construction References Index

vii ix 1 20 40 59 79 99 118 136 155 163 179

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page vi

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page vii

List of tables

1.1 2.1 4.1 4.2 4.3 4.4 4.5

A typology of compulsory voting Compulsory voting in the world today (2008) Multilevel logit models of political knowledge Multilevel logit models of efficacy Multilevel logit models of political conversation and persuasion Multilevel logit models of contacting politicians Multilevel logit models of participation in protests and demonstrations 4.6 Multilevel logit models of working with others 4.7 Multilevel logit models of engagement in any form of political participation 4.8 Multilevel logit models of party identification 4.9 Multilevel logit models of engagement in campaign activities 4.10 Multilevel logit models of contact by party or politician 5.1 Impact of the introduction of compulsory voting on turnout 5.2 Impact of the abolition of compulsory voting on turnout 5.3 Impact of the removal of sanctions on turnout 5.4 OLS regression models of the impact of compulsory voting on turnout (country averages) 6.1 Dimensions of electoral integrity 6.2 OLS regression models of the determinants of electoral integrity 6.3 Multilevel logit models of satisfaction with democracy 6.4 Multilevel logit models of quality of representation 7.1 OLS regression models of the impact of compulsory voting on female representation 7.2 OLS regression models of the impact of compulsory voting on support for small parties 7.3 OLS regression models of the impact of compulsory voting on support for far-right parties 7.4 OLS regression models of the impact of compulsory voting on support for left parties

4 36 67 68 69 70 71 72 73 74 76 77 82 87 88 94 102 108 113 114 121 124 126 128

M1546 - BIRCH PRE.qxp:ANDY Q7

viii 7.5 7.6 8.1 8.2

27/10/08

11:30

Page viii

List of tables OLS regression models of the impact of compulsory voting on inequality (Gini index) OLS regression models of the impact of compulsory voting on corruption A Balance sheet of empirical findings Comparative survey data on support for compulsory voting in voluntary voting states

131 132 140 144

M1546 - BIRCH PRE.qxp:ANDY Q7

27/10/08

11:30

Page ix

Preface

The impetus for writing this book can be traced to a council block in the innerLondon borough of Hackney. Hackney is one of the most deprived local authorities in the UK, and the block in question is not the most salubrious in the borough. At the time of the 2005 general election it was home to 37 registered electors, but local records showed that only four of those had indicated their preference at the polls. Being fairly familiar with this part of London, I knew that the residents of the block had more reasons than most to cast votes; they were more dependent on state services than the average resident, and more was therefore at stake for them in the elections. Yet the vast majority clearly saw no good reason to exercise their franchise. I was aware of the many complex reasons why this might be the case, but it struck me that the cultural and behavioural obstacles preventing these people from participating in the elections were far too great to be overcome by most of the commonly-touted solutions to low turnout. We often hear that politicians should engage more with ordinary people, that politics should become less adversarial, that grass-roots activism should be reinvigorated, that a wider range of people should be elected. Yet I could see that it would take far more than that substantially to increase the rates of electoral participation in this block of flats, and that the vicious cycle of low participation and underrepresentation preventing the interests of these residents from being served would continue unless there was some radical intervention to break it. I began to wonder then whether the enforcement of electoral obligation was perhaps the only way to overcome the collective action problem affecting these people. Coincidentally, compulsory voting was at the time receiving favourable public support from government ministers in both the UK and France, and it was being actively debated in a range of other countries as well. This seemed like an electoral institution worth exploring. Delving into the literature on compulsory voting, I was struck by two things: firstly how thin it was; approximately one-sixth of all the states in the modern world that hold elections hold them under a compulsory voting regime. This is a substantial number, in comparison to the number that employ certain other

M1546 - BIRCH PRE.qxp:ANDY Q7

x

27/10/08

11:30

Page x

Preface

electoral institutions – such as the single-transferable vote or the double ballot – which have received considerably more treatment in the academic literature. Secondly, I was struck that, beyond the widely acknowledged and widely proven fact that mandatory voting laws increase rates of electoral participation, the other claims commonly made about this institution had hardly been subject to any empirical analysis. The normative literature on electoral obligation abounds with assertions about the power of mandatory voting laws to alter behaviour and attitudes, but these claims are backed up by scant evidence. The idea behind the present study then took shape: to provide a serious consideration of compulsory voting, assessing the arguments for and against this institution on both normative and empirical grounds. In so doing, I have produced – to my knowledge – the first book-length study of compulsory voting to be published in English. Those who are interested in getting the gist of the arguments and findings of the book and considering their policy implications can go straight to the conclusion, which summarises the main empirical results of the study and considers the practical hurdles associated with the introduction of compulsory voting in a modern democracy. For those interested in the nitty-gritty of data, this is contained in chapters 4 to 7, which provide quantitative evidence to test the normative claims made about compulsory voting. The normative arguments themselves are laid out in chapter 3, while chapter 2 presents a brief history of the institution and an overview of its use in the contemporary world. I am indebted to a number of people who have helped me over the course of writing this book, including David Denver, Albert Weale, Ernesto Calvo, Joop van Holsteyn, as well as the participants in the international symposium on compulsory voting, Institut d’Etudes Politiques de Lille, Lille 20–21 October, 2005; the University of Essex Vice Chancellor’s Seminar, 20 November 2006; and the European Consortium for Political Research Joint Sessions Workshop on ‘Compulsory Voting: Principles and Practice’ in Helsinki, 7–12 May, 2007. I am also grateful to the Department of Government at the University of Essex for its support throughout this project. Finally, thanks to Keith for distracting me often enough to keep me sane. Hackney

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 1

1

Introduction

Why compulsory voting? Why write a book on compulsory voting? A total of 29 countries in the contemporary world legally oblige their citizens to participate in elections, including about a quarter of all democracies. Yet in many voluntary voting states, the common view of this institution is of an outdated and largely disused relic from the past that will eventually disappear altogether as voters flex their liberal muscles and struggle to free themselves from all forms of compulsion. In fact, this view is distinctly at odds with much contemporary political thought, which is increasingly coming to see rights and duties as going hand in hand. Moreover, in many states where participation in elections remains voluntary, falling turnout has led a growing number of voices to call for making it a legal requirement. In 2006 alone, three major reports were issued on the subject in the UK – by the Electoral Commission, the Hansard Society and the Institute for Public Policy Research (Ballinger, 2006; Electoral Commission, 2006; Keaney and Rogers, 2006). The situation is similar in countries such as France and Canada, where prominent members of the political élite have recently called for electoral participation to be made mandatory.1 The fact that compulsory voting has recently received so much attention from practicing politicians suggests that the time is ripe for a scholarly review of the institution. Oddly, there has not been a single English-language monograph on compulsory voting in over 50 years.2 This is not to say that the topic is not studied; it has been the subject of a range of scholarly journal articles (as detailed in subsequent chapters), and it is touched on in literatures as diverse as those on wealth inequality and support for the far right.Yet compulsory voting tends to be studied mainly in the context of analyses that have other principal objects of investigation. This volume aims to fill this gap in the scholarly literature by providing a detailed overview of the history, practice, causes and effects of the legal obligation to vote, as well as an analysis of the normative arguments surrounding it. Recent debates about the possibility of introducing mandatory voting in those states where going to the polls remains voluntary call for a detailed discussion

M1546 - BIRCH TXT.qxp:ANDY Q7

2

27/10/08

11:31

Page 2

Full participation

of the normative advantages and disadvantages of this institution. If compulsory voting is ever to be introduced in any of these polities, the normative debate will have to be won; it is thus useful to have a clear understanding of the various arguments for and against compulsory voting. One of the main functions of this study is thus evaluative. A second main function of the book is explanatory. Compulsory voting has been introduced in a variety of contexts to address a range of problems, from low turnout in Belgium in 1893 to electoral corruption in Thailand over a century later in 1997. Yet the focus of most academic literature on the subject has been on the question of turnout alone. This volume seeks to broaden the study of compulsory voting by elaborating the effects it is frequently held to have, and by systematically examining each of these effects against comparative evidence from around the world. Compulsory electoral participation significantly alters the incentive structures faced by all the actors in the electoral arena, from parties and candidates, to voters, to electoral administrators. Its proponents and detractors alike have pointed to a range of likely impacts on the way elections are carried out, on the choices that are made at election time by key actors, and the effects of electoral results on wider political outcomes. At the same time, there has been scant effort systematically to examine these conjectures in the light of empirical evidence. The investigation to be undertaken in this volume will thus seek to elucidate the impact of the institution on phenomena such as political engagement, party strategies, electoral integrity, electoral outcomes, and policy outcomes. Before moving on to these tasks, however, it is necessary to specify what exactly is meant by the term ‘compulsory voting’. That will be the role of this opening chapter, which will seek to conceptualise and construct a typology of electoral obligation, before examining variations in the way the institution of compulsory voting has been implemented in different states. The chapter will conclude with an overview of the structure of the book. What is in a term? Compulsory voting can be defined very simply as the legal obligation to attend the polls at election time3 and perform whatever duties are required there of electors. As is often recognised, the inherent constraints of the secret ballot mean that in most modern democracies (and even in many less-thandemocratic settings) compulsory voting is, strictly-speaking, impossible. The state cannot typically monitor the behaviour of the elector in the privacy of the polling booth and can therefore do nothing to prevent him or her from casting an invalid or blank ballot; in very few states is any legal effort made to do so.4 The Dutch language recognises this distinction by employing a term – opkomstplicht – which can be translated as compulsory (or obligatory) attendance at

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 3

Introduction

3

the polls,5 as does a recent Institute for Public Policy Research Report, which refers to ‘compulsory turnout’ (Keaney and Rogers, 2006). Most European languages fail to make this distinction, however, and use terms that translate roughly as ‘obligatory voting’. The French speak of le vote obligatoire, the Italians of il voto obbligatorio, the Spanish of el voto obligatorio and the Portuguese of o voto obrigatório. In German the terms employed are (gesetzliche) Wahlpflicht and Stimmpflicht, while most Slavic languages use variations on the Polish term głosowanie obowia˛zkowe.6 The terms ‘obligatory voting’ and ‘mandatory voting’ do make their appearances in the English-language literature, yet the most commonly used term to designate this practice is ‘compulsory voting’. This is somewhat unfortunate, given the pejorative connotations of the term ‘compulsion’ in English; certainly ‘obligation’ has a rather different sound. Use of the term ‘compulsion’ thus casts the institution in a negative light in many English-languages debates on the subject (despite the fact that the Australians have been happily using this term to describe their electoral system for over 80 years). This usage has the further consequence of precluding an automatic semantic link between the institution and the broader notion of political obligation. A more appropriate term might be ‘the legal obligation to participate in elections’, but this being cumbersome, the present study will employ the terms ‘compulsory voting’, ‘mandatory voting’, ‘compulsory electoral participation’ and ‘mandatory electoral participation’, which will be used interchangeably. Conceptualising compulsory voting As has long been recognised by electoral behaviouralists, there are a wide variety of factors that bring people to the polls. We can conceptualise the incentive to vote as falling into two broad categories; pull and push and factors. ‘Pull’ factors include the standard range of motivations for voting, including the desire to influence electoral outcomes, expressive aims, identification with political contestants, and perceptions of civic duty (e.g. Campbell et al., 1960; Riker and Ordeshook, 1968; Verba et al., 1978; Powell, 1980; 1982; 1986; Crewe, 1981; Rosenstone and Hansen, 1993; Dalton, 1996; Franklin, 1996; 2002; 2004; Gray and Caul, 2000; Blais, 2000; Norris 2002; 2004). The legal obligation to vote is a principal ‘push’ factor; voters are urged to the polls by the law, with the threat of sanctions if they do not comply. Yet there are also other types of pressure that can be exerted to encourage people to vote, including social and political influence which, operating outside the ambit of formal political institutions, can nevertheless be remarkably effective. Indeed it is through this type of pressure, rather than legal compulsion, that the highest known turnout rates have been achieved in the world – the USSR’s regularly reported 99.99% levels of electoral participation (Bruner, 1990).7 Coercive mobilisation may also take

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

4

11:31

Page 4

Full participation Table 1.1 A Typology of compulsory voting Sanctions

Form of obligation

Sanctioned

Unsanctioned

Formal

Sanctioned electoral compulsion (e.g. Australia)

Unsanctioned electoral compulsion (e.g. Venezuela)

Informal

Sanctions, benefits in the absence of formal compulsion (USSR)

No compulsion, little pressure to vote (USA)

the form of undue influence by political parties (see Cox and Kousser, 1981; Hasen, 2000; Lehoucq 2003). Finally, ‘ordinary’ social pressure can prove a powerful force in encouraging people to attend the polls (Campbell et al., 1960; Rosenstone and Hansen, 1993; Blais, 2000; Franklin, 2004). In considering compulsory electoral participation, we are mainly interested in cases where electors have a legal obligation to attend the polls, but legal and informal socio-political forces interact in complex ways. The legal obligation to participate in elections can be congruent with social and political norms, or it can be at odds with one or both of these. There may also be considerable variations within a state – along geographical, sub-cultural, or other lines – in patterns of congruence. In schematic terms, we can understand there to be two types of obligation to vote: informal (social, political) and formal (legal). It should be noted that the enforcement of formal compulsory electoral participation requirements is often linked to a political and/or cultural environment that helps reinforce voting (i.e. congruence between legal and socio-political forces). The interactions between these two dimensions of obligation yield a fourfold typology (see table 1.1). In the upper left-hand quadrant, we find the case of a formal obligation to vote combined with effective sanctions. The archetypal case of this is Australia where voting is compulsory, sanctions – though small – are effectively imposed, and there is considerable popular support for the institution. The upper right-hand quadrant represents a situation found in many Latin American countries, where compulsory electoral participation is a legal norm, but sanctions are either non-existent or are not applied. In the lower lefthand quadrant, we find cases where the formal obligation to vote is absent, but informal social and/or political pressures are so strong that a large proportion of the citizenry nevertheless exercise their franchise. The USSR was a good example of this type of system; on election day Communist Party activists would repeatedly knock on the doors of non-voters and urge them to the polls. Though there is little evidence that non-voting had nefarious consequences for

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 5

Introduction

5

Soviet citizens, there was a widespread perception that it might be harmful to one’s career or one’s chances of obtaining scarce goods (Mote, 1965: 76–83). There is a similar situation in contemporary North Korea; Mark Suh reports that ‘although voting is not compulsory by law, the political prescriptions laid down in the party catechism prescribe it as the correct behaviour of every citizen. A simple negligence, let alone denial [sic] to take part in the polls, would be followed by harsh discrimination in the living and working sphere of the person concerned’ (Suh, 2001: 399–400). Finally, in the lower right-hand quadrant, we find the situation in the vast majority of established democracies, but perhaps best represented by the United States, where voting is voluntary and there is little social pressure to vote. It should be noted that the categories in this typology are not mutually exclusive, in as much as the formal obligation to vote is often reinforced by informal sanctions. Legal obligation and sanction may therefore be viewed as two overlapping but not coextensive spheres. On the one hand, there may be a legal obligation to vote in the absence of any formal sanction for non-voting. Such an obligation does potentially have importance, as it often represents a touchstone in political debates, and it reflects an act of collective self-binding – it can therefore be seen as an indication of, and mechanism for perpetuating, a certain cultural attitude toward voting. On the other hand, a state can impose formal sanctions for non-participation in the absence of any formal requirement to vote. This is the case, for example, in Iran where ‘though voting is not compulsory, citizens may have to show the stamp impressed on voters’ identity cards in polling stations when applying for passports’ (Kauz et al., 2001: 64). Until recently, Italy provided a democratic example of this practice. Electoral participation is not obligatory, yet voting is described as a duty in the Italian constitution (Art. 48). This prescription was reflected in Italian electoral law between 1946 and 1992 (Caramani, 2000: 64f), when light sanctions were applied to non-voters (lists of non-voters were posted at polling stations).8 Another example is the US state of Illinois, which for a time put non-voters at the top of the list of those chosen for jury service, though voting was not technically compulsory (Abraham, 1955: 15–16). Alternatively, the state can provide positive incentives for participation, either in the form of ‘vote facilitation’ mechanisms such as automatic registration, the availability of proxy voting, holding voting on a rest day, holding elections over more than one day, allowing absent and/or early voting, covering the travel expenses of voters who are temporarily away from their places of residence (see Franklin, 1996; 2002; Norris, 2004: 171–74), or by offering selective incentives to voters (see below). Where voting is legally mandatory, a distinction is sometimes made between states that enforce the legal obligation to participate in elections strictly or weakly (e.g. Gratschew, 2004). The problem with this approach is that there

M1546 - BIRCH TXT.qxp:ANDY Q7

6

27/10/08

11:31

Page 6

Full participation

are virtually no instances in the contemporary world of truly ‘strict’ enforcement. Commonly cited examples of ‘strictly’ enforced compulsory electoral participation are Belgium and Australia, but closer examination of these two cases reveals that in fact enforcement mechanisms take a light touch. Stengers (1990: 105) writes that once compulsory electoral participation was introduced in Belgium it became a way of life, and compliance was due more to social norms than to actual sanctions, which were in any case only sparingly applied; he notes that ‘it is more a matter of habit than of obligation’9 (1990: 105). Belgians simply got used to voting and took it for granted that if an election were held, they would be expected to turn up at the polls. This view is confirmed by Pilet, who reports that in Belgium in 1985, only 62 of 450,000 non-voters were punished (Pilet, 2005: 20). Similarly, Lieven de Winter and colleagues point out that in 1999 the chances of being subject to a fine were only 1 in 10,000, and they note that ‘Given its scarcely civic behaviour in many domains, and given the tradition of impunity before the law, the massive electoral participation of Belgians is surprising to say the least’,10 (de Winter et al., 2003: 58). The fact that turnout in Belgian elections remains well over 90 per cent must therefore be accounted for by other factors and should not be taken as an indication that the law is strictly enforced. Likewise, in Australia those who study the supposedly efficient mechanisms for the enforcement of ‘compulsory voting’ often point out that relatively few people are penalised for non-participation, because a wide variety of potential excuses for non-voting are admitted, and no documentary evidence is required to support a justification. In the 2004 elections 685,937 citizens were recorded as not having voted; of these, 458,952 were issued with notices preliminary to penalties. A total of 234,552 of those issued with notices sought to avoid paying the $20 (£8) fine by supplying ‘valid and sufficient’ reasons for not voting, and only 52,796 paid the fine when it was first imposed; a further 847 paid the fine plus a penalty for late payment. Seven citizens appeared in court to contest their fines (Bennett, 2005: Appendix). Thus while enforcement would appear to be very ‘strict’ in Australia, the penalty for non-participation is modest, and the vast majority of non-voters avoid paying even this small amount by offering undocumented excuses for their behaviour.11 Descriptions of enforcement as ‘strict’ or ‘weak’ risk conflating administrative effort with success (compliance rates). In states with weak state capacity and low levels of rule of law (as is the case in many Latin American countries), efforts at enforcement are likely to yield a lower payoff than is the case in established democracies such as Australia and Belgium. The key factor is whether a state has any functioning administrative apparatus in place to enforce the legal requirement to vote. If it does have such a mechanism, we can speak of mandatory electoral participation with sanctions, whereas if it does not, we can speak of non-sanctioned mandatory electoral participation. Such an approach to

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 7

Introduction

7

categorisation makes it possible to measure the impact of various other institutions and social factors on the effectiveness of enforcement (i.e. turnout), which is not possible if rates of turnout are themselves taken as an indicator of how much effort the state puts into enforcing the legal obligation to vote. This volume will define compulsory electoral participation in terms of a legal obligation to participate in elections, with or without sanctions. In other words, we will mainly be concerned with the upper row in table 1.1 – states that have formal institutional mechanisms that legally oblige citizens to attend the polls (with or without the application of sanctions). Informal pressures are outside the scope of this study, for they fall either in the category of social structural determinants of electoral participation, or, if they take the form of coercion, in the category of electoral malpractice. States that penalise non-participation or reward participation through formal means will not be considered either, as these represent a somewhat heterogeneous category that cannot truly be said to institute electoral participation as a legal requirement. Modes of enforcement and sanctions for non-participation Compulsory voting is necessarily embedded in a complex set of regulations governing the conduct of elections, and these regulations shape each other. States where there is a legal obligation to attend the polls have reason to make voting as easy as possible for the citizenry, as this will lessen the costs of enforcement, and it will enhance the popular acceptability and legitimacy of the institution. It is therefore not surprising that mandatory voting is typically linked to an array of institutional mechanisms that facilitate electoral participation (though, as noted above, such mechanisms are not exclusive to states where electoral participation is required by law). But even when voting is made as easy as possible, there will still be those who refuse to comply with the legal obligation to participate, whether out of objection to the institution itself or for some other reason. In somewhat more than half the states where voting is compulsory, sanctions are imposed (the exceptions include Greece (since 2001), Argentina, Brazil, Costa Rica, the Dominican Republic, El Salvador, Guatemala, Honduras (since 2004), Mexico, Panama, Paraguay, and Venezuela). For the purposes of this analysis it is useful to group sanctions into broad categories in order of increasing severity (bearing in mind that many states impose combinations of sanctions or different types of sanction in different circumstances). Demand for an explanation In practice, the demand for an explanation forms part of contemporary enforcement procedures in virtually all systems, in as much as this is a necessary adjunct to the imposition of a penalty. How this is handled varies

M1546 - BIRCH TXT.qxp:ANDY Q7

8

27/10/08

11:31

Page 8

Full participation

from state to state; in Australia, non-voters are first asked to justify their nonparticipation before a decision is made whether or not to impose a fine, whereas in other cases (Belgium, Luxembourg, and most Latin American states), it is up to electors to take the initiative to explain their absence from the polls to the relevant authorities in order to avoid measures being taken to penalise them. Reprimand Slightly more severe than a simple demand for justification is the issuing of a formal reprimand to non-participants. When electoral participation was first made mandatory in Belgium, a formal reprimand was a widely-used alternative to a fine for first-time non-voters.12 Name-and-shame systems One step up from a formal reprimand issued personally to the voter is a public notification that an individual has failed to carry out his or her civic duty. Such ‘name-and-shame’ systems are commonly referred to by the French term affichage, and they typically involve the posting of a list of non-voters in a public place such as the polling station or a municipal building.13 Affichage was common in the nineteenth century, and it survived in Italy until 1993.14 Fines Fines are the most common form of sanction for non-participation in elections; these range from three Swiss francs (£1.30) in the canton of Schaffhausen to €100 or more in Luxembourg. Some polities have blanket fines for all instances of violation; in other cases fines are graduated, increasing from the first to the second violation, and so on. In Belgium, for example, a fine of €25 to €50 is imposed for the first instance of non-voting, and €50 to €125 for the second instance; in Luxembourg the fine for repeated violations can rise to as much as €1,000. Until relatively recently, some Swiss cantons went to the extreme of sending district officers (Waechter) house to house to collect fines from dilatory electors (Funk, 2004: 24), while in inter-war Hungary the fine took the form of an additional tax (Abraham, 1952: 89). Use-it-or-lose it systems: removal of rights of civic participation If voting is considered to be a civic duty, it would seem logical that non-voters should have their civic rights restricted. The most obvious right to restrict is the right to vote itself. In Belgium, if a voter has not voted in four elections over the course of 15 years, he or she is removed from the electoral rolls for ten years.15 Singapore operates an unusual system that combines the removal of voting rights with a fine. Non-voters have their names automatically removed from the electoral register and must pay a fee to have them reinstated unless they can produce a ‘valid and sufficient’ reason for not having participated. Not long

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 9

Introduction

9

after the 2006 elections, the Elections Department of Singapore posted the following a message on their website: ‘Names of persons who failed to vote in the General Election held on 6 May 2006 have been expunged from the registers of electors. The lists of these persons are now available for inspection. Persons whose names are expunged from the registers may apply for their names to be restored using Elections Department’s E-services. Click for more information.’16 The other relevant civic right is of course participation in elections as a candidate; prohibition on this activity is a form of sanction in some states with mandatory electoral participation, including Argentina, Bolivia, Brazil, Singapore and Thailand. Prohibition of public employment for non-voters Another civic right that is sometimes restricted for failure to attend the polls is the right to hold a post in the state administration. This is practiced in Belgium (those who have been removed from the electoral roll for failure to participate in successive elections cannot be hired to or promoted in public office, nor can they be awarded state honours).17 In Argentina, non-voters are in theory ineligible for public office for three years following the election. In Bolivia, citizens who have not either participated in the election or paid a fine for non-voting cannot obtain a ‘suffrage certificate’ and without such a certificate they are not able to be employed in the public sector, to be paid in public sector jobs, to carry out banking transactions, or to obtain a passport for 90 days following the election. Similar regulations are stipulated in Brazil, with the additional provision that non-voters can receive neither state education nor exemptions from military service or the payment of income tax. Though application of this sanction is no longer widespread in Brazil, a study of poor neighbourhoods in São Paulo in 1978 indicated the importance of this provision at that time: ‘The majority of those interviewed “saw” the título eleitoral as a necessary document, given that it is necessary in order to obtain employment’ (cit. Power and Roberts, 1995: 800f). Loss of services As noted above, Bolivians without a ‘suffrage certificate’ are forbidden from carrying out banking transactions or obtaining passports for 90 days following the election, and in Brazil such electors are in addition deprived by law of the right to state education. Until recently, non-voters in Greece were in theory unable to obtain driver’s licenses or passports (Gratschew, 2002: 108).18 Imprisonment Imprisonment for non-voting is rare in practice, though it has been formally incorporated into the sanctions regimes of a number of polities that have made electoral participation mandatory. In Australia, for example, non-voters who

M1546 - BIRCH TXT.qxp:ANDY Q7

10

27/10/08

11:31

Page 10

Full participation

refuse to pay the fine face prison sentences; following the 1993 election, 43 non-voters received prison sentences of one to two days (Bennett, 2005).19 Exemptions No polity has full compulsory electoral participation, strictly speaking, because all make exceptions for people in certain circumstances. Exemptions fall into two categories: exclusions from the right to participate and exclusions from the legal obligation to participate. Exclusions from the right to participate in elections All states restrict the right to vote to certain categories of people. In older times, restrictions commonly included property ownership, male gender, and literacy. Though such exclusions have largely fallen into disuse, there is nevertheless variation in the extent to which the suffrage is in practice ‘universal’ in the modern world; exclusions are generally based on two principles: political community membership and political competence. The concept of ‘universal suffrage’ implies that the franchise is granted to all members of the political community, with the proviso that the practice must be based on some notion of minimum competence, given that ‘the set of the ethically desirable is bounded by the set of the feasible’ (Weale, 1999: 9). But while Goodwin-Gill notes that franchise requirements have become increasingly relaxed among democracies (1994: 43), different polities still define political community membership and political competence variously, resulting in differing franchise qualifications. Age Children are in all modern states excluded from the franchise on grounds of lack of sufficient competence to participate in political decision-making. The age of electoral majority varies from 16 to 21 in the contemporary world (Massicotte et al., 2004: 17). Mental incompetence Competence-based restrictions also commonly include mental disability or insanity, variously defined. Massicotte et al. note that only four states – Canada, Ireland, Italy and Sweden – do not disenfanchise citizens on grounds of mental instability or deficiency, and they conclude that ‘only where democracy is strongly established does it seem to be possible to envisage ignoring serious mental deficiency as a cause for denying people the vote’ (2004: 27). Citizenship The most restrictive of the political-community based franchise exclusions is that of citizenship, which in certain states serves to deprive sizeable sectors of

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 11

Introduction

11

the population of the right to vote. In some cases, however, non-citizens are allowed to vote, if only in certain types of elections (for example, local elections, elections to the European Union parliament in EU member states). Imprisonment Many states impose additional franchise restrictions on the basis of criminal conviction; among those where electoral participation is obligatory are Australia, Brazil, Belgium and Venezuela (Goodwin-Gill, 1994: 42–46; Katz, 1997: 218–29; Massicotte et al., 2004: 18–26). Military service Some states also exclude from the political community – and therefore the franchise – those in active military service; this practice is particularly common in Latin America. (Goodwin-Gill, 1994: 44–45; Katz, 1997: 218–29; Massicotte et al., 2004: 18–26). Residence Another common category of persons excluded from the franchise are those not resident in the country at the time of an election. Though more and more states have made provision for voting by overseas residents in recent years, this is still a fairly common ground for disenfranchisement. In some cases only those living abroad for longer than a certain period lose their right to vote, while in other cases overseas residents can vote in some elections only; for example, overseas Bolivian and Brazilian citizens can vote in presidential but not other elections. Certain states also impose a length-of-residence requirement, with respect either to the country as a whole, as in Chile and Costa Rica, or to a particular constituency, as in Australia (Massicotte et al., 2004: 18–26). Finally, we must consider what is sometimes referred to as ‘administrative disenfranchisement’. Compulsory electoral participation in theory requires compulsory registration (though since 1988 Chile has operated an unusual system whereby electoral registration is voluntary but once registered, electors are legally required to vote; see Lapp, 2006 for details). In other cases, the two go hand-in-hand. There are three main types of registration – periodic list, continuous list, and civil registry systems;20 of these, civil registry systems (which involve automatic registration by the state on the basis of the records it holds) are the most effective means of maximising inclusion. It is therefore not surprising that this method is the most common among democracies (Massicotte et al., 2004: 67: Powell, 1986). By contrast, registration procedures that make demands on voters’ time tend to reduce the effectiveness of those procedures (Katz, 1997: 239). But whatever method is employed, there will always be some eligible voters who are left off the register and are therefore unable to vote.

M1546 - BIRCH TXT.qxp:ANDY Q7

12

27/10/08

11:31

Page 12

Full participation

Exemptions from the obligation to participate in elections The obligation to vote can only be applied to those who have a right to vote, but in states where electoral participation is in theory mandatory, there are a number of categories of persons who commonly have the right but not an obligation to take part in the electoral process. In some cases people in these categories are formally exempted from the duty of electoral participation, while in other cases sanctions are not applied for failure to participate. Gender When compulsory electoral participation was first introduced in many states, women had not yet been given the right to vote. But the subsequent enfranchisement of women did not in all cases entail a legal obligation to vote, especially in the Latin American context. In Ecuador electoral participation was made mandatory for registered men in 1929 but only in 1967 for women (Nohlen and Pachano, 2005: 374). Similarly, in Guatemala, electoral participation has been compulsory for men since the universal franchise was introduced in 1894, but women have only been obliged to attend the polls since 1981 (though they had gained the right to vote in 1954) (Somoza, 2005: 402). In Peru, electoral participation became compulsory in 1931 for all literate men over the age of 21, but only in 1955 for women (Tuesta Soldevilla, 2005: 450). However, the gender differences in Latin America have all been removed, and Egypt is the only state in the modern world where electoral participation remains mandatory for men only.21 Age Several Latin American and European countries, including Argentina, Bolivia, Brazil, Cyprus and Peru, make electoral participation voluntary for those over 70 years of age; in Ecuador and Schaffhausen (Switzerland) the relevant cutoff is 65, and in Luxembourg 75. In addition, sanctions in Brazil are not applied to non-voters aged 16 and 17. Infirmity Virtually all states with mandatory electoral participation exempt the ill or infirm from participating in elections or exempt them from sanctions for nonparticipation. Citizenship In Belgium, non-citizens have a choice as to whether they register, but once registered they are obliged to vote in those elections for which they are eligible (European Parliament elections for EU member state citizens, or local elections for all foreign nationals). Luxembourg operates a similar system for citizens of other EU member states.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 13

Introduction

13

Residence or location on election day Many states exempt from electoral participation those travelling out of the country (or away from their place of registration) on polling day (Australia, Bolivia, Brazil, Chile, Ecuador, Luxembourg, Singapore). In Argentina people who find themselves more than 500 kilometres from their place of registration on election day with a reasonable justification are exempt from voting. In Chile the same applies to those more than 200 kilometres from their place of registration, and in Cyprus to those residing 50 miles or more from their polling station. In Australia, those stationed in the Antarctic benefit from a similar exemption, as do itinerants. In Peru, only those out of the country for medical treatment or study are exempt from sanctions. Varying regulations apply to citizens living abroad. States that make provisions for foreign residents to vote do not always make it compulsory. In Belgium, citizens residing abroad have, since 1999, had the right to vote, but they are not obliged to do so. If they choose to register, they are then subject to the same requirements as other Belgians (i.e. they are obliged to vote), but they have a choice as to whether they register. Literacy Brazil and Ecuador make voting voluntary for illiterates. Public duty In Argentina election staff as well as judges and their deputies are exempt from participating in elections if they are carrying out official duties on election day. Similarly, Austro-Hungarian provinces that made electoral participation mandatory exempted those carrying out imperial professional duties (Jenks, 1950: 139). In Peru and Schaffhausen military conscripts are also exempt. Inability to reach the polls due to unforeseen circumstances Retroactive exemptions are commonly granted to those who were unable to reach the polling station on election day due to unforeseen circumstances such as lack of means of transport (Austro-Hungarian provinces), bereavement and ‘domestic calamities’ (Ecuador, Schaffhausen), or ‘fortuitous causes or circumstances beyond their control’ (Bolivia). Discretion of judge or court A number of states, including Australia, Belgium, Chile and Luxembourg grant exemption from sanctions retroactively if grounds are provided to the satisfaction of a judge or court. In such cases the relevant authority can exercise some discretion in determining cases, especially when they involve exceptional circumstances.

M1546 - BIRCH TXT.qxp:ANDY Q7

14

27/10/08

11:31

Page 14

Full participation

Religious belief or practice Eligible Australian citizens who object to electoral participation on the grounds of religious belief may apply to be ‘conscientious objectors’ and be exempted from voting on those grounds (Mackerras and McAllister, 1999: 224; Hill, 2002b: 443–8). Likewise, under the Spanish electoral law of 1907 clergymen were exempt from participation (Abraham, 1952: 86). Other options for institutional reform There are three principal institutional reform options that can be considered instead of compulsory voting: the constitutionalisation of electoral obligation, collective sanctions, and the provision of positive incentives to voters. These three institutions have distinctly different patterns of historical usage. Designation of voting as a duty in a country’s constitution has been employed both alongside compulsory electoral participation and as an alternative throughout the institution’s history. Collective sanctions constitute a historical device that has gradually been phased out in all but a handful of cases. The provision of positive incentives to participate in elections is, by contrast, a mechanism that is increasingly gaining in popularity. Each of these institutional mechanisms will be discussed in turn. Constitutionalisation The first alternative to formally requiring citizens to participate in elections is to embed in the constitution a provision that voting is a legal duty. This may seem like a small measure, but by formally recognising the obligation to participate, a state binds itself to a certain normative attitude toward elections. This formal binding can then provide a basis for civic education curricula in schools and voter education programmes by electoral management bodies. It is perhaps for this reason that the constitutions of a number of states describe voting as a civic duty without making it compulsory in law. These include the Central African Republic, Colombia, Cuba, East Timor, Haiti, Italy, Mozambique and Portugal. In Colombia, Article 258 of the constitution states that ‘Voting is a right and a civic duty’. Article 49 of the Portuguese constitution states that ‘The right to vote shall be exercised personally and shall constitute a civic duty’. Former Portuguese colonies Mozambique and East Timor have very similarly worded provisions.22 Article 48 of the Italian constitution describes voting as a civic duty, as does the law governing the conduct of the 2006 elections, while Liechtenstein defines electoral participation as a civic duty in the electoral law (Article 3) but not in the constitution. Collective sanctions Another institutional device bearing conceptual affinities with mandatory participation in elections is that of the turnout threshold. The requirement that

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 15

Introduction

15

electoral participation should reach a certain level for the election to be valid was not uncommon in the nineteenth century.23 At that time Bavaria and the Grand Duchy of Baden operated an unusual combination of collective and individual sanctions: if an election did not meet the turnout requirement, the costs of the re-held election would have to be borne directly by non-voters as a form of a special tax (Moreau, 1896: 27; cf Caramani, 2000: 64).24 More commonly, elections are simply declared invalid if they fail to meet the threshold set in law, and they are re-held at the expense of all. Though turnout requirements are common in contemporary referendums, they are now rare in elections outside Eastern Europe and the former Soviet Union. The modern practice of turnout thresholds in this region is a hang-over from the Soviet requirement that at least 50 per cent of eligible citizens were obliged to vote for the election to be valid (and, in addition, at least 50 per cent of the eligible electorate was required to vote for a candidate for that person to be declared elected (Carson, 1955: 76)). This practice was adopted also in Central Europe when most of the countries in the region introduced communist systems (or had such systems imposed on them). Interestingly, this provision has been retained in some form in the electoral legislation of a number of states in the region, despite the fact that most other aspects of Soviet-era electoral law have been abandoned. Until recently Russia imposed a turnout requirement of 50 per cent in presidential elections and 25 per cent for parliamentary polls. In Serbia and Montenegro, similar 50 per cent turnout requirements led some elections to be re-held in their entirety, while the 25 per cent requirement in Hungary has resulted in elections being re-held in individual constituencies. Elsewhere in the region, a 25 per cent turnout requirement applies in Lithuania, the threshold is 33 per cent for parliamentary elections in Uzbekistan, and the Soviet-era 50 per cent requirement has been retailed in Moldova, Belarus, Kazakhstan, Kyrgyzstan and Tajikistan.25 Given this ‘collectivist’ approach to electoral participation in the former communist region, it should come as no surprise that this is the only part of the world where compulsory electoral participation is not practiced at all.26 Incentives for voting The third main alternative to making electoral participation a legal requirement is to provide incentives that encourage citizens to go to the polls. Such incentives may take the form of sanctions or selective benefits (in as much as these can be distinguished). The application of sanctions without compulsion may seem odd, yet as noted above, both Iran and North Korea have such a system, and sanctions for non-voting were in place in Italy until the 1990s. The provision of selective benefits to voters is a rather newer idea that has recently begun to take hold in a variety of settings where increasing rates of electoral participation is deemed desirable but there is general reluctance to

M1546 - BIRCH TXT.qxp:ANDY Q7

16

27/10/08

11:31

Page 16

Full participation

make voting a legal obligation. This has often taken the form of the entry of voters into a lottery, as was done at the national level in Bulgaria in 2005 and at the local level in Norway (IDEA, 2006). This practice has also been employed in some US states (see Hasen, 1996; 2000). In Colombia – one of the only states in Latin America where voting is voluntary – voters have since 1997 been given preferential access to some educational and state employment opportunities, reduced military service, and reductions in fees for some state services (Lapp, 2004; Jaramillo and Franco-Cuervo, 2005: 302; www.votebien.com). Finally, in the UK the idea of exploring so-called ‘incentive voting’ has won support from a number of quarters, including Ken Ritchie of the Electoral Reform Society (Baston and Ritchie, 2004). There are arguments for and against the use of selective incentives as an alternative to legal compulsion, which will be discussed in the concluding chapter to this volume. Structure of the book Following the basic overview of the institution of mandatory electoral participation provided here, Chapter 2 details the history of the institution, tracing the idea back to medieval times and showing how it has developed in the modern state. The chapter goes on to examine common reasons why compulsory electoral participation has been introduced and it closes with an overview of the contemporary use of this institution for elections to lower or only houses of parliament. Chapter 3 outlines and assesses the principal normative arguments that have been made for and against compulsory electoral participation. These claims fall into three principal categories: arguments relating to rights and duties, legitimacy and collective rationality arguments, and evaluations of the practical consequences of making electoral participation mandatory. After reviewing these arguments, the chapter then summarises the empirical claims made by normative theorists in order to prepare the ground for subsequent chapters, whose main objective is to test these claims. Chapters 4 to 7 are structured around the key elements of the electoral cycle: campaigning, voting, conducting an election, and acting on electoral mandates. Chapter 4 is devoted to considering how compulsory electoral participation affects electoral campaigns and related attitudes. This chapter seeks to determine the impact of this institution on campaign-related behaviour and dispositions relevant to campaigns. Compulsory electoral participation can be expected to alter the incentive structure voters face when considering elections. It will thus invariably alter the way in which those seeking to win their votes appeal to the electorate, and, in turn, the way voters react to electoral campaigns. From the perspective of electoral contestants, the main difference between an election held under voluntary voting and one held under

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 17

Introduction

17

compulsory turnout is that in the former it is necessary to place emphasis on mobilisation to get one’s supporters to the polls, whereas in the latter this is taken care of (to a greater or a lesser extent) by the legal framework. The impact of this important difference is evaluated and its likely consequences are assessed against empirical evidence. In Chapter 5, attention turns to the topic that has received the greatest amount of attention from those who have studied the institution of compulsory voting: aggregate turnout. The question of concern in this chapter is how efficient compulsory voting really is in improving turnout levels. The overall aim of the chapter is to review previous analyses of the impact of mandatory voting laws on rates of participation, and to asses the conditions under which such laws are most likely to be associated with substantial change in rates of voting. Chapter 6 considers the influence of compulsory voting on electoral integrity and the overall legitimacy of the political system. As will have been demonstrated in Chapter 2, reducing electoral abuse and improving electoral integrity were among the most important reasons for the introduction of compulsory attendance at the polls in many states. It is therefore worth considering whether making electoral participation mandatory improves the conduct of elections and thereby improves confidence in the political process. This hypothesis is tested on the basis of case studies as well as aggregate data. A related question to be considered is whether mandatory voting enhances the legitimacy of the democratic process overall. Chapter 7 explores the impact of compulsory electoral participation on those political outcomes viewed as most significant, as well as those that can most intuitively be expected to result directly from the institution. Three such impacts are analysed: the impact of compulsory electoral participation on (1) rates of female participation (2) the distribution of partisan support (the balance between small and large parties, moderate and extremist parties, and left-right forces); (3) the fairness of policy outputs, understood in terms of redistributive policy and quality of governance. A final chapter concludes the volume, summarising the findings of the study and providing an overall evaluation of compulsory electoral participation on the basis of these results. It speculates as to the likelihood that electoral compulsion will be adopted in democracies where going to the polls is voluntary. The chapter also considers alternatives to making electoral participation a legal obligation, as well as possible modifications to the institution.

Notes 1 These cases are considered in greater detail in Chapter 8. 2 An extended review of the English-language literature has revealed only one other published volume on the subject, a short tract of 38 pages published in 1955 by

M1546 - BIRCH TXT.qxp:ANDY Q7

18

3

4

5 6

7

8 9 10

11

12 13 14 15

16 17

27/10/08

11:31

Page 18

Full participation Henry Abraham. This text is a summary of an unpublished doctoral dissertation (Abraham, 1952). ‘Attend the polls’ is used as a convenient short-hand to designate participation in person or in absentia, to the extent that the latter practice is allowed. It would therefore also include proxy, postal and other forms of remote voting. Tingsten makes the distinction between the obligation to vote (which involves simply attending the polling station) and the obligation to elect (which requires a positive choice to be made) (Tingsten, 1963 [1937]: 182). I am grateful to Joop van Holsteyn for providing this translation. The website of the Latvian parliament provides a useful list of translations of this term into the languages of all European Union member states at http://www. saeima.lv/pet/pls/DD?p_LN=mt&p_INTERFACE_LN=en&p_DESCRIPTEUR_ID=4 758. Official Soviet turnout figures are slightly misleading, in that voters could apply to have themselves removed from the register in their constituency on the grounds that they would be away on election day. Estimates of the number of people who avoided their electoral obligations in this way range from 2.5 to 10 per cent (Friedgut, 1979; Karklins, 1986; Roeder, 1989). The law governing the 2006 elections re-introduced voting as a civic duty, but did not prescribe sanctions for non-voting. ‘On peut parler d’automatisme plus que d’obligation’ [author’s translation from the French]. ‘Eu égard à son comportement peu civique dans nombreux domaines, et compte tenu de cette tradition de grande impunité du délit, la participation massive des Belges aux élections reste pour le moins surprenante’ [author’s translation from the French]. Another case of the supposed ‘strict’ enforcement of mandatory electoral participation is the Netherlands, where there was a legal obligation to vote between 1917 and 1970. Yet Irwin reports that in 1966 only 577 of 400,000 non-voters were brought to court (Irwin, 1974: 294). See Robson, 1923: 574 for relevant data. Affichage was proposed in France in 1889 and 1893, though it was never passed (Quiri, 1908: 78). Law Art 3, D.Lgs. 20 December 1993, n. 534 removed the requirement that names of not-voters should be posted at local municipal offices. The mandatory participation law employed in Hungary during the interwar period stipulated that a citizen should be removed from the electoral register after the first instance of non-voting (Abraham, 1952: 89; 1955: 19). In Cuba, non-voters were removed from the register under the 1919 electoral law, though electoral participation was not made formally compulsory until 1940 (Suter and Nohlen, 2005: 200). This message was posted on the website of the Singapore Elections Department at http://www.elections.gov.sg/; accessed on 23 November 2006. When compulsory electoral participation was initially introduced in Belgium, the restrictions on public service were more extensive, including a ban on bearing

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 19

Introduction

18

19

20 21 22

23

24

25 26

19

witness in court, serving on a jury, serving as legal counsel, serving on public commissions or being a teacher (Moreau, 1896: 28). Though not considered here under the rubric of mandatory electoral participation, Italy had until recently a similar system; before the removal of sanctions for nonvoting in the mid 1990s, the fact of not having voted was noted on official documents, and there are reports that this may have made it difficult to obtain services such as childcare (Jackman, 2001: 16315; Gratschew, 2002: 108). When electoral participation was made compulsory in the Philippines in 1973, the penalty for non-voting included a prison sentence of up to six-months; Hartmann et al. report that had this penalty been rigorously applied, over seven million citizens would have gone to jail following the 1981 elections. The measure was not systematically enforced, however, and electoral participation was made voluntary again in 1987 (Hartmann et al., 2001: 189). On varieties of registration procedures, see Massicotte et al., 2004: 66–76; ACE Project at http://www.aceproject.org/main/english/vr/vr.htm. Lebanon’s brief experiment with compulsory voting between 1952 and 1957 was likewise applied to men only (Scheffler, 2001: 174). In East Timor, section 47 of the constitution reads ‘The exercise of the right to vote is personal and constitutes a civic duty’, while Article 73(3) of the Mozambiquan constitution states that ‘The Right to vote shall be personal and shall constitute a civic duty’. Crampton (1983: 43) claimed that such a rule was applied in Bulgaria following 1879, and Moreau (1896: 22) cites turnout requirements in nineteenth century Russian local elections. Quiri (1908: 82) reports that bills to introduce turnout requirements were proposed (all unsuccessfully) in the French National Assembly in 1872, 1889, 1893, 1894 and 1900. For details of the regulations in the individual states discussed in this paragraph, see the Inter-Parliamentary Union Parline database at www.ipu.org. Indeed, the constitutions of some of the post-communist states actually emphasise the voluntary nature of voting. For example, article 65 of the Belarusian constitution states that ‘A voter shall decide personally whether to take part in elections and for whom to vote’.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 20

2

A history of compulsory voting and an overview of contemporary experience This chapter considers where, why and how compulsory voting was adopted for elections to lower (or only) houses of national parliaments. The first section examines the origins of compulsory voting as an idea; this is followed by an overview of its adoption and use since the nineteenth century. The third section investigates the reasons why voting was made mandatory, the arguments voiced in favour of and against this institution, and the identity of those who supported and opposed it. The chapter closes with overview of the use of compulsory voting in the world today. The origin of an idea and a practice The idea of obliging electors to attend the polls can be traced back to the Landsgemeinde, or citizen assemblies, in medieval Switzerland. Robson claims that ‘The idea appears first in the eighteenth century and seems to have evolved out of the remembrance of the ancient custom by which members of the Landesgemeinde of Schwyz were required to wear their swords at the Assembly Meetings; for apparently if a member did not put in an appearance, or turned up without his sword, he was subsequently refused admittance to the free dinner given on these occasions to the citizens of the canton’ (Robson, 1923: 570; cf Caramani, 2000: 63). Compulsory voting was first applied to modern elections in the seventeenth century American colonies. In Plymouth Colony a fine for non-voting was introduced in 1636, and over the course of the following decades several Massachusetts towns passed legislation compelling attendance at town meetings. The colony of Virginia made electoral participation mandatory throughout its history; the fine for non-participation introduced in 1649 was 100lb of tobacco (increased to 200lb in 1662). The colonies of Maryland and Delaware followed suit in 1715 and 1734 respectively, as did Georgia in its constitution of 1777 (Abraham, 1952: 107–18). After the American War of Independence electoral compulsion fell from favour in the US, and discussion of this measure was not revived until the early twentieth century. At this point there were

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 21

A history of compulsory voting

21

several efforts to introduce it at the local and state levels. Sanctions against nonvoters (in the form of taxes) were applied in Kansas City in 1889, before being ruled unconstitutional by the Missouri State Supreme Court in 1895 (Abraham, 1952: 113). This did not deter the states of North Dakota and Massachusetts which amended their constitutions in 1898 and 1918 respectively so as to allow for voting to be made compulsory, though this never in fact took place (Jackman, 2001).1 At the same time as mandatory electoral participation was being put to work on a small scale in what were then obscure and distant British colonies, it was also being discussed in the context of revolutionary France. The radical changes contemplated and realised during the Revolution sparked a debate about the suffrage as a right or a function that lasted until well into the twentieth century. Those who viewed voting as a right favoured making the franchise voluntary, whereas for those who saw it as a state function, it made sense for participation in elections to be obligatory.2 Moreau reports that compulsory electoral participation was demanded at the time of the French Revolution in the Cahiers of the commune of Aurons (Moreau, 1896: 3f), but the innovation does not appear to have been widely debated, and when in 1789 France became the first country in the world to hold elections under universal (male) suffrage, the franchise was voluntary. Voting in the indirect election for the French Senate (elected by local and regional deputies) has, however, been compulsory since 1875 (in some places since 1789). The French example of compulsory voting for the Senate but voluntary participation in popular elections provides an interesting case for comparison. The mandatory nature of voting in Senatorial elections is justified on the grounds that Senators are elected by public servants (mostly elected legislators), rather than by individuals. Electing Senators is therefore for these officials a public duty which cannot be shirked. This contrasts with the situation of ordinary citizens, who have not chosen to take on the duty of electing but have had it offered to them. A parallel point is made by Caramani, who points out that in early modern times, the representation of boroughs, counties and estates was generally compulsory in Europe, and that it was perceived as a burden. Only gradually when representation yielded greater advantages did it come to be perceived as a right worth struggling for (Caramani, 2000: 63).3 Regional trends in compulsory voting Following its pre-history in Europe and North America, compulsory electoral participation was introduced in the nineteenth and twentieth centuries in two main regional ‘waves’ in Europe and Latin America. The institution has also been employed more sporadically in Australasia, as well as in the Middle East

M1546 - BIRCH TXT.qxp:ANDY Q7

22

27/10/08

11:31

Page 22

Full participation

and Africa. Given that these waves were distinct in terms of influences and patterns, it makes sense to consider each of them separately. Europe As noted above, Switzerland was the first European state where compulsory electoral participation was introduced in modern times. True to the Swiss style of democracy, the decision as to whether voting should be voluntary or compulsory was devolved to individual cantons, and in some cases to individual communes. The practice was first introduced formally in the canton of St Gallen in 1835. Over the course of the next few decades, it was adopted in about half of the Swiss cantons. Solothurn was one of the first Swiss cantons to apply sanctions (first in the constitution of 1856, and then in the electoral law of 1860), but this was abandoned in 1887 due to enforcement difficulties, and turnout never reached more than 69 per cent. This experience is typical of Switzerland; many cantons never employed sanctions at all, and in those that did do so, turnout was markedly higher than in those where voting remained voluntary (Tingsten, 1963 [1937]; Funk, 2004). Compulsory voting gradually fell into disuse over the course of the late nineteenth and twentieth centuries, such that today Schaffhausen is the only Swiss canton where participation in elections remains mandatory. In 1878, Liechtenstein was the first state to make electoral participation compulsory at the national level. Mandatory voting was maintained for well over a century in the tiny Alpine state, until in 2004 Liechtenstein removed penalties for non-voting (while retaining in the electoral law the provision that voting was a civic duty). At the same time it introduced all-postal voting and initiated investigation of the possibility of extending the use of electronic voting.4 Thus the removal of penalties was accompanied by a replacement with positive incentives. The German principalities of Baden, Hesse, and Bavaria all made electoral participation obligatory for a period in the 1880s (Moreau, 1896). Caramani also reports that between 1924 and 1928 it was compulsory in the German Land of Lippe (Caramani, 2000: 64). Though the institution was debated at the national level in unified Germany (Tingsten, 1963 [1937]), it was never introduced. Belgium is perhaps the best known and best documented instance of compulsory electoral participation in Europe, not least because the practice has been maintained up until the present day. Voting was made mandatory in 1893, at the same time as universal manhood suffrage (see the next section for details). In the neighbouring Netherlands, a similar reform was brought about in 1917, also at the time of the introduction of universal manhood suffrage and proportional representation (Irwin, 1974; Irwin and van Holsteyn, 2005).5 The third of the low countries, Luxembourg, followed in 1919 by introducing universal

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 23

A history of compulsory voting

23

and compulsory manhood suffrage (Caramani, 2000: 63). Compulsory electoral participation was abolished in the Netherlands in 1970, despite its continuing popularity among the electorate (Irwin and van Holsteyn, 2005: 8–16). In the eastern part of the continent, Bulgaria took the lead by stipulating a requirement to participate in the indirect elections held under the law of 1882 (Moreau, 1896; Abraham, 1952: 89). Austro-Hungary followed in 1907, introducing compulsory voting at the same time as universal suffrage and the abolition of the weighted franchise (curia system). As in Switzerland, electoral participation in Austro-Hungary was never made mandatory throughout the entire territory; individual provinces were left to decide the matter on their own (Jenks, 1950: 134–9), and by 1911, eight had introduced the measure (Upper Austria, Lower Austria, Salzburg, Moravia, Vorarlberg, Silesia, Carniola, and Bukovina) (Robson, 1923: 571; Tingsten, 1963 [1937]: 199). Following the collapse of the Austro-Hungarian Empire after the First World War, the new Austrian republic also left it to individual provinces to decide whether or not they would make electoral participation obligatory. Vorarlberg, Styria and Tyrol took up the option, and it was practiced there for parliamentary elections until 1992, when compulsory attendance at the polls was abolished at federal level. Participation in presidential elections was mandatory in Austria until 1982; between 1982 and 2004 it was left to individual provinces to regulate, after which it was abolished completely (Hirczy, 1994; Bundesministerium für Inneres, 2006). Several of the other states that emerged from the Austro-Hungarian empire following the Treaty of Versailles also retained or adopted mandatory electoral participation, including Czechoslovakia (1920) and Hungary (1925) (Tingsten, 1963 [1937]; Abraham, 1952). Bulgaria re-introduced the practice in 1919 when universal suffrage was granted (Abraham, 1952: 89; Tingsten, 1963 [1937]), and neighbouring Romania followed suit by enshrining the legal obligation to participate in elections in its constitution of 1920 (Tingsten, 1963 [1937]; Abraham, 1952). The institution was abandoned in the Central European states after the Second World War, when informal political pressure to vote under the new communist regimes replaced legal obligation. In southern Europe, Spain introduced compulsory voting in 1907 at the same time as direct elections, but the measure lasted only till 1936, and enforcement was lax (Abraham, 1952: 87; Caramani, 2000: 63). Greece made electoral participation mandatory in 1926, when proportional representation was also introduced. Cyprus followed the Greek example upon gaining independence in 1960. Compulsory voting was adopted in Turkey at the time of the constitutional referendum of 1982 which preceded that country’s redemocratisation drive, and the institution has been retained since that time. This overview of the experience of mandatory electoral participation in Europe shows that the British Isles and Scandinavia are the only major

M1546 - BIRCH TXT.qxp:ANDY Q7

24

27/10/08

11:31

Page 24

Full participation

European regions where the legal obligation to vote never gained purchase. It was discussed in Sweden and Finland in the first years of the twentieth century, but not adopted (Tingsten, 1963 [1937]: 183; Alapuro, 2005). It was also debated extensively in the UK in the 1920s and 1930s, when half a dozen bills on the topic were introduced in Parliament. Yet none gained sufficient support to pass, and the topic faded from discussion in the 1950s when turnout increased (Sear and Strickland, 2003). Latin America The Latin American experience differs from that of Europe in that the legal obligation to participate in elections was introduced over a longer period (1893–1966), and it has not been formally abandoned nearly as frequently. Today Colombia, Nicaragua and Cuba are the only states in the region not officially to require their citizens to participate in elections.6 However, eight of the sixteen Latin American states that still practice mandatory attendance at the polls have removed sanctions for non-participation as an alternative to abolishing the institution altogether (see table 2.1). It is somewhat ironic that Nicaragua is one of the only Latin American states to have formally abolished mandatory electoral participation, as it was a pioneer in its introduction to the continent. In 1893, only a few months after the legal obligation to vote had been introduced in Belgium, Nicaragua embedded compulsory voting in its new Liberal constitution of 10 December, which also guaranteed universal male suffrage and ballot secrecy.7 The compulsory participation requirement was retained in the constitutions of 1939, 1948, 1950 and 1974, but removed after the Sandanista revolution of 1979. Honduras followed Nicaragua’s model, introducing in 1894 universal, secret and compulsory suffrage for men (Somoza, 2005: 402). Having been established in Central America, compulsory voting was introduced in Mexico as part of the reforms brought about at the time of the 1917 revolution (Nohlen, 2005: 12). The measure was then adopted in Panama in 1928; Costa Rica followed suit in 1936. Both these reforms were unusual in the Latin American context, in that they did not coincide with other major changes to the electoral system. Cuba introduced mandatory electoral participation in 1940, at the same time as it constitutionalised the right of women to vote and brought in proportional representation (though there had been sanctions for non-voting in Cuba since 1919 (Suter and Nohlen, 2005: 200)). As in Nicaragua, leftist revolution brought an end to formal compulsion to vote in Cuba (and in the Cuban case the revolution of 1959 also brought an end to direct elections until 1992). The legal obligation to participate in elections spread to South America when the institution was introduced in Argentina in 1912 as part of the Ley Sáenz Peña, which simultaneously adopted effective universal manhood

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 25

A history of compulsory voting

25

suffrage through the secret vote, and various other measures to eliminate electoral abuse (see the next section for details). Uruguay, Bolivia and Chile all made electoral participation mandatory over the course of the following decade, the first two in 1924 and Chile in 1925. In Chile 1925 this coincided with universal, direct, secret male suffrage and proportional representation (Nohlen, 2005b: 258–9). Ecuador made electoral participation compulsory in 1929 following the ‘Revolución Juliana’ in 1925; it also gave the vote to women at this point, though it did not oblige them to participate (Nohlen and Pachano, 2005: 374).8 Neighbouring Peru followed a similar path in 1931, introducing male suffrage for literate men under proportional representation, with secret and compulsory voting. The electoral reforms introduced in Brazil in 1932 following the Revoluçião de 30 were similar to the Peruvian changes made the previous year, save that Brazil also gave the vote to women at this time, and it established a system of electoral courts that resulted in a significant standardisation and amelioration of the electoral process (Lamounier and Amorim Neto, 2005: 164–5). Following the Second Word War, those Latin American states that had not yet made electoral participation a legal duty all did so, with the exception of Colombia. In Central America, compulsory voting had been entrenched in the constitution of the short-lived Federation of Central America that brought together El Salvador, Guatemala and Honduras between June 1921 and January 1922 (Escobar Armas, 1986: 139), but El Salvador properly adopted the institution (with fines for avoidance) only in 1950, together with universal suffrage and proportional representation (Samayoa, 1986). Venezuela made electoral participation compulsory in 1958 when that country returned to democracy following a decade of instability and authoritarianism. Sanctions related to denial of public services were imposed initially, but were subsequently abolished in 1993 (Molina and Thibaut, 2005: 548–9). During the 1960s, the introduction of mandatory electoral participation was associated with moves toward authoritarianism, rather than toward democracy. In Guatemala, voting was described in the constitution as a civic duty as early as 1945 (Escobar Armas, 186: 140),9 but it was made compulsory in 1965 following a US-backed right-wing coup. The Dominican Republic made voting mandatory in 1966 when the authoritarian Balaguer regime assumed power (Brea Franco, 2005: 240–1). Finally, Paraguay made participation in elections compulsory in 1967, also in the context of a move toward authoritarianism (León-Roesch and Ortiz Ortiz, 2005: 418). The norm of mandatory electoral participation is thus virtually ubiquitous in Latin America (though many of the states in the region make little effort in practice to apply sanctions). What is perhaps most interesting from the point of view of comparison with Europe is the fact that in Latin America mandatory electoral participation gained momentum over the course of the twentieth

M1546 - BIRCH TXT.qxp:ANDY Q7

26

27/10/08

11:31

Page 26

Full participation

century, whereas in Europe there have been only two new adoptions since the Second World War (Cyprus and Turkey). Australasia Compulsory voting has had a patchy history in Australasia. In Australia, electoral participation was first made compulsory in 1915 in the state of Queensland in advance of that year’s general election. The reform was extended to the federal level in 1924, and subsequently to the remaining states between 1926 and 1942 (Hughes, 1966: 81) (see below for full details). The Australian model then served as a basis for several Pacific countries. In Nauru elections were first conducted under the Australian electoral law in 1965, and when the tiny Pacific island state gained independence in 1968, electoral provisions were left largely unchanged, including the requirement to participate in elections (Reilly and Gratschew, 2001). The Australian example was undoubtedly also influential in Fiji’s decision to entrench mandatory voting in its new constitution in 1990. Elsewhere in the region, compulsory electoral participation was adopted in Singapore in 1958 when that state gained independence (Smith and Keith-Lucas, 1960: 230; Rieger, 2001: 242), and it was employed briefly in the Philippines during the authoritarian period between 1973 and 1987 (Hartmann et al., 2001: 189). It also made a fleeting appearance in communist Mongolia between 1980 and 1986 (Gluchowski and Grotz, 2001: 485). Likewise, communist Laos made attendance at the polls formally compulsory in 1989 (Hartmann, 2001a: 132–4). More recently, participation in elections was made mandatory in Thailand in 1997 as part of an effort to clean up elections (see below for more details). Africa and the Middle East There have been some experiments with making electoral participation mandatory in Africa and the Middle East. Lebanon made participation in elections compulsory very briefly, between 1952 and 1957, though the regulations applied to men only (Scheffler, 2001: 174). When Egypt introduced compulsory electoral participation in 1956, the male-only rule also applied and has been in force since. Upon gaining independence, the Republic of Congo10 and Gabon adopted mandatory electoral participation (1960 and 1961 respectively). In neither case was the measure long-lived, however.11 This brief history has primarily focused on geographic regions, but it is worth pausing to consider temporal patterns as well. The most popular periods for the introduction of compulsory electoral participation were following the two world wars; of the 29 states that have retained this institution, 11 adopted it between 1918 and 1939; it was also adopted during this period in about half a dozen states that have since abandoned it. Making electoral participation

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 27

A history of compulsory voting

27

mandatory was again flirted with following the Second World War; of the states where this institution is practiced in 2008, eight adopted it between 1940 and 1965. The period of the ‘third wave’ of democratisation beginning in 1974 and continuing to the present day has seen electoral participation made mandatory in a further four states. This temporal perspective reveals that the introduction of a legal obligation to vote is associated with times of political upheaval. The next section will consider this point in more detail by examining the contexts and motives that surrounded the introduction of compulsory voting. What accounts for the introduction of compulsory voting? Factors associated with the adoption of any institution vary significantly from state to state and from period to period. It may also be the case that regional factors play a part in this process; Baeza attributes the adoption of compulsory electoral participation in virtually all Latin American states to a legal tradition that favoured the formalisation of political development (Baeza, 1998). But country-specific and regional factors may well not tell the whole story. It is therefore worth considering whether there are common elements that led this institution to be introduced across the very diverse range of states throughout the world that have employed it. In this connection, it is instructive to examine the context of this move: who favoured it and who opposed, what arguments were advanced, and what impact was it expected to have? An exhaustive analysis of the debates in all countries is beyond the scope of this chapter. Instead, I have chosen four representative cases to analyse in some detail: Belgium (1893), Argentina (1912), Australia (1924) and Thailand (1997). These four cases span both the temporal and the geographic extent of the institution’s application. They also represent countries at different levels of political and socio-economic development. These differences make it possible to isolate common features in the debates. The concluding chapter will offer a consideration of selected cases where electoral participation was not made mandatory, but where it might potentially be introduced in future (the UK, Canada, and France), thereby providing a set of negative ‘control’ cases to counterbalance the assessment here of the ‘positive’ instances of adoption undertaken here. But first a few general comments about the context in which electoral participation has been made mandatory. It is worth noting that this measure has virtually always been introduced in established states with existing laws governing electoral suffrage. Though a large number of states have gained independence over the course of the past hundred and fifty years, there are very few instances of a new state adopting compulsory electoral participation from the start. The exceptions include Bulgaria (1879), Singapore (1958), Cyprus (1960), the Republic of the Congo (1960), Gabon (1961), and Nauru (1965).

M1546 - BIRCH TXT.qxp:ANDY Q7

28

27/10/08

11:31

Page 28

Full participation

Another generalisation that is evident from the overview of adoptions presented above is that compulsory electoral participation has in most cases been introduced as part of a package of measures designed to improve the democratic quality of the electoral process. In a number of cases, especially early ones, other elements in the ‘package’ often included franchise extensions, the introduction of the secret ballot, and the adoption of proportional representation. More recently, electoral participation has in some cases been made mandatory as part of a democratic transition or an effort to curb electoral abuse (Singapore, Fiji, Thailand). Given the frequent ‘bundling’ of mandatory electoral participation with other reforms, it is not surprising that it has often been adopted on the basis of cross-party support. In a number of cases compulsory voting was advocated by the established right-wing parties at the time of suffrage expansion, supported by the belief that newly-enfranchised workers would generate a highly mobilised and disciplined socialist vote. Making electoral participation mandatory was seen as a means of bolstering the sagging rates of electoral participation among ‘bourgeois’ parties. Writing in 1955, Henry Abraham noted that the right supported the introduction of compulsory voting in a variety of European states: ‘in almost all of the instances of concerted pressure on behalf of compulsory voting on record, especially in European countries, the conservative and centrist political parties have been the driving force behind the movement’ (Abraham, 1955: 10). Tingsten likewise reported that ‘the advocates of the institution have taken their stand on a general, conservative doctrine, while opposition has been forthcoming from radical and particularly from social-democratic quarters’ (Tingsten, 1963 [1937]: 182). A recent quantitative analysis of the adoption of compulsory voting by Gretchen Helmke and Bonnie Meguid confirms that it was more likely to be introduced under right-wing governments, especially in the context of rising support for the left (Helmke and Meguid, 2007). But if the right favoured making electoral participation mandatory on the grounds that the left was perceived as more organised and their electorate easier to mobilise, Tingsten points out that this assumption was often empirically wrong, as subsequent developments revealed that the working class was in fact less likely to turn out to vote voluntarily than more educated sectors of the electorate (Tingsten, 1963 [1937]: 184–5). It is therefore not surprising that in more recent times it has been the left that has most often advocated obliging citizens to go to the polls. The veracity of claims of partisan advantage will be studied in greater detail in Chapter 7; what is relevant from the point of view of the present discussion is politicians’ perceptions of the impact of this institution. It is striking the extent to which making electoral participation mandatory was viewed as a means of mitigating the impact of other electoral reforms, especially franchise

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 29

A history of compulsory voting

29

expansion. The evidence suggests that compulsory electoral participation was often sought as a means of protecting the right against the perceived threat of a larger electorate, and that its ‘bundling’ with alterations to the franchise was the result of bargaining between opposing political camps. Suffrage expansion made electoral compulsion attractive for another reason that parties were undoubtedly accurate in discerning: more potential voters meant more potential demands for traditional services provided in exchange for votes. Legally mandating participation was a convenient means for parties to avoid the prohibitive costs associated with informal (and legally dubious) forms of ‘treating’ or helping voters get to the polls. Given the anticipated burden on their resources which the addition of a large and relatively deprived group of new voters represented, compulsory voting was seen as being in the interest of all parties. This advantage of the institution was of course easy to justify in terms of reducing electoral abuse, which has been a theme in many debates.12 In an effort to explore and refine these generalisations, the four case studies that follow will trace the specific features of the electoral reform processes through which compulsory electoral participation was introduced. Belgium In Belgium mandatory attendance at elections was first demanded by the conservatives in 1858; in 1865 it was studied by a commission as a means of reducing electoral fraud; in 1887 the progressives introduced it into their programme; on 1 June 1893 it was passed as a constitutional amendment by 94 votes to 38 (with two abstentions) in the Chamber and in the Senate on 12 July by 58 votes to one with two abstentions) (Quiri, 1908). The advent of compulsory voting coincided with another modern ‘first’, proportional representation. A law of 28 June 1894 set out the details of implementation of the reforms; this was modified by a law of 1896. Enforcement of compulsory electoral participation was from the start patchy (Quiri, 1908), but the institution was accepted with little difficulty by the electorate, most of whom had never experienced voting as a voluntary activity. In this connection, one of the strongest arguments made at the time was that mandatory electoral participation would guarantee the legitimacy of elections by ensuring that their results reflected the popular will (Pilet, 2005: 5). Another argument often made was that rights and duties go together, and that citizens have a duty to the state to take part in constituting public decisionmaking bodies. Those holding this view often saw compulsory voting as a necessary corollary to universal suffrage (Pilet, 2005: 6). Robson (1923: 572) reports that compulsory voting was introduced in Belgium in 1893 in order to prevent parties from bribing opponents’ voters not to go to the polls. This observation is echoed by Stengers, who claims political parties could foresee that the extension of suffrage rights to the entire

M1546 - BIRCH TXT.qxp:ANDY Q7

30

27/10/08

11:31

Page 30

Full participation

population would be costly if current practices of helping voters to the polls were to persist; there was therefore a general consensus that compulsory voting would be of help to parties in removing this cost (Stengers, 1990: 105; cf Gosnell, 1930: 102). Partisan motives were prominent also. Writing on the Belgian case, Stengers (1990) describes mandatory electoral participation as a ‘corrective’ to universal suffrage, which granted the franchise to a large number of working-class voters who were expected to have views more radical than those of the established electorate. He claims that mandatory voting laws were viewed as a means of preventing radical forces from having a disproportionate influence on outcomes, and thereby moderating electoral results (1990; 105; cf Pilet, 2005: 4; Gosnell, 1930: 102). There was thus relatively widespread support for the package of alterations to the franchise that was adopted at this time. The right and the left were united in their support, while the liberals and some Catholic conservatives opposed the measures (Pilet, 2005: 5). However, Pilet also notes that many parties were divided on the issue of making electoral participation a legal obligation, suggesting that party advantage was by no means the only factor that entered into consideration. Following its initial introduction, mandatory electoral participation was little discussed in Belgium until the 1970s; the institution had been effective in increasing turnout, and was widely accepted by the population. In the 1970s, the subject was re-introduced into political dialogue as a result of the abolition of a similar measure in the Netherlands in 1970, and also due to the rise of a number of ‘new politics’ parties, which picked up the issue (Pilet, 2005: 6). It was not until the 1990s, however, that the debate over compulsory participation in elections really began to heat up. When in the early 1990s Guy Verhofstadt, leader of the Flemish Liberals, came out in favour of the abolition of mandatory electoral participation, the subject entered mainstream political debate (Pilet, 2005: 6–7). As at the time of the nineteenth-century debates, the liberal parties expressed their hostility to the institution (de Winter and Ackaert, 1994; Pilet, 2005). The far right also came out in favour of abolition, though it is unclear from empirical studies that it would have benefited from such a move (de Winter and Ackaert, 2003; Pilet, 2005). The partisan lines were not entirely clear-cut, however. The ecologist parties were initially also for abolition, but in 1995 the francophone Ecolo party reversed its stance and argued for the maintenance of the institution on grounds of the greater legitimacy it afforded (Pilet, 2005: 9). And as in the nineteenth century, there was also dissent within parties; Liberal Willy de Clercq was at odds with his party’s official policy when in 2004 he stated his defence of mandatory electoral participation thus: ‘out of respect for the principle that democracy is not simply a collection of rights, but also a series of

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 31

A history of compulsory voting

31

obligations, I will always remain resolutely in favour of compulsory voting’13 (cit. Pilet, 2005: 7). Many of his fellow party members held the same view, whereas 26.9 per cent of francophone Socialists were, in 1993, opposed to their party’s decision to keep compulsory voting (Pilet, 2005: 7). Finally, socialist forces were keen to maintain compulsory attendance at the polls, not least because it represented part of their ideological heritage, having been one of the main struggles they had engaged in when first becoming a major force in politics (Pilet, 2005: 10). Electoral participation was thus introduced in Belgium for a combination of normative and partisan reasons; it was initially viewed as a means of mitigating the expected effects of franchise expansion, and it has remained in force due to lack of a strongly co-ordinated opposition, as well as uncertainty as to the likely effects of making electoral participation voluntary. Argentina Electoral participation was made a legal obligation in Argentina in 1912 as part of the so-called Sáenz Peña Law that democratised several aspects of the suffrage and marked the beginning of modern electoral competition in the Latin American country. Prior to 1912, Argentine elections had been racked by corruption, and the parties on the left of the political spectrum (the principal such party at the time being Hipólito Yrigoyen’s Radicals) had little chance of electoral success. Electoral fraud and abuse had steadily eroded the legitimacy of the political system, which suffered regular and violent popular protests and electoral boycotts led by the Radicals (Díaz, 1983: chap. 3; Jones et al., 2005: 59).14 One of the main demands of the Radicals was the introduction of secret and compulsory voting. President Roque Sáenz Peña of the conservative National Autonomist Party saw electoral reform as a necessary means of restoring confidence in the political process and taming the leftist threat (Díaz, 1983: 16; McGree Deutsch, 1986: 30). A period of behind-the-scenes negotiation between Yrigoyen and Sáenz Peña helped to frame what finally emerged in 1912 as a law that for the first time established effective universal suffrage, together with secret voting and mandatory participation (Díaz, 1983: 66–78). Though a key demand of the Radicals, compulsory voting appears to have been readily accepted by Sáenz Peña; in his own words ‘it is not enough to guarantee the suffrage, we must create the electorate’;15 in other words, he wanted to make active citizens of Argentines. It is nevertheless worth noting that the obligatory nature of the franchise was one of the aspects of the law that was most contentious during parliamentary debates; it was claimed by some that mandating participation would only serve to mask the true extent of electoral indifference, though it was also argued that such a measure would help to overcome apathy. The latter argument finally won the day when penalties for nonparticipation were reduced in the final version of the law (Díaz 1983: 77–8).

M1546 - BIRCH TXT.qxp:ANDY Q7

32

27/10/08

11:31

Page 32

Full participation

The result of the reforms was, as expected, the integration of the mass public into institutionalised politics and increased political stability; the Radicals became the dominant political party, and Yrigoyen was elected president in 1916, 1922 and 1928. It is not clear, however, that the mandatory nature of electoral participation contributed significantly to this result, as effective sanctions for non-participation have never been applied. It is perhaps for this reason that Argentina has retained the institution with relatively little debate until the present time. The Argentine case is interesting in that it highlights the role of legitimacy crises in bringing about reforms that are not in the immediate political interest of those who introduce them. The situation had by the first years of the twentieth century grown so tense that legally incorporating the mass population into electoral politics was seen as preferable to retaining the status quo of violent protest and non-participation, even if it meant that those who introduced the measure lost politically as a result. Australia Australia adopted mandatory electoral participation in 1924, during the heyday of enthusiasm for the measure, and at a time when it had already become part of the established repertoire of options among democracies. As in Belgium, a combination of normative concerns motivated its introduction; practical issues of electoral administration were also prominent in the Australian debate. Australia is unusual in having introduced compulsory electoral participation as an isolated measure, not linked to other changes to the electoral system (though it was discussed at the time in connection with other possible reforms, such as the preferential ballot; Gow, 1971: 207). Following the introduction of compulsory voting in Queensland in 1914 and its use in a referendum the following year, the Labor party adopted the reform in its manifesto, undoubtedly at least in part due to their success at the state level under the institution. Hughes suggests that the reform of 1924 was strongly motivated by partisan concerns, being ‘the last desperate attempt of a government in the skids to save itself ’ (1966: 81). Yet ultimately compulsory voting gained cross-party support. There are several reasons for this. Firstly as in Europe, it was argued that the left (the Labor party) already used a form of social compulsion (via trade unions) to get its supporters to the polls, with the implication that legal compulsion would rectify the partisan imbalance (Hughes, 1966: 82). Secondly, there was a common perception that party costs would be reduced due to the fact that voters would no longer be in a position to demand transport to the polls (Hughes, 1966: 82; Gow, 1971: 207–9). The introduction of compulsory electoral participation thus appears to have been driven in part by a desire to avoid manipulative mobilisation, as in Belgium. This can be seen as a

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 33

A history of compulsory voting

33

self-serving desire by parties to reduce their costs, yet it also reflected normative concerns with the quality of elections. A number of other normative considerations also came into play. In reviewing the points made at the time of the reform, Hughes sets out 17 arguments for and 10 arguments against making electoral participation mandatory (Hughes, 1966: 81–3). Those in favour lauded it as a means of raising turnout, achieving greater legitimacy and better democratic decisions, as a tool of political education, and they viewed voting as a duty not only a right. Those against argued that it was an infringement of liberty, that it would be difficult to enforce and that it would lead to a higher number of invalid ballots (‘informal’ ballots in Australian parlance). A link was also made between compulsory registration and compulsory electoral participation; the one was seen to entail the other from a logical point of view, and it was also seen to make practical sense in that the expense of registering the entire population ought not to be wasted. Furthermore, it was argued that mandatory electoral participation provides a useful means of monitoring the accuracy of electoral registers. Those against the measure cited enforcement difficulties as well as the added cost of administering sanctions. Since its introduction, compulsory voting appears to have been widely accepted in Australia; as Hughes noted in 1966, ‘nowhere in Australia have there been significant attempts to abolish compulsory voting once it had been introduced’ (1966: 83). The same can be said 40 years later. In contradistinction to Belgium, where its abolition is frequently debated, the legal obligation to vote does not appear to be the subject of widespread discussion in Australia, and nor have there been any significant moves to loosen Australia’s strict enforcement apparatus. Thailand Thailand is the state most recently to make electoral participation a legal obligation. In 1997 a raft of electoral reforms was introduced as part of a major constitutional overhaul in an effort to improve the quality of democracy; among these was mandatory voting – viewed as a means of curbing electoral abuse and in particular vote-buying (Murray, 1998; Callahan, 2000: 34–7). Thai politics had gone through a period of political turbulence in the 1980s and early 1990s, characterised by military coups and frequent changes of government. In 1996, a Constitutional Drafting Assembly was set up following calls by intellectuals and NGOs for constitutional reform (Callahan, 2000: 99). The Assembly drafted a so-called ‘People’s Constitution’, adopted in the autumn of 1997 as an antidote to the emerging financial crisis, which was widely seen as the product of rampant corruption (Callahan, 2000: 165). In addition to establishing a number of independent monitoring bodies (including

M1546 - BIRCH TXT.qxp:ANDY Q7

34

27/10/08

11:31

Page 34

Full participation

an independent Electoral Commission), the new constitution also introduced several significant changes to electoral procedures. The desire to curb vote-buying was one shared by democracy activists and many politicians alike. The amount of money spent in elections had spiralled in recent years, and there was a common perception that vote-buying had ‘got out of hand’ (Murray, 1998: 527). As in the European countries where franchise expansions had increased the cost of rewarding voters, the rising cost of elections in Thailand provided politicians of all colours with a strong incentive to take action to reduce it. The institution of the constitutional assembly was another factor that led normative concerns with reducing corruption to take on a greater role than might otherwise have been the case. The new constitution provided for a mixed electoral system for the House of Representatives, a directly-elected Senate (to be elected under the single-non transferable vote system), and mandatory voting, together with a wide range of other reforms of the political system. Mandatory voting was seen as a means of reducing electoral corruption in that it would increase the number of people going to the polls, and thereby make vote buying prohibitively expensive (Murray, 1998: 533). Thus in Thailand, normative concerns with holding clean elections were among the most important determinants of reform, but the introduction of mandatory electoral participation also coincided with a common perception by political parties that the current electoral customs were becoming dysfunctional for them (as well as for the country as a whole). As in Argentina, a legitimacy crisis led political actors to look beyond their immediate interest and shape institutions designed to improve the tenor of politics – the constitutional assembly, and the electoral reforms it produced, including the legal obligation to participate in elections. Examination of these four cases has demonstrated that there were three common motives behind the introduction of compulsory voting: partisan advantage, strengthening democracy, and the prevention of electoral abuse. The first reason, which was most common in the European and Australian contexts, was an interest-based rationale; established parties feared that under mass enfranchisement highly-organised workers’ movements would be capable of mobilising their members efficiently, leading to higher rates of turnout on the left than among adherents of the more conservative parties. When it accompanied franchise expansion, mandatory voting was thus viewed as a ‘corrective’ that would go some way toward protecting the right. It is noteworthy that this was also one of the rationales for introducing proportional representation at the time of mass enfranchisement (Carstairs, 1980; Boix, 1999). Making electoral participation a legal obligation was therefore often an alternative or additional means of achieving this end. The second factor is a normative one; elections in which there was full participation would have more legitimacy than those held under a voluntary

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 35

A history of compulsory voting

35

system, and they would serve to create an active citizenry. There may well be practical reasons why leaders view it as being in their collective interest to enhance the democratic quality of elections, as this promotes the legitimacy of the electoral system and therefore their own legitimacy as politicians. Enhanced legitimacy in turn leads to greater governability, stability and ease in mobilising supporters. Another of the principal arguments advanced in favour of compulsory attendance at the polls was that it would clean up elections by reducing corruption and undue influence. The desire to minimise electoral abuse was evident particularly in the context of the widespread provision by parties of selective incentives to voters. On the face of it, this may seem like a normative argument, but it was also one that took considerable pressure off party resources, especially in the face of growing electorates or the growing cost of individual votes. Thus whether for self-interested or disinterested reasons, making electoral participation mandatory was seen as a means of replacing pressure to vote exercised by parties with legal pressure. A final factor that is not brought into relief in the case studies above is the role of diffusion. Though it is difficult to trace precise causal processes, it is undoubtedly not by chance that there have been ‘pockets’ of mandatory electoral participation in places such as the low countries, as well as the Alpine states of Switzerland, Liechtenstein and Austria (and nineteenth century Bavaria). The diffusion process is also evident in Latin America, with the gradual adoption of mandatory electoral participation first in Central America, and then spreading further south. Finally, the role of Australia in diffusing the use of its compulsory voting institutions in the Pacific region is seen in the relatively recent adoptions of this institution in Nauru and Fiji. The phenomenon of the regional diffusion of popular and apparently successful reforms is well-known to students of electoral systems. What is perhaps most interesting in the present case is the range of different contexts in which it occurred. This suggests the need for a more systematic analysis of the distribution of mandatory electoral participation in the world today, which is the subject of the next section. Compulsory voting in the world today With the exception of Eastern Europe and the former Soviet Union, all regions of the world contain states where participation in elections is compulsory (though the current use of this institution in Africa is confined to Egypt). The Eastern European exception is undoubtedly due to the fact that informally enforced mobilisation under communism made the citizens of these states averse to any kind of formal requirement to participate (Birch, 2003: chap. 3). Table 2.1 lists the 29 states that currently employ compulsory voting, together with the date of its adoption, whether there is a regime of sanctions

M1546 - BIRCH TXT.qxp:ANDY Q7

36

27/10/08

11:31

Page 36

Full participation Table 2.1 Compulsory voting in the world today (2008)

Country

Date adopted

Europe Belgium 1893 Cyprus 1960 Greece 1929 Luxembourg 1919 Switzerland 1903 (Schaffhausen) Turkey 1982 Americas Argentina 1912 Bolivia 1924 Brazil 1932 Chile 1925 Costa Rica 1936 Dominican Republic 1966 Ecuador 1929 El Salvador 1950 Guatemala 1965 Honduras 1894 Mexico 1917 Panama 1928 Paraguay 1967 Peru 1931 Uruguay 1924 Venezuela 1958 Asia and the Pacific Australia 1924 Fiji 1990 Laos 1989 Nauru 1965 Singapore 1958 Thailand 1997 Middle East and Africa Egypt 1956 (men only)

Sanction applied?

Constitutionalised?

Yes Yes No Yes Yes

Yes (Art 62) No Yes (Art. 51) No Yes (Art 23(2) )

Yes

Yes (Art. 175)

No Yes No Yes No No Yes No No No No No No Yes Yes No

Yes (Art. 37). Yes (Art. 219) Yes (Art. 14) Yes (Art. 15) Yes (Art. 93) Yes (Art. 88) Yes (Art. 27) Yes (Art. 73) Yes (Art. 136) Yes (Art. 5) Yes (Art 36) Yes (Art. 129) Yes (Art. 118) Yes (Art. 31) Yes (Art. 77) No

Yes Yes N/A Yes Yes Yes

No Yes (Art. 56) No No No Yes (Section 68)

Yes

No

for non-voting, and whether the institution is prescribed in the constitution. As can be seen in this table, just over half of all states that make electoral participation mandatory are located in Latin America. It is also noteworthy that about half the states that currently employ the institution have adopted it since the Second World War. Half impose sanctions for non-participation, and in

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 37

A history of compulsory voting

37

about two-thirds of all cases, the legal obligation to vote is entrenched in the constitution. These figures re-emphasise the considerable variety in the geographic and temporal contexts in which mandatory voting has been adopted. This variety makes it somewhat difficult to make general statements about the ‘types’ of states where this institution is used, but it is nevertheless worthwhile considering the factors they have in common. The diffusion effect mentioned above undoubtedly goes a long way toward explaining regional patterns, but the patchiness of the institution in most parts of the world suggests that there are other factors at play as well. Massicotte and colleagues claim that a disproportionate number of the states with mandatory electoral participation have Catholic majorities, and they speculate that ‘Perhaps it is natural for Catholics to believe that citizens have moral obligations not only towards God and the church, but also towards the state’ (Massicotte et al., 2004: 37). Yet closer consideration reveals that the majority of Catholic states among our sample of 29 is almost entirely due to the ‘Latin American effect’. Outside Latin America, the only mandatory electoral participation states that have Catholic majorities are Belgium and Luxembourg. Indeed, the quantitative analysis carried out by Helmke and Meguid did not find Catholicism to be a significant predictor of this reform (Helmke and Meguid, 2007). Massicotte et al. also cite colonial heritage as a factor in the distribution of compulsory voting in the contemporary world. They note that this institution tends not to be found in former British colonies, whereas former Spanish colonies are far more likely to have the institution (2004: 38). In fact, nearly half the non-Latin American states to make electoral participation mandatory are former British colonies (Cyprus, Australia, Fiji, Nauru, Singapore and Egypt), while the institution was introduced in Latin America prior to its adoption in Spain in 1907. Colonial heritage thus provides no better cue than religion as to the distribution of mandatory voting rules. Perhaps a more promising explanatory factor is the size of a state’s population. It may be the case that smaller states find it easier to implement and enforce a legal obligation to vote, whereas for more populous states the logistic challenge is seen as being too great (Giraud, 1931). This would explain why small states such as Cyprus, Luxembourg, Costa Rica, Panama, Uruguay, Laos, Fiji, Nauru and Singapore have adopted the institution, but it does not shed light on its maintenance in rather large states such as Brazil, Mexico, Argentina, Thailand, Egypt and Turkey. Median sate population size in the world was 5.6 million in 2005,16 so ‘small’ must be understood in this context. Of our sample of 29 states, 19 have populations larger than this, which suggests that in fact larger states are more likely than smaller ones to oblige their citizens to participate in elections. It is worth noting, however, that small state size emerges as a significant – though weak – predictor of legislating electoral

M1546 - BIRCH TXT.qxp:ANDY Q7

38

27/10/08

11:31

Page 38

Full participation

compulsion once other variables are controlled (Helmke and Meguid, 2007), suggesting that other factors condition the relationship. We are perhaps forced to conclude that, aside from the predominance of European and Latin American states among the 29 countries where electoral participation is mandatory, they have little in common. Genetic factors thus appear to be most important in determining use of this institution in the world today. It was adopted in different places at different times, usually in association with other democratic reforms. Frequently this involved a change to the constitution that was then difficult to alter (either because the constitution itself was difficult to alter or because it was difficult to justify removal of a civic obligation from the state’s founding law). And as noted above, the adoption of the institution took place mainly in the context of established rather than new states, which perhaps goes some way toward explaining why we find it in most widespread use on two of the continents with the longest traditions of statehood. These genetic stories, more than anything else, appear to account for why some states in today’s world make attendance at the polls a legal obligation and others do not. Conclusion This chapter has sought to provide a brief review of where and why mandatory electoral participation has been introduced and employed over the course of the past century and a half. It has become clear that the institution has been used in a very wide variety of contexts, it has been adopted under a wide variety of different political circumstances, and it has been seen as the solution to a range of different problems. This discussion has highlighted a number of normative arguments that have at various times been used in defence of making electoral participation a legal obligation, but it has not gone into detailed consideration of these points as such. This is the task of the next chapter, which will provide an analysis of the various arguments for and against, and will draw from this analysis a list of the empirical claims upon which the normative assessments in part rest. Notes 1 A similar attempt at constitutional amendment was made in Oregon in 1919, but the proposal was rejected by referendum. Abraham reports that between 1888 and 1951 there were attempts to introduce mandatory electoral participation in 48 US states (Abraham, 1952: 139–55). 2 The conceptual aspect of this debate is explored in greater detail in Chapter 3. 3 Caramani views the French Revolution as the key turning point in this process.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 39

A history of compulsory voting

39

4 ‘Regenschaftsbericht Lantag, Regierung, Gerichte, 2004’, Principality of Liechtenstein, http://www.llv.li/amtsstellen/llv-rk-amtsgeschaefte-rechenschafts bericht.htm. 5 The requirement to participate in elections was removed from the Dutch constitution in 1922, but retailed in law and practice for the following half century (Irwin and van Holsteyn, 2005: 3). 6 It is worth noting, however, that there is strong political pressure to vote in Cuba, and that Colombia provides positive material incentives to those who participate; see Chapter 1. Though compulsory electoral participation has frequently been proposed in Colombia, it has never been introduced (Hernandez Becerra, 1986: 23). 7 Article 23 of that constitution states that ‘voting is irrelinquishable and compulsory for all citizens’ (‘El voto activo es irrenunciable y obligatorio para los ciudadanos’) [author’s translation from the Spanish]. Compulsory registration was only in force from 1939, however (Krennerich, 2005: 483). 8 Registration was at this point voluntary and there were no sanctions for non-participation, so the fact of making electoral participation formally mandatory had a largely declaratory character. 9 As noted above, it had also been enshrined in the constitution of the Federation of Central America (of which Guatemala was a part) between 1921 and 1922. 10 The Republic of the Congo was subsequently renamed Zaire, and is now called the Democratic Republic of Congo. 11 Zaire/The Democratic Republic of Congo did formally reintroduce mandatory electoral participation between 1987 and 2005, but no elections were held during this period. 12 In Singapore, ‘The recommendation that compulsory voting should be introduced [. . .] came from the Commission of Inquiry into corrupt, illegal or undesirable practices at elections; they considered that the use of compulsory voting, coupled with a prohibition on the provision of private transport to take voters to the polling stations and a prohibition on any form of canvassing for votes on polling day, might counter the influence of secret societies’ (Smith and Keith-Lucas, 1960: 230). On the Belgian, Argentine, Australian and Thai examples, see below. 13 ‘Pour le respect du principe qu’une démocratie n’est pas uniquement une collection de droits mais aussi une série d’obligations, je resterai toujours résolument en faveur de l’oblgation de voter’ [author’s translation from the French]. 14 Díaz reports that turnout in the elections preceding the reforms had been a mere five per cent (1983: 78). 15 ‘No es bastante garantizar el sufragio, sino, que necesitamos crear al sufragante’ [author’s translation from the Spanish] (cit. Díaz, 1983: 111). 16 This figure is taken from examination of World Bank population data, found at www.worldbank.org.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 40

3

Normative arguments for and against compulsory voting

As a mechanism for collective decision-making, democracy assumes popular participation, yet democratic theorists are divided over how much and what forms of participation are necessary for a political system to be democratic. There is also a potential tension between the collectivist exigencies of democracy and the individualist orientation of liberal notions of rights, both of which are commonly invoked in normative discussions of institutional design. The views of those such as Schumpeter (1947) who warned of the dangers of excessive participation have been largely discredited, but many are still reluctant to concede the necessity of full participation in decision-making. The aim of this chapter is to consider the normative arguments for and against the legal obligation to participate in elections, and to identify in these arguments testable empirical assumptions that require further investigation. The normative arguments surrounding the institution of mandatory voting fall into three principal categories: arguments revolving around the notions of individual rights and duties; arguments about collective rationality and democratic legitimacy, and consequentialist arguments about the empirical effects of this institution. Rights and duties Rights and duties are concepts that are central to most understandings of political order, yet there is considerable disagreement over the relation between them. In the specific context of this study, there is a long-standing debate over whether the right to vote entails a corresponding obligation, and if so, whether that obligation should be reflected in law.1 The relationship between rights and duties has been a leitmotif in discussions of voting throughout its modern history, and in the early literature there was a tendency to see the two as mutually exclusive alternatives. The francophone and Ibero-American literature distinguished between vote-as-right and vote-as-function, the latter term referring to voting as a function performed by all citizens of a state, which can be roughly equated with the notion of voting as civic duty or public trust (Abraham, 1955: 5–6).2

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 41

Normative arguments for and against compulsory voting

41

As the French Constituent Assembly of 1791 was the first body in the modern world to grant universal suffrage, the modern notion of voting can usefully be traced to the debates held at this point. The notion of voting as a duty rather than a right was common in discourse of this period (despite the common association of the French Revolution with rights), Delegate Thouret argued (11 August) that ‘the status of elector is a public function whose delegation society must determine’3 (cit. Quiri, 1908: 27). This view is echoed in the words of Barnave (11 August), who argued that ‘the function of elector is not a right; it exists for society, which has the right to determine its conditions’4 (cit. Quiri, 1908: 28). It is noteworthy, however, that the Constituent Assembly did not go so far as to make voting compulsory in France. In favour of the idea of voting as a civic duty, Graux, a deputy in the Belgian parliament participating in the debates prior to the introduction of compulsory voting there a century later, argued that ‘one does not vote for oneself; the franchise is exercised in the social interest’5 (cit. Quiri, 1908: 25). Similarly, John Stuart Mill argued in ‘The Mode of Voting’ in Considerations on Representative Government: Those who say the suffrage is not a trust but a right, will scarcely accept the conclusions to which their doctrine leads. If it is a right, if it belongs to the voter for his own sake, on what ground can we blame him for selling it, or using it to recommend himself to any one to whom it is in his interest to please? A person is not expected to consult exclusively the public benefit when he makes use of his house, or his three per cent stock, or anything else to which he really has a right. The suffrage is indeed due to him, among other reasons, as a means to his own protection, but only against treatment from which he is equally bound, so far as depends on his own vote, to protect every one of his fellow citizens. His vote is not a thing in which he has an option; it has no more to do with his personal wishes than the verdict of a juryman. It is strictly a matter of duty; he is bound to give it according to his best and most conscientious opinion of the public good’ (Mill, 1991 [1861]: 354).

Mill’s view is echoed by that of Lord Bryce: ‘as individual liberty consists in exemption from political control, so political liberty consists in participation in legal control’ (1921, Vol. 1: 60). A more modern perspective, and one that prefigures later debates, is that voting-as-duty and voting-as-right are not incompatible, but that they go naturally together. Thus Félix Moreau, a French legal scholar, voiced the view in 1896 that: Liberty is only defended by those who cherish it, by those who appreciate its guarantees and accept the duties it imposes; the right of suffrage is among the guarantees and the duties of liberty. [. . .] Each person belongs to political society and receives benefits from it, perhaps against their will; they must bear the responsibility of those benefits. Democracy confers advantages and imposes duties; together they form an indivisible whole.6 (Moreau, 1896: 8, 19)

M1546 - BIRCH TXT.qxp:ANDY Q7

42

27/10/08

11:31

Page 42

Full participation

In recent discussions of the suffrage, there is a tendency to agree that electoral rights and electoral duties are in theory compatible; debate centres rather on whether suffrage rights entail obligations, and if so, whether those obligations ought to be made legally binding. Many would concede that though electors in a democracy have a moral obligation to vote, this ought not to be translated into a legal obligation. It is to the more recent debates that we now turn. Arguments in favour of compulsory voting The rights-based defence of mandatory electoral participation starts from the premise that duties and obligations are intimately connected to rights. Voting is a necessary attribute of citizenship; it is a public trust, a moral obligation, a duty of citizenship. Democratic obligations thus follow directly from democratic rights (see Lijphart, 1997; Hill, 2002a; Engelen, 2007). The obligation to participate in elections can be defended on the grounds of equality inherent in the definition of the democratic choice situation; all members of the community have a duty to contribute to collective decision-making if they are to enjoy its fruits, irrespective of any consequentialist arguments (considered below) as to the impact of such participation. The duty to vote is one that is tacitly and explicitly admitted in many empirical analyses of electoral participation. In their seminal argument on the rationality of voting, Riker and Ordeshook (1968) are obliged to invoke the perceived duty to vote as a term in their cost-benefit analysis (‘D’) in order to explain away the supposed irrationality of going to the polls. The notion of electoral obligation is thus not particularly contentious. What is more contentious is whether this widespread social norm ought to be made legally binding. In discussions of this question, it is commonly noted that legal compulsion is ubiquitous in modern democracies, yet most people have become accustomed to carrying out the duties required of them by the state and do not think of them as a burden. Parallels are often made between mandatory electoral participation and other legally binding duties such as paying taxes, serving in the military, serving on a jury and the mandatory education of children (Feeley, 1974; Lijphart, 1997: 11). There are also a host of other obligations that citizens in modern states fulfil in the interests of society, including wearing safety belts in cars, undergoing security checks in airports and other locations, disposing of household refuse in a responsible manner, etc. Completion of the census is perhaps the most apt comparison, as an election can be seen as a ‘political census’. Though enforcement of the duty to complete the census takes a light touch in most countries, participation is nevertheless generally close to universal.7 Another relevant comparison is electoral registration, which is likewise effectively compulsory in most countries. Finally, the common argument that the right to vote implies a right not to vote has been countered by Heather Lardy (2004), who argues that the right to

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 43

Normative arguments for and against compulsory voting

43

vote is a positive right: ‘The democratic standing which the right to vote confers is fully intelligible only as the award of an entitlement to participate actively, rather than as a merely passive possessor of the franchise’ (2004: 311). Moreover, ‘non-voting is not only a matter of public concern; it is, like voting itself, a public act’ (2004: 313). Lardy goes on to distinguish between liberty as non-interference and liberty as non-domination, arguing that the right to vote is a matter of freedom as non-domination, and as such it does not entail the right not to vote. It is worth quoting this argument at length: Those who argue for a right not to vote presume an underlying theory of liberty as non-interference. They assume this perhaps because the right to vote tends to be presented as an ordinary negative liberty, protecting against interferences by agencies of state, It has been argued here that compulsory voting may only be represented as an interference with liberty by mischaracterising the sort of freedom which the right to vote represents. The right to vote is about freedom as nondomination. It is for those who oppose compulsion to demonstrate how it would threaten the liberty to which the right to vote relates. The liberty which the right not to vote asserts can provide only an inaccurate and impoverished account of the relation between freedom and voting. The liberty which the right to vote represents protects individual decisions about how to vote, and about what value to attach to the act of voting. It does not, however, grant those whom it recognizes as electors a freedom not to vote at all. (Lardy, 2004: 315)

Finally, it is noteworthy that the European Court of Human Rights has ruled (X v. Austria, 1971) that compulsory voting does not violate freedom of thought, conscience and religion (Article 9 of the European Convention on Human Rights) so long as voters have the option of casting blank or spoilt ballots (Baston and Ritchie, 2004: 36; Electoral Commission, 2006: 10). Arguments against compulsory voting The most common rights-based argument against mandatory voting laws is that the right to vote implies a right not to vote (Mackenzie, 1958: 131; Katz, 1997: 244–5; Lever, 2007: 26–35). It is only one step from this argument to viewing compulsory voting as inherently undemocratic. Just such an argument is put forward by Henry Abraham: It is difficult to see how we can remain faithful to the principles of democracy by compelling people to exercise an ostensible privilege, that of voting, contrary to their will. Since the ability to vote is a privilege conferred upon the citizens, it ceases to be such when he [sic] is compelled to exercise it. How can he be deprived of his right not to vote? (Abraham, 1955: 33, emphasis in the original)

In this context, the analogy is frequently made between electoral compulsion and the styles of political persuasion employed in authoritarian or totalitarian states. As W. H. Morris Jones puts it, ‘It needs no demonstration that a

M1546 - BIRCH TXT.qxp:ANDY Q7

44

27/10/08

11:31

Page 44

Full participation

totalitarian view of political life easily involves an obligation not only to vote, but to do much more – and to do it all, moreover, in the ‘right’ direction’ (Morris Jones, 1954: 36; cf Jakee and Sun, 2006: 64). It has also been claimed that moral obligation does not necessarily imply legal obligation, and that though voting is a moral duty for those with informed opinions, it is an act which when translated into a legal requirement places undue restraint on individual liberty (Abraham, 1955: 6). This argument suggests that the benefits of compulsory voting must be weighed against the ‘costs’ to liberty that it represents. In a subtle argument, Richard Katz alludes to the classical liberal view that a person’s rights should only be infringed when the exercise of those rights impacts negatively on others, which he claims is not the case with compulsory voting: ‘If voting is a right, then compulsion would be justified in liberal terms only if failure to vote would impose substantial costs on others. But one could well argue that far from imposing costs, abstainers actually increase the political influence of those who do choose to vote’ (1997: 245). Lomasky and Brennan (2000) provide a cogent critique of the notion of voting as a personal duty on the grounds that (a) objectively speaking, voters have little chance of influencing the outcome in most cases: ‘On any list of the top thousand or so ways an ordinary citizen can usefully augment social wellbeing, casting a ballot on election day will either rank low or not be present at all’ (Lomasky and Brennan, 2000: 74), and (b) voting is not a sufficiently demanding form of civic engagement for it to qualify as a significant act of civic virtue: ‘the mere act of showing up at the polls every several years and grabbing some levers is palpably inadequate to qualify as a significant act of political expression’ (2000: 82). They also reject on principle the ‘superannuated’ republican argument for civic engagement, on the grounds that it is incompatible with the modern state, in which politics is so complicated and people’s lives are so busy, that most people do not have the time or inclination to devote themselves intensively to public affairs (2000: 84). The authors even go so far as to maintain, in a provocative conclusion to their argument, that inculcation in the people of the sense of a duty to vote serves political élites, in that it inclines voters to feel that by going to the polling station they are fulfilling their ‘necessary and sufficient’ (emphasis added) duties as citizens; for this reasons the authors claim that ‘belief in a duty to vote is the opiate of the democratic masses’ (2000: 86). It is worth noting, however, that Lomasky and Brennan distinguish between arguments about the duty to vote and compulsory voting as an institution. They note that there may be good grounds for adopting such an institution, aside from arguments for voting as a duty.8 In brief, the rights-based argument against the legal obligation to vote is that the notion of a ‘duty’ to vote is questionable in terms of individual morality, and that even were one to grant this notion, it would not imply that voting

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 45

Normative arguments for and against compulsory voting

45

should be made legally obligatory. To do so would be to place an unacceptable limitation on personal freedom, which would violate the right of citizens not to vote. Democratic legitimacy Closely linked to the argument about rights versus duties is that about legitimacy. If, as happened in Belgium in 1843, only 14 per cent of the eligible electorate present themselves at the polls (Pilet, 2005: 4), it is debatable whether one can consider the resulting assembly to have the same legitimacy as it would have if the entire eligible electorate had taken part in the process. This argument is perhaps the one which most troubles the political classes in polities with low turnout; politicians wonder if they have sufficient legitimacy to govern if they have been elected on the basis of only a small minority of eligible voters. This situation can be exacerbated when it is combined with an electoral system that grants a premium to the strongest parties. In this context, mandatory voting laws can be seen as a necessary (if insufficient) means of achieving true and accurate representation (and in this sense as a natural corollary of proportional representation, along side which it was commonly debated in Europe at the time of the franchise expansions that took place at the end of the nineteenth and the beginning of the twentieth centuries; see Chapter 2). This view is one that has been voiced since the origins of the institution; speaking in the Belgian national assembly in the debates that preceded the introduction of compulsory voting in that country, M. Beernaert argued that ‘in order that the law should be the expression of the national will, it is necessary that this will should be expressed, should manifest itself, and obviously this is not the case if electors neglect to go to the polls. Not only is the national will not known, but too many abstentions can distort expression [of the national will] by altering the majority’9 (cit. Quiri, 1908: 57). Likewise, deputy M. Smet de Nager argued that ‘if it is important that the chambers should be the adequate expression and the authorised body of the legal state, how can one deny that it is of general utility that the electoral function should really be exercised by those to whom it is entrusted’10 (cit. Quiri, 1908: 57–8). Arguments in favour of compulsory voting Many contemporary advocates of compulsory voting have also viewed it as a means of enhancing democratic legitimacy and collective rationality (see especially Hasen, 1996: 2137, 2178; Matsler, 2003: 961; Hill, 2006: 209; Engelen 2007). It has been held that the legal obligation to vote guarantees cooperative behaviour through self-binding in the collective interest; it thus ensures that the decisions taken have democratic legitimacy (cf Feeley, 1974; Wertheimer, 1975; Katz, 1997: 243; Hill, 2002a). There are two elements to

M1546 - BIRCH TXT.qxp:ANDY Q7

46

27/10/08

11:31

Page 46

Full participation

this argument, which often appeals in its modern formulation to concepts native to social choice theory.11 Firstly, participation can be seen as an equitable means of overcoming the collective action (free rider) problem inherent in voting (Feeley, 1974; Wertheimer, 1975; Katz, 1997: 243; Lijphart, 1997: 9). In what is perhaps the most persuasive argument in this vein, Wertheimer notes that a competitive electoral system is a public good from which all citizens derive benefit, but that ‘as with many public goods, it is irrational for an individual to contribute voluntarily to maintaining the electoral system, [therefore] to ensure the continued viability of the electoral system and to prevent unjust “free riding”, the obligation to vote should be enforced’ (1975: 278; cf Feeley, 1974: 237). In Richard Katz’s words: Because the chance that a single voter will alter the outcome of an election is trivial, it is never cost-effective for a singe individual to vote. But if no one voted, democracy would collapse and, short of that, many interests would go unprotected because the expected benefit for a single individual would never exceed the cost of voting, and individuals would not be able to organize to protect their aggregate and collective interest through collective action. Therefore, rational liberal individuals might choose to be coerced to vote so long as they were assured that others would be coerced as well in order to achieve the benefits of collective action. (1997: 243)

In a similar argument, Lisa Hill links the collective action argument to the ability of compulsory voting laws to overcome voters’ uncertainty about the intentions of their fellow electors and co-ordinate on collectively beneficial behaviour: Because compulsory voting ensures full turnout, it is able to overcome two of the most common causes of ‘rational abstention’: informational uncertainty about other potential voters’ intentions, on the one hand, and the transaction and opportunity costs of voting, on the other. Compulsion enables voters to overcome the problem of insufficient information, not only about the real value of an individual vote but also about the intentions of other voters. [. . .] Rather than perceiving the compulsion as yet another unwelcome form of state coercion, compulsory voting may be better understood as a coordination necessity in mass societies of individuated strangers unable to communicate and coordinate their preferences. (Hill, 2006: 214–15)

Wertheimer goes on to argue that non-voting is a ‘normal’, rational behaviour, and that those who vote out of a sense of efficacy are in some sense deluded: ‘[t]he alternative to legal compulsion is a system that encourages us to be irrational, in which we systematically delude ourselves about the extent of our efficacy and in which we need continually to promote and reinforce a strong sense of guilt’ (1975: 290). According to this line of reasoning, get-out-the-vote

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 47

Normative arguments for and against compulsory voting

47

campaigns give voters the false impression that by simply turning up at the polls, they have fulfilled their moral duty to participate in political life. Requiring citizens to vote would encourage them to consider more meaningful forms of participation (1975: 292). This last claim is relevant in the context of those who argue that voluntary voting might be thought to generate a benefit to the voter, in that voting would provide the citizen with a rewarding sense of having discharged their civic duty (see Lardy, 2004: 24 for an articulation of this view; cf Feeley, 1974: 238). By withdrawing this benefit, it might be supposed that citizens would seek to discharge their civic duty in other ways, and that those other ways might include more demanding forms of participation. Secondly, full participation is necessary in order for truly democratic decisions to be made, for only in these circumstances will the resulting decision reflect the views of the entire electorate (Lijphart, 1997; Engelen, 2007). For a decision to be made on an equal basis, all those who have a view on issues of collective decision must be obliged to express those views through the electoral process. In theory, non-participation by those who are completely indifferent between alternatives should not bias the outcome, but not all non-voters can be assumed to abstain out of indifference. It has be shown that unless the preference distribution of non-voters is symmetric and unimodal (which is unlikely to be the case), non-participation will most likely alter collective choice outcomes (Enelow and Hinich, 1984: 90–5; Mueller, 2003: 330–2). In this case the only valid exclusion would be one based on a principle, such as temporary infirmity, that could be seen to be largely unrelated to vote choice. Full participation does not guarantee an equitable outcome (which is in part dependent on the electoral system), but it may be seen as a necessary condition of such an outcome. This view reflects the inclusiveness requirements of republican conceptions of democracy, which suggest that compulsory voting may be necessary for the inclusion of all citizens in the representative process (e.g. Pettit, 1997: 191; Déloye, 2005). Lisa Hill sums up several of the arguments relating compulsory voting to democratic legitimacy thus: Low turnout impugns a number of fundamental democratic values such as popular sovereignty, legitimacy, representativeness, political equality and the minimization of elite power. Majority will is central to democratic rule, therefore lamenters of low turnout often argue that the more completely the preferences of the majority are registered, the more democratic the system will be. When a government’s mandate is informed by incomplete information about the wishes of the electorate, the legitimacy of its decisions may be in doubt. (Hill, 2006: 209)

In sum, the argument from legitimacy and collective rationality is that if elections are to serve as a tool for ensuring policy accountability, all members of the community have an obligation to make their views known. Full participation is

M1546 - BIRCH TXT.qxp:ANDY Q7

48

27/10/08

11:31

Page 48

Full participation

in this sense necessary for an election to be legitimate from a democratic point of view; it ensures that the electoral process is equitable in two senses: the costs of voting are distributed equitably across all beneficiaries of that process, and one of the conditions is fulfilled for the outcome to be equitable with respect to preferences.12 Arguments against compulsory voting Jakee and Sun charge that ‘equating higher turnout with greater “legitimacy” (or even improved “democracy”) is overly simplistic’ (2006: 69–70; cf Lever, 2007: 37–8), and there are a number of reasons for believing this to be the case. Opponents of making electoral participation mandatory can point to the fact that legitimacy is both theoretical and practical, and that the appeal of this institution to some social choice theorists might well not be shared by the population at large; on the contrary, there are reasons to believe that compulsory voting would actually depress the legitimacy of the electoral process. It can be imagined that the introduction of a legal obligation to vote would meet with organised opposition, which could well play on anti-compulsion attitudes even among those generally sympathetic to the idea that citizens have a duty to exercise their franchise. If this were to occur, it might well serve to delegitimate the institution in the eyes of considerable numbers of citizens (Watson and Tami, 2004). If compulsory voting were an institution unwanted and resisted by the majority – or even a substantial minority – of the electorate, it would be difficult to argue that it was democratic, and it would not have legitimacy in the descriptive sense, whatever normative arguments could be made in its favour. Another common fear is that requiring voters to attend the polls would increase the number of votes cast on a random basis. Concerns about a surge in random or uninformed votes are a leitmotif in arguments against requiring citizens to vote. Richard Katz highlights the potential of electoral compulsion to lead to inferior quality decisions due to the inclusion of unwilling voters: ‘the quality of decision is unlikely to be enhanced by the random contributions of the ignorant and the uninterested’ (Katz, 1997: 245). Furthermore, Jakee and Sun (2006) provide a formal proof that random voting could alter electoral outcomes, and that the larger the number of random votes, the greater the likelihood that electoral results would resemble a coin-toss. Enhancement of the stochastic element of the electoral process would lead to a deterioration in the democratic validity of the results, and it would impair the ability of elections to provide an accurate tool for translating popular opinion into political outcomes. It could be countered that for genuinely indifferent or alienated voters, the option of voting for ‘none of the above’ or simply spoiling their ballots provides a means of expressing lack of interest or protest. But this points to a common problem often associated with mandatory electoral participation: it leads

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 49

Normative arguments for and against compulsory voting

49

to a higher number of invalid and blank ballots (noted as an argument by Gratschew, 2002: 106). Consequences Arguments about the consequences of compulsory voting for the political process fall into two principal categories: those pertaining to popular attitudes and behaviours, and those relating to broader political outcomes. Given that they involve somewhat different arguments, it makes sense to consider these categories separately. Public engagement In many democracies there has recently been concern with declining levels of public engagement in political life. Falling turnout rates are often seen as one symptom of this development. The question that concerns us here is whether enforced electoral participation would serve to enhance public engagement, or whether it would merely mask the problem. In other words, is compulsory voting effective in raising levels of citizen awareness, interest in politics and involvement in public life (including involvement in and identification with political parties)? Does it mobilise and educate? As we saw in Chapter 2, compulsory voting was introduced in Argentina to root out wilful nonparticipation and conscious disengagement from politics, but this is a relatively unusual situation. The type of disengagement that commonly confronts modern democracies is linked rather to the perception that politics has little bearing on people’s lives and is of marginal relevance to them. Certainly ‘participationists’ from John Stuart Mill in the nineteenth century to Carole Pateman in the 1970s and Benjamin Barber in the 1980s have touted the educative effects of political participation (Pateman, 1970; Barber, 1984; Mill, 1991 [1861]). The relevant issue is whether voting is sufficiently ‘participatory’ to have this sort of educative effect, and if so, whether such an effect would result if participation were the result of compulsion rather than voluntary involvement in the electoral process. Arguments in favour of compulsory voting Studies of regime support have found that voters tend to be those citizens who are more supportive than non-voters of the political systems in which they take part, but that this is at least in part due to the conditioning effects of voting itself (Ginsberg and Weissberg, 1978; Finkel, 1985). This suggests that the act of voting can shape attitudes toward the political process. It is therefore not unrealistic to suppose that involving people in elections would increase levels of awareness of and interest in politics, as well as involvement in political parties, perceptions of efficacy, and regime support (Pateman, 1970: 63). In

M1546 - BIRCH TXT.qxp:ANDY Q7

50

27/10/08

11:31

Page 50

Full participation

this sense it could be argued that compulsory voting both mobilises and educates. Consideration of the empirical evidence in support of these claims will be left to Chapter 4. The goal of the present discussion is to elaborate the logic that underlies these assertions. Some have argued that compulsory voting would spur the uninformed and uninterested to acquire information about politics, in order to make their vote choice (Lijphart, 1997: 4, 10; Matsler, 2003: 955; Lardy, 2004: 314–5). It has also been claimed that the institution of compulsory voting would foster a norm of voting (Hasen, 1996; Matsler, 2003: 955, 961; Hill, 2006: 218–19). If this were true, we might expect that the institution would have the knock-on effect of increasing levels of political engagement more generally. It is also worth noting that such arguments represent an implicit critique of the position put forward above, that uninformed and uninterested voting under a compulsory regime would degrade the quality of democratic decision-making. The counter-factual assumption of such arguments is that voter attitudes and inclinations are exogenous, whereas the proponents of compulsory voting argue that the institution would alter those attitudes and thereby serve to improve the overall quality of democratic engagement, in addition to whatever quantitative increases it might yield in turnout figures. Arguments against compulsory voting Two arguments are typically advanced to counter the above assertions. Firstly, compulsory voting will not solve the underlying problem of disaffection. What are needed are other and more meaningful forms of participation. Secondly, compelling people to vote is more likely to lead to disgust with politics and an antipathetic attitude than it is to mobilise and engage. Firstly, compulsory voting is held not to solve the true problem of low levels of political participation and public engagement, but merely to address the symptoms. If the real problem in a society with low turnout levels is voter apathy or disillusion with politics, there is little point in ‘artificially’ raising turnout levels by compelling people to vote, as this will remove a valuable means of gauging the health of democracy (Abraham, 1955: 8, 33; Ballinger, 2006: 20; 2007). Mark Franklin articulates this argument eloquently when he points out that: People vote in order to affect the outputs of government in ways that are meaningful to them. Low turnout thus reflects a paucity of choices or a lack of evident connection between electoral choice and policy change. Raising turnout by making it compulsory does not directly affect either of these critical variables but may mask their effects. [. . .] [B]y focusing on the wrong remedy we may be ‘shooting the messenger’ with consequences for our ability to judge the success of other, more appropriate, reforms. (Franklin, 1999: 206)

He goes on to argue that requiring full participation would actually have the effect of removing a useful measure of popular disaffection:

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 51

Normative arguments for and against compulsory voting

51

Raising the salience of an election should have the effect of increasing the level of turnout. Turnout would then be serving, as it is now so often seen to do, as an indicator of the health of a democracy. An increase in turnout following reform would tell us that the reform had worked – so long as the usefulness of the indicator had not been compromised through the introduction of compulsory voting. (Franklin, 1999: 216)

Secondly, compulsion is likely to alienate further disengaged citizens rather than to involve them cognitively. The very act of being made to take part in an act in which they have no interest will turn benign indifference into active disenchantment, thereby causing further alienation and cognitive aversion to politics. Abraham speculates that compulsory voting ‘might gall the voter into disgust with everything political’ (1955: 8). It is also commonly pointed out that compulsory voting is costly to administer (noted as an argument by Abraham, 1955: 16–17; Gratschew, 2002: 106). In this context it is maintained that there are other more effective, more popular, and less costly means of increasing turnout, such as selective incentives to vote, vote facilitation mechanisms, etc. By the same token, there may well be other more effective ways of engaging the public and increasing efficacy. Political outcomes Related to the notion of procedural legitimacy is that of substantive legitimacy, and the practical consequences of a situation in which the decisions taken by an elected body do not accord with the popular will. This final set of concerns relates to the ‘quality’ and benefit of the vote choice of those who would otherwise not be inclined to go to the polls. The early literature reflects a certain elitism in addressing this issue. For example, Giraud suggests that the choices of those who would not vote under a voluntary system represent a peril to the public good, in as much as they may distort political outcomes by giving equal weight to the vote decisions of those who are passionately interested and involved in politics and those who take no interest in the subject whatsoever: The obligation to vote appears to me to be of such nature as to distort the play of representative institutions by introducing into it an element of disorder and confusion. [. . .] On the whole, non-voters represent, from the political point of view, no value, and [. . .] their non-participation in politics, so long as the causes of their abstention remain, represents not a loss for the country but a salutary purge, a purge that has the characteristic of being spontaneous and which, for this reason, is the only purge possible in a democratic regime.13 (Giraud, 1931: 7)

Writing a quarter of a century later, Henry Abraham averred that ‘without compulsion a very desirable sifting of the electorate takes place, since only those who have a genuine interest and knowledge of public affairs usually take the time and trouble to go to the polls’, and further on: ‘little would be gained by a mass of uninformed at the ballot booth’ (1955: 2, 33).

M1546 - BIRCH TXT.qxp:ANDY Q7

52

27/10/08

11:31

Page 52

Full participation

This view reflects the notion that intense preferences should be given more weight than non-intense preferences and those bordering on indifference. It should be noted that views on this point tend to be strongly coloured by assumptions about the vote choices of those with weak preferences. This is clear in the Belgian debate, as detailed in the previous chapter; compulsory voting was initially supported in the nineteenth century as a ‘moderating’ mechanism; the assumption being that those who would abstain under a voluntary system were in all likelihood ideological ‘moderates’. This accords with the contemporary view that those with more intense preferences are also likely to hold more extreme views, or at least those with extreme views are more likely to hold those views intensely. If intensity of preference equates with ideological extremism, and indifference translates into moderation at the polls, the logical conclusion is that mandatory electoral participation is a useful means of stemming extremism and promoting centrist outcomes. The ‘unconsidered’ votes of habitual non-voters would at worst have a benign impact on political outcomes, in that they would mostly reinforce the position of the median ‘thinking’ voter. If, on the other hand, it is held that those compelled to vote against their will might cast votes for extremist parties, either out of protest or because they are easily swayed by populist arguments, then compulsory voting might ‘artificially’ boost the electoral strength of extremist parties, in which case it might be held to be undesirable. This is of course an empirical question, and it may well be that the choices of those less inclined to vote will vary by country, by social group within countries, and over time. This question will be addressed in Chapter 7.14 Whatever the truth of the matter, it is clear that there has been interest expressed by the scholarly community in the partisan impacts of differential participation. It is only one step from this assertion to the conclusion that the extent of participation will also be related to the policies enacted, and specifically to the distributional impacts of those policies. If politics is about ‘who gets what, when and how’, an alteration in the participating ‘who’ can be expected to have ramifications for what different groups in society ‘get’. It is argued by some that non-voting is not entirely voluntary, in as much as it frequently stems from social disenfranchisement and exclusion (Wertheimer, 1975: 289–91; Lardy, 2004: 316). It is also worth considering the possibility that compulsory voting might have implications for descriptive representation, in that full participation might generate representative bodies which more faithfully mirror the electorate in socio-demographic terms. This aspect of political outcomes does not feature prominently in the normative literature on compulsory voting, but it is a logical corollary to the argument about partisan representation.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 53

Normative arguments for and against compulsory voting

53

Arguments in favour of compulsory voting One of the strongest ‘outcome’ arguments in favour of compulsory voting is the social justice rationale: full participation leads to more just policy outcomes; it is therefore not just procedurally fairer, but also substantively fairer (Feeley, 1974: 240; Wertheimer, 1975: 293; Lijphart 1997). Arend Lijphart puts it succinctly when he states that ‘low voter turnout means unequal and socio-economically biased turnout’ (1997: 2). Cross-national empirical work has demonstrated a clear link between turnout levels and patterns of partisan support (Crewe, 1981; Pacek and Radcliffe, 1995). It is because of the socio-political skew of electoral non-participation in many states and the resultant skew in policy influence that commentators have defended mandatory electoral participation as the only effective means of guaranteeing an outcome that is democratic in a substantive sense, where the policies enacted reflect the view of the population as a whole (Lijphart, 1997; Hill, 2002a). Under voluntary participation, ‘the production of government policies will be inefficient in that they will be formulated to appeal to this subset of the population whose preferences will diverge from the aggregated preferences of the larger public’ (Feeley, 1974: 240). As Chong and Olivera point out, it is those who benefit most from voting (through the public goods available from elected governments) who are least likely to vote under voluntary regimes. Mandatory voting therefore ‘redefines ownership of the electoral process’ to the benefit of the economically and politically disadvantaged (Chong and Olivera, 2005: 5). Heather Lardy makes a similar point with reference to ‘freedom as nondomination’ derived from Republican theories of democracy: The right to vote represents a freedom which cannot be diminished by being compelled because any exercise of the right, however it is initiated, contributes to the practice of freedom as non-domination. By voting, a person registers her status as a political equal, a full member of the democratic community. She does this even if she votes reluctantly or apathetically, or because compelled to do so by law. On this account, non-voting brings with it a serious risk of domination by those classes which do not vote regularly, and by the governors whom they elect. To prevent this, voting may be compelled consistently with the idea of liberty as nondomination, and, therefore, with the idea of the right to vote. (Lardy, 2004: 314)

If rates of electoral participation are associated with distributive outcomes, parties have a strong incentive to design their policy programmes so as to reflect the views and interests of the voting population, rather than the entire population. Given that the poor and socially marginalised are typically less likely to vote than the educated and affluent, the result of low turnout tends to be policies that are biased to the advantage of the affluent. This creates a vicious circle, in which the socially marginalised perceive that parties do not address their

M1546 - BIRCH TXT.qxp:ANDY Q7

54

27/10/08

11:31

Page 54

Full participation

concerns, so they disengage from politics and stay at home on election day; knowing this, parties continue to pay mere lip service to the concerns of these citizens, which reinforces the latter’s perceptions. Arguments against compulsory voting It can be argued that many people are ill-informed about issues of public policy and will therefore not be likely to make informed, ‘policy directed’ choices. If this is the case, their arbitrary choices, or the votes they cast for extremist parties in protest at being made to vote, will not actually serve their interest, but will instead only have a negative impact on the quality of decision-making. Another take on the issue would query the utility of imposing on everyone a duty to vote, when in practical terms it is not necessary that everyone participate in order for democratic decisions to meet the criteria which would be reasonably required of them. In the words of W. H. Morris Jones: Participation and consent may be useful and desirable, but only as aids to a complete and adequate debate. In this case, the duty to vote appears in a different and less glowing light. All that is imperative for the health of parliamentary democracy is that the right to vote should be exercised to the extent necessary to ensure that the play of ideas and clash of interests can take place. If a symphony is scored for fifty instruments, there is little to be gained by trebling the number; massed bands are neither here nor there so far as the quality of the music is concerned. In a similar way, heavy polls are largely irrelevant to the healthy conduct of political business. (Morris Jones, 1954: 35)

The point that a larger symphony does not make for better music may be countered by the argument that it is not the size of the orchestra per se so much as its composition that matters, and that a balanced and complete political ‘symphony’ can only be achieved in practical terms by requiring full participation. Nevertheless, there is a valid point here, in that full participation is by no means necessary for socially representative participation. Provided adequate numbers of relevant sectors of society exercise their franchise voluntarily, politicians will have an interest in catering to their needs. In this sense it might be argued that to focus on the over-representation of the privileged in the electorate is to obscure the real source of the disproportionate influence of the rich, which could well be due to their economic power pure and simple, especially as it translates into campaign donations, the funding of lobbies and think-tanks, and greater access to the corridors of power. If this is the source of the main political benefit of economic advantage, then it might be argued that the efforts of electoral engineers would be better spent on reforming campaign finance laws and other institutional mechanisms that could be used to curb the benefits of social privilege, than focusing on the relatively marginal impacts of differential turnout.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 55

Normative arguments for and against compulsory voting

55

Full assessment of the competing claims about consequences of mandatory electoral participation must await empirical analysis. But several points can nevertheless be made at this stage about the assumptions that underlie them. The notion that voluntary voting provides us with a convenient barometer of public disaffection (in the form of turnout percentages) requires comment. There are many ways of measuring public disaffection, from surveys and focus groups to levels of popular participation in a range of political activities. The act of voting is arguably too important to be employed as a mere measurement device. Moreover, the legal obligation to participate in elections does not deprive the elector of opportunities to express dissatisfaction with the political process. The proportion of spoilt ballots is regularly used in states with mandatory electoral participation as a means of assessing levels of popular disaffection. A none-of-the-above option on the ballot could be made to serve a similar function (Wertheimer, 1975: 279, 293). The advantage of relying on these measures is that they are far more accurate. We know that people fail to turn out to vote for a wide variety of reasons, not all of which have to do with lack of political engagement; turnout thus provides a very crude means of assessing levels of disengagement and disaffection (Lardy, 2004: 318). Against the argument that compulsory voting does not address the underlying causes of political malaise, it might therefore be pointed out that the institution is not designed to do this, but what it does do is to allow for a clearer articulation of that malaise by encouraging people to express it directly, rather from refraining from any expression at all. As Wertheimer argues, compulsory voting would provide more accurate information on true citizen preferences than the current system, where non-voting is difficult to interpret: ‘Does nonvoting indicate contentment or alienation? Under compulsory voting (assuming one could cast a blank ballot) we would have a precise count of those citizens who refuse to support any of the available candidates. In addition, those who abstain under the present system because they prefer a minor party or candidate who has no chance of winning are more likely to vote’ (1975: 293). It is noteworthy in this context that Mark Franklin has recently voiced doubts about his own frequently-cited view. In a 2001 review of voter participation in Canada, he is quoted as follows: Some years ago, I argued that compulsory voting would be a bit like shooting the messenger. The level of turnout tells us something about the health of democracy, and if we artificially raise turnout by means of compulsory voting then we no longer have a signal to tell us that something is wrong with our procedures. But I am not so sure about that any more. If we had compulsory voting in America, politicians would have an incentive to come forward with policies that simply are not dreamt about now, because the people who would vote for them

M1546 - BIRCH TXT.qxp:ANDY Q7

56

27/10/08

11:31

Page 56

Full participation

aren’t voting. If you forced those people to vote, then you would get policy entrepreneurs making proposals that currently are not on the agenda at all. Maybe political scientists ought to stand up and say it would be good for democracy to have compulsory voting. Maybe that would be a good role for us to play. (CRIC, 2001: 35)

If compulsory voting does achieve these ends, this does not mean that this institution is incompatible with the provision of vote-facilitation mechanisms. Indeed, states that make electoral participation mandatory typically have a wide range of measures in place to ensure that voting is not an onerous chore. Other forms of participation do not preclude mandatory electoral participation. At the same time, if compulsory voting can be shown empirically to lead to further alienation, resistance and disaffection, then this suggests that it would likely cause more harm than good. Empirical evidence is also needed to elucidate a number of the substantive outcome claims made about compulsory voting. It is unreasonable to expect parties to act other than in their electoral interest. The genius of a democratic political system is that it is set up in such a way that parties have an incentive to serve the social needs of the electorate, because that is the way they win power. It may be that the only means by which the vicious circle of social disenfranchisement can be overcome is by requiring all citizens to go to the polls. On the other hand, if uninformed voting by reluctant electors means that their needs are not met through full participation, then this institution may well be deemed undesirable in social terms. Conclusion As can be seen from the above arguments, debates about compulsory voting reach to the core of many central concerns of contemporary political theory, including the nature of rights, the nature of democracy, and the possible tension between these two key concepts. They also address the central concern of the legitimacy of political decision-making, and the capacity of individuals to take part therein. These competing claims will be weighed up in the Conclusion to this volume, which will seek to undertake an overall assessment of the institution of compulsory voting. Though many of these arguments are couched in normative terms, the discussion in this chapter has demonstrated that they rest on a number of empirical assertions that are often not supported by evidence. From the point of view of this study, many of the assertions made in this chapter can thus be thought of as hypotheses to be tested, and the remainder of the book will be devoted to this task. In drawing this chapter to a close, it is worth while setting out these hypotheses systematically. Claims have been made that the institution of mandatory electoral participation:

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 57

Normative arguments for and against compulsory voting

57

• increases levels of political knowledge among the citizenry; • either enhances or undermines citizen engagement with politics; • either enhances or undermines the legitimacy of the electoral process, and democracy more generally, in the eyes of the citizenry; • increases turnout; • increases the number of random and invalid votes cast; • influences the partisan distribution of support; • influences the composition of elected bodies in descriptive terms; • leads (or fails to lead) to substantively fairer social and political outcomes. In addition, the historical discussion in Chapter 2 indicated that the impact of compulsory voting on electoral integrity was an important concern for many political actors at the time of its adoption. Though considerations of the potential of this institution to generate fraud and electoral abuse have not been prominent in recent debates, this aspect of compulsory voting is also worth considering, given its relevance to historical discussions. We turn in Chapter 4 to the first set of empirical questions, namely those revolving around questions of political information, engagement and perceptions of regime legitimacy. Notes 1 This literature should be distinguished from theory in the so-called ‘politics of compulsion’ mould, which views compulsion, coercion and domination as real but unfortunate aspects of political life (see Stears, 2007). 2 The obligation view is reflected in the use-it-or-lose-it approach to compelling citizens to vote; see Chapter 2. 3 ‘La qualité d’électeur est une fonction publique, dont la société a le droit de déterminer la délégation’ [author’s translation from the French]. 4 ‘La fonction d’électeur n’est pas un droit; elle existe pour la société qui a le droit d’en déterminer les conditions’ [author’s translation from the French]. 5 ‘On ne vote pas pour soi-même; le vote s’exerce dans l’intérêt social’ [author’s translation from the French]. 6 ‘La liberté n’est défendue que par ceux qui l’aiment, par ceux qui apprécient ses garanties et acceptent les devoirs qu’elle impose; le droit de suffrage est parmi les garanties et les devoirs de la liberté. [. . .] Chacun appartient à une société politique, et en reçoit, contre son gré peut-être, des bienfaits; it doit en subir les charges. La démocratie confère des avantages et impose des devoirs; les uns et les autres forment un tout indivisible’ [author’s translation from the French]. 7 According to the UK Office of National Statistics, for example, the response rate for the 2001 census was 94 per cent, though only a ‘limited’ number of people have been prosecuted for failing to take part. See http://www.statistics.gov.uk/census2001/. 8 The parallel they make is the following: ‘A law requiring two witnesses for a will to be valid is not shown to be undesirable by a proof that there is no antecedent moral

M1546 - BIRCH TXT.qxp:ANDY Q7

58

9

10

11 12

13

14

27/10/08

11:31

Page 58

Full participation duty in the state of nature to have one’s testament doubly witnessed’ (2000: 86f). Though the authors do not elaborate on this parallel, the implication is that there may well be good pragmatic reasons for adopting compulsory voting. ‘Pour que la loi soit l’expression de la volonté nationale, il faut que cette volonté s’exprime, se manifeste, et évidemment, elle ne l’est pas, si les électeurs negligent de se rendre au scrutin. Non seulement la volonté nationale n’est pas connue, mais de trop nombreuses abstentions peuvent en fausser l’expression en déplaçant la majorité’ [author’s translation from the French]. ‘S’il importe que les chambres soient l’expression adéquate et l’organe autorisé du pays légal, comment nierait-on qu’il est d’utilté générale que la fonction électorale soit réellement exercée par ceux à qui elle est confiée?’ [author’s translation from the French]. It must, however, be borne in mind that social choice theorists have not always come down in favour of making voting mandatory. It is worth noting in passing that this argument suggests that participation ought properly to be thought of as a commitment problem rather than, as is conventionally the case, as a cost-benefit exercise. ‘L’obligation du vote nous paraît de nature à fausser le jeu des institutions représentatives en introduisant dans celui-ci un element de désordre et de confusion. [. . .] les abstentionnistes représentent, dans leur ensemble, au point de vue politique, une valeur nulle, et [. . .] leur non-participation à la politique, tant qu’existent les causes qui commandent leur abstention, représente non point une perte pour le pays mais une élimination salutaire, élimination qui a le caractère d’être spontanée et qui, de ce fait, est la seule possible en régime démocratique’ [author’s translation from the French]. It might, however be argued that such speculations ignore the educative and interest-sparking effects of mandatory voting laws; when people are obliged to vote, they may well become more interested in politics and begin to make informed choices. This is also an empirical question to be addressed in Chapter 4.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 59

4

Compulsory voting and election campaigns

Now that the theoretical and conceptual groundwork has been laid, this chapter will be the first of four to assess empirically the impact of compulsory voting on the electoral process. Logically and temporally, the electoral campaign is the first relevant point at which to begin such an investigation. ‘Campaign’ will be understood here both in the specific sense of the period immediately prior to an election in which parties and candidates actively seek the support of the electorate, but also the ‘permanent campaign’ that has developed in many democracies – in other words the regular messages that parties send to voters and the contacts they make with them in order to increase their chances of electoral success. In this sense the main focus of this chapter is how the institution of mandatory electoral participation conditions the relationship between political parties and citizens, and how the nature of this relationship in turn defines the tenor of electoral campaigns understood in the strict sense. The main hypothesis to be tested here is that, following the analysis laid out in Chapter 3, compulsory voting enhances citizen involvement in politics and tightens the links between parties and voters. Evidence to assess this claim will be drawn from the findings of previous studies, as well as the results of crossnational analyses of election survey data. The results of these analyses suggest that in some respects compulsory voting does enhance citizen involvement and engagement in the electoral process, but that in other respects the reverse is true, and that there is no clear link between this institution and most aspects of campaign behaviour. Mobilisation, conversion and electoral campaigns The most obvious impact of compulsory voting on campaign activities is the fact that institutionally-enforced universal participation shifts the main aim of political parties from mobilisation to conversion. An extensive body of literature has documented the extent to which political parties focus on mobilisation in elections held under voluntary voting regimes. The question that interests us here is how campaign styles differ under compulsory voting. Mark Franklin

M1546 - BIRCH TXT.qxp:ANDY Q7

60

27/10/08

11:31

Page 60

Full participation

notes that ‘compulsory voting may be viewed as a mobilizing tactic’ (1999: 207). With an institution in place to do the ‘work’ of mobilisation, parties are then able to focus on converting voters to their cause. If parties are freed from the task of getting voters to the polls and can concentrate their efforts instead on winning their hearts and minds, this ought in theory to have several consequences for campaign styles. The specific objective of the analyses that follow is to assess the impact of compulsory voting on electoral competition, in particular the impact of this institution on citizen-party linkage and citizen involvement in electoral campaigns. Electoral campaigns are complex and many-facetted events, and it is not possible in an investigation of this scope to do justice to all the possible ways in which the institution of compulsory voting might potentially intervene to structure or otherwise influence all the activities included in this rubric. The choice of empirical analyses undertaken here will perforce be constrained by the availability of comparable comparative data, as well as by the relative prominence of the normative arguments put forward in the previous chapter. It is nevertheless worth pausing to consider some of the ‘campaign effects’ that compulsory voting has been alleged to have. There is limited anecdotal evidence in the literature on compulsory voting of this institution’s impact on the degree of excitement and activity surrounding electoral campaigns. For example, Gosnell (1930: 120) attributes the orderly campaign style of early twentieth century Belgian elections in part to the existence of mandatory electoral participation: Compulsory voting and proportional representation have insured the voters that a full vote and a fair count are to be had by all. It is not necessary for the parties to have huge demonstrations. To be sure, there are public meetings and parades, but these are ordered and dignified. Propaganda is spread by the newspapers and by the use of posters and circulars. The Socialists may parade through the streets following a band and distributing leaflets, but they do not have clashes with their opponents as do the Socialists in Germany. The newspapers appeal to prejudices during the campaign as in other countries but the campaigns are conducted in a quiet manner. [. . .] The high standard of interest in elections which is maintained in Belgium is the result of the keenness of the economic struggle, the intensity of the party battles, the convenience of voting arrangements, and the success of compulsory voting and proportional representation.

The ‘orderly’ nature of Belgian electoral campaigns can also be viewed in a more negative light, however. At the time of debate over the maintenance of mandatory voting rules in Belgium in the 1990s the opinion was voiced that these rules made campaigns dreary and encouraged apathy, because parties had no incentive to try to mobilise their supporters (Pilet, 2005: 8). Other commentators have claimed that compulsory voting has a positive effect on campaigning styles. Jackman (2001: 16317) maintains that obliging

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 61

Compulsory voting and election compaigns

61

citizens to vote leads to higher ‘quality’ campaigns. Likewise, Lijphart (1997) notes that the incentives to use ‘attack’ advertising diminish when mobilisation (and demobilisation) are no longer parties’ primary goals. Finally, as discussed in Chapter 2, the perceived reduction of campaign spending under compulsory voting was one of the main motives for its adoption in a number of states. Perceptual and measurement considerations make it difficult to undertake cross-national analyses of the extent to which election campaigns in different countries are ‘orderly’ or ‘dreary’. At the same time, lack of global data on campaign advertising and spending hamper our ability to assess the hypothesised influence of compulsory voting on these aspects of campaigns. There are, however, a number of commonly-made claims about citizen involvement and attitudes toward parties and campaigns that can be tested on the basis of existing data. These topics will be the main focus of this chapter, though it should be borne in mind that there are many additional aspects of electoral campaigns that might profitably be assessed in connection with the institution of mandatory electoral participation. As suggested by the normative arguments detailed in Chapter 3, compulsory voting is commonly held to have consequences for several aspects of electoral competition: citizen political education, citizen political involvement, party identification, and party engagement of the citizenry. The reasoning behind each of these claims requires some elaboration. Political knowledge As has often been claimed by normative theorists and other advocates of compulsory voting, this institution can be expected to lead to higher levels of political knowledge (Lijphart, 1997: 10).1 There are three principal reasons for this. Firstly, if conversion replaces mobilisation as the main aim of parties during a campaign, they have an incentive to focus on policy rather than on ‘hype’, which should in theory lead to an electorate better-informed about issues and policies. Secondly, if voters know in advance that they will be voting on election day and that they will have to choose from among the options on offer, they have greater reason to pay attention to the campaign than is the case when they can toy with the idea of not voting only to change their minds shortly before the election. The third reason is related to the impact of the universality of electoral campaigns under compulsory voting; if the assumption can be made that everyone will be voting, the election is more likely to become a topic of conversation among friends, relatives and colleagues, which should serve to inform people of relevant issues and the policy positions of parties. Yet it may also be that compulsory voting fails to raise levels of political knowledge. Those who vote due to compulsion may have little incentive to inform themselves, and some may actively resist the acquisition of relevant information as a form of protest. As we saw in the previous chapter, this is one

M1546 - BIRCH TXT.qxp:ANDY Q7

62

27/10/08

11:31

Page 62

Full participation

of the main objections that opponents of compulsory voting have to the institution: that it brings uninformed voters to the polls, and that voting on a random basis by the uninformed can, under certain circumstances, have a substantive impact on the outcome which could well be construed as being undemocratic (Jackee and Sun, 2006). The empirical evidence of the impact of mandatory electoral participation on political knowledge is far from conclusive. There is some support from Belgian and Australian data for the theory that those in compulsory voting systems who would otherwise not vote are less well informed than those who would vote under a voluntary regime (Mackerras and McAlister, 1999; McAllister, 2001; Selb and Lachat, 2007). Yet there is also evidence to support the claim that compulsory voting increases levels of political knowledge. Gordon and Segura (1997) do not study levels of political knowledge directly, but they do find in their comparative study of European states that compulsory voting has a positive impact on the related concept of political sophistication, operationalised in this study in terms of the accuracy of citizen placements of political parties on the ideological spectrum (cf Berggren, 2001). David Brockington (2005) undertakes a direct test of the impact of compulsory voting on political information, using survey data from the Comparative Study of Electoral Systems (Module 1), and finds that compulsory voting has a positive impact on general levels of political information, but a negative impact on campaign-specific knowledge. It should be noted, however, that the latter is operationalised in terms of recall of candidate names, which may well not be highly salient information in many proportional representation systems. Not all studies have found mandatory electoral attendance to have a positive effect on political information levels. Using a dataset also derived from Module 1 of the CSES, Grönlund and Milner identify Belgium as a state with unusually low levels of ‘civic literacy’, a finding they interpret in terms of compulsory voting ‘leaving the less educated less motivated to acquire information, merely casting a vote at elections’ (2006: 397). Likewise, experimental counter-factual evidence from the voluntary voting state of Canada finds that the provision of material incentives to vote has no significant impact on levels of political attentiveness or interest among young voters (Loewen et al., 2007). It is not entirely clear that the incentives provided under experimental conditions are an adequate proxy for the legal obligation to participate in elections, but the results of this study do give pause for thought. Political engagement We might expect the universality of voting under a compulsory regime also to have implications for the degree to which citizens are engaged with electoral campaigns. If everyone in theory votes, then everyone is in theory involved at some level. The question of interest in this context is the extent to which

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 63

Compulsory voting and election compaigns

63

voting – including the anticipation of voting at an election and recollections of voting experiences in the past – has an impact on the degree to which citizens engage themselves psychologically and behaviourally in a campaign. Interest in politics is the most obvious form of psychological engagement. On the basis of the considerations voiced above, one would expect higher levels of political interest in countries practicing compulsory voting than in those where voting is voluntary (all else being equal). Political efficacy is a closely related aspect of psychological involvement that has also been hypothesised to be linked to compulsory voting. Lisa Hill views this institution as a means of breaking the cycle of low efficacy and non-participation that can form in voluntary voting systems (2006: 216). We would expect these higher levels of interest and efficacy to translate also into greater political participation in electoral campaigns, including political discussion, as well as direct contact in and involvement with campaign-related activities. This is because, one would anticipate that a larger proportion of the mass public would engage in other forms of political participation if a larger number were voting. Alternatively, it might be that voting is for some people a functional substitute for other forms of political engagement, such as protest politics (Dalton, 1996: 64; Norris, 2002: 205). Finally, the case can also be made that obliging people to vote may alienate the unengaged even further, and that for this reason the introduction of compulsory voting may be counter-productive. There is some evidence from previous studies that compulsory voting does indeed have an impact on levels of political interest. Comparing Dutch survey data from 1967, shortly before the repeal of compulsory voting legislation, and 1970, immediately after this change, Galen Irwin finds that those less interested in politics were substantially more likely to cease voting when voting was made voluntary than those with higher levels of political interest (1974: 299; cf Irwin and van Holsteyn, 2005: 39–40). He finds a similar shift in relation to efficacy, with turnout falling considerably more among the politically alienated than among others in society (1974: 301).2 There is reason to doubt that this relationship holds in other states, however. Engelen and Hooghe (2007) argue that political interest and voting should in theory be mutually reinforcing, echoing the findings of previous research (Finkel, 1985). Yet their analysis of political interest and efficacy finds no evidence in the European sample they employ that the institution of compulsory voting has any clear impact on either of these variables. Bilodeau and Blais find in their study of World Values Survey data that levels of interest in, and discussion of, politics are actually lower in countries where voting is mandatory (Bilodeau and Blais, 2005: 6). An analysis of political attitudes and behaviours among immigrants in New Zealand in the same study finds that, with the exception of those very recently arrived, immigrants from Australia have levels of political interest not significantly higher than immigrants from states where

M1546 - BIRCH TXT.qxp:ANDY Q7

64

27/10/08

11:31

Page 64

Full participation

voting is not compulsory, which suggests little or no socialisation effect for compulsory voting (Bilodeau and Blais, 2005: 7). Nor do they find that immigrants to Australia from other countries where voting is compulsory exhibit levels of political interest higher than those who have come from voluntary voting states (Bilodeau and Blais, 2005: 8). The impact of compulsory voting on other aspects of political engagement remains to be investigated. Party identification There are good reasons to believe that electoral institutions condition the strength of party identification (Bowler et al., 1994; Karp et al., 2008), and mandatory participation laws can be expected to have a measurable impact on the extent to which electors identify with political parties. It is well-known that Australia has exceptionally high levels of party identification, a fact that has been attributed to the institution of compulsory voting (Mackerras and McAllister, 1999: 229). The reasoning behind this claim is relatively straightforward: voting (and related forms of political involvement) reinforce identification with the party one habitually supports in an election. In Mackerras and McAllister’s words: ‘Compulsory voting ensures that voters cast a ballot and the act of voting means that they are forced to think, however superficially, about the major parties’ (1999: 229). The same authors note weaker party identification among those who say they would not vote under a voluntary regime than among those who say they would carry on going to the polls (1999: 229– 30). Similarly, Irwin found in his Dutch study that those with party identification were less likely to cease voting after compulsory voting was removed; he concludes that ‘having a party to support is itself a strong motivation to turn out at an election, and hardly needs the reinforcement of legislation. On the other hand, when compulsion was removed, one out of three non-identifiers dropped from the voting ranks’ (1974: 302). Irwin notes that strengh of party identifcation is also related to propensity to cease voting, with stronger identifers being more likely to continue attending the polls (1974: 307). Finally, Krister Lundell tests this hypothesis in the European setting and finds that compulsory voting has a positive impact on strength of partisanship (2007: 13). Previous research thus strongly suggests there is a positive relationship between the legal obligation to vote and party identification. Party engagement with citizens Party campaigning and engagement with individual members of the public can be expected to be affected by mandatory voting in several ways. It might be anticipated that campaigns conducted under compulsory voting rules are likely to be ‘dreary’ or lacklustre, due to the fact that the mobilisational incentive which drives parties under a voluntary regime is absent. Another

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 65

Compulsory voting and election compaigns

65

consideration relates to the cost-benefit analyses that parties invariably engage in when designing their campaigns and planning the types of activities to undertake. In compulsory voting states, the larger number of citizens requiring parties’ attention at election time might reduce their capacity to reach out directly to individual members of the public. High levels of electoral participation may therefore provide parties with an incentive to engage in selling themselves through ‘wholesale’ methods such as the mass media, rather than through face-to-face or other forms of individual contact. Indeed, as detailed in Chapter 2, it was the anticipation of an increase in particularistic rewards direct to individuals that led Thailand to make electoral participation mandatory in1997, in the hopes that this would reduce electoral abuse. Though it is not clear that this effort was successful in the Thai case (see Chapter 6) a similar logic applies to contact that involves persuasion rather than reward. More voters means more work, and the resulting strain on party capacity might militate in favour of other forms of campaigning. In a seven-country, study of the correlates of face-to-face party contacts, Jeffrey Karp and colleagues find that party contacts are generally higher in countries with candidate-based electoral systems than in those with party-list systems. Australia is an exception to this rule, however. Though Australia has candidate-based electoral systems for both its chambers of parliament, it has the lowest level of party contacts of any country in the study. The authors attribute this directly to the institution of compulsory voting (Karp et al., 2008). It is difficult to compare levels of excitement in electoral campaigns, but party contacting is something that can be measured and therefore compared. The expectation is that parties should engage in less face-to-face contacting under compulsory voting, because the motivational incentive is lacking. Empirical analysis The hypotheses elaborated above will be tested here through analysis of data from Module 2 of the Comparative Study of Electoral Systems (CSES) survey project. CSES Module 2 includes elections in 38 states, six of which (Australia, Belgium, Brazil Chile, Mexico and Peru) practice compulsory voting.3 In four of the six cases (Australia, Belgium, Chile and Peru), sanctions for non-voting are applied. This makes it possible to test for the effects of both compulsory voting per se and sanctioned compulsory voting. Details of data sources and variable construction can be found in the Appendix. Political knowledge Political knowledge is operationalised here in terms of three relevant questions asked in the CSES module. For ease of interpretation, answers to these questions are employed to develop a binary indicator contrasting those with

M1546 - BIRCH TXT.qxp:ANDY Q7

66

27/10/08

11:31

Page 66

Full participation

high knowledge and those with low knowledge (see the Appendix for details). Following previous authors, controls include individual-level factors such as the standard demographic variables – age, gender and education level – together with party identification and efficacy; at the macro-level, factors found to be relevant to political information are the effective number of parliamentary parties and the electoral system.4 An additional dummy variable designating Latin America was also included in the analysis, given the fact that four of the six compulsory voting states in this dataset are located in this region. Bivariate analysis at the macro-level indicates that states with mandatory voting have an average of 61.7 per cent of respondents with high levels of political knowledge, as opposed to 58.6 per cent in voluntary voting states. In states with sanctioned compulsory voting the difference is more pronounced; an average of 65.8 per cent have high levels of political knowledge, as opposed to 58.2 per cent in other states. Neither of these differences is statistically significant even at the 0.1 level, however. Multivariate analysis of the impact of compulsory voting on political knowledge is useful in providing a more fine-grained understanding of these relationships. Given the multilevel structure of this dataset, multilevel modelling techniques are employed to analyse the data.5 Table 4.1 presents the results of this analysis. As found in previous studies, older, better educated respondents and men generally have higher levels of political knowledge than others, as do respondents living in states with proportional electoral systems and more compact party systems. But while the coefficient for compulsory voting is positively signed in this model (and larger than its standard error) it is far from reaching statistical significance in Model 1. The same is true for sanctioned compulsory voting in Model 2. We can thus conclude that in the sample of elections analysed here, compulsory voting does not appear to be strongly related to overall levels of political knowledge, though there does seem to be a very weak positive association. Political engagement As conceptualised above, political engagement includes various cognitive and behavioural aspects of citizen involvement in the electoral process. These include political interest, efficacy, and electorally-relevant activity. Unfortunately, Module 2 of the CSES does not include an item that taps political interest directly. It does, however, include items on external efficacy and political discussion, as well as a trio of participation questions asking whether respondents have contacted politicians, participated in a protest or demonstration or worked with others who shared their concerns (see the Appendix for details). The control variables included in the multivariate models presented here are those most commonly found by previous authors to predict aspects of political

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 67

Compulsory voting and election compaigns

67

Table 4.1 Multilevel logit models of political knowledge Variable Individual-level variables: Age Education Female gender Party identification Efficacy Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant Intercept variance Chi squared – fixed effects Chi squared – random effects N: individuals N: elections

Model 1: CV

Model 2: Sanctioned CV

1.53*** (0.24) 2.83*** (0.21) ⫺0.67*** (0.04) 0.48*** (0.06) 0.08*** (0.01)

1.53*** (0.24) 2.83*** (0.21) ⫺0.67*** (0.04) 0.48*** (0.06) 0.08*** (0.01)

0.31 (0.27) 0.93*** (0.37) ⫺0.28*** (0.11) 0.49 (0.44) ⫺1.86*** (0.42)

0.17 (0.37) 0.89*** (0.35) ⫺0.26*** (0.10) 0.70 (0.39) ⫺1.89*** (0.42)

0.79*** (0.25) 605.69 (10 df) 10.19 (1df)

0.79*** (0.25) 624.57 (10 df) 10.43 (1df)

41,587 30

41,587 30

* p < .10; ** p < .05; *** p < .01; Cell entries are logit coefficients (standard errors).

engagement (Dalton, 1996: chaps. 3 & 4; Finkel, 1987; Bowler and Donovan, 2002; Norris, 2002; Bilodeau and Blais, 2005; Karp et al., 2008). These include, at the individual level, age, gender, education, and party identification. Macro-level controls include electoral system type, the effective number of parliamentary parties, and Latin American location. Starting with levels of external efficacy (operationalised here in terms of a question asking whether who people vote for makes a difference; see the Appendix for details), bivariate analysis reveals that the on average 63.9 per cent of respondents in states with compulsory voting have high levels of efficacy, as against 65.8 per cent in voluntary voting states. The same pattern manifests itself when we turn to sanctioned compulsory voting: 61.1 per cent of respondents in such states have high levels of efficacy, versus 66.0 per cent in other states. Thus, contrary to expectations, compulsory voting is associated with depressed efficacy levels (though the bivariate relationships in question fail to reach statistical significance). The multivariate models presented in table 4.2 confirm this finding; while the existence of compulsory voting alone has no significant impact on efficacy, sanctioned compulsory voting is negatively associated with this variable, once other relevant factors are taken into consideration. This relationship is only

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

68

11:31

Page 68

Full participation Table 4.2 Multilevel logit models of efficacy

Variable

Model 3: CV

Model 4: Sanctioned CV

Individual-level variables: Age Education Female gender Party identification

0.22** (0.10) 0.67*** (0.09) ⫺0.02 (0.02) 0.72*** (0.05)

0.22** (0.10) 0.62*** (0.09) ⫺0.02 (0.02) 0.72*** (0.05)

Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant Intercept variance Chi squared – fixed effects Chi squared – random effects N: individuals N: elections

⫺0.37 (0.29) 0.25* (0.15) 0.02 (0.05) 0.53* (0.31) ⫺0.56*** (0.20)

⫺0.34* (0.20) 0.31** (0.16) 0.00 (0.05) 0.34* (0.18) ⫺0.51*** (0.20)

0.16*** (0.04) 451.32 (9 df) 22.18 (1df)

0.16*** (0.04) 460.16 (9 df) 20.33 (1df)

48,804 33

48,804 33

* p < .10; ** p < .05; *** p < .01; Cell entries are logit coefficients (standard errors).

significant at the .1 level, but that small number of states included in this analysis warrants giving serious consideration to country-level variables with coefficients significant at this cut-off point. This result can be interpreted in several ways; it may be that an increase in the number of reluctant voters in states with mandatory voting laws heightens democratic disaffection, as the opponents of such laws argue should be the case. Alternatively, competition among parties may be dampened in compulsory voting states due to the fact that parties are not obliged to mobilise voters to go to the polls. This may engender more desultory electoral campaigns that fail to stir up the same amount of political passion as in states where parties have an incentive to do everything in their power to get out their vote. A final potential explanation relates to Lomasky and Brennan’s view, citied in Chapter 3, that ‘belief in a duty to vote is the opiate of the democratic masses’ (2000: 86), and the illusion of an unfulfilled potential efficacy this might breed in voluntary voting states (Wertheimer, 1975: 290). When voting is voluntary, nonvoters may think that their vote could make a difference were they inclined to exercise it; at the same time, voters might attribute that lack of a perceived ‘difference’ following an election to non-voting on the part of a substantial proportion of the population. But when everyone votes, these arguments are no

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 69

Compulsory voting and election compaigns

69

Table 4.3 Multilevel logit models of political conversation and persuasion Variable Individual-level variables: Age Education Female gender Party identification Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant Intercept variance Chi squared – fixed effects Chi squared – random effects N: individuals N: elections

Model 5: CV

Model 6: Sanctioned CV

⫺0.48*** (0.194) 0.88*** (0.103) ⫺0.32*** (0.040) 1.11*** (0.063)

⫺0.48*** (0.10) 0.88*** (0.10) ⫺0.32*** (0.04) 1.11*** (0.06)

⫺0.22 (0.33) ⫺1.11*** (0.43) 0.18** (0.10) 0.54 (0.59) ⫺2.40*** (0.31)

0.33 (0.41) ⫺1.11*** (0.42) 0.18* (0.09) 0.20 (0.61) ⫺2.42*** (0.34)

0.71** (0.30 676.64 (9 df) 5.71 (1df)

0.71** (0.30) 707.68 (9 df) 5.50 (1df)

51,483 34

51,483 34

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

longer tenable, and voters are forced to face up to the bare reality of what their collective decision has achieved in substantive terms. The situation with regard to political discussion and persuasion is less clear. The bivariate relationship reveals little difference between compulsory voting states and others: 23.3 per cent of survey respondents in the former report having talked to other people to persuade them to vote for a particular party or candidate, as against 22.2 per cent in the latter. In states with sanctioned compulsory voting, the figure is 24.0 per cent, as opposed to 22.1 per cent in all other states. These slight differences translate into insignificant coefficients in the multivariate models presented in table 4.3. Though the coefficient in Model 5 is negative, it is smaller than its standard error, which suggests that we should not attribute significance to this finding. The coefficient for sanctioned compulsory voting is positive, but again smaller than its standard error. It therefore seems that – at least in the sample of countries studied here – compulsory voting has no clear impact on people’s propensity to engage in political discussion and debate at election time. We next assess the impact of legally obliging electoral attendance on three aspects of political participation. The first is voter-initiated contact with a politician or government official. In states with compulsory voting, 13.4 per cent of

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

70

11:31

Page 70

Full participation Table 4.4 Multilevel logit models of contacting politicians

Variable Individual-level variables: Age Education Female gender Party identification Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant Intercept variance Chi squared – fixed effects Chi squared – random effects N: individuals N: elections

Model 7: CV

Model 8: Sanctioned CV

1.10*** (0.14) 1.66*** (0.17) ⫺0.29*** (0.05) 0.67*** (0.06)

1.11*** (0.14) 1.66*** (0.17) ⫺0.29*** (0.05) 0.67*** (0.06)

0.30 (0.32) ⫺0.72*** (0.26) 0.11** (0.06) ⫺0.08 (0.32) ⫺4.01*** (0.29)

0.30 (0.23) ⫺0.76*** (0.25) 0.13** (0.06) 0.07 (0.17) ⫺4.05*** (0.30)

0.36*** (0.09) 18.28 (9 df) 841.27 (1df)

0.36*** (0.09) 17.27 (9 df) 810.72 (1df)

50,060 34

50,060 34

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

those surveyed reported having engaged in this activity, whereas in voluntary voting states the figure was, at 13.9 per cent, almost identical. In states with sanctioned compulsory voting, the corresponding figure was a slightly higher 14.5 per cent, as against 13.7 per cent in all other countries. These differences again fail to manifest themselves as statistically significant coefficients in multivariate analysis. As shown in the models presented in table 4.4, the coefficients for both compulsory voting and sanctioned compulsory voting are positive, but they are not significant (the coefficient for sanctioned compulsory voting comes closest, being significant at 0.2). The situation is somewhat different when it comes to the potentially more demanding form of participation of attending a protest or demonstration. Though the bivariate relationships do not suggest much of an impact – 11.5 per cent in mandatory voting states reported having participated in a protest or demonstration versus 11.3 per cent in those where voting is voluntary, while 13.3 per cent in states with sanctioned compulsory voting protested or demonstrated as against 11.1 in other states – multivariate analysis reveals a strong positive relationship between the legal obligation to vote and this form of political participation. When other factors are controlled, respondents in compulsory voting states were significantly more likely to protest or demonstrate than

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 71

Compulsory voting and election compaigns

71

Table 4.5 Multilevel logit models of participation in protests and demonstrations Variable Individual-level variables: Age Education Female gender Party identification

Model 9: CV

Model 10: Sanctioned CV

⫺1.60*** (0.24) 1.56*** (0.14) ⫺0.15*** (0.06) 0.87*** (0.09)

⫺1.60*** (0.24) 1.56*** (0.14) ⫺0.15*** (0.06) 0.87*** (0.09)

Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant

0.15 (0.38) 0.01 (0.08) ⫺0.17 (0.21) ⫺3.45*** (0.41)

0.52*** (0.11) 0.08 (0.36) 0.04 (0.08) 0.06 (0.11) ⫺3.51*** (0.42)

Intercept variance Chi squared – fixed effects Chi squared – random effects

0.60*** (0.19) 16217.80 (9 df) 9.60 (1df)

0.59*** (0.20) 2484.81 (9 df) 9.14 (1df)

49,979 34

49,979 34

N: individuals N: elections

0.49** (0.21)

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

was the case in other states, and the relationship is even more significant when the contrast is between states with sanctioned compulsory voting and the rest (see table 4.5). This might plausibly be interpreted in terms of frustration on the part of citizens compelled to participate in electoral events that they do not find meaningful, but a more compelling explanation is that the fact of being obliged to vote makes citizens more likely to hold the resultant government to account for its actions. The second form of participation to consider is working with people who share the same concerns. In this case, the bivariate relationships suggest a different story: 17.2 per cent of those surveyed in compulsory voting states claimed to have worked with others, whereas the figure is 20.1 per cent for states where voting is not mandated by law. A total of 18.5 per cent of respondents in states with sanctioned compulsory voting reported having worked with others, while 19.7 per cent of those in other states did the same. Yet these differences fail to carry over into multivariate analysis, where the coefficients for both compulsory voting and sanctioned compulsory voting are positive but statistically insignificant (see table 4.6). Finally, it makes sense to consider the fact that different people are attracted to different forms of participation. As can be seen from a cursory overview of

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

72

11:31

Page 72

Full participation Table 4.6 Multilevel logit models of working with others

Variable Individual-level variables: Age Education Female gender Party identification

Model 11: CV

Model 12: Sanctioned CV

⫺0.17 (0.14) 1.29*** (0.09) ⫺0.24*** (0.05) 0.77*** (0.08)

⫺0.17 (0.14) 1.29*** (0.093) ⫺0.24*** (0.05) 0.77*** (0.08)

Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant

⫺0.32 (0.33) 0.06 (0.09) 0.01 (0.25) ⫺2.83*** (0.33)

0.22 (0.16) ⫺0.34 (0.31) 0.07 (0.08) ⫺0.01 (0.25) ⫺2.85*** (0.34)

Intercept variance Chi squared – fixed effects Chi squared – random effects

0.53*** (0.13) 264634.50 (9 df) 16.33 (1df)

0.53*** (0.13) 876.77 (9 df) 16.23 (1df)

49,839 34

49,839 34

N: individuals N: elections

0.06 (0.23)

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

tables 4.4 to 4.6, age is positively associated with contacting politicians, but negatively related to the other two forms of participation; it can be expected that variations in personal dispositions will also translate into different modes of political engagement in any population. A summary variable was therefore constructed to indicate whether a respondent had engaged in any of the three aforementioned types of political participation. When it comes to overall means, the difference in participation levels between those in compulsory voting and voluntary voting states is negligible: 26.5 per cent reported engaging in at least one act of participation in the former and 26.9 per cent in the latter. The difference between states with sanctioned compulsory voting and the rest is slightly greater: 28.5 per cent in the former and 26.6 per cent in the latter. This difference comes into relief in the multivariate models presented in table 4.7. Whereas the coefficient for compulsory voting is insignificant (though positive), that for legally-mandated voting backed up by sanctions is both positive and weakly significant (at the .1 level). This suggests that overall, participation is slightly higher in states where non-voting is both illegal and subject to sanction, all else being equal. The picture painted by these findings is a mixed one, suggesting that political engagement is a complex phenomenon. People in states with compulsory

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 73

Compulsory voting and election compaigns

73

Table 4.7 Multilevel logit models of engagement in any form of political participation Variable Individual-level variables: Age Education Female gender Party identification Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant Intercept variance Chi squared – fixed effects Chi squared – random effects N: individuals N: elections

Model 13: CV

Model 14: Sanctioned CV

⫺0.13 (0.16) 1.45*** (0.10) ⫺0.28*** (0.04) 0.78*** (0.07)

⫺0.13 (0.16) 1.45*** (0.10) ⫺0.28*** (0.04) 0.76*** (0.07)

0.28 (0.27) ⫺0.20 (0.30) 0.09 (0.07) ⫺0.15 (0.24) ⫺2.62*** (0.32)

0.30* (0.17) ⫺0.24 (0.29) 0.10 (0.07) ⫺0.03 (0.14) ⫺2.66*** (0.33)

0.46*** (0.11) 710.97 (9 df) 17.17 (1df)

0.45*** (0.110) 842.69 (9 df) 16.85 (1df)

51,754 34

51,754 34

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

voting (especially with sanctions for non-voting) feel less efficacious on the whole than those living in places where going to the polls is voluntary, yet they participate more, especially in protest activities. This could be interpreted in terms of the possible frustration bred by being obliged by law to attend polls. At the same time, those in compulsory voting states could simply be closer to Pippa Norris’s (1999) model of ‘critical citizens’ who increasingly voice discontent with the political system but are also quite heavily engaged in participating to change that way politics works. Party identification Party identification is the third of the categories of political behaviour relevant to electoral campaigns. We have seen in the previous models that identification with political parties is closely linked empirically to both political knowledge and engagement. What concerns us here is whether levels of party identification are also affected by the institution of mandatory electoral participation. If this were the case, it would suggest that obliging citizens to vote works in the above models both directly and indirectly via its impact on party identification. A cursory overview of the relationship between party identification and legal electoral obligation suggests that, contrary to expectations, the citizens of

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

74

11:31

Page 74

Full participation Table 4.8 Multilevel logit models of party identification

Variable Individual-level variables: Age Education Female gender Efficacy Political knowledge

Model 15: CV

Model 16: Sanctioned CV

1.14*** (0.16) 0.72*** (0.15) ⫺0.16*** (0.03) 0.27*** (0.03) 0.48*** (0.06)

1.14*** (0.16) 0.72*** (0.15) ⫺0.16*** (0.03) 0.27*** (0.03) 0.48*** (0.06)

Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Effective No. parliamentary parties Latin America Constant

⫺0.07 (0.42) ⫺0.06 (0.13) 0.01 (0.62) ⫺1.01** (0.45)

⫺0.36 (0.44) ⫺0.02 (0.44) ⫺0.08 (0.13) ⫺0.07 (0.28) ⫺0.96 (0.45)

Intercept variance Chi squared – fixed effects Chi squared – random effects

0.83** (0.42) 240.61 (10 df) 3.85 (1df)

0.82** (0.42) 251.48 (10 df) 3.91 (1df)

41,587 30

41,587 30

N: individuals N: elections

⫺0.25 (0.61)

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

mandatory voting states have if anything lower levels of party identification than those in states where voting is voluntary. The average level of party identification in compulsory voting states is 60.6 per cent, whereas it is 69.5 per cent in states with voluntary voting. The difference is even more pronounced in states where non-voting is sanctioned; here rates of party identification are 57.8 per cent, as against 69.3 per cent in other states. These patterns are not borne out in multivariate analysis, however. Party identification has been found in previous studies to be influenced by a variety of individual and aggregate-level factors that are included as controls in the multivariate analysis conducted here. These include age, gender, education, political knowledge, efficacy, the effective number of parliamentary parties, and electoral system type (Campbell et al., 1960; Dalton, 1996: chap. 9; 2000; Brader and Tucker, 2001; 2002; Huber et al., 2005). Latin American location was also included as a control variable. The multivariate models presented in table 4.8 indicate that though there is a negative relationship between party identification and both compulsory voting and sanctioned compulsory voting, it does not reach statistical significance. As noted above, it is received wisdom in Australia that compulsory voting is one of the reasons for the high levels of party identification in that country. The

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 75

Compulsory voting and election compaigns

75

findings presented here call the causal link in this relationship into question, and they demonstrate that Australia is if anything an exception to the general rule of lower-than-average levels of party identification in mandatory voting states. The lack of statistical significance for this variable cautions against placing too much emphasis on this finding, however, and we are obliged to conclude that there does not appear to be a strong link in the cross-national setting between identification with political parties and the legal obligation to participate in elections. Party engagement with citizens Party engagement with citizens is the final aspect of campaign-related activities to be considered in this chapter. Two types of behaviour are relevant in this context: citizen participation in campaign activities on behalf of a party or candidate, and party- or candidate-initiated contact with electors. Previous literature has found party engagement with citizens to be affected by a number of individual-level and macro-level variables, including age, gender, education, party identification, efficacy and electoral system type (Ware, 1996: chap. 2; Whiteley and Seyd, 1998; Cox, 1999; Karp et al., 2008). These variables, together with the Latin American dummy, are included in the multivariate models presented here. But before moving to multivariate analysis, it is worth pausing once more to assess the bivariate relationships. When it comes to participation in campaign activities, compulsory voting does not appear to make a tremendous amount of difference, though the bivariate findings do run counter to our expectation that electoral compulsion would depress activism by parties and electors. The average rate of campaign activism in states with this institution is 13.6 per cent, and 12.1 per cent in those without. In states where non-voting is sanctioned, the corresponding figure is 12.9 per cent, whereas it is 12.3 per cent in all other states. Unsurprisingly, neither difference is statistically significant. These slight differences are reflected in the multivariate models presented in table 4.9, though the coefficients corresponding to the relevant variables are well below statistical significance. We find a similar picture when it comes to party- and candidate-initiated contacts with voters. In states with compulsory voting, an average of 27.6 per cent of survey respondents reported having been contacted during the electoral campaign, whereas 25.0 per cent reported contact in states where voting is voluntary. Under sanctioned compulsory voting the situation is reversed, though the difference is very slight: an average of 25.0 per cent of respondents in sanctioned compulsory voting states were contacted, as against 25.5 per cent in all other countries. Again, neither difference is significant in statistical terms. Table 4.10 shows the results of multivariate analysis of contacts. As in the previous models, the coefficients for both compulsory voting and sanctioned compulsory voting are positive but not statistically significant.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

76

11:31

Page 76

Full participation Table 4.9 Multilevel logit models of engagement in campaign activities

Variable Individual-level variables: Age Education Female gender Party identification Efficacy Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Latin America Constant Intercept variance Chi squared – fixed effects Chi squared – random effects N: individuals N: elections

Model 17: CV

Model 18: Sanctioned CV

0.53** (0.25) 0.66*** (0.12) ⫺0.29*** (0.07) 1.21*** (0.05) 0.19*** (0.03)

049** (0.23) 0.61*** (0.11) ⫺0.27*** (0.06) 1.12*** (0.06) 0.18*** (0.03)

0.21 (0.20) ⫺0.48* (0.30) 0.47** (0.21) ⫺4.14*** (0.41)

0.06 (0.18) ⫺0.67** (0.30) 0.51** (0.25) ⫺3.65*** (0.41)

0.55*** (0.14) 2561.99 (9 df) 16.05 (1df)

0.56*** (0.16) 1285.45 (9 df) 12.13 (1df)

46,029 32

46,029 32

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

We had predicted on the basis of case studies that the institution of mandatory electoral participation would represent a disincentive for parties and candidates to engage voters actively in their campaigns and to contact them at election time, given the reduced need for voter mobilisation. The evidence reported here shows no evidence of such an effect, however. If anything, this practice is associated with higher levels of contact and campaign activity than voting under a voluntary regime, though the lack of statistically significant differences warrants caution when interpreting this finding. Conclusion If compulsory voting can be expected to alter the incentive structure citizens face at election time, it will invariably alter the way in which those seeking to win their votes appeal to the electorate, and, in turn, the way citizens react to electoral campaigns. From the perspective of electoral contestants, the main difference between an election held under voluntary voting and one held under mandatory participation is that in the former it is necessary to focus on mobilisation to get one’s supporters to the polls, whereas in the latter this is taken care of (to greater or a lesser extent) by the legal framework, and the main emphasis ought therefore to be on conversion.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 77

Compulsory voting and election compaigns

77

Table 4.10 Multilevel logit models of contact by party or politician Variable Individual-level variables: Age Education Female gender Party identification Efficacy Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Latin America Constant Intercept variance Chi squared – fixed effects Chi squared – random effects N: individuals N: elections

Model 19: CV

Model 20: Sanctioned CV

015 (0.22) 0.59*** (0.13) ⫺0.11*** (0.04) 0.33*** (0.05) 0.06*** (0.01)

0.15 (0.22) 0.59*** (0.13) ⫺0.11*** (0.04) 0.33*** (0.05) 0.06*** (0.01)

0.21 (0.35) ⫺0.62** (0.31) 0.12 (0.52) ⫺1.58*** (0.34)

⫺0.01 (0.39) ⫺0.62** (0.31) 0.32 (0.54) ⫺1.56*** (0.32)

0.64*** (0.18) 196.30 (9 df) 12.85 (1df)

0.64*** (0.17) 187.56 (9 df) 13.57 (1df)

46,645 32

46,645 32

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

This chapter has tested the implications of this difference in a number of ways, but it has provided scant evidence of a clear link between compulsory voting and citizen engagement in electoral campaigns. Citizens under compulsory voting do not have significantly higher (or lower) levels of political knowledge, party identification or direct engagement with political actors during election campaigns. They are somewhat more likely to engage in other acts of political participation, and significantly more likely to take part in protest activities. By contrast, there is weak evidence to suggest that they feel less efficacious than their counterparts in voluntary voting regimes. Several caveats are in order. Firstly, the dataset on which this chapter is based includes only six states with compulsory voting, not the entire universe of states. The effects we observe here might therefore not apply if a wider range of cases were included. Secondly, four of these states (and two of the four states with sanctioned compulsory voting) are located in Latin America, so it may be that specific features of the Latin American context have a bearing on the results of the analysis. Finally, with a maximum of 34 macro-level cases (and fewer in some analyses, due to missing data), it is often difficult to find statistically significant relations among variables. That said, the sample of sanctioned compulsory voting states for which the strongest results were found are broadly reflective of the universe of mandatory voting regimes; half of them are located

M1546 - BIRCH TXT.qxp:ANDY Q7

78

27/10/08

11:31

Page 78

Full participation

in Latin America, one in Europe, and one in Oceania. There is thus reason to believe that these findings are indeed tapping inherent features of compulsory voting. It is worth noting that the specific impact of compulsory voting on electoral campaigns has not featured prominently in debates over whether to make electoral participation mandatory (except in as much as manipulative and abusive practices carried out during campaigns were thought to be susceptible to reduction through this measure – a topic to be dealt with in Chapter 6). In some senses therefore, the effects demonstrated in this chapter are ‘by-products’ of the institution. The following chapter addresses the effect that compulsory voting was most commonly and most directly designed to address: overall rates of electoral participation. Notes 1 Lijphart suggests that the educative effect of compulsory voting may be particularly pronounced shortly after the institution is introduced, in that ‘under compulsory voting, parties and candidates have a strong incentive to get information to previous non-voters’ (1997: 10). Unfortunately empirical data to test this proposition are not readily available. 2 Interestingly, Dutch survey evidence from 1966 and 1999 suggests that there is little or no link between popular support for compulsory voting and levels of political interest, and support for compulsory voting is actually higher among those with lower levels of efficacy (Irwin and van Holsteyn, 2007: 7–8). 3 Switzerland is also included in this dataset, but given that compulsory voting is practiced only in the canton of Schaffhausen, which makes up less than one per cent of the Swiss population, Switzerland is considered to be a voluntary voting state for the purposes of the analyses undertaken here. 4 See Gordon and Segura, 1997; Milner, 2002: chap. 3; Brockington, 2005; Grönlund and Milner, 2006. 5 The analyses reported here were carried out using MLwiN version 2.02 software with restricted iterative generalised least squares (RIGLS) estimation, which is most appropriate with restricted numbers of level-two cases (Steenbergen and Jones, 2002; Rasbash et al., 2004), and predictive quasi-likelihood (PQL) approximation for the estimation of equations with discrete dependent variables. The data were weighted by sample size.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 79

5

Compulsory voting and electoral turnout

Though there may be debates as to the normative merits of imposing a legal obligation to vote, there is a general consensus among contemporary politicians and political scientists alike that low rates of participation are unhealthy for a modern democracy. It is not surprising, then, that the recent wave of calls for the (re)introduction of compulsory voting should have coincided with a general decline in turnout rates in established democracies. Whatever else it accomplishes, obliging people to exercise their right to vote ought, in theory, to have the effect of increasing rates of electoral participation. In the academic domain, it is in the context of turnout that compulsory voting has received the most extensive treatment. The question that concerns us in this chapter is how efficient compulsory voting really is in improving turnout levels. The overall aim of the chapter is thus to review previous analyses of the impact of compulsory voting on electoral participation levels, and to assess the conditions under which mandatory voting rules are most likely to be associated with a substantial change in rates of voting. Does it matter whether people are obliged to vote? How important are formal sanctions? There are three principal strategies in the comparative politics toolkit for addressing this question: longitudinal analysis of the impact of institutional change (introduction or abolition of compulsory voting) in individual countries, cross-sectional analysis of variations in institutional design across states, and counterfactual analysis based on survey research. All three strategies will be employed here. The first avenue for exploring the impact of compulsory voting will be an examination of individual cases of states that have introduced or abolished this institution at various points in their history, as well as those that have altered their enforcement mechanisms. The chapter will then proceed to a cross-national quantitative analysis of the impact of mandatory voting on aggregate turnout levels, drawing on previous studies, but employing more nuanced measures of the institution. In both cases the analyses will be confined to an examination of elections to lower (or only) chambers of parliament in order to ensure comparability across states. Before concluding, the chapter will provide an overview of previous counterfactual

M1546 - BIRCH TXT.qxp:ANDY Q7

80

27/10/08

11:31

Page 80

Full participation

studies of the likely impact of abolishing compulsory voting in states that currently have it. The evidence presented here largely corroborates that of existing research,1 which shows that compulsory voting – especially when accompanied by sanctions – is associated with higher overall turnout levels than voluntary voting. The longitudinal evidence from countries that have altered their voting regimes also provides some indication of the direct impact of such changes on turnout level. Longitudinal analysis: changes in voting regimes The main advantage of examining the impact of changes in electoral regimes over time is that such analysis allows us to address directly the question of what would happen if compulsory voting were introduced (or abolished) in a modern democracy. Indeed, investigation of individual cases of institutional change is the mainstay of much of the popular and semi-academic literature on this topic. Yet there are three principal difficulties associated with longitudinal analysis of the impact of compulsory voting on turnout: firstly, the introduction of this institution in many cases took place several generations ago in the context of a limited franchise. This is true, for example in many of the Latin American states. Evidence of changes in rates of electoral participation from such cases are arguably of very limited relevance to the contemporary world, given the substantial socio-demographic and political changes that have taken place in the intervening period. The second problem commonly associated with over-time analysis of individual cases is that that the introduction of compulsory voting was often accompanied by other changes in franchise provisions, or by a change of electoral system, making it difficult to isolate the impact of enforced electoral obligation per se. Finally, reliable data on turnout change is too limited to enable multivariate statistical analysis, so the assessment of longitudinal change must be confined to ‘eyeballing’ data from individual countries. Historical data from the best-known cases of compulsory voting suggest that its introduction was accompanied by substantial rises in rates of electoral participation. In Australia there was an increase following this reform from 64.2 per cent in 1924 to an average of 94.6 per cent in the nine elections following the change (Jackman, 2001: 16316). In the Netherlands and Belgium, the introduction of this institution accompanied that of universal manhood suffrage, which blurs the impact on turnout considerably. Nevertheless, the change in levels of electoral participation is in the expected direction. Tingsten reports that in the Netherlands turnout rose from 84.5 per cent in the election immediately preceding the 1917 reforms to 91.4 per cent in the election following, while Belgium witnessed a similar shift, with turnout rising from

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 81

Compulsory voting and electoral turnout

81

between 70 and 85 per cent to about 95 per cent (Tingsten, 1963 [1937]: 190– 3). As we saw in Chapter 2, the introduction of mandatory electoral participation also accompanied a host of other electoral and franchise reforms in Argentina. The joint effect of these reforms was linked to the removal of the Radicals’ electoral boycott; turnout rose from a mere five per cent prior to the Sáenz Peña reforms to 62.9 per cent in the elections of 1914 (Díaz, 1983: 78). More relevant for analysis of the contemporary world, however, are more recent cases. Taking the post-Second World War world as our frame of reference for contemporary comparison, we will consider only those cases where the adoption of mandatory voting was not accompanied by either independence or substantial alterations in the suffrage (these criteria exclude Cyprus, Egypt, and El Salvador). This leaves the following eight cases available for analysis: Singapore (1958), Guatemala (1965), Dominican Republic (1966), Paraguay (1967), Turkey (1982), Laos (1989), Fiji (1990), and Thailand (1997). Mark Franklin argues that institutional changes such as the introduction of compulsory voting have cumulative effects over time; according to this argument, compulsory voting should lead to an immediate upward shift in turnout, followed by additional rises over time as the population adjusts to the new institution and the effects of this reform are consolidated when successive generations are socialised into the new regime. Franklin claims that institutional changes such as this require 10 elections for their effects to be fully measured (2004: 84–5). For this reason we consider not only the immediate effects of institutional change, but also subsequent variations in turnout (where data are available). Assessing these cases one by one, we can discern evidence of the impact on turnout of the shift to mandatory electoral participation (see table 5.1). Singapore held only one voluntary voting election (in 1954), but on the basis of the evidence, it does appear that the introduction of mandatory voting in 1958 had a dramatic impact on rates of electoral participation in the new state. Turnout went up from 52.7 per cent in the first post-independence election to 90.1 per cent following the reforms, and it has remained at approximately this level ever since (rising slightly in subsequent elections to about 95 per cent). The difference between the average turnout in the election preceding change and the average for the 10 elections following change is 40.1 per cent. The stringent sanctions regime in Singapore undoubtedly accounts for this effect. In Guatemala, turnout levels before the introduction of compulsory voting in 1965 are somewhat lower than those after, suggesting that the shift to mandatory voting did have an impact. The absence of sanctions in this country is probably the reason why the effect of the change is less pronounced than in Singapore; the difference in pre- and post-change average turnout is only 5.2 per cent. In the Dominican Republic, lack of sufficient data on the elections held prior to the introduction of compulsory voting in 1966 make it difficult to interpret

8-post

7-post

6-post

5-post

4-post

1958: 46.7 1959: 44.9 1961: 44.5 1966: 50.0 1970: 53.3 1974: N/A 1978: N/A 1982: N/A 1985: 69.3 1990: 56.4 1994: 21.0

1965 Yes 1952: N/A 1957: N/A 1962: 64.7 1966: 75.6 1970: 63.5 1974: 71.7 1978: 75.8 1982: 71.5 1986: 72.2 1990: 59.9 1994: 87.6

1966 ?

Dominican Rep.

1958: N/A 1960: N/A 1963: 85.1 1968: 73.1 1973: 77.4 1978: 86.0 1983: 92.6 1988: 92.2 1989: 52.0 1993: 67.6 1998: 80.5

1967 No

Paraguay

1969: 64.3 1973: 66.8 1977: 70.4 1983: 92.3 1987: 93.3 1991: 83.9 1995: 85.2 1999: 87.1 2002: 76.9

1982 Yes

Turkey

1965: N/A 1967: N/A 1972: 67.8 1989: 98.4 1992: 99.3 1997: 99.4 2002: 99.9 2006: 99.8

1989 ?

Laos

1977: 73.8 1982: 86.7 1987: 71.0 1992: 79.0 1994: 73.6 1999: 91.5 2001: N/A 2006: 64

1990 No

Fiji

1992: 59.3 1995: 62.0 1996: 62.4 2001: 69.9 2005: 75.1

1997 Yes

Thailand

11:31

3-post

2-post

1954: 52.7 1959: 90.1 1962: 90.6 1963: 95.1 1968: 94.1 1972: 82.1 1976: 95.1 1980: 95.5 1984: 95.6

1958 Yes

Guatemala

27/10/08

Election after

Election before

2-pre

3-pre

Year of adoption Expected impact?

Singapore

Table 5.1 Impact of the introduction of compulsory voting on turnout (%)

M1546 - BIRCH TXT.qxp:ANDY Q7 Page 82

40.1

Initial change

Cumulative change

5.2

1995: N/A 1999: 53.8 5.5 3.5

1998: 52.9 2002: 51.0 10.9 21.9 19.3

⫺12.0 ⫺8.9

2003: 64.1

31.6

30.6 ⫺0.2

8.0

11.3

7.5

Sources: IPU Parline database at http://www.ipu.org/parline-e/parlinesearch.asp; IDEA, 2002; IDEA Voter Turnout Database at http://www.idea.int/vt/index.cfm; Hartmann, 2001a: 135; Hartmann, 2001b; Bendel and Krennerich, 2005: 324; Brea Franco, 2005: 248; León-Roesch and Ortiz, 2005: 426.

NB: Cell values are turnout levels, reported as the proportion of registered electors who participated.

10-post

1988: 94.7 1991: 95.0 37.4

9-post

M1546 - BIRCH TXT.qxp:ANDY Q7 27/10/08 11:31 Page 83

M1546 - BIRCH TXT.qxp:ANDY Q7

84

27/10/08

11:31

Page 84

Full participation

the effect of this change. True, turnout increased from 64.7 per cent to 75.6 per cent following the reform, but it then fell back to 63.5 per cent at the following election, and the average turnout in post-reform elections is, at 68.2 per cent, only 3.5 per cent higher than that of the election immediately preceding the shift to compulsory voting. It is unclear what to make of these figures, so we have to conclude that this is an ambiguous case. In Paraguay, the introduction of mandatory voting was accompanied by a move toward authoritarianism, as detailed in Chapter 2. The change in voting rules was not, however, successful in increasing turnout; rates of electoral participation actually fell during this period. Lack of sanctions for non-participation may account for the fact that average turnout in the ten elections following the introduction of compulsory voting was 8.9 per cent lower than the average in the parliamentary poll that preceded the change.2 In Turkey turnout had already been on the rise prior to the introduction of the legal obligation to participate in elections in 1982, but this change prompted a dramatic increase in turnout of 22 percentage points, an increase that cannot plausibly be explained by non-institutional changes. It should be noted that electoral participation declined by about 10 percentage points two elections later, but it nevertheless remained at a level significantly above those prior to the change. The average for the six post-change elections is, at 86.5 per cent, 19.3 percentage points higher than the average for the three elections that preceded reform. Again, the existence of sanctions may well account for this increase in reported rates of participation. The Laotian case is somewhat difficult to interpret, in that the dramatic increase in turnout that accompanied the advent of electoral compulsion in 1989 could well be due also to the existence of a hard-line communist regime (the previous elections had been held in 1972, prior to the abdication of the king in 1975). Nevertheless, the 99-per-cent-plus turnout rates that have been witnessed in this state under compulsory voting are consistent with expectations (the precise sanctions regime in Laos is unclear, but whether or not there are legal sanctions for non-voting, political sanctions most probably account for these figures). Unfortunately data on turnout prior to Fiji’s adoption of mandatory voting in 1990 are patchy. The existing evidence suggests that the introduction of electoral compulsion was initially successful at increasing turnout, but that electoral participation subsequently declined. The first post-independence elections in 1972 witnessed a reported turnout of 99.3 per cent. In 1982 (the next election for which data are available), the figure fell to approximately 86 per cent. In 1992, the first election held under a compulsory voting regime, turnout was down to 78.8 per cent, and despite some subsequent rises, it dropped further to 64 per cent in the most recent poll (2006). (It is worth noting that the introduction of mandatory voting in Fiji in 1990 was also

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 85

Compulsory voting and electoral turnout

85

accompanied by significant changes to the electoral system (see Hartmann, 2001b: 649)). In this case, the existence of sanctions does not appear to have been successful in bringing about behavioural change. This may, in part, be due to the turbulent political experience of this country in recent years. In Thailand the impact of mandatory voting was, as expected, to increase turnout. Though the rise was a modest 7.5 percentage points at the first parliamentary election following the change, there was a subsequent further increase of about five percentage points at the next election. Turnout has not reached particularly impressive rates in this case, yet the sanctions regime in force does appear to have ensured that the new institution has had an effect on rates of electoral participation. We can draw three main conclusions from these longitudinal figures. Firstly, the data provide some evidence to suggest that the introduction of mandatory electoral participation does make a difference to turnout in the contemporary world: in four of the eight cases considered here, the introduction of compulsory voting had the anticipated effect of increasing turnout levels; two cases are ambiguous, and in the remaining two cases, the shift clearly did not lead to a rise in reported rates of participation. Another way of viewing these data is that the average turnout change between the last election before the adoption of mandatory electoral participation and the first election following this change was 13.7 per cent, and the difference between the average turnout in the last three pre-reform elections and the first post-reform polls is 12.7 per cent. Secondly, sanctions also matter. The average pre-post difference in the four states with sanctions is 17.6 per cent, whereas the average difference in those without sanctions is approximately zero (these calculations exclude the Laotian case). Though sanctions may not always be effective in generating truly ‘full’ participation, they do seem to ensure that participation rises following the introduction of mandatory electoral participation. Thirdly, it appears from these examples that – contrary to Franklin’s predictions – where there is a positive impact it is generally immediate. Subsequent elections sometimes exhibit an increase in turnout, sometimes not. It therefore does not seem that this institution is one that takes time to ‘bed in’; nor is there evidence of a cumulative effect as successive generations are socialised into the institution. In none of the countries considered here do we observe clear evidence of this tendency.3 In the second election following the change, turnout was just as likely to fall again as it was to rise further, though it generally stabilised at a level higher than that observed under voluntary voting. This suggests that the first election following reform is crucial and that if electoral engineers miss the chance in this election to make compulsory voting ‘work’, they may well miss the chance altogether. Turning now to the impact of the abolition of compulsory voting, we see similar effects. The relevant post-war cases include the Netherlands, which

M1546 - BIRCH TXT.qxp:ANDY Q7

86

27/10/08

11:31

Page 86

Full participation

abolished the institution in 1970 after over 50 years of use, and Liechtenstein, which did the same in 2004. In addition to these state-wide shifts, it is also worth considering the partial abolition of the practice in the remaining three Austrian provinces which still made voting mandatory at the time of the institution’s state-wide abolition in 1992, as well as several Swiss cantons, which gradually made voting voluntary over the course of the twentieth century. While the impact of introducing compulsory voting was generally found above to be immediate, we might expect that the effect of removing electoral compulsion would be slower; generations socialised into voting during the mandatory voting period might well still vote in higher numbers due to habituation, as has indeed been found by Irwin (1974: 310). When compulsory voting was abolished in the Netherlands in 1970, there was a fall in turnout levels of nearly 16 per cent between the last mandatory voting election and the first election where voting was voluntary (see table 5.2). It seems that once the novelty effect of voluntary voting wore off, a number of Dutch voters returned to voting, for turnout had risen by the late 1970s to a level only about seven percentage points below what it had been when participation in the polls had been mandatory. It subsequently fell again, however, such that the average for the 10 post-abolition elections was, at 81.6 per cent, 13.6 percentage points below the average for the three elections that preceded change. In Austria, the situation is less clear. As can be see from the data in table 5.2, turnout had been falling prior to the change anyway, so it cannot be claimed with any confidence that the abolition of compulsory voting in the provinces of Styria, Tyrol and Vorarlberg (which collectively represent only 27.6 per cent of the Austrian population) had any discernible influence on overall rates of electoral participation. Yet the overall trends may mask within-country variations in turnout decline that suggest compulsory voting could have had the expected impact. The within-country variations among Austrian provinces approximate the conditions of a ‘natural experiment’, in that the controlled setting of similar populations, similar political circumstances and identical elections facilitates insight into the impact of variations in the application of electoral compulsion. Such an analysis was undertaken by Wolfgang Hirczy (1994) on turnout differences among the Austrian provinces between 1953 and 1987; he finds evidence that where it exists, mandatory voting has had the effect of attenuating turnout declines in Austria, and changes in the provisions regarding compulsory voting led to significant differences in turnout levels (1994: 67–74). As noted in Chapter 2, Liechtenstein was the first country in the world to adopt compulsory voting at national level in 1878. After 126 years of use, this state abolished the practice while at the same time introducing all-postal voting for its seventeen thousand registered electors. Turnout levels exhibited virtually no change in the 2005 election, as postal voting evidently served to

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 87

Compulsory voting and electoral turnout

87

Table 5.2 Impact of the abolition of compulsory voting on turnout (%)

Year of abolition Expected impact? 3-pre 2-pre Election before Election after 2-post 3-post 4-post 5-post 6-post 7-post 8-post 9-post 10-post Initial change Cumulative change

Netherlands

Austria

Liechtenstein

1970 Yes

1992 ?

2004 No

1959: 95.6 1963: 95.1 1967: 94.9 1971: 79.1 1972: 83.5 1977: 88.0 1981: 87.0 1982: 81.0 1986: 85.8 1989: 80.3 1994: 78.7 1998: 73.2 2002: 79.1 ⫺15.8 ⫺13.6

1983: 92.6 1986: 90.5 1990: 86.1 1994: 82.5 1999: 80.4 2002: 84.3 2006: 74.2

1993: 85.3 1997: 86.9 2001: 86.7 2005: 86.5

⫺3.6 ⫺9.3

⫺0.2 0.2

NB: Cell values are turnout levels, reported as the proportion of registered electors who participated. Sources: IDEA, 2002; IDEA Voter Turnout Database at http://www.idea.int/vt/index.cfm; 2006: IPU Parline database at http://www.ipu.org/parline-e/parlinesearch.asp.

counteract the decline that might have been anticipated with the removal of electoral compulsion. The Swiss case is more difficult to assess, due to the fact that compulsory voting has always been applied at the cantonal level, and that the number of cantons mandating electoral participation varied significantly over the course of the twentieth century. Patricia Funk (2004) assesses the effect of abolitions of compulsory voting regulations over time and across cantons. The overall finding of this study is that the removal of compulsory voting resulted in a decrease in turnout of between six and ten per cent (2004: 6). Interestingly, Funk also finds a correlation of 0.9 between the degree of compliance with the law prior to its removal (as measured by the turnout level), and the decline in turnout following abolition of the requirement (2004: 5), suggesting that citizens in relatively compliant cantons were more responsive to alterations in legal regulations than was the case in cantons where aggregate levels of compliance were lower.

M1546 - BIRCH TXT.qxp:ANDY Q7

88

27/10/08

11:31

Page 88

Full participation Table 5.3 Impact of the removal of sanctions on turnout (%)

Year of removal Expected impact? 3-pre 2-pre Election before Election after 2-post 3-post 4-post Initial change Cumulative change

Venezuela

Greece

1993 Yes

2001 No

1978: 87.6 1983: 87.8 1988: 81.7 1993: 60% (approx) 1998: 52.6 2000: 56.6 2005: 25.3 21.7 37.1

1993: 79.2 1996: 76.3 2000: 76.0 2004: 75.6

0.4 1.5

NB: Cell values are turnout levels, reported as the proportion of registered electors who participated. Sources: IDEA, 2002; IPU Parline database at http://www.ipu.org/parlinee/parlinesearch.asp.

We would expect the impact of abolishing compulsory voting to be less clearcut than that of its introduction, and that is indeed what we observe in the small number of cases analysed here. Nevertheless, the overall figures are consistent with what we would anticipate: the average change between the last elections conducted under compulsory voting (or partial compulsory voting) in the Netherlands, Austria and Liechtenstein and the first polls held under a voluntary system was a decline of 6.5 per cent, and the average difference in turnout between the last three compulsory or semi-compulsory elections and the first voluntary ones is 7.6 per cent. Finally, it is worth considering the effect of changes in sanctions regimes on voting levels. These include Venezuela, which formally abolished sanctions for non-voting in 1993, and Greece which did the same in 2001. We would expect the removal of sanctions to depress turnout, all else being equal. Table 5.3 presents the turnout figures for these two cases before and after the change. In Venezuela the reform was part of a package of measures brought in to address the legitimacy crisis that had developed in the 1980s and early 1990s. This period of political turbulence had been accompanied by a slight decline in turnout, though overall rates of electoral participation were still high by international and regional standards.4 Following the reforms, turnout dropped suddenly and dramatically, providing clear evidence that the removal of sanctions had an impact in the expected direction.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 89

Compulsory voting and electoral turnout

89

In Greece, a constitutional reform in 2001 removed existing penalties for non-voting, though voting still remains compulsory in theory (Pantelis et al., 2008; Malkopoulou, 2007). These penalties had mainly involved restrictions on access to state documents such as drivers licences and passports, and enforcement had been weak (Gratschew, 2002). The data in table 5.3 bear this out: prior to the reform, turnout had been by no means universal. Moreover, it had been declining slightly over the course of the 1990s, and the constitutional change does not seem to have interfered in this election-on-election decline one way or the other. The Greek case appears to have been one of bringing the law into line with civic norms (see also below), representing a tacit acknowledgement that sanctions had not been effective. Thus the removal of sanctions had the expected effect in Venezuela where they had been relatively successful in maintaining levels of electoral participation, but they had no evident impact in Greece, where they had not been so successful. This finding comports with the Swiss evidence put forward by Funk that the abolition of compulsory voting had a greater influence where the law had been doing more ‘work’. The longitudinal evidence presented in this section provides considerable support for the supposition that compulsory voting should have a positive impact on turnout. Though some of the cases are ambiguous, and the small number of countries under consideration makes it difficult to draw firm conclusions, certainly the evidence is broadly in line with expectations; the introduction of compulsory voting was associated in the majority of cases with an increase in turnout, while a reversion to voluntary voting was associated with a decline in the rate of electoral participation, as was also true for the removal of sanctions. Though there are reasons to doubt the causal efficacy of institutional reform in certain cases, the data overall support the conclusion that compulsory voting increases and/or maintains turnout. The major flaw in election-on-election comparisons such as those conducted here is that they do not enable us to take other factors into consideration. It is therefore necessary to turn to a large-N cross-national framework in order to engage in multivariate analysis. This is the task of the next section. Cross-sectional analysis: Comparative evidence for the impact of compulsory voting Cross-sectional analyses of the impact of compulsory voting on turnout have been employed since political scientists began examining the institution. Tingsten compares turnout in early twentieth century Switzerland (1919– 1931) in cantons with and without compulsory voting, and finds an average variation of between 5.4 and 9.8 per cent, depending on the election (1963 [1937]: 197). A similar analysis of Austrian provinces in 1907 and 1911,

M1546 - BIRCH TXT.qxp:ANDY Q7

90

27/10/08

11:31

Page 90

Full participation

shortly after the introduction of mandatory voting laws, finds a difference in average turnout levels of between 20 and 25 per cent (Tingsten, 1963 [1937]: 199–200). The Hirczy and Funk studies discussed in the preceding section also include cross-sectional elements in their intra-state comparisons. But it is the cross-national variant of cross-sectional analysis that has most frequently been used to assess the impact of mandatory voting on aggregate rates of electoral participation. The literature on the determinants of turnout is vast, and compulsory voting is a variable that figures regularly in comparative studies of this phenomenon. This approach has nevertheless been criticised by Hirczy (1994: 65) on two main counts, and it is worth pausing to assess these objections. Hirczy’s first point is that the establishment of cross-sectional associations does not demonstrate a causal relationship, and that the causal pathway could in fact be from high turnout to compulsory voting, in as much as ‘A society that practices near-universal participation may simply enshrine its civic norm in law’ (1994: 65). It is clear from the historical analysis undertaken in Chapter 2 that this was rarely the case. Countries adopted compulsory voting for a variety of reasons, but combating low turnout featured prominently among them. Though theoretically possible, it is in practice unclear that any state has adopted compulsory voting in order to ‘enshrine its civic norm in law’. Moreover, most of the reforms through which compulsory voting were introduced were carried out sufficiently far in the past that it is not appropriate to invoke current norms to account for them. On the contrary, most case studies of compulsory voting have found that the institution has been responsible for creating the norm. For example, Galen Irwin comments that in the Netherlands compliance with electoral compulsion was not achieved through fines (which were applied only sporadically), but that ‘obedience was simply recognition that the law was the law and the law should be obeyed’ (1974: 294). Likewise, in a longitudinal analysis of compulsory voting in Swiss cantons, Patricia Funk (2004) finds that the law regarding voting was an important enforcer of social norms, not the other way round. The main cases of abolition – the Netherlands, Austria and Liechtenstein – all took place in the context of turnout levels that had remained high by international standards. The cases of abolition in the Swiss cantons and the removal of sanctions in Greece are admittedly more ambiguous, but overall, it appears from the longitudinal analyses carried out above that the causal arrow flows mainly from institutional requirements to behaviour, and not the other way round. This objection to crosssectional analysis can therefore be largely discounted; if we do find a statistically significant association between turnout and compulsory voting, we can be confident that it is the institution of mandatory electoral participation that is responsible for variations in turnout, and not the other way round. Hirczy’s second objection is that the institution of compulsory voting often coincides with proportional representation, which has also been found to

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 91

Compulsory voting and electoral turnout

91

increase turnout, making it difficult to disentangle the two effects. In fact, eight of the 29 countries currently practicing compulsory voting have nonproportional electoral systems, and there is no evident problem with multicollinearity between these two variables in the analyses carried out below. Hirczy’s final objection is that most of the extant cross-national studies at the time he was writing suffered from a problem of few cases and many variables. This problem is also remedied in the analyses carried out here, which are undertaken on a global dataset. Hirczy’s main point, that the analysis of compulsory voting should include a longitudinal as well as a cross-sectional component, is well-taken, and the preceding section provides evidence of the benefits of the longitudinal approach. There are nevertheless significant advantages to crosssectional analyses, as they allow us to test hypotheses in a multivariate context on the full range of modern states. In his meta-analysis of the aggregate-level determinants of turnout, Benny Geys reports that 13 of 15 studies including a variable designating compulsory voting find a relationship between this variable and turnout (2006: 652). Most investigations of this phenomenon have been undertaken in the context of democracies, and it has generally been found that compulsory voting has a positive impact on turnout in this context (e.g. Powell, 1980: 25; 1982: 113; Crewe, 1981: 240–1; Franklin, 1996: 227; 1999: 215; 2002; 2004: 135; Jackman, 1987: 412; Jackman and Miller, 1995: 474; Katz, 1997: 241).5 Some global cross-national turnout studies have not found compulsory voting to have a significant effect, however. These include two analyses of advanced industrial democracies (Crepaz, 1990; Gray and Caul, 2000: 1110) as well as Pippa Norris’s global study (2002: 65). In a subsequent global comparative analysis of the determinants of aggregate electoral participation, Norris explores this finding further, arguing that the impact of compulsory voting is considerably greater in established democracies than in other states. She provides evidence that the relationship is actually negative in semi-democracies, while it is severely attenuated in new democracies and in non-democratic states (Norris, 2004: 169). It is, however, worth mentioning that some studies based on old, new and semi-democracies have found a positive relationship between compulsory voting and turnout, including those by Blais (2000: 28), Blais and Dobrzynska (1998: 246), and Karp and Banducci (2008). These investigations have variously estimated that compulsory voting increases turnout by between six and twenty percentage points. Moreover, electoral compulsion is generally found to be the most powerful institutional variable in such studies, having a greater impact on turnout than voting provisions or the electoral system. As suggested by the above account, most research has operationalised compulsory voting as a binary (dummy) variable, but there have been exceptions. In their study of turnout in Latin America, Fornos and colleagues operationalise compulsory voting in terms of a four-point scale designating degrees

M1546 - BIRCH TXT.qxp:ANDY Q7

92

27/10/08

11:31

Page 92

Full participation

of compliance (2004: 920). Their analysis finds the impact of compulsory voting is particularly marked when effective sanctions are applied (2004: 923–5). Siaroff and Merer (2002: 919) include a variable for ‘strictly-enforced compulsory voting’ (which proves highly significant in their European study). The database employed in the investigation conducted here includes direct elections to lower (or only) houses of parliament in all United Nations member states between 1990 and 2006. This represents exactly 700 elections in 181 countries, perhaps the broadest ranging sample with which the impact of compulsory voting on turnout has yet been tested. The 17-year time span reduces the likelihood that fluctuations in the global economy or the global political situation might account for the turnout trends observed. At the same time, the period under analysis falls into the post-war domain of contemporary political configurations. This sample is therefore in theory well suited to providing insight as to the effect of compulsory voting on turnout in the contemporary world. For the purposes of this analysis, aggregate electoral participation is measured as a proportion of the registered electorate. There have been numerous discussions of the pros and cons of basing turnout on the voting age population or the registered electorate (see, for example, Blais, 2000: 23; López Pintor et al., 2002; Franklin, 2004: 86–7; Norris, 2004: 168–70; Geys, 2006). The choice of variable is determined in the current analysis largely by data availability – data on turnout as a proportion of the voting age population are simply not available for a large proportion of the elections considered here, so this analysis relies instead on turnout as a percentage of the registered electorate. The choice of control variables to include is informed by the useful metaanalysis of aggregate-level studies of the determinants of turnout conducted by Geys (2006). Geys analyses 83 aggregate-level studies of turnout and identifies the variables that most often prove significant. Of these, two – previous turnout and registration requirements – are excluded from the present analysis on theoretical grounds. In the case of previous turnout, the relationship is close to being tautological and it makes little sense to include this variable in the type of pooled data structure employed here. Registration requirements are excluded due to the fact that turnout is measured as a proportion of the registered electorate. Two other variables are excluded because of lack of usable data for the full range of cases included here: population stability and campaign spending. The remaining variables found to be significant in the Geys study are included as control variables in this analysis: population size (logged), closeness of the race (margin of victory), concurrent elections (operationalised here as a concurrent presidential election), and electoral system type. Per capita GDP (logged) is also included, because though Geys does not assess it in his investigation, it has been found in a number of other studies to be an important determinant of aggregate-level turnout (Powell, 1980: 25; 1982: 121; Blais and

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 93

Compulsory voting and electoral turnout

93

Dobrzynska, 1998: 247; Blais, 2000: 28; Norris, 2002: 65; 2004: 160). Level of democracy was considered for inclusion for similar reasons (Norris, 2002: 65; 2004: 160). Details of variable construction and data sources can be found in the Appendix. Mean turnout for the 122 elections in this sample conducted under compulsory voting was 75.3 per cent, as against 69.2 per cent in the 506 elections held under voluntary regimes.6 This provides prima facia evidence that, as reported in most previous studies, compulsory voting has a significant impact on turnout, even when other relevant factors are taken into consideration. Moving now to multivariate analysis, it is necessary to choose between the variety of different estimation methods currently available for analysing pooled cross-sectional time series data of the type assembled here. The well-known difficulties associated with assessing the effect of rarely-changing institutional variables when using data of this type suggest that the best strategy is ‘between’ estimation, which is based on the country means for each variable. This allows overall relations between relatively stable country-specific variables to come to the fore, it prevents short-term effects from unduly affecting the results, and it overcomes problems associated with temporal dependencies. Admittedly, this estimation strategy does not allow the analyst to take advantage of the information that might be gained from over-time variation in the institutional variable of interest, but compulsory voting varies so little over this time period that the country-specific longitudinal analyses conducted above were deemed preferable to trying to make sense of temporal patterns in the pooled data. Table 5.4 presents the results of the cross-national multivariate analysis. Model 1 is designed to asses the impact of all forms of compulsory voting – sanctioned and unsanctioned – on aggregate turnout. The insignificant (though correctly signed) coefficient for compulsory voting suggests that legally obliging citizens to attend the polls has no strong impact on turnout. This may be due to the fact that, unlike many previous studies, this sample includes a relatively large proportion of non-democratic states whose high levels of turnout are achieved by means of coercion. In order to probe this possibility, Model 2 includes only those cases that were in the top half of the democratic scale (they had a combined Freedom House rating of 9 or greater on the inverted scale employed here; see the Appendix for details). In this model the coefficient for compulsory voting rises somewhat and its standard error decreases (despite the fact that the number of cases is cut by about a third), but the coefficient still fails to reach statistical significance. It stands to reason that only compulsory voting with sanctions can be expected to have a strong effect on rates of electoral participation in the wide range of countries included here. The limited evidence presented in the previous section suggested that the impact of removing sanctions or the legal obligation to participate in elections altogether was greatest when compliance was

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

94

11:31

Page 94

Full participation

Table 5.4 OLS regression models of the impact of compulsory voting on turnout (country averages) Variable Compulsory voting Sanctioned compulsory voting Electoral obligation constitutionalised

Model 1

Model 2

2.55 (3.27)

3.74 (3.18)

Model 3

Model 4

11.82*** (3.91) ⫺0.20 (3.23)

Controls: Per capita GDP (log) 2.10*** (0.78) 2.55*** (0.82) 1.72** (0.95) 2.25*** (0.95) Population (log) ⫺.88* (0.56) ⫺1.10** (0.55) ⫺0.88* (0.55) ⫺0.84* (0.56) Proportional electoral system 2.75 (2.55) 1.37 (2.79) 2.47 (2.45) 3.14 (2.60) Concurrent presidential elections 5.12 (3.59) 3.71 (4.19) 4.90 (3.39) 5.87* (3.58) Margin of victory 0.06 (0.06) ⫺0.00 (0.08) 0.06 (0.06) 0.06 (0.06) Constant

56.93 (9.59)

57.00 (9.89)

59.33 (9.29)

55.68 (9.50)

N: Adj. R2:

159 .07

107 .13

159 .12

159 .07

* p ⬎ .10; ** p ⬎ .05; *** p ⬎ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

highest, and the only sure way of ensuring compliance is by means of sanctions. Model 3, which includes sanctioned compulsory voting as a variable, demonstrates that this is indeed the case. The impact of this variable on turnout is significant at the .003 level, and it is clear from this equation that enforcing a legal obligation to vote is strongly associated with turnout levels, even in the global sample employed here. All else being equal, states with sanctioned compulsory voting are predicted by the model to have turnout that is 11.8 percentage points higher than those without. A final model (Model 4) tests the supposition that the mere fact of having the obligation to vote written into the constitution might have an impact of rates of electoral participation. It is not particularly surprising that this is not found to be the case; the variable designating constitutionalised electoral obligation fails to reach anything near statistical significance, and it even has the wrong sign. A word is in order about the control variables included in these models. As in many previous studies, smaller and richer states are found to have higher levels of turnout than their poorer, bigger counterparts. The relatively small number of cases included and the global scale of the sample mean that it is not

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 95

Compulsory voting and electoral turnout

95

surprising that variables found in other more homogenous samples fail to reach statistical significance. This is true for the controls included for electoral systems and – with one exception – concurrent presidential elections. Level of democracy also proved insignificant in preliminary models, and muticollinearity between this variable and per capita GDP necessitated its exclusion from the final models reported.7 All in all, the cross-national evidence provided here strongly supports both the findings of the longitudinal explorations undertaken above and the previous literature on turnout. Compulsory voting is an institution that ‘matters’ more than most, but its impact is directly related to the enforcement regime associated with it. Counterfactual evidence The final strategy used to examine the effect of mandating participation is the analysis of counterfactual evidence drawn from bespoke surveys. Such surveys are most commonly carried out in countries that already have compulsory voting. The two most intensively-studied cases are those of Belgium and Australia, which will provide the focus of analysis here. A question in the Belgian election survey (‘If compulsory voting were abolished in Belgium, would you still go to vote?’) has been employed to undertake counterfactual simulations of the scale of potential non-voting. In the 1991 survey, 30 per cent of respondents answered this question in the negative (Hooghe and Pelleriaux, 1998: 420–1). Studies of Wallonia have found that only 44.3 per cent in 1991 and 44.8 per cent in 1995 would always vote if electoral participation were not mandatory, though the proportion who claim they would ‘never’ vote is 22.3 per cent and 24.4 per cent respectively, leaving a significant number of respondents somewhere in between these two positions (de Winter and Ackaert, 1994: 83; Schiffino and Aish, 1999: 102). By the time of the 1999 elections, the proportion claiming they would ‘never’ vote under a voluntary regime had risen to 25.8 per cent, while only 46.7 per cent said they would ‘always’ turn up at the polls in the absence of compulsion. A further 14.9 per cent said they would ‘generally’ vote, while 12.5 per cent said they would vote only ‘sometimes’ (de Winter et al., 2003: 60). In Australia, the proportion of the population voicing support for compulsory voting has generally been between 60 and 70 per cent (McAllister, 1986; Orr, 1998: 577; Mackerras and McAllister, 1999: 221). Survey data from 1996 analysed by Mackerras and McAllister show that 68 per cent of respondents would have voted under a voluntary system, and a further 19 per cent ‘probably’ would have voted. Only three per cent ‘definitely’ would not have voted (Mackerras and McAllister, 1999: 227).8 Simon Jackman (1999) notes that survey data such as this are likely to over-estimate actual vote intentions,

M1546 - BIRCH TXT.qxp:ANDY Q7

96

27/10/08

11:31

Page 96

Full participation

in as much as election surveys tend to under-sample those with low levels of political interest (a phenomenon that Mackerras and McAllister also acknowledge), and social desirability effects inflate estimates of voluntary voting. In the context of compulsory voting, this is compounded by the further problem of lack of reflection: ‘responding to a survey about politics is probably one of the few times that most people are required to think about and offer opinions about political participation: over seventy years of compulsory voting in Federal Elections means the question “are you going to vote?” seldom comes up in everyday conversation about politics in Australia’ (Jackman, 1999: 37).9 Jackman’s adjusted estimates put the ‘true’ figures of likely participation under a voluntary regime at 63 per cent (plus or minus 10 per cent) in Australia (1999: 40).10 Despite considerable uncertainty, there is a remarkable similarity in the overall conclusions that can be drawn from these studies. The findings reviewed here suggest that in both Belgium and Australia there would be a fall of at least 20 per cent and up to 40 per cent if compulsory voting were repealed. Conclusion This chapter has considered the influence of mandatory electoral participation on the most obvious of the factors it is expected to affect: aggregate turnout levels. As Hirczy (1994) points out, one cannot always extrapolate directly from cross-national variations to the impact of adopting compulsory voting in a given context. For this reason, both longitudinal and cross-sectional analysis was carried out here, and the results of counter-factual analyses were also considered. The three approaches to the problem have converged on similar findings: compulsory voting increases turnout. The longitudinal analyses conducted in this chapter estimate that the introduction and the abolition of compulsory voting were each ‘worth’ a turnout differential of approximately 13 per cent, whereas the cross-sectional analysis suggests that sanctioned compulsory voting makes a difference of about 12 per cent to aggregate rates of electoral participation – a remarkably similar figure, given the differences in the two analytic approaches. The counterfactual survey evidence from Australia and Belgium indicate that the repeal of compulsory voting laws would have a larger impact, but it is worth noting that the ‘novelty effect’ of such a change in the Netherlands resulted in a sudden drop, followed by a subsequent rise in participation rates. It may therefore be that the counter-factual studies overestimate the medium-term impact of the abolition of compulsory voting (see also Hill and Louth, 2004). As we saw in Chapter 2, compulsory voting was introduced in the aim of accomplishing a variety of goals, but among those goals increasing turnout invariably figured prominently. It is therefore not surprising that the existing

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 97

Compulsory voting and electoral turnout

97

literature on this institution has focused primarily on its impact on rates of electoral participation. This focus has been that adopted also in this chapter, and the overall conclusions of the analyses conducted here are not particularly novel. The main contribution of the chapter has been to examine the question at hand from several different methodological angles, and to test the impact of compulsory voting on as large a number of cases as possible, so as better to understand both the extent and the limits of the empirical relationship between this institution and turnout. The next chapter brings us back to the main focus of this volume, which is the influence of compulsory voting beyond turnout. Chapter 2 demonstrated that in addition to increasing turnout, one of the most common motives for adopting this institution was to improve the quality of elections, understood in terms of electoral integrity and legitimacy. This raises the issue of potential compromises involved in any act of institutional design. It may, for example, be that there are trade-offs to be made between the quantity of participation (understood in terms of turnout) and the quality of that participation. Alternatively, good things may go together, and an institution that increases turnout may also be beneficial in other ways. The following chapter will therefore be devoted to an assessment of the relationship between compulsory voting and electoral quality. Notes 1 Blais (2006) and Geys (2006) provide useful overviews of this research. 2 Data are unavailable for parliamentary elections prior to 1963. 3 The evidence provided by Franklin was from only two cases – the Netherlands and Italy (the latter is not classified here as having had compulsory voting). In both cases the change observed was the abolition, rather than the introduction, of electoral compulsion. It must be admitted that Franklin’s analysis includes controls for a variety of other factors, while the present analysis does not do so for the reasons indicated above. 4 On the ‘crisis’ of electoral participation in Venezuela, see Ellner (1993/4); Lapp (2004); Molina and Thibaut (2005). 5 Regional analyses that have found such an effect include those by Franklin (1999: 210); Anduiza Perea (2002: 657); Siaroff and Merer (2002: 919) (Europe); and Fornos et al. (2004: 923) (Latin America). 6 Excluded from these calculations are Austria 1990, Switzerland and Egypt, which have fractional values on the compulsory voting variable; see the Appendix for details. 7 Lack of significance of the margin of victory variable is undoubtedly due to the estimation strategy employed here, which averages variables across elections; this strategy will perforce reduce the impact of ‘conjunctural’ factors such as this, which exhibit variation from one election to the next. In a model including only the most recent elections (not shown), this variable is highly significant, as expected, but the other variables are hardly altered.

M1546 - BIRCH TXT.qxp:ANDY Q7

98

27/10/08

11:31

Page 98

Full participation

8 The survey item was ‘Would you have voted in the election if voting had not been compulsory?’. 9 de Winter and Ackaert (1994: 83) acknowledge the drawbacks of counterfactual surveys on similar grounds. 10 Unusually, Australia held the election to its 1998 Constitutional Convention under a voluntary voting regime. Turnout in this poll was only 53.5 per cent (Orr, 1998: 578), but this can be accounted for by a variety of factors, including the fact that this was an election to an atypical body, the election was conducted by post rather than in the usual manner, and the electoral system employed was unfamiliar (Orr, 1998: 577–9). It is thus not clear that this electoral event provides an indication of likely turnout for general elections under a voluntary rule.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 99

6

Compulsory voting, electoral integrity and democratic legitimacy

Chapter 5 confirmed the well-established tendency of mandatory voting regimes to increase rates of electoral participation, but the ‘amount’ of participation is not necessarily the only relevant aspect of voting in a democracy. It is also of interest to consider the impact of compulsory voting on the meaningfulness of the electoral process. This is because in a modern democracy any proposed innovation to increase electoral participation must concern itself not only with the quantity of electoral participation but also its quality. Quality can be understood here in two senses: both the integrity of the electoral process – how ‘free and fair’ elections are – and the meaningfulness of vote expression. In connection with the second understanding, it is also worth considering the impact of this institution on the legitimacy of the democratic process as a whole. These are the aims of this chapter. The first understanding of electoral ‘quality’ is the extent to which elections held under a compulsory voting regime are free from manipulation and abuse. If such measures increase turnout but at the same time raise the level of electoral fraud, vote-buying, and other forms of malpractice, then most would agree that they are not worth the price. We saw in Chapter 2 that compulsory voting was often introduced in order to improve the conduct of elections. This chapter will investigate the extent to which the institution achieves this end. Are elections held under compulsory voting likely to be associated with higher standards of electoral integrity than those held under voluntary regimes? Which aspects of electoral integrity are likely to be particularly affected by compulsory voting? The second understanding of electoral quality relates to question of popular legitimacy. Questions relevant in this context include the impact of the institution on the proportion of invalid and arbitrary votes, as well as its effect on popular perceptions of the quality and legitimacy of the democratic process. The first section of the chapter will thus address general considerations relevant to the relationship between compulsory voting and the quality of electoral administration. This will be followed in the second section by evidence from case studies, with a particular focus on the cases of Belgium and

M1546 - BIRCH TXT.qxp:ANDY Q7

100

27/10/08

11:31

Page 100

Full participation

Australia. The third section will undertake a large-N cross-sectional analysis of the impact of compulsory voting on electoral integrity. The fourth section of the chapter will turn attention to the effect of compulsory voting on invalid ballots and arbitrary votes, while the fifth section will consider the impact of this institution on perceptions of democratic legitimacy. The empirical analyses undertaken in this chapter will not provide clear-cut answers to all of the questions addressed: though there are convincing arguments that compulsory voting ‘ought’ to have a variety of effects on different aspects of electoral probity, the evidence available does not allow us to come to any firm conclusions about the overall influence of this institution on the quality of elections understood in this sense. The second finding is that compulsory voting can lead to a moderate increase in the proportion of blank, invalid and arbitrary votes cast, but that in most contexts this does not represent a serious challenge to the legitimacy of the democratic system. Finally, there is some evidence to suggest that compulsory voting has a positive impact on overall democratic legitimacy. Electoral integrity: definitions and categorisations Electoral integrity is a matter of the extent to which the proper conduct of elections is ensured by the electoral administrative apparatus. It therefore makes sense to pause to examine its implications for electoral administration in general. The introduction of compulsory voting gives added responsibility to electoral authorities to enable every citizen to vote. Whereas the right and the duty to vote are largely matters of institutional design, the opportunity to vote reflects the extent to which electoral administrative procedures facilitate voting. In order to realise both the right and the duty to vote, electors must have adequate opportunity, which in turn requires the electoral management body to ensure the supply of sufficient accessible polling places (or mechanisms) whose location is well-publicised in advance and which are available for use at convenient times.1 Ballots must also be designed and explained such that all sectors of the population can use them effectively without undue difficulty.2 From the start, compulsory voting was accompanied by vote-facilitating mechanisms. Quiri (1908: 76) reports that turn-of-the-century Belgium provided free rail transport to voters living away from home in order that they might exercise the franchise. Similarly, Gosnell (1930: 108) notes that vote facilitation was paired in early twentieth century Belgium by measures to ensure the integrity of the electoral process: ‘one reason for the high participation in Belgian national legislative elections is the convenience of the voting system. [. . .] The election process itself is safeguarded in every possible way’. In modern states vote facilitation is also taken quite seriously under compulsory

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 101

Compulsory voting, electoral integrity and democratic legitimacy

101

voting, with devices ranging from discounted travel fares to elaborate remote voting provisions. Given that making voting easier can at times entail additional risks (especially when remote voting is involved), precautions are necessary to minimise abuse. Finally, where compulsory voting is enforced by means of sanctions, an effective system needs to be put in place to identify infringements and to apply the sanctions in question. It goes without saying that the combination of vote-facilitation mechanisms and sanctions regimes places considerable burdens on the electoral administration of countries that practice enforced compulsory voting. This additional administrative demand can be expected to have different consequences under different circumstances. Where administrative capacity can be met by the state, compulsory voting ought to work to professionalise electoral conduct. Simon Jackman comments that ‘one (perhaps) unintended consequence of an aggressive system of CV [compulsory voting] is the centralization and professionalization of election administration. [. . .] This many mitigate the problems that accompany decentralized and non-professional election administration, clearly evident in the aftermath of the 2000 US presidential election’ (2001: 16317). Yet in cases where administrative capacity is lacking, we might expect to observe the opposite result. Indeed, Maria Gratschew notes that ‘The cost of enforcement [of compulsory voting] may lead some electoral administrations to lower their standards of enforcement’ (2002: 106). Provisions enabling voting may also be difficult for states with weak capacity to ensure effectively. For example, in El Salvador where voting is technically compulsory, Krennerich (2005: 278) reports that many voters do not receive the electoral card without which they cannot vote. The overall impact of compulsory voting on standards of electoral administration is difficult to predict. It is therefore worth examining the likely effect of this institution on different aspects of electoral integrity. We must first distinguish between ‘benign’ deficits in electoral conduct such as unintended procedural problems, and malefic violations of electoral integrity such as fraud, manipulation, and various other forms of electoral corruption. Though the distinction may not in all cases be easy to make, it is an important conceptual difference. Elections are extremely complicated and large-scale logistical events, and in any even moderately large polity there are bound to be ways in which their conduct fails, for purely logistical reasons, to live up to the ideal laid out in the electoral law (Mozaffar and Schedler, 2002). The introduction of compulsory voting will by definition increase the level of electoral violation, if this concept is understood in a narrow legal sense, for the simple reason that, by introducing an additional regulation, compulsory voting adds an additional regulation to be violated.3 Similarly, compulsory voting generally entails compulsory registration, and if electoral registration

M1546 - BIRCH TXT.qxp:ANDY Q7

102

27/10/08

11:31

Page 102

Full participation Table 6.1 Dimensions of electoral integrity

Inclusion of eligible actors in the electoral process

Vote preference formulation and expression

Vote aggregation aggregation

Adjudication of disputes

– Administrative inclusion of candidates/ parties – Administrative enfranchisement of electors

Regulation of the electoral campaign: – regulations governing polling facilities – regulations governing vote expression

Regulations governing: – boundary delimitation – vote counting, tabulation and publication of the results – conversion of votes into seats

– Administration of adjudication procedures – Imposition of sanctions

has not previously been compulsory, its introduction will invariably lead to a certain number of violations of that requirement as well. These are, however, somewhat trivial relationships. The purpose of this analysis is to explore the way in which mandatory voting impacts on other aspects of electoral integrity. The first issue we have to address in this context is how electoral integrity is to be defined. In the broad sense, electoral integrity can be taken to include questions of institutional design; certain types of rules and regulations may be held to violate international electoral standards, and/or the democratic rights guaranteed by the state in question. But it is electoral integrity in the narrower sense of rule implementation, or electoral conduct, which will concern us here. Integrity in the conduct of elections has four main dimensions, set out in schematic form in table 6.1. The first dimension is that of inclusion: all eligible actors – electoral, candidates, campaign activists, and election observers – must have their rights of participation guaranteed according to the law, and they must all be provided with equal opportunities to participate in the activities legally open to them. Vote preference formation and expression is the second dimension. Eligible electors must have the opportunity to assess electoral options under conditions of fairness and equality, they must have ample and equal opportunities to express their vote choice, and the expression of their choice must not be affected by undue influence. The third dimension is vote aggregation: votes must be aggregated in an accurate and transparent manner, and the result of this process must be reported accurately within a reasonable period of time following the vote. Finally, electoral disputes must be adjudicated in a timely and unbiased manner according to the law.4

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 103

Compulsory voting, electoral integrity and democratic legitimacy

103

Electoral integrity: evidence from case studies This typology provides a basis from which to identify the possible influence of compulsory voting on electoral integrity. In many cases it is difficult to identify any specific impact, despite the fact that the prevention of abuse was one of the motives for introducing this measure. In Argentina, for example, pre-reform elections had been characterised by fraud and manipulation so serious that elections were effectively non-competitive (Díaz, 1983: 30, 54). The Sáenz Peña reforms that introduced compulsory voting as part of a package of measures (see Chapter 2) did not eliminate electoral irregularities, but they are widely recognised as having ushered in the age of modern electoral competition. Yet due to the wide-ranging changes that occurred in the Argentine political system at this time, it is difficult to attribute the reduction in abuse to compulsory voting per se. In Thailand, where making voting mandatory was also seen as a means of reducing abuse, it is likewise hard to see evidence that this measure – or the others with which it was packaged – were successful in this regard. Commenting on the 2005 Thai elections, the Asian Network for Free Elections observation report notes that ‘structural obstacles embedded in traditional Thai politics’ had prevented a notable decrease in irregularities (ANFREL, 2005: 6). Because of the difficulties associated with tracing the precise influence of compulsory voting on electoral integrity in many cases, the discussion that follows will focus mainly on Belgium and Australia, the cases for which there is the greatest amount of evidence. Inclusion of eligible actors in the electoral process As noted above, compulsory voting logically necessitates compulsory electoral registration (though Chile is a curious case where voting is compulsory but electoral registration is not).The bureaucracy required to maintain a full electoral roll may lead to greater resources being devoted to electoral registration, with the consequence that it is better performed. A more up-to-date and accurate electoral register should reduce the opportunities for electoral fraud, as will a better-resourced administrative apparatus. Furthermore, post-electoral checks on non-voters could potentially flush out inaccuracies in the register. If the electoral administration is efficient, those who are recorded as not having voted will be chased for fines, and if these individuals are found not to exist – or are found to have their details wrongly recorded – this information can be used to correct the electoral register. Finally, in as much as selective facilitation of voting is a means whereby electoral authorities have traditionally manipulated elections, compulsory electoral participation ought in theory to enhance access to the polls for individuals desiring to vote.5 Yet if voting is obligatory, those wishing not to vote are likely to avoid registration. We can therefore expect to see more violations of registration

M1546 - BIRCH TXT.qxp:ANDY Q7

104

27/10/08

11:31

Page 104

Full participation

regulations under compulsory voting than is the case where voting is voluntary. In addition, because personation is less likely to be a viable strategy under compulsory voting (see below), efforts to fiddle electoral results may well focus on the manufacture of ‘dead souls’. ‘Register-stuffing’ could replace ballot-box stuffing. There is little reason to believe that other aspects of inclusion – the inclusion of candidates and other actors – would be affected by compulsory voting requirements. In Belgium the quality of the electoral register appears to be high overall, and it has improved significantly since the introduction of a rolling register in 1976 (Ackaert et al., 1992: 211). It is of course difficult to say whether this high quality is because of, or despite, the existence of compulsory voting, but it is undoubtedly due in part to the fact that electoral registration is initiated by the state, rather by the citizen. The situation in Australia appears to be somewhat different, however, as here electoral registration is initiated by the voter. Hughes (1966) notes that one of the arguments in favour of mandatory voting in Australia at the time of its introduction in 1924 was that it would help check on the accuracy of the electoral rolls (Hughes, 1966: 82). However, Mackerras and McAllister cite a parliamentary committee report from 1994 in which there was a minority view that compulsory voting encouraged electoral fraud, due to the fact that it required the facilitation of voter registration, which meant that the registration process included insufficient checks (Mackerras and McAllister, 1999: 223). They note that the principal limit to fraud is a generally low level of political interest on the part of the electorate, suggesting that under different circumstances, voter-friendly registration procedures could lead to abuse: ‘Because voting is compulsory, politicians and electoral officials have gone to considerable lengths to make the system voter friendly. Indeed, Australia is probably the most voter-friendly country in the world. If Australians felt more strongly about election results than they do then there probably would be more abuse of the system’ (Mackerras and McAllister, 1999: 223). The same authors report that levels of registration in Australia are lower than those in other democracies, a phenomenon they link to compulsory voting (1999: 219). This discussion suggests that the legal obligation to participate in elections may have an effect on the inclusion of voters on the electoral register, but that this is only likely to be a significant impact where registration is voter-initiated, as is the case in Argentina and Chile. In the vast majority of contemporary mandatory voting states, by contrast, the state takes the initiative to register its citizens to vote. Vote preference formulation and expression As argued in Chapter 4, campaigns in countries with voluntary voting are as much about getting out the vote (mobilisation) as they are about converting

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 105

Compulsory voting, electoral integrity and democratic legitimacy

105

voters. Mobilisation ceases to be such an issue under compulsory voting, and parties have an incentive to focus their energies on converting those least interested in (and probably least informed about) politics, who most likely would have been non-voters under voluntary rules. This provides an incentive for parties to supply selective benefits to non-politically aware (unattached or apathetic) voters who do not care so much about their vote and are not so concerned to maintain its integrity. Emphasis can therefore be expected to switch from mobilisation (when voters may be given selective incentives to get to, or to stay away from, the polls) to vote-buying (where voters are given selective incentives to vote for a certain party). We might therefore anticipate a rise in electoral clientelism under compulsory voting, if social conditions are favourable to this phenomenon. Political apathy on the part of reluctant voters under mandatory voting may also lead to an increase in the exchange of votes for particularistic rewards. Abraham poses the rhetorical question: ‘Which of the following voters would be more likely to accept an ‘honorarium’ for his vote: He who goes [to the polls] of his own free will and choice or he who goes grudgingly, feeling that he is giving time for which he ought to be recompensed? Certainly the latter!’ (1955: 9; cf Abraham, 1952: 201). Though mandatory voting was introduced in Thailand precisely to combat the purchase of votes, the measure has not been successful in this respect, as noted above, and it may be speculated that this is at least in part due to the alteration in the incentive structure induced by a reduction in opportunities for selective mobilisation and the increased availability of apathetic voters. Reluctant voters may also be somewhat more ‘careless’ with their votes, and less vigilant in ensuring that they are properly cast; they may even be willing to collude in acts of personation and other forms of electoral fraud, to save them the bother of voting. At the same time, the fact that parties do not have to mount expensive mobilisation campaigns often means that their overall costs are less (Lijphart, 1997; Gratschew, 2002: 105–6). Given that the need for campaign finance is one of the motors of political corruption of various sorts, a reduction in such costs could lead to an overall reduction in corrupt activities. A reduced need for money on the part of political parties should translate into a reduced tendency to resort to rent-seeking and other forms of illicit activity. As far as preference expression is concerned, compulsory voting has the potential to decrease the likelihood of personation, because known non-voters’ names cannot be borrowed for this purpose undetected. If everyone goes to the polls, personation is far more likely to result in detection, and this fact may well have a deterrent effect, regardless of how vigilant voters are inclined to be on partisan grounds. In practice there appears to be variation in the extent to which mandatory voting has been associated with the development of electoral clientelism.

M1546 - BIRCH TXT.qxp:ANDY Q7

106

27/10/08

11:31

Page 106

Full participation

It is well-known that Belgium has experienced levels of electoral clientelism equalled in Western Europe only by Italy, where there were long sanctions for non-voting (e.g. Rihoux, 1996: 272; Deschouwer and de Winter, 1998). Though the evidence linking clientelism to compulsory voting is circumstantial, and a wide range of other factors undoubtedly come into play, it may well be that Belgian voters’ obligation to go to the polls has put pressure on parties to retain weakly politicised members of the electorate through the provision of selective incentives, and that this institution has served to help preserve a form of political linkage that has social origins. It is perhaps not surprising that it has generally been the smaller, less ‘pillarised’, parties, that have advocated the abolition of compulsory voting in Belgium (Rihoux, 1996: 258–9). The phenomenon of electoral clientelism does not appear to be prominent in Australia, however, and the difference between the two countries is most probably due to the fact that Belgium is a far more pillarised society than Australia, so the conditions for clientelism are more favourable in the former. In Australia there is evidence to suggest that compulsory voting has lowered campaign costs, as predicted above (Hughes, 1966; McAllister, 1986; Aitkin, Jinks, and Warhurst, 1989). Hughes compares overall party expenditure levels in Australia with those of other established democracies, and finds that Australian parties spend less on average than their counterparts elsewhere (Hughes, 1966: 94). Belgium has, by contrast, the third highest level of campaign expenditure of ten major Western European countries for which data are available (Farrell and Webb, 2000: 114).6 It was not possible to obtain evidence on the frequency of personation in either Belgium or Australia (such data are, in any case, anecdotal at best, even when they do exist). The hypotheses with regard to this form of electoral malpractice can therefore not be tested. In sum, this discussion indicates that there is some evidence of a switch in the use of particularistic incentives from mobilisation to vote choice as a result of the introduction of compulsory voting, but only where the social context is conducive to such practices. Clientelistic societies, and particularly those with patronage-based party ‘machines’ are potentially susceptible to an increase in vote-buying following the introduction of mandatory electoral participation laws, whereas full participation should reduce the need for material resources in societies where vote choice is based more on party programmes and public goods that benefit society at large. Vote aggregation and adjudication of disputes It is unclear that compulsory voting will have any impact on vote aggregation, except in as much as it tends to necessitate a larger and therefore more professional electoral administration. Likewise, there will be more disputes to adjudicate under compulsory voting (which may put pressure on the courts and other

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 107

Compulsory voting, electoral integrity and democratic legitimacy

107

agencies involved in dispute adjudication), but otherwise any effect of this institution on dispute adjudication is difficult to conceive. No clear-cut evidence was found that mandatory participation was associated in any meaningful way with the quality of either vote aggregation or the adjudication of disputes. The case-study evidence for the impact of compulsory voting on electoral conduct is therefore somewhat ambiguous; the institution would appear to have a number of countervailing effects, such that its net impact depends on the context in question. Which of the relevant influences dominates as a rule is a question that can potentially be elucidated by means of large-N statistical analysis. Electoral integrity: evidence from aggregate data To this end, a large-N test of the impact of compulsory voting on electoral integrity was conducted using the 2006 Freedom House ‘Electoral Process’ scores for each state. In 2006, Freedom House began to release the component scores of their overall rankings of states in their annual Freedom in the World survey, including the ‘electoral process’ variable, which provides a useful indicator of electoral integrity.7 The indicator ranges from 0 to 12, where higher scores indicate ‘cleaner’ elections. This variable was matched to the most recent election in the cross-national dataset employed in Chapter 5. A variety of control variables known or suspected to be associated with electoral integrity were included in multivariate analysis. These include corruption, as proxied by the Transparency International Corruptions Perception Index score (corruption in the public sector is expected to be associated with electoral manipulation and abuse); per capita GDP (poorer states are expected to experience more problems with electoral irregularity); a proportional representation electoral system (found in previous elections to be associated with electoral probity); and concurrent presidential elections (also found in previous research to be associated with electoral conduct).8 Table 6.2 presents multivariate regression models of determinants of the quality of the electoral process. Though the control variables behave largely as expected, with less corrupt, richer countries and those with proportional representation electoral systems holding fairer elections, compulsory voting does not appear to have a significant effect on electoral quality in any of its guises – compulsory voting alone, sanctioned compulsory voting, or a constitionalised electoral obligation. This null finding may be due to the fact that the various impacts of compulsory voting on electoral integrity cancel each other out, as suggested in the previous section. It was unfortunately not possible to control in this model for some of the contextual variables believed to modulate the influence of mandatory electoral participation on electoral conduct, such as clientelism and state

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

108

11:31

Page 108

Full participation Table 6.2 OLS regression models of the determinants of electoral integrity

Variable

Model 1

Compulsory voting Sanctioned compulsory voting Electoral obligation constitutionalised Controls: Per capita GDP (log) Corruption Perceptions Index Proportional electoral system Concurrent presidential elections

Model 2

Model 3

0.789 (0.89) ⫺1.41 (1.17) 0.91 (0.88) 0.47 (0.33) 0.54* (0.32) 0.47* (0.32) 0.65*** (0.24) 0.66*** (0.24) 0.66*** (0.24) 1.57*** (0.62) 1.69*** (0.61) 1.48** (0.64) 1.14 (0.81) 1.55** (0.79) 1.04 (0.82)

Constant

0.24 (1.84)

⫺0.21 (1.79)

0.19 (1.80)

N: Adj. R2:

149 0.31

148 0.31

149 0.31

* p ⬎ .10; ** p ⬎ .05; *** p ⬎ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

administrative capacity; nor was it possible, with the data available here, to test for the sub-types of electoral conduct identified above. The results of this analysis are therefore somewhat indeterminate, but there is no clear evidence that compulsory voting has a strong impact one way or the other on the overall quality of electoral conduct in a state. Legitimacy: Invalid and arbitrary voting It was noted above that elections generate ‘benign’ problems as well as those resulting from intentional manipulation. Invalid and arbitrary voting can in some sense be seen to be just such phenomena. When voters are obliged to go to the polls, there will invariably be some who are genuinely indifferent between electoral options, or who resent electoral compulsion per se. The only practical option available to such voters is to cast a blank or invalid ballot,9 or to cast their vote arbitrarily for one of the options. Invalid ballots are a universal phenomenon that arises for a combination of reasons, but since the introduction of compulsory voting in the nineteenth century, commentators have frequently remarked on the fact that they appear to be greater in number under compulsory voting than is the case where attendance at the polls is voluntary. Arbitrary voting has been less studied, because it is more difficult to gather evidence for it, but it is held to be a relatively widespread phenomenon in some compulsory voting states. In Australia, where the legal obligation to participate in elections is accompanied by the compulsory indication of an exhaustive preference ranking on that country’s preferential (rank-ordered) ballots, arbitrary

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 109

Compulsory voting, electoral integrity and democratic legitimacy

109

voting often takes the form of the so-called ‘donkey vote’, where voters start at the top of the ballot paper with a ‘1’ and number the candidates consecutively as they work their way down the paper. In many other contexts, arbitrary voting tends to take the form of voting for the first (or last) option on the ballot paper, which gives an advantage to candidates or parties with names at the top of the alphabet when ballot listings are ordered in this manner. Such arbitrariness is often seen to be a pathology of compulsory voting, in that it concentrates support on an electoral option for non-political reasons, and in some cases the additional support thereby gained is sufficient to alter the outcome of the election. Both invalid and arbitrary votes may be considered to be vote expressions of inferior ‘quality’ to the casting of a valid vote for the candidate of one’s choice. This is due to the fact that they carry no precise partisan meaning, though in some cases they may have an impact on the result of the election. In 1996 Kim Beazley, Australian leader of the opposition, was elected on a vote margin which suggested that he could well have benefited from the so-called ‘donkey vote’ (King and Leigh, 2006: 16). If the introduction of compulsory voting radically increases the proportion of votes that are in this sense ‘meaningless’, this would call the legitimacy of the outcome seriously into question. This section therefore seeks to asses the prevalence of invalid and arbitrary voting in states with compulsory voting, and to assess the role of electoral compulsion in generating these results. Invalid votes There is considerable evidence to suggest that states with mandatory voting rules have higher levels of invalid votes that those where going to the polls is voluntary (Tingsten, [1937] 1963: 191; Hirczy, 1994: 72–3; Jackman, 2001; Gratschew, 2002: 106; 2004: 57; Reynolds and Steenbergen, 2006), undoubtedly due to the fact that some voters will intentionally spoil their ballots if they are not attracted to any of the vote options they find there. Indeed, some scholars have criticised the institution on precisely these grounds (Mackerras and McAllister, 1999). In a recent multivariate analysis of elections around the world, Reynolds and Steenbergen find that states with compulsory voting have on average 6.4 per cent more spoilt ballots than those with voluntary regimes, when other relevant factors are taken into consideration (Reynolds and Steenergen, 2006: 582). Reynolds and Steenbergen also point to the exceptionally high rate of ballot spoilage on the Latin American continent; though most Latin American countries have compulsory voting, a dummy variable for this geographic location still has a coefficient of 4.2 in their multivariate analysis (which controls also for compulsory voting). This phenomenon is clearly evident in recent elections in Argentina, Brazil, and Chile, all countries with (unsanctioned) mandatory electoral participation, in which casting a blank or invalid ballot has become a

M1546 - BIRCH TXT.qxp:ANDY Q7

110

27/10/08

11:31

Page 110

Full participation

form of political protest akin to abstention under many voluntary regimes. In Argentina this type of voting is termed voto bronca (‘rage vote’), and in the 2001 elections over 40 per cent of ballots were blank or cast for write-in candidates, due to the ‘crisis of representation’ afflicting the country during that period (Lapp, 2004: 8; Rodríguez et al., 2004). This was a relatively new phenomenon, in that in the1980s blank votes had almost always made up less than five per cent of the total, despite the considerable political instability of this period (Fraga, 1992: 20). Power and Roberts investigate the high and rising rate of invalid ballots in Brazil between 1945 and 1990, which topped 40 per cent in the late 1980s (Power and Roberts, 1995). They link ballot spoilage to a variety of factors, some social, some institutional and some political, and they note that this phenomenon is a functional equivalent of abstention in states with voluntary voting (1995: 803). Carlin (2006) investigates the correlates of invalid votes in Chile, and finds that disinterest in and dissatisfaction with the political system, together with younger age (interpreted in cohort terms), are the strongest individual-level determinants of casting an invalid ballot in that country (where registration is voluntary, but voting is mandatory for registered electors). Finally, Power and Garand (2007) conduct a multivariate analysis of the factors influencing invalid votes across the Latin American region, and find that institutional design feature (including compulsory voting) combine with socio-economic and political variables to determine the overall rate of vote invalidation. As these studies suggest, ballot spoilage results from a combination of factors, of which compulsory voting is just one. In Australia, simplification of the ordinal ballot (through the introduction of ‘above the line’ voting) led to a dramatic reduction of invalid ballots from an average of 8.9 per cent in the 1949– 1983 period to 3.6 per cent on average between 1984 and 1996 (Mackerras and McAllister, 1999: 225), and the incidence of invalid ballots has been linked to a variety of additional factors, including ballot paper complexity and socioeconomic structure (McAllister and Makkai, 1993). In Belgium, the gradual introduction of electronic voting from 1991 onwards appears to have considerably lowered the number of invalid votes in those areas where it is employed (Rihoux, 1996: 262; Fraeys, 2003: 381–390), whereas the sum of blank and invalid ballots had surpassed 7.5 per cent in Brussels in the 1970s, it fell to 3.6 per cent in 2003 in areas served by electronic voting (Fraeys, 2003: 392). We can thus conclude that electoral compulsion does indeed lead a minority of people to a situation where they are unable to make a choice. Yet, the number who cast blank ballots or who intentionally spoil their ballot papers in countries with mandatory electoral participation is notably lower than those who fail to vote in most countries with voluntary voting. The Dutch case is instructive in this regard. Invalid votes had never represented a large proportion in the Netherlands, but the advent of voluntary voting brought the toll down from 2.8

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 111

Compulsory voting, electoral integrity and democratic legitimacy

111

to 0.7 per cent of the total.10 By contrast, the decline in turnout in the election immediately following the abolition of compulsion was – as we saw in the last chapter – nearly 16 per cent. Arbitrary votes Arbitrary voting must be distinguished from voting for a minor party as an act of protest (a phenomenon that will be considered in Chapter 7). Arbitrary voting is typically studied in the academic literature under the rubric of ‘ballot order effects’, though referred to colloquially as the ‘donkey vote’, and known to affect vote choice in Australia (Mackerras, 1970; Kelley and McAllister, 1984). Jackee and Sun (2006) have provided formal proofs that under a firstpast-the post electoral system, random or arbitrary voting can alter electoral outcomes (though it must be noted that no contemporary mandatory voting state employs such an electoral system), and an empirical study by Brockington has found that ballot order effects are influenced in the US context by the amount of information in the electorate (Brockington, 2003). Examining a state with compulsory voting, Selb and Lachat have found that in Belgium, non-voluntary voters tend to cast votes less consistent with their own preferences than those who would vote even if electoral participation were not mandatory (Selb and Lachat, 2007). The introduction of random ballot ordering in Australia in 1984 provides a useful means of assessing the extent of arbitrary voting in that country. King and Leigh (2006) demonstrate that arbitrary decisions based on ballot order account for only about one per cent of the vote share in Australia, a lower figure than in many polities where voting is voluntary. This suggests that while real and sometimes politically significant,11 this potential effect of compulsory voting is relatively minor, and it is in any case difficult to ascertain that it is indeed a consequence of this institution. The elevation of invalid and arbitrary voting can in most cases be considered a minor defect of compulsory voting. The number of voters who choose to ‘waste’ their vote in this way is typically not so large as to delegitimise this institution. Moreover, it could be argued that the option of spoiling one’s ballot or casting an arbitrary vote is a normatively desirable means of expressing a political view. Certainly in Argentina and Brazil this practice has taken on a strong political dimension, and this type of behaviour may in many respects be seen as preferable to outright abstention, in that it sends a more direct and precise signal of intent. A frequently mooted alternative means of reducing invalid and arbitrary votes is an option such as ‘none of the above’, ‘against all’ or ‘abstain’ on ballot papers for voters who are disenchanted with all the options on offer and want a means of expressing that sentiment directly. Such a practice is available in a number of states in Eastern Europe and the former Soviet Union (though none of these has compulsory voting). This option will be considered in greater detail in the Conclusion.

M1546 - BIRCH TXT.qxp:ANDY Q7

112

27/10/08

11:31

Page 112

Full participation Legitimacy: perceptions of the democratic process

It is also worth exploring the impact of mandatory electoral participation on other forms of regime legitimacy. It was noted in Chapter 3 that one of the arguments against obliging citizens to attend the polls is that it addresses symptoms rather than underlying problems; by ‘artificially’ increasing rates of electoral participation it can obscure deeper societal disillusion with politics that is expressed in countries with voluntary voting through low turnout. At the same time, low levels of electoral participation under voluntary voting rules are often interpreted as an indication that the resulting government is lacking in legitimacy; full participation may be seen as a means of enhancing the legitimacy of the officials thereby elected. This question can addressed in purely normative terms, an exercise that was undertaken in Chapter 3, but the legitimacy issue can also be conceived in terms of popular perceptions of and support for the democratic system. This is the task of the present section. There is limited comparative evidence both for and against the proposition that compulsory voting enhances popular perceptions of political legitimacy. In a recent analysis of European Social Survey data from 2004, Krister Lundell finds that trust in politicians and trust in political parties is higher in the mandatory voting states of Belgium, Luxembourg and Greece than in European states where voting is voluntary, even when other relevant factors are controlled; trust in parliament is however, lower in such states. It is worth noting that in all three cases, the effects were very small but statistically significant (Lundell, 2007: 14–15). This proposition will be examined here by means of the Comparative Study of Electoral Systems dataset employed in Chapter 4. Two measures of system legitimacy are used: a standard question on satisfaction with democracy, and a question on the perceived quality of representation (see the Appendix for details of question wording and operationalisation).12 As far as satisfaction with democracy is concerned, an average of 51.7 per cent of survey respondents in the 34 CSES states in this dataset reported high levels of satisfaction with democracy, while the figure in voluntary voting states was, at 58.7 per cent, somewhat higher (though this difference is not statistically significant). The situation is reversed when it comes to sanctioned compulsory voting, however: an average of 61.4 per cent of respondents in states with this institution report high levels of satisfaction with democracy, as against 56.9 per cent of those in other states (again this difference is not statistically significant). This curious finding may well be due to he fact that most of the compulsory voting states in this dataset are located in Latin America, which has notably lower levels of satisfaction with democracy than other parts of the world (39.6 per cent of Latin American respondents are generally satisfied with the democracy in their countries, whereas the figure for non-Latin American states is 59.8 per cent).

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 113

Compulsory voting, electoral integrity and democratic legitimacy

113

Table 6.3 Multilevel logit models of satisfaction with democracy Variable Individual-level variables: Age Education Female gender Party identification

Model 4

Model 5

⫺0.00 (0.15) 0.28*** (0.11) ⫺0.00 (0.03) 0.34*** (0.05)

⫺0.00 (0.15) 0.28*** (0.11) ⫺0.00 (0.03) 0.34*** (0.05)

Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Per capita GDP (logged) Corruption Perceptions Index Latin America Constant

⫺0.14 (0.17) 0.02 (0.21) 0.25*** (0.07) ⫺0.63** (0.26) ⫺1.77 (1.74)

0.39*** (0.10) ⫺0.16 (0.17) 0.04 (0.21) 0.24*** (0.07) ⫺0.38* (0.22) ⫺1.88 (1.72)

Intercept variance Chi squared – fixed effects Chi squared – random effects

0.25*** (0.06) 2879.18 (10 df) 18.10 (1 df)

0.25*** (0.06) 1896.69 (10 df) 17.34 (1 df)

48,845 34

48,845 34

N: individuals N: elections

0.44*** (0.12)

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

This suggests that multivariate analysis is necessary to tease out the precise relationship between this variable and mandatory electoral participation laws. Table 6.3 reports the results of a multilevel model of satisfaction with democracy. Following the most common findings of previous cross-national studies of system legitimacy and regime support (Anderson and Guillory, 1997; Norris, 1999; Newton and Norris, 2000; Seligson, 2002; Anderson and Tverdova, 2003; Banducci and Karp, 2003; Dalton, 2004), control variables include at the individual level: age, education, gender, and party identification; and at the aggregate level: electoral system type, corruption, and per capita GDP as well as a control for Latin American location (see the Appendix for details of variable construction). It is clear from these results that compulsory voting has a strong and significant impact on satisfaction with democracy, once other factors (including Latin American location) are controlled; the residents of mandatory electoral participation states are happier with the way democracy works in their systems than those in states where voting is voluntary. Interestingly, sanctions for non-voting do not appear to enhance democratic satisfaction, suggesting that it is not the level of participation per se, but the act of collective self-binding to encourage participation that is relevant.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

114

11:31

Page 114

Full participation Table 6.4 Multilevel logit models of quality of representation

Variable Individual-level variables: Age Education Female gender Party identification

Model 6

Model 7

⫺0.21* (0.13) ⫺0.12 (0.10) 0.09*** (0.02) 0.53*** (0.04)

⫺0.21* (0.13) ⫺0.12 (0.100) 0.09*** (0.02) 0.52*** (0.04)

Election-level variables: Compulsory voting Sanctioned compulsory voting PR electoral system Per capita GDP (logged) Corruption Perceptions Index Latin America Constant

0.02 (0.18) ⫺0.15 (0.16) 0.13** (0.06) ⫺0.69** (0.35) 0.37 (1.22)

0.25 (0.22) ⫺0.00 (0.18) ⫺0.14 (0.16) 0.12** (0.06) ⫺0.42 (0.31) 0.26 (1.22)

Intercept variance Chi squared – fixed effects Chi squared – random effects

0.20*** (0.04) 326.11 (10 df) 21.52 (1 df)

0.21*** (0.04) 300.10 (10 df) 21.68 (1 df)

43,971 32

43,971 32

N: individuals N: elections

0.39* (0.26)

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01; Cell entries are logit coefficients (standard errors).

The second legitimacy variable considered is the perceived quality of representation. The bivariate relationships follow a similar pattern to that observed for satisfaction with democracy. A total of 46.6 per cent of respondents in compulsory voting states report that elections ensure that the views of voters are represented very well or quite well by majority parties, whereas the corresponding figure for voluntary voting states is, at 49.8 per cent, slightly higher. In states with sanctioned compulsory voting, however, 50.6 per cent of respondents believe that elections deliver high quality representation, while the figure for other states is a slightly lower 49.0 per cent. Table 6.4 reports the results of multivariate analysis of this variable, employing controls as in the above analysis of satisfaction with democracy. These models suggest a positive relationship between compulsory voting the extent to which elections deliver representation, but the link is far weaker than is the case for satisfaction with democracy; the coefficient for sanctioned compulsory voting fails to reach statistical significance (though it is larger than its standard error). This negative finding may perhaps be interpreted in terms of countervailing influences. Parties’ lack of incentive to mobilise voters under compulsory voting may reduce the extent to which they convey to voters their ability to

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 115

Compulsory voting, electoral integrity and democratic legitimacy

115

represent popular preferences; at the same time, generally satisfaction with democracy in such states may make respondents reluctant to criticise representative institutions. Overall, the findings of the CSES survey data analysis provide some evidence that compulsory voting is associated with higher levels of system legitimacy than voluntary voting, at least among those states included in this dataset. The impact of the institution on overall satisfaction with the democratic process is greater, however, than that on the quality of representation, which we might expect to be more closely linked to an electoral institution. It is not entirely clear why this might be the case, but we can speculate that more low-key campaigns in states with compulsory voting reduce partisan antagonisms and thus voter perceptions of efficacy and representation. At the same time, the inclusiveness of the institution of compulsory electoral participation may well explain the fact that survey respondents in states that practice it are generally more satisfied with the quality of democracy they enjoy. On balance, requiring citizens to attend the polls appears to have a positive impact on popular perceptions of system legitimacy. Conclusion Whereas Chapter 4 considered the impact of compulsory voting on participation from a ‘quantitative’ perspective; in others words – the role of compulsory voting in raising aggregate levels of voter turnout – this chapter has sought to assess several of the ‘qualitative’ aspects of this practice. In so doing it has revealed evidence that overall, legally obliging attendance at the polls has a number of countervailing impacts on electoral fraud and malpractice: it may increase the accuracy of the electoral register or it may lower it; it provides deterrents to personation, but also incentives; it lays the ground for political parties to engage in electoral clientelism, but at the same time it reduces their need for campaign finance. The overall impact of compulsory voting on electoral integrity will therefore depend on the political and social context of the polity in which it operates. We can also expect that the introduction of this institution in a state that did not previously have it is likely to shift the form malfeasance takes, because it alters the opportunities to interfere in the electoral process. The quality of electoral registers may be lower under compulsory voting, but only where registration is initiated by the voter. This finding suggests that states thinking of introducing compulsory voting should also consider changing to a system of state-initiated voter registration if they do not already have such a system. The second main finding is that states with socio-political conditions that favour electoral clientelism may discover that this practice is enhanced by compulsory voting. Though this finding is based on somewhat weak evidence,

M1546 - BIRCH TXT.qxp:ANDY Q7

116

27/10/08

11:31

Page 116

Full participation

it deserves further research and consideration. If this is the case, it would suggest that states with a history of electoral clientelism should think twice about introducing compulsory voting. The common wisdom that compulsory voting is associated with elevated levels of invalid and arbitrary voting also receives limited support, though details of electoral systems and the mechanics of voting are also strong conditioning factors. Perhaps the most relevant finding is that the numbers casting invalid and arbitrary votes is still relatively low in most compulsory voting states, which suggests that the existence of these phenomena should not be considered a strong argument against the institution. Indeed, a normative argument can also be made for substituting non-voting with some form of nonpreference expression at the polls, be it a blank or invalid ballot, or a vote for ‘none of the above’. The final relevant finding in this chapter is that the institution of mandatory electoral participation appears to have a positive impact on overall system legitimacy, as predicted by many proponents of the institution. This suggests that overall, compulsory voting has either a neutral or a positive impact on the quality of the electoral process in those states where it is practiced, but that the nature of this impact is conditioned by a variety of other social and institutional factors. Notes 1 In theory this implies that all eligible members of the community should be able to vote, regardless of their location. Though the Soviet Union went to the extreme of enabling astronauts in space to cast ballots, logistical and resource considerations will often restrict the extent to which universal availability is achieved. There is also a trade-off in some cases between quantitative and qualitative aspects of ballot provision; postal ballots (or some electronic equivalent) may be the only way in which isolated voters can exercise their franchise, but it is difficult if not impossible for the state to guarantee secrecy under these conditions. 2 This may require, for example, that ballots be multilingual, that special provisions be made for voters with disabilities, and that voter education campaigns be conducted prior to the election. 3 Non-voting appears to have been on the rise in some countries with compulsory voting, including (until very recently) Belgium (Ackaert et al., 1992; Rihoux, 1996: 271; Hooghe and Pelleriaux, 1998: 420; IDEA, 2002: 129) and Australia (LópezPintor and Gratschew, 2002: 127). 4 These requirements are drawn from international legal standards laid out in a variety of sources. For an overview see Goodwin-Gill, 1994. 5 In practice, however, this is not always the case, as the Salvadoran example cited above demonstrates. 6 Unfortunately, as noted in Chapter 4, there is insufficient data on campaign expenditure levels across the range of cases considered here for multivariate analysis to be carried out on the impact of this variable.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 117

Compulsory voting, electoral integrity and democratic legitimacy

117

7 The methodology subtending this score can be found at: http://www.freedomhouse.org/template.cfm?page=351&ana_page=333&year=2007. 8 For a fuller justification of the inclusion of these control variables see Birch, 2007. See the Appendix for details of variable construction. 9 These practices are technically legal in most states with compulsory voting; see Hill 2002b on the ambiguities of the Australian case. 10 Data are taken from the ‘Dutch Election Results Since 1918’ database at http://www.nlverkiezingen.com/index_en.html. 11 The authors identify five electoral contests since 1996 where candidates appear to have won due to ballot order effects (2006: 16). 12 There have been reservations voiced about the interpretation of satisfaction with democracy as a variable (Canache et al., 2001), but it nevertheless remains one of the most common indicators of democratic legitimacy. Moreover, the more specific wording of the ‘quality of representation’ question should provide a useful complement to the general nature of the ‘satisfaction with democracy’ item.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 118

7

Compulsory voting and political outcomes

So far we have been concerned mainly with the relationship between compulsory voting and aspects of the electoral process itself, including the attitudes and behaviour of citizens before and during voting. This chapter will turn instead to the impact of this institution on a variety of different ‘outcomes’ of elections. The range of possible outcomes to consider is large, and it is not possible within the space available in a book of this length to take them all into account. Attention will be focused on what are viewed as the most substantively relevant effects, as well as those that can intuitively be expected to result directly from making electoral participation mandatory. The issue of precisely who is elected constitutes the most obvious ‘outcome’ of the political process. There is a variety of different ways of understanding this ‘who’, depending on how representation is conceptualised. The two most common understandings of representation are descriptive and partisan: which types of people are elected, and what is the distribution of political support in the body constituted as a result of the election. The other main way in which political outcomes can be viewed is in terms of the ‘what’ of the political process: what are the concrete political outcomes generated by an elected body, and what impact do they have on society at large? Bearing these distinctions in mind, three types of electoral outcome variable will analysed in this chapter, two bearing on the ‘who’ and one concerning the ‘what’. We will first consider descriptive representation, specifically the impact of mandatory electoral participation on the proportion of women elected to national parliaments; secondly the impact of this practice on the distribution of partisan support, and on the consequences of this institution for the electoral fortunes of small parties, far-right parties and the balance between left and right forces in national parliaments; and thirdly the fairness of political outcomes, as measured by income equality and levels of corruption. The findings of this chapter are mainly negative; compulsory voting does not seem to have any significant impact on the proportion of women who gain seats in national parliaments, nor does it appear systematically to affect the distribution of support among parties – be they big or small, moderate or extremist,

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 119

Compulsory voting and political outcomes

119

left or right. The two results that do stand out in this chapter are the evident tendencies of mandatory attendance at the polls to reduce both income inequality and corruption. This suggests that a number of the fears of the opponents of compulsory voting are somewhat ill-founded, and that the institution has considerable benefits in terms of the substantive results of the electoral process. Female representation Female representation has been increasing dramatically across the world in recent years as cultural changes have led to a greater emphasis on the role of women in politics, particularly on the part of inter-governmental organisations involved in electoral monitoring and assistance in many countries. The inclusion of targets for gender representation in the UN Millennium Development Goals agenda represented an additional spur to the promotion of women to seats in national parliaments. The substantive importance of this aspect of representation suggests that is an appropriate object of investigation. Methodological concerns are also relevant in this connection. Of all the aspects of descriptive representation that one might be interested in considering, female representation is the most straightforward to compare across countries, and the attribute for which comparative data are most readily available. Compulsory voting can be anticipated to have a positive influence on female representation for two reasons. Firstly, in many socio-political contexts, women vote less frequently than men. This is especially true in the developing world, where traditional cultural roles inhibit women from becoming involved in politics. Full participation can be expected to favour female candidates in as much as the additional women mobilised to vote under such a regime may be more likely to cast their votes for their fellow women. The second reason to anticipate a positive impact of compulsory voting on female representation is that encouraging women to vote may have knock-on effects for other forms of participation, including standing for election. The bivariate relationship between these two variables does little to confirm this hypothesis, however. The mean proportion of women in parliaments with compulsory voting is 12.5 per cent, while the equivalent proportion in voluntary voting states is an almost identical 12.3 per cent. In states with sanctioned compulsory voting, this figure actually drops to 11.0 per cent, as against 12.5 per cent on average in other states. In states where the obligation to participate in elections is constitutionalised, the mean proportion of women in national parliaments is, at 13.4 per cent, slightly higher than the 12.1 per cent in those with no such constitutional provision. These differences are slight, however, and none of them is statistically significant. Multivariate analysis can determine whether other factors may be masking a significant relationship. The comparative literature on female representation

M1546 - BIRCH TXT.qxp:ANDY Q7

120

27/10/08

11:31

Page 120

Full participation

has identified a range of institutional, socio-economic and cultural variables that predict the proportion of female parliamentarians in a state. As in previous analyses, the focus here is on those variables most commonly found to be significant in large-N comparative analyses. Institutional factors that have been widely found to be positively associated with women’s representation include gender quotas (either formalised in the legal framework governing elections, or operated by political parties) (Yoon, 2001; Norris, 2006), as well as proportional representation electoral systems, particularly those of the list-PR variety (Norris, 1985; 206; Rule, 1987; Matland, 1998; Kenworthy and Malami, 1999; Reynolds, 1999; Yoon, 2001; Schwindt-Bayer and Mishler, 2005). High levels of socio-economic development are also known to boost the number of women elected to parliament (Matland, 1998; Kenworthy and Malami, 1999). Finally, cultural factors related to the social standing of women, measured by variables such as levels of female education and workforce participation have been found to be linked to levels of female representation as well (Rule, 1987; Matland, 1998; Keworthy and Malami, 1999; Reynolds, 1999). A modified version of the dataset employed in Chapter 5 is used here to test the hypothesis that compulsory voting should increase female representation. Information on female representation from the Inter-Parliamentary Union Parline database provides a useful source of data on the proportion of women in lower houses of parliament at each election. Following the studies referred to above, control variables include the existence of a gender quota, electoral system type, female education rates, female labour force participation, and per capita GDP (logged). Because accurate data on gender quotas are only available for the most recent elections, only these elections are included in the models constructed. Table 7.1 presents the results of this analysis. These models tell a story similar to that suggested by the bivariate relationships. None of the three variables employed as measures of mandatory voting comes even close to exhibiting a statistically significant relationship with the female share of legislative seats. We are therefore obliged to conclude that, contrary to expectations, the legal obligation to participate in elections does not affect levels of female representation. Distribution of partisan support Unlike the impact of compulsory voting on female representation, the influence of this institution on the distribution of partisan support in elected assemblies is a topic that has received considerable attention from academics, political commentators and practicing politicians since the earliest use of mandatory voting laws. Two aspects of this subject have been particularly prominent in debates: the impact of mandatory voting on right versus left forces, and its

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 121

Compulsory voting and political outcomes

121

Table 7.1 OLS regression models of the impact of compulsory voting on female representation Variable Compulsory voting

Model 1

Model 3

⫺1.32 (2.12) ⫺2.11 (2.77)

Sanctioned compulsory voting Electoral obligation constitutionalised Controls: Gender quota Proportional electoral system Per capita GDP (log) Female education

Model 2

⫺0.35 (0.87) 4.85*** (1.63) 3.60** (1.68) 1.34*** (0.51) ⫺0.02 (0.15) 0.34*** (0.09)

5.03*** (1.62) 3.54** (1.65) 1.41*** (0.51) ⫺0.02 (0.15) 0.33*** (0.09)

4.71*** (1.63) 3.48** (1.71) 1.33*** (0.51) ⫺0.03 (0.15) 0.35*** (2.16)

Constant

⫺11.74 (7.39)

⫺12.02 (7.35)

⫺11.46 (7.40)

N: Adj. R2:

140 0.24

139 0.25

140 0.24

Female labour force participation

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

implications for the seat share of minor and extremist parties, particularly those of the far right. The literature on this topic has almost all been countrybased, however; there has been little systematic comparative investigation of the impact of compulsory voting on the relative fortunes of political parties. This section will consider the relationship between this institution and the electoral success of small and far-right parties, as well as the overall left-right balance of forces in parliament. We deal here with an area that touches on fears often raised by proposals to make electoral participation compulsory. Claims that involuntary voters will support minor or far-right parties are commonly used as an argument against this institution (Makerras and McAllister, 1999: 229), as are allegations that it will disrupt the overall balance of party support. Small parties If, as was suggested in Chapter 6, more people are inclined to cast random votes when voting is mandatory, this would have the effect of increasing the vote

M1546 - BIRCH TXT.qxp:ANDY Q7

122

27/10/08

11:31

Page 122

Full participation

share of small parties. The same is even more likely to be the case if reluctant electors intentionally cast votes for minor parties in protest at having been obliged to attend the polls. These considerations suggest that mandating full participation could well inflate the vote share of minor parties. If small parties did indeed benefit from compulsory voting, this would in many contexts have the effect of fragmenting parliamentary party systems and possibly making stable governments more difficult to form and sustain. This might therefore be viewed as a possible negative consequence of compulsory voting laws. But there are also reasons for believing that the opposite effect might result from such rules. Inattentive voters brought to the polls under compulsory schemes could well be expected to cast votes that reflect their familiarity with the various options on the ballot. If this were the case, they would be more likely to favour larger parties, with which they would gain greater familiarity through casual contact with political media coverage and through the campaigning strategies of the parties themselves (assuming that larger parties are better positioned to run stronger campaigns – an assumption that may not always be true, but is most likely a valid assumption in the majority of cases). If this were the case, then compulsory voting would serve to inflate the vote share of large parties, at the expense of smaller organisations. There are thus reasons for believing that the legal obligation to vote could both increase and depress vote totals for small parties; were random and protest voting higher under compulsory voting, we might expect small parties to do better, whereas if the voting patterns of reluctant voters is dominated by the supply of information, large parties would benefit. It is of course also possible that ‘supply side’ and ‘demand side’ factors would cancel each other out, in which case full participation might have a negligible net impact on the electoral success of small parties. This is of course an empirical question that can only by answered by recourse to data. There is some comparative evidence on this question, however; simulations by Bernhagen and Marsh (2007) of the impact of full turnout on partisan support suggest that small parties would benefit from higher rates of electoral participation. Likewise in the Australian context, Jackman claims that small parties are better off under compulsory voting (1999: 43), a conclusion supported by Bélanger (2004). Makkeras and McAllister corroborate this view when they assert that ‘Minor or protest parties benefit [from compulsory voting] because high turnout also mobilises disproportionately more swinging and uncommitted voters, who have usually defected from the major parties’ (Makerras and McAllister, 1999: 229). At the same time, they also suggest that this impact may be outweighed by the benefit compulsory voting gives to large parties, as ‘compulsory voting fosters widespread loyalties toward the major parties across the electorate, thus making it much more difficult for independent and minor party candidates to attract the electoral support they require for successful election’ (Mackerras and McAllister, 1999: 230).

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 123

Compulsory voting and political outcomes

123

The bivariate relationships in the cross-national dataset used here suggest that small parties suffer in states where electoral participation is mandatory. In multiparty elections globally between 1990 and 2006, an average of 10.3 per cent of the vote went to parties polling less than five per cent in compulsory voting states, as against 11.4 per cent in states with voluntary electoral participation. In states with sanctions for non-voting, the relevant figure is 8.9 per cent, as opposed to 11.5 per cent in all other states (the first of these differences is not significant; the second is significant at the .05 level). This relationship is best explored through multivariate analysis, for which the database presented in Chapter 4 – pooled data on multiparty legislative elections held throughout the world between 1990 and 2006 – proves convenient. As in the analyses carried out in previous chapters, the models constructed here are based on country means for the period under consideration. Control variables are included for a variety of factors known or suspected to be associated with party system configuration, including single-member district electoral systems (Taagepera and Shugart, 1989 Lijphart, 1994; Cox, 1997; Katz, 1997), levels of ethnic fractionalisation (Ordeshook and Shvetsova, 1994; Cox, 1997), level of democracy (Birch, 2003), and population size (on the grounds that small countries are less likely to be able to support large numbers of small parties). The analyses presented in table 7.2 show that, contrary to the hypothesis suggested by the studies referred to above, compulsory electoral participation is not of benefit to small parties. In fact, the coefficients for the various compulsory voting indicators are in all cases negative, suggesting that, if anything, obliging voters to attend the polls actually has a depressant effect on small party success. Only the variable for a constitutionalised obligation to vote is significant, however. This result is somewhat unexpected, but it can perhaps be explained in relation to the greater degree of national integration that could be associated with states that agree to constitutionalise electoral participation. In any case, the fear that mandatory voting would have the effect of leading to the proliferation of minor parties appears to be unfounded. Far-right parties The rise of far-right parties in recent years has been of considerable concern to political commentators, especially in the European context, where the inclusion of such parties in ruling coalitions in several countries has sparked fears of a revival of xenophobic and authoritarian politics. For this reason, any proposed reform of electoral institutions is invariably assessed at least in part on the basis of whether it would be likely to encourage or benefit the far right. A variety of arguments has been put forward by political actors and media commentators about the effect of compulsory voting on the electoral fortunes of parties in this portion of the political spectrum. Yet there is little consensus as to what this impact is likely to be. De Winter et al. (2003: 59) note that the

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

124

11:31

Page 124

Full participation

Table 7.2 OLS regression models of the impact of compulsory voting on support for small parties Variable Compulsory voting Sanctioned compulsory voting Electoral obligation constitutionalised Controls: SMD electoral system Ethnic fractionalisation Democracy Population (log)

Model 4 ⫺3.59 (2.24)

Model 5

Model 6

⫺2.877 (2.92) ⫺4.81** (2.34)

⫺5.73*** (2.00) 1.85 (3.82) ⫺0.27 (0.31) 1.75*** (0.43)

⫺5.27*** (1.19) 1.94 (3.85) ⫺0.29 (0.31) 1.70*** (0.43)

⫺6.17*** (2.02) 2.04 (3.80) ⫺0.25 (0.31) 1.83 (0.43)

Constant

0.88 (6.16)

0.97 (6.19)

0.21 (6.13)

N: Adj. R2:

129 0.19

129 0.18

129 0.20

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

rise of the far right in France has led to a call for the introduction of electoral compulsion on the grounds that supporters of extremist parties were more likely to vote than the supporters of moderate parties, while in Belgium there have been calls for the abolition of the same institution as a means of stemming the rise of the far right due to the perception that such right parties have weaker support bases than those of the mainstream. It is, of course, possible that both views are correct if the support bases of the far right in France differ from those in Belgium, but this may also be a case of the grass always being greener on the other side of the fence. It might be argued that compulsory attendance at the polls would benefit any party that is the object of protest voting, which is often the case for extremist parties. At the same time, it should work to the disadvantage of parties whose supporters are intensely committed to the cause of their party, as the mobilisational advantage enjoyed by these parties under a voluntary regime will be lost. In as much as extremist parties are often thought to have highly committed supporters, compulsory voting might be expected to harm their electoral chances. In practice, extremist parties tend to have heterogeneous support bases, winning votes from both the disaffected and the intensely committed. The impact of the mandatory participation in elections on any given extremist party therefore depends on the precise composition of that party’s support base, and specifically on the balance of protest and committed votes the party gains.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 125

Compulsory voting and political outcomes

125

While there is little concrete empirical research to elucidate this question, de Winter and colleagues have provided extensive evidence to suggest that compulsory voting does not particularly benefit the far right in Belgium, contrary to views often expressed in that country’s media (de Winter et al., 2003).1 Yet comparative bivariate analysis by Pippa Norris (2005: 122–3) shows that in elections held in a selection of 39 (mostly democratic) states between 1990 and 2004, the far right performed slightly better in states with compulsory voting than in those where voting was voluntary; the overall difference in votes was 2.7 per cent during this period, which translated into an increase in seats of 2.4 per cent. Though these numbers are small, they nevertheless represent a substantial proportion of the overall vote for the far right, suggesting that if compulsory voting does serve to inflate the vote share of radical right parties, this could considerably alter their political position in many contexts. To assess this hypothesis in a cross-national setting, aggregate data from the Benoit-Laver dataset of expert placements of parties on the left-right spectrum were employed (see Benoit and Laver, 2006; see the Appendix for details of coding). Using the Benoit-Laver data, the proportion of the vote given to farright parties was calculated for each of the 41 elections that correspond to those in the aggregate dataset employed in this volume.2 The bivariate relationships suggest that obliging people to participate in elections does indeed have the effect of magnifying support for the far right – in states in this dataset that practice this institution, the average combined vote for far right parties is 4.2 per cent, nearly double the 2.3 per cent of the vote that goes to such parties in states where voting is voluntary. It is interesting to note that this difference of just under two per cent is broadly similar to that produced by Norris, though the calculations presented here rely on different data. Yet this finding is accounted for in large part by the substantial Belgian vote for the far right. If Belgium is excluded, the figures for mandatory and voluntary voting states are nearly identical: 2.3 per cent on average in states with sanctioned compulsory voting as opposed to 2.8 per cent in all others. Multivariate analysis allows us to incorporate established aggregate-level control variables for support for the far right, including electoral system type (Jackman and Volpert, 1996; Golder, 2003; van der Brug et al., 2005),3 levels of unemployment (Jackman and Volpert, 1996; Golder, 2003), immigration levels (Betz, 1994; Golder, 2003), and the interaction between unemployment and immigration (Golder, 2003). The results of this analysis suggest that though compulsory voting on its own does not have a significant impact on far right support, sanctioned compulsory voting increases the vote share of the far right by over four per cent, all else being equal (see Model 8, table 7.3). However, further analysis suggests that this effect is driven largely by the inclusion of Belgium. When Belgium is removed from the dataset and this model is re-run, the coefficient for sanctioned compulsory voting

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

126

11:31

Page 126

Full participation

Table 7.3 OLS regression models of the impact of compulsory voting on support for far-right parties Variable Compulsory voting Sanctioned compulsory voting Electoral obligation constitutionalised

Model 7

Model 8

Model 9

4.11* (2.08)

1.92 (2.22)

Model 10

2.61 (1.95)

2.23 (1.93)

Controls: SMD electoral system Immigrant population

⫺2.60 (2.29) ⫺2.71 (2.22) ⫺2.33 (2.11) ⫺2.05 (2.35) ⫺0.46* ⫺0.52** ⫺0.43* ⫺0.42* (0.23) (0.23) (0.22) (0.23) Unemployment ⫺0.41** ⫺0.44** ⫺0.39** ⫺0.39* (0.20) (0.19) (0.19) (0.20) Immigrant population x 0.08*** 0.08*** 0.07*** 0.074*** unemployment (0.03) (0.03) (0.02) (0.03) Constant 4.17* (2.24) 4.51** 3.75 (2.10) 3.67 (2.26) (2.19) N: 35 35 35 34 Adj. R2: .22 .27 .24 .21 * p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

is considerably reduced in magnitude, and it also falls well below statistical significance (see Model 9, table 7.3). Finally, Model 10 in this table shows that constitutionalisation of the legal obligation to vote has no significant impact on the vote share of the far right. This evidence suggests that the Belgian example is exceptional, and that there is no clear systematic relationship between far right support and the institution of mandatory electoral participation. Left and right forces The impact of institutional reform is perhaps the greatest concern of practicing politicians, who may either welcome or shun such a change if it is expected to have a partisan effect. At the same time, the perception of partisan advantage may well diminish the legitimacy of the institution – whatever the steadystate impacts this institution has on political legitimacy – if it is perceived as the pet reform of a particular party. Even arguments that compulsory voting will bring in its wake greater social equality will have a suspicious air to them if the politicians voicing such views are anticipated to benefit from full participation. There has been a long-running debate in the US over the impact of that country’s low levels of electoral participation on partisan outcomes, with

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 127

Compulsory voting and political outcomes

127

some scholars claiming no effect, some claiming a significant impact, and others finding that the partisan advantage of full participation would be small (see Highton and Wolfinger, 2001 and Martinez and Gill, 2005 for overviews of this literature, which generally shows that the Democrats gain a small advantage from higher turnout). Several comparative analyses have also found that higher turnout is associated with greater support for left-wing parties (Crewe, 1981; Pacek and Radcliffe, 1995), though others have found little or no impact (Bernhagen and Marsh, 2007; Czesnik, 2007; van der Eijk and van Egmond, 2007; Fisher, 2007; Pettersen and Rose, 2007; Rubenson et al., 2007). In the Australian context, it is widely held that compulsory voting benefits the left (Jackman, 1999: 30–3). Mackerras and McAllister (1999: 227–9; cf McAllister, 1986) show that compulsory voting ‘disadvantages’ right-wing parties,4 though Jackman (1999: 42) argues persuasively that the effect is less than raw survey results indicate, due to the effects of social desirability bias and non-response. In Belgium, Hooghe and Pelleriaux (1998) find that abolishing compulsory voting in Belgium would lead to the over-representation of those with higher education, higher socio-economic status, and male gender, but interestingly – that it would have little partisan effect. De Winter and Ackaert (1998) agree that the partisan effects of electoral compulsion are difficult to gauge, though they also note that small shifts in vote share can have major consequences for the dynamics of coalition formation (see also de Winter et al., 2003). In Brazil, survey evidence suggests that the left would benefit slightly from a voluntary voting regime, but the effect is marginal at best and it would not have altered recent election results (Elkins, 2000). Using the Benoit-Laver dataset, the proportion of the vote given to left parties was calculated for each of the 41 countries that correspond to those in the aggregate dataset employed above. The bivariate relationships suggest that the left vote is actually lower in states with compulsory voting than in those where going to the polls is voluntary. The share of the vote won by left parties is 40.9 per cent on average in the former, but only 32.6 per cent in the latter. If we restrict the comparison to the five of six states in this dataset that impose sanctions on non-voters,5 we find that the mean vote share of left parties is 37.3 per cent, as against 40.0 per cent in all other states in the dataset (though neither of these differences is significant). These figures confound the common wisdom that compulsory electoral participation benefits the left, and they warrant further investigation by means of multivariate analysis. The multivariate models presented in table 7.4 include controls for the overall level of political and economic development of the country in question, as well as a control for short-term economic conditions, proxied here by GDP growth and the rate of unemployment in the year of the election.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

128

11:31

Page 128

Full participation

Table 7.4 OLS regression models of the impact of compulsory voting on support for left parties Variable Compulsory voting Sanctioned compulsory voting Electoral obligation constitutionalised Controls: Per capita GDP (log) Democracy

Model 11

Model 12

Model 13

⫺1.78 (5.95) 3.75 (6.25) ⫺3.58 (5.79) ⫺9.21*** (3.17) 4.60** (2.29) ⫺1.97* (1.04) ⫺0.02 (0.40)

⫺9.92*** (3.11) 4.87** (2.28) ⫺2.19** (1.02) ⫺0.03 (0.40)

⫺8.93*** (3.146) 4.35* (2.33) ⫺2.01** (0.99) ⫺0.03 (0.40)

Constant

72.82*** (26.69)

76.15*** (26.35)

74.06*** (26.15)

N: Adj. R2:

39 .18

39 .18

39 .18

GDP growth Unemployment

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

These models show that in the dataset used here, the more democratic the country, the more likely it is that the left will do well at elections, whereas the richer it is and the faster the economy is growing, the more likely the right will benefit.6 Yet in none of its forms does compulsory voting affect the overall balance of left-right support. This null finding may reflect the fact that parties adapt to serve the needs of their voters, and that left parties in states with mandatory electoral participation are more willing to cater to the interests of underprivileged groups which might not otherwise vote. This suggests that to understand the impact of mandatory voting laws on political outcomes, we need to move beyond partisan trends to actual policy outputs. It is to this task that the next section of this chapter is devoted. Political outcomes It might be supposed that if compulsory voting has no evident partisan impact on election results, as the evidence presented above indicates, then it is not an institution that is particularly relevant in political outcome terms. This is a fallacious argument, however, as it rests on the assumption that parties are fixed entities that do not respond to the expectations of voters. If parties in a given party system were suddenly given an incentive to cater to the interests of a

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 129

Compulsory voting and political outcomes

129

previously under-represented sector of the electorate, their efforts to do so could cancel each other out, and subsequent elections could register no significant shift in patterns of party support. At the same time, all the parties in the system could well have altered their policy programmes to reflect the preferences of the new group, such that the policies delivered following the change could differ markedly from those previously implemented. The partisan distribution of support is therefore not the only outcome variable that needs to be considered when investigating the political impact of compulsory voting; it is also necessary to examine possible variations in the outcomes of the political process. In his seminal paper advocating compulsory voting, Arend Lijphart (1997) refers to a variety of literature documenting the policy consequences of participation differentials, claiming that the under-represented would benefit materially from full participation. Studies of the Netherlands (Irwin, 1974; Irwin and van Holsteyn, 2005) and Belgium (de Winter and Ackaert, 1998; de Winter et al., 2003) have provided evidence that compulsory voting has an impact on the social composition of the electorate, in that it reduces/prevents the over-representation of the more affluent which has resulted or would result under voluntary voting. Similarly, a cross-national analysis of European data by Gallego (2007) suggests that raising turnout reduces educational biases in the composition of the electorate. Yet social biases in the composition of the electorate do not necessarily translate into policy outcomes that favour voters over non-voters. To demonstrate that this is the case, it would need to be shown that legislators discriminate in policy-making against non-voters. There is some evidence from the US that this type of discrimination does indeed take place in that country (Martin, 2003; Griffin and Newman, 2005), that higher turnout favours welfare spending in industrialised countries (Hicks and Swank, 1992), and that there is a positive relationship between turnout and income distribution (Mueller and Stratmann, 2003). The question that concerns us here is what specific impact the institution of compulsory voting has on political outcomes. If all eligible electors vote, including the generally less-well-off groups that tend to vote less regularly when going to the polls is optional, would their participation actually translate into policies more likely to benefit the less advantaged in society? Or would they – as Selb and Lachat (2007) suggest – be less able than voluntary voters to identify political options that would be to their advantage, and therefore unlikely to have a clear impact on the outputs of elected governments? Though social justice arguments about compulsory voting have often been made in the context of normative analysis, these claims have rarely been subject to empirical investigation.7 Two measures of outcome fairness will be employed here to test the social justice claim: the economic measure of income (in)equality, and the political measure of good governance/corruption. The argument with regard to political equality is that if all participate in political decision-making, economic

M1546 - BIRCH TXT.qxp:ANDY Q7

130

27/10/08

11:31

Page 130

Full participation

outcomes should be more equal than if the advantaged participate more than the disadvantaged. Likewise with political outcomes: full participation should provide a superior means of holding politicians to account, in that those disaffected by corruption will not have the choice of opting out, but will instead have to make a conscious choice among ballot options (even if that choice is to spoil the ballot in question). We start with income inequality. In a recent cross-national analysis, Chong and Olivera (2005) find that compulsory voting does indeed depress income inequality. The only measure of compulsory voting found in their paper to be significant, however, was the strictly-enforced version. As argued in Chapter 1, this conceptualisation of the institution is flawed, in that it conflates state capacity with institutional design.8 In fact, levels of wealth inequality are actually higher in states that formally require their citizens to attend the polls than in those that do not do so. The average Gini index for compulsory voting states in the early 1990s was 45.2, whereas it was 38.0 in countries where voting is voluntary (a difference significant at the .01 level). When it comes to sanctioned compulsory voting, however, the difference is drastically reduced: the Gini index in those states that enforce compulsory voting laws is 39.8 on average, while it is 39.5 in all other states. It may be speculated that Latin American exceptionalism once more provides an explanation for the first-reported difference; the average Gini index in Latin American states is, at 49.9, far higher than the average of 37.3 elsewhere in the world (bearing in mind that this dataset includes only the 109 cases for which usable data were available). Turning to multivariate analysis, models were constructed on the basis of the country-averaged data employed in previous regressions. The dependent variable was a measure of the Gini index from the World Bank World Development Indicators database. Following previous studies, control variables include regional dummies designed to capture cultural factors, as well as those designating the level of social and economic development (Deininger and Squire, 1996; Chong and Olivera, 2005). The proportion of the legislature made up of women was employed as a proxy for social development, and per capita GDP was used to measure economic development. The first three models presented in table 7.5 do not suggest that compulsory voting has a significant impact on overall income inequality. None of the coefficients for compulsory voting is statistically significant.9 As expected, the coefficient for sanctioned compulsory voting comes closest to being significant (it is somewhat larger than its standard error), and it is associated with a negative impact on inequality, yet these results are not encouraging. It may be that the Gini index – the most readily comparable of cross-national indicators – is too crude a measure to tap certain efforts at wealth distribution, or it may be that the cultural and historical factors associated with inequality are too complex to be adequately captured in this model.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 131

Compulsory voting and political outcomes

131

Table 7.5 OLS regression models of the impact of compulsory voting on inequality (Gini index) Variable

Model 14

Model 15

Compulsory voting ⫺0.01 (2.64) Sanctioned compulsory voting ⫺3.35 (2.54) Electoral obligation constitutionalised Controls: Per capita GDP (log) Proportion women in parliament Africa Asia and the Pacific Eastern Europe and Central Asia Western Europe Latin America

Model 16

Model 17

⫺5.52*** (1.66) 2.99 (2.52)

⫺0.22 (0.65) ⫺0.13 (0.65) ⫺0.13 (0.64) ⫺0.59 (1.15) ⫺0.23*** ⫺0.25*** ⫺0.21** ⫺0.48*** (0.09) (0.09) (0.09) (0.09) 7.88** 7.57** 8.10** (3.79) (3.78) (3.76) ⫺2.13 (3.61) ⫺2.57 (3.61) ⫺1.80 (3.59) ⫺7.37** ⫺7.86** ⫺7.07** (3.47) (3.46) (3.43) ⫺4.28 (3.36) ⫺4.170 ⫺5.32 (3.45) ⫺17.47*** (3.35) (2.58) 13.25*** 13.72*** 10.98*** (3.97) (3.48) (3.94)

Constant

43.18*** (6.75)

43.26*** (6.76)

41.99*** (6.78)

55.04*** (8.60)

N: Adj. R2:

121 0.56

120 0.56

121 0.56

34 0.90

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

In order to hold as many of these factors as possible constant, it makes sense to conduct a separate analysis of those parts of the world where the legal obligation to vote is most prevalent: Latin America and Western Europe. Model 17 in table 7.5 is designed to test the impact of sanctioned compulsory voting in these two regions. This model lends considerable support to the hypothesis that the institution of compulsory voting should be associated with a substantial reduction in wealth inequalities; the coefficient for sanctioned compulsory voting is negative and highly significant. It does appear that in Western Europe and Latin America, at least, mandatory attendance at the polls promotes social equality. Finally we consider the extent to which full electoral participation might place a check on the abuse of power. If the entire electorate is called on to express its opinion of political leaders, disaffected and disillusioned citizens who

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

132

11:31

Page 132

Full participation

Table 7.6 OLS regression models of the impact of compulsory voting on corruption Variable Compulsory voting Sanctioned compulsory voting Electoral obligation constitutionalised Controls: Per capita GDP (log) Africa Asia and the Pacific Eastern Europe and Central Asia Western Europe Latin America Democracy Proportional electoral system

Model 18

Model 19

Model 20

⫺0.13 (0.32) 0.70** (0.34) ⫺0.44 (0.32) 0.83*** (0.08) 0.40 (0.42) 0.12 (0.38) ⫺0.64* (0.38)

0.81*** (0.08) 0.44 (0.42) 0.13 (0.38) ⫺0.57* (0.37)

0.81*** (0.08) 0.38 (0.42) 0.11 (0.38) ⫺0.66* (0.37)

1.12*** (0.40) ⫺0.63 (0.46) 0.14*** (0.03) ⫺0.31* (0.20)

1.04*** (0.37) ⫺0.82** (0.41) 0.14*** (0.03) ⫺0.32* (0.19)

1.22*** (0.41) ⫺0.40 (0.47) 0.14*** (0.03) ⫺0.28 (0.20)

Constant

⫺3.58*** (0.77)

⫺3.56*** (0.77)

⫺3.46*** (0.77)

N: Adj. R2:

149 0.78

148 0.78

149 0.78

* p ⬍ .10; ** p ⬍ .05; *** p ⬍ .01 Unstandardised coefficients are reported; figures in brackets are standard errors.

might not otherwise vote may be more likely to cast their ballots in favour of the options they perceive to be most likely to curb abuse. We might therefore expect compulsory voting to be associated with lower overall levels of corruption. This hypothesis gains some support from the bivariate analysis of the relationship between compulsory voting and the Transparency International Corruption Perceptions Index (the most widely used cross-national indicator of corruption, available in 2006 for a total of 163 states, and 149 of those included in the database used in this study). Bearing in mind that a high score on the 1–10 Corruption Perceptions Index (CPI) indicates less corruption, the mean 2006 CPI score for states with compulsory voting was 4.4, as against 3.9 for all other states (though this difference is not statistically significant). When we consider sanctioned compulsory voting, the gap widens further to a mean of 6.0 in states that enforce the obligation to vote, and only 3.9 in other states – a difference that is significant at the .05 level. Multivariate analysis presented in table 7.6 confirms this finding. These models control for variables most consistently found in previous studies to be associated with cross-national variations in levels of corruption, including per

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 133

Compulsory voting and political outcomes

133

capita GDP, level of democracy, electoral system type, and regional location (Triesman, 2000; Montinola and Jackman, 2002; Persson and Tabellini, 2003; Persson, Tabellini and Trebbi, 2003; Gerring and Thacker, 2004; Kunicova and Rose-Ackerman, 2005; Lederman et al., 2005). These models indicate that mandatory attendance at the polls provides a means of checking the abuse of power by politicians, in that levels of corruption appear to be lowered by this institution (at least when it has sanctions associated with it). Sanctioned compulsory voting has a significant positive effect on the CPI, controlling for other relevant variables. Full participation appears to help hold leaders to account. Conclusion The impact of compulsory voting on political outcomes is arguably the ‘makeor-break’ consequence that is most likely to shape views of the institution among political elites and mass public alike. Politicians are naturally sensitive to the partisan consequences of any institutional reform proposal, and ordinary citizens are likely to be more swayed by arguments about the concrete social consequences of a reform than by less tangible potential effects such as an increase in levels of satisfaction with democracy. This chapter has found that a number of the fears commonly voiced at the mention of compulsory voting are unfounded, while at least two of the potential benefits of this institution receive empirical support. There is no clear evidence that mandatory participation in the electoral process has any systematic impact on the support won by different types of party. This should be good news to those who fear that the introduction of mandatory voting might upset the political equilibrium of a country, or that it might lead to an up-surge in the number of seats won by politically and ideologically minor political forces (though of course the general cross-national patterns examined here do not preclude the possibility of partisan effects in individual cases). Nor has this chapter discovered clear evidence that full participation increases female representation, but it has found the institution to be associated with the desirable outcomes of greater income equality (at least in Latin America and Western Europe) and reduced corruption. This indicates that even if electoral malpractice in the narrow sense is unaffected by making electoral participation mandatory (as was found in Chapter 6), other aspects of governance are positively influenced by this institution, and it ‘delivers’ on what is perhaps one of the strongest claims made for it: that it should serve the ends of social justice. This chapter concludes the empirical analysis of compulsory voting undertaken in this volume. But what are we to make of all these regression coefficients? The conclusion to follow represents an attempt to draw up a balance

M1546 - BIRCH TXT.qxp:ANDY Q7

134

27/10/08

11:31

Page 134

Full participation

sheet out of the findings presented in this and the previous three chapters, and to assess some of the practical factors associated with the potential adoption of mandatory voting rules. Notes 1 Pilet notes that Belgian opinion polls provide contradictory evidence on this point (2005: 11–12). 2 The elections included in the resulting dataset are: Albania 2001, Australia 2001, Austria 2002, Belgium 2003, Bosnia and Herzegovina 2002, Canada 2000, Croatia 2003, Cyprus 2001, Czech Republic 2002, Denmark 2001, Estonia 2003, Finland 2003, Germany 2002, Greece 2004, Hungary 2002, Iceland 2003, Ireland 2002, Israel 2003, Italy 2001, Japan 2003, Latvia 2002, Lithuania 2000, Luxembourg 2004, Macedonia 2002, Malta 2003, Moldova 2001, Netherlands 2003, New Zealand 2002, Norway 2001, Poland 2001, Portugal 2002, Romania 2000, Slovakia 2002, Slovenia 2000, Spain 2004, Sweden 2002, Switzerland 2003, Turkey 2002, Ukraine 2002, United Kingdom 2001, and the USA 2002. The compulsory voting states include Australia, Belgium, Cyprus, Greece, Luxembourg and Turkey. For the purposes of this analysis, Switzerland is considered a voluntary voting state, given that fewer than one per cent of the country’s voters live in the single canton where going to the polls is mandatory. 3 Note that some studies (e.g. Carter, 2002; Norris, 2005) have found that the impact of the electoral system influences the level of electoral success of far-right parties, not their vote share, suggesting that the effect of these institutions is, in Duvergerian terminology, mechanical rather than psychological. This distinction will not be explored here, as the analyses carried out in this chapter employ vote share (rather than seat share) as the dependent variable. 4 One could equally argue that the turnout differential which would result under voluntary voting would distort the true level of party support to the disadvantage of the left, whereas the result achieved under mandatory voting is a more accurate reflection of the distribution of partisan support. 5 The only state in this dataset with non-sanctioned compulsory voting is Greece. 6 It might be thought that these patterns are an artefact of the inclusion of a number of post-communist states in the datasets, but dummy variables for both Western and Post-Communist Europe included in earlier versions of these models failed to achieve statistical significance. 7 There was a debate in the mid-1990s in the journal Economics and Politics on the impact of compulsory voting on government expenditure (see Crain and Leonard, 1993; Yeret, 1995; O’Toole and Strobl, 1995; and Crain 1995). This debate was somewhat inconclusive, but it appears that though compulsory voting depresses growth in overall government expenditure (Crain and Leonard, 1993; Crain, 1995), it also has the effect of shifting government spending from defence and economic services to health and housing (O’Toole and Strobl, 1995). This accords with the view that greater inclusion in the franchise should be associated with more redistributive welfare outcomes.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 135

Compulsory voting and political outcomes

135

8 Another problem with the Chong and Olivera analysis is that they include regional control variables only for Latin America and the OECD countries. The inclusion of an Eastern European control in the analysis presented below demonstrates that levels of income inequality are significantly lower in this region. It is also noteworthy that this is the only region of the world in which the institution of compulsory voting is entirely absent. 9 The positive coefficient for the variable designating constitutionalisation of the obligation to vote is unexpected, and it is difficult to see what might be driving this. Certainly there is no particular reason to believe that such a provision would increase levels of inequality.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 136

8

Conclusion

The institution of compulsory voting is but one of a range of devices that mould the incentives faced by voters and parties at election time. Vote facilitating mechanisms, the extent to which votes ‘count’ under the electoral system in place, the overall freeness and fairness of the election, and a wide range of other institutional factors also shape this context (in addition to the political, socioeconomic and individual-level forces that are associated with whether or not individuals vote). Nevertheless, mandatory voting laws are arguably the institution that has the most direct and straightforward relationship with electoral participation. This final chapter sets out to examine the ‘balance sheet’ of compulsory voting, as reflected in the findings of the previous chapters, so as to evaluate the desirability of the institution as well as the circumstances under which it is likely to have positive effects. The chapter will then draw out the implications of this evaluation and consider the prospects for the (re)adoption of mandatory electoral participation in the near future, the circumstances under which this is likely to take place, and the various modifications that could be made to this institution to render it more palatable to the citizens of contemporary democracies. Several options that could be used instead of mandatory voting will also be assessed, including the constitutionalisation of electoral obligation, collective sanctions for non-voting and ‘incentive voting’. Finally, the chapter will provide a brief summary of the policy implications of the book’s main findings. Compulsory voting: a balance sheet This study has surveyed mandatory electoral participation from a variety of perspectives. The first chapter introduced the concept of the legal obligation to vote and summarised the various forms such an obligation has taken in states that have employed it. Chapter 2 provided a brief history of the use of compulsory voting laws and sought to discover commonalities in the reasons for their adoption; the most frequent of these appeared to be calculations of partisan advantage, considerations related to legitimacy (including the desire to increase rates

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 137

Conclusion

137

of electoral participation and strengthen democracy), and the perceived need to reduce electoral fraud and abuse. The normative arguments commonly made for and against mandatory electoral participation were laid out in Chapter 3, where it was shown that many of these arguments rest on empirical assertions for which there has heretofore been scant evidence. The following four chapters sought to provide evidence to test these assertions. These chapters were structured in such a way as to reflect both logically and chronologically the various stages of the electoral process. Campaigns and related behaviour were the focus of Chapter 4, which assessed the impact of mandatory electoral participation on various aspects of citizen engagement in politics and party engagement with citizens. Chapter 5 analysed the effect of compulsory electoral participation on aggregate turnout, while Chapter 6 considered the variety of ways in which this institution might influence electoral integrity and legitimacy. Finally, Chapter 7 evaluated the potential for mandatory voting rules to alter outcomes of the political process, including female representation, the distribution of partisan support, wealth inequality and corruption. One of the tasks of this concluding chapter is to provide an overall assessment of compulsory voting. We first summarise and assess the normative claims for compulsory voting, before going on to review the empirical evidence. As outlined in the Introduction, the arguments for and against the legal obligation to vote can be divided into three principal categories: arguments that appeal to normative conceptions of democratic rights and duties, those that focus on the relationship between participation, democratic legitimacy, and collective rationality, and consequentialist arguments related to the impact of this measure on satisfaction with democracy, public engagement, and policy outcomes. Before turning to the consequentialist arguments evaluated empirically, it is worth pausing to assess the first two sets of claims. These arguments can be summarised as follows: 1. Rights and duties: Proponents of compulsory voting argue that democratic obligations follow rights, and that voting is a duty of citizenship. Mandatory electoral participation ensures co-operative behaviour through self-binding in the collective interest. Opponents of compulsory voting argue that the right to vote implies a right not to vote. A right is not a right if one does not have the ability to choose whether or not to exercise it. Political duties therefore ought not to be made legally binding. On the question rights and duties, it must be granted that rights are in practice subject to considerable limitations in contemporary democratic states, in that legal compulsion is deeply embedded in democratic systems. State agencies in democracies compel citizens to engage in a range of actions, from completing census forms, sending their children to school and serving in the military to recycling, carrying state-issued identification papers, and registering to vote.

M1546 - BIRCH TXT.qxp:ANDY Q7

138

27/10/08

11:31

Page 138

Full participation

Most of these behaviours are based on social norms that have been institutionalised in the form of law. In as much as norms vary from state to state, the range of activities subject to compulsion is also subject to considerable variation. This suggests that the question of whether rights should be limited in the name of democratic participation depends on the extent to which the duty to vote is an accepted norm in a given society. If it is a norm, then a strong argument can be made that electoral participation should be made mandatory. In the modern world, societal pressure is no longer capable of serving as the social glue that would enforce collective norms, so it may be necessary to rely on legal compulsion to achieve that end. Whereas in previous generations cohesive communities provided a mechanism for enforcing norms of participation, this mechanism has broken down in many modern settings. In the absence of traditional means of norm enforcement, citizens are in need of an alternative method of self-binding that can in contemporary society only be achieved by means of legal regulations. Compulsory voting is just such a regulation. Adding yet another duty – and a crucial duty – to the long list of our legally binding obligations would not in principle be incompatible with the modern practice of rights observation or with the liberal notions underlying it. 2. Democratic legitimacy and collective rationality: Proponents of compulsory voting argue that mandatory electoral participation ensures that the decisions taken by representative institutions have democratic legitimacy. Democratic decisions must, at a minimum, reflect the views of the majority, but unless we have some means of determining everyone’s view (even if that view is that they do not hold a view), decisions will be flawed in democratic terms, and therefore less legitimate. Mandatory electoral participation is the only way of effectively taking all views into account in a way that shares the costs of voting equally among all those who benefit from electoral institutions. The opponents of compulsory voting argue that mandatory electoral participation does not truly solve the problem of low levels of political participation and the corresponding lack of legitimacy of democratic institutions, it merely addresses the symptoms. If the real problems are voter apathy, disengagement and disillusion with politics – the fact that too many people simply don’t have views on the options from which they are asked to choose at election time – there is little point in ‘artificially’ raising turnout levels by compelling people to attend the polls. The collective rationality argument for compulsory voting is a strong one. Making electoral participation mandatory is the only equitable and effective means of overcoming the free rider problem inherent in voting: most people view voting as a public good and recognise the importance of democratic elections. But individuals often see no point in voting themselves, because they know their own contribution to both turnout and electoral results will be

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 139

Conclusion

139

minuscule. The ‘irrationality’ of any individual going to the polls, given the infinitesimal chance that their vote will be decisive, ought in theory to discourage voting altogether and undermine democracy, generating a collectively irrational outcome. Requiring voters to vote is thus a means of guaranteeing the decisiveness of the collective will in a way that distributes the cost of voting equally among all citizens. It can thus be viewed as a convenient co-ordination device for voters. Those who have no view can spoil or fail to complete their ballots; the evidence presented in Chapter 6 suggests that these practices are indeed relatively widely used in compulsory voting states. This discussion suggests that on balance the normative arguments favour compulsory voting. But what of the consequences of this institution? As elaborated in Chapter 3, proponents of compulsory voting argue that the legal obligation to vote will raise levels of citizen awareness, interest in politics and involvement in public life. It will thus have both mobilising and educative effects, increasing levels of public engagement and political information. Moreover, given that mandatory electoral participation will mean elections take account of the views of all, rather than the socio-demographically skewed selection of those who vote under voluntary schemes, the policies delivered by the resulting governments will more accurately reflect the needs of the entire population. The opponents of compulsory voting argue that mandatory electoral participation is in practice costly to administer and is likely to be unpopular. There are other more effective, more popular, and less costly means of increasing public engagement, such as selective incentives to vote, vote facilitation mechanisms, and other changes to the political process. Furthermore, many people are ill-informed about issues of public policy and will therefore not make meaningful choices. Their votes will be arbitrary, and they may even vote for extremist parties in protest against electoral compulsion, which will have a negative impact on the quality of policy outcomes. Assessment of these competing claims has taken up the bulk of this volume. Table 8.1 presents a ‘balance sheet’ of the empirical findings from Chapters 4, 5, 6 and 7. Descriptions of the 22 impact variables have been formulated such that ‘positive’ denotes a normatively preferred effect, while ‘negative’ indicates something most people would rather not see. In simple terms, the ‘positives’ outweigh the ‘negatives’ in this table by a ratio of 3:1. The only truly ‘negative’ findings are that mandatory attendance at the polls appears to depress levels of perceived political efficacy, and that it is often associated with higher numbers of spoilt ballot papers. This should be good news to the proponents of laws mandating full participation, and it should also serve to allay some of the fears of those hesitant to consider this institution seriously. It is worth noting that virtually all of these effects are based on the evaluation of sanctioned compulsory voting. Thus, the empirical analyses carried out here confirm the conventional wisdom that mandating electoral participation

M1546 - BIRCH TXT.qxp:ANDY Q7

140

27/10/08

11:31

Page 140

Full participation Table 8.1 A Balance sheet of empirical findings

Object of impact

Direction of impact

Individual-level: Political knowledge Political efficacy Political conversation and persuasion Propensity to contact politicians Propensity to participate in protests and demonstrations Propensity to work with others to address concerns Political participation (general) Party identification Propensity to participate in campaign activities Likelihood of being contacted by a party or politician Satisfaction with democracy Quality of representation

No significant effect Negative No significant effect No significant effect Positive No significant effect Positive No significant effect No significant effect No significant effect Positive No significant effect

Polity-level: Turnout Electoral integrity Proportion of valid ballots cast Proportion of non-random votes cast Female representation Support for small parties Support for the left Support for the far right Wealth equality Corruption reduction

Positive No significant effect Negative Unclear No significant effect No significant effect No significant effect No significant effect Positive* Positive

* In Latin America and Western Europe.

through legal means is not likely to have a significant impact unless the legal requirement to vote is backed up by an effective enforcement regime. That said, this table suggests that in most cases (14 out of 22) mandatory participation in elections has no significant effect whatsoever, or that its impact is unclear. There are a number of possible explanations for this, the most convincing of which are data limitations and political adaptation. Several of the analyses carried out here are restricted to a limited range of countries for reasons to do with data availability, and it may be that the compulsory voting states included in these datasets are not representative of all states that use this device. A further potential limitation is the fact that the bulk of the statistical analyses reported in this volume are cross-sectional in nature, and it is wellknown that variations across states do not necessarily reflect the likely impact of an institution in a given polity. Unfortunately, it was only possible to conduct detailed longitudinal analysis in the case of turnout, for which data are available

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 141

Conclusion

141

for lengthy time period. In the case of the majority of the other variables studied, there are simply no readily available data for the periods immediately before and after compulsory voting was adopted in those states that employ it. The fact that the longitudinal and cross-sectional analyses of the turnout data point to similar effects is encouraging, but we cannot necessarily assume that the variations across states observed here can predict in any direct way the consequences of adopting mandatory voting laws in any particular state. There are also political reasons why full participation might not generate marked impacts on the variables considered here. Institutions undoubtedly ‘matter’, but at the same time, both politicians and citizens adapt to institutions and adjust their behaviour accordingly. It may also be that an institution such as compulsory voting has countervailing effects that cancel each other out; this could well be the case, for example, when it comes to some of the political engagement and party support variables analysed in Chapters 4 and 7 respectively. But whatever the reasons for the non-significance of a number of the variables investigated in this study, it remains the case that some of the main objects of analysis have proved to be influenced by mandatory voting regimes. Prospects for the adoption of mandatory electoral participation laws Regardless of how we might evaluate mandatory electoral participation in theory, the practicalities of the adoption of such an institution may represent a serious hurdle in many contexts. It is noteworthy that heated recent discussions over such laws in several established democracies have not yet led to any concrete changes. In considering this question, it makes sense first to address the issue of the conditions under which mandatory electoral participation should be adopted, and then the thorny question of how it could be put into practice. The ‘should’ question in fact breaks down into three questions: that pertaining to need, that pertaining to capacity, and that pertaining to the likelihood of abuse. These three questions will be of varying importance in different types of state, but all three issues are worth at least considering. Firstly, it can be expected that the attraction of compulsory voting will vary inversely with the degree to which the current level of electoral participation is seen as problematic. Some voluntary voting states – Malta being the classic example – regularly achieve rates of electoral participation as high as has been obtained under compulsory voting. In such cases there is no evident need for such an institution (though a case could be made that even here that it makes sense to ‘lock in’ the current widespread normative and behavioural adherence to participation in order to guard against a possible future decline). The question of whether a state has the capacity successfully to implement and to operate an effective mandatory voting regime is also highly relevant in this context. We have seen that compulsory voting laws have little impact on

M1546 - BIRCH TXT.qxp:ANDY Q7

142

27/10/08

11:31

Page 142

Full participation

political outcomes in the absence of sanctions. Whether a state has the administrative ability to implement sanctions can therefore be expected to be an important consideration when the potential adoption of this institution is assessed. Lack of state administrative capacity is a problem that plagues many less developed countries, and in this context compulsory voting is probably less likely to have positive political benefits than more fundamental changes to the electoral administrative system which might be made, such as improvements in voter registration and the provision of polling facilities. A final relevant concern in this context is the likelihood of abuse of mandatory voting requirements. In her book Electoral Engineering, Pippa Norris notes that ‘Mandatory voting regulations may be genuine attempts to increase widespread public involvement in the political process, or they may be employed by less democratic regimes to compel the public to vote, in the attempt to legitimize one-party contests’ (2004: 168). It goes without saying that when compulsory voting serves as a handmaiden to authoritarianism, it fails all normative tests. This indicates that the institution might be appropriate for democracies, but not for non-democracies or semi-democratic contexts where it could be a tool for coercion and forced mobilisation (cf Hill, 2002c: 2–3). Since Somoza introduced compulsory voting in Nicaragua in order to enhance his legitimacy, this institution has been associated with repression (Lapp, 2004: 5). Likewise, provisions for mandatory voting were exploited in El Salvador to ensure cooperation with the junta that assumed power in the 1979 coup. Non-voting was declared subversive and treasonous in the 1982 constituent assembly elections, in order to counter an election boycott called by the Front for National Liberation/Revolutionary Democratic Front in protest against the violence and repression of the regime (Barnes, 1998: 94f). Compulsory voting may also not be appropriate when statehood problems mean that a substantial proportion of the population is unwilling to participate in elections due to the fact that they do not recognise the legitimacy of the state in which they reside. Sear and Strickland (2003: 2) claim that Sudeten Germans in interwar Czechoslovakia voted for the communist party in protest against their inclusion in the newly-created state. Such behaviour is likely to skew electoral results and thereby distort the democratic process. It can also be viewed as a coercive approach to inclusion and national integration, to which there are preferable alternatives. These examples suggest that compulsory voting may not be suitable for some emerging and semi-democracies, not to mention authoritarian states. But in the context of most democracies, the institution is one that merits serious consideration. The most relevant question in these contexts is undoubtedly whether compulsory voting could be introduced, in other words whether efforts to introduce it would stand a reasonable chance of surmounting the twin hurdles of public confidence and the support of political elites.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 143

Conclusion

143

It is worth bearing in mind the fact that support for mandatory electoral participation is generally high in countries that have it. Mackerras and McAllister (1999) trace Australian opinion polls back as early as 1943, and find that the proportion supporting the institution never fell below 60 per cent, and that this level increased between the 1940s and 1960s to a high of 76 per cent in 1969 before falling to the 60–70 per cent range in the 1980s and 90s (1999: 221). Hill (2002b) reports figures of 74 per cent support in 1996 and 67 per cent in 1997. Of Belgian non-voters polled in 1978 and 1981, only 8.5 per cent and 12.6 per cent respectively claimed to abstain because they were against the institution of compulsory voting (Bouillin-Dartevelle et al., 1984, 217). The corresponding figure for 1999 in Wallonia was, at 2.4 per cent, even lower (de Winter et al., 2003), suggesting that opposition to the institution is not great. The same was true in the Netherlands, even shortly before that country switched to voluntary voting. Irwin and van Holsteyn report that support for mandatory attendance at the polls was found in a 1966 survey to be 69 per cent, and 70 per cent in a survey conducted the following year. Debates over abolition evidently had the effect of altering public opinion, however. On the eve of abolition in 1969, 53 per cent of survey respondents were in favour of this move, as against 29 per cent who were for maintaining the practice (2007: 3–4). Still, the general conclusion must be that citizens in mandatory voting states are for the most part satisfied with the institution. What are the prospects of compulsory voting being introduced in a democracy near you? Wolfgang Hirczy notes a recent increase in enthusiasm for compulsory voting among academics and electoral reformers, but notes also that ‘it is generally conceded that the chances for adoption are slim’ (2000: 45). Yet given that such a move has only sporadically risen to the surface in most contemporary voluntary voting states, popular support for such a measure is remarkably high in many places, with significant numbers of ordinary citizens favouring it. Table 8.2 contains a selection of survey results conducted at various times in various places in states that do not currently have compulsory voting. This sample is not meant to be representative in any way; it simply includes those surveys I came across in the course of researching this book.1 Inevitable variations in question wording somewhat limit comparability across countries, but the findings suggest that in three of four cases support for this institution is slightly over a third, while in France the 2004 survey finding cited here put the figure at nearly double that rate (though the French figure could well reflect a temporary surge in support following a call for compulsory voting from French Socialist Laurent Fabius – see below). Nearly two in five Canadians were either ‘very’ or ‘somewhat’ favourably disposed to compulsory voting in 2002 (following the 2000 general election, when turnout had declined to 61.2 per cent from 67.0 per cent in the previous election), while in the Netherlands – a former compulsory voting state – the figure of 35 per cent support is almost

M1546 - BIRCH TXT.qxp:ANDY Q7

144

27/10/08

11:31

Page 144

Full participation

Table 8.2 Comparative survey data on support for compulsory voting in voluntary voting states Country (year)

Source

Question wording

Results

Canada (2002)

Pammet and LeDuc, 2003, Table 41, p 48.

N/A (table heading: ‘Support for Compulsory Voting’)

Very supportive: 16.7% Somewhat supportive: 22.8% Somewhat opposed: 23.0% Very opposed: 37.4% (39.5% support)

France (2004)

IFOP polla

‘Would you be for or against voting at elections being made compulsory?’b

Very much in favour: 29%; somewhat in favour: 33%, somewhat opposed: 24%, very opposed: 13% (62% support)

Irwin and van Holsteyn, 2007: 4.

N/A

MORI pollc

N/A

35% support the reintroduction of compulsory voting 36% support 57% against

Netherlands (1999)

UK (2005)

a

Details of this poll are available at http://www.ifop.com/europe/sondages/opinionf/voteoblig.asp. b ‘Seriez-vous favourable ou opposé à ce que le vote aux élections devienne obligatoire?’ [author’s translation from the French]. c Cit. Electoral Commission 2006, p. 11.

identical to that found in the UK in 2005.2 It is noteworthy that the 36 per cent figure cited here represents a substantial decline from a UK survey conducted in July 2001, shortly after the 2001 election, when support for making electoral participation mandatory had risen to 47 per cent.3 These data suggest that a significant minority of citizens in voluntary voting states are open to the idea of making electoral participation mandatory, and that others could quite possibly be convinced of the virtues of the institution if it could be portrayed to them as a convenient means of collective co-ordination and self-binding in the public interest in a context where traditional social pressures are no longer serving that function effectively. The role of convincing

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 145

Conclusion

145

citizens would in all probability need to be played by politicians. It is thus instructive to consider the opinions of political élites on the question of compulsory voting. In a recent study of the attitudes of politicians in New Zealand, Germany, the Netherlands and Australia toward proposed reforms of electoral institutions, Bowler, Donovan and Karp found that, as with other proposals for reform, those who had successfully won elections were significantly less likely to favour alteration of the status quo when it came to the question of compulsory voting (in Australia such a change would entail abolition of the institution, whereas in the three other states it would involve introduction). Members of small parties were also significantly less likely to favour reform in this sphere, as were members of left-wing parties (which is at odds with the attitudes of both groups toward other reform proposals, including the adoption of term limits or the use of referenda and initiatives) (Bowler et al., 2006: 442).4 Satisfaction of the winners with the status quo is generally understandable, but the greater affinity with this institution of larger parties and those on the right of the political spectrum is somewhat puzzling, given that academic research has, if anything, portrayed compulsory voting as favouring small parties and those of the left. As we saw in Chapter 7, neither of these claims stands up to large-N statistical analysis, but the results of the study by Bowler and colleagues nevertheless merit further probing. With this in mind, it is worth considering debates about compulsory voting in three countries where this topic has been discussed recently, and which are not included in the Bowler et al. study: the UK, Canada and France. The United Kingdom Compulsory voting was considered at various times in twentieth century Britain (see Sear and Strickland, 2003), and its advocates have included such illustrious figures as Winston Churchill (Watson and Tami, 2004). Yet there was little serious debate of the subject in the latter half of the century. This changed in the early years of the twenty-first century when a dramatic decline in turnout at Westminster elections, from 71.6 per cent in 1997 to 59.2 per cent in 2001, brought public engagement into the spotlight. Among the many proposals put forward to deal with this problem, compulsory voting has surfaced more than once. Recent proposals by politicians to introduce or consider this institution have generally come from members of the Labour party. In November 2001 Labour Gareth Thomas MP presented a private member’s bill to introduce compulsory voting, though the bill made it no further than the first reading. In 2005, Labour minister Geoff Hoon caught the attention of the ‘chattering classes’ when he also called for the introduction of mandatory electoral participation to be seriously considered.5 There followed in 2006 no fewer than three major reports on the topic, one by the Electoral Commission, and two by think tanks (Ballinger, 2006; Electoral Commission, 2006; Keaney and

M1546 - BIRCH TXT.qxp:ANDY Q7

146

27/10/08

11:31

Page 146

Full participation

Rogers, 2006). There is little evidence that the Labour party as a whole supports making voting mandatory, however, and none of the other main parties views the institution with favour. Canada A Royal Commission on Electoral Reform and Party Financing rejected compulsory voting in 1991 as being inconsistent with Canadian values, but interest in the topic received a sudden injection not from a political source, but from newly-appointed Chief Electoral Officer, Jean-Pierre Kingsley, who made the desirability of compulsory voting one of his first topics of public comment following the parliamentary elections of 2000 (Hill, 2002c: 1). The issue was taken up by Canadian Senator Mac Harb (Liberal), who introduced a bill in the Senate in support of compulsory voting in late 2004, on the grounds that the ‘obligation to vote must be accepted as one of the necessary duties citizens carry out in order to maintain our system of democracy and the benefits that go with it’.6 This bill managed to reach a second reading, where it was extensively debated, but it ultimately failed in June 2005.7 It is worth noting that again, it was a member of the main centre-left party that took forward the initiative on making voting mandatory. France In France it has also been the left-of-centre political forces that have in recent years most seriously considered making voting compulsory, due to a common perception that it would prevent those with intense preferences on the extremes of the political spectrum (and notably the far right) from expanding their electoral base. Falling turnout among left-wing voters has also been a consideration. These two rationales converged at the time of the 2002 presidential elections, where low turnout among Socialist voters allowed far-right candidate Jean-Marie Le Pen to come second to Jacques Chirac in the first round of voting and thereby to enter the run-off, when many chagrined Socialists found themselves voting for right-wing Chirac in order to keep out the far right. Socialist Laurent Fabius announced in 2003 that he favoured the introduction of compulsory voting for National Assembly elections (Service des Etudes Juridiques, 2003: 5). The bill introduced at that time was defeated, however. The evidence from New Zealand, the Netherlands and Germany that support for compulsory voting is associated with right-wing parties therefore appears to be an anomaly, though the finding that right-wing Australian politicians favour the repeal of that country’s compulsory voting laws is consistent with the British, Canadian and French situations reviewed here. The greater support for compulsory voting among larger parties found by Bowler et al. is, however, consistent with the British, Canadian and French data: in all cases, it is larger parties that have taken up this issue. This evidence provides some insight into

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 147

Conclusion

147

the views of politicians on the matter of electoral obligation, but the prospects of this institution being adopted in a real-world context will depend invariably on the form any specific proposal takes. It is to the specific institutional design context of mandatory voting that we now turn. Modalities of mandatory electoral participation It goes without saying that the introduction of mandatory electoral participation will never be straightforward in a state where attendance at the polls has previously been voluntary. At the same time, the challenge is by no means insuperable, and there are several institutional mechanisms that can be employed to make full participation more palatable to the citizens of contemporary states: • Include a ‘none-of-the-above’ or’ abstain’ option on the ballot paper to ensure that those unattracted to any of the political options on offer can voice their disaffection without being obliged to spoil their ballot papers.8 • Provide exemptions on grounds of age, infirmity, lack of access to polling facilities,9 as well as for valid excuses. • Allow for conscientious objection.10 Though it should require some effort to apply for this status to ensure that it is not taken advantage of, those who genuinely do not believe in voting should be allowed to refuse the franchise. Objector status must in be reversible in theory, though opting back into the franchise should be made difficult as well, in order to prevent frivolous use of the procedure. • Facilitate voting, in as far as this is possible without endangering ballot secrecy and other guarantees of the integrity of the electoral process. For all practical purposes, voting must be done under the supervision of electoral authorities in order for ballot secrecy to be ensured. This means that the extensive use of postal or remote electronic voting would not be compatible with democratic norms (though such practices might be the only feasible way of enfranchising people living in very remote locations). However, there are a number of ways in which voting can be facilitated, including the siting of polling facilities in post offices, libraries and other public places (most likely in the form of automated ‘secrecy booths’ or kiosks). Enabling voters to vote at any polling station, through the use of an electronic register and electronic voting, would also make it considerably more likely that people could reach a polling place on election day. Other possible vote-facilitation mechanisms include holding elections over two rest days and the synchronisation of election cycles. • Provide stringent guarantees that information gathered for the purposes of electoral registration and voting would not be shared with other state agencies, and that vote choices would remain secret from all.

M1546 - BIRCH TXT.qxp:ANDY Q7

148

27/10/08

11:31

Page 148

Full participation

• Mount a comprehensive public information and education campaign to make people aware of their legal obligation to vote and the reasons behind this obligation. Yet even were such measures to accompany a mandatory voting regime, it can be envisaged that significant numbers of citizens might still object to making voting compulsory. For the institution to have the requisite legitimacy that would enable it to function effectively, it would require a popular mandate. This suggests that compulsory voting would be most effectively introduced were it to be first piloted and then approved by referendum.11 Passage of a referendum question on compulsory voting might be easier than believed, in that the referendum itself would be held under voluntary voting rules, and those who participate in referendums are, in the majority of cases, regular voters who would be least inclined to object to obliging their fellow citizens to follow their example. Once introduced, the legal obligation to vote would most likely be internalised by the vast majority of the population, as ‘legal control works principally through a law’s enactment, not its enforcement’ (Hasen, 1996: 2168). Nevertheless, the evidence presented in the preceding chapters indicates that the establishment of an effective enforcement regime is key to the success of compulsory voting rules. The existence of sanctions for non-voting signals to the citizenry that the state takes the institution seriously. At the same time, ‘strict’ enforcement need not take the form of universal monitoring. To see why it is not necessary to enforce compulsory electoral participation assiduously for the legal obligation to vote to be a deterrent, one need only think of public transport systems where passengers are in theory obliged to buy tickets, but the tickets are checked only sporadically. If the penalty for being caught multiplied by the probability of being caught is high enough, it will not be rational for voters to violate the law. This indicates that a cheaper means of increasing compliance may be to raise the penalty for non-voting, rather than to increase the extent of monitoring. It is worth noting, however, that very high penalties for non-voting are typically not politically palatable, and that compliance is highest where there is a combination of an active enforcement mechanism, high levels of voting in general, and a positive – or acquiescent – attitude toward the institution. These considerations suggest that where the political will is present at the level of the political elite, compulsory voting could well be successfully introduced. At the same time, there are undeniably many contexts in which either the elite or the public are sufficiently resistant to this idea that it is unlikely to feature on the political horizon for more than brief moments. Under these circumstances, it makes sense to consider possible alternatives to the legal obligation to participate in elections.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 149

Conclusion

149

Other options The Introduction to this book identified three possible institutional reform options that could be considered instead of mandatory electoral participation: the constitutionalisation of electoral obligation, collective sanctions in the form of turnout requirements, and the provision of incentives for voting, All three mechanisms have been employed in the past, and it is worthwhile considering whether any of them represents a feasible alternative to compulsory voting. In the case of a constitutional requirement to participate in elections, the answer is a resounding ‘no’; the impact of this device was tested in Chapter 4, and it was not associated with a statistically significant increase in rates of electoral participation. We are therefore obliged to conclude that such clauses in the constitution are primarily of declamatory value. They may, at the margins, provide electoral and educational authorities with slightly stronger grounds on which to promote electoral participation in their voter information and education campaigns, but in actuality this potential does not seem to translate into substantial increases in rates of voting. This is certainly not a reason for removing such provisions from existing constitutions, but it does suggest that on its own, the constitutionalisation of electoral obligation cannot normally be expected to boost turnout to any significant degree. Nor does it have any appreciable effects on other aspects of the political process.12 The recent Eastern European experience with turnout requirements – turnout thresholds below which an election is declared invalid – indicates that this institution is not promising either. Those states that have employed such thresholds tend to have turnout figures lower than the average for this region (Birch, 2003: 60–61). It is quite possible that these countries have retained legal thresholds due to fears generated by particularly alarming declines in rates of electoral participation, not that the thresholds themselves have actually depressed turnout, but there is no evidence that this type of legal requirement has had any positive influence on participation rates in this region. It is not difficult to see why this might be the case; collective sanctions for non-voting are unlikely to influence individual behaviour in communities as large as states or even constituencies (which is in some cases the level at which thresholds operate). Electoral authorities may be under greater pressure to ensure high turnout in order to prevent the election from having to be re-run, but in the absence of any concrete means of ensuring voters go to the polls, chronically low levels of electoral participation can persist over long periods of time. Thus collective sanctions in the form of turnout requirements are unlikely to be successful at raising turnout. They would have the disadvantage of requiring repeat elections to be held in the event that the initial poll failed to meet the requirement set. The final alternative to mandatory voting laws is what is commonly termed ‘incentive voting’. This device operates on the basis of the same individual

M1546 - BIRCH TXT.qxp:ANDY Q7

150

27/10/08

11:31

Page 150

Full participation

incentive principle as mandatory voting, but it relies on the use of ‘carrots’ rather than ‘sticks’. The idea is quite simple: rather than sanctioning voters for non-participation, the state rewards them for voting. Suggested rewards have included a reduction in tax (which would in practice be functionally equivalent to the imposition of fines for non-voting),13 lottery tickets for voters, or the direct provision of small gifts. Various types of reward have been used in different states at different times to entice voters to the polls. Citizens in ancient Athens were provided monetary inducements to encourage them to participate in the Athenian Assembly (Staveley, 1972: 78–9), and a glass of wine has traditionally been offered to Swiss voters in some cantons upon completion of their electoral duty. As noted in Chapter 1, Colombian voters have since 1997 been given preferential access to some state services, and there have recently been experiments with incentive voting elsewhere, including Bulgaria, as well as sub-national authorities in the US and Norway. The appeal of incentive voting is that it involves a positive encouragement rather than a negative punishment. Voters invariably go to some trouble to reach the polling station on election day (not to mention the cost associated with collecting and assessing the information necessary to decide among the options on the ballot). It perhaps stands to reason that they should receive at least a token reward for their troubles. It is therefore not surprising that incentive voting should have received some support from the proponents of compulsory voting (Feeley, 1974: 241), as well as from its opponents (Baston and Ritchie, 2004). Yet there are also a number of arguments that have been made against incentive voting. The principal argument is that as a civic duty, voting is not something for which voters should expect a material reward. A parallel is sometimes made with ‘vote-buying’, and all the negative connotations associated with that practice (see Hasen, 2000 for a detailed discussion of the ethics of vote-buying, including incentive voting of the type described here). Paying voters not to vote has, after all, been a frequent form of vote manipulation in many political contexts (Cox and Kousser 1981; Schaffer 2002; 2005). An additional problem noted by Hasen (2000: 1355–6) is that incentives can quite easily be targeted at specific groups if efforts are not made to ensure they are given to all voters, and that even if all voters benefit, the reward will be of greater value to the poor than to the rich (though the same argument can be made in reverse form about electoral fines). It is also often pointed out that the use of token rewards trivialises the voting act. When Jimmy Hood MP (Labour member from Clydesdale) tabled a private members bill to establish a commission to investigate incentive voting in the UK, he did so on the grounds that ‘[t]he stick of compulsion is too authoritarian, too dictatorial and too alien to our parliamentary democracy’ and that incentivising voting would provide a more appealing alternative in the British context, but John Gummer MP

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 151

Conclusion

151

criticised the idea of incentive voting as equating democracy with ‘the little man at the bottom of the cornflakes packet’ (Hansard, 29 November 2000).14 Finally, incentives do not appear to work nearly as well in practice as electoral compulsion, as the evidence cited in Chapter 1 suggests. Commenting on an incentive scheme used in the US state of California, where voters in one area were given donuts as a reward for exercising their right to vote, Richard Hasen analyses the normative advantages of sanctions over incentives: ‘sticks have the [. . .] advantage over carrots of making internalization of the voting requirement more likely. A half-dozen Yum-Yum doughnuts does not send the same message to a voter as does a government law compelling all to vote. The doughnut inspires an outcome-oriented calculus; the law suggests moral authority or social consensus’ (Hasen, 1996: 2172; see also Hasen, 2000: 1359; Hill, 2006: 211). The conclusion must be that none of these options appears to be as effective as compulsory voting in achieving the end of full participation. However, there are other possibilities that have not yet been explored, and which might prove more promising. These include requiring citizens to vote only at the first election for which they are eligible. Franklin (2004) makes a strong argument that early experiences with electoral participation are important in shaping behaviour later in life. If young people were required to vote at least once upon achieving the age of electoral majority, it is quite possible that this would inculcate in them a habit of participation that would be retained later in life. Another option would be to require a certain proportion of the electorate – chosen by lot – to participate. This is a version of the sort of thresholds employed in many Eastern European states, but in this case individuals would be identified for participation, and sanctions could be given to non-voters. The rationale behind this proposal is that universal turnout is not necessarily required for democratic decision-making and equitable policy outcomes. What is arguably required is that a sufficient number of people vote, and that they are sufficiently representative of the population at large. Whether either of these alternative proposals would work in practice is a moot point; the main purpose of suggesting them is to note that there are ways of ‘watering down’ compulsory voting so as to make it more acceptable to larger numbers of people, while at the same time potentially preserving some of its principal benefits. Policy implications The implications of this book’s findings for policy fall into two main categories: the conditions under which compulsory voting would be particularly beneficial, and the means through which it could most effectively be introduced. Both the normative arguments presented in Chapter 3 and the empirical findings summarised at the outset of this chapter suggest that on balance, compulsory

M1546 - BIRCH TXT.qxp:ANDY Q7

152

27/10/08

11:31

Page 152

Full participation

voting has positive impacts on the political process. In addition to increasing turnout, the evidence indicates that it fosters greater political participation, higher levels of satisfaction with democracy, greater equality in wealth distribution, and that it provides a check on the abuse of power. Yet this institution will be of more benefit in some contexts than in others. Countries experiencing low turnout and high levels of disaffection with democracy stand to benefit most from making electoral participation mandatory. This applies not only to many established democracies, but also to a large number of new and emerging democratic states. Subject to the caveats mentioned above, there are also a number of other reasons for believing that compulsory voting would be particularly beneficial in the context of emerging democracies. Political corruption is a problem that troubles many nascent democratic regimes, and the finding in Chapter 7 that compulsory voting is associated with reduced corruption suggests that countries experiencing problems with the abuse of power by public officials might benefit from this institution (in addition to other reforms). The chances of compulsory voting being successfully introduced obviously depend on the extent to which it is accepted by the population and perceived as legitimate. Political will on the part of the élites in government and civil society is undeniably crucial in this context. The existing evidence suggests that popular opinion about compulsory electoral participation can be quite volatile, which indicates that it is susceptible to shaping by opinion formers. In seeking to introduce this institution, it is also important to design the adoption so as to take into account possible obstacles. Legitimation through a popular referendum may well go a long way toward taking the wind out of the sails of possible organised opposition groups, as will allowing for a ‘none-of-the-above’ option on the ballot paper and a conscientious objector status for those who refuse to vote on principle. It is also advisable to ensure that making voting mandatory is accompanied by measures to ensure that the electoral act is as easy as possible. The range of possible vote-facilitation measures listed above ought, in combination, to enable virtually all those obliged to participate to be able to do so with relatively little personal inconvenience. In short, the introduction of compulsory voting is by no means unrealistic in most democratic contexts, and the institution has many potential benefits to offer contemporary states. The aim of this book has been to provide a conceptual and empirical overview of the institution of compulsory voting. Though this has been a longstanding topic of debate and deliberation in most democracies, many of the arguments most commonly made both for and against the legal obligation to participate in elections have taken the form of hypothetical speculation, given the paucity of evidence either to back the up or to refute them. This volume has sought to delineate the main arguments for and against, and to subject each of them to empirical tests as possible and appropriate. These tests have been

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 153

Conclusion

153

conducted on large-scale datasets including countries from across the world, and though they have not provided definitive answers to many of the questions posed by those curious about mandatory voting rules, they do together represent the most systematic and comprehensive analysis to date of the impact of such rules. That said, further research is clearly required, both to flesh out the bare bones of the admittedly rather cursory empirical investigations conducted here, as well as to explore the impact of compulsory voting in the context of specific polities. Rather than being the final word on the topic, this study will hopefully be a sign-post on the way to a renewed interest in bringing about full participation. Notes 1 There have also been reports of compulsory voting being debated in recent years in states as diverse as New Zealand, Jordan and India (see Hill and Louth, 2004: 4). 2 The 2005 British Election Study found that 77.1 per cent of respondents agreed with the proposition that ‘It is every citizen’s duty to vote in an election’. The British Election Study codebook and data can be found at www.essex.ac.uk/bes. The figure cited here is from variable bq22f, and it represents the combined frequencies for the responses ‘strongly agree’ and ‘agree’. 3 MORI website, http://www.ipsos-mori.com/polls/2—1/elec_comm1–6.shtml. 4 The other significant finding in this study was that men and those less satisfied with democracy were more likely to be in favour of change in this area. None of the country variables was significant. 5 See Patrick Wintour, ‘Hoon Calls for Compulsory Voting’, Guardian 4 July 2005; Hoon was supported in this call by fellow cabinet minister Peter Hain. 6 Debates of the Senate (Hansard), 1st Session, 38th Parliament, Vol. 142, issue 33, Wednesday 9 February 2005; Vol. 142, issue 61, Tuesday May 17, 2005. See also Harb, 2005. The proposal was accompanied by provision for a ‘none-of -the-above’ option on the ballot paper, as well as by the introduction of conscientious objector status. 7 Debates of the Senate (Hansard), 1st Session, 38th Parliament, Vol. 142, issue 33, Wednesday 9 February 2005; Vol. 142, issue 61. 8 Those advocating compulsory voting often favour a ‘none-of-the-above’ (or similar) option on the ballot (e.g. Feeley, 1974: 241–2; Wertheimer, 1975: 279, 293; Matsler, 2003: 955, 974–6; Watson and Tami, 2004; Hill, 2006: 223; Keaney and Rogers, 2006: 31–2). 9 Under some circumstances compulsory voting is unsuitable for logistic reasons; stories of Brazilians travelling two days in a canoe to vote (Power and Roberts, 1995: 819) suggest that requiring all citizens to vote may be asking more than is reasonable. 10 As noted in Chapter 1, Australia has a version of conscientious objector status; being a member of a religious order that prohibits voting is considered by the Australian Electoral Commission a valid reason for not voting.

M1546 - BIRCH TXT.qxp:ANDY Q7

154

27/10/08

11:31

Page 154

Full participation

11 In the UK context, the need for piloting is stressed by Watson and Tami (2004) and by Keaney and Rogers (2006: 32). Keaney and Rogers (2006: 32) also support the introduction of mandatory electoral participation following passage in a referendum. 12 The only significant coefficient for this variable was in the model of electoral support for small parties (Chapter 7, table 7.2, Model 6), where a constitutional obligation to vote is associated with a reduction in the support for parties polling less than five per cent of the vote. It is difficult to believe this finding is the result of a direct causal impact, however, as there are no theoretical grounds for believing this to be the case (the expectation based on previous findings was that, if anything, compulsory voting would increase support for small parties). As argued in Chapter 7, this result is most likely an artefact of country-specific factors. 13 In the UK context, Graham Allen MP (Lab) proposed the possibility of voters receiving a small tax or council tax credit in 2001 (cit. Lardy, 2004: 307f). 14 The bill was defeated by 143 votes to 79.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 155

Appendix: sources of data and variable construction

Legal regulations Sources: ACE Project Comparative Data database at http://aceproject.org/epic-en Inter-Parliamentary Union Parline database at http://www.ipu.org/parline-e/ parlinesearch.asp Legislationline database at http://www.legislationline.org Political Database of the Americas, Georgetown University at http://pdba. georgetown.edu/ Political Transformation and the Electoral Process in Post-Communist Europe database at www.essex.ac.uk/elections University of richmond.edu/

Richmond

Constitution

Finder

at

http://confinder.

Nohlen et al., 1999; Rose, 2000; Nohlen et al., 2001; Gratschew, 2002; 2004; Birch, 2003; Massicotte et al., 2004; Nohlen, 2005a; Reynolds et al., 2005; Payne et al., 2006. Individual-level data Sources The individual-level data used in this volume are taken from the June 2007 release of Module 2 of the Comparative Study of Electoral Systems Project. Data details can be found at http://www.cses.org/download/module2/ module2.htm. The June 2007 full release of the Module 2 dataset includes 41 cases. Of these, four (France 2002, Kyrgyzstan 2005, Russia 2004, and Taiwan 2004) were removed because they are presidential elections only, which lessens comparability and precludes testing of the impact of parliamentary electoral system variables. Hong Kong 2004 was removed because Hong Kong is not a

M1546 - BIRCH TXT.qxp:ANDY Q7

156

27/10/08

11:31

Page 156

Full participation

sovereign state, which means that a number of relevant macro-level variables are not available for this case. Two Portuguese elections are included in this dataset – 2002 and 2005. The 2005 elections only are included in the dataset used for analysis, so as to avoid possible temporal dependencies in the data. Finally, two German surveys are included in the CSES dataset: a telephone survey and a mail-back survey. The telephone survey only is included in the dataset analysed here, given its greater variable coverage. The resulting dataset thus includes 34 surveys and 55,029 cases.1 Variable construction Political information: Three political information questions were asked in the CSES surveys. The questions themselves varied from country to country; they were designed to be of equivalent difficulty (see Brockington, 2005 and Milner and Grönlund, 2006; for details). Correct answers to the three questions (B3047_1, B3047_2 and B3047_3) were first combined to form a fourpoint (0–3) scale (as is conventional, ‘don’t know’ answers were treated as incorrect). This scale was then dichotomised (0,1=0; 2,3=1) for ease of model interpretation. Efficacy (B3014): The measure of (external) efficacy was constructed on the basis of the question ‘Some people say that no matter who people vote for, it won’t make any difference to what happens. Others say that who people vote for can make a difference to what happens. Using the scale on this card (where ONE means that voting won’t make a difference to what happens and FIVE means that voting can make a difference), where would you place yourself?’ As suggested by the question wording, this variable was constructed as a five-point scale. A binary version of the variable (in which responses ‘4’ and ‘5’ are coded ‘1’ and responses ‘1’, ‘2’ and ‘3’ are coded ‘0’ was created for use as a dependent variable, in order to facilitate interpretation. Political discussion and persuasion (B3001_1): This variable was constructed on the basis of a question worded as follows: ‘Here is a list of things some people do during elections. Which if any did you do during the most recent election?. . . talked to other people to persuade them to vote for a particular party or candidate?. . . showed your support for a particular party or candidate by, for example, attending a meeting, putting up a poster, or in some other way?’ Answers to the first of these questions were coded as a dummy variable. Voter-initiated contact with politicians (B3042_1); Work with others (B3042_2); Participation in protests or demonstrations (B3042_3): These political participation indicators were coded as dummy variables on the basis of the following question: ‘Over the past five years or so, have you done

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 157

Appendix: sources of data and variable construction

157

any of the following things to express your views about something the government should or should not be doing? (Have you) contacted a politician or government official either in person, or in writing or some other way? (Have you) taken part in a protest, march or demonstration? (Have you) worked with people who shared the same concern?’ Summary participation indicator: In addition, a summary dummy variable was constructed to designate that a respondent engaged in at least one of the three aforementioned types of political participation (contacting politicians, working with others or participation in protests or demonstrations). Party identification (B3028 & B3034): A dummy variable designating party identification was constructed on the basis of the following two questions: ‘Do you usually think of yourself as close to any particular political party?’, and the follow-on question asked of those respondents who answered ‘no’ or ‘don’t know’ to this question: ‘Do you feel yourself a little closer to one of the political parties than the others?’. Participation in campaign activities (B3001_2): Political discussion (B3001_1): This variable was constructed on the basis of a question worded as follows: ‘Here is a list of things some people do during elections. Which if any did you do during the most recent election?. . . talked to other people to persuade them to vote for a particular party or candidate?. . . showed your support for a particular party or candidate by, for example, attending a meeting, putting up a poster, or in some other way?’ Answers to the second of these questions were coded as a dummy variable. Contacted by party or politician (B3003): This indicator was constructed as a dummy variable on the basis of the following question: ‘During the last campaign did a candidate or anyone from a political party contact you to persuade you to vote for them?’. Satisfaction with democracy (B3012): This indicator was constructed as a dummy variable on the basis of responses to the following question: ‘On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied with the way democracy works in [country]?’ Replies of ‘very satisfied’ and ‘fairly satisfied’ were coded ‘1’, while replies of ‘not very satisfied’ and ‘not at all satisfied’ were coded ‘0’. Quality of representation (B3022): This indicator was constructed as a dummy variable on the basis of responses to the following question: ‘Thinking about how elections in [country] work in practice, how well do elections

M1546 - BIRCH TXT.qxp:ANDY Q7

158

27/10/08

11:31

Page 158

Full participation

ensure that the views of voters are represented by Majority Parties: very well, quite well, not very well, or not well at all?’ Replies of ‘very well’ and ‘quite well’ were coded ‘1’, while replies of ‘not very well’ and ‘not well at all’ were coded ‘0’. The demographic variables age (B2001), gender (B2002), education (B2003) and religiosity (B2024) were all standardised to a 0–1 scale. Gender was coded such that 0=male, 1=female. Aggregate-level data Cases: elections to lower or only chambers of parliament, 1990–2006, except where otherwise noted. Sources of electoral data: IDEA Voter Turnout Database at http://www.idea. int/vt/index.cfm; Inter-Parliamentary Union Parline database at http://www. ipu.org/parline-e/parlinesearch.asp; IFES Election Guide at www.electionguide.org; Political Database of the Americas, Georgetown University, at http://pdba.georgetown.edu/; Election Results Archive of the Center on Democratic Performance, University of Binghamton at http://www.binghamton .edu/cdp/era/index.html; African Elections Database at http://africanelections.tripod.com/; Parties and Elections in Europe database at http://www. parties-and-elections.de/index.html; Political Transformation and the Electoral Process in Post-Communist Europe database at www.essex.ac.uk/elections; IDEA, 2002; Birch, 2003; Shafferer, 2003; Reynolds et al., 2005; Benoit and Laver, 2006. Variable construction of electoral data Compulsory voting was operationalised in terms of three dummy variables, one designating the existence or absence of compulsory voting with or without sanctions, a second variable indicating the existence or absence of compulsory voting with sanctions, and a third representing the presence or absence of a constitutional provision describing voting as a duty or obligation (note that this includes a number of states that are not considered compulsory voting under the definition adopted here: the Central African Republic, Colombia, Cuba, East Timor, Haiti, Italy, Mozambique and Portugal). For Austria 1990 and Switzerland, the value of each variable is weighted to reflect the proportion of the country that practices the institution in question (based on data, respectively, from Demographisches Jahrbuch 2005, Statistik Austria, Vienna, 2006 and Demographisches Porträt der Schweiz, Ausgabe 2004, Swiss Federal Statistical Office, Neuchâtel, 2004). For Egypt, the variables designating compulsory voting are set at 0.5, as voting is only compulsory for men.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 159

Appendix: sources of data and variable construction

159

Effective number of parliamentary parties was the conventional measure derived from Laakso and Taagepera (1979) as the inverse of the sum of squares of party seat proportions. Electoral system was operationalised in terms of a dummy variable designating an election conducted under a proportional rule. Party block vote systems were coded as non-proportional, whereas other multi-member systems were coded as proportional. Mixed and majoritarian systems were initially coded separately, but the two types of system were found in preliminary models to have similar effects, so they were combined in to a single (non-proportional) category in the final models reported here. Turnout: Aggregate turnout is calculated as a proportion of the registered electorate. In multi-round systems, first-round turnout figures are employed. Concurrent presidential elections was coded as a dummy variable indicating whether (one of the rounds of the) parliamentary elections coincided with (one of the rounds of) a presidential election. Margin of victory: Following Geys, 2006, margin of victory was operationalised as the difference in vote shares between the larges party and the second largest party. However, because vote share data were not available for all races included in this study, difference in seat shares is employed instead of difference in vote shares. In mixed compensatory systems, the proportional representation vote share was employed. In two-round systems, first round vote data were used. Directly elected seats only were included in calculations. The following cases were excluded because during the period under analysis elections were not conducted on a multiparty basis: Afghanistan, Bahrain, Cuba, Iran, Kuwait, Laos, Maldives, Marshall Islands, Micronesia, Nauru, Oman, Palau, Swaziland, Tonga, Turkmenistan, Tuvalu, Vietnam, Chad, 1990, Djibouti 2003, Iraq 1996 and 2000, Jordan 1997 2003, Kiribati 1991, Malawi 1992, Sudan 1996, Tanzania 1990, Togo 1990, Uganda 1996, 2001. Proportion of the vote for small parties: Small parties were defined for the purpose of this analysis as those that won less than five per cent of the popular vote. Proportion of the vote for left parties: Dimension 13 of the Benoit-Laver (2006) dataset codes political parties in 47 countries according to a 1–20 scale, where ’1’ is the left extreme and ‘20’ is the right extreme. The parties in this dataset were coded either ‘left’ or ‘right’, depending whether they fell in the 1–10 of 11–20 portion of the spectrum. The vote share of left parties was then

M1546 - BIRCH TXT.qxp:ANDY Q7

160

27/10/08

11:31

Page 160

Full participation

calculated on the basis of data provided in the Benoit-Lever dataset. This dataset includes all parties that meet one of three criteria: they have representation in the national parliament, they gained at least one per cent of the vote in the relevant elections, or they met neither of the first two criteria but were nevertheless deemed ‘politically relevant’ by the expert assessors (see Benoit and Laver, 2006 for details). Proportion of the vote for far-right parties: This variable was constructed on the basis of the Benoit-Laver (2006) dataset described above. Parties with a mean left-right score of greater than 18 were coded as ‘far right’. The proportion of the vote share gained by such parties was calculated on the basis of data provided in the Benoit-Laver dataset. It should be noted that the inclusion criteria (see above) exclude a number of very small far-right parties in several countries. Given that the total vote for these parties rarely exceed one per cent, this should not have a major impact on the findings reported here. Sources and variable construction of other data Freedom House ‘Electoral Process’ score: This score ranges from 0 to 12. The methodology subtending this score can be found at: http://www.freedomhouse.org. Corruption: This was measured by the Transparency International Corruption Perceptions Index, see www.transparency.org for details.This index is ranges from 0 (most corrupt) to 10 (least corrupt); see the TI website for methodological details. Data were taken from 2006 to maximise country coverage. Per capita Gross Domestic Product (logged): Data are in US dollars; due to data availability constraints, 2005 data are used for elections occurring in 2006. Sources: United Nations Monthly Bulletin of Statistics Online, at http://unstats.un.org/unsd/mbs/. Per capita GDP for Taiwan is taken from the Taiwanese office of National Statistics website at http://eng.stat.gov. tw/mp.asp?mp=5. Population size: Figures are thousands of inhabitants, and they are taken from the nearest year for which data are available. Sources: World Bank World Development Indicators at http://web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/0,,contentMDK:2053528 5~menuPK:1192694~pagePK:64133150~piPK:64133175~theSitePK:23 9419,00.html; United Nations World Population Prospects at http://esa.un.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 161

Appendix: sources of data and variable construction

161

org/unpp/; Population data for Taiwan are taken from the Taiwan Yearbook 2006, at http://www.gio.gov.tw/taiwan-website/5-gp/yearbook. Level of democracy was measured as the combined Political Rights and Civil Liberties ratings for a country in a given year. The combined score was then inverted to create a scale, ranging from 2 to 14, on which higher scores represent higher levels of democracy. Source: Freedom House Freedom in the World archive available at http://www.freedomhouse.org/template.cfm?page=15. Female representation: The proportion members of the lower or only house of the national assembly who are female. Source: Inter-Parliamentary Union Parline database at http://www.ipu.org/ parline-e/parlinesearch.asp. Gender quota: This indicator was constructed as a dummy variable designating the existence of a gender quota either in the formal legal system or operated by political parties. Source: Global Database of Quotas for Women, International IDEA and Stockholm University, www.quotaproject.org. Female education: This variable indicates girls’ share of secondary enrolment. These data are from 2001–2005 depending on the country. The data points do not in most cases correspond to individual election years, but as this indicator is used as a proxy for the cultural standing of women, it is an indicator of longterm cultural trends that are unlikely to be affected by short-term fluctuations. Source: United Nations Demographic and Social Statistics at http://unstats.un. org/unsd/demographic/. Female labour force participation: This variable indicates that share of women in the adult labour force. These data are from various time points at or around the millennium, depending on the country. The data points do not in most cases correspond to individual election years, but as this indicator is used as a proxy for the cultural standing of women, it is an indicator of long-term cultural trends that are unlikely to be affected by short-term fluctuations. Source: Source: United Nations Demographic and Social Statistics at http://unstats.un.org/unsd/demographic/.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

162

11:31

Page 162

Full participation

Immigrant population: The proportion of foreign-born residents. Data are in the main from 2000–2002. Sources: UN Demographic Yearbook Special Census Topics, Vol. 3, March 2006, at http://unstats.un.org/unsd/demographic; Jean-Christophe Dumont and Georges Lemaître, ‘Counting Immigrants and Expatriates in OECD Countries: A New Perspective’, OECD, 2004, at http://www.oecd.org/ dataoecd/27/5/33868740.pdf. Unemployment: The proportion of the workforce unemployed in the year of the election. Source: International http://laborsta.ilo.org/.

Labour

Organization

Laborsta

database

at

Ethnic fractionalisation: See Alesina et al. 2003 for details. Source: Alesina et al., 2003. Income inequality (Gini Index): Data are taken from the World Bank World Development Indicators Dataset, 2005. The data are in virtually all cases taken from the period 1995–2004. Source: World Bank website at http://devdata.worldbank.org/wdi2005/ Section2.htm. Note 1 The Japanese election of 2004 was for (half of) the upper chamber; all the other elections were for lower or lower and upper chambers, and in some cases for presidents as well.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 163

References

Abraham, Henry J. (1952), ‘Compulsory Voting: Its Practice and Theory’, PhD Dissertation, University of Pennsylvania, Philadelphia, Pennsylvania. Abraham, Henry J. (1955), Compulsory Voting, Annals of American Government, Washington, DC: Public Affairs Press. Ackaert, J. (1992), ‘Electoral Absenteeism and the Blank and Invalid Vote in Belgium’, Res Publica 34.2, pp. 209–26. Ackaert, Johan, de Winter, Lieven, Aish, Anne-Marie and Frognier, André-Paul (1992), ‘L’Abstentionnisme électoral et vote blanc et nul en Belgique’, Res Publica. Aitkin, Don, Jinks, Brian and Warhurst, John (1986), Australian Political Institutions, Melbourne: Longman Cheshire. Alapuro, Risto (2005), ‘Suffrage and the Construction of the Voter in Finland (and other Nordic Countries)’, Paper presented at the international symposium on compulsory voting, Institut d’Etudes Politiques de Lille, Lille 20–21 October, 2005. Alesina, Alberto, Devleeschauwer, Arnaud, Easterly, William and Kurlat, Sergio (2003), ‘Fractionalization’, Journal of Economic Growth 8, pp. 155–94. Anderson, Christopher J. and Guillory, Christine A. (1997), ‘Political Institutions and Satisfaction with Democracy: A Cross-National Analysis of Consensus and Majoritarian Systems’, American Political Science Review 91.1, pp. 66–81. Anderson, Christopher J. and Tverdova, Yulia V. (2003), ‘Corruption, Political Allegiances, and Attitudes Toward Government in Contemporary Democracies’, American Journal of Political Science 47.1, pp. 91–109. Anduiza Perea, Eva (2002), ‘Individual Characteristics, Institutional Incentives and Electoral Abstention in Western Europe’, European Journal of Political Research 41, pp. 643–73. Asian Network for Free Elections [ANFREL] (2005), ‘Thailand General Election 2005: Report of International Observation Missions January-February 2005’, Bangkok. Baeza, Mario Fernández (1998), ‘El Voto Obligatorio en América Latina’, in Nohlen, Dieter, Picado, Sonia and Zovatto, Daniel (eds.), Tratado de Derecho Electoral Comparado de América Latina, Mexico City: IIDH/CAPEL. Ballinger, Chris (2006), Democracy and Voting, London: Hansard Society. Ballinger, Chris (2007), ‘Compulsory Voting: Palliative Care for Democracy in the UK?’, Paper presented at the ‘Compulsory Voting: Principles and Practice’ workshop,

M1546 - BIRCH TXT.qxp:ANDY Q7

164

27/10/08

11:31

Page 164

References

European Consortium for Political Research Joint Sessions of Workshops, Helsinki, 7–12 May 2007. Banducci, Susan A. and Karp, Jeffrey (2003), ‘How Elections Change the Way Citizens View the Political System: Campaigns, Media Effects and Electoral Outcomes in Comparative Perspective’, British Journal of Political Science 33.3, pp. 443–67. Barber, Benjamin (1984), Strong Democracy: Participatory Politics for a New Age, Berkeley, Los Angeles and London: University of California Press. Barnes, William A. (1998), ‘Incomplete Democracy in Central America: Polarization and Voter Turnout in Nicaragua and El Salvador’, Journal of Interamerican Studies and World Affairs 40.3, pp. 63–101. Baston, Lewis and Ritchie, Ken with Mortimore, Roger and Gill, Mark (2004), Turning Out or Turning Off? An Analysis of Political Disengagement and What Can Be Done About It, London: Electoral Reform Society. Bélanger, Eric (2004), ‘Antipartyism and Third-Party Vote Choice: A Comparison of Canada, Britain and Australia’, Comparative Political Studies 37.9, pp. 1054–1078. Bendel, Petra and Krennerich, Michael (2005), ‘Guatemala’, in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. I North America, Central America, and the Caribbean, Oxford: Oxford University Press, pp. 317–52. Bennett, Scott (2005), ‘Compulsory Voting in Australian National Elections’, Research Brief No. 6 2005–06, Parliament of Australia Library. Benoit, Kenneth and Laver, Michael (2006), Party Policy in Modern Democracies, London: Routledge. Berggren, Heidi M. (2001), ‘Institutional Context and Reduction of Resource Bias in Political Sophistication’, Political Research Quarterly 54.3, pp. 531–52. Bernhagen, Patrick and Marsh, Michael (2007), ‘The Partisan Effects of Low Turnout: Analyzing Vote Abstention as a Missing Data Problem’, Electoral Studies 26.3, pp. 548–60. Betz, Hans-Georg (1994), Radicalism and Right-Wing Populism in Western Europe, Basingstoke: Macmillan. Bilodeau, Antoine and Blais, André (2005), ‘Le vote obligatoire exerce-t-il un effet de socialisation politique?’, Paper presented at the international symposium on compulsory voting, Institut d’Etudes Politiques de Lille, Lille 20–21 October, 2005. Birch, Sarah (2003), Electoral Systems and Political Transformation in Post-Communist Europe, Basingstoke: Palgrave-Macmillan. Birch, Sarah (2007), ‘Electoral Systems and the Manipulation of Elections’, Comparative Political Studies, 40.12, pp. 1533–56. Blais, André (2000), To Vote to Not to Vote: The Merits and Limits of Rational Choice Theory, Pittsburgh, PA: University of Pittsburgh Press. Blais, André (2006), ‘What Affects Voter Turnout?’, Annual Review of Political Science 9, pp. 111–25. Blais, André and Dobrz-ynska, Agnieszka (1998), ‘Turnout in Electoral Democracies’, European Journal of Political Research 33, pp. 239–61. Boix, Carles (1999), ‘Setting the Rules of the Game: The Choice of Electoral Systems in Advanced Democracies’, American Political Science Review 93.3, pp. 609–24. Bouillin-Dartevelle, R., Claeys, P.-H., Dawance-Goossens, J., Delruelle-Vosswinkel, N.,

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 165

References

165

Loeb-Mayer, N., Noël-Aranda, M. C., Thoveron, G. (1984), Les Elections législatives du 8 novembre 1981: la Nouvelle rupture, Brussels: Editions de l’Université de Bruxelles. Bowler, Shaun and Donovan, Todd (2002), ‘Democracy, Institutions and Attitudes about Citizen Influence on Government’, British Journal of Political Science 32.2, pp. 371–90. Bowler, Shaun and Donovan, Todd and Karp, Jeffrey A. (2006), ‘Why Politicians Like Electoral Institutions: Self-Interest, Values, or Ideology?’, Journal of Politics 68.2, pp. 434–46. Bowler, Shaun, Lanoue, David J. and Savoie, Paul (1994), ‘Electoral Systems, Party Competition and Strength of Partisan Attachment: Evidence from Three Countries’, Journal of Politics 56.4, pp. 991–1007. Brader, Ted and Tucker, Joshua A. (2001), ‘The Emergence of Mass Partisanship in Russia, 1993–1996’, American Journal of Political Science 45.1, pp. 69–83. Brader, Ted and Tucker, Joshua A. (2002), ‘Pathways to Partisanship in New Democracies: Evidence from Russia’, unpublished manuscript. Brea Franco, Julio (2005), ‘Dominican Republic’, in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. I North America, Central America, and the Caribbean, Oxford: Oxford University Press, pp. 239–68. Brockington, David (2003), ‘A Low Information Theory of Ballot Position Effect’, Political Behavior 25.1, pp. 1–27. Brockington, David (2005), ‘Compulsory Voting and Political Information: A CrossNational Examination’, paper delivered at the Annual Meeting of the American Political Science Association Washington, DC, 1–4 September, 2005. van der Brug, Wouter, Fennema, Meindert and Tillie, Jean (2005), ‘Why Some AntiImmigrant Parties Fail and Others Succeed’, Comparative Political Studies 38.5, pp. 537–73. Brunner, Georg (1990), ‘Elections in the Soviet Union’, in Furtak, Robert K. (ed.), Elections in Socialist States, New York and London: Harvester Wheatsheaf, pp. 20–52. Bryce, James (1921), Modern Democracies, New York: Macmillan. Bundesministerium für Inneres [Austria] ‘Elections in Austria – Compulsory Voting’, www.bmi.gv.at/wahlen/elections_compulsory_voting.asp, 2006. Callahan, William A. (2000), Pollwatching, Elections and Civil Society in Southeast Asia, Aldershot: Ashgate. Campbell, Angus, Converse, Philip, Miller, William and Stokes, Donald (1960), The American Voter, New York: Wiley. Canache, Damarys, Mondak, Jeffrey J. and Seligson, Mitchell A. (2001), ‘Meaning and Measurement in Cross-National Research on Satisfaction with Democracy’, Public Opinion Quarterly 65, pp. 506–28. Caramani, Daniele (2000), The Societies of Europe: Elections in Western Europe since 1815: Electoral Results by Constituencies, Basingstoke: Macmillan. Carlin, Ryan E. (2006), ‘The Decline of Citizen Participation in Electoral Politics in Postauthoritarian Chile’, Democratization 13.4, pp. 632–51. Carson, George Barr (1955), Electoral Practices in the U.S.S.R., New York: Frederick A. Praeger. Carstairs, Andrew McLaren (1980), A Short History of Electoral Systems in Western Europe, London: George Allen and Unwin.

M1546 - BIRCH TXT.qxp:ANDY Q7

166

27/10/08

11:31

Page 166

References

Carter, Elisabeth (2002), ‘Proportional Representation and the Fortunes of Right-Wing Extremist Parties’, West European Politics 25.3, pp. 125–46. Centre for Research and Information on Canada/Centre de Recherche et d’Information sur le Canada [CRIC] (2001), ‘Voter Participation in Canada: Is Canadian Democracy in Crisis?’, CRIC Paper No. 3, Montréal: Centre for Research and Information on Canada/Centre de Recherche et d’Information sur le Canada, 2001. Chong, Alberto and Olivera, Mauricio (2005), ‘On Compulsory Voting and Income Inequality in a Cross-Section of Countries’, Inter-American Development Bank Research Department, Working Paper No. 533. Cox, Gary (1997), Making Votes Count: Strategic Coordination in the World’s Electoral Systems, Cambridge: Cambridge University Press. Cox, Gary (1999), ‘Electoral Rules and the Calculus of Mobilization’, Legislative Studies Quarterly 24.3, pp. 387–417. Cox, Gary W. and Kousser, J. Morgan (1981), ‘Turnout and Rural Corruption: New York as a Test Case’, American Journal of Political Science 25.4, pp. 646–63. Crain, W. Mark (1995), ‘The Right Versus the Obligation to Vote: Rejoinder to Yeret, and O’Toole and Strobl’, Economics and Politics 7.3, pp. 281–7. Crain, W. Mark and Leonard, Mary L. (1993), ‘The Right Versus the Obligation to Vote: Effects on Cross-Country Government Growth’, Economics and Politics 5.1, pp. 43–51. Crampton, R. J. (1983), Bulgaria 1878–1918: A History, New York, NY: Colombia University Press. Crepaz, Markus M. L. (1990), ‘The Impact of Party Polarization and Postmaterialism on Voter Turnout: A Comparative Study of 16 Industrial Democracies’, European Journal of Political Research 18.2, pp. 183–205. Crewe, Ivor, ‘Electoral Participation’ in Butler, David, Penniman, Howard R. and Ranney, Austin (eds.) (1981), Democracy at the Polls: A Comparative Study of Competitive National Elections, Washington and London: American Enterprise for Public Policy Research, pp. 216–63. Czesnik, Mikolaj (2007), ‘Is Compulsory Voting a Remedy? Evidence from the 2001 Polish Parliamentary Elections’, Paper presented at the ‘Compulsory Voting: Principles and Practice’ workshop, European Consortium for Political Research Joint Sessions of Workshops, Helsinki, 7–12 May 2007. Dalton, Russell J. (1996), Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies, 2nd ed., Chatham, NJ: Chatham House. Dalton, Russell J. (2000), ‘The Decline of Party Identifications’, in Dalton, Russell J. and Wattenberg, Martin P. (eds.), Parties without Partisans: Political Change in Advanced Industrial Democracies, Oxford: Oxford University Press, pp. 19–36. Dalton, Russell J. (2004), Democratic Challenges, Democratic Choices: The Erosion of Political Support in Advanced Industrial Democracies, Oxford: Oxford University Press. Deininger, Klaus W. and Squire, Lyn (1996), ‘A New Dataset Measuring Income Inequality’, World Bank Economic Review 10.3, pp. 565–91. Déloye, Yves (2005), ‘Chronique d’une allergie républicaine. L’obligation électorale contre le vote obligatoire’, paper delivered at the symposium on ‘Le Vote obligatoire’, CEPEN, Lille, 20–21 October, 2005. Deschouwer, Kris and de Winter, Lieven (1998), ‘La corruption politique et le clien-

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 167

References

167

télisme; le spectre italien?’, in Marco Martiniello ad Marc Swyngedouw (eds.), Où va la Belgique?, Paris: L’Harmattan, pp. 139–52. Díaz, Honorario A. (1983), Ley Sáenz Peña, Buenos Aires, Centro Editor de America Latina. van der Eijk, Cees and van Egmond, Marcel (2007), ‘Political Effects of Low Turnout in National and European Elections’, Electoral Studies 26.3, pp. 561–73. The Electoral Commission (2006), Compulsory Voting Around the World, London: The Electoral Commission. Elkins, Zachary (2000), ‘Quem iria votar? Conhecendo as conseqüências do voto obrigatório no Brasil’, Opinião pública 6.1, pp. 109–36. Ellner, Steve (1993), ‘The Deepening of Democracy in a Crisis Setting: Political Reform and the Electoral Process in Venezuela’, Journal of Interamerican Studies and World Affairs 53.4 (1993–1994), pp. 1–42. Enelow, James and Hinich, Melvin J. (1984), The Spatial Theory of Voting, Cambridge: Cambridge University Press. Engelen, Bart (2007), ‘Why Compulsory Voting Can Enhance Democracy’, Acta Politica 42, pp. 23–39. Engelen, Bart and Hooghe, Marc (2007), ‘Compulsory Voting and its Effects on Political Participation, Interest and Efficacy’, Paper presented at the ‘Compulsory Voting: Principles and Practice’ workshop, European Consortium for Political Research Joint Sessions of Workshops, Helsinki, 7–12 May 2007. Escobar Armas, Carlos (1986), ‘Guatemala’, in Hernández Becerra, Augusto (ed.), Legislación Electoral Comparada : Colombia, México, Panamá, Venezuela y Centroamérica, San José, Costa Rica: Centro de Asesoría y Promoción Electoral, Instituto Interamericano de Derechos Humanos, Instituto de Investigaciones Jurídicas, Universidad Nacional Autónoma de México. Farrell, David M. and Webb, Paul (2000), ‘Political Parties as Campaign Organizations’, in Dalton, Russell J. and Wattenberg, Martin P. (eds.), Parties without Partisans; Political Change in Advanced Industrial Democracies, Oxford: Oxford University Press, 102–28. Feeley, Malcolm M. (1974), ‘A Solution to the “Voting Dilemma” in Modern Democratic Theory’, Ethics 84.3, pp. 235–42. Finkel, Steven E. (1985), ‘Reciprocal Effects of Participation and Political Efficacy: A Panel Analysis’, American Journal of Political Science 29.4, pp. 891–913. Finkel, Steven E. (1987), ‘The Effects of Participation on Political Efficacy and Political Support: Evidence from a West German Panel’, Journal of Politics 49.2, pp. 441–64. Finkel, Steven E. (2002), ‘Civic Education and the Mobilization of Political Participation in Developing Democracies’, Journal of Politics 64.4, pp. 994–1020. Fisher, Stephen D. (2007), ‘(Change in) Turnout and (Change in) the Left Share of the Vote’, Electoral Studies 26.3, pp. 598–611. Fornos, Carolina, Power, Timothy and Garand, James (2004), ‘Explaining Voter Turnout in Latin America, 1980 to 2000’, Comparative Political Studies 37.8, pp. 909–40. Fraeys, William (2003), ‘Les élections législatives du 18 mai 2003; Anayse des résultats’, Res Publica 2–3, pp. 379–400.

M1546 - BIRCH TXT.qxp:ANDY Q7

168

27/10/08

11:31

Page 168

References

Fraga, Rosendo (1992), Argentina en las Urnas 1931–1991, Buenos Aires: Centro de Estudios Unión para la Nueva Mayoría. Franklin, Mark N. (1996), ‘Electoral Participation’, in LeDuc, Lawrence, Niemi Richard G. and Norris, Pippa (eds.), Comparing Democracies: Elections and Voting in Global Perspective, Thousand Oaks, CA, London and New Delhi: Sage, pp. 216–35. Franklin, Mark N. (1999), ‘Electoral Engineering and Cross-National Turnout Differences: What Role for Compulsory Voting?’, British Journal of Political Science 29.1, pp. 205–16. Franklin, Mark N. (2002), ‘The Dynamics of Electoral Participation’, in LeDuc, Lawrence, Niemi, Richard G. and Norris, Pippa (eds.), Comparing Democracies 2: New Challenges in the Study of Elections and Voting, London, Thousand Oaks, CA and New Delhi: Sage, pp. 148–66. Franklin, Mark N. (2004), Voter Turnout and the Dynamics of Electoral Competition in Established Democracies Since 1945, Cambridge: Cambridge University Press. Friedgut, Theodore H. (1979), Political Participation in the USSR, Princeton, NJ: Princeton University Press. Funk, Patricia (2004), ‘Is There an Expressive Function of Law? An Empirical Analysis of Voting Laws with Symbolic Fines’, Social Science Research Network paper, available at http://papers.ssrn.com. Gallego, Aina (2007), ‘Does Compulsory Voting Reduce Inequality in Turnout? An Examination of its Impact on Educational Biases’, Paper presented at the ECPR Joint Sessions, Helsinki, 7–12 May, 2007. Gerring John and Thacker, Strom C. (2004), ‘Political Institutions and Corruption: The Role of Unitarism and Parliamentarism’, British Journal of Political Science, 34.2, pp. 295–330. Geys, Benny (2006), ‘Explaining Voter Turnout: A Review of Aggregate-Level Research’, Electoral Studies 25.4, pp. 637–663. Ginsberg, Benjamin and Weissberg, Robert (1978), ‘Elections and the Mobilization of Popular Support’, American Journal f Political Science 22.1, pp. 31–55. Giraud, Emile (1931), ‘Le Vote obligatoire du point de vue des principes et du bon fonctionnement des institutions représentatives’ Review du Droit Public et de la Science Politique en France et a l’Etranger, July-August-September, 1931. Gluchowski, Peter and Grotz, Florian (2001), ‘Mongolia’, in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. II South East Asia, East Asia and the South Pacific, Oxford: Oxford University Press, pp. 481–523. Golder, Matthew (2003), ‘Explaining Variation in the Success of Extreme Rights Parties in Western Europe’, Comparative Political Studies 36.4, pp. 432–66. Goodwin-Gill, Guy S. (1994), Free and Fair Elections: International Law and Practice, Geneva: Inter-Parliamentary Union. Gordon, Stacy B. and Segura, Gary M. (1997), ‘Cross-National Variation in the Political Sophistication of Individuals: Capability or Choice?’, Journal of Politics 59.1, pp. 126–47. Gosnell, Harold F. (1930), Why Europe Votes, Chicago: University of Chicago Press.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 169

References

169

Gow, Neil (1971), ‘The Introduction of Compulsory Voting in the Australian Commonwealth’, Politics 6.2, pp. 201–10. Gratschew, Maria (2002), ‘Compulsory Voting’, in Rafael López Pintor and Maria Gratschew (eds.), Voter Turnout since 1945: A Global Report, Stockholm: International IDEA, pp. 105–110. Gratschew, Maria (2004), ‘Compulsory Voting in Western Europe’, in Rafael López Pintor and Maria Gratschew (eds.), Voter Turnout in Western Europe since 1945, Stockholm: International IDEA, pp. 25–31. Gray, Mark and Caul, Miki (2000), ‘Declining Voter Turnout in Advanced Industrial Democracies, 1950 to 1997: The Effects of Declining Group Mobilization’, Comparative Political Studies 33.9, pp. 1091–1122. Griffin, John D. and Newman, Brian (2005), ‘Are Voters Better Represented?’, Journal of Politics 67.4, pp. 1206–1227. Grönlund, Kimmo and Milner, Henry (2006), ‘The Determinants of Political Knowledge in Comparative Perspective’, Scandinavian Political Studies 29.4, pp. 386–406. Harb, Mac (2005), ‘The Case for Mandatory Voting in Canada’, Canadian Parliamentary Review, Summer 2005, pp. 4–6. Hartmann, Christof (2001a), ‘Laos’, in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. II South East Asia, East Asia and the South Pacific, Oxford: Oxford University Press, pp. 129–41. Hartmann, Christof (2001b), ‘Fiji Islands’, in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. II South East Asia, East Asia and the South Pacific, Oxford: Oxford University Press, 643–72. Hartmann, Christof, Hassall, Graham and Santos, Soliman M. Jr., ‘Philippines’ in Nohlen, Dieter, Grotz, Florian and Hartmann Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. II South East Asia, East Asia and the South Pacific, Oxford; Oxford University Press, 2001, pp. 185–238. Hasen, Richard (1996), ‘Voting without Law?’ University of Pennsylvania Law Review 144.5, pp. 2135–2179. Hasen, Richard (2000), ‘Vote Buying’, California Law Review 88.5, pp. 1323–71. Helmke, Gretchen and Meguid, Bonnie (2007), ‘Endogenous Institutions: The Origins of Compulsory Voting Laws’, paper presented at the Annual Meeting of the American Political Science Association, Chicago, 30 August – 2 September 2007. Hernández Becerra, Augusto (1986), ‘Colombia’, in Hernández Becerra, Augusto (ed.), Legislación Electoral Comparada : Colombia, México, Panamá, Venezuela y Centroamérica, San José, Costa Rica: Centro de Asesoría y Promoción Electoral, Instituto Interamericano de Derechos Humanos, Instituto de Investigaciones Jurídicas, Universidad Nacional Autónoma de México. Hicks, Alexander M. and Swank, Duane H. (1992), ‘Politics, Institutions, and Welfare Spending in Industrialized Democracies, 1960–82’, American Political Science Review 86.3, pp. 658–74. Highton, Benjamin and Wolfinger, Raymond E. (2001), ‘The Political Implications of Higher Turnout’, British Journal of Political Science 31.1, pp. 179–92. Hill, Lisa (2002a), ‘On the Reasonableness of Compelling Citizens to ‘Vote’: The Australian Case’, Political Studies, pp. 80–101.

M1546 - BIRCH TXT.qxp:ANDY Q7

170

27/10/08

11:31

Page 170

References

Hill, Lisa (2002b), ‘Compulsory Voting: Residual Problems and Potential Solutions’, Australian Journal of Political Science 37.3. Hill, Lisa (2002c), ‘Democratic Assistance: A Compulsory Voting Template’, paper presented at the Jubilee conference of the Australia Political Studies Association, Australia National University, Canberra, October 2002. Hill, Lisa (2006), ‘Low Voter Turnout in the United States: Is Compulsory Voting a Viable Solution’, Journal of Theoretical Politics 18.2, pp. 207–32. Hill, Lisa and Louth, Jonathon (2004), ‘Compulsory Voting Laws and Turnout: Efficacy and Appropriateness’, Paper presented at the Australasian Political Studies Association Conference, University of Adelaide, 29 September – 1 October, 2004. Hirczy, Wolfgang (1994), ‘The Impact of Mandatory Voting Laws on Turnout: A QuasiExperimental Approach’, Electoral Studies 13.1, pp. 64–76. Hirczy, Wolfgang (2000), ‘Compulsory Voting’, in Richard Rose (ed.), International Encyclopedia of Elections, Basingstoke: Macmillan, pp. 44–7. Hooghe, Marc and Pelleriaux, Koen (1998), ‘Compulsory Voting in Belgium: An Application of the Lijphart Thesis’, Electoral Studies 17.4, pp. 419–24. Huber, John D., Kernell, Georgia and Leoni, Eduardo L. (2005), ‘Institutional Context, Cognitive Resources and Party Attachments Across Democracies’, Political Analysis 13, pp. 365–86. Hughes, Colin (1966), ‘Compulsory Voting’, Politics 1.2, pp. 81–95. International IDEA [IDEA] (2002), Voter Turnout Since 1945: A Global Report, Stockholm: International IDEA. International IDEA [IDEA] (2006), Engaging the Electorate: Initiatives to Promote Voter Turnout from Around the World, Stockholm: International IDEA. Irwin, Galen (1974), ‘Compulsory Voting Legislation’s Impact on Turnout in the Netherlands’, Comparative Political Studies 7.3, pp. 292–315. Irwin, Galen A. and van Holsteyn, Joop (2005), ‘Scarfman’s Parcel; Old and New Thoughts on the Abolition of Compulsory Voting in the Netherlands’, Paper presented at the international symposium on compulsory voting, Institut d’Etudes Politiques de Lille, Lille 20–21 October, 2005. Irwin, Galen A. and van Holsteyn, Joop (2007), ‘May We Have the Votes Please? Attitudes towards Compulsory Voting’, Paper presented at the ‘Compulsory Voting: Principles and Practice’ workshop, European Consortium for Political Research Joint Sessions of Workshops, Helsinki, 7–12 May 2007. Jackee, Keith and Sun, Guang-Zhen (2006), ‘Is Compulsory Voting More Democratic?’, Public Choice 129, pp. 61–75. Jackman, Robert W. (1987), ‘Political Institutions and Voter Turnout in Industrial Democracies’, American Political Science Review 81.2, pp. 405–23. Jackman, Robert W. and Miller, Ross A. (1995), ‘Voter Turnout in the Industrial Democracies During the 1980s’, Comparative Political Studies 27.4, pp. 467–92. Jackman, Robert W. and Volpert, Karin (1996), ‘Conditions Favouring Parties of the Extreme Right in Western Europe’, British Journal of Political Science 26.4, pp. 501–22. Jackman, Simon (1999), ‘Electoral Enquiry: Non-Compulsory Voting in Australia? What Surveys Can (and Can’t) Tell Us’, Electoral Studies 18.1, pp. 29–48. Jackman, Simon (2001), ‘Voting: Compulsory’, in Smelser, Neil J. and Bates, Paul B.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 171

References

171

(eds.), International Encyclopedia of the Social and Behavioral Sciences, vol. 24, Amsterdam: Elsevier, pp. 16314–8. Jaramillo, Juan and Franco-Cuervo, Beatriz (2005), ‘Colombia’, in Dieter Nohlen (ed.), Elections in the Americas: A Data Handbook, Vol. II: South America, Oxford: Oxford University Press, pp. 295–364. Jenks, William Alexander (1950), The Austrian Electoral Reform of 1907, New York: Colombia University Press. Jones, Mark P., Lauga, Martín and León-Roesch, Marta (2005), ‘Argentina’ in Nohlen, Dieter, Elections in the Americas: A Data Handbook: Vol. 2 South America, Oxford: Oxford University Press, pp. 59–122. Karp, Jeffrey A. and Banducci, Susan A. (2008), ‘Political Efficacy and Participation in Twenty-Seven Democracies: How Electoral Systems Shape Political Behaviour’, British Journal of Political Science 38.1, pp. 11–112. Karp, Jeffrey A., Banducci, Susan A. and Bowler, Shaun (2008), ‘Getting Out the Vote: Party Mobilization in a Comparative Perspective’, British Journal of Political Science 38.2, pp. 311–34. Karklins, Rasma (1986), ‘Soviet Elections Revisited: Voter Abstention in NonCompetitive Voting’, American Political Science Review 80.2, pp. 449–69. Katz, Richard, S. (1997), Democracy and Elections, Oxford: Oxford University Press. Kauz, Ralph, Sharoudi, Hamid Khosravi and Rieck, Andreas (2001), ‘Iran’, in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. I The Middle East, Central Asia and South Asia, Oxford: Oxford University Press, pp. 57–83. Keaney, Emily and Rogers, Ben (2006), A Citizen’s Duty: Voter Inequality and the Case for Compulsory Turnout, London: IPPR. Kelley, Jonathan and McAllister, Ian (1984), ‘Ballot Paper Cues and the Vote in Australia and Britain: Alphabetic Voting, Sex and Title’, Public Opinion Quarterly 48.2, pp. 452–66. Kenworthy, Lane and Malami, Melissa (1999), ‘Gender Inequality in Political Representation: A Worldwide Comparative Analysis’, Social Forces 78.1, pp. 235–68. King, Amy and Leigh, Andrew (2006), ‘Ballot Order Effects under Compulsory Voting’, unpublished manuscript. Krennerich, Michael (2005), ‘Nicaragua’ in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. I North America, Central America, and the Caribbean, Oxford: Oxford University Press, pp. 479–510. Kunicova, Jana and Rose-Ackerman, Susan (2005), ‘Electoral Rules and Constitutional Structures as Constraints on Corruption’, British Journal of Political Science, 35.4, pp. 573–606. Laakso, Markku U. and Taagepera, Rein (1979), ‘ “Effective” Number of Parties: A Measure with Application to Western Europe’, Comparative Political Studies 12.3, pp. 3–27. Lamounier, Bolívar and Amorim Neto, Octavio (2005), ‘Brazil’, in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. 2 South America, Oxford: Oxford University Press, pp. 163–252. Lapp, Nancy D. (2004), ‘To Vote or Not to Vote: Compulsory Voting in Latin America’,

M1546 - BIRCH TXT.qxp:ANDY Q7

172

27/10/08

11:31

Page 172

References

Paper presented at the Western Political Science Association meeting, 11–13 March, 2004. Lapp, Nancy D. (2006), ‘Compulsory Voting in Chile: Democratization and Electoral Reform’, paper presented at the American Political Science Annual Meeting, Philadelphia, 31 August-3 September, 2006. Lardy, Heather (2004), ‘Is There a Right Not to Vote?’, Oxford Journal of Legal Studies 24.2. Lederman, D. L., Loayza, N. V and Soares, R. R. (2005), ‘Accountability and Corruption: Political Institutions Matter’, Economics and Politics 17.1, pp. 1–35. Lehoucq, Fabrice E. (2003), “Electoral Fraud: Causes, Types, and Consequences.” Annual Review of Political Science 6, pp. 233–56. León-Roesch, Marta and Richard Ortiz Ortiz (2005), ‘Paraguay’ in Dieter Nohlen (ed.), Elections in the Americas: A Data Handbook: Vol. II South America, Oxford: Oxford University Press, pp. 411–44. Lever, Annabelle (2007), ‘The Case for Compulsory Voting: A Critical Perspective’, Paper presented at the ‘Compulsory Voting: Principles and Practice’ workshop, European Consortium for Political Research Joint Sessions of Workshops, Helsinki, 7–12 May 2007. Lijphart, Arend (1994), Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies, 1945–1990. Oxford: Oxford University Press. Lijphart, Arend (1997), ‘Unequal Participation: Democracy’s Unresolved Dilemma’, American Political Science Review 91.1, pp. 1–14. Loewen, Peter, Milner, Henry and Hicks, Bruce M. (2007), ‘Compulsory Voting and Informed Youth: A Field Experiment’, Paper presented at the ‘Compulsory Voting: Principles and Practice’ workshop, European Consortium for Political Research Joint Sessions of Workshops, Helsinki, 7–12 May 2007. Lomasky, Loren E. and Brennan, Geoffrey (2000), ‘Is There a Duty to Vote?’ Social Philosophy and Policy 17.1, pp. 62–86. López Pintor, Rafael, Gratschew, Maria and Sullivan, Kate (2002), ‘Voter Turnout Rates in Comparative Perspective’, in International IDEA, Voter Turnout since 1945: A Global Report, Stockholm: International IDEA, pp. 75–91. Lundell, Krister (2007), ‘Partisanship and Political Support: The Impact of Compulsory Voting’, Paper delivered at the European Consortium for Political Research workshop on Compulsory Voting: Principles and Practice, Helsinki, 7–12 May, 2007. McAllister, Ian (1986), ‘Compulsory Voting, Turnout, and Party Advantage in Australia’, Politics 21.1. McAllister, Ian (2001), ‘Elections without Cues: The 1999 Australian Republic Referendum’, Australian Journal of Political Science 36.2, pp. 247–69. McAllister, Ian and Makkai, Toni (1993), ‘Institutions, Society or Protest? Explaining Invalid Votes in Australian Elections’, Electoral Studies 12.1, pp. 23–40. McGree Deutsch, Sandra (1986), Counterevolution in Argentina, 1990–1932: The Argentine Patriotic League, Lincoln and London: University of Nebraska Press. Mackenzie, W. J. M. (1958), Free Elections, London: George Allen & Unwin. Mackerras, Malcolm (1970), ‘Preference Voting and the “Donkey Vote” ‘, Politics 5.1, pp. 69–76.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 173

References

173

Mackerras, M. and McAllister, I. (1999), ‘Compulsory Voting, Party Stability and Electoral Advantage in Australia’, Electoral Studies 18.2, pp. 217–34. Malkopoulou, Anthoula (2007), ‘Compulsory Voting in Greece: A History of Concepts in Motion’, Paper delivered at the European Consortium for Political Research workshop on Compulsory Voting: Principles and Practice, Helsinki, 7–12 May, 2007. Martin, Paul S. (2003), ‘Voting’s Rewards: Voter Turnout, Attentive Publics, and Congressional Allocation of Federal Money’, American Journal of Political Science 47.1, pp. 110–27. Martinez, Michael D. and Gill, Jeff (2005), ‘The Effects of Turnout on Partisan Outcomes in US Presidential Elections 1960–2000’, Journal of Politics 67.4, pp. 1248–74. Massicotte, Louis, Blais, André and Yoshinaka, Antoine (2004), Establishing the Rules of the Game: Election Laws in Democracies, Toronto, Buffalo and London; University of Toronto Press. Matland, Richard E. (1998), ‘Women’s Representation in National Legislatures: Developed and Developing Countries’, Legislative Studies Quarterly 23.1, pp. 109–25. Matsler, Sean (2003), ‘Compulsory Voting in America’, Southern California Law Review 76.4, pp. 953–78. Mill, John Stuart (1861), ‘Consideration on Representative Government’ in Mill, John Stuart, On Liberty and Other Essays, ed. John Gray, Oxford: Oxford University Press, 1991, pp. 203–467. Milner, Henry (2002), Civic Literacy: How Informed Citizens Make Democracy Work, Hanover, NH: University Press of New England. Molina, José and Thibaut, Bernard (2005), ‘Venezuela’, in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. 2 South America, Oxford: Oxford University Press, pp. 535–92. Montinola, Gabriella and Jackman, Robert (1992), ‘Sources of Corruption: A CrossCountry Study’, British Journal of Political Science 32.1, pp. 147–70. Moreau, Félix (1896), Le Vote obligatoire: principe et sanctions, extrait de la Revue Politique et Parlementaire, Paris: Bureau de la Review Politique et Parlementaire. Morris Jones, W. H. (1954), ‘In Defence of Apathy: Some Doubts on the Duty to Vote’, Political Studies 2.1, pp. 25–37. Mote, Max E. (1965), Soviet Local and Republic Elections, Stanford, CA: Hoover Institution Press. Mozaffar, Shaheen and Schedler, Andreas (2002), ‘The Comparative Study of Electoral Governance – Introduction’, International Political Science Review 23.1, pp. 5–27. Mueller, Dennis C. (2003), Public Choice III, Cambridge: Cambridge University Press. Mueller, Dennis C. and Stratmann, Thomas (2003), ‘The Economic Effects of Democratic Participation’, Journal of Public Economics 87, pp. 2129–55. Murray, David (1998), ‘Thailand’s Recent Electoral Reforms’, Electoral Studies 17.4, pp. 525–36. Newton, Kenneth and Norris, Pippa (2000), ‘Confidence in Public Institutions: Faith, Culture, or Performance’, in Pharr, Susan J. and Putnam, Robert D. (eds.), Disaffected Democracies: What’s Troubling the Trilateral Countries, Princeton, NJ: Princeton University Press, pp. 52–73.

M1546 - BIRCH TXT.qxp:ANDY Q7

174

27/10/08

11:31

Page 174

References

Nohlen, Dieter (ed.) (2005a), Elections in the Americas: A Data Handbook (2 vols.), Oxford: Oxford University Press. Nohlen, Dieter (ed.) (2005b), ‘Chile’ in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. 2 South America, Oxford: Oxford University Press, pp. 253–93. Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.) (2001), Elections in Asia and the Pacific: A Data Handbook (2 vols.), Oxford: Oxford University Press. Nohlen, Dieter, Krennerich, Michael and Thibaut, Bernhard (eds.) (1999), Elections in Africa: A Data Handbook, Oxford: Oxford University Press. Nohlen, Dieter and Pachano, Simon (2005), ‘Ecuador’, in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. II South America, Oxford: Oxford University Press, pp. 365–410. Norris, Pippa (1985), ‘Women’s Legislative Participation in Western Europe’ West European Politics 8.4, pp. 90–101. Norris, Pippa (1999), ‘Institutional Explanations for Political Support’, in Pippa Norris (ed.), Critical Citizens: Global Support for Democratic Governance, Oxford: Oxford University Press, pp. 217–35. Norris, Pippa (2002), Democratic Phoenix: Reinventing Political Activism, Cambridge: Cambridge University Press. Norris, Pippa (2004), Electoral Engineering: Voting Rules and Political Behaviour, Cambridge: Cambridge University Press. Norris, Pippa (2005), Radical Right: Voters and Parties in the Electoral Market, Cambridge: Cambridge University Press. Norris, Pippa (2006), ‘The Impact of Electoral Reform on Women’s Representation’, Acta Politica 41.2, pp. 197–213. Ordeshook, Peter and Shvetsova, Olga (1994), ‘Ethnic Heterogeneity, District Magnitude, and the Number of Parties’, American Journal of Political Science 38, pp. 100–23. Orr, Graeme (1998), ‘Tinkering with Convention: Voluntary Voting at Australia’s 1997 Constitutional Convention Election’, Electoral Studies 17.4, pp. 575–81. O’Toole, Francis and Strobl, Eric (1995), ‘Compulsory Voting and Government Spending’, Economics and Politics 7.3, pp. 271–7. Pacek, Alexander and Radcliffe, Benjamin (1995), ‘Turnout and Vote for Left-of-Centre Parties: A Cross-National Analysis, British Journal of Political Science, 25.1, pp. 137–43. Pammet, John H. and LeDuc, Lawrence (2003), Explaining the Turnout Decline in Canadian Federal Elections: A New Survey of Non-Voters, Elections Canada. Pantelis, A., Koutsoubinas, S. and Gerapetritis, G. (2008), ’Greece’, in Nohlen, D., Caton, M. and Stöver, P. ( eds.), Elections in Europe: A Data Handbook, Oxford: Oxford University Press. Pateman, Carole (1970), Participation and Democratic Theory, Cambridge: Cambridge University Press. Payne, J. Mark, Zovatto, Daniel G., Mateo Díaz, Mercedes with Allamand Zavala, Andrés, Carrillo Flórez, Fernando, Echebarría, Koldo, Freidenberg, Flavia and Jarquín, Edmundo (2006), La Polícita Importa: Democracia y Desarollo en América Latina, Washington, DC: Interamerican Bank for Development.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 175

References

175

Persson, Torsten and Tabellini, Guido (2003), The Economic Effects of Constitutions. Cambridge, MA and London: MIT Press. Persson, Torsten, Tabellini, Guido and Trebbi, F. (2003), ‘Electoral Rules and Corruption’ European Economic Review, 1.4, pp. 958–89. Pettersen, Per Arnt and Rose, Lawrence E. (2007), ‘The Dog that Didn’t Bark: Would Increased Electoral Turnout Make a Difference?’ Electoral Studies 26.3, pp. 574–88. Pettit, Philip (1997), Republicanism: A Theory of Freedom and Government, Oxford: Oxford University Press. Pilet, Jean-Benoit (2005), ‘Comprendre les positions sur le vote obligatoire en Belgique: entre conviction et stratégie’, Paper presented at the international symposium on compulsory voting, Institut d’Etudes Politiques de Lille, Lille 20–21 October, 2005. Powell, G. Bingham (1980), ‘Voting Turnout in Thirty Democracies: Partisan, Legal, and Socio-Economic Influences’, in Richard Rose (ed.), Electoral Participation: A Comparative Analysis, Beverly Hills and London: Sage, pp. 5–34. Powell, G. Bingham (1982), Contemporary Democracies: Participation, Stability, and Violence, Cambridge, MA: Harvard University Press. Powell, G. Bingham (1986), ‘American Turnout in Comparative Perspective’, American Political Science Review 80.1, pp. 17–37. Power, T., Roberts, T. (1995), ‘Compulsory Voting: Invalid Ballots and Abstention in Brazil’, Political Research Quarterly 48.4. Power, Timothy J. and Garand, James C. (2007), ‘Determinants of Invalid Voting In Latin America’, Electoral Studies 26.2, pp. 432–444. Quiri, N. (1908), Le Vote obligatoire, Doctoral thesis, Paris: La Librairie des Facultés A. Michalon. Rasbash, Jon, Fiona Steele, Browne, William and Prosser, Bob (2004), A User’s Guide to MLwiN version 2.0, London: Centre for Multilevel Modelling, Institute of Education, University of London. Reilly, Benjamin and Gratschew, Maria (2001), ‘Nauru’, in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. II South East Asia, East Asia and the South Pacific, Oxford: Oxford University Press, pp. 697–704. Reynolds, Andrew (1999), ‘Women in Legislatures and Executives of the World: Knocking at the Highest Ceiling’, World Politics 51.4, pp. 547–72. Reynolds, Andrew, Reilly, Ben and Ellis, Andrew (2005), Electoral System Design: The New International Handbook, Stockholm: International IDEA. Reynolds, Andrew and Steenbergen, Marco (2006), ‘How the World Votes: The Political Consequences of Ballot Design, Innovation and Manipulation’, Electoral Studies 25.3, pp. 570–98. Rieger, Hans Christoph (2001), ‘Singapore’, in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. II South East Asia, East Asia and the South Pacific, Oxford: Oxford University Press, pp. 239–59. Rihoux, Benoit (1996), ‘Electoral Reform and Electoral Behaviour in Belgium: Change within Continuity. . . or Conversely’, Res Publica 2, pp. 255–78. Riker, William H. and Ordeshook, Peter C. (1968), ‘A Theory of the Calculus of Voting’, American Political Science Review 62, pp. 25–43.

M1546 - BIRCH TXT.qxp:ANDY Q7

176

27/10/08

11:31

Page 176

References

Robson, William (1923), ‘Compulsory Voting’, Political Science Quarterly 38, pp. 569–77. Rodríguez Blanco, Maricel, Entín, Gabriel and Rodríguez, Darío (2004), ‘Ciudadanía Exclusión: Un Problema para el análisis del voto’, in Cheresky, Isidoro and Blaquer, Jean-Michel (eds.), ¿Qué cambió en la política Argentina? Elecciones, instituciones y ciudadanía en perspectiva comparada, Buenos Aires: Homo Sapiens. Roeder, Philip G. (1989), ‘Electoral Avoidance in the Soviet Union’, Soviet Studies 41.3, pp. 462–83. Rose, Richard (ed.) (2000), International Encyclopedia of Elections, Basingstoke: Macmillan. Rosenstone, Steven and Hansen, Mark (1993), Mobilization, Participation and Democracy in America, New York: Macmillan. Rubenson, Daniel, Blais, André, Fournier, Patrick, Gidengil, Elisabeth and Nevitte, Neil (2007), ‘Does Low Turnout Matter? Evidence from the 2000 Canadian Federal Election’, Electoral Studies 26.3, pp. 589–97. Rule, Wilma (1987), ‘Electoral Systems, Contextual Factors and Women’s Opportunity for Election to Parliament in Twenty-Three Democracies’, Western Political Quarterly 40.3, pp. 477–98. Samayoa, Salvador (1986), ‘El Salvador’ in Hernández Becerra, Augusto (ed.), Legislación Electoral Comparada : Colombia, México, Panamá, Venezuela y Centroamérica, San José, Costa Rica : Centro de Asesoría y Promoción Electoral, Instituto Interamericano de Derechos Humanos, Instituto de Investigaciones Jurídicas, Universidad Nacional Autónoma de México. Schaffer, Frederic Charles (2002), ‘Might Cleaning Up Elections Keep People Away from the Polls? Historical and Comparative Perspectives’, International Political Science Review 23.1, pp. 69–84. Schaffer, Frederic Charles (2005), ‘Clean Elections and the Great Unwashed: Vote Buying and Voter Education in the Philippines’, Occasional Paper No. 21, School of Social Science, Institute for Advanced Study, Princeton, NJ. Schafferer, Christian (2003), ‘The Legislative Yuan Election, Taiwan 2001’, Electoral Studies 22.3, pp. 532–7. Scheffler, Thomas (2001), ‘Lebanon’ in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. 1 The Middle East, Central Asia and South Asia, Oxford: Oxford University Press, pp. 169–98. Schiffino, Nathalie and Aish, Anne-Marie (1999), ‘Et si le vote n’était plus obligatoire?’ in Frognier, André-Paul and Aish, Anne-Marie (eds.), Des élections en trompe-l’oeil: Enquête sur le comportement électoral des Wallons et des Francophones, Brussels: DeBoeck, pp. 97–120. Schumpeter, J. (1974), Capitalism, Socialism, and Democracy, 3rd ed., New York: Harper. Schwindt-Bayer, Leslie A. and Mishler, William (2005), ‘An Integrated Model of Women’s Representation’, Journal of Politics 67.2, pp. 407–28. Sear, Chris and Strickland, Pat (2003), ‘Compulsory Voting’, House of Commons Library Standard Note SN/PC/954, 3 July 2003. Selb, Peter and Lachat, Romain (2007), ‘The More the Better? Counterfactual Evidence on the Effect of Compulsory Voting on the Consistency of Party Choice’, paper presented at the ECPR Joint Sessions of Workshops, Helsinki, 7–12 May 2007.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 177

References

177

Seligson, Mitchell A. (2002), ‘The Impact of Corruption on Regime Legitimacy: A Comparative Study of Four Latin American Countries’, Journal of Politics 64.2, pp. 408–33. Service des Etudes Juridiques, French Senate (2003), ‘Le Vote Obligatoire’, Document de Travail du Sénat, série Législation Comparée, available at http://www.senat. fr/lc/lc121/lc121.html. Siaroff, Alan and Merer, John W. A. (2002), ‘Parliamentary Election Turnout in Europe since 1990’, Political Studies 50, pp. 916–27. Smith, T. E. and Keith-Lucas, B. (1960), Elections in Developing Countries: A Study of Electoral Procedures Used in Tropical Africa, South-East Asia and the British Caribbean, London: Macmillan. Somoza, Alexander, ‘Honduras’, in Nohlen, Dieter (ed.) (2005), Elections in the Americas: A Data Handbook: Vol. I North America, Central America, and the Caribbean, Oxford: Oxford University Press, pp. 399–421. Staveley, E. S. (1972), Greek and Roman Voting and Elections, London: Thames and Hudson. Stears, Mark (2007), ‘Review Article: Liberalism and the Politics of Compulsion’, British Journal of Political Science 37.3, pp. 533–53. Steenbergen, Marco R. and Jones, Bradford S. (2002), ‘Modeling Multilevel Data Structures’, American Journal of Political Science 46.1, pp. 218–37. Stengers, Jean (1990), ‘Histoire de la législation électorale en Belgique’, in Serge Noiret (ed.), Political Strategies and Electoral Reforms: Origins of Voting Systems in Europe in the 19th and 20th Centuries, Baden-Baden: Nomos. Suh, Mark B. (2001), ‘Korea (Democratic People’s Republic/North Korea)’, in Nohlen, Dieter, Grotz, Florian and Hartmann, Christof (eds.), Elections in Asia and the Pacific: A Data Handbook: Vol. II South East Asia, East Asia and the South Pacific, Oxford: Oxford University Press, pp. 395–409. Suter, Jan and Nohlen, Dieter, ‘Cuba’, in Nohlen, Dieter (ed.) (2005), Elections in the Americas: A Data Handbook: Vol. I North America, Central America, and the Caribbean, Oxford: Oxford University Press, pp. 195–222. Taagepera, Rein and Matthew Soberg Shugart (1989), Seats and Votes: The Effects and Determinants of Electoral Systems, New Haven and London: Yale University Press. Tingsten, Herbert (1963) [1937], Political Behavior: Studies in Election Statistics, Totowa, NJ: Bedminster Press, 1963 [1937]. Triesman, Daniel (2000), ‘The Causes of Corruption: A Cross-National Study’, Journal of Public Economics 76, pp. 399–457. Tuesta Soldevilla, Fernando (2005), ‘Peru’, in Nohlen, Dieter (ed.), Elections in the Americas: A Data Handbook: Vol. II South America, Oxford: Oxford University Press, pp. 445–86. Verba, Sidney, Nie, Norman H. and Kim, Jae-on (1978), Participation and Political Equality: A Seven-Nation Comparison, Cambridge: Cambridge University Press. Ware, Alan (1996), Political Parties and Party Systems, Oxford: Oxford University Press. Watson, Tom and Tami, Mark (2001, updated 2004), Make Voting Compulsory, London: Fabian Society. Weale, Albert (1999), Democracy, Basingstoke: Macmillan.

M1546 - BIRCH TXT.qxp:ANDY Q7

178

27/10/08

11:31

Page 178

References

Wertheimer, Allen (1975), ‘In Defense of Compulsory Voting’, in Pennock, Ronald and Chapman, John (eds.), Participation in Politics, New York: Lieber-Atherton. Whiteley, Paul and Seyd, Patrick (1998), ‘The Dynamics of Party Activism in Britain: A Spiral of Demobilization?’, British Journal of Political Science 28.1, pp. 113–37. de Winter, Lieven and Ackaert, Johan (1994), ‘Abstentionnisme électoral et vote blanc ou nul: le ‘non-vote’ en Wallonie’, in Frognier, André-Paul and Aish-Van Vaerenbergh, Ann-Marie (eds.), Elections la fêlure?: Enquête sur le comportement electoral des Wallons et des Francophones, Brussels: De Boeck, pp. 77–97. de Winter, Lieven and Ackaert, Johan (1998) ‘Compulsory Voting in Belgium: A Reply to Hooghe and Pelleriaux’, Electoral Studies 17.4, pp. 425–28. de Winter, Lieven, Dumont, Patrick and Ackaert, Johan (2003), ‘La Participation électorale réelle et potentielle: des vertus du vote obligatoire’, in Frongier, André-Paul and Aish A.-M. (eds.), Elections: La Rupture? Le Comportement des Belges face aux élections de 1999, Brussels: De Boeck, pp. 54–69. Yeret, Eliyahu (1995), ‘Dictators, Democracies, and the Voters: A Comment on the Right Versus the Obligation to Vote’, Economics and Politics 7.3, pp. 263–70. Yoon, Mi Yung (2001), ‘Democratization and Women’s Legislative Representation in Sub-Saharan Africa’, Democratization 8.2, pp. 169–90.

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 179

Index

Note: ‘n.’ after a page reference indicates the number of the note on that page. Abraham, Henry 5, 8, 14, 17–18n.2, 18n.15, 20, 21, 23, 28, 38n.1, 40, 43, 44, 50, 51, 105, 163 Ackaert, Johan 30, 95, 98n.9, 104, 116n.3, 127, 129, 163, 178 affichage 8, 18n.13 see also sanctions for non-voting Aish, Anne-Marie 95, 163, 176 Aitkin, Don 106, 163 Alapuro, Risto 24, 163 Albania 134n.2 Alesina, Alberto 163 Allen, Graham 154n.13 Amorim Neto, Octavio 25, 171 Anderson, Christopher J. 113, 163 Anduiza Perea, Eva 97n.5, 163 Argentina 7, 8, 12, 13, 24, 27, 31–2, 34, 36, 37, 39n.12, 81, 103, 104, 109–11 Asian Network for Free Elections 103, 163 Australia 3, 4, 6, 8, 9, 11, 13, 14, 26, 27, 32–3, 34, 35, 36, 37, 39n.12, 62, 63–4, 65, 74–5, 80, 95–6, 98n.10, 100–11, 116n.3, 117n.9, 122, 134n.2, 143, 145, 146, 153n.10 Austria 13, 23, 35, 43, 86–90, 97n.6 Baden 15, 22 Baeza, Mario Fernández 27, 163

Ballinger, Chris 1, 50, 145, 163 ballot secrecy see secret ballot ballot spoilage see invalid ballots Banducci, Susan A. 91, 113, 164, 171 Barber, Benjamin 49, 164 Barnes, William 142, 164 Baston, Lewis 16, 43, 150, 164 Bavaria 15, 22, 35 Beazley, Kim 109 Bélanger, Eric 122, 164 Belarus 15, 19n.26 Belgium 2, 6, 8, 9, 11, 12, 13, 18–19n.17, 22, 24, 27, 29–31, 32, 36, 37, 39n.12, 45, 52, 60, 62, 65, 80, 95–6, 99, 100–7, 112, 116n.3, 124, 125, 126, 127, 129, 134n.1, 134n.2, 143 Bendel, Petra 83, 164 Bennett, Scott 6, 9, 164 Benoit, Kenneth 125, 127, 164 Berggren, Heidi 62, 164 Bernhagen, Patrick 122, 127, 164 Betz, Hans-Georg 125, 164 Bilodeau, Antoine 63–4, 67, 164 Blais, André 3, 4, 63–4, 67, 91, 92–3, 97n.1, 164, 173, 176 Boix, Carles 34, 164 Bolivia 9, 11, 12, 13, 25, 36 Bosnia and Herzegovina 134n.2 Bouillin-Dartevelle, R. 143, 164 Bowler, Shaun 64, 67, 145, 146, 165

M1546 - BIRCH TXT.qxp:ANDY Q7

180

27/10/08

11:31

Page 180

Index

Brader, Ted 74, 165 Brazil 7, 8, 11, 12, 13, 25, 36, 37, 65, 109–11, 127, 153n.9 Brea Franco, Julio 25, 83, 165 Brennan, Geoffrey 44, 68, 172 Brockington, David 62, 78n.4, 111, 165 Browne, William 175 van der Brug, Wouter 125, 165 Brunner, Georg 3, 165 Bryce, James 41, 165 Bukovina 23 Bulgaria 16, 19n.23, 23, 27, 150 California 151 Callahan, William 33, 165 Campbell, Angus 3, 4, 74, 165 Canache, Damarys 117n.12, 165 Canada 1, 10, 27, 55, 62, 134n.2, 143–5, 146 Caramani, Daniele 5, 15, 20, 21, 22, 23, 38n.3, 165 Carlin, Ryan 110, 165 Carniola 23 Carson, George Barr 15, 165 Carstairs, Andrew McLaren 34, 165 Carter, Elisabeth 134n.3, 166 Caul, Miki 3, 91, 169 Central African Republic 14 Chile 11, 13, 25, 36, 65, 103, 104, 109–10 Chirac, Jacques 146 Chong, Alberto 53, 130, 135n.8, 166 Churchill, Winston 145 civic duty see political duties Claeys, R. 164 de Clercq, Willy 30–1 collective sanctions for non-voting see turnout thresholds Colombia 14, 16, 24, 25, 39n.6, 150 Congo, Democratic Republic of 39n.11 Congo, the Republic of the 26, 27, 39n.10 constitutionalisation of the obligation to vote 14, 36–8, 94, 107–8, 119–32, 135n.9, 149

Converse, Philip 165 corruption 112–14, 118, 129, 131–4, 140, 152 Costa Rica 7, 11, 24, 36, 37 Cox, Gary 4, 75, 123, 150, 166 Crain, W. Mark 134n.7, 166 Crampton, R. J. 19n.23, 166 Crepaz, Markus M. L. 91, 166 Crewe, Ivor 3, 53, 91, 127, 166 Croatia 134n.2 Cuba 14, 18n.15, 24, 39n.6 Cyprus 12, 13, 23, 26, 27, 36, 37, 81, 134n.2 Czechoslovakia 23, 142 Czech Republic 134n.2 Czesnik, Mikolaj 127, 166 Dalton, Russell J. 3, 63, 67, 74, 113, 166 Dawance-Goossens, J. 164 Deininger, Klaus W. 130, 166 Delaware 20 Déloye, Yves 47, 166 Delruelle-Vosswinkel, N. 164 democratic legitimacy see legitimacy Democratic People’s Republic of Korea, see North Korea Denmark 134n.2 Deschouwer, Kris 106, 166 Devleeschauwer, Arnaud 163 Díaz, Honorario 31, 39n.14, 39n.15, 81, 103, 167 Dobrzynska, Agnieszka 91, 92–3, 164 Dominican Republic 7, 25, 36, 81–4 Donovan, Todd 67, 145, 165 Dumont, Patrick 178 duty see political duties Easterley, William 163 East Timor 14, 19n.22 Ecuador 12, 13, 25, 36 van Egmond, Marcel 127, 167 Egypt 12, 26, 35, 36, 37, 81, 97n.6 van der Eijk, Cees 127, 167 Electoral Commission [of the United Kingdom] 1, 43, 144, 145, 167

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Page 181

Index electoral integrity 2, 29–35, 57, 97, 99–108, 115–16, 140, 142, 147 Electoral Reform Society 16 electoral systems 24–5, 28–34, 47, 62, 66–74, 80, 107–8, 113–14, 120–1, 123–4, 125–6 electoral turnout see turnout El Salvador 7, 25, 36, 81, 101, 116n.5, 142 Elkins, Zachary 127, 167 Ellis, Andrew 175 Ellner, Steve 97n.4, 167 Enelow, James 47, 167 enforcement of the obligation to vote 5–7, 148 see also sanctions for non-voting Engelen, Bart 42, 45, 47, 63, 167 Entín, Gabriel 176 equality see inequality Escobar Armas, Carlos 25, 167 Estonia 134n.2 European Court of Human Rights 43 European Union 11, 12, 18n.6 exemptions from the obligation to vote 10–14 Fabius, Laurent 146 facilitation of voting see vote facilitation Farrell, David M. 106, 167 Feeley, Malcolm M. 42, 45, 46, 47, 53, 150, 153n.8, 167 female representation 118–21, 140 Fennema, Meindert 165 Fiji 26, 28, 35, 36, 37, 81–5 fines for non-voting 8, 9 see also sanctions for non-voting Finkel, Steven E. 49, 63, 67, 167 Finland 24, 134n.2 Fisher, Steven D. 127, 167 Fornos, Carolina 91–2, 97n.5, 167 Fournier, Patrick 176 Fraeys, William 110, 167 Fraga, Rosendo 110, 168 France 1, 3, 18, 19n.24, 21, 27, 41, 124, 143–5, 146–7

181

franchise restrictions 8–9, 10–11, 23–7, 28–34, 81 Franco-Cuervo, Beatriz 16, 171 Franklin, Mark 3, 4, 5, 50–1, 55–6, 59–60, 81, 85, 91, 92, 97n.3, 97n.5, 151, 168 Friedgut, Theodore 18n.7, 168 Frognier, André-Paul 163 Funk, Patricia 8, 22, 87, 90, 168 Gabon 26, 27 Gallego, Aina 129, 168 Garand, James C. 110, 167, 175 Georgia 20 Gerapetritis, G. 174 Germany 3, 22, 134n.2, 145, 146 Gerring, John 133, 168 Geys, Benny 91–2, 97n.1, 168 Gidengil, Elisabeth 176 Gill, Jeff 127, 173 Gini Index see inequality Ginsberg, Benjamin 49, 168 Giraud, Emile 37, 51, 168 Gluchowski, Peter 26, 168 Golder, Matthew 125, 168 Goodwin-Gill, Guy S. 10, 11, 116n.4, 168 Gordon, Stacy B. 62, 78n.4, 168 Gosnell, Harold F. 30, 60, 100, 168 Gow, Neil 32, 169 Gratschew, Maria 5, 9, 19n.18, 26, 49, 51, 89, 101, 105, 109, 116n.3, 169, 172, 175 Gray, Mark 3, 91, 169 Greece 7, 9, 23, 36, 88–9, 90, 112, 134n.2, 134n.5 Griffin, John D. 129, 169 Grönlund, Kimmo 62, 78n.4, 169 Grotz, Florian 26, 168, 174 Guatemala 7, 12, 25, 36, 39n.9, 81–3 Guillory, Christine A. 113, 163 Gummer, John 150 Haiti 14 Hansen, Mark 3, 4, 176

M1546 - BIRCH TXT.qxp:ANDY Q7

182

27/10/08

11:31

Page 182

Index

Harb, Mac 146, 152n.6, 169 Hartmann, Christof 19n.19, 26, 83, 169, 174 Hasen, Richard 4, 16, 45, 50, 148, 150, 151, 169 Hassall, Graham 169 Helmke, Gretchen 28, 37, 38, 169, 36 Hernández Becerra, Augusto 39n.6, 169 Hesse 22 Hicks, Alexander M. 169 Hicks, Bruce M. 129, 172 Highton, Benjamin 127, 169 Hill, Lisa 14, 41, 45, 46, 47, 50, 53, 63, 96, 117n.9, 142, 143, 146, 151, 153n.1, 153n.8, 169, 170 Hinich, Melvin J. 47, 167 Hirczy, Wolfgang 23, 86, 90–1, 96, 109, 143, 170 van Holsteyn, Joop 18n.5, 22, 23, 39n.5, 63, 78n.2, 129, 143, 144, 170 Hood, Jimmy 150 Hooghe, Marc 63, 95, 116n.3, 127, 167, 170 Hoon, Geoff 145 Honduras 7, 24, 25, 36 Huber, John D. 74, 170 Hughes, Colin 26, 32, 33, 104, 106, 170 Hungary 15, 18n.15, 23, 134n.2 Iceland 134n.2 Illinois 5 incentive voting 14, 15–16, 149–50 India 153n.1 inequality 51–6, 128–31, 133–4, 140 Institute for Public Policy Research 1, 3 International IDEA 16, 83, 87, 116n.3, 170 invalid ballots 2, 49, 55, 57, 100, 108–11, 116, 139–40 Iran 5, 15 Ireland 10, 134n.2 Irwin, Galen 18n.11, 22, 23, 39n.5, 63, 64, 78n.2, 86, 129, 143, 144, 170 Israel 134n.2

Italy 5, 8, 10, 14, 19n.18, 97n.3, 106, 134n.2 Jackee, Keith 44, 48, 62, 111, 170 Jackman, Robert W. 91, 125, 133, 170, 173 Jackman, Simon 19n.18, 21, 60, 95–6, 11, 109, 127, 170 Japan 134n.2 Jaramillo, Juan 16, 171 Jenks, William Alexander 13, 23, 171 Jinks, Brian 106, 163 Jones, Bradford S. 78n.5, 177 Jones, Mark P. 31, 171 Jordan 153n.1 Kansas City 21 Karklins, Rasma 18n.7, 171 Karp, Jeffrey 64, 65, 67, 75, 91, 113, 145, 164, 165, 171 Katz, Richard S. 11, 43, 44, 45, 46, 48, 91, 123, 171 Kauz, Ralph 5, 171 Kazakhstan 15 Keaney, Emily 1, 3, 145–6, 153n.8, 154n.11, 171 Keith-Lucas, B. 26, 39n.12, 177 Kelley, Jonathan 111, 171 Kenworthy, Lane 120, 171 Kernell, Georgia 170 Kim, Jae-on 177 King, Amy 109, 111, 171 Kingsley, Jean-Pierre 146 Kousser, J. Morgan 4, 150, 166 Koutsoubinas, S. 174 Krennerich, Michael 39n.7, 83, 101, 164, 171, 174 Kunicova, Jana 133, 171 Kurlat, Sergio 163 Kyrgyzstan 15 Laakso, Markku U. 171 Lachat, Romain 62, 111, 129, 176 Lamounier, Bolívar 25, 171 Lanoue, David J. 165

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Index Laos 26, 36, 37, 81–5 Lapp, Nancy 11, 16, 97n.4, 110, 142, 171, 172 Lardy, Heather 42–3, 47, 50, 52, 53, 55, 154n.13, 172 Latvia 134n.2 Lauga, Martín 171 Laver, Michael 125, 127, 164 Lebanon 19n.21, 26 Lederman, D. L. 133, 172 LeDuc, Lawrence 144, 174 Lehoucq, Fabrice E. 4, 172 legitimacy 45–9, 57, 97, 99–100, 108–16, 138–9, 152 Leigh, Andrew 109, 111, 171 Leonard, Mary L. 134n.7, 166 Leoni, Eduardo L. 170 León-Roesch, Marta 25, 83, 171, 172 Lever, Annabelle 43, 48, 172 Liechtenstein 14, 22, 35, 39n.4, 86–8, 90 Lijphart, Arend 42, 46, 47, 50, 53, 61, 78n.1, 105, 123, 129, 172 Lippe 22 Lithuania 15, 134n.2 Loeb-Mayer, N. 165 Loewen, Peter 62, 172 Loayza, N. V. 172 Lomasky, Loren E. 44, 68, 172 López-Pintor, Rafael 92, 116n.3, 172 Louth, Jonathon 96, 153n.1, 169 Lower Austria 23 Lundell, Krister 64, 112, 172 Luxembourg 8, 12, 13, 22, 36, 37, 112, 134n.2 McAllister, Ian 14, 62, 64, 95–6, 104, 106, 109, 110, 111, 121, 122, 127, 143, 171, 172, 173 Macedonia 134n.2 McGree Deutsch, Sandra 31, 172 Mackenzie, W. J. M. 43, 172 Mackerras, Malcolm 14, 62, 64, 95–6, 104, 109, 110, 111, 121, 122, 127, 143, 172, 173

Page 183

183

Makkai, Toni 110, 172 Malami, Melissa 120, 171 Malkopoulou, Anthoula 89, 173 Malta 134n.2, 141 Marsh, Michael 122, 127, 164 Martin, Paul S. 129, 173 Martinez, Michael D. 127, 173 Maryland 20 Massachusetts 20, 21 Massicotte, Louis 10, 11, 19n.20, 37, 173 Mateo Díaz, Mercedes 174 Matland, Richard E. 120, 173 Matsler, Sean 45, 50, 153n.8, 173 Meguid, Bonnie 28, 37, 38, 169 Merer, John W. A. 92, 97n.5, 176 Mexico 7, 24, 36, 37 Mill, John Stuart 41, 49, 173 Miller, Ross A. 91, 170 Miller, William 165 Milner, Henry 62, 78n.4, 169, 172, 173 Mishler, William 120, 176 Missouri 21 mobilisation 3, 49–51, 59–78, 104–5, 114, 139 Moldova 15, 134n.2 Molina, José 25, 97n.4, 173 Mondak, Jeffrey J. 165 Mongolia 26 Montinola, Gabriella 133, 173 Moravia 23 Moreau, Félix 15, 19n.17, 19n.23, 21, 22, 23, 41, 173 Morris Jones, W. H. 43–4, 54, 173 Mote, Max E. 5, 173 Mozaffar, Shaheen 101, 173 Mozambique 14, 19n.22 Mueller, Dennis C. 47, 129, 173 Murray, David 33, 34, 173 Nauru 26, 27, 35, 36, 37 Netherlands 2, 22, 23, 30, 39n.5, 63, 64, 78n.3, 80, 85–8, 90, 97n.3, 110–11, 117n.10, 129, 134n.2, 143–4 145, 146

M1546 - BIRCH TXT.qxp:ANDY Q7

184

27/10/08

11:31

Page 184

Index

Nevitte, Neil 176 Newman, Brian 129, 169 Newton, Kenneth 113, 173 New Zealand 63, 134n.2, 145, 146, 153n.1 Nicaragua 24, 142 Nie, Norman H. 177 Noël-Aranda, M. C. 165 Nohlen, Dieter 12, 18n.15, 24, 25, 174, 177 Norris, Pippa 3, 5, 63, 67, 91, 92, 93, 113, 120, 125, 134n.3, 142, 173, 174 North Dakota 21 North Korea 5, 15 Norway 15, 134n.2, 150 Olivera, Mauricio 53, 130, 135n.8, 166 Ordeshook, Peter C. 3, 42, 123, 174, 175 Oregon 38 Orr, Graeme 95, 98n.10, 174 Ortiz Ortiz, Richard 25, 83, 172 O’Toole, Francis 134n.7, 174 Pacek, Alexander 53, 127, 174 Pachano, Simon 12, 25, 174 Pammet, John H. 144, 174 Panama 7, 24, 36, 37 Pantelis, A. 89, 174 Paraguay 7, 25, 36, 81–4 parties see political parties see also party identification; small parties party identification 66, 67, 68–74, 140 Pateman, Carole 49, 174 Payne, Mark J. 174 Pelleriaux, Koen 95, 116n.3, 127, 170 Le Pen, Jean-Marie 146 penalties for non-voting see sanctions for non-voting Persson, Torsten 133, 175 Peru 12, 13, 25, 36, 65 Pettersen, Per Arnt 127, 175 Pettit, Philip 47, 175 Philippines 19n.19, 26

Pilet, Jean-Benoit 6, 29, 30, 31, 45, 60, 134n.1, 175 Plymouth Colony 20 Poland 3, 134n.2 policy outcomes see political outcomes political duties 1, 3, 14, 21, 40–5, 137–8 political efficacy 46–7, 49, 62–4, 66–8, 73, 77, 139–40 political engagement 2, 49–51, 57, 59–78, 139–40 political knowledge 49–51, 57, 61–2, 65–7, 77 political obligation see political duties political outcomes 2, 118–35 political parties 28–35, 45, 49–52, 56, 59–78, 105, 106, 114, 118–19, 120–4, 126, 128, 140, 145–7 see also party identification; small parties political protest 71 political rights 1, 21, 40–5, 137–8 Portugal 3, 14, 134n.2 Powell, G. Bingham 3, 11, 91, 92, 175 Power, Timothy J. 9, 110, 53n.9, 167, 175 Prosser, Bob 175 Queensland 26, 32 Quiri, N. 18n.13, 19n.24, 29, 41, 45, 100, 175 Radcliffe, Benjamin 53, 127, 174 Rasbash, Jon 78n.5, 175 Reilly, Benjamin 26, 175 Reynolds, Andrew 109, 120, 175 Rieck, Andreas 171 Rieger, Hans Christoph 26, 175 rights see political rights Rihoux, Benoit 106, 110, 116n.3, 175 Riker, William H. 3, 42, 175 Ritchie, Ken 16, 43, 150, 164 Roberts, J. Timmons 9, 110, 153n.9, 175 Robson, William 18n.12, 20, 23, 29, 176 Rodríguez, Dario 176 Rodríguez Blanco, Maricel 110, 176

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

11:31

Index Roeder, Philip G. 18n.7, 176 Rogers, Ben 1, 3, 145–6, 153n.8, 154n.11, 171 Romania 23, 134n.2 Rose, Lawrence E. 127, 175 Rose, Richard 176 Rose-Ackerman, Susan 133, 171 Rosenstone, Steven 3, 4, 176 Rubenson, Daniel 127, 176 Rule, Wilma 120, 176 Russia 15, 19n.23 Sáenz Peña, Roque 31 St Gallen 22 Salzburg 23 Samayoa, Salvador 25, 176 sanctions for non-voting 3, 4–10, 19n.18–19, 24, 36, 79, 83, 88, 101, 148 see also affichage; fines for non-voting Santos, Soliman M. Jr. 169 satisfaction with democracy 113 Savoie, Paul 165 Schaffer, Frederic Charles 150, 176 Schafferer, Christian 176 Schaffhausen 8, 12, 13, 22, 36, 78n.3 Schedler, Andreas 101, 173 Scheffler, Thomas 19n.21, 26, 176 Schiffino, Nathalie 95, 176 Schumpeter, J. 40, 176 Schwindt-Bayer, Leslie A. 120, 176 Sear, Chris 24, 142, 145, 176 secret ballot 2, 28, 147 Segura, Gary M. 62, 78n.4, 168 Selb, Peter 62, 111, 129, 176 Seligson, Mitchell A. 113, 165, 177 Serbia and Montenegro 15 Seyd, Patrick 75, 178 Sharoudi, Hamid Khosravi 171 Shugart, Matthew Soberg 123, 177 Shvetsova, Olga 123, 174 Siaroff, Alan 92, 97n.5, 177 Silesia 23 Singapore 9, 13, 18n.16, 26, 27, 28, 36, 37, 39n.12, 81–3

Page 185

185

Slovakia 134n.2 Slovenia 134n.2 small parties 126 see also party identification; political parties Smith, T. E. 26, 39n.12, 177 Soares, R. R. 172 Solothurn 22 Somoza, Alexander 12, 24, 177 Soviet Union see Union of Soviet Socialist Republics Spain 3, 14, 23, 37, 134n.2 spoilt ballots see invalid ballots Squire, Lyn 130, 166 Staveley, E. S. 150, 176 Stears, Mark 57, 177 Steele, Fiona 175 Steenbergen, Marco 78n.5, 109, 175, 177 Stengers, Jean 6, 29, 30, 177 Stokes, Donald 165 Stratmann, Thomas 129, 173 Strickland, Pat 24, 142, 145, 176 Strobl, Eric 134n.7, 174 Styria 23, 86 Suh, Mark B. 5, 177 Sullivan, Kate 172 Sun, Guang-Zhen 44, 48, 62, 111, 170 Suter, Jan 18n.15, 24, 177 Swank, Duane H. 129, 169 Sweden 10, 24, 134n.2 Switzerland 8, 12, 20, 22, 35, 36, 78n.3, 86–7, 89, 90, 97n.6, 134n.2, 150 Taagepera, Rein 123, 171, 177 Tabellini, Guido 133, 175 Tajikistan 15 Tami, Mark 48, 145, 153n.8, 154n.11, 177 Thacker, Strom C. 133, 168 Thailand 2, 9, 26, 27, 28, 33–4, 36, 37, 39n.12, 65, 81–5, 103, 105 Thibaut, Bernard 25, 97n.4, 173, 174 Thomas, Gareth 145 Thoveron, G. 165

M1546 - BIRCH TXT.qxp:ANDY Q7

27/10/08

186

11:31

Index

Tillie, Jean 165 Tingsten, Herbert 18n.4, 22, 23, 24, 28, 80–1, 109, 177 Trebbi, F. 133, 175 Triesman, Daniel 133, 177 Tucker, Joshua A. 74, 165 Tuesta Soldevilla, Fernando 12, 177 Turkey 23, 26, 36, 37, 81–4, 134n.2 turnout 1, 2, 3, 45, 50, 53, 57, 79–98 turnout thresholds 14–15, 149 see also sanctions for non-voting Tverdova, Yulia V. 113, 163 Tyrol 23, 86 Ukraine 134n.2 undue influence see electoral integrity Union of Soviet Socialist Republics 3, 4–5, 15, 18n.7, 35, 111, 116n.1 United Kingdom 24, 27, 134n.2, 144–6, 153, 154n.11, 154n.13 United States of America 4–5, 16, 20–1, 101, 126–7, 129, 134n.2, 150 Upper Austria 23 Uruguay 25, 36, 37 Uzbekistan 15 Venezuela 4, 7, 11, 25, 36, 88–9, 97n.4 Verba, Sidney 3, 177 Verhoftstadt, Guy 30

Page 186

Virginia 20 Volpert, Karin 125, 170 Vorarlberg 23, 86 vote facilitation 5, 7, 51, 56, 100–1, 152 Wallonia 95, 143 Ware, Alan 75, 177 Warhurst, John 106, 163 Watson, Tom 48, 145, 153n.8, 154n.11, 177 Weale, Albert 10, 177 wealth inequality see inequality Webb, Paul 106, 167 Weissberg, Robert 49, 168 Wertheimer, Allen 45, 46–7, 52, 53, 55, 68, 153n.8, 178 Whiteley, Paul 75, 178 de Winter, Lieven 6, 30, 95, 98n.9, 106, 123, 125, 127, 129, 143, 163, 166, 178 Wintour, Patrick 153n.5 Wolfinger, Raymond E. 127, 169 Yeret, Eliyahu 134n.7, 178 Yoon, Mi Yung 120, 178 Yoshinaka, Antoine 173 Yrigoyen, Hipólito 31–2 Zaire see Congo, Democratic Republic of Zovatto, Daniel G. 174