Protecting the Ballot: How First-Wave Democracies Ended Electoral Corruption 9780691240046

Citation preview

P ROTEC TI N G TH E BALLOT

Protecting the Ballot How First-­Wave Democracies Ended Electoral Corruption Isabela Mares

P R I N C E ­T O N U N I V E R S I T Y P R E S S P R I N C E ­T O N A N D OX F O R D

Copyright © 2022 by Prince­ton University Press Prince­ton University Press is committed to the protection of copyright and the intellectual property our authors entrust to us. Copyright promotes the pro­gress and integrity of knowledge. Thank you for supporting ­free speech and the global exchange of ideas by purchasing an authorized edition of this book. If you wish to reproduce or distribute any part of it in any form, please obtain permission. Requests for permission to reproduce material from this work should be sent to permissions@press​.­princeton​.­edu Published by Prince­ton University Press 41 William Street, Prince­ton, New Jersey 08540 99 Banbury Road, Oxford OX2 6JX press​.­princeton​.­edu All Rights Reserved ISBN: 978-0-691-24002-2 ISBN (pbk.): 978-0-691-24003-9 ISBN (e-­book): 978-0-691-24004-6 British Library Cataloging-­in-­Publication Data is available Editorial: Bridget Flannery-­McCoy and Alena Chekanov Production Editorial: Jenny Wolkowicki Cover design: Karl Spurzem Production: Lauren ­Reese Publicity: Kate Hensley and Charlotte Coyne Cover image: Svintage Archive / Alamy Stock Photo This book has been composed in Adobe Text Pro and Gotham Printed on acid-­free paper. ∞ Printed in the United States of Amer­i­ca 10 ​9 ​8 ​7 ​6 ​5 ​4 ​3 ​2 ​1

For Andrew

CONTENTS

Illustrations ix Abbreviations xi Acknowl­edgments xiii 1

How Did First-­Wave Democracies End Electoral Corruption?  1

2

The Democ­ratization of Electoral Practices  25

3

Reforming Campaigns: Reducing the Illicit Use of State Resources 69

4

Limiting Vote-­Buying and Treating  111

5

Protecting Voter Autonomy: Ballots, Envelopes, and Isolating Spaces  145

6

Curbing Electoral Fraud  179

7 Conclusion 207

Appendix 217 References 219 Index 231

vii

ILLUSTRATIONS

Figures



1.1. 2.1. 2.2. 3.1.

3.2. 3.3. 3.4.

3.5.

4.1. 4.2. 4.3. 4.4. 4.5. 5.1. 5.2.

Pathways to electoral reform 9 The determinants of demand for electoral reforms 39 Elite splits and demand for electoral reforms 47 The relative incidence of electoral strategies politicizing state resources in French and German elections 75 Determinants of support for vari­ous amendments seeking to limit the use of state resources during campaigns 95 Determinants of support for proposals to limit electoral clientelism 97 Changing electoral competition for candidates on the right following the breakdown of the electoral cartel (Kartellbündnis) 103 Correlates of support for the invalidation of elections that politicize state resources: divergence among parties on the right 107 Relative incidence of vote-­buying in French and German elections 116 The determinants of support for reforms to limit vote-­buying in the French Chamber of Deputies 124 Average prices of bribes in contested British elections relative to the price of weekly wages 128 Correlates of resource constraints for Liberal and Conservative candidates 132 The determinants of demand for reforms to limit campaign expenditures 142 Support for the introduction of the isoloir during early deliberations 162 The composition of the electoral co­ali­tion supporting the isoloir in 1907 166 ix

x I llu s tr ati o n s



5.3. 6.1. 6.2. 6.3.

Support for reforms to protect voter secrecy in Germany Incidents of fraud in German national elections Partisan strongholds in the Third Republic The determinants of support for reforms to limit fraud

176 189 203 204

­Tables

2.1. A typology of electoral irregularities 37 2.2. The composition of reform co­ali­tions 66 3.1. Proposals to limit the candidature officielle considered by the French Chamber of Deputies in 1902 87 3.2. The programmatic orientation of candidates of dif­fer­ent parties competing in the 1898 and 1906 parliamentary elections in France 94 4.1. Convictions and acquittals for bribery in German national elections 118 4.2. Partisan asymmetry in vote-­buying expenses 131 5.1. Proposals for better protection of voter secrecy considered by the German Reichstag 174

ABBREVIATIONS

H.C. Parliamentary Papers House of Commons, Parliamentary Papers

H.C. Debates House of Commons, Debates



JORF ChamberDeb Journal Officiel de la République Française, Débats parlementaires, Chambre des députés



JORF ChamberDoc Journal Officiel de la République Française, Documents parlementaires, Chambre des députés



JORF SenateDeb Journal Officiel de la République Française, Débats parlementaires, Sénat



JORF SenateDoc Journal Officiel de la République Française, Documents parlementaires, Sénat



StBer DtRt Stenographische Berichte des Deutschen Reichstages

xi

ACKNOWL­E DGMENTS

I have worked on this proj­ect for over a de­cade. In the m ­ iddle of it, I took a detour from the study of nineteenth-­century electoral irregularities to examine w ­ hether some of the forms of electoral malpractice I uncovered in nineteenth-­century elections persist ­today. The resulting book, Conditionality and Coercion: Electoral Clientelism in Eastern Eu­rope, coauthored with Lauren Young, documented practices of electoral coercion and vote-­ buying in recent democracies. The study of vari­ous forms of malpractice in con­temporary elections generated many new insights I took with me back to the study of nineteenth-­century po­liti­cal developments. During the research for this book, I benefited im­mensely from the guidance of colleagues in archives who have kindly shared their expertise. I am grateful to Joris Vanderborght for introducing me to Belgian parliamentary sources and for guiding my research. In France, I am grateful to Dominique Anglès d’Auriac and Jean-­Antonin Caheric for their kind help in uncovering parliamentary sources and for their tutorials on the po­liti­cal intricacies of the Third Republic. The first wave of the Covid pandemic hit when I was preparing the manuscript for publication. I am indebted to Mr. Caheric for helping me access parliamentary sources I was still missing at a time when travel between the United States and France was restricted. Without his help, the publication of this book would have been delayed considerably. I am also grateful to Christopher Kam for generously sharing some of his British historical data. Over the course of several years, many colleagues and friends have graciously commented on vari­ous parts of the proj­ect. For feedback and encouragement on this proj­ect, I would like to thank Susan Rose-­Ackerman, Julia Adams, Hubertus Buchstein, Robert Braun, Martin Dimitrov, Miriam Golden, Antoinette Handley, Kimuli Kasara, Andrew ­Little, Raúl Madrid, Didac Queralt, David Rueda, Maria Paula Saffon, Susan Stokes, Nadia Urbinati, Kurt Weyland, Steven Wilkinson, and Lauren Young. Two of my doctoral students at Yale, Moritz Bondeli and Alexander Trubowitz, read xiii

xiv Ac k n ow le d g m e nt s

and commented on the final version of the manuscript. I am very grateful for their feedback. I have presented the book at vari­ous seminars at the London School of Economics, the Kennedy School at Harvard, Oxford University, the University of Texas at Austin, Prince­ton University, the University of Toronto, the University of Geneva, the University of California, Berkeley, and Yale University. Many thanks to the participants in t­ hese workshops for their feedback. I am also grateful to Sarah Goldberg for her editorial assistance in preparing the manuscript for publication. ­Unless other­wise noted, all the translations from French or German parliamentary sources are my own. At Prince­ton University Press, I am grateful to Eric Crahan, Bridget Flannery-­McCoy, and Alena Chekanov for enthusiastically recognizing the importance of the manuscript and for their help in turning it into a book. Alena Chekanov skillfully managed the review pro­cess and kept me on track. Many thanks to the anonymous reviewers for providing feedback on the manuscript. I am also grateful to Jenn Backer for her superb help editing the book. I would like to thank the John Simon Guggenheim Memorial Foundation for financial support in the form of a fellowship that allowed me to work on this proj­ect. I am also grateful to Yale University for providing relief from teaching and administrative work that made it pos­si­ble for me to complete the study. Fi­nally, I could not have engaged in the long pro­cess of research that resulted in this book without the support of my f­ amily. My husband, Radu, has shared my passion for this subject and prevented me from getting lost in the details of the proj­ect and losing the bigger picture. Over the past years, my son, Andrew, has accompanied me on many research trips and has been my informal “helper” in many archives. I hope I conveyed to him the sheer happiness of research and discovery. I dedicate this book to Andrew, with love.

P ROTEC TI N G TH E BALLOT

1 How Did First-­Wave Democracies End Electoral Corruption?

The extension of voting rights during the first wave of democ­ratization created lofty hopes and expectations. A character in Flaubert’s Sentimental Education who had recently returned from the barricades of the 1848 revolution in Paris expressed the exuberance of the immediate period following the extension of suffrage as follows. “The Republic is proclaimed! ­We’ll be happy henceforth! . . . ​No more kings! You understand? The entire land ­free! The entire land ­free!” (Flaubert 2004, 2:235; Rosanvallon 2018: 261). Demo­cratic elites of the time shared this hopefulness about the transformative power of elections and anticipated that the adoption of universal suffrage would usher in far-­reaching economic and social transformations. Léon Gambetta believed that universal suffrage had “the power to establish all freedoms and all. Universal suffrage can . . . ​establish for all time real order, absolute justice, full liberty and genuine equality” (Tourneur 1904: 41). Jules Ferry, a prominent Republican politician and one of the first Republican prime ministers of the Third Republic, hailed the introduction of universal suffrage as the “law of laws” (la loi des lois). According to Ferry, universal suffrage “was a sacred and sovereign institution. . . . ​Universal suffrage is not only right, legitimate and just, but is also inevitable. It is the entire pre­sent and the entire ­future. Universal suffrage is the honor of the masses, the security of the disadvantaged, the reconciliation of classes, the law for all” (1893: 92; cf. Rosanvallon 1992: 450). Other prominent Republican politicians shared 1

2 C H A P TER 1

Ferry’s almost religious fervor for the institution of the suffrage. Louis Blanc, a historian and representative in the Chamber of Deputies, referred to the suffrage as “the triumphal arch through which all redemptive [sauveur] princi­ples ­will pass” (Rosanvallon 1992: 449), while Gambetta called it the “sacred arch of democracy” (1881, 2:223). The everyday, mundane real­ity of the first elections following suffrage expansion was a far cry from t­ hese lofty expectations. As soon as the period of canvassing began, candidates distributed money, gifts, or offers of food and drink. In Britain, where such practices of treating voters w ­ ere more pronounced, poor voters spent most of their time wandering from one public ­house to another (Seymour 1915: 134). An electoral report from Lancaster vividly described such endless pro­cessions of voters during campaigns, noting that “nothing could be more degrading than the effect of this sort of canvass on voters. They struck work, they spent the nights in public ­houses and the days in wandering about, begging from the assistant on e­ ither side or a few shillings to enable them to continue their debauch” (Hansard, P.P. 1867: 3777). In addition to bribing and treating, candidates used state employees at vari­ous levels of government during campaigns. Th ­ ese included mayors, policemen, and tax collectors or other employees of the local administration. In German national elections, candidates affiliated with the governmental majority deployed policemen to prevent opponents from distributing election material or from campaigning. In Prus­sia, Landräte state employees responsible for tax assessment and military conscription ­were routinely deployed during campaigns. At election time, the Landräte threatened voters with higher tax assessments during the following year if they supported the “incorrect po­liti­cal candidate” (Kühne 1994: 157). In France, mayors and employees of the local administration played a crucial role in mobilizing voters on election day by using a mix of promises of administrative ­favors and threats to cut local social assistance ( JORF ChamberDeb, 24 June 1902). Economic coercion was widespread at the time. On election day, enterprises turned into po­liti­cal battlegrounds. Candidates deployed rural employers to mobilize voters, equip them with the “correct ballot,” and attempt to monitor their votes by taking advantage of imperfections in voting technology. As a report of the French parliament commented with regard to ­these electoral practices, “the large landowners, the power­ful industry leaders, and, in general terms, the influential men who hold in their hands ­labor, in other words, the bread of families, . . . ​have used influences of all kinds, threats, repression, intrigues. They have attempted to distort elections

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 3

through a close surveillance of the vote, by controlling the ballots . . . ​at the time t­ hese ballots are handed over to the president of the voting station” ( JORF ChamberDoc 1890, no. 107). The widespread presence of vari­ous forms of malfeasance contributed to disillusionment with elections. Criticisms contending that electoral rights ­were only hollow promises and that elections w ­ ere a form of hy­poc­r isy intensified. Commenting on the first elections held in the Third Republic, one publication noted that “the universal suffrage, the expression of National Consciousness, was at the mercy of the cupidity and ambitions of its guardians. While appearing sovereign, France is actually enslaved.” Such a state of affairs was “humiliating, immoral and dangerous” (Legrand 1877: 10). Paul Granier de Cassagnac, a legislator from Gers, warned that the “universal suffrage has lost his prestige and authority” ( JORF ChamberDeb, 18 December 1901). In France, this criticism was shared by both opponents and the strongest supporters of the Republican regime. A report of the Commission of Universal Suffrage of the French Chamber of Deputies noted that elections “failed to represent the true ­will of the ­people” ( JORF ChamberDoc 1902, no. 181). Charles Benoist, a Republican legislator from Paris, commented on the tension between the promise of universal suffrage and the practical realities of elections: “In this country . . . ​po­liti­cal philosophy requires that one considers the ­people as the exclusive, inexhaustible, and constantly renewed source of power and law. If the universal suffrage is not ­free, and I say fully and absolutely ­free, . . . ​this philosophy ­will be nothing but a lie, this power, nothing but an usurpation, this law nothing but a scorn” ( JORF ChamberDeb, 23 May 1907). At the same time, this recognition of the shallowness of the demo­cratic achievement set in motion an arduous po­liti­cal pro­cess that sought to remedy the multitude of electoral imperfections. In many countries, this pro­ cess unfolded over several de­cades. Long periods of lull w ­ ere punctured by surges in interest in reforming elections, only to be followed by neglect. Given that many of the legislators participating in ­these deliberations ­were beneficiaries of the electoral status quo, the formation of an encompassing co­ali­tion that supported reforms to de­moc­ra­tize electoral practices was a highly difficult pro­cess. In The Spirit of Laws, Montesquieu remarked that the law “which determines the manner in which ballots [billets de suffrage] are distributed is a fundamental law in a democracy” (2002: 284). H ­ ere, Montesquieu highlights the importance of laws that or­ga­nize practical, concrete details of elections for the functioning of democracies. Nineteenth-­century po­liti­cal handbooks

4 C H A P TER 1

also stressed the importance of material guarantees as safeguards of the voting pro­cess. The entry on voting in the General Dictionary of Politics published in 1867 noted that “the freer a ­people is, the more carefully it surrounds the vote with the most meticulous guarantees” (Block 1880: 1123). In an 1896 study detailing the menu of electoral irregularities, an official of the French Ministry of Interior remarked that the freedom of the vote depended on its materiality (Uzé 1896: 9; cf. Ihl 1998: 84). While the comparative lit­er­a­ture on democ­ratization has sidestepped questions about the practical organ­ization of elections, the goal of this study is to fill this theoretical and empirical gap. This book examines the adoption of electoral reforms that sought to protect the integrity of the electoral pro­cess and reduce opportunities for malfeasance. By addressing material imperfections in the voting pro­cess, ­these reforms attempted to complete the demo­cratic proj­ect begun with the expansion of suffrage. By eliminating undue influences on voters’ autonomy, as well as opportunities for electoral malfeasance, the reforms w ­ ere designed to transform nominal voting rights granted as a result of suffrage expansion into de facto po­liti­cal rights. In this study, I examine reforms that affect ­every stage of the electoral pro­cess: the organ­ization of campaigns, the moment of the vote, and the counting of ballots. One set of changes in electoral laws concerned the organ­ization of campaigns. Th ­ ese reforms attempted to sanction candidates’ ability to reduce the incidence of vote-­buying or treating. The Napoleonic Code that was ­adopted by a large number of Eu­ro­pean countries during the first half of the nineteenth ­century included one early effort to limit electoral fraud. This approach subjected exchanges that involved offers of money or goods to penal laws and established harsh punishments for bribery. Such provisions of the penal code met with a l­imited degree of success in curbing t­ hese illicit strategies ­because the ­legal burden of prosecuting bribery remained prohibitively high. As a result, countries that ­adopted the Napoleonic Code attempted to find other ways to sanction fraud throughout the nineteenth ­century. Po­liti­cal developments in France illustrate this trajectory of reform. ­Here, politicians attempted to introduce additional laws that complemented the penal code, but at the same time made it easier to sanction fraud, by relying on courts. By contrast, reformers in Britain a­ dopted a dif­fer­ent approach by placing stricter limits on campaign expenditures and imposing stronger punishments on candidates who exceeded ­these limits. The second set of changes concerned the voting pro­cess itself. ­Here reformers attempted to reduce the ability of candidates and their brokers

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 5

to take advantage of imperfections in voting technology and pierce voter secrecy. The first component of ­these reforms was the adoption of ballot envelopes. Initially, this electoral innovation was a­ dopted with g­ reat hesitation. In 1868, Württemberg introduced ballot envelopes but eliminated this provision in 1882 (Duplantier 1901: 245). Canada introduced ballot envelopes in 1876 but rescinded this provision in 1888 (ibid., 246). Swiss cantons played a pioneering role in the introduction of this legislation. Waadt (or Vaudt), a canton in western Switzerland, mandated the introduction of ballot envelopes in 1881. This was followed by Tessin in 1889, St. Gall in 1890, and Luzern in 1893 (ibid., 244–46). Legislation introduced in Norway in 1884 mandated ballot envelopes but required voters to sign their ballots. ­After a long period of negotiations, the German Reichstag mandated the introduction of ballot envelopes for national elections in 1903. A related set of provisions to protect voter secrecy mandated the introduction of isolating spaces. The goal of ­these reforms was to allow voters to escape the gaze of candidate representatives located in the vicinity of ballot containers, known at the time as electoral urns. In 1856, Australia pioneered the introduction of isolating spaces (Duplantier 1901: 259). Britain introduced this reform as part of the Ballot Act enacted in 1872 (35 & 36 Vict., Article 26). The Ballot Act, in turn, served as a model for comprehensive electoral reform a­ dopted in Belgium in 1877, the loi Malou. Luxembourg and the Netherlands introduced isolating spaces in 1879 and 1896, respectively. Romania a­ dopted them as part of its electoral reform of 1899. Imperial Germany mandated isolating spaces together with ballot envelopes in 1903. France a­ dopted the reforms in 1914. A third set of reforms attempted to minimize the ability of election administrators to modify the outcome of elections at the time votes w ­ ere counted. One strategy to reduce this electoral irregularity was to introduce harsh punishments ­under the penal code. This solution was a­ dopted in Germany. An alternative ave­nue of reform considered by nineteenth-­century parliaments mandated the presence of representatives of all candidates when ballots w ­ ere counted. Th ­ ese electoral reforms w ­ ere a­ dopted in Britain in 1872 and in Belgium in 1877. The French parliament considered ­these proposals during multiple rounds of deliberations, spanning nearly three de­cades, but a po­liti­cal majority in ­favor of this reform could not be assembled. Electoral reforms to protect electoral integrity w ­ ere introduced at vari­ ous stages in the pro­cess of democ­ratization. Consequently, universal manhood suffrage was not a prerequisite for electoral reforms. Britain ­adopted reforms demo­cratizing elections in parallel with reforms extending suffrage.

6 C H A P TER 1

In Belgium, ­these significant electoral reforms ­were ­adopted prior to the introduction of mass suffrage. Developments in Imperial Germany contrast with t­ hose of other Eu­ro­pean countries. Despite its early extension of universal manhood suffrage, Germany granted elected legislators no responsibilities in forming governmental majorities. Even though Germany remained an electoral authoritarian regime ­until 1914, the Reichstag ­adopted electoral reforms protecting voter autonomy in 1903. France a­ dopted t­ hese electoral reforms with significant delay ­after the introduction of universal suffrage. Over seventy years lapsed between the introduction of universal manhood suffrage in 1848 and the ac­cep­tance of electoral reforms to reduce opportunities for electoral malfeasance. The goal of this study is to explain the po­liti­cal origin of electoral reforms designed to guarantee greater integrity of the electoral pro­cess by removing opportunities for illicit influence and fraud. I seek to understand the motivations of politicians who demanded t­ hese reforms and the pro­cess by which parliamentary majorities that supported reforms came about. I attempt to chronicle the initial rise in demand for t­ hese reforms and explain the motivations of legislators who supported electoral change. Who ­were the legislators who supported the introduction of ­these changes in electoral law? What explains their dissatisfaction with the status quo? Did this initial parliamentary impetus in support of electoral reforms vary across countries and across policy areas? Next, I turn to the negotiation of electoral reforms and seek to characterize the composition of the po­liti­cal majorities that supported reforms. My objective is to understand why the composition of the legislative majorities supporting t­ hese electoral reforms varied across countries and across policy areas. The Argument in Brief

In a 2009 article, Adam Przeworski presented the central puzzle examined by the democ­ratization lit­er­a­ture as follows: “Why would p ­ eople who monopolize po­liti­cal power ever decide to put their interests or values at risk by sharing it with o ­ thers? Specifically, why would t­ hose who hold po­liti­cal rights in the form of suffrage decide to extend t­ hese rights to anyone e­ lse?” (2009: 291). While Przeworski considers dilemmas associated with the expansion of suffrage, his formulation naturally applies to the demo­cratizing reforms examined in this study. Why do legislators elected ­under rules that permit some form of malfeasance choose to modify the laws that made their victory pos­si­ble? Similar to the extension of suffrage, the choice to replace a

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 7

familiar institution with a novel one involves significant risk-­taking by elected lawmakers. Reforms intended to reduce illicit strategies redistribute influence within the existing electorate by allowing the translation of preferences of some constituents into votes with less noise and bias. One of the goals of this study is to provide an account of the considerations politicians brought to bear on t­ hese decisions and the formation of majorities that supported a change in electoral rules. In this book I propose a straightforward solution to Przeworski’s democ­ ratization puzzle. I show that legislators turned to electoral reforms as economic and po­liti­cal changes such as elite splits increased the costs associated with dif­fer­ent forms of electoral malfeasance. For legislators who faced such rising economic and electoral costs, electoral reforms presented an opportunity to level the playing field and impose constraints on the electoral strategies of their competitors. Understanding demand for reform requires us to unpack ­these economic and electoral costs that reduced the attractiveness of the status quo. With regard to the reforms examined in this study, legislators had to choose between maintaining the status quo that provided opportunities for electoral malfeasance and supporting an alternative state that foreclosed such opportunities. My first assumption is that support for the status quo is conditioned by politicians’ access to resources that can be politicized during campaigns. Such assets may include public resources of the state or private resources financed by the candidate or by the party. The former may include access to employees of the state administration who can be deployed during campaigns or access to the policy resources that can be offered to voters in exchange for their support. Private resources may include wealth or access to private employers as brokers. The distribution of resources that can be politicized affects the initial po­liti­cal cleavage over the desirability of electoral reforms and the size of initial majorities supporting reforms. As long as a majority of legislators can access such resources and deploy t­ hese seamlessly during elections, the impetus for po­liti­cal change is low. B ­ ecause the distribution of resources determines support for the status quo, the composition of po­liti­cal majorities that oppose reforms is likely to vary across reform dimensions. The composition of the majority supporting the status quo is likely to differ across irregularities that politicize private or public resources. ­These initial majorities are, however, not static. This majority of resource-­ endowed politicians may disintegrate and give way to a new majority if po­liti­cal or economic changes increase the costs associated with vari­ous

8 C H A P TER 1

illicit strategies. I consider two constraints on the use of illicit strategies. The first is the economic costs associated with dif­fer­ent forms of malfeasance. The use of illicit strategies is likely to vary across space and time and be influenced by the economic characteristics of a par­tic­u­lar constituency. All ­things being equal, a candidate who promises money or gifts in exchange for his vote ­will face higher economic costs if the number of poor voters that are necessary for this victory is higher. The economic costs of bribing are likely to be higher in regions with more generous anti-­poverty programs or higher minimum wages. Electoral strategies that deploy employers or other private actors as brokers are also constrained by economic conditions in a par­tic­u­lar region. The costs of economic coercion are higher in districts where l­abor is scarce or where a number of competitors can rehire voters targeted by coercive strategies. In addition to economic costs, a politician who engages in bribery or coercion may also incur electoral costs. A candidate incurs electoral costs if the use of an illicit strategy reduces, rather than increases, the total number of votes received by the candidate in the respective constituency. A candidate incurs electoral costs if voters sanction electoral malfeasance. He may also incur electoral costs if the use of such a strategy reduces his opportunities to form electoral co­ali­tions with other candidates. Consider two examples of electoral costs. A politician who competes on the basis of programmatic policies but engages in clientelistic promises may incur electoral costs. Voters are likely to punish a candidate who seeks to combine such appeals as “schizophrenic” (Kitschelt 2000). In an electoral system based on runoffs, a candidate who uses illicit strategies may incur electoral costs. Potential co­ali­tion partners may be deterred by t­ hese strategies and support his opponent during runoffs. As t­ hese examples illustrate, electoral costs may result from the responses of both the voters and the other candidates in electoral systems that require coordination during runoffs. Such increases in economic and electoral costs are significant b ­ ecause they may bring about a co­ali­tional realignment in support of a policy that reduces opportunities for malfeasance. A majority supporting reforms emerges if a group of resource-­endowed politicians reconsiders the attractiveness of the status quo. The pivotal legislators in the new electoral majority supporting reforms are resource-­endowed legislators who encounter high economic or electoral costs associated with illicit strategies. For legislators facing high economic or electoral costs, electoral reforms are preferrable to the status quo, b ­ ecause they constrain resource-­endowed legislators who do not face such constraints from using illicit strategies in ­future elections.

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 9

Economic development

Initial cleavage line based on relative resource advantage

Economic costs of illicit strategies increase

Political majority supporting reforms

Electoral costs of using illicit strategies increase

Elite splits FIGURE 1.1.

Pathways to electoral reform.

Figure 1.1 outlines the f­actors that contribute to the rise in electoral and economic costs, respectively. Economic development, on the one hand, and elite splits, on the other hand, reduce the attractiveness of the status quo even among legislators with an initial resource advantage. Taken together, ­these developments facilitate the formation of po­liti­cal majorities that support electoral reforms. Th ­ ese electoral majorities are likely to include both resource-­constrained legislators and resource-­endowed legislators who face high economic and electoral costs. The nineteenth c­ entury was a period of massive economic transformations. ­These economic developments included industrialization, urbanization, and rising income levels. All of ­these changes increased the economic costs of vari­ous forms of electoral malfeasance, reducing their electoral desirability. ­These economic developments w ­ ere more consequential for clientelistic strategies that used private economic resources. Such strategies included both vote-­buying and economic coercion that uses private actors. A historian of German elections commented on the electoral consequences of ­labor scarcity experienced by many Prus­sian districts during the 1890s as follows: “In recent periods, landowners had to use this means of power [Machtmittel] very carefully ­because of the ­labor shortage that existed in the countryside.

10 C H A P TER 1

One was happy if one could keep one’s employees and one was careful to not antagonize the employees through electoral harassments and to not drive them to the cities” (Wulff 1922: 13). As this example suggests, l­ abor scarcity resulting from economic development and migration to cities increased the economic costs of repressive electoral strategies. In addition to economic developments, elite splits also have significant consequences in affecting legislators’ preferences about the desirability of electoral reforms. Elite splits ­were a common po­liti­cal development in many first-­wave democracies. In this study, I define elite splits in electoral terms, not as economic conflicts between ­owners of dif­fer­ent ­factors. Such electoral splits among elites may come in a variety of forms. A party may break up into dif­fer­ent factions that decide to field their own candidates. Alternatively, an informal pre-­electoral alliance among parties may disintegrate, leading to the fielding of competing candidates. Elite splits raise the electoral costs associated with vari­ous forms of malfeasance. In this study, I ­will document two distinct pathways by which elite splits raised the electoral costs associated with illicit strategies. The first of ­these mechanisms is pre­sent in countries whose electoral systems require electoral coordination in the second round. During the period examined in this study, many Eu­ro­pean countries, including Belgium, France, Germany, and the Netherlands, had majoritarian elections with runoffs. ­These electoral systems allowed a large number of candidates to enter in the first round but necessitated high levels of electoral coordination among candidates during a runoff. To win in the second round of a race, a candidate needed endorsements from former competitors. Such anticipation of electoral coordination in the second round increased the electoral costs associated with illicit strategies. Using dif­fer­ent forms of malfeasance in the first round was likely to increase the wrath of opponents and reduce their willingness to endorse a candidate who had used illicit strategies. As such, a candidate who relied on dif­fer­ent forms of malfeasance in the first round could find himself without co­ali­tion partners in the runoff, a situation that could lead to electoral loss. Such considerations about electoral costs affected legislators’ views about the desirability of the status quo in the aftermath of elite splits. Following such splits, the pivotal group of legislators who embraced electoral reforms ­were resource-­endowed legislators who now faced increased competition in runoffs. Illustrating this logic, the book ­will document the shift in the position of German National Liberals about the desirability of electoral reforms following the breakdown of the electoral cartel with Conservatives.

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 11

Second, elite splits also contributed to the formation of ­either parties or groups of politicians competing on the basis of programmatic promises. The most prominent example of such an elite split was that between Republicans and Radicals in the Third Republic. The split between vari­ous Liberal groups in Germany was also based on programmatic differentiation among ­these splinter groups. Elite splits altered candidates’ ability to access resources necessary for the production of electoral malfeasance and at the same time increased incentives for programmatic differentiation. For candidates who ­were competing on the basis of programmatic promises, illicit strategies became an increasing liability, as they could undermine the credibility of their promises in the eyes of voters. In this po­liti­cal environment, elite splits also contributed to an increase in electoral costs. The shift in the position of ­these legislators facilitated the formation of a new parliamentary majority supporting electoral reforms. Legislators competing on the basis of programmatic promises became the pivotal group in the majority supporting reforms. Let us consider some examples that illustrate the po­liti­cal consequences of elite splits. During Bismarck’s chancellorship, Germany’s parties on the right—­the Conservatives, National Liberals, and ­Free Conservatives—­ formed an electoral alliance (Kartellbündnis). Additional alliances established by party organ­izations in each province underpinned this national cartel. As a result, the alliance fielded only one candidate in each district, who could benefit from the electoral support of the local state apparatus. Around 1900, this national electoral cartel disintegrated as a result of conflict among the parties on the right over trade issues. The breakdown of the Kartellbündnis changed the electoral considerations of candidates on the right about the attractiveness of using state resources. In this more fragmented po­liti­cal environment, candidates on the right found themselves in the novel situation of competing against each other while being, at the same time, dependent on co­ali­tions in runoffs. This increase in electoral competition raised the electoral costs of strategies that politicized state resources. The use of such resources as policemen or state officials during campaigns by one candidate reduced the willingness of his opponents to support him during runoffs. This cartel split affected the electoral strategies and the preferences for reforms of some legislators previously affiliated with t­ hese parties. Sensing ­these growing electoral costs, officials of Prus­sia’s interior minister began to obfuscate their electoral strategy of politicizing state resources, attempting to pursue a policy of “noiseless” deployment of the state apparatus (Müller 1963; Mares 2015: 51). For National Liberals, rising electoral costs

12 C H A P TER 1

contributed to an open indictment of electoral strategies that politicized state resources. Following the breakup of the Kartellbündnis, National Liberals supported the invalidation of candidates who politicized state resources. On the floor of the Reichstag and in the commission verifying electoral irregularities, National Liberals supported harsher punishments for candidates who used state resources during campaigns. The National Liberals’ shift in position contributed to the formation of a new majority willing to sanction electoral malfeasance. As a result, the electoral jurisprudence ­adopted by the Reichstag included increasingly sharper sanctions for candidates who used state resources during campaigns (Hatschek 1920). Following the breakdown of the electoral cartel, the National Liberals also supported electoral reforms intended to ensure voter secrecy by cosponsoring legislation that mandated uniformly designed ballot containers. In neighboring France, the elite split between Republicans and Radicals also contributed to the formation of po­liti­cal majorities that supported electoral reforms. Although disagreement between ­these factions goes back to the first elections of the Third Republic, the split resulted in the formation of an in­de­pen­dent Radical Party in 1902. As Radical Socialists competed on the basis of programmatic appeals—­including demands for the introduction of progressive income taxes, social insurance, and ­labor market reforms—­ clientelistic strategies became po­liti­cally counterproductive and ­were perceived more as a liability than as a source of electoral advantage. As Radicals began to endorse electoral reforms, the disagreement between Radicals and Republicans over the desirability of ­these reforms intensified. This repositioning of Radical legislators on the question of electoral reforms was the decisive f­ actor in the creation of a po­liti­cal majority that supported reforms. In the case of electoral reforms that l­imited the use of state resources during campaigns, Radicals’ change in position about the desirability of ­these reforms contributed to the adoption of reforms that sanctioned electoral clientelism. A similar po­liti­cal dynamic was at play in the case of vote-­buying reforms and contributed to the adoption of ­these reforms. Fi­nally, Radical legislators w ­ ere a pivotal group in facilitating the formation of an electoral majority supporting better protection of voter secrecy. B ­ ecause ­these reforms w ­ ere opposed by both centrist Republicans and Monarchist legislators, the repositioning of Radicals was of crucial importance in guaranteeing the adoption of ­these reforms. The foregoing discussion has outlined an answer to Przeworski’s “democ­ ratization dilemma” for reforms that limit vari­ous types of electoral malfeasance. In cases where both economic costs and electoral costs are low, the

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 13

initial resource asymmetry explains the size and composition of the majority opposing change. In such cases, I expect that a po­liti­cal majority of resource-­ endowed politicians w ­ ill reject proposals for reform. By contrast, rising electoral and economic costs are likely to reduce the desirability of the status quo for a number of legislators and contribute to the formation of a dif­fer­ent po­liti­cal majority favoring reforms. In my explanation, resource-­endowed legislators who encounter high economic or electoral costs are po­liti­cally pivotal. In Third Republic France, this role was played by Radical politicians who broke away from the Republican majority. In Imperial Germany, National Liberals played a similar role during the reforms associated with the protection of voter secrecy. In addition, this book generates new insights about the position of parties on the right about electoral reforms. A number of studies infer such parties’ positions ­toward electoral reforms from their affiliation with par­tic­ u­lar ­factors of production, such as land or capital (Boix 2003; Ziblatt 2017). Reflecting the interests of immobile ­factors, such as land, ­these parties are expected to oppose democ­ratization. The preferences of ­these legislators are also expected to be stable over time and across dimensions. In contrast to t­ hese explanations, I suggest that the position of parties on the right with regard to electoral reforms ­will vary across policy areas, depending on the available resources that can be politicized during elections. Consider the positions of Monarchists and centrist Republican legislators in France during the negotiation of electoral reforms limiting dif­fer­ent forms of malfeasance. B ­ ecause Monarchists could only access private resources during campaigns, they opposed electoral reforms to limit bribery and treating of voters. At the same time, Monarchists and other non-­Republican legislators on the right supported reforms that would limit the politicization of state resources during campaigns. Republican candidates, who could access po­liti­cal resources of the state, embraced the opposite position during ­these reforms. ­These legislators supported reforms that ­limited vote-­buying but opposed reforms that attempted to rein in the use of state resources. In short, the overarching argument of this book can account for the dif­fer­ent positions ­toward electoral reforms held by parties on the right. It can also explain why the composition of the electoral co­ali­tions favoring reforms w ­ ill vary across policy areas. The crucial variables that explain t­ hese dif­fer­ent electoral majorities are the distribution of resources and the magnitude of economic and electoral costs. The analy­sis presented in the book thus joins a growing lit­er­a­ture that has examined the importance of elite splits for demo­cratic reforms (Ansell and

14 C H A P TER 1

Samuels 2014; Madrid 2019a, 2019b). The lit­er­a­ture on third-­wave democ­ ratization anticipated the importance of elite splits. Scholars such as Guillermo O’Donnell and Philippe Schmitter (1986) and Adam Przeworski (1991) argued that an elite split sets in motion a demo­cratic transition. In recent years, explanations that stress the importance of elite splits for demo­cratic reforms have experienced a comeback. Disagreement persists, however, as to how to conceptualize the salient elites as well as the sources of the split. Ben Ansell and David Samuels conceptualize elites as economic classes rather than partisan groups. In this framework, landholding elites oppose the extension of suffrage b ­ ecause it encroaches on their rights, while the ­middle classes support democ­ratization in order to limit expropriation by state authorities (Ansell and Samuels 2014). In a number of articles analyzing the adoption of demo­cratic reforms in Latin Amer­ic­ a, Raúl Madrid (2019a, 2019b) also focuses on the importance of elite splits but conceptualizes them in partisan terms. According to Madrid, splits within a ruling party weaken the control of insiders, while providing opportunities for the outside group to enact reforms (Madrid 2019b: 1540). Madrid shows how intra-­elite splits within the ruling PAN party in Argentina contributed to the adoption of secret ballot reforms in 1912. Such reforms ­were supported by legislators who defected from the party in power in an effort to weaken the po­liti­cal machine of the incumbent. As w ­ ill be discussed in chapter 2, the analy­sis of this book emphasizes partisan rather than economic elites. Fi­nally, this book speaks to theories that attribute an impor­tant role to Social Demo­cratic parties as agents responsible for demo­cratic reforms (Rueschemeyer, Huber Stephens, and Stephens 1992). In contrast to ­these explanations, I show that the importance of such parties representing lower-­ income voters during ­these demo­cratic breakthroughs was relatively ­limited. While Social Demo­cratic legislators supported the adoption of reforms to limit vari­ous forms of electoral malfeasance, their participation was not a sufficient condition for a­ ctual policy change. By documenting the formation of electoral majorities that favored electoral reforms, I show that decisive po­liti­cal majorities form only ­after legislators affiliated with parties on the right who often had the resources to engage in vari­ous forms of electoral malpractice shifted their position in ­favor of reforms. To understand the adoption of ­these demo­cratic reforms, we need to unpack the calculations of legislators associated with parties in the po­liti­cal center and the po­liti­cal right. The shift in ­these legislators’ position explains the timing of the adoption of the dif­fer­ent reforms demo­cratizing electoral practices examined in this book.

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 15

Relationship to the Lit­er­a­ture

In deriving its theoretical expectations about the sources of demand for electoral reforms, this study considers first and foremost the electoral motivations of politicians. Politicians choose electoral reforms that improve the likelihood of their f­ uture electoral victory by considering their available po­liti­cal resources and the economic and electoral costs to deploy such resources. In this section, I show how the theoretical predictions of this explanation build on, but also challenge, two dominant perspectives in the study of democ­ratization. The first of ­these approaches is the modernization perspective, which examines the consequences of economic development for the adoption of demo­cratic reforms. The second approach, the redistributive approach, derives predictions for demo­cratic reforms—­such as decisions to extend suffrage—­from expectations about the ­future level of taxes and spending. While ­these perspectives provide impor­tant insights that are integrated in this study, neither of ­these approaches can explain the complex co­ali­tional dynamics that led to the adoption of demo­cratizing reforms limiting electoral malfeasance. ECONOMIC MODERNIZATION AND DEMO­CRATIC REFORM

The most enduring theoretical perspective in the study of democ­ratization is the modernization approach. Seymour Lipset’s study “Some Social Requisites of Democracy,” published in 1959 (see also Lipset 1994), was one of the earliest contributions that developed this perspective. Lipset uncovered a positive relationship between levels of economic development and the presence of a demo­cratic regime. The modernization perspective was revived during the 1990s by Adam Przeworski and his collaborators. In a number of studies, ­these scholars have examined the correlations between economic development and regime transition and survival in a large number of regime changes in the period ­after World War II (Przeworski and Limongi 1997; Przeworski et al. 2000). Boix and Stokes (2003) extended ­these studies for first-­wave democracies. In recent years, a number of studies have extended predictions from the modernization perspective to the study of limiting electoral malfeasance. ­These studies have examined the consequences of economic modernization on reforms that limit vote-­buying, focusing predominantly on the British experience (Cox 1987; Kam 2017; Stokes et al. 2015). One consequence of Britain’s economic development throughout the nineteenth c­ entury was the

16 C H A P TER 1

increase in the size of the constituency. Gary Cox examined several pro­cesses by which the growth in the size of constituencies reduced the attractiveness of electoral bribery. On the one hand, this increase stretched the bud­get of politicians, as “a fixed amount of money would buy a smaller proportion of total votes in small towns.” On the other hand, “even if the price of votes was less (in proportion to the greater number of voters), the costs of arranging to bribe many more electors, not to mention the increased risk of being caught, made bribery a less attractive option” (Cox 1987: 57). In analyzing the British case, Susan Stokes and colleagues (2015) propose an additional mechanism linking economic development and the adoption of reforms that limit vote-­buying. They conjecture that modernization exposed the monitoring difficulties endemic to vote-­buying exchanges. Stokes and her coauthors argue that clientelistic exchanges such as bribery provide no economies of scale to candidates (ibid., 216). As a result, it became increasingly difficult for brokers to structure clientelistic exchanges with voters and for candidates to monitor the activities of brokers. They note that “not just ‘arranging to bribe’ but holding the bribe’s recipient to account was a costly ­matter, one that was labor-­intensive, requiring close and continuous contact between large numbers of electoral agents and individual voters” (ibid.). The growth of the size of the electorate increased candidates’ incentives to invest more in programmatic linkages to voters. “When the national electorate and local constituencies grew, party programs and print appeals became well worth the investment they required” (ibid.). Empirically, the modernization perspective has identified macro-­level correlations between broad structural changes and demo­cratic reforms using cross-­national data. While Cox (1987) and Stokes et al. (2015) spell out the implications of economic development for the strategies of politicians, ­these studies do not test their theoretical implications with individual-­level data. In empirical terms, my study contributes to the modernization perspective by providing one of the first tests of this lit­er­a­ture using legislator-­level data. In theoretical terms, I propose an explanation that examines how economic development interacts with the resources available to legislators, on the one hand, and with their electoral costs, on the other hand, to affect demand for reform. While economic development may increase the costs of vari­ous illicit strategies, the importance of ­these pro­cesses in explaining the success of reforms needs to be assessed in conjunction with ­these additional variables. As discussed above, economic development is a decisive ­factor in the formation of a po­liti­cal majority that supports reforms only if it raises the economic costs of electoral malfeasance of resource-­endowed legislators.

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 17

As subsequent chapters ­will document, in a significant number of cases, co­ali­tional realignment in support of electoral reforms did not come about as a result of economic changes but as a result of elite splits that increased politicians’ electoral costs associated with vari­ous forms of malfeasance. REDISTRIBUTIVE THREATS

Other prominent theories explain elites’ decisions to extend suffrage and de­moc­ra­tize as responses to threats of mass unrest and massive upheaval (Acemoglu and Robinson 2000, 2006). In this explanation, po­liti­cal elites decide to extend suffrage and engage in other far-­reaching po­liti­cal reforms in an effort to prevent massive unrest, radicalization, and revolution. Daron Acemoglu and James Robinson formulated the central proposition of this explanation: democ­ratization occurs “­because the disenfranchised citizens can threaten the elite and force it to make concessions. Th ­ ese threats can take the form of strikes, demonstrations, riots, and—in the limit—­a revolution. B ­ ecause ­these actions impose costs on the elite, it ­will try to prevent them. . . . ​The elite must de­moc­ra­tize—­create a credible commitment to ­future majoritarian policies—if it wishes to avoid more demo­cratic outcomes” (2006: xii–­xiii). Incumbent po­liti­cal elites are the crucial actors who choose the level and scope of suffrage extension. In developing their predictions about suffrage extension, Acemoglu and Robinson derive the preferences of elites from considerations about their ­future tax burden ­under new electoral rules (2006: 36). The change relative to the current tax burden of elites is a function of the level of in­equality in a society. Consequently, elites incur higher ­future costs in terms of taxation in socie­ties with higher levels of in­equality. As a result, elites in socie­ties with higher levels of in­equality are more reluctant to fully extend suffrage and ­settle, instead, for more repressive po­liti­cal regimes. The most impor­tant limitation of this theory is the extremely long causal chain between the level of in­equality and elites’ position ­toward reform. This opens up a high indeterminacy that is never tested empirically. Carles Boix (2003) pre­sents a variation of this explanation. Similar to Acemoglu and Robinson, Boix also derives predictions about elites’ decision to extend suffrage from their considerations about their expected tax rates ­under demo­cratic or authoritarian regimes. Boix adds the twist that not all assets are fully taxable, as some elites retain their ability to shift their assets to a dif­fer­ent constituency. This reduces the redistributive threats of a demo­ cratic transition even in a country with high levels of income in­equality.

18 C H A P TER 1

It follows that the relationship between asset in­equality and demo­cratic transitions is mediated by the asset specificity of an economy. The most impor­tant aspect of ­these explanations is their ability to link structural inequalities, on the one hand, and differences in demo­cratic outcomes, on the other hand. Th ­ ese theories capture some well-­known stylized facts: highly unequal countries are more likely to remain authoritarian, while countries with lower levels of asset inequalities are more likely to experience a demo­cratic transition. Similarly, the high mobility of capital assets may have facilitated demo­cratic transitions even in countries with other­wise high levels of in­equality. Britain’s democ­ratization during the nineteenth c­ entury may illustrate this development. Nevertheless, while the “end points” of ­these explanations reflect known empirical correlations, the intermediate steps connecting differences in in­equality and regime types proposed by ­these theories are less persuasive and untested. Redistributive theories of democ­ratization have impor­tant limitations. First, as discussed above, the causal chain connecting structural economic conditions and elites’ support for regime outcomes is extremely long. Th ­ ese intermediate steps, which are never explic­itly tested or addressed, involve a lot of additional assumptions about the adoption of ­future levels of taxation and re­distribution and the implementation of ­these policies in the new democracy. Surreptitiously, ­these assumptions shift the analytical emphasis away from bargains over electoral rules ­toward bargains about levels of taxes and spending. The assumptions made about ­these additional steps of the po­liti­cal pro­cess are never tested empirically. Why should one assume that in a recent democracy out­going elites have no influence over the formulation of tax policy? Why should we assume that they have no role in the design of the institutions implementing social policy? Why should we follow the implications of the Meltzer-­Richard model that higher in­equality results in higher levels of re­distribution, given the overwhelming empirical evidence contradicting the predictions of this model? The level of redistributive threat, which is the crucial intervening variable in t­ hese theories of democ­ ratization, is never sufficiently unpacked theoretically and empirically. In this study, I challenge the premise made by redistributive theories of democ­ratization that we need to sneak in considerations about ­future levels of taxation to explain choices over electoral rules. Rather, I suggest that we need to consider electoral motivations and examine how t­ hese considerations interact with elites’ economic positions to explain support for reforms. In reintroducing electoral calculations into the study of demo­cratic reforms, I join a po­liti­cal science lit­er­a­ture that goes back to Schattschneider

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 19

(1960) but that has been revived in recent years (Valelly 2004; Teele 2018). Once we consider the electoral motivations of legislators, we can develop a range of additional predictions about demand for demo­cratizing reforms. The analytical superiority of the account presented in this book—­relative to modernization theories or redistributivist theories of democ­ratization—­lies in its ability to account for the dynamic changes in the composition of electoral co­ali­tions supporting electoral reforms as well as for the heterogeneity in the composition of ­these co­ali­tions. In recent years, scholars have proposed “big bang” theories of democ­ ratization. Such theories presuppose that all demo­cratic outcomes—be it the expansion of suffrage, the protection of ballot secrecy, or the elimination of fraud—­follow a similar po­liti­cal logic and that the same variables can seamlessly explain the adoption of ­these reforms. In contrast to t­ hese studies, I suggest that it is essential to disaggregate dif­fer­ent reform dimensions. The winners and losers of electoral reforms differ across vari­ous types of electoral reforms. As a result, the po­liti­cal co­ali­tions that support electoral reforms also vary across dif­fer­ent dimensions of democ­ratization. This study of electoral reforms in first-­wave democracies provides a useful blueprint for the analy­sis of demo­cratic reforms in other historical or con­temporary contexts. Empirical Strategy

The empirical strategy of this study differs from existing approaches to the study of democ­ratization. Foundational studies in the lit­er­a­ture explain demo­cratic transitions as a result of conflicts between dif­fer­ent po­liti­cal classes. Barrington Moore’s seminal contribution (1966) to regime transitions explains the historical trajectories leading to liberal democracy, fascism, and communism as a result of interactions between landlords and commercial elites, on the one hand, and landlords and peasants, on the other. Gregory Luebbert (1991) explains regime outcomes in interwar Eu­rope as a result of the po­liti­cal mobilization of the working classes and landless and middle-­class peasants. Dietrich Rueschemeyer, Evelyne Huber Stephens, and John Stephens (1992) explain differences in demo­cratic transitions as a result of differences in working-­class mobilization. By operating with highly aggregated po­liti­cal actors, t­ hese approaches sidestep impor­tant questions of po­liti­cal agency. A second empirical strategy used in the lit­er­a­ture on democ­ratization has considered individual countries as the main unit of analy­sis. Th ­ ese analyses of demo­cratic transitions using cross-­national data sets have been widely

20 C H A P TER 1

used (Boix 2003; Przeworski and Limongi 1997; Przeworski et al. 2000; Houle 2009; Ansell and Samuels 2014; Acemoglu and Robinson 2006; Haggard and Kaufman 2016). ­These authors have attempted to unpack correlates between a variety of economic ­factors and demo­cratic transitions. Both Boix (2003) and Ansell and Samuels (2014) examine demo­cratic transitions using both historical and postwar samples, covering the period between 1850 and 2004. In a recent example of this line of research, Stephan Haggard and Robert Kaufman (2016) have examined demo­cratic transitions and demo­ cratic backsliding during the period 1980–2008. In this study, I pursue a dif­fer­ent empirical strategy. I examine decisions to adopt demo­cratic reforms from the perspective of individual legislators who vote on proposals that seek to reduce vari­ous opportunities for electoral malfeasance (for similar studies, see Mares 2015; Aidt and Franck 2015; Teele 2018). In an ­earlier study (2015), I referred to this approach as “micro-­historical institutionalism” and discussed the empirical advantages of this strategy. This research strategy is micro-­historical ­because the unit of analy­sis is legislators rather than more aggregated units, such as countries or parties. The approach is “institutional” b ­ ecause it pays close attention to the electoral and parliamentary rules that affect the strategies of politicians. This methodological choice pre­sents us with numerous advantages relative to alternative empirical strategies. The first is that we mea­sure po­liti­cal decisions at the appropriate level. Therefore, we avoid fallacies of aggregation that infer individual be­hav­ior from the observation of larger units. This empirical strategy also allows me to take advantage of the nested nature of the data and examine how partisan affiliations, on the one hand, and the individual attributes of candidates, on the other hand, help explain support for or opposition to electoral reforms. At the same time, the analy­sis allows me to consider the relative importance of economic conditions and electoral considerations in terms of their effect on legislators’ decision making. This enables me to provide a richer account of ­these po­liti­cal decisions as compared to modernization theories of democ­ratization or approaches that invoke redistributive threats. Employing this strategy and fulfilling its empirical promise required what Jeffrey Williamson (2006) has called “empirical heavy-­lifting.” The empirical analy­sis of the book entailed collecting and digitizing multiple economic censuses, electoral data, and roll-­call votes. To understand the motivations of actors considering vari­ous reforms, I draw on an analy­sis of additional archival evidence that includes ministerial sources and deliberations in parliamentary commissions and on the floor of the parliaments at

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 21

the time. Altogether, the new quantitative and historical sources allow us to characterize the individual and partisan motivations that facilitated the adoption of demo­cratizing reforms and clarify why the composition of the electoral co­ali­tions supporting reforms varied across countries and reform dimensions. A Road Map of the Book

In chapter 2, I formulate the theoretical hypotheses of the study. I begin by describing the repertoire of electoral irregularities that occur during campaigns, at the moment of the vote, and at the time ballots are counted. I discuss both the heterogeneity in the types of resources that are politicized—­which include both private resources financed by the individual candidates and public resources—­and the variety of brokers or intermediaries deployed by candidates. Next, I unpack the considerations of legislators about the desirability of electoral reforms and look at both the relative advantages associated with access to resources that can be politicized during elections and the economic and electoral costs associated with dif­fer­ent illicit strategies. The remaining part of the chapter examines the partisan po­liti­cal landscape in the dif­fer­ent countries included in this study. I describe both the resource asymmetries across parties and the consequences of economic changes and elite splits for the formation of electoral majorities surrounding electoral reforms. Chapter 3 examines efforts to adopt electoral reforms that limit the ability of candidates to politicize state resources during campaigns. I examine proposals to adopt such electoral reforms in France and Germany, two countries where this irregularity was pervasive, ­going back to the authoritarian period. In the initial stages of reforms, legislators affiliated with the governmental majorities opposed t­ hese proposals and used their sizable electoral majorities in the parliaments of the period to block the introduction of such reforms. The chapter examines the consequences of an elite split in the dominant co­ali­tion in both France and Germany for the introduction of ­these electoral reforms. I document how considerations about the rising electoral costs of ­these illicit strategies contributed to a partisan realignment during the period when reforms ­were considered by parliaments. In France, this partisan realignment facilitated the adoption of reforms that imposed stronger limits on the use of state resources during campaigns. ­These reforms ­were ­adopted immediately prior to World War I. In Imperial Germany, the elite split also contributed to the formation of a new po­liti­cal

22 C H A P TER 1

majority in support of stronger sanctions for the use of state resources during campaigns. While Imperial Germany did not adopt new legislation that ­limited the electoral use of state resources, I document impor­tant changes in electoral jurisprudence ­adopted during the de­cade before World War I that imposed stronger sanctions on this electoral malfeasance. In chapter 4, I examine electoral reforms that attempted to limit the use of vote-­buying or treating. Candidates’ use of ­these strategies differed significantly across the countries examined in this book. In Britain, candidates from both po­liti­cal parties engaged in vote-­buying. This development contributed to the dramatic increase in candidates’ private campaign expenditures during the second de­cade of the nineteenth ­century. In France, by contrast, vote-­buying was a residual electoral irregularity and used primarily by Monarchists and other non-­Republican candidates who ­were unable to access state resources. Vote-­buying was a relatively underutilized campaign strategy in German elections as well. ­Here, the availability of state resources that could be politicized during elections and the much more severe punishment imposed on candidates who offered money or goods strongly reduced electoral incentives to use this strategy. The chapter examines po­liti­cal initiatives to introduce electoral reforms to limit vote-­buying in France and Britain. Since vote-­buying was an electoral irregularity financed by private resources, in both countries, the initial cleavage line over the introduction of ­these reforms pitted resource-­endowed against resource-­constrained candidates. Individual wealth remained an impor­tant predictor of a candidate’s opposition to reforms limiting vote-­buying. With re­spect to the partisan cleavages that developed during this electoral reform, I show that parties with a relative resource disadvantage supported electoral reforms to limit vote-­buying, while parties that benefited from a relative electoral advantage opposed ­these reforms. This stands in in­ter­est­ing contrast to the co­ali­tional dynamics over the reforms that attempted to reduce the ability of candidates to politicize state resources discussed in chapter 3. While Monarchists and other Conservative legislators spearheaded po­liti­cal initiatives to reduce the candidature officielle, ­these legislators strongly opposed efforts to reform vote-­buying. While Republican candidates opposed reforms to limit the use of state resources during campaigns, they supported reforms to limit the use of vote-­buying. Chapter 5 turns to electoral reforms that attempted to protect the secrecy of the vote. While electoral laws in most countries included commitments to secret voting, this remained only a distant aspiration rather than a po­liti­ cal real­ity. Imperfections in voting technology undermined voting secrecy.

E lec to r a l Co r r u p ti o n i n Fi r s t-Wav e D e m o c r ac i e s 23

The large variability in the shapes or colors of ballots made it pos­si­ble for brokers stationed in the immediate vicinity of the urn to ascertain voters’ choices. The chapter examines the adoption of electoral reforms to uphold voter secrecy in France and Germany. The proposed solutions that sought to remedy this situation included the introduction of ballot envelopes and isolating spaces. Such proposals for electoral reforms generated remarkably strong opposition. I examine the adoption of reforms to protect voter secrecy in France and Germany using a combination of roll-­call votes and an analy­sis of parliamentary deliberations. I document the importance of elite splits in facilitating the formation of parliamentary majorities favoring reforms. Both the breakdown of the Kartellbündnis in Germany and the split between centrist Republicans and Radicals in France contributed to the formation of a parliamentary majority that endorsed electoral reforms. In chapter 6, I examine reforms to combat electoral fraud. When counting ballots, presidents of polling places could undermine the results of an election by adding ballots to the urn, by subtracting ballots from it, or by replacing all the ballots. The chapter begins by documenting the incidence of fraud in nineteenth-­century elections, using reports of electoral irregularities. Fraud was a relatively rare electoral occurrence, which resulted from the decentralized initiatives of voting bureau presidents. Efforts to coordinate fraud ­were rarely done so across districts. During nineteenth-­century elections, we do not find cases of orchestrated fraud coordinated by regional or national party organ­izations. In addressing the issue of electoral fraud, nineteenth-­century reformers considered two pos­si­ble solutions. The first was to subject such offenses to penal law and thus increase the relative punishment for acts of fraud relative to other campaign irregularities. Germany pursued this approach and imposed harsh punishments for fraud in its penal code. The second solution was to allow candidates’ representatives to supervise ballot counting. This policy was a­ dopted by Britain in 1872 as part of the Ballot Act and in Belgium in 1877. In both Belgium and Britain, the low-­resource asymmetries among the two large parties facilitated the formation of an electoral majority in support of ­these reforms with relative ease. By contrast, in France, po­liti­cal negotiations around proposals to allow candidate representatives to supervise the counting of ballots lasted for over thirty years without yielding a decisive majority in ­favor of reforms. I document the formation of a po­liti­ cal majority to reform fraud in nineteenth-­century France, which brought together a co­ali­tion of extremes against the center. I show that both Monarchist and Socialist legislators who lacked connections to voting bureau

24 C H A P TER 1

presidents and thus lacked the opportunities to engage in electoral fraud ­were the driving force ­behind the adoption of ­these reforms. The concluding chapter discusses the implications of reforms a­ dopted during the first wave of democ­ratization for recent democracies. Bribery, intimidation, and fraud are recurrent prob­lems in many con­temporary democracies, where efforts to introduce electoral reforms have proven elusive and disappointing. I conclude the book by discussing the pos­si­ble po­liti­cal interventions that increase the electoral costs of illicit exchanges of politicians and create electoral alliances in support of reforms to limit vari­ous types of electoral malfeasance.

2 The Democ­ratization of Electoral Practices Electoral rules in many nineteenth-­century countries required each incoming parliament to confirm the results of elections that brought each successive cohort of politicians into office. In France, this pro­cess, which was known as the vérification des pouvoirs, was conducted at the beginning of each new parliamentary session by commissions that included fellow deputies who w ­ ere randomly assigned to investigate vari­ous races. In Imperial Germany, this pro­cess was conducted by an election-­verifying commission (Wahlprüfungskommission), whose composition reflected the distribution of power in the Reichstag. ­These investigations documented a variety of strategies by which candidates coerced, bribed, and intimidated voters, ­whether during campaigns, during voter mobilization on election day, or at the time of voting. Another range of irregularities occurred a­ fter the voting pro­cess concluded, when voting station presidents replaced some or all of the ballots in ballot boxes, known at the time as electoral urns. One recurrent reaction during parliamentary discussions of ­these irregularities was moral outrage. Many politicians viewed the detailed descriptions of malpractice in remote villages and large urban centers as demonstrating the hollowness of the promises made by electoral laws. Lamenting imperfections in voting technology that facilitated electoral coercion, Heinrich Rickert, a ­Free Liberal politician, remarked: “The freedom of the vote is . . . ​ the foundation of our constitutional life. If we do not provide sufficient guarantees for it, then our public life w ­ ill be based on hy­poc­risy, on pressure 25

26 C H A P TER 2

and on force, and ­those have never led to a good outcome” (StBer DtRt, 15 January 1896). A commission investigating electoral irregularities in France during the 1903 election reached similar conclusions: “One cannot say that universal suffrage is f­ ree. B ­ ecause voting is not entirely secret, it cannot be exercised with entire freedom. The voter, if he is employed, ­faces the pressure of employers, if he is a shop­keeper, he ­faces the pressure of customers, and if he is a state employee, he ­faces the pressure of the state” ( JORF ChamberDeb, 27 October 1904). Most of the time, however, this moral outrage was insufficient to spur legislative action in support of electoral reform. Such wrath and disapproval ­were short-­lived and often dissipated into apathy. Even politicians who expressed indignation about the electoral practices that ­were common at the time ­were quick to backpedal on ambitious proposals for reform ­because they ­were concerned about the impact of changing electoral rules on their po­liti­cal fate. By contrast, opponents of reform ­were quick to invoke myriad practical difficulties associated with changes to electoral rules. A French senator of the Third Republic commented on ­these strategies of delay as follows: “The order of this discussion is all too familiar to us. ­After making beautiful proclamations about po­liti­cal princi­ples, one makes them dis­appear in a variety of secondary objections. One examines with a microscope all the pos­si­ble abuses that can result from the introduction of reforms” ( JORF SenateDeb, 23 June 1905). Such legislative procrastination considerably delayed the introduction of reforms. In France, for example, the first proposal to reduce the incidence of electoral fraud was submitted to the parliament in 1877. It took, however, nearly four additional de­cades of negotiations u ­ ntil a new electoral law was a­ dopted. In Germany, proposals to better protect the secrecy of the vote ­were discussed for the first time in 1871. However, an electoral law mandating the use of ballot envelopes was ­adopted only in 1903. How did electoral majorities that supported reforms to limit malfeasance come about? What motivated legislators to endorse ­these reforms and overcome their initial re­sis­tance? Who supported and who opposed ­these reforms? Did the composition of the po­liti­cal co­ali­tion supporting electoral reforms vary across countries and across reform dimensions? If so, how can we understand this variation? The goal of this chapter is to develop the theoretical building blocks that ­will allow me to answer ­these questions. I begin this chapter by examining the most salient electoral irregularities pre­sent in nineteenth-­century elections, as well as the ways in which penal codes and electoral laws introduced during the first half of the ­century sanctioned such irregularities. In subsequent sections of this chapter, I examine

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 27

the most impor­tant ­factors that ­shaped individual legislators’ preferences for reforms. I argue that ­these preferences ­were affected by legislators’ access to the resources needed to produce par­tic­u­lar irregularities and by the economic and electoral costs incurred by legislators who engaged in dif­fer­ent forms of malfeasance. While access to resources creates the initial cleavage line over the desirability of electoral reforms, economic and po­liti­cal developments may increase the economic and electoral costs associated with illicit strategies and reduce the attractiveness of the status quo even for resource-­endowed politicians. The next section examines the po­liti­cal and economic conditions that affect the in­equality in access to po­liti­cal resources, on the one hand, and the electoral and economic costs associated with vari­ous forms of malfeasance, on the other. In countries with duopolistic po­liti­cal competition, the resource asymmetry among candidates is low. ­There, when resource-­constrained politicians push for electoral reforms, such reforms are likely to lead to cross-­party consensus. The negotiation of electoral reforms is likely to be dif­fer­ent in countries where one party dominates the po­liti­cal landscape. ­Here, enduring po­liti­cal asymmetries in the distribution of po­liti­cal resources reduce the willingness of politicians affiliated with the incumbent party to forgo their initial resource advantage. I conjecture that in ­these contexts, elite splits play an impor­tant role in breaking the po­liti­cal stalemate over reforms. Th ­ ese splits increase the electoral costs associated with illicit strategies and facilitate the formation of a parliamentary majority in f­ avor of reforms. Using evidence from Third Republic France and Imperial Germany, I show that elite splits—­such as the Republican-­Radical split in France and the breakdown of the Kartellbündnis between Conservatives and National Liberals in Germany—­modified the calculations of resource-­endowed legislators about the desirability of the status quo and facilitated the formation of a po­liti­cal majority in support of reforms. In the final part of the chapter, I show how this explanation illuminates differences in the composition of the electoral co­ali­tions across countries and across reform dimensions. The Universe of Electoral Irregularities

Electoral practices in nineteenth-­century Eu­rope encompassed a variety of forms of malfeasance. Politicians appealed to voters by offering gifts, money, food, or access to government benefits in exchange for their vote. Candidates also relied on a broad variety of brokers to mediate their relationship

28 C H A P TER 2

to voters, such as mayors, local government officials, and intermediaries employed briefly before a campaign began. During campaigns, ­these brokers distributed vari­ous goods to voters and engaged in sustained efforts to prevent opponents from campaigning. On election day, t­ hese brokers mobilized voters, brought them to the voting place, and actively attempted to pierce the secrecy of the vote. A ­ fter the ballots w ­ ere cast, candidates engaged in efforts to modify the result of the popu­lar ­will by exchanging the ballots cast by voters with dif­fer­ent ballots. VOTE- ­B UYING

In vote-­buying exchanges, candidates or intermediaries working on their behalf offered voters money or in-­kind goods in exchange for their electoral support. In nineteenth-­century elections, vote-­buying was an extremely colorful practice, and candidates offered voters a mix of highly heterogeneous goods. A complaint about electoral irregularities in Chambéry during the 1910 elections reported that candidates had offered the following goods during the campaigns: “One voter had been offered five francs, the other two francs, a third a felt hat, a fourth a vest, a fifth a handkerchief, a sixth firewood, and the seventh a dinner” ( JORF ChamberDeb, 9 June 1910). The goods or bribes offered to voters ­were financed with candidates’ private resources. This distinguishes vote-­buying exchanges from other electoral irregularities examined in this study that politicize state resources. Describing this strategy’s private financing, H. J. Hanham notes that “the amount of money spent by candidates depended more on the length of the purse than the needs of the constituency” (1959: 54). Wealthier candidates thus had an easier time financing vote-­buying exchanges compared to less affluent candidates. One example of such a candidate was W. H. Smith, a “very wealthy man whose Westminster constituency lived on his doorstep and expected him to give dinners as well as donations” (ibid.). W. H. Leatham, an affluent candidate competing in the Wakefield election of 1870, declared to the commission investigating bribery that “if he had a sackful of money, he could have disposed of it” ­because “the appetite of Wakefield to be bribed” was high (Inverness Courier, 26 July 1870). In France, wealthy candidates who engaged in vote-­buying exchanges with voters ­were known as candidatures d’argent, the moneyed candidacy. One example of this type of candidate was Achille Fould, a deputy elected in Tarbes in 1903. Fould, who campaigned by lavishly displaying his wealth,

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 29

was described by contemporaries as the “arch-­millionaire.” Election reports described him as “campaigning with an open wallet” and distributing three hundred francs during each encounter with voters ( JORF ChamberDeb, 24 October 1903). Raphaël Louis Bischoffsheim, another wealthy candidate who was elected in the first electoral district of Nice in 1890, campaigned by throwing pieces of twenty, ten, and five francs at crowds that followed him during his visits to the constituency (ibid., 21 January 1890). A second type of vote-­buying exchange was known as “treating.” In contrast to vote-­buying, treating involved offers of goods that ­were not directly targeted to specific voters, primarily food and drink, and occurred most commonly in inns, restaurants, and public ­houses. Treating was a mainstay of British elections; the 1874 election was described as a “torrent of gin and beer.” Masses of drunken voters moving from place to place ­were one emblematic image of nineteenth-­century British elections. Charles Seymour vividly describes ­these voters: “from the moment the canvass began, in many of the boroughs, the poor spent most their time wandering from one public ­house to another in search of refreshment” (1915: 400). In France, such electoral events took the form of banquets, known as rastells ( JORF ChamberDeb, 26 November 1889, 6 June 1902; Lefranc 1868). On t­ hese occasions, candidates and their electoral agents offered abundant ­free drinks, wine, and bistrouille, a mixture of eau-­de-­vie and coffee. A candidate competing in Montreuil during the 1902 election offered voters bistrouille in over one hundred forty electoral meetings ( JORF ChamberDeb, 30 June 1902). Wine was also offered abundantly during ­these banquets. Describing the campaign occurring in a settlement near Civray (Vienne), observers noted that “an unlimited flow of wine” gushed from barrels stationed in the center of the locality, in immediate proximity to the voting place (ibid., 29 June 1898). The lit­er­a­ture examining vote-­buying in con­temporary elections has largely focused on quid pro quo exchanges between brokers and voters that presuppose electoral reciprocity (Hicken 2011). Treating defies this classification. The use of ­these exchanges in nineteenth-­century elections and the large amounts of resources devoted to t­ hese strategies raise questions about the electoral advantage gained from such exchanges. Jean-­Paul Charnay considers treating as a “childish” (puerile) attempt to influence voters (1965: 288). Contemporaries disagreed with this characterization of treating. They considered it a rational strategy, giving candidates the opportunity to signal their electoral strength and competitiveness. Commenting on the

30 C H A P TER 2

use of treating in Hexham, a county in North­umberland, during the 1892 election, a judge remarked: They are perfectly ready to vote for the man who is popu­lar, and if by reason of treats and picnics you can produce a general feeling that the candidate is a good fellow, and that he is willing to give a poor man a supper or treat or entertainment of this kind, and if that idea gets generally spread over the division an enormous amount of popularity is produced by it as against another association which does not resort to that sort of ­thing. (O’Malley and Hardcastle 1893: 145) Treating was, thus, a relatively sophisticated strategy used by candidates to signal personal competency and to improve their electoral viability (Kramon 2016; Mares and Young 2018a). THE USE OF STATE RESOURCES DURING CAMPAIGNS

The second electoral irregularity I ­will examine in this study involves the use of state resources during campaigns. In France, electoral strategies that politicized state resources ­were known as the candidature officielle (Pierre 1893; Voilliot 2005). While the term originated during the Second Empire, it was also used to characterize Third Republic electoral practices, when the politicization of state resources continued unabated (Pilenco 1928). In Germany, this electoral irregularity was known as amtliche Wahlbeeinflussung, a term that can be translated as the use of official resources to influence election outcomes (Hatschek 1920: 202). Throughout the book, I ­will use terms such as “state ­favors,” “policy ­favors,” and “state coercion” to describe electoral strategies that politicize state resources during campaigns. Candidates could politicize resources at vari­ous levels of the administration to shape their relationship to voters. Alphonse Gourd, a Republican legislator from Rhône, described the variety of state employees that w ­ ere mobilized in French elections as follows: “Prefects or deputy prefects, in a tour of review of the district or on another occasion, would recommend ­either to mayors or to influential voters to convince ­others to vote for a par­tic­ul­ ar candidate. Naturally and ordinarily, this candidate was the candidate of the government” ( JORF ChamberDeb, 29 March 1902). In German national elections, the hierarchy of brokers also originated at the highest levels of the national government. The Prus­sian Ministry of Interior played a significant role in coordinating the electoral responsibilities of lower-­level employees of the Prus­sian administration (Arsenscheck 2003; Kühne 1994,

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 31

1995). As “­these employees played a critical role in affecting ­whether the armies of low-­level employees could be set in motion,” Thomas Kühne, a historian of German electoral practices, refers to them as “multiplicators of voters in a district” (1994: 61). In France, the central pillar of the electoral machinery of the incumbent party was the prefect. At the onset of campaigns, prefects intervened by announcing that candidates held impor­tant connections to vari­ous ministries in Paris, thus signaling the promise of largesse in the case of a governmental victory. Discussing the importance of prefects in French elections, Charles Seymour and Donald Frary remarked that they ­were viewed as “dispensers of an infinite number of ­favors. Their power and their credit with the central government w ­ ere apparently limitless. ­Every village wanted its road, ­every city was seeking a spur of railway or a canal” (1918: 358). Contemporaries noted that voters “­were promised paradise on earth, a paradise which was financed by contributions of taxpayers in exchange for a correct vote” ( JORF ChamberDeb, 17 March 1910). In addition to announcing the immediate po­liti­cal advantages that could be enjoyed by voters who made the correct choice on their ballot, prefects played an impor­tant role in mobilizing brokers at lower levels of the administrative hierarchy. Th ­ ese included deputy prefects but also, most significantly, mayors. Mayors, in turn, coordinated the electoral activities of a variety of other local state employees. Th ­ ese local brokers included “teachers [who] could become agents of the most active electoral propaganda, postal employees, local policemen, road menders [cantônniers], secretaries and employees of the city halls” ( JORF ChamberDeb, 29 March 1902). State employees ­were also used extensively in German national elections. A crucial player of the Prus­sian bureaucracy who was repeatedly used as electoral broker was the Landrat, an official responsible for tax collection, the assessment of tax liabilities, and military conscription. The list of state officials deployed in Prus­sian elections included postal workers, school inspectors, policemen, village mayors, and nightwatchmen. Candidates’ policy promises of f­ avors and governmental transfers in Germany w ­ ere much more subdued than in France. Consider the 1884 election in Elbing, a district in East Prus­sia that reportedly displayed “a use of the official candidacy similar to the situation encountered in neighboring France.” ­Here, Landräte ­were reported to have provided significant support to a Conservative candidate by participating in electoral meetings, accompanying the candidate on trips in the district, and distributing Conservative ballots to voters (StBer DtRt 1884, encl. no. 180). Yet even in such a district—­depicted as one of the

32 C H A P TER 2

most problematic cases of amtliche Wahlbeeinflussung—we do not find evidence of the governmental largesse one encounters in French elections. By contrast, the modal electoral strategy used by employees of the state in German elections was repression and intimidation. ­Here, candidates enlisted support from state employees to deter opposition candidates from reaching voters. Policemen, Landräte, or other state employees routinely intervened during campaigns by intimidating and often arresting ballot con­ vey­ors and other brokers employed by opposition candidates. An event that unfolded in Klein Roris, a locality in Potsdam’s tenth district, during the 1891 election illustrates the repressive use of the state apparatus. H ­ ere the mayor was reported to have “forcefully ripped the ballots and copies of the Berliner Volksblatt from the hands of the Social Demo­cratic ballot con­ vey­or” (StBer DtRt 1892, encl. no. 165). In Neuhoff bei Zossen, a neighboring locality, policemen on ­horse­back attacked ballot con­vey­ors and drove them out of the area (ibid., encl. no. 182). In Potsdam’s tenth district, the ballot con­vey­ors “­were punched, thrown to the ground, stomped upon and slandered” (ibid., encl. no. 165). Altogether, ­these examples demonstrate the more extensive use of electoral coercion in German national elections compared to ­those of neighboring countries. Economic Coercion and Violations of Voting Secrecy

Vote-­buying and the use of state resources ­were illicit campaign strategies. Voting itself created new opportunities for electoral malfeasance. At the time of voting, significant imperfections in voting technology allowed candidates and their brokers, stationed by the electoral urns, to observe voters’ choices. At the same time, t­ hese brokers developed sophisticated ways of recording voters’ choices. Following the vote, employers sanctioned ­those who had voted “incorrectly” through reduction of wages, worsening employment conditions, or even layoffs (Mares 2015). Most nineteenth-­century electoral laws mandated uniform ballots. In practice, however, this represented an aspiration rather than a po­liti­cal real­ity. Politicians manipulated the shape, color, and size of ballots to track voters’ electoral choices. In France, electoral reports noted the presence of ballots printed on papers of vari­ous shades of white, on pieces of newspaper, or on transparent paper ( JORF ChamberDeb, 16 February 1882, 27 July 1882, 19 June 1890, 25 February 1899, 20 June 1898, 13 June 1910, 28 June 1910). In Germany, the use of ballots of dif­fer­ent shapes and colors

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 33

was also widely documented. In a discussion in the Reichstag, Ignaz Auer, a Social Demo­cratic legislator, described ­these irregularities as follows: We know that t­ here are ballots as large as half of a sheet of paper and also ballots the size of stamps; we know that especially in the western industrial areas, candidates formulate clear indications to distribute at specific times ballots of par­tic­u­lar sizes and shapes that are made out of paper with dif­fer­ent levels of thickness. We have seen ballots printed on cardboard paper and ballots printed on the thinnest transparent paper. (StBer DtRt, 1 February 1899) The design of ballot boxes, known as electoral urns, also contributed to fraud during the voting pro­cess. French electoral law required ballot boxes to be closed by two dissimilar locks (Article 41 of law of 1849; Article 22 of 1852 regulatory decree). In practice, out of administrative con­ve­nience, vari­ous objects w ­ ere used to collect ballots, including crystal bowls, soup tureens, and vases covered by pieces of paper ( JORF ChamberDeb, 20 June 1898; Charnay 1965: 598). German electoral law did not include specific regulations for the design of ballot boxes, leaving such choices to the discretion of local election officials. Con­temporary accounts documented the wide variety of objects used as ballot boxes. Th ­ ese included “cigar boxes, drawers, suitcases, hat boxes, cooking pots, earthen bowls, beer mugs, plates and washtubs” (Siegfried 1903). Small voting urns ­were particularly conducive to electoral fraud. As ballots would lie flat, one on top of each other, it was extremely easy to identify voters’ choices (Fairbairn 1997: 818). ­These imperfections in voting technology created opportunities for candidates to monitor voters’ choices. Brokers placed in the immediate vicinity of electoral urns could observe the shape and color of the ballot inserted into the urn. ­These brokers could also exert last-­minute pressure on voters to exchange their ballots before casting their vote. Given this opportunity to observe and modify voters’ choices, candidates selected brokers who could impose harsh post-­electoral sanctions on voters for making the wrong electoral choice. Candidates’ preferred brokers tasked with monitoring the vote ­were rural or urban employers. ­Because they controlled the employment and wages of voters, employers could impose the harshest post-­electoral sanctions. In nineteenth-­century France and Germany, employers ­were often stationed in the vicinity of ballot boxes during voting. In a debate in the German Reichstag, Heinrich Rickert, a F ­ ree Liberal legislator, presented the following account of economic intimidation: “employers lead voters

34 C H A P TER 2

to voting places like animals to slaughter. In front of the door of the voting precinct some supervisor or inspector pushes a ballot in the hands of the worker. As [the worker] does not want to lose his job, he is in no position to vote according to his conviction” (StBer DtRt, 15 May 1895). In France, employers’ control of voters’ choices using ballots of dif­fer­ent shapes was reported as late as 1914. In Lille, employers distributed ballots that w ­ ere folded in dif­fer­ent ways: “Some w ­ ere folded in the shape of a triangle, o ­ thers in the shape of a diamond, a hat or an accordion. Naturally, they did not hesitate to inform t­ hese voters that they would keep an eye on them when the votes ­were counted and that it was imperative to find ­these par­tic­u­ lar ballots” ( JORF ChamberDeb, 30 June 1914). Observing voters’ choices gave employers the opportunity to punish their employees a­ fter the election with layoffs or the loss of discretionary social policy benefits provided by the com­pany. Following the 1884 election in Arnsberg’s fifth district, voters who refused to take the ballot handed to them by the com­pany’s overseer, who was stationed in the vicinity of the urns, lost their com­pany housing (StBer DtRt 1884, encl. no. 57). In an article “Who Wants to Hear, He Can Hear!” Egmon Tielsch, a porcelain manufacturer, expressed his disappointment at the victory of the ­Free Liberal candidate in his constituency and threatened to retaliate with layoffs. “My workers! The election to the Reichstag on 21 February has demonstrated that you do not follow my admonitions and that you are supporting ­Free Liberal ­causes, which cannot represent your true interests. . . . ​ I ­will be reluctant to employ workers who are opposed to the noble aspirations of our Emperor” (StBer DtRt 1887, encl. no. 105). Thus, layoffs remained the harshest form of post-­electoral punishment imposed by employers. ELECTORAL FRAUD

A final set of irregularities I consider in the book are ­those that occur when ballots are counted. Penal codes and electoral laws of the period referred to ­these irregularities as fraud. The term “fraud” encompasses a variety of practices, including the falsification of ballots, the subtraction or addition of ballots when they ­were counted, and the modification of the name of a candidate on a ballot (de Kéguelin 1904). The Napoleonic Penal Code of 1810 introduced strong punishments for fraud, sanctioning it with prison, monetary fines, and a temporary loss of po­liti­cal rights. Electoral laws ­adopted during the first half of the nineteenth ­century added protections against fraud, which complemented the penal sanctions at the time. Th ­ ese included strong commitments to the publicity

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 35

of elections, guaranteeing that both electoral operations and ballot counting could be observed by all t­ hose pre­sent at the voting place. Electoral jurisprudence followed up on t­ hese commitments. Elections could be invalidated if the ballot counting took place in an alcove, if the ­tables on which ballots ­were found w ­ ere too close to a wall, or if the distance between t­ ables holding ballots was too large (Charnay 1965: 620). ­These ­legal provisions ­were an effective deterrent in reducing the incidence of fraud; throughout the nineteenth ­century, other electoral irregularities ­were much more common by comparison. In addition, the strong penalties for fraud ­were effective in discouraging this be­hav­ior. Electoral fraud occurred on a very small scale that could often be attributed to the cunning and resourcefulness of individual mayors. ­There was very ­little coordination across districts in orchestrating this irregularity. This distinguishes nineteenth-­century po­liti­cal developments from the relatively common occurrence of this phenomenon in con­temporary autocracies (Magaloni 2006; Simpser 2013). The coordination of fraud across constituencies, regions, and, sometimes, even an entire country was remarkably absent in nineteenth-­century elections. To illustrate the l­imited scope of fraud, let us consider a few examples discussed in electoral reports at the time. A summary of French court investigations noted that the simplest way to commit electoral fraud was to count ballots ­behind closed doors (Conseil d’État 1909: 1257). Defying provisions about the publicity of elections, election officials sometimes counted ballots in darkness ( JORF ChamberDeb, 20 June 1910). During the 1914 election in Nîmes, the mayor required all ­those pre­sent to leave the room before ballots ­were counted in order “not to disturb public order” (ibid., 12 October 1914). On other occasions, the presidents of voting places chose to engineer disorder to justify the evacuation of voters. Consider the following incident, reported in Maubec during the 1878 elections, depicting the strategies used by a mayor in order to justify the expulsion of voters from the voting place before the ballots could be counted: The mayor invites the policemen and guards inside the voting place and then ­orders in a loud voice “Move around.” The policemen proceed by making voters execute an “electoral round” around the voting ­table. Following the caprices of the mayor or the gendarmes, the round speeds up or slows down. ­There is no revolt against this order. Move along, move along. ­After a while all ­these ­people clash, crush, murmur, and cry. At that point, they are all forced out of the voting place and one becomes

36 C H A P TER 2

the master of the urns. The winner can dispose of the contents of the urns as he desires. (ibid., 2 December 1878) When ballots ­were counted, voting station presidents could modify the outcome of the vote by destroying, substituting, or adding ballots to the urn. In France, ­these strategies ­were colloquially known as garnir l’urne (garnishing the urn). Presidents of voting places could invalidate ballots by staining them. During the 1890 election in Hérault, the president of the electoral bureau put a batch of ballots that did not support his party to the side. Ignoring the presence of a witness in the voting area, the mayor of this locality quietly stated, “now we w ­ ill go ahead and stain them” ( JORF ChamberDeb, 4 February 1890). In other cases, election officials tore apart ballots of the undesirable candidate (ibid., 21 October 1902) or burned them (ibid., 16 November 1881, 17 June 1890). ­Table 2.1 pre­sents the dif­fer­ent forms of electoral malfeasance I ­w ill examine in this study. First, I distinguish among electoral irregularities that occur during campaigns, on the one hand, and irregularities that occur during the organ­ization of the voting pro­cess and the counting of ballots, on the other. The second distinction introduced in this ­table concerns the types of resources deployed by candidates for electoral malfeasance. H ­ ere I w ­ ill distinguish among illicit electoral strategies that involve state resources and ­those involving primarily private economic resources. Th ­ ese strategies differ both in the resources deployed by candidates and in their visibility to voters. I conjecture that ­these differences w ­ ill shape legislators’ views about the desirability of electoral reforms.

The Demand for Electoral Reforms: A Candidate-­Level Perspective

The main goal of this chapter is to characterize the electoral co­ali­tions that supported vari­ous reforms to end electoral malfeasance. In this section, I provide the building blocks of this analy­sis. I examine the trade-­offs encountered by legislators faced with a proposal for electoral reform. I seek to identify the most impor­tant ­factors that affected legislators’ preferences about the desirability of changes to the status quo and examine the conditions ­under which legislators ­were willing to endorse reforms. This analy­sis provides the necessary microfoundations that allow me to describe the composition of legislative co­ali­tions that supported electoral reforms and examine how ­these co­ali­tions varied across reform dimensions and countries.

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 37

­TABLE 2.1.

A typology of electoral irregularities Timing of irregularity

Resources used for electoral malfeasance

During campaigns

On election day

Private

Vote-­buying

Breaches of voting secrecy and electoral coercion

Public

State f­ avors and pressure

Fraud

In deciding ­whether to support electoral reforms, legislators compare the status quo with a post-­reform state that imposes higher costs on some illicit strategies in order to reduce opportunities to carry them out. From a legislator’s perspective, the first ­factor affecting the attractiveness of the status quo relative to this alternative is the availability of the resources necessary for the production of the illicit strategy ­under consideration. All ­things equal, legislators who enjoy ample access to po­liti­cal resources are likely to oppose changes to the status quo. By contrast, resource-­constrained politicians are likely to f­ avor the adoption of electoral reforms. For the latter, electoral reforms that increase the cost of par­tic­u­lar irregularities are advantageous ­because they constrain the f­ uture electoral be­hav­ior of other candidates who may benefit from a resource advantage. As the previous section documented, the types of resources vary across dif­fer­ent forms of electoral malfeasance. Vote-­buying exchanges ­were financed predominantly with politicians’ personal resources. In this case, private wealth is a likely source of cleavage among politicians considering the desirability of reforms. It follows that wealthy candidates are likely to oppose reforms that impose higher sanctions on vote-­buying or that limit the use of campaign expenditures. By contrast, less affluent candidates are likely to demand electoral reforms. In contrast to vote-­buying, the main resource in electoral strategies that deploy state resources during campaigns—in the form of policy transfers or access to brokers—is the connection to the central and local administration. Candidates with power­f ul connections to the government are likely to oppose proposals to limit the politicization of state resources. By contrast, candidates lacking such connections w ­ ill regard such reforms favorably. In the case of electoral irregularities that involve economic coercion, candidates’ access to employers determines the initial cleavage line among legislators. Landowners or industrial employers who have a large workforce and the opportunity to control their vote are likely to oppose reforms

38 C H A P TER 2

protecting voter secrecy. Candidates with other occupational backgrounds lack such resource advantages and are likely to support the adoption of such reforms. In the case of po­liti­cal irregularities that deploy state resources, partisan competition affects the magnitude of this initial resource asymmetry. In countries with duopolistic competition that experience robust po­liti­cal alternation, resource asymmetries are small. Even in opposition, legislators from parties with a recent governmental experience share similar expectations about their f­ uture ability to access state resources. This is likely to facilitate compromise over the adoption of electoral reforms. The situation is dif­fer­ent in cases where po­liti­cal alternation is low and where one hegemonic party enjoys uninterrupted access to state resources. This situation is likely to strengthen the incentive for resource-­constrained politicians to demand electoral reforms, while entrenching the opposition of resource-­endowed politicians to any change to the status quo. Considerations about resource availability are, however, only one ­factor that enters into legislators’ calculations about the desirability of electoral reforms. Even resource-­endowed politicians are not entirely unconstrained in using existing resources to engage in the production of electoral irregularities. If ­these constraints are high, politicians may find the status quo less desirable and may choose to forgo their initial resource advantage. In figure 2.1, I pre­sent the decision-­making pro­cess of a legislator who is evaluating the attractiveness of the status quo relative to a proposal for electoral reform. I distinguish between two types of constraints that impose dif­fer­ent costs for the politician who engages in an illicit strategy. A politician incurs electoral costs if the use of an illicit strategy brings about a decrease in his vote share. By contrast, a politician incurs economic costs if changing economic conditions in a district limit his ability to engage in an illicit strategy. Both electoral and economic constraints are likely to reduce the attractiveness of the status quo even for resource-­endowed legislators and increase their willingness to support an electoral reform. Th ­ ese changes in the preferences of this pivotal group of legislators may give rise to a new po­liti­cal majority that ­favors reforms. This new po­liti­cal majority is likely to include politicians with an initial resource advantage who now encounter high electoral or economic costs associated with the illicit strategy in addition to resource-­constrained politicians. The key goal of the theory is to examine the f­ actors that are conducive to this co­ali­tional realignment. The use of illicit strategies is not entirely costless for politicians who enjoy access to resources that allow them to engage in vari­ous forms of

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 39

Can the legislator access political resources allowing him to employ illicit electoral strategies?

No Supports electoral reform

Yes Opposes electoral reform

Does the use of illicit strategies generate electoral costs for the politician?

Yes Supports electoral reform

No Opposes electoral reform

Do economic conditions in the constituency impose economic constraints for the use of an illicit strategy?

Yes Supports electoral reform

FIGURE 2.1.

No Opposes electoral reform

The determinants of demand for electoral reforms.

malfeasance. A candidate incurs electoral costs if the use of an illicit strategy reduces rather than increases his vote share. A decline in votes can be a result of voters deciding to sanction candidates who engage in vari­ous forms of malfeasance or be the response of other competitors once they observe the use of electoral malfeasance. Electoral costs play a decisive role in shaping

40 C H A P TER 2

the views about the desirability of the status quo of legislators with access to the resources necessary to engage in electoral malfeasance. If such electoral costs are high, ­these legislators may forgo their resource advantage and instead endorse proposals to change the status quo. Consider first how a candidate who engages in electoral malfeasance may incur electoral costs as a result of decisions made by other competitors during the race. One electoral context in which competitors may react to the use of illicit electoral strategies and sanction their use is a majoritarian electoral system with runoffs. This electoral system was common among many first-­wave democracies. In this system, the path to victory for a candidate often necessitated coordination during runoffs with other candidates he competed against in the first round. Yet first-­round competitors could respond to a candidate who used dif­fer­ent forms of malfeasance during the first round by endorsing a dif­fer­ent candidate during runoffs. In this case, the loss of co­ali­ tion partners during runoffs could result in electoral defeat for the candidate who had engaged in electoral malfeasance. I conjecture that ­these strategic considerations about coordination in runoffs w ­ ere likely to reduce the attractiveness of the status quo for resource-­endowed legislators. A candidate who uses dif­fer­ent forms of malfeasance may incur electoral costs if voters sanction such strategies. In evaluating dif­fer­ent candidates, voters consider the immediate benefits of monetary transfers, policy benefits promised by the candidate, and the consistency and coherence of the candidate’s electoral commitment. Recent lit­er­a­ture in po­liti­cal science has analyzed the importance of ­these dif­fer­ent ­factors in explaining voters’ evaluation of dif­fer­ent illicit strategies using a combination of survey data and survey-­based experiments (Weitz-­Shapiro 2012, 2014; Weitz-­Shapiro and Winters 2013; Stokes et al. 2015; Kramon 2016; Mares and Young 2018a). This lit­er­a­ture has identified significant individual-­level differences across voters in their evaluation of clientelistic strategies—­such as vote-­buying or policy ­favors—­w ith lower-­income voters less likely to sanction ­these strategies as compared to higher-­income voters (Weitz-­Shapiro 2014). At the same time, this lit­er­a­ture has found differences in the willingness of voters to sanction dif­fer­ent types forms of malfeasance. Generally, strategies involving electoral coercion are sanctioned more strongly than strategies based on positive inducements (Mares and Visconti 2020; Mares and Young 2018a). A related result is that voters evaluate the consistency of a candidate’s profile. Vicki Randall (1988) and Herbert Kitschelt (2000: 854) have noted that voters are likely to punish a candidate who seeks to combine both

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 41

programmatic appeals and illicit strategies as “schizophrenic.” The use of illicit strategies—­such as bribery or coercion—by a candidate competing on the basis of programmatic promises is likely to undercut, in the eyes of voters, the credibility of the programmatic appeal. As voters perceive the inconsistency between ­these two strategies, they are likely to sanction a candidate combining programmatic and non-programmatic appeals. This lit­er­a­ture highlights the fact that the electoral costs incurred by candidates who deploy non-­programmatic promises are not negligible. I ­will not be able to test hypotheses about the electoral costs of vari­ous illicit strategies using individual-­level data, as survey data is unavailable for the period. However, I w ­ ill examine the implications of this lit­er­a­ture about the importance of consistency in a candidate’s profile for the electoral costs associated with vari­ous illicit strategies. The direct implication of this lit­ er­a­ture is that voters are more likely to sanction politicians who combine programmatic and non-­programmatic promises relative to politicians who compete on the basis of programmatic promises alone. In other words, politicians who compete on the basis of programmatic promises incur higher electoral costs as compared to politicians who do not include such programmatic appeals. This suggests that the growing programmatic differentiation of the period examined in this study contributed to an increase in electoral costs associated with dif­fer­ent forms of malfeasance such as vote-­buying and intimidation. By increasing electoral costs, such programmatic differentiation reduced the attractiveness of the status quo of politicians competing on the basis of programmatic promises. ­These ­factors also facilitated the formation of an electoral majority that favored reforms even if a majority of legislators still enjoyed access to po­liti­cal resources. Resource-­endowed legislators who incur high electoral costs are, thus, the pivotal group who contribute to the formation of a majority that supports electoral reforms. Politicians incur electoral costs only when illicit electoral strategies are vis­i­ble to voters. By contrast, illicit strategies that occur ­after the ballots are counted, such as fraud, are likely to be less consequential for electoral costs. This difference in the visibility of illicit strategies provides an additional opportunity to test ­whether the relationship between the programmatic orientation of a candidate and support for electoral reforms is mediated by electoral costs. If electoral costs are the mediating variable, we should see no relationship between the programmatic orientation of a candidate and support for reforms limiting the incidence of fraud. However, we should see a positive relationship between programmatic orientation and

42 C H A P TER 2

electoral costs in the case of vis­i­ble electoral strategies, such as vote-­buying and intimidation. Resource-­endowed politicians may also experience economic constraints in using illicit strategies. A politician who campaigns by offering bribes to or treating voters is likely to find such strategies prohibitively costly in wealthy or populous constituencies, as the lit­er­a­ture on clientelism suggests. Commenting on the difficulty of bribing voters in the constituency of Yorkshire, the Duke of Wellington remarked that “no gentleman could bear the expense—­the m ­ iddle classes had it all to themselves” (Law 1881, 2:329). Similarly, a strategy of electoral coercion may be costly in constituencies where voters have many employment opportunities. In such environments, po­liti­cal pressure is potentially costly to the employer, as the employer may lose valuable employees. ­These economic constraints are likely to reduce the attractiveness of the status quo for resource-­endowed politicians and increase their willingness to support electoral reforms. In sum, a parsimonious account of legislators’ preferences with regard to electoral reforms needs to examine how considerations about access to resources explain the initial cleavage lines about the desirability of such electoral reforms. However, ­these initial cleavages between resource-­endowed and resource-­constrained politicians are not static. Economic and electoral costs may reduce the desirability of the status quo even among resource-­ endowed politicians and may increase po­liti­cal support for electoral reforms. The magnitude of ­these economic and electoral constraints affects the size and the po­liti­cal composition of the majority supporting electoral reforms. In the following section, I ­will shift perspectives from the micro to the macro level and examine how po­liti­cal and economic developments affect resource asymmetries and economic and electoral costs. The Po­liti­cal and Economic Environment

The configuration of the party system differs significantly among the countries included in this study. This, in turn, had impor­tant consequences for the distribution of resources required for electoral irregularities and for the formation of po­liti­cal majorities favoring reforms. Britain and Belgium experienced relatively robust competition among two parties, which resulted in significant alternation. In Belgium, following a period of party unity ­after in­de­pen­dence, Liberals and Conservatives alternated in power. Britain also experienced robust alternation during the period following the First Reform Act of 1831. Between 1831 and 1880, Conservatives appointed

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 43

eight prime ministers, while Whigs and, ­later, Liberals appointed six prime ministers. The situation was dif­fer­ent in France and Germany. While the party system remained relatively fragmented in both countries, po­liti­cal alternation was low. In Imperial Germany, the Conservative Party formed ­every government between 1870 and 1914. In France, the Republican Party also enjoyed a dominant position and formed the overwhelming majority of governments between 1877 and 1906. In ­these po­liti­cal environments, incumbent parties—­such as the German Conservatives and French Republicans—­ mounted a strong re­sis­tance to the adoption of electoral reforms. ­Here, splits among incumbent elites w ­ ere an impor­tant development that facilitated the formation of electoral majorities favoring reforms. The goal of this section is to clarify how t­ hese broader po­liti­cal configurations and the changing economic environment affect the types of cleavages over electoral reforms. The temporal starting point for my analy­sis is the adoption of nominal voting secrecy. Such reforms replaced the system of viva voce or open voting with voting that included ballots. France took a pioneering role in the adoption of ballots, introducing them in 1831. Germany a­ dopted secret voting in 1867, as part of the electoral law of the German Federation (Buchstein 2000: 356; Huber 1978, 2:227). The Constitution of the German Empire, a­ dopted in 1871, took up this provision of the North German Federation (Huber 1978, 2:289–306). Open voting was ­adopted in ­England in 1872 as part of the Secret Ballot Act. Belgium ­adopted secret voting in 1877 (Caramani 2000: 155). By 1914, most Eu­ro­pean countries had introduced remarkably similar reforms that addressed both irregularities in the organ­ization of campaigns and imperfections in the voting pro­cess. Of course, electoral corruption was not entirely eradicated in 1914. However, the po­liti­cal environment changed dramatically in 1918 a­ fter the introduction of proportional repre­sen­ta­tion and the creation of much larger electoral constituencies. During the period covered in this book, countries found themselves at dif­fer­ent stages in the pro­cess of democ­ratization. To begin with, the scope of suffrage differed across countries. In Britain and Belgium, suffrage was still restricted. Belgium maintained census requirements ­until 1892, requiring the payment of 20 gulden as a precondition of the vote (Caramani 2000: 151). By contrast, France and Germany guaranteed universal manhood suffrage. The countries included in this study also differ in the second dimension of democ­ratization: namely, the mechanism of executive se­lection (Dahl 1971). In Britain, France, and Belgium parliamentary majorities selected the executive. By contrast, in Imperial Germany the monarch retained the prerogative to appoint the government. Thus, on Robert Dahl’s dimension of “executive

44 C H A P TER 2

se­lection” Germany is less demo­cratic than the other countries included in this study. Including cases that take dif­fer­ent values on distinct indicators of the democ­ratization score in the analy­sis allows me to examine w ­ hether such regime attributes have different consequences for the formation of po­liti­cal majorities supporting electoral reforms. During the po­liti­cal revolt that led to Belgium’s in­de­pen­dence in 1830, Conservatives and Liberals forged an alliance against William I, the Protestant Dutch king. In the years following in­de­pen­dence, the two parties governed in a unity co­ali­tion, jointly overcoming the economic crisis that resulted from the war over Luxembourg and the loss of the Dutch market. This Unionist co­ali­tion disintegrated, however, ­after 1839. Middle-­class anticlericals left the co­ali­tion as a result of their dissatisfaction with the growing po­liti­cal influence of the Catholic Church and established the Liberal Party. Led by Walthère Frère-­Orban, Liberals governed Belgium, pursuing a policy of ­free trade that attempted to spur industrial development. The Conservative Party was established in 1852 and renamed in 1864 as the Federation of Catholic Circles and Conservative Associations. Initially, Conservative Catholics appealed for a higher tariff protection of rural districts that suffered ­under the ­free trade policy of the Liberals. As Stathis Kalyvas has pointed out, the politicization of religion occurred only ­after the economic split (1996: 172). Even though Frère-­Orban repeatedly asserted that “liberalism would commit suicide by descending into the religious field,” Liberals began to compete on an anticlerical platform in an effort to hold together the dif­ fer­ent factions within the party (ibid.; De Meeus 1962: 324). Britain also experienced regular po­liti­cal turnover despite the dramatic extension of suffrage during the period. During the 1840s, Conservatives experienced a po­liti­cal split over the adoption of the Corn Laws (Blake 1985: 57). The defection of a large group of politicians committed to ­free trade to the Liberal Party weakened the Tories temporarily, leaving b ­ ehind a small protectionist “rump party” (Blake 1985; Ziblatt 2017). Beginning in the 1870s, u ­ nder the leadership of Benjamin Disraeli, the Conservative Party experienced significant reor­ga­ni­za­tion. Disraeli established the Conservative Central Office and the National Union of Conservative Associations, an organ­ ization that brought together local voluntary associations. An additional ­factor that gave stability to the two-­party system in Britain was the religious cleavage: Conservatives w ­ ere defenders of the established Anglican Church and drew strong support from its members. By contrast, Liberals drew support from Catholics, Freethinkers, and non-­conformist voters (Parry 1993: 128–49).

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 45

In contrast to Belgium and Britain, the po­liti­cal system in France was dominated by a single party, the Republican Party. During the Third Republic, the initial po­liti­cal conflict was between supporters of the Republic and supporters of the monarchy. The Third Republic started off with the electoral defeat of the Republican Party (Rudelle 1982: 18). During the 1871 election, Monarchists achieved their last po­liti­cal victory. As voters sanctioned Republicans for the peace treaty with Germany, Monarchists gained a resounding triumph, winning four hundred of the five hundred fifty seats of the Chamber of Deputies (Berstein 2006: 12). The situation reversed beginning with the 1876 election. The Republicans won e­ very successive election between 1877 and 1902 and appointed twenty-­four out of the twenty-­seven prime ministers of the period. The two exceptions to this uninterrupted po­liti­ cal Republican dominance w ­ ere Pierre Tirard, an in­de­pen­dent politician who served as prime minister in early 1888 and in 1889, and Léon Bourgeois, a Radical, who occupied the role of head of government in early 1896. Throughout the Third Republic, the anti-­Republican Right or Conservative faction in the Assemblée encompassed a highly heterogeneous group of politicians. At the beginning of the Third Republic, this group included both Bonapartist and Monarchist legislators (Rémond 1954: 125–29; Hanson 2010). While Bonapartists w ­ ere followers of Napoleon, Monarchists w ­ ere divided between Legitimists and Orleanists. While the former followed the Count of Chambord, the last heir of the Bourbon dynasty, the latter ­were followers of the Count of Paris, the heir of the Orléans ­family. The anti-­Republican Monarchist opposition was profoundly divided. The division reflected not only disagreement about the branch of the royal ­family that could legitimately claim the monarchical succession in France but also dif­fer­ent emphases on the need to defend Catholic princi­ples in the new Republic. Following the adoption of the constitutional reforms of 1885, the importance of anti-­Republican forces waned. René Rémond, a historian of the po­liti­cal right in France, characterizes the po­liti­cal influence of this group as follows: “from t­ here on, incapable of stopping the vote on legislation they oppose, but hoping to recapture the majority, politicians on the right often give in to the temptations faced by all oppositions: they contest the legitimacy of the majority, by stirring up public opinion against the parliamentary regime” (1954: 158). Beginning in the 1890s, the non-­Republican right embraced new themes that created opportunities to express their opposition to Republican institutions. Th ­ ese themes included the defense of religion, the military, and nationalism. In the last de­cade of the nineteenth ­century and the first de­cade of the twentieth, new parliamentary groups

46 C H A P TER 2

representing anti-­Republican sentiment sprang up. The Patrie Française, a nationalist movement that brought together legislators previously affiliated with the Bonapartist group, won fifty legislative seats during the 1902 election. In 1901, Alfred Le Mun founded the Action Libérale, a po­liti­cal movement defending Catholic values and institutions. The Action Libérale remained the most significant parliamentary grouping of anti-­Republican deputies ­until 1914. On the left, Republicans encountered Socialist opposition. Léon Blum, the Socialist prime minister of the interwar period, noted that the development of Socialism in France had been slowed down or accelerated by the regression of Republican ideas. This statement recognizes the ability of the Republican Party to co-­opt and incorporate some of the first Socialist movements. In 1876, Albert Naguet established a parliamentary faction that called itself “the vanguard of the demo­cratic field” and criticized the governmental majority (Candar 2004: 117). During the 1890s, Socialist politicians criticized Third Republic institutions—­including calling for the elimination of the Senate—­and demanded an increased presence of the state in the economy through nationalization and the development of social protection policies. The Section Française de l’Internationale Ouvrière (SFIO), formed in 1905, unified the French Socialist Party and the Socialist Party of France. ­Under the leadership of Jean Jaurès, Socialists took up a revolutionary set of demands, objecting to all policies that might assure the domination of the bourgeoisie. ­These policies included the allocation of military credits and many outlays included in the oversized bud­gets of the period (Sirinelli 1994: 973). While Germany shared with France low po­liti­cal turnover, its party system was much more fragmented. Due to its disproportionate influence on the formation of the government, the German Conservative Party (Deutsche Konservative Partei) occupied the central role in the party system. Representing the interests of Protestant landowning elites, Conservatives stood for a protectionist economy and a staunch defense of the monarchy. As ­will be discussed below, during the first two de­cades following Germany’s unification, Conservatives ruled as part of an electoral co­ali­tion with National Liberals, a party representing the interests of large manufacturing elites and supporting Bismarck’s protectionist policies. The opposition to the Conservative–­National Liberal bloc encompassed a variety of parties, mobilizing ­either religious or economic cleavages or ethnic minorities. The Zentrum mobilized Germany’s Catholic minorities in opposition to the Kulturkampf. Embracing a Marxist radical critique of capitalism, the SPD,

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 47

Political fragmentation/ Increased need for cooperation in runoffs Elite splits

Electoral costs increase

Programmatic differentiation FIGURE 2.2.

Increase in support for reforms

Elite splits and demand for electoral reforms.

Germany’s Social Demo­cratic Party, mobilized the nascent working class. A variety of small liberal parties, including the ­Free Liberal Party and the German P ­ eople’s Party, opposed the government’s protectionist policy. Fi­nally, we also find a variety of parties representing the ethnic minorities of the empire, such as Poles, Danes, and Guelphs. In both France and Germany, the po­liti­cal dominance of Republicans and Conservatives established a profound asymmetry among parties in their ability to access state resources. Due to their dominance, Republicans and Conservatives ­were able to appropriate a large share of administrative state resources to compete in elections. In France, successive governments increasingly took for granted their ability to politicize the resources of the state apparatus. The politicization of state resources by incumbent parties reduced the willingness of dif­fer­ent actors to cooperate during the reform pro­cess and created a stalemate during reforms. As the demand for electoral reform among opposition parties increased, the re­sis­tance of Conservatives or Republicans to electoral changes also hardened. This polarized dynamic also translated to other policy areas. Parties that could not access state resources, such as the Monarchists, campaigned primarily using private resources. Republican re­sis­tance to reforms to limit governmental pressure (pression gouvernementale) increased Monarchists’ incentives to resist reforms to limit vote-­buying. How did politicians in France and Germany overcome this initial stalemate? One impor­tant po­liti­cal development that broke this stalemate was a split in the incumbent co­ali­tion. In figure 2.2, I summarize the pathways by which elite splits facilitate the formation of a po­liti­cal majority favoring reforms. On the one hand, elite splits increased electoral fragmentation and the need for cooperation in runoffs. On the other hand, elite splits contributed to programmatic differentiation among politicians in the resource-­endowed majority. Both developments contributed to a rise in the electoral costs associated with illicit strategies, reducing the attractiveness of the status quo for legislators with access to state resources. Legislators

48 C H A P TER 2

experiencing the effects of ­these costs joined the parliamentary co­ali­tion supporting a change in electoral rules, which had previously consisted of resource-­constrained politicians alone. This electoral realignment facilitated the formation of a po­liti­cal majority favoring reform. Let us consider now the elite splits in the two countries with long-­ standing po­liti­cal majorities. I define elite splits in electoral and partisan terms alone. Most commonly, elite splits take place through the breakdown of one of the parties represented in parliament. This was the case in France, with the separation of the Republican Party into a centrist faction and Radical Socialists. Elite splits also occur if an electoral co­ali­tion of two or more candidates breaks down. This was what happened to the Kartellbündnis, an electoral alliance between Conservatives, F ­ ree Conservatives, and National Liberals. While existing studies extend the term “elite splits” to include non-­ parliamentary actors (Ansell and Samuels 2014), I ­will restrict the analy­sis to parties or electoral alliances. In France, the incumbent Republican Party began to fracture beginning in the 1890s. This was, at first, an informal split between a centrist group—­ the Opportunist Republicans—­and a programmatic faction, the Radicals (Daniel 1897). In 1901, Radicals established their own po­liti­cal party, the Republican-­Radical and Radical-­Socialist Party (Parti républicain radical et radical-­socialiste) (Kayser 1962: 291). Both Opportunist Republicans and Radicals shared a commitment to the po­liti­cal institutions of the Third Republic. The two parties disagreed, however, on the appropriate role of state intervention in the economy, an issue that dominated French parliamentary deliberations at the time. Republicans embraced a liberal position on the role of the state, arguing that “it was not the role of the state to solve social prob­lems, but only to create the necessary conditions to solve this prob­lem” (Berstein 2006: 11). By contrast, Radicals pushed for an expansion of state intervention in the economy and for the introduction of social insurance, such as disability and old age assistance. In terms of ­labor market policy, Radicals supported legislation affirming the rights of u ­ nions. During the debates over changes to the tax system, which occupied a significant amount of the electoral agenda before World War I, Radical deputies ­were the strongest advocates for the introduction of a progressive income tax (Kayser 1962; Delalande 2011). During the first two de­cades a­ fter the unification of the German Empire, the Conservative, F ­ ree Conservative, and National Liberal parties formed an electoral alliance, the Kartellbündnis. This alliance reflected efforts to uphold the common economic interests of landed and industrial elites in support

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 49

of agricultural and industrial tariffs. In practice, this electoral coordination among cartel parties could be established at dif­fer­ent levels of electoral competition. In some cases, this occurred at the level of a province by fielding a common candidate, known as the Kartellkandidaten. Prior to 1890, this was often the case in many Prus­sian provinces. At other times, parties could trade off support across districts. In other words, candidates affiliated with cartel parties could support a Conservative candidate in one district in exchange for common support for a National Liberal in another. The electoral coordination during the period of the Kartellbündnis went hand in hand with an intensive use of the state apparatus during elections. The architect of this electoral strategy was Robert von Puttkamer, Prus­sia’s minister of interior from 1881 to 1888. Puttkamer established a sophisticated system of bureaucratic rewards and punishments for officials who aided the Kartellkandidaten. ­These included threats of disciplinary mea­sures and promises of awards (Arsenscheck 2003: 194). Thomas Kühne discussed the strategy of state mobilization as follows: “In par­tic­u­lar, the mid-­level officials that displayed indolence during the elections w ­ ere u ­ nder regular attack from above. ­After all, ­these employees played a critical role in affecting w ­ hether the armies of low-­level employees could be set in motion as multiplicators or as voters in a district” (1994: 61). Otto von Bismarck himself endorsed the politicization of state employees. In an intervention in the Reichstag, Bismarck noted that “employees of the state need to serve a par­tic­u­lar administration and to carry out the directives of the latter” (StBer DtRt 1881, encl. no. 135). This electoral alliance disintegrated, however, at the beginning of the 1890s. Two ­factors contributed to this breakdown. At the time of the renewal of the anti-­Socialist legislation in 1889, Chancellor von Bismarck proposed a version of the law that imposed extremely harsh punishments for civil liberties violations. The National Liberals rejected the proposal in protest, breaking the co­ali­tion with Conservatives (Sperber 1997: 204). The economic policies pursued by Leo von Caprivi, who succeeded Bismarck as chancellor, fed the po­liti­cal disagreement among National Liberals and Conservatives. As part of the New Course, Caprivi advocated for a decrease in tariffs for grains, a proposal that met with Conservative opposition (ibid., 205). This disagreement among parliamentary elites had implications for electoral competition across Germany. As a result of their disagreement over the new policy, Conservatives and National Liberals broke up their regional alliances and began to field in­de­pen­dent candidates. During the 1890 elections, Conservatives and National Liberals who had previously

50 C H A P TER 2

fielded one candidate competed against each other in seventy-­three dif­fer­ ent constituencies (ibid., 204). In both France and Germany, such elite splits had significant implications for electoral be­hav­ior and the use of illicit electoral strategies: they increased the electoral costs resource-­endowed candidates could incur in using such strategies. The campaign promises announced by individual candidates reveal an increasing use of programmatic promises by Radical candidates. As early as the 1898 election, one-­third of the Radical candidates competed on the basis of programmatic appeals ( JORF 1899). As a result of their programmatic positioning, Radicals ­were increasingly able to form electoral co­ali­tions during runoffs with Socialist deputies (Daniel 1898). The investment in such programmatic promises widened the divide between Radicals and Republicans regarding the appropriateness of illicit strategies. Radicals spearheaded parliamentary investigations of Republican legislators’ electoral malfeasance, which led to the downfall of the government of Jules Méline. One chronicler of the parliament noted how Radicals “challenged Méline both in the parliament, in the po­liti­cal press and private meetings” (Daniel 1897: 1). In the following years, the conflict between Radicals and Republicans over electoral reform shifted to the floor of the Assemblée and the Commission of Universal Suffrage of the Chamber. Joseph Ruau, a Radical legislator from Haute-­Garonne and president of this commission, spearheaded proposals to reform the electoral law and reduce opportunities for intimidation. Members of the government vehemently opposed such proposals. Commenting on the opposition of Republicans, Ruau noted that they “considered the proj­ect to be a weapon addressed against the government” ( JORF ChamberDeb, 17 December 1901). This conflict between Radicals and Republicans became the driving force ­behind the adoption of electoral reforms in France. In Germany, the disintegration of the Kartellbündnis increased the electoral costs associated with the use of illicit strategies. One of the immediate consequences of this elite split was an increase in electoral competition in each district. When the electoral cartel had been in place, Conservatives, National Liberals, and F ­ ree Conservatives had agreed to support only one candidate in each district. Following the breakdown of the cartel, electoral fragmentation in each district increased. At the same time, the greater number of candidates competing in the first round also increased the need for coordination during runoffs. The rapid increase in the number of runoffs illustrates this growing fragmentation on the right. Prior to 1890, less than 15  ­percent of the Conservative politicians elected to the Reichstag ­were

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 51

elected in runoffs. For candidates of the German Imperial Party (Deutsche Reichspartei), the ratio increased from 15 ­percent before 1890 to around 35 ­percent in subsequent elections. Among the Kartellparteien, National Liberals found themselves in the most competitive situation following the breakdown of the cartel. In 1887, around 13 ­percent of National Liberals ­were elected during runoffs. This increased to 71 ­percent in 1893 and 91 ­percent in 1912, the last election prior to World War I. The need for runoffs decreased the attractiveness of illicit strategies. In this new competitive environment, it became costlier to deploy policemen or mayors as brokers. The strategy could backfire, as first-­round opponents could withhold support during runoffs. This rise in electoral costs contributed to a change in the electoral strategy pursued by high-­level German officials. The new strategy would be a policy of “noiseless deployment” of the bureaucratic apparatus (Müller 1963; Mares 2015), with successive governments taking all pos­si­ble steps to conceal evidence of their po­liti­cal intervention from the public. Beginning in the 1890s, Prus­sian officials in the central ministry refrained from using the ample rewards and punishments that had been in place u ­ nder Puttkamer. This represented an extraordinary change in policy (Arsenscheck 2003: 197). The electoral strategy pursued by the Caprivi administration during the 1903 election reflects this new policy of “noiseless intervention.” The government recommended taking par­tic­u­lar precautions when competing against candidates who could become potential co­ali­tion partners during runoffs (Arsenscheck 2003: 198). The government established a special office in Berlin led by a military official whose goal was to disseminate electoral material to voters (Mares 2015: 51). In a secret memorandum, the Ministry of Interior instructed local officials to take unpre­ce­dented caution to hide any evidence of its intervention. To avoid exposing the participation of officials, the Landräte had to provide the addresses of specific contacts (Vertrauensmänner) whom the central office in Berlin could communicate with (Müller 1963). This noiseless strategy strongly contrasts with the heavy-­ handed intervention of e­ arlier elections. Reflecting this increase in electoral costs, the preferences of resource-­ endowed politicians t­ oward electoral reforms changed during this period. Beginning in the 1890s, National Liberals began to endorse proposals for such reforms. Faced with increasing competition from Conservatives on the right and F ­ ree Liberals on the left, National Liberals threw their support ­behind reforms. ­After 1896, National Liberals in the Reichstag supported legislation to provide better protection for voter secrecy and cosponsored

52 C H A P TER 2

legislation mandating the standardization of electoral urns, which was ­adopted on the eve of World War I. In decisions concerning the invalidation of elections due to the politicization of state resources, National Liberals increasingly sanctioned the use of ­these strategies. Th ­ ese decisions led to the invalidation of some of ­these races and to changes in electoral jurisprudence concerning the electoral use of state resources. Fi­nally, the period ­under study ­here is characterized by dramatic economic developments in all countries as they underwent pro­cesses of modernization. The size of the population and implicitly the population density in dif­fer­ent constituencies ­rose over the course of the period ­under study. The percentage of the population employed in manufacturing, a standard mea­ sure of economic development, also increased during this period. As ­these developments increased the costs of vari­ous illicit strategies, they imposed constraints on the use of ­these strategies even for resource-­endowed politicians. Discussing the implications of economic development for vote-­buying strategies, Susan Stokes and her coauthors (2015) have conjectured that candidates competing in wealthier and more populous constituencies incur higher costs in financing campaign expenditures. By increasing the “economic costs” of illicit strategies, such developments may increase support for electoral reforms even among resource-­endowed politicians. Such economic changes may contribute to the formation of a more encompassing co­ali­tion supporting electoral reforms. More specifically, economic development may reduce the ability of employers to use coercive or punitive strategies against voters who disobey their instructions. If economic opportunity in a district increases, the employers’ threats of layoffs may lose their bite. In such conditions, workers coerced to vote for a par­tic­ul­ ar candidate may also choose to look for employment elsewhere, which also may impose costs on employers. The economic costs of illicit strategies are higher in districts with higher economic fragmentation (Mares and Zhu 2015). An analy­sis of po­liti­cal developments in Prus­sia during this time documented that the increase in ­labor scarcity weakened the ability of landlords to influence voters’ choices (Ardanaz and Mares 2014). This, in turn, decreased the opposition of politicians in ­these districts to the adoption of reforms mandating the introduction of secret voting. The explanation proposed in this chapter refines the predictions of the modernization perspective, a leading approach in the study of democ­ ratization. According to this perspective, economic changes play an impor­tant role in the democ­ratization of electoral practices. I conjecture, however,

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 53

that the consequences of ­these economic changes for demand for electoral reforms vary across policy areas. Economic developments are likely to play a more impor­tant role in the formation of a pro-­reform majority in the case of illicit strategies that politicize private resources. Candidates who use private resources internalize the increase in the economic costs of using illicit strategies more strongly than do candidates who compete using state-­financed resources. Economic developments affect the private campaign of a candidate or the economic profitability of a firm directly, leading the candidate to reevaluate the attractiveness of illicit strategies. I conjecture, however, that economic developments are less consequential for the formation of a pro-­ reform majority in the case of reforms that politicize state resources. B ­ ecause ­these illicit strategies are financed by public resources, candidates do not internalize the increasing economic costs associated with an increase in transfers or state resources to mobilize voters. Th ­ ese predictions run against ­those from economic models of electoral reforms (e.g., Lizzeri and Persico 2004). While economic development may lead to a decrease in bribery, the relationship between economic development and electoral strategies that use state resources is more ambiguous. In summary, this discussion has developed a number of theoretical expectations that ­will inform the analy­sis presented in the book. I approached this question first from a micro-­level perspective, arguing that legislators’ preferences with regard to electoral reforms can be explained by a combination of three dif­fer­ent types of considerations. Th ­ ese include concerns about relative resource advantage, economic constraints, and electoral constraints. Macro-­ level po­liti­cal ­factors as well as economic changes influence the significance of ­these dif­fer­ent ­factors and set in motion dif­fer­ent co­ali­tional dynamics surrounding electoral reforms. In cases of symmetric competition between two parties of equal size and strength, the resource asymmetry is relatively low. In such contexts, I expect politicians who encounter a relative resource disadvantage to initiate proposals for reform. Rising economic or electoral costs may increase the impetus for reforms experienced by such politicians. Due to the low asymmetry in accessing resources, po­liti­cal consensus over electoral reforms is likely to be rapid in such cases. Th ­ ese developments characterize the co­ali­tional formation in Britain in the case of vote-­buying reforms and in Belgium and Britain in the case of reforms to limit fraud. A po­liti­cal majority that supports electoral reforms develops with greater difficulty in countries with a hegemonic party system. ­Here, the enduring po­liti­cal asymmetry in the distribution of po­liti­cal resources reduces the willingness of the incumbent party to forgo this resource advantage. In ­these

54 C H A P TER 2

cases, elite splits are particularly impor­tant in breaking this po­liti­cal stalemate. By increasing the electoral costs associated with illicit strategies, elite splits facilitate a co­ali­tion realignment, making a parliamentary majority that supports reforms pos­si­ble. In France, following the elite split between Radicals and centrist Republicans, Radicals joined resource-­constrained politicians such as the Socialists and, in some instances, Monarchists in supporting electoral reforms. In Germany, following the breakdown of the Kartellbündnis, National Liberals joined the co­ali­tion of resource-­constrained politicians favoring electoral reforms. ­These considerations about electoral costs help explain how po­liti­cal majorities favoring reforms ­were pos­si­ble even in countries with hegemonic po­liti­cal parties and low po­liti­cal turnover. Co­ali­tions for Electoral Reforms

Reforms to decrease the use of illicit strategies ­were not ­adopted in one fell swoop, in a big bang of democ­ratization. Rather, ­these reforms ­were ­adopted through a series of piecemeal changes to electoral laws. The trade-­offs confronting legislators voting on ­these reforms varied across dif­fer­ent dimensions, which led to the formation of dif­fer­ent po­liti­cal majorities across t­ hese reform dimensions. Reforms to limit vote-­buying presented dif­fer­ent trade-­ offs as compared to reforms seeking to reduce opportunities for the use of state resources during elections. Both the distribution of po­liti­cal resources and the salience of economic and electoral costs varied significantly across ­these dimensions. Using the predictions outlined above, I ­will examine the importance of considerations about access to resources, electoral costs, and economic costs in shaping demand for electoral reforms. This ­will allow me to analyze the makeup of the electoral co­ali­tion that supports the adoption of dif­fer­ent reforms. REDUCING ACCESS TO STATE RESOURCES DURING CAMPAIGNS

Prior to the unification of the German Empire in 1871, high-­ranking bureaucrats and politicians competing in dif­fer­ent principalities shared a similar understanding that the electoral use of state resources was a legitimate form of campaigning (Rejewski 1973). In the aftermath of the 1848 revolution, the Prus­sian Ministry of Interior issued a number of decrees codifying such expectations. Th ­ ese decrees affirmed the obligations of state employees to be loyal to the government and the intention of the administration “to lay

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 55

off mercilessly any state employees guilty of taking an adversarial stance against the government” (ibid., 19). During the first elections ­after 1870, high-­ranking German politicians condoned the use of state resources during elections. In a speech in the Reichstag in 1881, Otto von Bismarck, Germany’s chancellor, considered that “employees of the state could not be entirely nonpartisan. Their position in the administration presupposes a certain level of partisanship at elections” (StBer DtRt, 3 March 1881). Robert von Puttkamer, Prus­sia’s minister of interior, expressed a similar position. “The government,” Puttkamer argued, “expects and supports an open intervention of state officials during elections” (ibid., 17 March 1881). In France, the expectation that state employees should actively support the government during elections originated during the Second Empire. At the time, the candidature officielle was legally recognized as an official theory of state power (Pierre 1893; Pilenco 1928). In one formulation of this po­liti­cal princi­ple, the emperor had the right to intervene on behalf of a candidate, ­because he “submits his power to a popu­lar decision during each election” (Charnay 1965: 216). The idea that the monarch could use any means at his disposal to support the po­liti­cal candidate of his choice went unchallenged during this period. Summarizing the use of the resources of the administration during the Second Empire, Alexandre Pilenco noted that “beginning with the first Napoleonic elections, all the candidates who ­were ­either supported or not actively opposed by prefects ­were elected with strong majorities. In 1863, the police chief of Chinon, a municipality in Indre-­et-­ Loire, exclaimed ­after the election, ‘It is us, the police, who have determined the results’ ” (1930: 171). The transition to the Third Republic brought about a change in the politicization of state resources. The approach of successive Republican administrations to this question is characterized by internal inconsistencies and hy­poc­risy. Officially, the government condemned the use of state resources during campaigns as a vestige of the monarchical past. Nonetheless, the government did not follow through with its promises to break with past practices (Pilenco 1930: 171). As one Republican minister of the period forcefully asserted, “the government has not only the right but also the obligation to communicate to voters the names of candidates who are favorable to the government and ­those who oppose it” (Weil 1895: 230). Full participation by state employees in elections was expected and neutrality was frowned upon. A remark by a prefect summarized this expectation as follows: “neutrality is a hidden hostility” (la neutralité, c’est une hostilité qui se cache) (ibid., 231).

56 C H A P TER 2

French legislators pursued two strategies to limit the electoral use of state resources. The first attempted to enumerate the ways in which candidates politicized state resources in an effort to provide l­ egal clarity on the strategies that ­were, in fact, punishable offenses. This raised, of course, countless questions. Was the participation of a minister at a rally a punishable offense? The endorsement of a candidate by a minister? When was the use of employees of the local administration as brokers an illicit offense? The second strategy attempted to increase the severity of l­ egal punishments for such irregularities. An impor­tant goal pursued by reformers was to calibrate the punishment for irregularities that used state resources and for other campaign-­related irregularities in order to reduce opportunities for substitution. In the initial stages of the reform pro­cess, demand for t­ hese reforms originated with resource-­constrained politicians. This co­ali­tion included both Socialist and Monarchist legislators. Cuneo d’Ornano, a legislator affiliated with the anti-­Republican Appel au Peuple, and René Viviani, a Socialist from Seine, cosponsored some of the earliest legislation sanctioning the use of state resources in elections ( JORF ChamberDoc 1890, no. 843). Socialists portrayed this initiative as a mea­sure necessary for their survival. We need “to defend ourselves against the candidature officielle and have against us all the forces of power” ( JORF ChamberDeb, 17 March 1910). Conservative legislators also supported proposals to limit the use of state resources in elections, affirming that “the government does not have the right to use the weight of its entire force . . . ​in elections” ( JORF ChamberDoc 1890, no. 843). A po­liti­cal majority favoring reforms was only established a­ fter the elite split between Republicans and Radicals. Beginning in 1898, Radicals criticized the deployment of state agents as brokers and asked the government to ensure that “state employees re­spect the in­de­pen­dence of universal suffrage” (Daniel 1898: 436). Facing opposition from centrist Republicans, prominent Radical politicians, such as Joseph Ruau and Félix Defontaine, presided over the Commission of Universal Suffrage of the Chamber of Deputies and drafted successive versions of electoral reform legislation. Following this realignment, the electoral majority favoring reforms included not only Radicals and Socialists but also Conservative anti-­Republican legislators. A compromise favoring electoral reforms was reached in 1910. The legislation emerging from this compromise introduced stronger sanctions for what was known at the time as “collective corruption”—­namely, offers of access to resources directed to communes.

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 57

Similar po­liti­cal developments took place in Germany. The split of the previous electoral alliance contributed to the creation of a new po­liti­cal majority in f­ avor of stronger sanctions against the electoral use of state resources. During the last electoral cycles prior to World War I, National Liberals voted increasingly to invalidate the elections of candidates who politicized state resources during campaigns, joining the progressive co­ali­tion of the Reichstag (StBer DtRt, encl. no. 1497). In the commissions investigating electoral irregularities, National Liberals increasingly supported stronger sanctions against candidates who used state resources. ­These included, among ­others, sanctions for the use of official “symbols of state power” during campaigns and for the reliance on local administrative employees to distribute ballots. Even though Imperial Germany did not adopt a new law sanctioning the use of state resources during the period prior to World War I, German lawmakers modified electoral jurisprudence to punish the use of state resources (Hatschek 1920). In agreement with the propositions outlined in this chapter, the new po­liti­cal majority favoring ­these changes included National Liberals in addition to the other resource-­constrained parties in the Reichstag. REFORMING VOTE-­B UYING

Many Eu­ro­pean countries introduced l­ egal provisions to sanction offers of money or goods in exchange for votes prior to the introduction of mass suffrage. The French Penal Code, a­ dopted in 1810, defined bribery as a criminal offense and recommended sanctioning any citizen who had bought or sold a vote with the loss of po­liti­cal rights (Code Pénal 1815: 43). The French law became the blueprint for the civil codes ­adopted in several other Eu­ro­pean countries during the first de­cades of the nineteenth ­century. The Prus­sian, Belgian, Swedish, Finnish, Danish, and Austrian civil codes took up the provisions sanctioning bribery that had been introduced in the Napoleonic Code (Mayer 1906: 342; Nypels 1868: 336). Electoral laws a­ dopted during the second half of the nineteenth c­ entury, during a period of suffrage expansion, incorporated t­ hese sanctions of bribery. In France, the Organic Decree of 1852, which regulated the organ­ization of elections throughout the Third Republic, defined bribery as a promise of “money, ­things [effets] or any assets [valeurs] in exchange for the promise to give a vote or to abstain from voting” ( JORF ChamberDoc no. 1852, Article 18). The law included the punishments for ­these offenses that had been established by the Napoleonic Code. Both persons who made such

58 C H A P TER 2

promises and persons who accepted ­these offers ­were sanctioned with a prison sentence of three months to two years and a fine of five hundred to five thousand francs. ­Because it lacked a criminal code, Britain relied on a variety of electoral laws to sanction vote-­buying. The earliest efforts to do so date to the Treating Act of 1677 (Rogers and Willoughby 1906, 2:294). The Corrupt Practices Act, enacted in 1854, was an attempt to rein in bribery (Seymour 1915: 385). The Act defined corrupt exchanges as efforts to “give, lend, agree to give, agree to lend, offer, promise, promise to procure money or valuable consideration” (Rogers 1880, 2:296). To limit electoral irregularities, the Corrupt Practices Act imposed punishments both for giving and for receiving bribes. While the sanctions for ­these illicit strategies ­were highly variable, the highest punishment was ­either one year of prison with or without hard ­labor or a fine of two hundred pounds (Rogers and Willoughby 1906, 2:395). The Corrupt Practices Act was, however, in­effec­tive in reducing the incidence of bribery. Politicians of the period denounced the Act as a “pompous profession,” “a mockery,” and “a delusion” (Seymour 1915: 386). The most impor­tant weakness of this law was the poor calibration between the punishment of bribery and that of other campaign irregularities, such as treating. By imposing significantly higher punishments on bribery relative to treating, the law created perverse incentives for actors to substitute one irregularity with another that carried a lesser punishment. The significance of treating in British elections increased dramatically ­after the adoption of the Corrupt Practices Act. ­These differences between Britain and continental Eu­ro­pean countries had consequences for the importance of vote-­buying within the mix of electoral irregularities. In Britain, vote-­buying remained pervasive, straining the resources of individual candidates. By contrast, the penalization of vote-­buying in continental Eu­ro­pean countries was relatively successful in deterring candidates from using this strategy. In both France and Germany, vote-­buying played only a negligible role in the mix of campaign strategies used by candidates. Indeed, less than 10 ­percent of the petitions alleging electoral irregularities submitted to the German Reichstag contained allegations of vote-­buying. ­These differences in the status quo between continental Eu­ro­pean countries and Britain also affected the strategies pursued by reformers. In Britain, the failure of the Corrupt Practices Act to sanction vote-­buying led to a dramatic increase in election expenses. The House of Commons estimated that the campaign expenditures during the 1880 election exceeded £3,000,000, which represented a 200 ­percent increase over the total from 1874 (Rix 2017: 70). As a result, reformers’ main priority was to

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 59

limit the growth of election expenditures. However, in countries that relied on the penal code to sanction corruption, reformers faced an uphill b ­ attle. For example, in France, courts adjudicating cases of bribery successively raised the evidentiary requirements from plaintiffs. A successful conviction required not only the documentation of bribery or treating but also evidence that ­those offering a bribe intended to influence the vote (Pierre 1893: 169). At the same time, courts required plaintiffs to document “the existence of a declared or implied market” in the exchange of money or gifts for votes ( JORF ChamberDeb, 18 November 1899). As a result, reformers on the continent pursued alternative ways to limit corruption. ­These included eliminating the evidentiary standard to document corrupt intent and sanctioning promises—­and not only of transactions where voters had accepted candidates’ offers. Who opposed and who supported vote-­buying reforms? Similar to other electoral irregularities analyzed in this book, relative resource advantages explain legislators’ initial preferences with regard to the desirability of reforms: resource-­constrained candidates favored the adoption of electoral reforms, while resource-­endowed candidates opposed changes to the status quo. The strategy of vote-­buying primarily deployed candidates’ private resources. As such, the initial cleavage line over vote-­buying reforms was between rich and poor politicians. ­Because of the importance of private resources in this illicit strategy, the composition of the parliamentary co­ali­ tion that supported ­these reforms differs from that of the co­ali­tion seeking to limit the use of state resources. While Monarchist and Conservative legislators supported reforms to limit the electoral use of state resources, ­these non-­Republican Conservative legislators opposed reforms to limit vote-­buying. Monarchists believed limiting the ability of candidates to offer private goods to voters could leave candidates entirely defenseless against their Republican opponents who benefited from access to state resources ( JORF SenateDeb, 16 January 1906). In contrast to Monarchists, other po­liti­cal parties on the right w ­ ere divided about the advantages of reforms to limit vote-­buying. The electoral majority supporting the adoption of vote-­buying reforms was a heterogeneous co­ali­tion of legislators from a variety of parties. This co­ali­tion also included resource-­constrained Republican and Radical legislators, as well as politicians competing in more eco­nom­ically developed constituencies. Both resource-­endowed and resource-­constrained legislators competing in ­these environments concurred that a reform that sanctioned vote-­buying more stringently could limit the ability of wealthier candidates to deploy

60 C H A P TER 2

t­ hese strategies in ­future elections. Electoral reforms w ­ ere a shield to limit the ­future entry of “moneyed candidates,” the candidatures d’argent. Unlike other reforms examined in this study, considerations about electoral costs are insignificant in explaining support for vote-­buying reforms. In Britain, the impetus for electoral reforms also originated with resource-­ constrained politicians. The “saturnalia” of the 1880 election (where the campaign expenditures for vote-­buying and treating exceeded £3,000,000) provided the immediate impetus for ­these reforms. Reports of electoral irregularities documented that Conservatives significantly outspent Liberals on agents, bribes, and overall campaign expenses. Importantly, however, Liberals experienced a more stringent resource constraint in constituencies that saw more rapid population growth and higher manufacturing employment growth. Liberals competing in ­these constituencies spent a significant share of their wealth on campaign expenditures. The Corrupt and Illegal Practices Act, a­ dopted in Britain in 1883, addressed the prob­lem of rampant campaign expenditures by imposing strict ceilings on the amount that could be spent by politicians during elections and by increasing sanctions on politicians who exceeded ­these limits. At the same time, the reform ­limited “indiscriminate treating” by candidates, a practice that resulted from candidates’ decisions to keep public ­houses open during the entire campaign. Another impor­tant source of campaign expenditures in Britain was the overemployment of brokers. The 1883 reforms sought to limit ­these practices as well as to encourage candidates to rely on volunteers rather than paid agents. All in all, the reforms ­adopted in 1883 significantly reduced spending on bribes, treating, and agents and created incentives for candidates to invest in modern party organ­izations. REFORMS PROTECTING VOTERS’ AUTONOMY

As discussed above, candidates competing in nineteenth-­century elections took advantage of imperfections in voting technology to monitor voters’ choices when they cast their ballots. To this end, candidates placed their representatives in the immediate vicinity of ballot boxes, where they could observe and attempt to modify voters’ choices. At the same time, t­ hese brokers could impose strong punishments on voters who had cast an “incorrect” ballot. A statement by Alfred Krupp, a prominent German manufacturer, illustrates how nineteenth-­century employers took such economic coercion for granted: “We are on our own property, as the school has been built by the Krupp ­family and we have the right to look each voter in the eye

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 61

and appeal to his conscience so that he votes for the person from whom he has received his bread” (Paul 1987: 35). Electoral reforms attempted to remedy this prob­lem by better protecting voter secrecy at the ballot box. The proposals that attempted to reduce violations of voter autonomy w ­ ere remarkably similar across countries. Reformers’ first attempt to protect voter secrecy entailed the introduction of ballot envelopes. Reformers hoped that ballot envelopes would limit the ability of brokers located in the vicinity of the electoral urn to keep a tally of voters’ choices. In the German Reichstag, Adolf Gröber, a Zentrum politician, justified ballot envelopes as follows: “The question of ­whether a ballot is white or not entirely white, or w ­ hether it has a blue, gray, yellow or red shade, w ­ ill no longer be a valid question a­ fter the introduction of ballot envelopes.” Other legislators concurred that “controlling votes by controlling the color of ballots ­will no longer be pos­si­ble” (StBer DtRt, 29 January 1896). In addition to ballot envelopes, reformers advocated isolating spaces. German reformers regarded such private voting spaces as innovations that guaranteed that “the vote expressed the true conviction of the voter” (StBer DtRt, 15 January 1890). Isolating spaces w ­ ere regarded as a necessary complement to ballot envelopes, giving voters the opportunity to “be alone and unobserved for a few seconds, so that they could insert in the envelope the ballot that corresponded to their true convictions” (ibid.). In ­these private spaces “voters who ­were pressured from all directions could reflect in silence before casting their vote” (StBer DtRt, 1 February 1899). The crucial resource for the production of this irregularity was access to employers. Economic brokers such as employers ­were uniquely positioned to take advantage of violations in voting secrecy ­because they controlled the wages and employment of voters and thus could impose power­ful post-­ electoral punishments on ­those who made “incorrect” choices. This resource asymmetry had both individual-­level and partisan implications for support of the initial proposals for electoral reforms. At the individual level, the wealth of a candidate or his occupational status was likely to shape his preferences for electoral reforms. Wealthy candidates or employers w ­ ere likely to support the status quo. Indeed, it was not uncommon at the time for the main local employer to stand for elections in the constituency where he was the dominant “bread-­lord.” This was the case for both rural and industrial employers. Parties differed in their ability to co-­opt t­ hese private brokers. In Germany, both Conservatives and National Liberals enjoyed a relative resource advantage in enlisting employers as brokers. Nearly 45 ­percent of Conservative

62 C H A P TER 2

legislators and 40 ­percent of National Liberal legislators w ­ ere employers. In France, a large number of Republican and Monarchist legislators w ­ ere employers. By contrast, resource-­constrained parties in this regard included Socialists, in both countries, and, in Germany, the Catholic Zentrum and ­Free Liberals. The asymmetry in the distribution of resources necessary to create this irregularity generates a much stronger left-­r ight distinction as compared to other reforms examined in this study. Over time, this initial electoral co­ali­tion that opposed electoral reforms disintegrated, giving rise to a new majority that supported the protection of voter secrecy. Two developments reduced the attractiveness of the status quo for legislators, which contributed to the co­ali­tional realignment. First, a number of economic developments increased private brokers’ costs of coercion. Employers’ threats of economic sanctions for voters who defied coercion w ­ ere weaker if voters had more outside opportunities. As a result, increased economic diversification or l­ abor scarcity increased the costs of electoral repression for employers. Commenting on the consequences of l­abor scarcity for electoral politics in Prus­sian districts, one con­temporary noted: In the early days, the electoral pressures of landowners on rural workers ­were certainly not insignificant. In ­later periods, landowners had to use this means of power [Machtmittel] very carefully b ­ ecause of the ­labor shortage that existed in the countryside. One was happy if one could keep one’s employees and one was careful to not antagonize one’s employees through electoral harassment and to not drive them to the cities. (Wulff 1922: 13) Additional po­liti­cal developments contributed to the formation of a po­liti­cal majority that supported electoral reforms by raising the electoral costs associated with coercion. In France, the elite split between centrist Republicans and Radicals created power­ful incentives for candidates to invest in programmatic promises in an effort to differentiate themselves from their competitors. For Radicals and some Republican candidates, economic coercion became incompatible with their new programmatic appeals. As a result, they began to endorse reforms to protect electoral secrecy. In May 1907, Charles Benoist, a Republican legislator and the president of the Commission of Universal Suffrage in the Senate, commented on the formation of this new po­liti­cal majority as follows: “At this point in time, no ­matter which side of the Chamber we are on, we agree on the po­liti­cal and moral necessity of making the theoretical secrecy of the vote a real­ity and ensuring the freedom of the voter” ( JORF ChamberDeb, 21 May 1907).

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 63

In Germany, one can also document a shift in the position of National Liberal politicians on the question of electoral reforms following the breakdown of their alliance with Conservatives. During the 1880s, National Liberals voted with Conservatives to oppose the introduction of ballot envelopes and isolating spaces. However, during the 1890s, the position of National Liberals about the desirability of electoral reforms changed. In January 1896, Ernst Bassermann, the leader of the National Liberal faction of the Reichstag, reported that a large majority of his faction supported the introduction of ballot envelopes and private voting spaces (StBer DtRt, 29 January 1896). National Liberals’ support for greater protections for voting secrecy continued even a­ fter the introduction of ballot envelopes in 1903. An additional obstacle to voting secrecy resulted from the heterogeneity in the design of electoral urns. Small urns allowed ballots to be stacked, enabling election officials to identify voters’ choices. In 1909, National Liberal politicians cosponsored legislation supporting the introduction of standardized electoral urns (StBer DtRt 1909, encl. no. 21). Th ­ ese legislative initiatives w ­ ere successful; Germany ­adopted a standardized urn on the eve of World War I. In both France and Germany, the electoral majority in support of greater protections for voter secrecy brought together politicians who lacked access to economic brokers and resource-­endowed legislators who faced high economic or electoral costs associated with illicit strategies. By 1910, the French electoral co­ali­tion that supported the introduction of reforms to protect voter autonomy included Republicans competing on the basis of programmatic promises in addition to Socialist and Conservative legislators. In Germany, National Liberals joined a co­ali­tion for reform that included Catholics, Socialists, and F ­ ree Liberals. On the eve of World War I, Conservatives remained the only party holding out against electoral reforms. LIMITING THE INCIDENCE OF FRAUD

Commenting on the prob­lem of electoral fraud, Joseph Stalin famously quipped, “It’s not the p ­ eople who vote that count, it’s the p ­ eople who count the ballots.” Reforms to curtail electoral fraud attempted to limit the discretion that could be exercised by the “­people who counted the ballots” and ensure that elections represented the popu­lar w ­ ill as expressed by the ballots that ­were cast. The earliest efforts to sanction fraud date to the Napoleonic Code of 1810, which imposed strong punishments on any modification of electoral results associated with the counting of ballots. It sanctioned fraud with

64 C H A P TER 2

imprisonment and the loss of po­liti­cal rights, such as the right to vote and the right to be elected. The provisions a­ dopted in the Napoleonic Code spread to neighboring countries such as Belgium and Germany and ­were incorporated as part of penal codes and electoral laws a­ dopted during the first half of the nineteenth ­century. Electoral laws ­adopted in France in the first half of the nineteenth c­ entury went well beyond the provisions of the Napoleonic Code, sanctioning additional irregularities that occurred when ballots ­were counted. ­These included “the removal of urns containing ballots that had not been counted yet” and the “modification of the votes made ­either by members of the voting bureau or by authorities designated to count the ballots” ( JORF Lois et Décrets, 15 March 1849). ­These penal provisions ­were relatively successful in reducing the incidence of fraud in Eu­ro­pean countries at the time. However, the intensification of electoral competition spurred legislators’ demands for additional reforms. The main proposal to limit fraud considered during the second half of the nineteenth c­ entury recommended allowing candidates to appoint representatives to supervise ballot counting. Among the countries included in this study, Belgium, Britain, and France introduced legislation mandating the presence of candidate representatives. While Imperial Germany did not introduce such legislation, the electoral jurisprudence developed during this time also affirmed the importance of candidate representatives as an immediate protection against fraud (Hatschek 1915: 174). The po­liti­cal debates around the introduction of candidate representatives reveal both similarities to and significant differences from the other reforms examined in this study. Let me begin by illustrating the most salient cleavages over the adoption of ­these electoral reforms in France. Considerations about access to resources—­such as access to mayors who acted as voting bureau presidents—­explain the initial cleavage line that came about during fraud reform studied in this book. Demand for electoral reform originated among Socialist deputies who lacked access to t­ hese mayors ( JORF ChamberDoc 1894, no. 336). Legislators associated with Conservative parties such as the Action Libérale also supported proposals to grant candidate representatives extensive powers during the counting of ballots. Legislation to limit fraud stands in contrast to other electoral reforms examined in this study. Considerations about electoral costs are entirely absent in the case of fraud. Given that fraud occurs a­ fter voting has ended, a candidate does not incur any electoral costs associated with this illicit strategy. The po­liti­cal deliberations about t­ hese electoral reforms reflect the absence of such concerns. Radical politicians who competed on the basis of

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 65

programmatic promises had been a pivotal legislative group supporting the adoption of electoral reforms limiting the candidature officielle, vote-­buying, and electoral intimidation. However, Radicals lacked the electoral incentive to support fraud reform due to the low electoral costs associated with this electoral irregularity. The absence of Radical support for reforms introducing candidate representatives was noted by other legislators in the Chamber of Deputies. Commenting on this inconsistency, a Republican deputy noted, “electoral reform concerning the sincerity of the vote was one of the main princi­ples of the Radical party. ­Today, what do we see? Only the Radical party opposes ­these reforms” ( JORF ChamberDeb, 27 May 1907). One can characterize the electoral co­ali­tion that supported the introduction of candidate representatives in France as one of extremes against the center. Legislators situated at the extreme of the po­liti­cal spectrum—­ Socialists and Conservatives—­supported the introduction of candidate representatives. By contrast, centrist Radicals opposed ­these reforms. Republican legislators w ­ ere divided: ­those with po­liti­cal connections to the state apparatus (as mea­sured by their previous experience as mayors) opposed the introduction of candidate representatives, while Republicans lacking such resources supported ­these reforms. Thus, in the absence of electoral costs, the composition of the electoral co­ali­tion that supported fraud reform largely reflects considerations about resource advantages. In contrast to France, Belgium and Britain ­adopted reforms that introduced candidate representatives with relative ease. Britain included candidate representatives as part of the Ballot Act of 1872 (Kinzer 1982; Seymour 1915), with an amendment mandating that candidates should appoint “agents” to supervise electoral operations. Belgium followed Britain’s lead. Belgium introduced provisions allowing candidate representatives in 1877 at the same time as secret voting in a reform package known as the loi Malou. Candidate representatives also known as witnesses (témoins) w ­ ere given the right to monitor the transportation and counting of ballots as well as to challenge the results. In both countries, robust po­liti­cal alternation reduced asymmetries among parties in accessing resources to engage in electoral fraud. At the same time, this alternation mitigated concerns that t­ hese new electoral rules could disproportionately aid one party. In the case of this low resource asymmetry, reforms enjoyed the po­liti­cal support of both parties. As the discussion in this section has suggested, the composition of the legislative co­ali­tions supporting the stringent punishment of corruption varies across reform areas. Th ­ ese differences are conditioned, on the one hand, by the type of resources that are relevant for the par­tic­u­lar irregularity

66 C H A P TER 2

­TABLE 2.2.

The composition of reform co­ali­tions Timing of irregularity

Resources used for electoral malfeasance

Private

Public

Campaigns

Election day

Economic costs: salient

Economic costs: salient

Electoral costs: salient

Electoral costs: salient

Reform co­ali­tion: Cleavage line conditioned by private access to resources (rich versus poor)

Reform co­ali­tion: Facilitated by co­ali­tional realignment due to rising electoral and economic costs

Economic costs: not salient

Economic costs: not salient

Electoral costs: salient

Electoral costs: not salient

Reform co­ali­tion: Facilitated by co­ali­tional realignment due to rising electoral costs

Reform co­ali­tion: “Extremes against the center”

and, on the other, by the relative importance of economic and electoral costs. In ­table 2.2, I summarize the implications of this analy­sis of the composition of co­ali­tions supporting electoral reforms. Let me conclude by discussing some differences between the analy­sis in this chapter and recent approaches in the study of democ­ratization. Redistributive theories of democ­ratization, exemplified in research by Acemoglu and Robinson (2000, 2006) and Boix (2003), derive elites’ preferences for electoral reforms from changes in taxation and wealth re­distribution they expect to occur following changes in the scope of suffrage. According to Boix, elites who own fixed assets are likely to experience the highest tax burden ­after the enfranchisement of low-­income voters. As a result, ­these actors are likely to mount the strongest opposition to electoral reforms. This assumption presupposes a long causal chain, whose connections are difficult, if not impossible, to test empirically. This study has instead attempted to provide an alternative motivation for legislators’ preferences for electoral reforms that stresses the importance of electoral calculations. In this account, legislators are willing to forgo their immediate resource advantage and accept reforms that sanction electoral irregularities if they experience high economic and electoral costs for using illicit strategies or if they face a relative resource disadvantage. ­These electoral considerations are rooted in legislators’ concerns for their immediate electoral advantage and provide a theoretically more compelling

D e m o c r atiz ati o n o f E lec to r a l P r ac ti c e s 67

account of support for reforms than explanations featuring non-­electoral motivations. Taking into account the importance of resources that can be monetized in elections as well as the in­equality in access to such resources allows me to generate dif­fer­ent empirical predictions about the composition of electoral co­ali­tions during the electoral reform pro­cess. In existing redistributive theories of democ­ratization, access to fixed assets such as land is identified as the main source of cleavage over electoral reforms, such that landowners or politicians representing districts with high levels of in­equality w ­ ill always oppose electoral reforms. ­These studies cannot account for the variation in the mix of support and opposition by right-­wing parties documented in this study. As discussed above, Monarchists and other Conservative politicians in France found themselves on both sides of the cleavage around electoral reforms, supporting reforms limiting the pression gouvernementale and the adoption of candidate representatives but opposing vote-­buying reforms. Fi­nally, in contrast to modernization theories, I predict that the relative importance of economic constraints is also likely to vary across reforms. Specifically, I expect such conditions to be more salient in shaping legislators’ considerations about the desirability of reform where irregularities deploy private resources. In ­these scenarios, politicians are more likely to shoulder the economic costs associated with illicit strategies. As a result, changes in economic conditions are more likely to facilitate co­ali­tional realignments in the case of reforms to limit irregularities that involve the use of a candidate’s private resources. Conclusion

In this chapter, I have developed the theoretical building blocks that ­will inform the empirical analy­sis in the remainder of the book. I have formulated a legislator-­centered explanation that seeks to identify the most impor­tant considerations that ­shaped politicians’ demands for reforms to eliminate electoral irregularities. As the nature of resources deployed to engage in irregularities and the types of constraints faced by candidates vary significantly across multiple dimensions of reform, I expect the composition of the electoral co­ali­tion that supports electoral reform to vary across policy areas. The unequal distribution of resources is likely to affect the intensity of the conflict over electoral reforms. I have suggested that the intensity of this conflict and the pattern of po­liti­cal co­ali­tions are likely to vary across countries depending on the nature of party competition. In countries with

68 C H A P TER 2

robust po­liti­cal competition and electoral turnover, the in­equality in the access to resources is low. This creates a stronger incentive for parties to agree on reform and may lead to bipartisan co­ali­tions. The adoption of candidate representatives as a solution for limiting fraud in Belgium and Britain illustrates this po­liti­cal dynamic. In countries with a hegemonic party system, it is much more difficult to assem­ble a po­liti­cal majority that f­ avors electoral reforms. H ­ ere, the decisive f­ actor that facilitates the formation of a po­liti­cal majority in ­favor of reforms is a split of the incumbent party. One additional implication of my analy­sis is that the electoral co­ali­tion that supports the adoption of each electoral reform is highly heterogeneous. Neither redistributive nor economic theories of democ­ratization can explain the variation across reform dimensions in the composition of electoral co­ali­ tions supporting reforms. Such co­ali­tions are likely to include legislators from parties that lack access to resources to engage in par­tic­u­lar irregularities alongside resource-­endowed candidates who demand electoral reforms ­because the economic or electoral costs of using such strategies is too high. Fi­nally, departing from existing explanations of democ­ratization, I suggest that parties on the right—­such as the rightist anti-­Republican parties in France—­may take dif­fer­ent positions across dif­fer­ent reform dimensions. ­Because of their relative resource disadvantage, such parties may support reforms that reduce the use of state resources during elections. At the same time, ­these legislators are likely to oppose reforms that limit the incidence of vote-­buying. The combination of considerations about resource asymmetries, economic costs, and electoral costs can explain the variation in the composition of ­these reform co­ali­tions. The remaining chapters ­will test ­these predictions.

3 Reforming Campaigns REDUCING THE ILLICIT USE OF STATE RESOURCES

In nineteenth-­century campaigns and elections, candidates deployed a variety of state resources. Politicians representing governing parties campaigned using promises of collective or individualized benefits in exchange for electoral support. The rapid increase in the scope of the state during the period and public investments in myriad infrastructure proj­ects opened up ample opportunities to politicize state resources. In addition to monetizing current or promises of f­ uture transfers, candidates also deployed employees of the state as electoral brokers. Policemen, tax collectors, mayors, and employees of local governments distributed electoral materials and ballots, mobilized voters on election day, and demobilized opponents. The use of ­these electoral strategies significantly undermined the princi­ple of the neutrality of the state. In this chapter, I examine efforts to limit electoral strategies that politicized state resources in nineteenth-­century elections and the consequences of such efforts for changes in e­ ither electoral law or jurisprudence. Since the ability to access the administrative apparatus of the state during campaigns created im­mense resource advantages for candidates, it also created power­ful incentives to resist changes to the status quo. As a result, proposals to limit the use of state resources during elections w ­ ere the most contested electoral reforms examined in this study. In contrast to other reforms 69

70 C H A P TER 3

analyzed in this book—­limiting vote-­buying, protecting the secrecy of the vote, and reducing electoral fraud—­reforms to limit the use of state resources during campaigns generated the strongest po­liti­cal re­sis­tance. And even when a new po­liti­cal majority in f­ avor of electoral reforms came about, strong opposition, grounded in the desire to retain this resource advantage, muddled the scope of the resulting policy change. The changes in legislation or jurisprudence to limit the use of state resources ­were remarkably small in scope when compared to the other dimensions of reform examined in this study. I examine ­these questions through a comparative analy­sis of po­liti­cal developments in Third Republic France and Imperial Germany during the period between 1870 and 1914. One finds impor­tant similarities in the politics of ­these electoral reforms across ­these cases. In both France and Germany, relative resource asymmetries among legislators can explain the initial cleavage lines concerning the desirability of electoral reforms. Politicians affiliated with governing parties could access state resources with relative ease and opposed changes to the status quo. In Germany, this initial po­liti­cal majority that opposed electoral reforms included Conservatives, National Liberals, and ­Free Conservatives. In France, the initial electoral co­ali­tion that opposed a change in the status quo included Republican legislators who became the governing majority a­ fter 1877. By contrast, the co­ali­tion that supported electoral reforms was extremely heterogeneous, including resource-­constrained legislators with ideologically distant preferences. In Germany, ­Free Liberals, the Catholic Zentrum, and Social Demo­crats supported the adoption of reforms. In France, the electoral co­ali­tion advocating for a change in electoral law brought together Monarchists and other Conservative candidates alongside Socialists. While vocally opposing the status quo, t­ hese proponents of reform remained a minority and lacked the po­liti­cal strength to push through their preferred outcome. In both France and Germany, elite splits within the incumbent po­liti­cal majorities created opportunities for the formation of a po­liti­cal majority that favored reforms. As a result of ­these elite splits, the use of clientelistic strategies that politicized state resources became increasingly costly for some legislators. Th ­ ese cross-­pressured legislators experienced a conflict between the desire to maintain their resource advantage and the electoral costs they incurred as they deployed state employees as brokers. A new electoral majority that supported reforms emerged as ­these cross-­pressured legislators switched their position. In France, I document the elite split between Radicals and Republican legislators and the rise of programmatic

R e fo r m i n g C a m pa i g n s 71

competition beginning in the first de­cades of the twentieth c­ entury. A po­liti­ cal majority that supported electoral reforms emerged beginning in 1907, ­after politicians competing on the basis of programmatic promises began to endorse electoral reforms. A law limiting electoral corruption was ­adopted in 1913. In Germany, the breakdown of the electoral cartel between National Liberals and Conservatives created a new electoral environment for politicians on the right. The mix of electoral competition and cooperation changed in the aftermath of the elite split. Increased fragmentation heightened competition during the first round of elections. At the same time, the importance of electoral coordination for the formation of electoral alliances in runoffs also increased. Following the elite split, the ability of candidates who politicized state resources to form co­ali­tions during runoffs declined. National Liberal legislators experienced rising electoral costs associated with the use of state resources during campaigns b ­ ecause of their stronger dependence on runoffs. As a result, the preferences of National Liberal legislators t­ oward the use of state resources during campaigns changed during this period. In the two de­cades prior to World War I, National Liberals increasingly joined other parties—­such as Catholics, Social Demo­crats, and ­Free Liberals—in demanding more severe sanctions on candidates who campaigned by deploying state resources. I illustrate this change in the position of National Liberal legislators using evidence from roll-­call votes demanding the invalidation of any electoral victory by a candidate who had engaged in clientelistic practices. The chapter concludes by documenting changes in electoral jurisprudence regarding the use of state resources ­adopted by the German Reichstag prior to World War I. Altogether, t­ hese changes contributed to a ­silent democ­ratization of electoral practices (Hatschek 1920). The Mobilization of State Resources during Campaigns in Nineteenth-­Century Elections

In several Eu­ro­pean countries, the use of state resources for po­liti­cal campaigns originated during a period of electoral authoritarianism. In Second Empire France, the government took its right to intervene in elections for granted (Pilenco 1930). In 1831, Prime Minister Casimir Périer summarized this ­legal princi­ple as follows: “The goverment is not ­going to take a neutral position during the elections and it does not mandate neutrality on the part of the administration ­either” (Charnay 1965: 212). Other politicians at the time argued that the “right of the government to choose candidates is a right

72 C H A P TER 3

written in the law itself which cannot be contested” (Pilenco 1930: 68). In Germany, the administrative law during the period of l­ imited constitutionalism also affirmed that state employees had a “duty of loyalty” (Treuepflicht) ­toward the government and that this duty also extended to electoral periods (Rejewski 1973: 11). Friedrich Graf von Eulenburg, Bismarck’s interior minister, noted in 1862 that state employees’ obligation to serve the administration “cannot be severed during an election.” Von Eulenburg regarded this duty as one of the “basic princi­ples” (Grundpfeiler) of the Prus­sian state (ibid., 63). Discussing his expectations regarding state employees, Otto von Manteuffel, Prus­sia’s prime minister during the 1850s, affirmed his intention to “remove from office mercilessly any state employee who offends the princi­ ple of loyalty, who cannot muster the courage required by their occupation or who takes sides in opposition to the government” (ibid., 41). In France, electoral practices that politicized state resources during campaigns w ­ ere also known as the candidature officielle (official candidacy). The candidature officielle resurfaced as an electoral strategy during the first elections of the Third Republic, at a time of intense confrontation between Republicans and Conservative or Monarchist candidates. In ­these highly contested elections, Monarchists took advantage of their access to the state and deployed state resources heavi­ly. Out of the 531 Conservative candidates competing in the 1877 election, 49 ­were sanctioned as “official candidates” (Voilliot 2005: 228; Mayeur 1973: 41). A memoir of the period recalls the mobilization of prefects, deputy prefects, and the departmental administration in support of Conservative candidates (de Meaux 1905: 362). Due to the massive politicization of the state apparatus, a number of Republican politicians in the Chamber of Deputies called for the immediate invalidation of the results of the election. As a legislator commented on the excesses at the time, “­there was, in ­these elections, an official candidacy admirably or­ga­ nized, orchestrated at the higher echelons of the administration, in other words the prefecture, and descending to the lowest level of the hierarchy, in other words, the humblest city hall of the department” ( JORF ChamberDeb, 7 March 1877). The Republicans, who emerged victorious in 1877, denounced the candidature officielle as the imperial candidacy (candidature impériale), in other words, as a relic of the Second Empire. This represented, however, only a rhetorical positioning and not a decision to forfeit this impor­tant po­liti­cal resource. Republicans understood the im­mense electoral advantage that could be derived from the deployment of state resources during campaigns. Consequently, Republicans did not abandon state resources but continued

R e fo r m i n g C a m pai g n s 73

to use them in a more hidden manner as compared to how they had been used during the Second Empire (Pourcher 1987). Christophe Voilliot, a historian of the Third Republic, refers to this change in electoral reliance on the resources of the state as the “euphemisation” of the official candidacy (2005: 230). Opponents of Republicans also noted such continuity in the use of state resources between the Second Empire and the Third Republic ( JORF ChamberDeb, 18 December 1901). The law guiding the regulation of elections during the Third Republic asserted that promises of goods w ­ ere electoral offenses. Article 38 of the Organic Decree of 2 February 1852 made such strategies punishable by a fine, which could total as much as 5,000 francs, and by a prison sentence of up to two years. Article 39 of the same law also sanctioned the use of threats or vio­lence intended to influence the vote ( JORF Lois et Décrets, 2 February 1852). If an employee of the state was a perpetrator of such offenses, then the punishment was doubled. The most impor­tant weakness of the electoral law remained the abstract formulation of the irregularities that could occur during campaigns. The law singled out promises of employment in the local administration in exchange for the vote as a punishable offense. At the same time, the law was ­silent about myriad other electoral strategies by which candidates could draw on the resources of the state during elections. Proponents of reforms took up the task of listing the additional strategies by which candidates politicized state resources during campaigns. In Imperial Germany, the electoral jurisprudence considered campaign strategies that politicized state resources as violations of the electoral law, as ­these interventions undermined voters’ ability to exercise their vote freely (Hatschek 1920: 209; Hatschek 1915: 550). In contrast to France, the penalization of ­these electoral strategies resulted from a number of decisions made by the commission verifying electoral integrity in the Reichstag but could not be traced to a provision of the electoral law. Julius Hatschek, a leading expert of electoral law of the period, outlined the interventions by state employees that qualified as electoral offenses (Wahldelikt). Th ­ ese included: A generally strong partiality of the government for a par­tic­u­lar candidacy, the so-­called “official candidacy,” the calling of a meeting of voters by an official person, the publication of an electoral call in an official publication, even if the latter was only an indirect announcement of a candidacy, the sending of ballots by official persons, the distribution of ballots by employees of the state, the illegal bans of electoral meetings, the illegal confiscation of ballots and the illegal pressuring of the ­owners

74 C H A P TER 3

of certain public spaces not to make the space available to par­tic­u­lar parties. (Hatschek 1920: 552) Despite t­ hese restrictions, candidates continued to use state resources with impunity. In the following section I describe the multitude of ways in which candidates politicized the resources of the state. The reliance on and use of state resources during campaigns differed significantly across nineteenth-­century Eu­ro­pean countries. Such electoral strategies w ­ ere relatively uncommon in British elections. Among the races investigated by the House of Commons for illicit practices between 1867 and 1883, 80 ­percent of reports mention the incidence of vote-­buying or treating. By contrast, less than 10 ­percent of the reports mention candidates who deployed state resources for electoral gain (Kasara and Mares 2016). In Belgium, information documenting candidates’ electoral strategies is relatively sparse. One can draw on several investigations of electoral corruption conducted by the Belgian parliament in 1859 and 1863 to piece together an account of candidates’ electoral strategies (Pagnoul, Vleeschouwers, and Bourmanne 2000). The most common irregularity reported at the time involves clerical influence. Similar to Britain, the deployment of state resources during campaigns was relatively rare. The situation was entirely dif­fer­ent in Germany and France. In both countries, state resources ­were heavi­ly used during campaigns. Discussing the electoral use of state resources in Germany, Robert Arsenscheck, a historian of German electoral practices, noted that “such practices w ­ ere not the exception, but rather the rule” (2003: 173). In figure 3.1, I draw on data from parliamentary reports of electoral irregularities in both France and Germany to pre­sent a systematic account of the relative incidence of complaints about the use of state resources in elections. The descriptive pattern presented in figure 3.1 largely confirms the importance of this strategy in the menu of electoral irregularities. The sources of the French data are reports of the commissions de vérification des pouvoirs, the parliamentary commissions established in the French Chamber of Deputies ­after each election to investigate the validity of each parliamentary race. Using multiple teams of research assistants, I coded all reports for elections held between 1881 and 1914. In each report, I identified complaints raised by one of the parties in the race alleging the use of state resources by one or several opponents in the race. For Germany, the data source is the reports of the commission investigating electoral irregularities, the Wahlprüfungskommission. ­Here, I selected a random sample of

Share of electoral reports invoking candidature officielle Share of fraud reports invoking use of state resources

France 100

80

60

40

20

0

100

1881

1885

1889

1893

1898

1902

1906

1910

1914

Germany

80

60

40

20

0

1871 1874 1877 1878 1881 1884 1887 1890 1893 1898 1903 1907 1912

The relative incidence of electoral strategies politicizing state resources in French and German elections. Sources: France: JORF ChamberDeb (vari­ous years); Rapports des commissions de vérification des pouvoirs; Germany: StBer DtRt (vari­ous years). FIGURE 3.1.

76 C H A P TER 3

120 election reports distributed across elections held between 1874 and 1912. As I did for France, I coded all incidents detailing the use of state resources during campaigns. Due to the smaller sample of coded election reports, the estimates of the incidence of contested elections in Germany presented in figure 3.1 have a higher mea­sure­ment error. In both France and Germany, candidates commonly politicized state resources during elections. In France, on average, 70 ­percent of election reports contest an election b ­ ecause a candidate politicized state resources. For Germany, the average share of contested elections where candidates deployed state resources is slightly lower at around 60 ­percent. This stands in contrast to other electoral irregularities, such as vote-­buying and fraud, which w ­ ere much less frequently used in German elections. The reports of ­these contested elections pre­sent extremely rich evidence that allows us to reconstruct the variety of electoral brokers deployed by the candidates and the variety of policies and programs brokers used to mobilize voters and influence their electoral choice. Consider first the hierarchy of brokers one finds in French elections. Government ministers ­were situated at the top of this hierarchy. In a letter addressed to prefects during the highly contested election of 1877, the minister of interior justified his intervention in unambiguous terms: “The government has not only the right, but has also the obligation to inform voters about the candidates who support and the candidates who oppose its policy” (Voilliot 2005: 229; Challeton 1891: 262). Similarly, a minister of public works during the Third Republic characterized the support given by the government to candidates as a signal: “The government can signal his friends and enemies. By making their preferences known to voters, they are strictly within the bounds of the law, moreover, they fulfill an obligation” (Weil 1895: 229). Republican candidates often started their campaign by displaying their connections to ­these high-­ranking politicians. Boniface Boni de Castellane, a Republican candidate competing in Basses Alpes in 1902, introduced himself to voters in the local newspaper as an “official candidate” supported by Pierre Waldeck-­Rousseau, the prime minister at the time ( JORF ChamberDeb, 7 September 1902). On other occasions, candidates displayed their po­liti­cal connections in the form of notes of support signed by senators and other prominent Republican politicians that ­were published in local newspapers (ibid., 7 December 1893). Candidates’ public display of their po­liti­cal ties was designed to activate voters’ expectations of ­future transfers in return for supporting the governmental candidate (Pilenco 1930). Such f­ avors included funds for the construction of roads and bridges, funds for local infrastructure

R e fo r m i n g C a m pai g n s 77

proj­ects, or relief to victims of natu­ral disasters ( JORF ChamberDeb, 12 December 1885). Candidates who enjoyed the support of the government ­were sometimes referred to as “living bud­gets”: “bud­gets of public works, for religious affairs, bud­gets for lakes and waterfalls and even bud­gets for fine arts” (ibid., 14 December 1889). This display of po­liti­cal connections gave a tremendous boost to official government candidates relative to their opponents. One candidate who enjoyed the support of the government campaigned by warning voters that “if you vote for my opponent, you ­won’t be able to ask for anything when you w ­ ill need something for the locality” ( JORF ChamberDeb, 3 June 1910). In 1909, the Republican Party machine in Aveyron mocked candidates who did not have the backing of the government in several local newspapers. ­These publications drew a contrast between Étienne Fournol, a Republican candidate, portrayed as a “friend of the government,” and his opponent. One entire page of this newspaper contained a ­table with only two columns. One of the columns reminded voters of the f­ avors enjoyed by a deputy that was agreeable to the government: subventions for school buildings, the creation of jobs as teachers, the creation of post offices, the allocation of benefits to hospices, the creation of bridges, the offers of books to the school. The other column had to enumerate the f­ avors obtained by the candidate of the opposition. It contained only one word, written in large letters. Nothing. This was followed by the remark: Long live Étienne Fournol, the friend of the government. (ibid., 20 June 1909) The display of connections was often followed by the delivery of subsidies (Pourcher 1987: 210). Candidates strategically timed the delivery of such subsidies to allow them to take electoral credit. In Gironde, mayors distributed disaster relief sixteen months ­after a flood. The benefits ­were accompanied by a letter that stated, “This generous donation is attributable to numerous and insistent approaches on your behalf made by Mr. Abel Surchamp, your devoted deputy” ( JORF ChamberDeb, 30 June 1902). In 1902, the administration delivered disaster relief in Ardèche only on the eve of the vote (ibid., 6 June 1902). A mayor in a locality in Haute-­Garonne prefaced the distribution of disaster relief funds with the statement, “You all need to vote for Bepmale: you owe him all the benefits that have been distributed” (ibid., 1 June 1898). This politicized distribution of subsidies by ministries was one of the most impor­tant electoral strategies used by successive Republican governments. Such distributions are repeatedly documented in reports of electoral

78 C H A P TER 3

irregularities. Unfortunately, we cannot conduct a systematic quantitative analy­sis of this strategy, as the full-­time series rec­ords of ­these subsidies have not been preserved. The relevant archival series of the Archives Nationales is interrupted between 1864 and 1944 (Voilliot 2005: 237). As Christophe Voilliot comments on this break in documentation, “archival traces of such transactions have rarely been preserved, as Ministers of the Interior of the Third Republic are fascinated with the purifying fire” (ibid., 238). Prefects and deputy prefects occupied an intermediate position in the chain of brokers mobilized by Republican candidates during the Third Republic between the central echelons of the administration and individual mayors who coordinated electoral support at the local level. Historians of the period have noted, however, that electoral intervention of prefects during the Third Republic was less intense and less vis­ib ­ le as compared to ­earlier po­liti­cal regimes, such as the Second Empire. Paul de Cassagnac, a conservative politician in the National Assembly, commented on this more subdued intervention: “I ­will not deny that in certain circumstances, the prefects of the Empire had pushed their zeal and intervention too far. . . . ​ ­Under the Empire, the official candidacy was clear-­cut, was candid, was loyal. It was displayed without hy­poc­risy on the white poster which did not deceive anybody. By contrast, during the Third Republic, posters of all colors hide it, imposing it on servile voters” ( JORF ChamberDeb, 17 December 1901). The first responsibility of prefects during campaigns was to select and endorse the “correct” Republican candidate. In 1893 in Nièvre, the prefect or­ga­nized a “Republican Congress” to select one candidate to compete against Jules Jaluzot, the incumbent Conservative representative of the district (Voilliot 2005: 237). Jacques Cata­logne, a Republican candidate competing in Basses Pyrénées, was described as a candidate who had “grown up in the hot­house of the prefect” ( JORF ChamberDeb, 7 September 1902). To endorse candidates, prefects accompanied politicians during campaigns and attended official events with them. Opponents denounced t­ hese rallies as opportunities for the administration to “shower decorations” and sprinkle “kilo­meters of purple ribbon” (ibid., 24 June 1902, 7 June 1906). In addition to supporting loyal candidates, prefects tried to create hurdles for opposition candidates. During the 1906 election in Oise, for example, the prefect threatened the opposition candidate: “I’ll do every­thing I can to thwart your candidacy, as I represent the administration” (ibid., 26 June 1906). While prefects rewarded loyal mayors, they also sanctioned disloyal mayors or t­ hose who exerted insufficient effort. During the 1885 election in Charente-­Inférieure, the prefect removed from office mayors and employees

R e fo r m i n g C a m pa i g n s 79

of the local administration who expressed anti-­Republican positions in their publications ( JORF ChamberDeb, 28 November 1885). Similar incidents ­were reported during the 1898 elections in Corsica and Aveyron, where the prefect punished disloyal mayors by dismissing them (ibid., 27 June 1898). During the 1898 election in Nyons (Drôme), the prefect suspended two mayors for intervening on behalf of a Conservative candidate, Arthur de Suarès d’Aulan (ibid., 14 November 1889, 19 November 1889). In Germany, the hierarchy of brokers was structured differently than in France. One impor­tant f­ actor that accounted for this difference was the federal nature of the German Empire. As a result of this constitutional design, we do not find a clear line of command connecting the central government, the regional authorities, and candidates’ campaigns in the districts (Mares 2018). The federal government also controlled only a very small share of revenues that could be directed to individual electoral districts to reward or incentivize electoral be­hav­ior. In German elections, the hierarchy of brokers originated with ministries of dif­fer­ent regions who mobilized state employees located at dif­fer­ent levels of the government. The most significant ministry that coordinated electoral operations was the Prus­sian Ministry of Interior (Kehr 1965; Arsenscheck 2003). The period of greatest administrative politicization occurred between 1881 and 1888, a time when Robert von Puttkamer served as minister of interior. As historians of the period noted, Puttkamer presided over a well-­designed system of sanctions and rewards to incentivize electoral intervention by members of the bureaucracy. Th ­ ese incentives included, among other t­ hings, promotions for diligent employees as well as threats of transfers or other disciplinary mea­sures (Arsenscheck 2003: 194). Describing Puttkamer’s strategy, Kühne noted that “mid-­level officials who displayed indolence during elections ­were ­under regular attack from above. ­After all, ­these employees played a critical role in affecting w ­ hether the armies of low-­level employees could be set in motion as multiplicators or as voters in a district” (1994: 61). The Landrat, the head of an administrative district—­known in some provinces as Kreisdirektor—­was a pivotal broker in German elections. The main administrative responsibilities of Landräte included tax collection, tax assessment, and military conscription. Landräte provided crucial electoral support to candidates especially during campaigns. First and most importantly, Landräte ­were responsible for promoting the central electoral purpose b ­ ehind which the entire state apparatus was or­ga­nized. During the 1907 elections in Hannover’s first district, the Landrat announced in a meeting of mayors that “he had received an order from higher levels

80 C H A P TER 3

that the [representatives of ethnic minorities such as] Guelphs have to dis­ appear from the Reichstag” (StBer DtRt 1907, encl. no. 801). He followed this announcement with efforts to pressure local-­level employees to deter Guelph candidates from reaching voters. In advance of the 1878 election, the Landrat in Oppeln’s fifth district met with mayors and ordered them to mobilize teachers and policemen to vote for “candidates loyal to the Empire,” in other words, Conservative candidates. At the same time, the Landrat threatened localities considering voting for the Catholic Zentrum with a reduction in public support (StBer DtRt 1883, encl. no. 1152). The election in Elbing in 1884 was contested between Bernhard von Puttkammer-­Plauth, a Conservative candidate, and a ­Free Liberal challenger. Members of the electoral commission characterized this election as the clearest example of an official candidacy similar to t­ hose encountered in neighboring France. The Landräte in Elbing and Marienburg have provided massive support to Puttkammer. They have recommended his candidacy in public meetings, have accompanied Puttkammer on trips in the district, and have participated in his electoral meetings. With a handshake, they obtained the commitment of agricultural workers to support his candidacy. Even electoral officers have distributed ballots with Puttkammer’s name and have acted in similar ways as the Landräte. (StBer DtRt 1884, encl. no. 180) In both France and Germany, state employees at the local level ­were the most instrumental players in day-­to-­day electoral operations. In France, the candidature officielle invariably went through the mayoral office. Mayors provided multiple electoral ser­vices to candidates, which included, among other t­ hings, lists of voters’ addresses, access to the governmental offices that could be used as campaign headquarters, and transportation for voters on election day (Charnay 1965: 230). Mayors also mobilized state employees to interact with voters. Typical brokers included teachers ( JORF ChamberDeb, 20 June 1898), tax collectors (ibid., 23 June 1902, 7 September 1902), mailmen (ibid., 23 June 1910), justices of the peace (juges de paix) (ibid., 12 December 1885, 7 September 1902), road-­menders (cantônniers) (ibid., 28 March 1890), village guards (gardes champêtres) (ibid., 30 January 1890), policemen (ibid., 22 November 1898), and employees of state-­ owned enterprises (ibid., 23 November 1889). As ­these examples illustrate, brokers ­were selected from all levels of governmental bureaucracy and interacted with voters using a variety of material resources.

R e fo r m i n g C a m pa i g n s 81

Employees of the local administration distributed election materials and ballots for candidates and mobilized voters at rallies ( JORF ChamberDeb, 12 February 1889, 30 January 1890, 21 November 1893). The most valuable inducement state employees offered voters in exchange for their electoral support was employment. Mayors in Corsica routinely offered positions in the tax administration or public works (such as road-­building) in exchange for electoral support (ibid., 8 May 1898). To monitor voters’ choices, mayors encouraged voters to submit ballots with internally recognizable signs, which ­were known as bulletins-­à-­clef. Such ballots included additional handwritten indications that allowed presidents of voting places to identify individual voters (ibid., 29 March 1890). The alternative strategy of electoral mobilization involved the use of coercion—­threats to withdraw valuable benefits from voters who supported the “wrong” candidate. In Landes, a department in southwestern France, mayors campaigned accompanied by the village guard (garde champêtre) and threatened to cut poor relief to voters who d ­ idn’t support a Republican candidate ( JORF ChamberDeb, 24 June 1902). In a locality in the Charente Inférieure, the mayor threatened to close the parish workshops (ateliers communaux) if the Republican candidate was not elected (ibid., 10 June 1898). Such threats ­were commonly reported in French elections and continued as late as 1910 (ibid., 23 June 1910). In Germany, anti-­poverty programs ­were not exploited electorally, as recipients of social assistance benefits ­were not allowed to vote. As a result, candidates and their brokers politicized taxes rather than transfers. During the 1877 election in Altona, tax collectors threatened voters with higher city taxes if they voted for the Social Demo­cratic candidate (StBer DtRt 1877, encl. no. 11). During same election candidates competing in Nord Dithmarschen, a district in Schleswig Hohlstein, deployed tax collectors to mobilize voters. The same threats of higher taxes ­were used (ibid., encl. no. 85). Similar to France, in Germany, mayors w ­ ere an impor­tant node in the network of state employees that w ­ ere mobilized during elections. They would coordinate a governmental candidate’s entire campaign in a locality. As such the playing field among candidates was never a level one. Rather, elections ­were strug­gles between the government and individual candidates who lacked this support. Mayors introduced candidates to voters, distributed electoral material, and mobilized voters on the day of the vote. During the 1878 election in Sorau, the mayor recommended the conservative Schön as the “candidate who represents the intentions and desires of the government in the best way” (StBer DtRt 1878, encl. no. 167). During the same

82 C H A P TER 3

election in Osterode, a district in Prus­sia, mayors distributed ballots for a Conservative candidate while asking voters to burn the ballots distributed by the challenger (ibid., encl. no. 16). Mayors complemented t­ hese strategies with threats of punishment for noncompliant voters. The mayor of Harburg, a locality in Hannover, told voters that they would be employed as forced laborers in road construction if they did not support the Conservative candidate, Graf Corte (ibid., encl. no. 67). Mayors provided impor­tant logistical support to candidates during campaigns. This included material resources and access to a variety of additional brokers. In Sachsen Altenburg, mayors gave Herrmann von Blödau, a Conservative candidate, access to the official list of voters but did not grant similar permission to the Social Demo­cratic candidate (StBer DtRt 1903, encl. no. 193). Mayors also facilitated candidates’ access to other employees of the local administration. ­These included postal workers, school inspectors, department heads, heads of localities and communes, and nightwatchmen. Eugen Richter, a prominent ­Free Liberal politician and a member of the electoral commission of the Reichstag, compared the Prus­sian bureaucratic apparatus at election time to a hydra: new heads sprang up as the old heads ­were chopped off (StBer DtRt 1878, encl. no. 16). The electoral strategies of state employees in Germany ­were much more coercive than the strategies in neighboring France. One tactic used frequently was to prevent opposition candidates from distributing ballots or other electoral materials. Consider some examples from an extensive repertoire of such repressive interventions. During the 1907 elections in Wahrenbrück and Liebenswerda, two localities in Merseburg, policemen forcibly prevented Social Demo­cratic ballot con­vey­ors from distributing their ballots, ripping the bags from their hands. Afterward, the mayor fined ­these con­vey­ors (StBer DtRt 1909, encl. no. 738). A report of electoral irregularities that occurred during the 1891 election in Klein Roris, a locality in Potsdam’s tenth district, depicts similar events. A witness described the efforts of the mayor of the locality to prevent Social Demo­cratic ballot con­vey­ors from entering the village as follows: The mayor forcefully ripped the ballots and copies of the Berliner Volksblatt from the hands of Kaiser, the Social Demo­cratic ballot con­vey­or. The latter protested and showed the mayor a copy of the electoral law, but the mayor ripped this document to pieces. Then, he stuck his hand in Kaiser’s pocket, removed the ballots, and also ripped t­ hose to pieces. He threw Kaiser to the ground and then threw him down the staircase onto

R e fo r m i n g C a m pa i g n s 83

the street. At the end, he shouted: “Let the dogs loose.” Somebody released two dogs, and they attacked Kaiser. (StBer DtRt 1892, encl. no. 165) Such events occurred not only in Prus­sia but also in the southern German states. In Württemberg’s first district during the 1878 election, policemen seized ballots from Social Demo­cratic ballot con­vey­ors, saying, “we already have ballots in the city hall and ­don’t need ballots from riffraff ” (StBer DtRt 1878, encl. no. 74). During the 1912 election in Baden’s seventh district, two policemen in uniform detained ballot con­vey­ors and ripped from their hands the electoral materials for the Social Demo­cratic candidate (StBer DtRt 1912, encl. no. 753). This section has demonstrated that state resources ­were used extensively during electoral campaigns in both French and German elections. In both countries, po­liti­cal candidates aligned with the incumbent electoral party or co­ali­tion could take advantage of a large network of state employees and deploy them as brokers. ­These brokers fulfilled impor­tant electoral responsibilities on behalf of candidates, distributing informational materials and ballots, mobilizing voters, and demobilizing opponents. I have also documented two differences across ­these two countries in the use of state resources. The use of patronage, the promise of transfers directed to the locality, was much greater in France than in Germany. The promise of transfers directed to the municipality was a central implicit promise in all French elections but was virtually absent in Germany. Second, the use of coercion—­and sometimes vio­lence—in the interaction between brokers and voters appears to have been greater in Germany than in France (Mares 2018).

The Adoption of Reforms Limiting the Electoral Use of State Resources FRANCE: SANCTIONING THE PRESSION GOUVERNEMENTALE

At the moment when we are discussing this legislation, in ­every constituency, we can find prefects who have received 20,000, 30,000, or 40,000 francs in the form of subsidies from the state as the immediate result of an intervention by a par­tic­u­lar candidate. It appears to me that this is a form of corruption. However, this is not corruption that involves the resources and money of the individual candidate but a form of corruption that involves every­body’s money, the money of taxpayers ( JORF ChamberDeb, 17 March 1910).

84 C H A P TER 3

As the previous examples have documented, the politicization of state resources during elections was a multifaceted phenomenon involving myriad strategies. Given the complexity of ­these interactions and the variety of brokers, the electoral law of the Third Republic was ill-­equipped to design appropriate punishments for t­ hese strategies. The legislation that was in place at the time (the Organic Decree of February 1852) considered offers of employment in exchange for the vote and other promises as punishable offenses ( JORF Lois et Décrets, 1852). However, this provision was rarely enforced (Pierre 1893: 82; Voilliot 2005: 212). The leading l­ egal treatise on electoral law of the period, Eugène Pierre’s Traité de droit politique, électoral et parlementaire, summarized the jurisprudence on questions of the candidature officielle as follows: “It has been established that an election cannot be voided only as a result of the intervention or by the patronage of the administration and the presence of an official candidacy if it can be proven that neither pressure nor vio­lence was pre­sent to modify the outcome of the vote” (1893: 82). As late as 17 March 1910, both Conservative and Socialist legislators in the Chamber noted that “a member of the majority has never been invalidated.” During the period 1900–1914, the French parliament considered a variety of proposals designed to remedy this shortcoming in the electoral law. One set of proposals attempted to define more precisely the range of electoral strategies that ­violated the provisions of the existing law. Reformers hoped that such clarifications could prevent legislators in the Chamber from turning a blind eye to blatant instances of electoral corruption. The second ave­ nue of reform involved changes to the procedure leading to the invalidation of elections. Reformers proposed empowering additional actors—­such as courts or juries—to decide w ­ hether an election should be invalidated. Th ­ ese reforms could break the grip the Republican majority held over the definition of illicit strategies and could lead to the invalidation of races where candidates politicized state resources. ­These proposals for reform date to the first de­cade of the Third Republic and originated as responses to reports that investigated irregularities ( JORF ChamberDoc 1877, no. 526). At the beginning, t­ hese recommendations included only vague exhortations to end governmental pressure without specifying a set of credible punishments for ­these strategies. To illustrate this logic, consider one of the earliest amendments to the electoral law formulated by Gustave Cuneo d’Ornano, a politician from Charente affiliated with Appel au Peuple. D’Ornano’s proposal was a long-­winded tirade against the candidature officielle ( JORF ChamberDoc 1890, no. 843). Invoking the

R e fo r m i n g C a m pa i g n s 85

princi­ple that the “government does not have the right to use the weight of its entire force, which is im­mense, on the conscience of the voters,” D’Ornano recommended introducing a “moral punishment on state employees of all ranks who had abused their authority in the interests of an administrative candidacy.” He also proposed punishing candidates who engaged in t­ hese illicit be­hav­iors with the loss of the right to run for office for a period of five years (ibid.). Unsurprisingly, proposals sponsored by Conservative legislators such as D’Ornano targeted only campaign irregularities that involved state resources and avoided taking a position on electoral irregularities that involved private resources, such as vote-­buying or treating. Conservatives’ positions on questions of electoral reforms clearly reflected considerations about relative resource advantage. By contrast, the recommendations for reform formulated by Socialist deputies targeted both irregularities. Socialists w ­ ere the most resource-­constrained candidates seeking election to the French Chamber of Deputies. They competed against ­either Conservative candidates who could deploy private resources or Republican candidates who could politicize the resources of the state. Commenting on this resource disadvantage, Paul Constans, a Socialist deputy from Allier, noted: “We are continuously obliged to defend ourselves against the official candidacies and have against us all the forces of capitalism and all the forces of power” ( JORF ChamberDeb, 17 March 1910). In a fiery intervention in the Chamber of Deputies, Jean Jaurès, the leader of the Socialist faction in the Chamber, urged other legislators to support proposals to limit the ability of candidates to use state resources during elections. Jaurès linked the Socialist demand for ­these reforms to the relative resource disadvantage of his party. We have no vocation for the antechambers of prefects. (Applause on the far left.) But we have the right to say that if the candidature officielle is everywhere blameworthy and reprehensible—­and it is everywhere so—it is especially so when it is used against this vanguard party. Having never accessed power before, this party has not been able to create a clientèle among civil servants, and it has neither the support of the church nor the support of money. (ibid., 7 December 1893) Beginning in 1900, the recommendations for electoral reform became much more specific. The Chamber of Deputies had taken up deliberations about electoral reforms for the first time only briefly in 1898 and returned to the question in 1902. René Viviani, a Socialist from Seine, and Jean Odilon-­ Barrot, an in­de­pen­dent from Ardèche, cosponsored the proposals discussed

86 C H A P TER 3

on ­these occasions. Republican legislators seized this opportunity to set the agenda of deliberations in the Chamber and preempt demands for reform that involved dramatic changes to the status quo. As part of this defensive strategy, Republicans advanced a series of proposals that consolidated the existing legislation without proposing additional sanctions for the use of state resources. By contrast, opposition candidates on the right and the left of the po­liti­cal spectrum formulated a variety of amendments listing dif­ fer­ent interventions of the state at elections and proposing vari­ous sanctions for t­ hese incidents. One can summarize the strategy as “reform by enumeration.” In this section, I provide a more systematic analy­sis of the correlates of support for ­these amendments to provide more severe sanctions for vari­ous electoral interventions by state employees. During the early discussion of ­these proposals, the cleavage line in the Chamber can be predicted by the relative resource advantage of politicians. Republican and Radical legislators overwhelmingly voted in support of the status quo. By contrast, resource-­ constrained Conservative legislators strongly supported all amendments for reform. At the time, considerations about the electoral costs of illicit strategies ­were insignificant in explaining demand for electoral reforms. In 1902, an overwhelming majority in the Chamber voted to oppose amendments imposing harsher punishments for electoral interventions by state employees. ­Table 3.1 lists the amendments to the electoral law ­under discussion in the French Chamber in 1902. The sponsors of ­these initiatives believed that the existing legislation did not go far enough in punishing such offenses but only “platonically condemned the candidature officielle.” As Alphonse Gourd, a proponent of one of ­these amendments, remarked, “every­body complains about pression officielle, yet they refuse to take the necessary mea­sures to assure its suppression” ( JORF ChamberDeb, 26 March 1902). To correct the imprecision of the electoral law, ­these proposals listed recurrent electoral strategies and proposed punishments for candidates who engaged in such practices. The sponsors of ­these proposals ­were affiliated with parties spanning the entire po­liti­cal landscape at the time. ­These included deputies affiliated with the Catholic Action Libérale (Adam Régis de l’Estourbeillon), Socialists (Albert Chiché), and even Republicans (Alphonse Gourd). The main justification for t­ hese amendments was that the electoral law failed to sanction a vast number of campaign strategies that w ­ ere extremely common at the time. ­These included interventions by “prefects or deputy prefects, urging ­either mayors or influential voters to vote for the candidate of the government. The teachers become agents of the most active electoral

R e fo r m i n g C a m pa i g n s 87

­TABLE 3.1.

Proposals to limit the candidature officielle considered by the French Chamber of Deputies in 1902

Sponsor of Amendment

Proposal

Result

Adam Régis de l’Estourbeillon

“The following acts ­will be considered as acts of corruption. All ministers or employees of the state who offer ­favors or exceptional donations during the electoral period and the month that precedes it based on the demand of a candidate or incumbent politician.”

Defeated on 26 March 1902

Albert Chiché

“The punishments ­will be doubled if the offenses described above have been committed by a minister.”

Defeated on 26 March 1902

Alphonse Gourd

“The civil servants and employees of the state, departments and communes who . . . ​have influenced ­either directly or by using a third party have influenced or attempted to influence the vote of one or multiple voters ­will be punished with a prison sentence of at least one year and a fine between 100 and 500 francs. If this is a repeated offense, the punishment ­will be doubled.”

Defeated on 29 March 1902

Source: JORF ChamberDeb 1902.

propaganda. Th ­ ese are facts that the proj­ect does not yet sanction. It should sanction ­these” ( JORF ChamberDeb, 29 March 1902). The proponents of ­these amendments questioned w ­ hether the intention of the Republican majority in the Chamber to reduce electoral corruption was genuine. If lawmakers sincerely embraced the goal of reducing electoral intervention by state employees, then it was imperative to adopt the additional amendments. Without ­these additional clarifications, the electoral law would continue to enable all forms of official pressure to operate with impunity (ibid.). The legislators who proposed ­these amendments believed that the use of state resources during campaigns was qualitatively dif­fer­ent from and normatively more undesirable than strategies that politicized private resources. Régis de l’Estourbeillon, a legislator affiliated with Action Libérale, said that electoral promises made by ministers are the most “treacherous and dangerous forms of corruption” ( JORF ChamberDeb, 26 March 1902). One impor­tant difference between electoral intervention by private brokers and state employees’ involvement was the authoritative nature of the former.

88 C H A P TER 3

As Régis de l’Estourbeillon argued, all interventions of public employees have an authoritative nature and are likely to elicit greater voter compliance (ibid.). The difference in the severity of ­these interventions necessitated more stringent punishments for interventions by state employees as compared to ­those by private actors. The proposals submitted in the Chamber at the time recommended doubling the sanctions for interventions by state employees relative to acts of corruption that involved private brokers. During the 1902 discussion of ­these proposals, prominent Republican politicians rejected ­these additional amendments to the electoral law. The president of the Commission of Universal Suffrage of the Chamber at the time opposed t­ hese amendments and acknowledged that the commission was only interested in providing a definition of corruption that was “not excessive.” Pierre Waldeck-­Rousseau, prime minister from 1899 to 1902, represented the position of the government during the debates in the Chamber. In this capacity, Waldeck-­Rousseau expressed his willingness to support a law that synthesized existing provisions without taking additional risky steps. At the same time, he rejected any additional proposals that sought to “regulate the actions of state employees.” To deflect demands for ­these additional reforms, Waldeck-­Rousseau argued that ­these provisions ­were already covered by other laws, such as the law on the press that had been ­adopted in 1881. The cleavage lines over ­these electoral reforms in the Senate mirrored the divisions in the Chamber and can be explained by relative resource advantages. On the floor of the Senate, the strongest proponents of electoral reforms to limit the use of state resources ­were Monarchists. Several senators affiliated with Monarchist parliamentary groups supported electoral reforms limiting administrative pressure and “a stronger punishment of all state employees and their accomplices” ( JORF SenateDeb, 16 January 1906). Similar to Conservative legislators in the Chamber, ­these senators argued that the law did not go far enough in punishing the candidature officielle. One unintended consequence of French electoral jurisprudence was that opposition candidates ­were sanctioned more harshly than the candidates affiliated with the governing majority. Gustave de Lamarzelle, a senator from Morbihan and one of the prominent Conservative speakers in the Senate, characterized this asymmetric punishment as follows: Every­body knows that the official candidacy exists. . . . ​The government does not hesitate to tell voters of each department: if you vote for the opposition candidates, you w ­ ill not be able to count on anything e­ lse, not on a road, not on a railway line. . . . ​According to you, this is not

R e fo r m i n g C a m pa i g n s 89

corruption. On the other hand, if the candidate offers a meal to his voters during an electoral campaign, the law ­will be applied and he ­will be punished. You label as corrupt this insignificant act which takes place ­every time, while the law is not applicable for scandalous, incredible acts of corruption. (ibid.) During its debates, the Senate advanced a compromise solution that extended the definition of punishable offenses to include interventions that politicized state resources. This article of the law, formulated by the Senate, read as follows: “Whoever has offered gifts, promises of gifts or administrative f­ avors, e­ ither to a commune or to any collectivity of citizens in order to influence the vote of a constituency or a fraction of this constituency, w ­ ill be punished by a prison sentence between three months and two years and by a fine between five hundred and five thousand francs.” This decision originated in the commission of the Senate presided over by Eugène Lintilhac, a Radical senator from Cantal. In its report, the commission noted that “collective corruption was as immoral and even more dangerous than individual corruption and that it has to be punished more severely” ( JORF SenateDeb, 18 January 1906). Conservative senators, who considered such collective promises as “the most significant risk for the functioning of universal suffrage,” also supported such proposals (ibid., 23 January 1906). The solution proposed by the Senate responded to the concerns raised by Conservatives who demanded symmetry in the ways in which the electoral law viewed illicit strategies using private and public resources. The amendment listed both types of irregularities and proposed similar sanctions for both. It also recognized the importance of “collective corruption,” or what the current lit­er­a­ture on electoral clientelism would call “patronage” (Stokes et al. 2015). Legislators in the Chamber supported the amendment; several concurred that such “collective corruption” was the most severe and most shameless form of corruption ( JORF ChamberDeb, 21 March 1910). The amendment was approved by the Chamber on 18 March 1910 and included in the electoral law a­ dopted in France in 1914, the Law Repressing Acts of Corruption in the Organ­ization of Elections ( JORF Lois et Décrets, 1 April 1914, Article 3). I w ­ ill delve more deeply into the analy­sis of the 1910 vote in the Chamber in an effort to identify the predictors of the support for ­these reforms. In addition to sanctioning strategies that enlisted state resources during campaigns, legislators also submitted a variety of proposals that attempted

90 C H A P TER 3

to modify the procedures for invalidating elections. Reformers hoped that ­these additional players would be more willing to sanction the use of state resources during campaigns and thus contribute to a change in jurisprudence. One encounters in the Chamber documents a variety of such proposals that e­ ither complemented the parliamentary review of elections with a judicial review or that minimized the role of the parliamentary review altogether. During the 1890s, the proposals by Jean Odilon-­Barrot and René Viviani recommended modifying the pro­cess of investigating irregularities by allowing the Ministry of Justice to appoint judges to investigate contested elections ( JORF ChamberDoc 1890, no. 2749, Article 5). A proposal introduced in 1902 recommended allowing candidates who had been prejudiced by acts of corruption to directly subpoena employees of the state. Paul Bertrand, a legislator from Marne, was the strongest supporter of this proposal and contended that the existing legislation created only ­limited opportunities for candidates to challenge opponents who had politicized state resources. Bertrand hoped that greater opportunities for ­legal contestation would increase the public visibility of illicit strategies and reduce their incidence ( JORF ChamberDeb, 26 March 1902). An additional proposal recommended introducing a separate judicial investigation (instruction judiciaire) tasked with assessing allegations of corruption before the Chamber of Deputies de­cided on the validity of an election. This proposal was cosponsored by two legislators situated at po­liti­cal extremes in the Chamber, Pierre Richard, a Socialist from Seine, and Paul Henri Ferrette, a legislator affiliated with the conservative Action Libérale. The cosponsors of this amendment justified it as striking a ­middle ground in the pro­cess of investigating fraud. The proposal maintained the influence of both legislative and judicial branches during ­these investigations without granting excessive authority to ­either actor ( JORF ChamberDeb, 26 March 1902). During the 1910 discussion of ­these reforms, Louis Ollivier, a legislator affiliated with the Catholic Action Libérale, submitted to the Chamber of Deputies on 21 March an additional amendment recommending a change in the procedure of investigating electoral irregularities. This amendment proposed that all offenses (délits) ­either anticipated or punished by the existing law be submitted to a jury. Ollivier invoked the l­ egal pre­ce­dent whereby juries could decide on cases of defamation. Motivating this change in jurisdiction, he argued that the investigation of electoral irregularities had to consider both the “imperative letter of the law” and electoral habits and norms. For proponents of ­these reforms, juries w ­ ere in a unique position

R e fo r m i n g C a m pai g n s 91

to reach impartial decisions in cases of corruption, as “they are above any suspicion, impartial, highly in­de­pen­dent and capable of judging the facts humanely and equitably.” Such proposals met with strong opposition from Republican legislators and government representatives who participated in the parliamentary deliberations. One line of criticism invoked the dangers of ­legal anarchy. Opponents argued that ­these proposals established two dif­fer­ent types of jurisdiction, blurring po­liti­cal responsibilities among the dif­fer­ent actors ( JORF ChamberDeb, 26 March 1902). Lucien Millevoye, a Republican from Seine, argued that t­ hese proposals generated a genuine confusion of power: “By allowing judges to evaluate our po­liti­cal honor and the direction of po­liti­cal operations, you ­will compromise both the judicial and the legislative power. You ­will establish in this country a horrifying strug­gle of passions” (ibid.). Another objection to this proposal was that it gave excessive power to judges. Jules Antoine Roche, a Republican from Var, argued that the proposals introduced the absolute arbitrariness of the judges. For Roche, ­these amendments empowered “magistrates to become not only arbiters of a par­tic­ul­ ar race, but arbiters of the entire election. Using the pretext to identify electoral irregularities, judges w ­ ill be able to examine all the incidents affecting an election and they may use po­liti­cal motives to decide ­whether a par­tic­ul­ ar fact constitutes an electoral irregularity” (ibid.). The opponents of ­these changes prevailed. As a result, proposals recommending bold changes in the pro­cess of investigating electoral irregularities ­were defeated. The po­liti­cal compromise that was ­adopted allowed judges to investigate electoral irregularities only a­ fter the conclusion of parliamentary inquiries and only ­after the Chamber de­cided to invalidate a race ( JORF Lois et Décrets, 1 April 1914, Article 5). In practice, this represented only a cosmetic change that was unlikely to lead to more stringent punishment of patronage and clientelism. During this episode, the Republican majority in the Chamber successfully resisted changes to the pro­cess of investigating electoral irregularities. VOTING FOR ELECTORAL REFORMS

During the 1902 deliberations in the Chamber, legislators voted on vari­ous amendments recommending more severe punishments for candidates who politicized state resources during campaigns. The decisive vote on this legislation took place in 1910. As the draft bill ­adopted by the Senate returned to the lower ­house, legislators in the Chamber recommended using this

92 C H A P TER 3

draft bill as the basis for the final legislation. This proposal was approved on 16 May 1910. Th ­ ese votes provide us with an opportunity to examine more closely the determinants of the demand for reform and to understand the sources of disagreement among legislators about ­these reforms. The analy­sis of the parliamentary votes for electoral reforms presented in this and subsequent chapters takes advantage of recently digitized data of the Third Republic. This data set brings together information on the partisanship, electoral competitiveness, and personal attributes of politicians with information about the economic conditions of each district. (In the online appendix [http://­press​.­princeton​.­edu​/­protecting​-­the​-­ballot], I pre­sent more detailed information about the sources used to construct this data.) The goal of the analy­sis is to provide empirical tests of the main propositions presented in chapter 2 using individual-­level data. My empirical analy­sis seeks to assess how resource asymmetries, economic constraints, and electoral constraints shape the demand for electoral reforms. In chapter 2, I conjectured that demand for electoral reforms reflects considerations about relative resource advantage, on the one hand, and about the economic and electoral costs of illicit strategies, on the other hand. Consider first the relative resource advantage. Both Republican and Radical legislators could access state resources during campaigns. Reflecting this relative resource advantage, legislators affiliated with ­these parties ­were likely to support the status quo and reject proposals for change. By contrast, both Conservative and Socialist legislators ­were unable to access resources of the state, such as brokers or public policies. Consequently, I expect t­ hese resource-­constrained legislators to support electoral reforms. More severe punishments for strategies that politicized state resources as electorally advantageous would constrain Republican and Radical opponents and level the electoral playing field. A second f­ actor that affected legislators’ demand for reform was the electoral costs a candidate could incur as a result of the strategies deployed during campaigns. Candidates who competed on the basis of programmatic policies w ­ ere likely to incur electoral costs when using clientelistic strategies, as ­these strategies undermined the credibility of ­these promises. For resource-­constrained politicians—­such as Socialists—­who competed on the basis of programmatic promises, ­these considerations about the electoral costs reinforced concerns about their relative disadvantage, increasing demand for reform. By contrast, resource-­endowed politicians facing electoral costs became cross-­pressured. If considerations about electoral costs outweigh the importance of relative resource advantage, such cross-­pressured

R e fo r m i n g C a m pa i g n s 93

legislators would be likely to join the parliamentary majority supporting reforms. This switch of ­these cross-­pressured politicians is a necessary condition for the formation of a majority supporting reforms. To mea­sure the programmatic promises made by candidates, I use information in a significant, hitherto unused historical source of French electoral politics. Beginning in 1885, the French Chamber of Deputies systematically collected the electoral platforms of all French legislators. This collection of documents is also known as “the Barodet,” named a­ fter Désiré Barodet, a Republican legislator who recommended creating this publication. For each campaign, the collection includes all electoral platforms, also known as professions de foi. ­These documents—­which range from one to eight pages in length—­pre­sent the main promises made by politicians during their campaigns. The professions de foi thus allow us to mea­sure the presence of programmatic promises during campaigns and proxy the programmatic orientation of a candidate. In the empirical analy­sis presented below and in subsequent chapters, I proxy the programmatic orientation of a candidate by using the commitment of a candidate’s platform to the introduction of progressive income taxes. Reform of taxation—­and proposals to introduce a progressive income tax—­was a central fiscal question on the French parliament’s agenda at the time. While the first proposals to introduce an income tax ­were considered as early as 1880, the French Chamber ­adopted an income tax only in 1914. ­Because of the importance of the new tax and the fact that it had the potential to restructure the entire fiscal architecture of the French state, a large number of the professions de foi took issue with t­ hese proposals. Candidates took a variety of positions ­toward the income tax. While some rejected the income tax altogether, considering it a form of extortion, other candidates competed on the basis of promises to introduce a progressive income tax, subjecting higher incomes to a higher rate. I proxy the programmatic orientation of a candidate by using the commitment to the introduction of progressive income taxes in a candidate’s platform. ­Table 3.2 pre­sents descriptive statistics on the distribution of ­these programmatic promises of legislators affiliated with dif­fer­ent parties competing during the 1898 and 1906 parliamentary elections in France. The data documents both that the relative share of candidates supporting tax progressivity varies across parties and significant temporal variation in the trends. Unsurprisingly, Socialist deputies ­were the most likely to mention the promise of a progressive income tax in their professions de foi. In both 1888 and 1906, nearly 50 ­percent of Socialist deputies advocated for the introduction of a

94 C H A P TER 3

­TABLE 3.2.

The programmatic orientation of candidates of dif­fer­ent parties competing in the 1898 and 1906 parliamentary elections in France Percentage of professions de foi demanding the introduction of progressive income tax

Party Republican

1898

1906

6.21%

27.39%

Radical

33.5%

29.03%

Socialist

44.8%

48.2%

Monarchists and Catholics

8.47%

1.66%

No partisan affiliation

5.46%

5.71%

Source: JORF 1899, 1907.

progressive income tax in their electoral platform. In contrast to Socialists, Monarchists and other Republican legislators campaigned in strong opposition to the progressive income tax. We find a strong difference among Republicans and Radical legislators in their support for programmatic policies during the 1898 election, but the difference in programmatic commitment between t­ hese two groups of legislators attenuated over this period. By 1906, about 30 ­percent of Radicals and centrist Republicans supported the introduction of an income tax. Using information from the professions de foi as a proxy of the programmatic commitment of a candidate, we are now ready to examine the determinants of support for electoral reforms. I begin with an analy­sis on the votes on amendments to sanction vari­ous electoral interventions involving state resources discussed in the Chamber in 1902. Figure 3.2 shows the determinants of support for the proposals to sanction the use of state resources during campaigns proposed by Régis de l’Estourbeillon, Louis Gauthier, and Albert Chiché, respectively. The results presented in the figure illustrate that the sources of division w ­ ere largely similar across t­ hese four dif­fer­ent votes. Conservative and Monarchist politicians showed strong support for ­these proposals, favoring all four amendments to the electoral law. For t­ hese legislators, who lacked access to state resources, electoral reforms created an opportunity to level the playing field during campaigns and limit the electoral advantage of Republican candidates. By contrast, resource-­endowed Republican and Radical legislators opposed electoral reforms. I have conjectured that resource-­endowed legislators competing on the basis of programmatic promises are cross-­pressured. At this early stage of the reform

5

10

–10

–5

0

Gauthier amendment

0

5

15

Economic constraints Rural employment Urbanization Landholding Inequality Voting population (ln)

Electoral constraints Programmatic promises

Partisan resources Nonpartisan Socialist Radical Republican Right

Individual resources Mayor

–5

0

5

10

Economic constraints Rural employment Urbanization Landholding Inequality Voting population (ln)

Economic constraints Rural employment Urbanization Landholding Inequality Voting population (ln) –10

Electoral constraints Programmatic promises

Electoral constraints Programmatic promises

Individual resources Mayor Partisan resources Nonpartisan Socialist Radical Republican Right

Chiché amendment

Partisan resources Nonpartisan Socialist Radical Republican Right

Individual resources Mayor

FIGURE 3.2.

Determinants of support for vari­ous amendments seeking to limit the use of state resources during campaigns. Notes: Results pre­sent correlates of support for vari­ous proposals limiting the electoral use of state resources considered by the French Chamber of Deputies in 1902. All models are estimated using Firth logistic regression. See the appendix for the sources used to construct the variables in the analy­sis.

–15

–5

Estourbeillon amendment

96 C H A P TER 3

pro­cess, it appears that concerns about electoral costs ­were not ­great enough to sway ­these legislators’ preferences in ­favor of reforms. The relationship between programmatic promises and support for the proposal does not reach statistical significance at conventional levels. Other results disconfirm some of the predictions formulated in chapter 2. Surprisingly, Socialist candidates voted in opposition to all four amendments. The position of Socialists during ­these votes stands in contrast with other parliamentary efforts of the Socialist faction at the time—­such as the Viviani proposal—­which supported more severe sanctions on candidates who politicized state resources. We also find no evidence that the individual-­ level resources of a legislator—as proxied by his previous experience as mayor—­correlate with his position on this amendment. Fi­nally, let’s consider the correlates between economic development, as proxied by economic conditions in a district, and support for electoral reforms. In chapter 2, I conjectured that for electoral strategies that politicize state resources, economic constraints are negligible in shaping demand for reforms. Such strategies are financed by tax revenues and not by candidate resources. As a result, changing economic conditions are less likely to impact support for electoral reforms. The results presented in figure 3.3 lend support to ­these conjectures. Vari­ous indicators of economic development—­ such as urbanization or the size of the population in a constituency—­are uncorrelated with support for reforms. The level of industrial employment is negatively correlated with support for reform, but this relationship is not robust across all models. In contrast to the predictions made by modernization scholars, economic development and urbanization do not correlate with increased demand for reform. We also find no statistical relationship between landholding in­equality and support for reforms. As discussed e­ arlier, the Chamber of Representatives voted on t­ hese proposals on a second occasion in 1910. At the time, the proposal that had been approved by the Senate returned to the Chamber. The solution that emerged from the deliberations of the upper chamber recommended sanctioning candidates who had made promises of administrative ­favors to individuals or communes in exchange for the vote. This was not a minor concession to reformers but a recommendation that attacked a central electoral practice of the Third Republic. During a vote that took place on 18 March 1910, the proposal was ­adopted by a majority of votes. The results presented in figure 3.3 show that the composition of the parliamentary co­ali­tion supporting reform changed compared to 1902. Importantly, both Republican and Radical legislators supported this proposal for

R e fo r m i n g C a m pai g n s 97

(a) Individual resources

(b) Economic constraints Economic constraints

Individual resources

Agricultural employment

Mayor

Urbanization Landholding inequality

Owner –2

–1.5

–1

–0.5

0

0.5

Population (ln) 0

(c) Electoral constraints

5

10

15

(d) Full model Individual resources Mayor Partisan resources Nonpartisan Socialist Radical Republican Right

Programmatic promises

–1

0

1

2

3

4

Electoral constraints Programmatic promises Economic constraints Agricultural employment Urbanization Landholding inequality Population (ln) –5

0

5

10

15

Determinants of support for proposals to limit electoral clientelism. Notes: Results pre­sent correlates of support for reforms limiting electoral clientelism during a vote taken in the French Chamber of Deputies on 18 March 1910. All models are estimated using Firth logistic regression. See the appendix for the sources used to construct the variables in the analy­sis. FIGURE 3.3.

reform a­ dopted in the Senate. On this occasion, Socialist legislators supported the proposal to punish patronage. At the same time, the position of Conservative legislators about the desirability of reform changed relative to the position taken during the 1902 vote. For many Conservative legislators, the solution accepted by the Senate did not go far enough in limiting electoral strategies that politicized state resources. The proposal also did not introduce the modifications to the pro­cess of investigating irregularities requested by Conservatives. The Conservative opposition during this vote reflects this discontent with the compromise that had been hammered out in the Senate. Together with the analy­sis of the parliamentary deliberations, the results of ­these roll-­call votes for reforms generate two conclusions. First, they illustrate the importance of considerations about po­liti­cal resources in explaining disagreement over provisions to limit the use of state resources. During both the 1902 and 1910 votes, legislators affiliated with Monarchist and other

98 C H A P TER 3

conservative factions supported t­ hese electoral reforms. By contrast, legislators who enjoyed an advantage in accessing state resources—­such as Republicans and Radicals—­initially resisted proposals for reform. Second, the change from the 1902 to the 1910 vote demonstrates the co­ali­tional realignment that took place in France during this period. In par­ tic­ul­ar, the position of Republican and Radical politicians who competed on the basis of programmatic promises shifted over this period. While in 1902 t­ hese legislators rejected proposals for reform, by 1907, t­ hese politicians ­were ready to discount their relative resource advantage and support proposals that sanctioned the use of state resources. The results demonstrate the pivotality of such cross-­pressured politicians in the electoral co­ali­tions that supported reforms. The electoral changes introduced in France in 1914 demonstrate that the adoption of reforms limiting the ability of incumbent elites to politicize state resources is pos­si­ble even in countries without significant electoral turnover. Commenting on ­these reforms, Jean Bienvenu-­Martin, the rapporteur of the Commission of Universal Suffrage in the Chamber, noted that legislation enacted in other countries was much more “thorough and rigorous” as compared to the French reforms. At the same time, he acknowledged that the reforms reached the po­liti­cal possibility frontier of the French po­liti­cal system at the time. The engine for electoral change was politicians competing on the basis of programmatic promises and who experienced electoral costs associated with clientelistic strategies. For ­these politicians, an electoral reform that closed off opportunities for patronage was preferable to the status quo b ­ ecause it ­limited opportunities for other politicians to deploy state resources during campaigns. A QUIET SHIFT IN POLICY IN IMPERIAL GERMANY

As discussed above, French electoral law enacted at the time of the Second Empire included some provisions that sought to limit the use of state resources during elections, but such provisions w ­ ere rarely enforced in the Third Republic. The situation was dif­fer­ent in Imperial Germany. The 1870 electoral law did not include any provisions that sanctioned the electoral use of state resources. It was pos­si­ble to sanction candidates who politicized state resources during campaigns only when such strategies came into conflict with other princi­ples that ­were upheld by the electoral law, such as the princi­ple of electoral freedom of voters. In Germany, the arbiter of ­these impor­tant po­liti­cal questions was the commission verifying electoral

R e fo r m i n g C a m pa i g n s 99

irregularities of the Reichstag, known as the Wahlprüfungskommission. The commission had been established in 1876 and was given the prerogative to examine all violations of the electoral law (Arsenscheck 2003: 57). The electoral commission played an extremely impor­tant role in deciding w ­ hether the involvement of state officials in campaigns came into conflict with other princi­ples of electoral law. The commission’s decisions formed a dynamic and rapidly changing jurisprudence (Hatschek 1920). ­These modifications reflected the gradual emergence of a po­liti­cal majority willing to sanction the use of state resources during campaigns. Notwithstanding this change in jurisprudence, before 1914, the Reichstag did not adopt additional legislation that sought to limit ­these electoral strategies. During the first elections of the Imperial period, the electoral commission took a very permissive approach ­toward electoral strategies that deployed state resources. ­These decisions reflected the preferences of the main po­liti­ cal parties on the right: Conservatives, ­Free Liberals, and National Liberals. During the early elections of this period, politicians representing t­ hese dominant parties competed as part of an electoral alliance, the Kartellbündnis, and faced low electoral costs associated with the use of state resources. The electoral commission refused to sanction candidates who deployed state employees as brokers. However, the Kartellbündnis broke down at the beginning of the 1890s. Following this elite split, the preferences of National Liberals about the desirability of electoral strategies that politicized state resources changed. The increasing po­liti­cal fragmentation elevated the importance of runoffs in districts that had been previously won easily by a candidate enjoying support from other candidates on the right. Greater fragmentation, in turn, increased the electoral costs associated with the use of state resources during campaigns. ­These electoral changes impacted the preferences of legislators regarding the appropriateness of strategies that politicized state resources. The electoral commission’s decisions during the elections prior to World War I reflected this change in preferences. Consequently, the commission increasingly sanctioned candidates who used state resources during campaigns and engaged in electoral strategies that had been considered permissible during previous elections. ­These developments suggest that a gradual democ­ratization of electoral practices took place in Imperial Germany during this period. ­Earlier I discussed the variety of strategies by which candidates competing in elections politicized state resources. A wide range of governmental employees, including the Landräte, county secretaries, and policemen, ­were used to distribute ballots and mobilize voters during elections but

100 C H A P TER 3

also to harass and deter opposition candidates from campaigning. Such electoral interventions by state employees gave rise to electoral complaints and calls for the invalidation of ­these races, which ­were considered by the electoral commission of the Reichstag. The investigation of the election in Rothenburg-­Hoyerswerda in 1874 was instrumental in bringing to the forefront of the discussion questions pertaining to the use of state resources during campaigns (StBer DtRt 1874, encl. no. 56). The election was won by Carl Friedrich von Seydewitz, a Conservative candidate who defeated his opponent by a narrow margin of only eighty-­two votes. A protest contesting the election objected to von Seydewitz’s deployment of Landräte to distribute ballots (ibid.). In the discussion of this case in the Reichstag, Edward Banks, a ­Free Liberal deputy, criticized the election as a “drastic case of extreme official candidacy” (StBer DtRt, 30 November 1875; Klein 2003: 145). Both Conservatives and National Liberals defended von Seydewitz’s actions and argued that they did not jeopardize voters’ electoral freedom. Opposing the decision to invalidate the election, Karl Bär, a National Liberal legislator from Offenburg, argued that “the intervention of the employees of the state did not pose any threat on the ability of voters to express their opinion, as the voter was protected by secret voting. The voter does not learn any new information from the fact that the policeman is used other than the government assessment of this candidate. The voter is entirely f­ ree to choose ­whether he wants to select a candidate that is friendly to the government or his opponent; he is not in a coerced state [Zwangsnotwendigkeit]” (StBer DtRt, 30 November 1875). During the following years, the electoral commission investigated ­whether other interventions by state employees ­violated the electoral law. ­These included the use of official publications to promote par­tic­u­lar candidates, promises of goods and ser­vices, the mobilization of voters, and the demobilization of opponents. Legislators affiliated with non-­governing parties—­such as ­Free Liberals, Zentrum, and Social Democrats—­argued that such strategies ­violated voters’ electoral freedom and must be sanctioned. Clemens Heereman von Zuydwyk, a deputy representing the Zentrum, said that “the premise of all elections is that they take place in freedom and voters are f­ ree, to the extent that it is pos­si­ble, of all direct influences but also from official influences of the state and administration. If the government pre­sents itself as a party and intervenes officially in the election, then this is a violation [Einverschiebung] of all princi­ples of the constitution, transforming the entire parliamentary repre­sen­ta­tion into a farce” (StBer DtRt 1881, encl. no. 16). In their defense of electoral intervention by state employees, both

R e fo r m i n g C a m pa i g n s 101

Conservative and National Liberal legislators argued for the po­liti­cal right of state employees to participate in elections (StBer DtRt, 11 March 1880; Klein 2003: 195). In its decision, the electoral commission attempted to strike a balance between ­these competing claims. On the one hand, the commission agreed with some of the arguments made by Conservative and National Liberal politicians, concluding that “participation of state employees during elections is permissible. Similar to other voters, state employees have the right to be active during elections.” However, the commission acknowledged that the electoral involvement of state employees may transgress the bound­aries of the permissible if such employees use the powers of their office during elections or if “state employees exert pressure on voters that are in an administratively subordinate position.” This seminal decision of the electoral commission opened up the possibility for an election to be invalidated if a candidate had used state employees and “the bound­aries of permissible ­were crossed” (Klein 2003: 197). In a variety of decisions during the first electoral cycles of the Reichstag, the electoral commission refused to invalidate races where candidates had deployed state resources. In the 1877 election in Ost-­Priegnitz in Potsdam, for example, several official publications printed electoral announcements on behalf of Hugo von Graevenitz, a Conservative candidate. The commission de­cided, however, that “this type of announcement had no influence on the outcome of the race” (StBer DtRt 1877, encl. no. 151). The commission handed down a similar decision during the 1878 election in Reichenbach Neurode. On this occasion, the commission ruled that the decision of a Landrat to endorse a Conservative candidate “did not influence the outcome of the election” and that ­there ­were no grounds for invalidation (StBer DtRt 1882, encl. no. 194). In a decision concerning the 1881 election in Württemberg’s first district, the commission determined that the deployment of policemen as brokers “did not affect the validity of the race ­because lower-­level employees in a commune cannot influence an election by distributing electoral materials” (ibid., encl. no. 74). In a decision concerning the 1881 election in Halberstadt, the commission concluded that the mayor and policemen’s eviction of Social Demo­cratic ballot con­vey­ors from the voting place was “not decisive in influencing the outcome of the race.” The commission argued that ballot con­vey­ors ­were “only temporarily unable to reach the polling place” and as such the intervention of policemen did not influence the outcome of the race (StBer DtRt 1883, encl. no. 141). At the beginning of the 1890s, the formidable electoral cartel of Conservatives and National Liberals disintegrated. Bismarck’s successors as

102 C H A P TER 3

chancellors—­von Caprivi, Hohenlohe-­Schillingfürst, von Bülow, and von Bethmann-­Hollweg—­were unable to hold together an electoral co­ali­tion representing agrarian and industrial elites who held divergent views regarding a variety of policies that ­were ­under discussion in the Reichstag. The immediate impetus for the breakdown of the electoral cartel was growing disagreement among parties representing agrarian and industrial sectors over the liberalization of trade. While National Liberals pushed for a change in trade policy, Conservatives resisted the reduction in tariff levels. Figure 3.4 documents the changes in electoral competition following the breakdown of the electoral cartel in 1890. Prior to the breakdown of the cartel about 80 ­percent of Conservative legislators and 70 ­percent of National Liberals w ­ ere elected in the first round. The electoral cartel was effective in lowering electoral competition among right-­wing candidates. In this competitive environment, the electoral costs incurred by candidates for deploying the state apparatus during campaigns ­were low. Following the breakdown of the electoral cartel, the majority of National Liberal and Conservative candidates faced a very dif­fer­ent type of electoral competition, characterized by runoffs. The share of candidates competing in the first round declined dramatically. While this change in the character of competition affected legislators of both parties, it was particularly pronounced for National Liberals. By 1903 only 5 ­percent of National Liberal politicians competing in national elections ­were elected in the first round (Milatz 1967; Fenske 1981: 118–19). As conjectured in chapter 2, the co­ali­tional breakdown increased the electoral costs associated with the deployment of the state apparatus during campaigns. The increased electoral fragmentation required coordination with other candidates on the right during runoffs. The heavy-­handed reliance on policemen or other employees of the local administration during the first round of a race could reduce the willingness of other candidates to provide electoral support during runoffs and thus backfire. In this new competitive environment, all candidates on the right reevaluated their electoral strategies, on the one hand, and the desirability of the use of state resources during campaigns, on the other. Conservative and National Liberal candidates responded differently to this change in the electoral environment. While Conservative politicians and key bureaucrats in the administration recognized the increase in electoral costs, they responded by pursuing more subdued forms of electoral interference. Historians of the period have referred to ­these strategies as “noiseless intervention” (Müller 1963). Faced with the growing trade-­off between resource advantage and

Percent of candidates competing in runoffs

Importance of runoffs for Conservative (K) candidates 80

60

40

20 1885

1890

1895

1900 Year

Percent of candidates competing in runoffs

Share K win first round

1905

1910

Share K win runoffs

Importance of runoffs for National Liberal candidates 100 80 60 40 20 0 1885

1890

1895

Share NL win first round

1900 Year

1905

1910

Share NL runoff

Changing electoral competition for candidates on the right following the breakdown of the electoral cartel (Kartellbündnis). Source: Milatz 1967: 215. FIGURE 3.4.

104 C H A P TER 3

rising electoral costs, Conservatives chose to preserve their relative resource advantage and did not push for a change in legislation. In the cases of invalidation considered at the time, Conservatives did not support invalidation. By contrast, National Liberals w ­ ere more willing to sanction the use of illicit strategies. This change in the position of National Liberal deputies generated a new po­liti­cal majority in the Reichstag that supported greater sanctions on this form of intervention. In its decisions during the last elections prior to 1914, the commission increasingly sanctioned the use of state resources during campaigns. Let us consider first the consequences of the cartel breakdown for the po­liti­cal strategies pursued by bureaucrats in the Prus­sian Ministry of Interior, the central institution coordinating the involvement of state employees during elections. The archival rec­ords of the Prus­sian Ministry of Interior document that Prus­sian bureaucrats ­were aware that the changing po­liti­ cal environment imposed higher electoral costs in using ­these strategies. In response to the increased po­liti­cal fragmentation of the 1890s, Prus­sian bureaucrats began to pursue a strategy of “noiseless” and l­imited po­liti­cal intervention (Müller 1963). Beginning in the 1890s, the Prus­sian ministry formulated a number of directives or guidelines (Richtlinien) outlining new rules for electoral intervention (Geheimes Staatsarchiv Berlin Dahlem 90a, B; Puhle 1967: 329). Th ­ ese directives w ­ ere distributed down the ranks of the Prus­sian bureaucracy with a high degree of secrecy. This change in the visibility of bureaucratic involvement during elections demonstrates that in the aftermath of the breakdown of the Kartellbündnis, Conservatives also internalized the electoral costs associated with the electoral deployment of state resources. The Richtlinien clarified which candidates would “require” po­liti­cal support by the bureaucracy, based on their partisan affiliation. This clarification was necessary, given the increase in the po­liti­cal fragmentation of the Reichstag. Thus, during the elections up to 1903, bureaucrats ­were urged to support Conservatives, ­Free Conservatives, and National Liberal candidates (Fairbairn 1997: 70). During the 1907 election, another Liberal party—­the Deutsche Freisinnige Partei—­was added to the list of parties that needed bureaucratic support. A second recommendation made to bureaucrats was to exercise care ­toward candidates who could be co­ali­tion partners during runoffs. This demonstrates that considerations about co­ali­tion partners in runoffs ­were increasingly salient for Conservative candidates. The new guidelines recommended supporting Zentrum candidates during runoffs against Social Demo­cratic candidates, a move that departed from

R e fo r m i n g C a m pa i g n s 105

the position held during the first elections of the period, which labeled the Zentrum as a party hostile to the government (regierungsfeindliche Partei). However, bureaucrats w ­ ere asked to consult with their superiors if uncertainty existed as to ­whether par­tic­u­lar candidates ­were worthy of po­liti­cal support (Arsenscheck 2003: 198). Reflecting concerns about the electoral costs associated with the use of state resources, t­ hese directives also attempted to redefine the range of permissible electoral strategies. Officials of the Prus­sian Ministry of Interior advised Landräte and other local-­level officials not to sign the candidacy announcements (Wahlaufruf). Bureaucrats w ­ ere also advised to refrain from taking clear positions during governmental rallies (Arsenscheck 2003: 198). On the other hand, officials in Berlin endorsed electoral activities, such as the control and multiplication of voter lists. Thus, while concerns about electoral costs contributed to a reduction of vis­i­ble electoral involvement of the bureaucracy, such considerations did not lead to a full pullback of bureaucratic involvement. During the last elections prior to World War I, the Prus­sian ministry took a number of steps to reduce the visibility of its electoral engagement. During the 1903 election, the Caprivi government took a hitherto unpre­ ce­dented precaution to hide any trace of its intervention. The government relied on the extensive dissemination of leaflets and other electoral materials attempting to justify its position on issues of taxation and reassure low-­income voters that they would not be disproportionately affected by any tax increase. In a secret memorandum, the Prus­sian Ministry of Interior instructed local government officials to oppose the policy positions of its critics. To disseminate electoral material to voters, the government established a special office in Berlin, led by a military official. To minimize the visibility of the electoral involvement of government employees during campaigns, the government asked the Landräte to provide the addresses of lower-­level brokers (Vertrauensmänner) who could be contacted during elections. This electoral strategy contrasts with the heavy-­handed involvement of the Ministry of Interior during ­earlier elections. The leaders of the German Conservative Party also ­were aware that the use of state resources during campaigns could be electorally counterproductive. During the elections immediately before World War I, the Conservative Party distributed electoral publications to its candidates, updating them on the changes in electoral jurisprudence (Anderson 2000). However, even though Conservatives experienced a tension between increased electoral costs and their resource advantage, this conflict did not change their preferences with

106 C H A P TER 3

regard to electoral reforms. ­Until World War I, Conservatives did not support changes in electoral law or jurisprudence that sanctioned the electoral use of state resources. Also, Conservatives routinely refused to sanction candidates who deployed state resources during their campaigns. In contrast to Conservatives, National Liberal legislators’ preferences about electoral interventions that politicized state resources changed a­ fter the breakdown of the electoral cartel. During this period, National Liberals ­were also cross-­pressured between their relative resource advantage and the electoral costs associated with illicit strategies. Unlike Conservatives, National Liberals increasingly supported stronger sanctions on electoral strategies that politicized state resources. This shift in National Liberals’ preferences on questions regarding the appropriateness of the use of state resources can be documented using roll-­call votes on decisions to invalidate the mandate of a candidate who politicized state resources during campaigns. In figure 3.5, I examine the results of one such vote, taken in the Reichstag in 1914. On this occasion, members of the Reichstag voted on the recommendation of the electoral commission to invalidate the election of Felix Hoesch, a Conservative candidate. Hoesch, a landowner, had been elected in Osterburg-­Stendal, a district in Magdeburg, defeating his opponent in a narrow race. The commission investigating irregularities found ample evidence of the use of state resources during the campaign. Hoesch had benefited from the electoral support of the Landräte and countless other employees of the administration during the election (StBer DtRt, 1912, encl. no. 1497). The electoral commission of the Reichstag voted to invalidate the race and sent the decision to the floor of the lower h ­ ouse. In a meeting held on 26 March 1914, the Reichstag reaffirmed the decision of the commission, invalidating Hoesch’s election (StBer DtRt, 24 March 1914). The results presented in figure 3.5 document the shift in the preferences of National Liberals about the use of state resources during campaigns. At this time, National Liberals stood alongside other progressive po­liti­cal parties in the Reichstag—­such as F ­ ree Liberals and Social Democrats—in opposing the politicization of state resources. Surprisingly, Zentrum candidates voted on this occasion with Conservatives against Hoesch’s invalidation. The shift in the position of National Liberals documents the new electoral majority supporting electoral reforms in the German parliament prior to World War I. The National Liberals’ shift in position also changed the majority in the electoral commission of the Reichstag ­toward electoral interventions that politicized state resources. Reflecting this new po­liti­cal majority, the commission

R e fo r m i n g C a m pa i g n s 107

Partisanship Others Conservatives National liberals Free liberals Zentrum Social Democrats

Economic conditions Agriculture (percentage) Landholding inequality Population (ln) Urbanization (percentage) –10

–5

0

5

10

Correlates of support for the invalidation of elections that politicize state resources: divergence among parties on the right. Notes: The results pre­sent the correlates of support of a decision to invalidate the election of Felix Hoesch (German Conservative Party) in Osterburg/Stendal during the 1912 election. Reichstag vote took place on 26 March 1914. FIGURE 3.5.

made a number of decisions that increasingly sanctioned candidates who used state resources during campaigns. During the last election cycles of the Wilhelmine period, the electoral commission of the Reichstag modified the electoral jurisprudence in several decisions concerning the use of state resources during campaigns. During the first election cycles, the electoral commission refused to sanction electoral interventions by employees of the state, ruling that they had the “right as citizens and voters to intervene actively in elections” (in die Wahlagitation tätig eingreifen) (StBer DtRt 1907, encl. no. 251). However, during the last two elections prior to World War I, the electoral commission formulated a stricter standard that distinguished between electoral interventions of state employees as private citizens and activities that reflected the “attributes of the office” (amtliche Eigenschaften) (ibid.). In a 1907 decision that changed jurisprudence, the commission ruled that electoral interventions that used the “authority of the office” or other means of state power ­were illicit electoral activities (ibid.). Relatedly, the electoral commission

108 C H A P TER 3

ruled that the use of vis­i­ble symbols of state power by a candidate also constituted electoral irregularity (ibid., encl. no. 367). ­These illicit activities included, among ­others, the distribution of ballots with official symbols of state power by policemen (StBer DtRt, 15 November 1906; StBer DtRt 1907, encl. no. 702, p. 4471). In one of the electoral commission’s last decisions before World War I, it investigated irregularities that occurred in Lyck Allenstein, a district in East Prus­sia. H ­ ere, Landräte distributed election materials in envelopes with official seals. The envelopes also included notes that reminded addressees of their official oath (Diensteid). The commission ruled that “reminding state employees of their official oath represents an inadmissible pressure by the Landräte to achieve a conservative majority” (StBer DtRt 1912, encl. no. 1061). The position of the electoral commission with regard to electoral interventions by lower-­level employees in the state administration—­such as mayors and city hall employees—­also changed during the last elections of the period. In examining events that occurred in Dittmarsch-­Steinburg, Schleswig Hohlstein’s fifth district, the commission concluded that the distribution of ballots by employees of the local administration represented an illicit electoral interference (StBer DtRt 1912, encl. no. 851). The commission also concluded that the organ­ization of rallies by mayors or employees of the local administration constituted an illicit electoral strategy (StBer DtRt 1903, encl. no. 589; StBer DtRt 1912, encl. no. 851). If the mayor used “the connections of the office and the influence connected to the office for electoral gain,” then such strategies w ­ ere illicit (StBer DtRt 1907, encl. no. 252). Following up on t­ hese decisions, the commission de­cided to annul votes cast in localities that had experienced electoral irregularities and subtract t­ hese from the total number of votes received by the winning candidate. This section has documented the consequences of the breakdown of the electoral alliance between Conservatives and National Liberals with regard to electoral reforms. This elite split increased the electoral costs of strategies that politicized state resources. I documented the rising awareness of bureaucrats in the Prus­sian administration and of Conservative politicians to t­ hese electoral costs. Nevertheless, while Conservatives w ­ ere cross-­pressured between considerations about their po­liti­cal advantage and rising electoral costs, their position on the desirability of electoral reforms did not change during this period. By contrast, the position of National Liberals underwent a change as a result of rising electoral costs. Prior to World War I, National Liberals endorsed proposals to limit the electoral use of state resources and voted to sanction candidates who engaged in ­these illicit strategies. This

R e fo r m i n g C a m pa i g n s 109

change in the preferences of National Liberals contributed to the formation of a new electoral majority favoring reforms. The changes in German electoral jurisprudence led to stronger sanctions for strategies that politicized state resources prior to World War I. The results of this section confirm one of the main predictions developed in chapter 2, which points to the importance of elite splits in contributing to the formation of a new po­liti­cal majority favoring stronger sanctions on different forms of electoral malfeasance. Conclusion

In this chapter, I have documented the variety of ways in which state resources w ­ ere politicized in nineteenth-­century elections. Candidates affiliated with the governing majorities competed by politicizing the entire governmental infrastructure during campaigns. Th ­ ese candidates delegated to government employees such tasks as distributing election materials and mobilizing voters on election day. The importance and magnitude of state resources for electoral campaigns generated strong po­liti­cal re­sis­tance to electoral reforms from politicians who could access such resources. The initial cleavage line over t­ hese electoral reforms reflected this resource asymmetry. In both France and Germany, legislators affiliated with governing parties opposed the introduction of reforms. By contrast, candidates lacking access to state resources supported a variety of sanctions regarding the use of state resources. The chapter documents how elite splits within the governing co­ali­tion modified the preferences of legislators about the desirability of electoral interventions. In both France and Germany, elite splits played an impor­tant role in facilitating the formation of a po­liti­cal majority favoring reforms. However, elite splits affected legislators’ preferences about the desirability of the status quo through dif­fer­ent mechanisms. In France, programmatic considerations w ­ ere decisive in changing the preferences of resource-­ endowed politicians about the desirability of electoral reforms. For legislators competing on the basis of programmatic promises, it was electorally too costly to combine such appeals to voters and patronage strategies that politicized state resources. This change in preferences of resource-­endowed politicians who encountered high electoral costs facilitated the formation of an encompassing po­liti­cal majority favoring stronger sanctions for ­those who used state resources during campaigns. This co­ali­tional realignment contributed to the adoption of the law limiting electoral corruption by the French parliament in 1913.

110 C H A P TER 3

In Germany, the breakdown of the electoral cartel of National Liberals and Conservatives also contributed to a change in the preferences of resource-­endowed politicians about the desirability of the status quo. Following this electoral split, the programmatic shift of Conservative and National Liberal politicians was less pronounced than in France. ­Here, the increased po­liti­cal fragmentation—­which necessitated coordination during runoffs—­raised the electoral costs incurred by candidates who deployed state resources during campaigns. In Imperial Germany, National Liberals who ­were dependent primarily on runoffs changed their preferences about the desirability of the status quo and ­were increasingly likely to sanction candidates who deployed state resources during campaigns. During the last elections before World War I, this shift contributed to the formation of a new po­liti­cal majority that was favorable ­toward sanctioning the electoral use of state resources. The change in the electoral jurisprudence sanctioning the use of state resources during campaigns that took place in Germany before 1918 reflects the preferences of this new po­liti­cal majority.

4 Limiting Vote-­Buying and Treating

Candidates competed in nineteenth-­century elections by offering goods, money, and food and drink to voters in exchange for their vote. Such exchanges w ­ ere financed by the private resources of candidates, not by public resources. In some cases, candidates (or their representatives) targeted voters directly, extracting promises of electoral support in exchange for ­these goods. In nineteenth-­century French elections, goods offered in exchange for the vote included cigarettes, bread, meat vouchers, and clothes ( JORF ChamberDeb, 7 July 1898, 10 February 1881, 27 June 1902, 23 June 1898). In other cases, candidates offered food and drink freely, without clear expectations of electoral reciprocity. In Britain, most of the treating occurred in public h ­ ouses, which w ­ ere used as the committee rooms of parties. In France, such untargeted exchanges took place in rural inns (auberges), in small rest areas, known also as buvettes, and sometimes in the homes of candidates (ibid., 1 February 1886, 21 November 1885). On other occasions, candidates placed barrels of wine in the immediate vicinity of the voting place (ibid., 29 June 1898). Many Eu­ro­pean countries ­adopted l­ egal sanctions against bribery prior to the introduction of mass suffrage. The French Civil Code categorized bribery as a criminal offense and recommended stripping offenders of po­liti­cal rights. In a pro­cess spurred on by the Napoleonic invasion, t­ hese provisions of the French Civil Code diffused rapidly across Eu­ro­pean countries. Both 111

112 C H A P TER 4

civil codes and electoral laws introduced in Eu­rope during the nineteenth ­century imposed strong punishments on vote-­buying. Electoral laws in Belgium, Italy, Denmark, and Austria incorporated provisions regarding vote-­buying that ­were similar to the French Civil Code (Mayer 1906: 340–45). Civil codes and electoral laws in nearly all other German principalities, as well as the law governing national elections in Germany, also considered vote-­buying to be a criminal offense, imposing both monetary fines and prison for candidates who engaged in such exchanges. Lacking a codified criminal law, Britain remained an exception to this po­liti­cal development. ­These early differences in the sanctioning of vote-­buying had consequences for the electoral importance of bribery. In continental Eu­ro­pean countries, the penalization of bribery in the criminal code deterred candidates from using ­these strategies, reducing the importance of vote-­buying during campaigns. The use of vote-­buying was much less widespread in ­these countries than in Britain. However, the attempt to sanction vote-­buying through the penal code was not without limits. The ­legal requirements to sanction offers of gifts and money as a criminal offense ­were very high. As a result, vote-­buying incidents that w ­ ere reported by electoral commissions remained often unpunished in courts. In ­these countries, the most impor­tant objective of electoral reforms was simplifying the pro­cess by which courts could sanction bribery. The situation was entirely dif­fer­ent in Britain. Minimal penal sanctions created a permissive environment for candidates to use promises of goods and money during campaigns, leading to an exponential growth in expenses incurred by candidates through bribing and treating. Successive electoral reforms seeking to limit vote-­buying considered by the British parliament w ­ ere intended to reduce the levels and growth of campaign expenditures incurred by candidates. In this chapter, I explore the adoption of electoral reforms that sought to limit vote-­buying and reduce the l­ egal burden of prosecuting such offenses. Similar to other cases examined in this study, I explore only reform initiatives considered a­ fter the introduction of nominal voting secrecy. The goal of my analy­sis is to identify the f­ actors that facilitated the formation of po­liti­cal majorities in support of reforms. Considerations about access to the resources required to finance such exchanges and about relative resource advantages informed legislators’ preferences about the desirability of electoral reforms. Economic developments amplified resource asymmetries among candidates and increased the salience of electoral reforms for resource-­constrained legislators and parties. In Britain, economic developments magnified partisan resource asymmetries and enhanced the

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 113

incentives of resource-­constrained Liberal politicians to push for the adoption of reforms limiting campaign expenditures. In contrast to the other electoral reforms examined in this book, concerns about electoral costs did not play a significant role in influencing legislative demand for reforms to limit vote-­buying. The chapter ­will proceed as follows. I begin by discussing the penalization of bribery in the French Civil Code and the diffusion of this ­legal solution across continental Eu­ro­pean countries. Next, I examine the adoption of electoral reforms that attempted to simplify the penalties for bribery through the courts. The reforms ­adopted in France in 1914 enjoyed broad cross-­partisan support and ­were only opposed by Conservative legislators. Then I turn to an analy­sis of reforms to limit vote-­buying in Britain. Due to the absence of criminal sanctions for bribery, the l­egal environment was much more permissive ­toward candidates who used offers of money or goods to interact with voters, which led to the rapid rise in electoral expenses. In analyzing a combination of roll-­call votes and parliamentary deliberations, I examine the motivations of politicians for supporting or opposing reforms to limit campaign expenditures and the most salient cleavage lines formed during ­these deliberations. The Legacy of the Napoleonic Code

Efforts to sanction vote-­buying as a criminal offense date to the French Revolution. The Constitution of 16 Thermidor Year X, ­adopted in France in 1802, considered vote-­buying “a loathsome offense” (Freudenthal 1895: 37; Chauveau and Hélié 1852: 191). The authors of this constitution recommended sanctioning offers of money or goods by taking away the offender’s po­liti­cal rights. Article 32 decreed that any citizen who had bought or sold votes should be excluded from local meetings and stripped of the right to occupy a public role for a period of twenty years. The punishment was escalated for repeated offenses. In this case, the law recommended taking away irrevocably the right to occupy a public office (Freudenthal 1895: 37). The Napoleonic Code, ­adopted in 1810, took up ­these l­egal princi­ples and solutions. The penal code introduced very strong punishments both for ­people who offered money in exchange for votes and for ­people who accepted t­ hese offers. Article 113 of the penal code recommended sanctioning any citizen who at time of elections had bought or sold a vote for any price with the loss of citizenship rights and any public duty or employment for a period of five to ten years. In addition, the code imposed fines that ­were double the amount offered in the bribe. In establishing such harsh

114 C H A P TER 4

punishments, lawmakers drew a contrast between venal po­liti­cal practices in Britain and an idealized French society that embraced honor and dignity. The preface to the penal code read as follows: “Let us leave to the British the scandalous privilege to solicit the votes of their fellow citizens with the price of money and through expenses; French honor rejects such means” (Code Pénal 1815: 43). Spurred by the Napoleonic invasion, the solution a­ dopted in the French Penal Code spread across continental Eu­ro­pean countries (Mayer 1906). Civil codes ­adopted at the time sanctioned bribery and treating with a loss of po­liti­cal rights and imprisonment. The Swedish, Finnish, Danish, and Austrian civil codes introduced provisions that closely mirrored the French Civil Code (ibid., 342). In some cases, civil codes introduced deviations from the French model. For instance, the Danish Civil Code sanctioned a person who bought a vote with a harsher punishment than a person who sold a vote, while the Austrian Civil Code imposed the same penalty for transactions w ­ hether or not voters had accepted an offer of bribery (ibid., 347). Other countries incorporated provisions that sanctioned bribery harshly in their electoral laws. The electoral law of Luxembourg punished both the person making the offer and the person who accepted money or gifts in exchange for the vote with the temporary loss of electoral rights. The Dutch electoral law sanctioned promises or gifts made with the goal of preventing voting or changing voters’ intentions with prison sentences of up to six months and with a monetary sanction of 300 gulden (ibid.). A Belgian electoral law, ­adopted in 1867, sanctioned vote-­buying with fines of fifty to five hundred francs (Nypels 1868: 336). Dif­fer­ent German principalities also a­ dopted the provisions of the French code in the aftermath of the Napoleonic invasion. The Prus­sian Criminal Code of 1851 sanctioned selling or buying votes with a prison term of three months to two years. The Prus­sian government also retained the right to limit the po­liti­cal rights of ­people who engaged in vote-­buying exchanges (Freudenthal 1895: 35). In Hessen, the penal code sanctioned both promises of money or gifts and the ac­cep­tance of ­these offers with imprisonment (Korrektionshaus) for a period of up to a year (ibid., 30). The penal code of Hannover, enacted in 1840, distinguished between active and passive bribery. Active bribery consisted of offering gifts or other advantages in order to obtain a public office; candidates who engaged in passive vote-­buying accepted such offers e­ ither directly or through intermediaries. The penal code of Braunschweig sanctioned any offer of “gifts or other advantages with prison or fines” (ibid.). Among large German principalities, the Bavarian Criminal Code was the only code that did not sanction vote-­buying.

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 115

With re­spect to the punishment of bribery, the provisions of the Napoleonic Code influenced the Criminal Code of the German Empire via the codes of individual principalities. Ernst Mayer, a German expert on criminal law at the time, documented how the Prus­sian Civil Code took up the provisions of the Napoleonic Code without any modification. Through this pro­cess, provisions of the Napoleonic Code made their way to the Criminal Code of the German Empire (Mayer 1910). Commenting on this ­legal borrowing, Mayer noted that the Prus­sian and German criminal codes copied titles of vari­ous sections of the Napoleonic Code, even though ­these ­were not in immediate agreement with the specific provisions of the German law (ibid.). Article 109 of the German Criminal Code sanctioned bribery with prison terms of one month to two years and a pos­si­ble loss of po­liti­ cal rights (ibid.). The strong sanctions for bribery and treating through civil codes ­were relatively successful in deterring candidates from using t­ hese electoral strategies. In both France and Germany, vote-­buying was used less often than strategies that politicized state resources. Figure 4.1 uses information from the electoral complaints submitted to the commission de vérification des pouvoirs in France and the Wahlprüfungskommission in Germany to assess the relative importance of vote-­buying in the mix of electoral irregularities. In France, around 30 ­percent of the election reports included complaints about vote-­buying. In Germany, averaging across all elections, around 15 ­percent of electoral reports included complaints about bribery. Historians of the period offer similar accounts of the ­limited use of vote-­buying in German elections. Consider Margaret Lavinia Anderson’s characterization of the ­limited role of bribery in German elections. Cotters in the east Elbian Herrschaft Pnuwo ­were given a half bushel of dried peas for voting Conservative. Voters in the Rhenisch village of Rodenkirchen who accepted “loyal” ballots w ­ ere handed a sausage from a large hamper at the door of the polls. A dram of schnapps at the beginning of our period and much larger quantities of “Freibier” at the end ­were common, although not always considered respectable features of German elections. . . . ​What­ever the incidence of ­these ­little reciprocities, it is clear that German voters w ­ ere not considered venal even by parties hoping to overturn an election, nor did enough money change hands to give Reichstag deputies, unlike their British counter­parts, any stake in regulating it. Bribery simply did not play a role in what Germans thought to be wrong with their po­liti­cal pro­cess. (2000: 27)

Share of electoral reports reporting vote-buying

France 60

40

20

0

1881

1885

1889

1893

1898

1902

1906

1910

1914

Share of electoral reports invoking vote-buying

Germany 25

20

15

10

5

0

1871 1874 1877 1878 1881 1884 1887 1890 1893 1898 1903 1907 1912

Relative incidence of vote-­buying in French and German elections. Sources: The entries for French elections are based on a coding of the content of all electoral petitions considered by the commissions de vérification des pouvoirs. JORF Chamber Deb, vari­ous years. The entries for German elections reflect a coding of a random subset comprising 20 ­percent of electoral reports, stratified by election. StBer DtRt (vari­ous years). FIGURE 4.1.

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 117

Though the penal codes’ provisions ­were effective in deterring vote-­buying, German and French courts encountered insurmountable difficulties in implementing the harsh sanctions dictated by the code (Pierre 1893). Cases of vote-­buying that arrived in court rarely led to sanctions. On the one hand, it was difficult to document the existence of a market for votes, as both buyers and sellers had an incentive to obfuscate or deny they had participated in such transactions. At the same time, to prosecute vote-­buying as a criminal offense, it was necessary to prove that the sale and purchase of votes had actually taken place. Promises could not be sanctioned using the provisions of the penal code. Fi­nally, courts also required evidence that candidates or brokers who had offered money to voters did so with the intention of buying votes. Commenting on ­these difficulties, Pierre Waldeck-­Rousseau noted that “it is very difficult to prove intention, since ­these transactions are not certified by a notary” ( JORF ChamberDeb, 26 March 1902). The information presented in ­table 4.1 illustrates ­these difficulties. ­Here, I pre­sent information on the number of convictions and acquittals for bribery in Germany for the period 1892–1901. This data, collected by Mayer, reports information on all cases that fell u ­ nder the provisions of Paragraph 109 of the German Penal Code, which sanctioned bribing. The data documents two consequences of the penalization of bribery. On the one hand, the number of cases that resulted in ­either penalties or acquittals was low. At the same time, the number of acquitted cases relative to cases that w ­ ere sanctioned was remarkably high. This demonstrates the inability of courts to carry out the punishments designed by the penal code, a situation that ­legal experts of the time called the “weakness of possibilities of prosecution.” In continental Eu­ro­pean countries, reformers encountered a par­tic­ul­ ar set of challenges that w ­ ere a by-­product of the reliance on the penal code to sanction vote-­buying. This criminalization of bribery was effective in deterring vote-­buying but completely in­effec­tive in prosecuting it. One pos­si­ble solution to this prob­lem was to reform the criminal code and decriminalize vote-­buying altogether. Another possibility was to lower the l­ egal hurdles to sanction vote-­buying. This was the path chosen by reformers in both France and Germany. In Germany, l­ egal experts militated for a reform of the penal code during the first de­cades of the twentieth ­century (Mayer 1906, 1910). ­These considerations of experts, however, remained insufficient in placing vote-­buying reform on the agenda of the Reichstag. Germany did not adopt additional reforms sanctioning vote-­buying prior to World War I. In France, proponents of electoral reforms also pushed back against the constraining legacy of the civil code. The penal code sanctioned the buying

118 C H A P TER 4

­TABLE 4.1.

Convictions and acquittals for bribery in German national

elections Correctional Fa­cil­i­ty 1892 1893 1894

1

Penalties Prison

Loss of po­liti­cal rights

Acquittals

15

10

9

6

13

6

1895

19

1896

4

1

11 6

1897

2

5

1898

5

6

1899

10

28

1900

3

5

1901

4

4

Source: Mayer 1906: 289.

or selling of votes very harshly. Article 119 of the penal code imposed severe punishments on vote-­buying, sanctioning it with two years’ imprisonment. The electoral law of the Third Republic reaffirmed this provision of the penal code. It sanctioned vote-­buying strategies with e­ ither a three-­month prison term or a fine that could be as high as five thousand francs ( JORF Lois et Décrets, February 1852, Article 38). Additional electoral laws a­ dopted in 1875 and 1884 reaffirmed t­ hese provisions and extended them to municipal and senatorial elections (ibid., 2 August 1875, 2 August 1884). A number of impor­tant decisions in French courts made it difficult to sanction vote-­buying. ­These decisions raised the ­legal burden for plaintiffs who brought forward complaints about corruption. In a number of seminal cases that s­ haped subsequent jurisdiction of questions of electoral corruption, courts ruled that plaintiffs had to document not just the existence of a corrupt exchange but also the intention of both parties to the exchange to buy or sell votes. In a 1889 decision, the court of Toulouse ruled that “offers of ­free drinks made by candidates to voters cannot be sanctioned by the penal law ­unless it can be established with certainty that the goal of ­these offers was to obtain votes” (Pierre 1893: 169). In a dif­fer­ent decision, the courts ruled that the statement of a single voter that he had received an amount of money in exchange for his vote was not sufficient proof of an illicit strategy if the candidate formally denied this statement (ibid.). Altogether, ­these decisions made it difficult, if not impossible, to sanction vote-­buying.

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 119

­These court decisions also influenced considerations of the parliamentary commissions that investigated electoral corruption. In examining accusations of treating or bribery, t­ hese commissions followed the restrictive rules established by the courts. The electoral commissions viewed testimonies from voters who received promises of goods and ser­vices as, by themselves, insufficient to meet courts’ evidentiary standards ( JORF ChamberDeb, 18 November 1899, p. 80). Incorporating the higher evidentiary standard of the penal law, electoral commissions investigating accusations of bribery ruled that it was necessary to establish “the existence of a declared or implied market in the exchange of money or gifts for votes” ( JORF ChamberDeb, 18 November 1899). Plaintiffs had to prove that voters had accepted ­these offers and that the party making the offer had the intention to subvert the election. Proving both ac­cep­tance and the intention of corruption was notoriously difficult. Similarly, proving voters’ compliance with brokers’ strategies was an additional evidentiary hurdle that was nearly impossible to clear (ibid., 26 March 1902). As a result, electoral commissions rarely recommended the invalidation of elections on accusations of vote-­buying alone. The penalization of vote-­buying ­under the criminal law created unique challenges when it came to sanctioning vote-­buying exchanges. It was nearly impossible to meet the evidentiary standards of the penal code in order to sanction vote-­buying. Faced with ­these constraints, French lawmakers had two opportunities for reform. One pos­si­ble solution involved changing the penal code itself, by removing any provisions discussing bribery or treating. The second solution was to modify the definition of corrupt practices in electoral law. This redefinition sought to accomplish two goals. On the one hand, it attempted to eliminate the requirement that plaintiffs prove intent. On the other hand, it also attempted to sanction promises and not only transactions where voters had accepted the offer made by the candidate. The solution to both prob­lems consisted in jettisoning language implying a corrupt intent and providing a detailed description of offers that represented examples of corrupt exchanges. French legislators who demanded electoral reforms pursued both objectives. Inequalities among candidates and parties in terms of their access to the private resources necessary to finance vote-­buying take us a long way in explaining the demand for or opposition to electoral reforms. Proposals for electoral reforms originated with Socialist politicians who lacked both private resources and access to the resources of the state. Socialists defended the importance of reforms to limit vote-­buying as a necessary extension of the expansion of suffrage. One commonly used argument in support of

120 C H A P TER 4

reforms was that the existing legislation favored rich candidates and wealthy landowners who introduced to France electoral habits found on the other side of the Atlantic ( JORF ChamberDeb, 18 March 1910). Jean Jaurès, the Socialist leader, became familiar with cases of excessive vote-­buying as a member of the commissions de vérification des pouvoirs. One of the races he investigated was the 1893 election in Bagnères-­de-­Bigorre, a district in Hautes Pyrénées. The election had been won by Edmond Blanc, an extremely wealthy candidate who used his private resources lavishly. Jaurès’s passionate intervention in the Chamber of Deputies on 17 March 1894 played a crucial role in the invalidation of Blanc’s election. In his interventions, Jaurès forcefully asserted that Blanc’s victory underscored a fundamental weakness of elections: they ­were no longer ­free if they always guaranteed the victory of ­those with the most abundant resources (Ihl 1998: 107). The position of Republican legislators ­toward reforms to limit vote-­ buying reflected their relative resource endowment. While a majority of Republican legislators could count on support from the higher levels of the state administration, only a small number of Republicans could rely on their private resources during campaigns. During the early elections of the Third Republic, Republican legislators attacked vote-­buying as vestiges of “aristocratic electoral practices” and as an attack on the Republic itself ( JORF ChamberDeb, 20 November 1881). Republicans depicted offers of food and drink to voters as an “aristocratic plea­sure” and as “a calculation of the monarchical reaction allowing certain voters to let themselves be bought by a few coins or to lose their reason in libations or odious orgies” (ibid.). Republicans considered the use of vote-­buying by Monarchist candidates an attempt to undermine the princi­ple of popu­lar sovereignty by ridiculing it. As one Republican legislator put it: “I understand why this is good for the aristocracy. This offers a good argument for them. Look, they could say, the sovereign ­people! The sovereign ­people which we can buy with a few pieces of silver, which we can intoxicate with a few glasses of wine” (ibid.). The initial opposition to t­ hese reform proposals came from Conservative legislators. Lacking access to state resources, ­these candidates had no choice but to use private resources to finance their campaigns. Legislators affiliated with non-­Republican parties on the right such as the Action Libérale denounced proposals to limit the use of vote-­buying as legislation “made against the candidates of the opposition” ( JORF SenateDeb, 18 January 1906). Opponents of reforms formulated three arguments to defend existing vote-­buying practices. First, they argued that such strategies w ­ ere justifiable electoral interventions in a context where their opponents could use

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 121

the full weight of the state apparatus during elections. Jules de Cuverville, a Monarchist senator from Finistère, argued that limiting the ability of candidates to promise private resources to voters would leave in­de­pen­dent candidates completely defenseless against their opponents who benefited from state resources (ibid., 16 January 1906). A second argument in defense of the status quo was that practices of treating ­were consistent with existing norms and w ­ ere not sanctioned by voters. Candidates should be allowed “to have the imprudence to be ­human, generous, follow local traditions and ‘let go’ in the communicative heat of a meeting” ( JORF ChamberDeb, 17 March 1910). Fi­nally, other candidates defended vote-­buying as a philanthropic initiative, as a “plea­sure which is the greatest joy of rich politicians.” Joseph Lasies, a legislator from Gers who competed as an in­de­pen­dent candidate, advanced this defense of vote-­buying. Lasies drew a contrast between “the bad rich . . . ​ [and] the ‘good rich,’ t­ hose who generously contribute during campaigns to mutual aid socie­ties, orphans, orchestras, or veterans’ associations.” Efforts to punish vote-­buying, Lasies argued, penalized the good rich; ­doing so could eliminate all generous forms of giving (ibid.). The first proposals for electoral reforms ­were submitted by Jean Odilon-­ Barrot and René Viviani in 1902. By this time, the majority of French legislators viewed additional electoral reforms as necessary to limit strategies that “infringed on the sacred rights of voters” ( JORF ChamberDoc 1902, no. 2919). Both Odilon-­Barrot and Viviani w ­ ere l­ awyers of relatively modest means who had experienced electoral defeats in races where they encountered rich opponents. Both politicians had weak partisan linkages: Odilon-­ Barrot initially campaigned as a Republican, but he chose to compete as a non-­affiliated politician in ­later elections. Viviani also changed his partisan affiliation on multiple occasions during his ­career. The Odilon-­Barrot and Viviani proposal was submitted to the Chamber of Deputies on 27 January 1902. ­After a brief discussion in the Commission of Universal Suffrage, the Chamber took up the deliberations on 26 March 1902. Odilon-­Barrot and Viviani recommended making it easier to sanction bribery by eliminating from the electoral law wording that implied the corrupt intent of politicians. At the same time, the proposal attempted to sanction not only transactions that had been accepted by voters but also promises made to them by specifically listing all strategies that w ­ ere punishable. Offers of gifts or money ­were considered corrupt offenses. The sanctions ­were prison for two months to three years and a fine of two hundred to a thousand francs ( JORF ChamberDeb, 20 March 1902, p. 1516). Albert de Clagny, an in­de­pen­dent politician from Seine-­et-­Oise, summarized the

122 C H A P TER 4

intentions of the law as follows: “We have made only a s­ imple innovation. We are no longer only demanding sanctions against acts of corruption that have a real effect, where the voter who has been offered gifts or promises accepts the offer. We simply demand applying sanctions even to cases where the attempt to corrupt the voter has been unsuccessful” ( JORF ChamberDeb, 20 March 1902). The debates in the Chamber focused on the electoral strategies that should be sanctioned. Both Radical and Republican legislators in the Chamber ­were ready to support an expansive definition of illicit electoral exchanges. In the words of Jean Bienvenu-­Martin, a Radical legislator from Yonne and president of the Commission of Universal Suffrage, the ideal solution would provide a “synthetic” definition of corrupt exchanges without fully enumerating all pos­si­ble strategies pursued by candidates. A large number of Republican deputies also supported this approach ­because it sidestepped the thorny po­liti­cal question of comparing and contrasting the severity of dif­fer­ent irregularities. The opposition to ­these proposals came primarily from Monarchists and legislators affiliated with other parties on the right. As ­these legislators lacked access to state resources, they benefited electorally from a status quo that left bribery or treating unpunished. Voicing the minority viewpoint of the Commission of Universal Suffrage, Régis de l’Estourbeillon, a legislator affiliated with the Conservative Action Libérale, opined that the law “placed a Damocles’ sword above the head of any deputy. No one can escape the suspicion of having committed illicit offenses” ( JORF ChamberDeb, 17 March 1910). Other deputies concurred, arguing that “the law risked sending half of the population of the country to prison.” Monarchist legislators also objected to the proposal to sanction promises and not only exchanges where voters had accepted bribes. Opposing this change, legislators argued that the term “promises” was ill-­defined and that it could be interpreted broadly and arbitrarily. Promises ­were central to electoral campaigns: “every­thing is a promise and candidates are often required to make promises that appeal to the material and pecuniary interests of voters” (ibid., 18 March 1910). Opponents also worried that such an expansive definition of corruption gave the governmental majority the means to invalidate races won by opposition candidates. The stringent punishment of vote-­buying was viewed as a ploy to weaken opposition parties while strengthening dominant Republicans. Monarchists argued that “­there is no single candidate of the opposition who is not susceptible to fall ­under the incidence of the new law. The text which we are voting on gives the majority the power to invalidate all

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 123

elections won by adversaries.” ­Others offered that this “was not a law that ends corruption that uses money” but rather an attempt to use this opportunity to weaken the opposition ( JORF ChamberDeb, 17 March 1910). A high priority for legislators who drafted the new law was to “prevent rich candidates from hiring ballot con­vey­ors and from transforming ­these ballot con­vey­ors into agents of propaganda” ( JORF ChamberDeb, 18 December 1901). The solution to this prob­lem was to impose the same punishment for irregularities committed by both candidates and brokers. Conservative legislators but also some Republican politicians objected to this recommendation. In an intervention in the Chamber of Deputies, Aristide Briand, a Republican from Loire, argued that this solution did not fully consider candidates’ ­limited agency over and inability to constrain the actions of their brokers. “If candidates act by themselves, they can take all desirable precautions about their acts, gestures, and words. However, one cannot always control persons who are excessively active or zealous” (ibid., 18 March 1910). ­Others, by contrast, argued that by sanctioning promises, the law l­ imited candidates’ ability to recruit brokers. Joseph Lasiez, a Republican from Gers, worried that “with ­these excessive formulations, nobody ­will agree to carry our ballots. But they w ­ ill also not agree to carry your ballots, Socialist gentlemen! Be careful! [Messieurs les Socialistes! Prenez y garde!]” (ibid., 17 March 1910). Opponents of this solution remained unsuccessful in bringing about a change to the proposal during the parliamentary deliberations. The electoral law that was ­adopted imposed the same punishment on both candidates and their brokers. In the French parliament, the discussion of the legislation to limit vote-­ buying is brief and provides us with a single opportunity to use roll-­call data to characterize demand for reforms. The vote took place on 29 March 1902 ­after the first discussion of the Odilon-­Viviani proposal in the Chamber of Deputies. The results presented in the four panels of figure 4.2 examine how considerations about relative resource advantages, on the one hand, and about the economic and electoral costs of vote-­buying, on the other, explain demand for electoral reforms. Panel (a) of figure 4.2 considers relative resource advantages. As bribery and treating w ­ ere financed by candidates using their own resources, one expects that individual-­level attributes mea­sur­ing the resource asymmetries w ­ ill predict support for electoral reforms. Wealthy candidates are expected to ­favor the status quo, while resource-­constrained candidates are expected to support the adoption of reforms. I proxy the relative resource advantage with a variable that takes the value 1 if the legislator is an industrial or agricultural owner.

124 C H A P TER 4

(b) Economic constraints

(a) Individual and partisan resources Individual resources Owner

Economic constraints Rural employment

Partisan resources Nonpartisan

Urbanization

Socialist

Voting population (ln)

Radical Republican

Landholding inequality

Right –5

0

–5

5

0

5

10

(d) Full model

(c) Electoral constraints

Partisan resources Nonpartisan Socialist Radical Republican Right

Electoral constraints

Electoral constraints Programmatic promises Economic constraints Rural employment Urbanization Landholding inequality Voting population (ln) Competitiveness Runoff Mayor

Programmatic promises

Individual resources Owner –1

0

1

2

3

–10

–5

0

5

10

15

The determinants of support for reforms to limit vote-­buying in the French Chamber of Deputies. Notes: Results pre­sent correlates of support for reforms to limit vote-­buying during a vote taken in the French Chamber of Deputies on 29 March 1902 . All models are estimated using Firth logistic regression. See the appendix for the sources used to construct the variables used in the analy­sis. FIGURE 4.2.

Partisan differences about the desirability of reforms also reflect considerations about resource endowments. Resource-­endowed Conservatives and Republicans are likely to oppose reforms, while resource-­constrained Radicals and Socialists are likely to support reforms. We find support for ­these hypotheses. Individual-­level resource differences among candidates predict support for—or opposition to—­reform. Non-­Republican parties on the right opposed proposals to impose higher sanctions on vote-­buying. By contrast, “resource-­constrained” Socialist and Radical legislators supported proposals to sanction vote-­buying. Centrist Republicans ­were divided about the desirability of reforms. In panel (b) of figure 4.2, I examine the importance of economic constraints for support for reforms. This analy­sis tests the hypothesis advanced by Susan Stokes and her coauthors (2015) that economic development ­limited the use of vote-­buying. We find mixed support for this hypothesis.

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 125

While legislators representing more populous districts are more supportive of electoral reforms, the variable does not reach statistical significance at conventional levels. Consistent with the predictions of the modernization perspective, however, we find that legislators representing districts with higher levels of agricultural employment are more likely to oppose ­these reforms. Other variables mea­sur­ing economic development—­such as the level of urbanization—­are not systematically correlated with support for electoral reforms. The results reported in panel (c) examine the importance of electoral costs in shaping support for reforms. Electoral costs are proxied h ­ ere using a mea­sure of the programmatic promises of candidates in their professions de foi. I conjectured that candidates competing on the basis of programmatic promises find it electorally costlier to use illicit strategies. For ­these candidates, reforms that impose higher sanctions on illicit exchanges are advantageous ­because they constrain competitors who may incur lower electoral costs from using such strategies. The results presented in figure 4.2 test this conjecture. The bivariate correlation shows a positive relationship between the programmatic commitment of a candidate and support for electoral reforms. However, the relationship is not robust to the inclusion of additional controls. Panel (d) in figure 4.2 includes the full model. Electoral reforms limiting vote-­buying w ­ ere supported by resource-­constrained politicians and resisted by resource-­endowed legislators. We find support for the conjectures of the modernization perspective. Politicians from more populous and eco­nom­ically developed districts ­were more supportive of reforms to limit vote-­buying. The composition of the electoral co­ali­tion supporting vote-­buying reforms contrasts with the co­ali­tion that came about during the reform of the candidature officielle discussed in chapter 3. In both cases, considerations about access to po­liti­cal resources predicted support for reforms, leading to dif­fer­ent positions taken by parties on the right ­toward ­these reforms. While Conservative legislators supported reforms to limit the use of state resources, they resisted the introduction of reforms to limit vote-­buying. By contrast, Radicals and some centrist Republicans supported reforms to limit vote-­buying but resisted reforms to limit patronage. By documenting heterogeneity in the position of rightist parties t­ oward electoral reforms, t­ hese findings contradict an impor­tant assumption in the lit­er­a­ture on democ­ratization that assumes that politicians connected with wealthy assets consistently oppose demo­cratizing reforms. This heterogeneity in the positions of parties on the right also demonstrates that one cannot infer legislators’ positions ­toward electoral reform

126 C H A P TER 4

from their connection to vari­ous economic ­factors—­such as land—or assume that wealth always translates into opposition to reforms. The electoral reform to limit bribery was ­adopted by the French parliament in March 1914 and went into effect during the same year ( JORF Lois et Décrets, 1 April 1914). The law departed from the status quo in two ways. First, it reformed some of the contradictions of the existing jurisprudence by removing the requirement to prove an intention to buy votes. At the same time, the new law made it pos­si­ble to sanction as corrupt exchanges offers that had been made by a candidate but had not yet been accepted by voters. Second, the electoral reform made it pos­si­ble to sanction non-­ targeted exchanges between brokers and voters. This was an impor­tant l­ egal modernization that extended sanctions for an electoral irregularity that had grown in importance during the last de­cades of the nineteenth ­century. Reforming Vote-­Buying in Britain

Bribery and treating w ­ ere common electoral practices in British elections, which continued ­after the adoption of secret voting. Candidates competing in British elections offered voters highly heterogeneous goods, including meat, coal, and groceries (Rogers 1880: 232). In an effort to ascertain voters’ demands, candidates canvassed individual voters to inquire about their needs and their willingness to engage in a transaction with the candidate. Crompton Roberts, a candidate competing in Sandwich, talked to voters in person about their needs. His handwritten notes include the following information written next to the names of voters: “Very favorable and poor,” “Promised, wants a ­little drop,” “Wants to be seen, cash,” “Wants much assistance; had much illness in the ­family, half a year’s rent at 3s = 2l. 18s. 6d.” ­Because some of ­these entries also included notes written by agents, one can also infer that brokers acted upon ­these suggestions (H.C. Parliamentary Papers 1881, c2776). In addition to in-­kind goods, candidates offered voters money in exchange for their vote. Some of ­these practices originated with ceremonial offers made by candidates to voters in Elizabethan elections, which then “mutated into a system of head money in which candidates paid their supporters a fixed amount of money” (Kam 2017: 597). ­These amounts had dif­fer­ent names. At Berwick-­Upon-­Tweed, the customary payments made by candidates to voters w ­ ere called “gooseberries” and ranged from £2 to £3 during the 1850s (ibid.). Candidates and their brokers developed a number of strategies to monitor voters’ compliance with t­ hese offers. One strategy brokers used to ensure

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 127

that voters would not renege on their promises was to offer them money immediately before the vote. In Macclesfield, during the 1874 election, election agents secured empty ­houses or sometimes pieces of land in the vicinity of the polling booth. H ­ ere, agents sat during the day of the vote, distributing money to voters. As the electoral report describes ­these transactions, “voters ­were being brought to the tent, bargained with and paid and then sent on straight into the polling booth” (H.C. Parliamentary Papers 1881, c2853). Candidates monitored t­ hese vote-­buying exchanges, trying to reduce voters’ ability to renege. Consider the sophisticated monitoring strategies developed in Gloucester during the 1874 election. Canvassers brought voters to the committee room, the voters stated their names and number on the register, or produced their cards. The person in charge examined the register to see that the number was correct, and then, marking off the name of the voter on the register, gave him a slip of paper with his number on one side of it, and on the other side of it a mark, signifying the amount of money he was to receive. The amount was usually a sovereign, but in some cases as much as 2l. and in ­others as l­ittle as 5s. The voter took this ticket and ­either handed it through a hole cut in the door or some other opening into an inner room, or himself went into the inner room, in which a person in disguise was sitting, and ­there received the sum of money marked upon his ticket. He was then conducted by messengers for the purpose to the poll. In one of the committee rooms, No. 1A, the methods ­were so far varied that the voter on his first visit obtained his ticket and was taken to the poll and did not obtain his money ­until his return. (ibid., c2841) While targeted vote-­buying was a common practice, the average price of a bribe declined significantly during the second half of the nineteenth ­century (Kam 2017). Figure 4.3 shows the average price of bribes recorded in the electoral petitions at the time, expressed as a function of weekly earnings. Christopher Kam collected and generously shared this data using information from the petitions filed in the House of Commons. The price of bribes experienced a significant decline ­after the introduction of secret voting in 1872. One pos­si­ble explanation for t­ hese trends is that a­ fter the secret ballot, bribery became a much riskier electoral investment. The new price reflects politicians’ estimate of the increased riskiness of ­these investments. In addition to bribery, candidates engaged in non-­targeted treating of voters. Treating came in the form of offers of food and drink freely and without disguise during campaigns. In Britain, most of the treating occurred in

128 C H A P TER 4

15

Average bribe

10

5

0 1865

1870

1875

1880

1885

Year Average prices of bribes in contested British elections relative to the price of weekly wages. Note: The average bribe price in elections as a function of 1906 earnings. Source: Kam 2017 using data generously provided by the author. FIGURE 4.3.

public ­houses, which w ­ ere engaged as the committee rooms of parties. The parliamentary report on electoral irregularities in Sandwich describes the vigorous competition for control of ­these spaces among candidates: “The first step taken by Mr. Edwin Hughes, for which indeed the ground had been prepared before his arrival, was to secure a majority of the public ­houses, and with them the l­ittle coterie of customers which each landlord could influence. Seventy-­one committee rooms in public h ­ ouses are engaged in Deal and Walmer and eigh­teen in Sandwich” (H.C. Parliamentary Papers 1881, c2796). Public ­houses “swarmed with fighting, cursing and drunken electors. . . . ​Electors ­were welcomed to drink as much as they chose to at the expense of the candidate. This practice culminated on the day of election, in a carnival of drunkenness and crime” (Wilkinson 1884: 130). The importance of treating relative to bribery increased over time. A report examining electoral irregularities in Macclesfield during the 1880 election noted that “this election seems to have been mainly distinguished by the open and undisguised treating which prevailed throughout the borough on behalf of all the candidates. Public h ­ ouses appear also to have

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 129

been opened on behalf of all the candidates on e­ very side, and t­ here was indiscriminate treating of voters and non-­voters both before and ­after the poll” (H.C. Parliamentary Papers 1881, c2853). Several historians of British electoral practices attribute this increase in the relative importance of treating to an increase in the size of constituencies (O’Malley and Hardcastle 1893, vol. 4; Day 1894: 150). Charles Seymour argues that “with the growth of the electorate, collective corruption becomes the eco­nom­ical means of influencing voters and ­under the ballot proved as productive of results as individual bribery. Treats, picnics and entertainment became the custom” (1915: 438). A statement made by a justice discussing irregularities in North­umberland describes the substitution that occurred between t­ hese two strategies as follows: Now that the constituencies are so large it becomes impossible successfully to bribe. Bribery does not produce any result except on the part of ­those who actually are bribed. No one is prepared to go to vote for a man ­because he is said to be willing to pay for votes ­unless he has got an agreement beforehand. With reference to treating it is other­wise, a very small amount is sufficient to procure a g­ reat deal of popularity, b ­ ecause, looking at the very wide extension of the franchise, ­there are in ­every constituency a considerable number of men who do not take politics seriously at all, or attach much importance to one side or the other. They are perfectly ready to vote for the man who is popu­lar, and if by reason of treats and picnics you can produce a general feeling that the candidate is a good fellow and that he is willing to give a poor man a supper or treat or entertainment of this kind, and if that idea gets generally spread over the division an enormous amount of popularity is produced by it as against another association which does not resort to that sort of ­thing. (O’Malley and Hardcastle 1893, 4:145) A second interpretation of the rise of non-­targeted expenses suggests that they ­were a form of turnout-­buying. Turnout-­buying is a clientelistic strategy premised on the indirect monitoring of turnout rather than the vote (Nichter 2008). Christopher Kam has argued that the transition from open to secret voting in 1872 created incentives for candidates to invest more resources in turnout-­buying (Kam 2017: 628). Kam shows that when controlling for candidates’ expenditures, turnout increased by about 3 ­percent ­after the adoption of the secret ballot.

130 C H A P TER 4

This increase in the importance of turnout-­buying required a high number of brokers. To mediate their relationship to voters, candidates maintained oversized networks of brokers. Candidates campaigned by hiring an army of canvassers, messengers, and bill posters (O’Malley and Hardcastle 1893, 3:155). Commenting on ­these practices, Seymour noted that “the corrupt employment of voters was not unusual and had the same practical effect as direct purchase” (1915: 436). The royal commission that investigated electoral irregularities in Norwich in 1874 commented on the “replacement of cash bribery by an increase in the number of brokers,” a practice referred to at the time as “colorful employment” (Kam 2017: 622; H.C. Parliamentary Papers 1876, c1442). For example, during an election in Cambridge, one agent employed as many as seventy-­four messengers (Rogers 1880: 303). In Salisbury, the candidate employed over two hundred persons as clerks, messengers, or bill posters (O’Malley and Hardcastle 1893, 4:22). Altogether, the increased importance of treating and of “colorful employment” contributed to a rise in campaign expenditures devoted to vote-­ buying. Noting this increase in campaign expenditures, a parliamentary commission wrote, “Your commissioners have to state that while corrupt practices extensively prevailed at ­every election into which they felt it their duty to inquire . . . ​, the corruption at the late election was far more widespread, and far more open than had been the case at any previous parliamentary election, at all events of recent years, though the bribes ­were, in most cases, trifling in amount” (H.C. Parliamentary Papers 1881, c2853). ­Table 4.2 documents the changes in campaign expenditures during elections following the adoption of the secret ballot. Vote-­buying and treating ­were electoral strategies pursued by both parties. Nevertheless, reports of electoral expenses assembled by the electoral commission at the time reveal that Conservatives devoted more resources to vote-­buying as compared to Liberals. While the average expenses on agents of Liberal candidates stood at £755.57, Conservatives’ expenses on brokers stood at £1,063.03. Overall the average campaign expenditures incurred by Conservatives exceeded the expenditures of Liberal candidates by 50 ­percent. This asymmetry in resources remains a leading ­factor in accounting for Liberals’ decision to take an additional stab at vote-­buying reforms. We can probe deeper and investigate the economic and po­liti­cal correlates of this resource asymmetry. Where was this resource asymmetry most pronounced? Who w ­ ere the candidates experiencing it most strongly? To address t­ hese questions, I constructed a mea­sure of the resource constraints experienced by dif­fer­ent legislators. The denominator of this variable

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 131

­TABLE 4.2.

Partisan asymmetry in vote-­buying expenses Liberal candidates

Conservative candidates

Average agent expenses

£ 755.57

£1,063.03

Average printing expenses

£ 463.13

£ 509.04

£1,208.26

£1,856.27

Average total expenses

Source: H.C. Parliamentary Papers 1880, no. 282. Election charges. Return to charges made to the candidates at the late elections by returning officers.

represents the wealth of a legislator, which is mea­sured using information from the ­will of each politician. The numerator is a mea­sure of the campaign expenditures incurred by candidates during the 1874 election. ­These include expenditures incurred for agents, including clerks, messengers, and canvassers, expenditures for printing and advertising, and so on. The source of this information is mandatory returns filed by all candidates elected to the House of Commons (H.C. Parliamentary Papers 1880, no. 382). This additional variable documents another dimension of the resource asymmetry among parties. During the 1874 election, Liberal candidates ­were spending less than Conservatives on campaigns. At the same time, Liberals ­were spending a higher share of their wealth. On average, Liberal candidates ­were spending 25 ­percent more of their wealth than Conservatives during a race, and this difference is significant at conventional levels. In figure 4.4, I examine the economic, po­liti­cal, and demographic correlates of this resource constraint. ­Here I regress the mea­sure of resource constraint experienced by a candidate against vari­ous district-­level covariates, such as economic development, landholding in­equality, and population density. The results indicate that Liberal and Conservative candidates experienced this resource constraint in dif­fer­ent constituencies. For Conservatives, the resource constraint was less pronounced in constituencies with high levels of landholding in­equality. In ­these constituencies, the presence of “deference voting” reduced the need to bribe and treat. By contrast, Liberals experienced stronger resource constraint in constituencies with a higher share of manufacturing employment and higher population growth. The ongoing rapid economic development thus worked against Liberal candidates and amplified the resource constraint they experienced. ­These perceptions of a widening resource asymmetry explain the urgency with which Liberals approached the question of electoral reforms. The Liberal victory during the 1880 election created a new opportunity to implement such reforms.

Correlates of resource constraints for liberals Landholding inequality Proportion urban Wealth of constituency Voting population (ln) Manufacturing employment Population growth Population density Literacy –5

–10

5

0

10

Correlates of resource constraints for conservatives Landholding inequality Proportion urban Wealth of constituency Voting population (ln) Manufacturing employment Population growth Population density Literacy –4

–2

0

2

Correlates of resource constraints for Liberal and Conservative candidates. Notes: Results pre­sent differences across the Conservative and Liberal parties in the district-­ level conditions that predict a higher resource constraint for candidates who rely on vote-­ buying. The resource constraint is mea­sured as the ratio between a candidate’s campaign expenditures and his wealth. FIGURE 4.4.

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 133

Negotiating Electoral Reforms

Even during the period of l­imited suffrage, bribery played an impor­tant role in British elections. The earliest effort to limit vote-­buying dates to the Treating Act of 1677 (Rogers 1880, 2:294). The Corrupt Practices Act, ­adopted in 1854, was the first attempt to define corrupt electoral practices. The Act placed decisions concerning the voiding of elections in the hands of parliamentary commissions (Seymour 1915: chap. 8). Another reform, introduced in 1868, transferred the responsibility to the courts to rule in cases of electoral malfeasance. Commenting on this decision, Benjamin Disraeli, who was prime minister at the time, expressed his hope that by “insuring a local investigation of the m ­ atter” the law would “terminate corruption altogether” (Hanham 1959: 274). The Ballot Act of 1872, which introduced secret voting, played a significant role in reducing electoral intimidation. As discussed above, in response to the increased uncertainty as to w ­ hether voters would deliver on their promises, candidates reduced the level of bribes and invested in monitoring institutions. However, bribery continued even ­after the adoption of the Ballot Act. Commenting on the continuation of vote-­ buying, a commissioner investigating electoral irregularities noted that “while the Ballot Act enabled many voters to take bribes on both sides, it did not, as far as we could ascertain, render a single person unwilling to bribe for fear or bribery in vain” (H.C. Parliamentary Papers 1881, c2796). Developments during the 1880 election—or what historians called the “saturnalia of 1880”—­were the impetus for additional reforms seeking to clean up malfeasance (Shannon 1996: 76). Liberals who won the election appointed several royal commissions tasked with examining electoral irregularities. According to the official returns of candidates’ election expenses, the total amount of expenditures in 1880 stood at £1,737,300 (O’Leary 1962: 206). This was an underestimate; the investigations carried out by the House of Commons estimated a dramatic increase in campaign expenditures relative to 1874 (Rix 2017: 70). The investigations of the royal commissions also focused on several districts that experienced outrageously high levels of malfeasance—­Boston, Canterbury, Chester, Macclesfield, Oxford, and Sandwich—­and recommended temporarily disenfranchising ­these constituencies. In Macclesfield and Sandwich, about half of the voters had been found guilty of corruption (H.C. Parliamentary Papers 1881, c2796, c2853). I expect this resource asymmetry to translate into strong disagreement regarding the desirability of electoral reforms to limit campaign expenditures.

134 C H A P TER 4

Prominent Liberal candidates supported the adoption of electoral reforms to limit electoral corruption (O’Leary 1962). Many Liberal politicians voiced disappointment about voters who would, in Labouchere’s words, “forego their birthright for the sake of the ­free teas, the conjurors, the rope-­dancers, the fire-­works and the comic singers . . . ​provided [by the Conservatives]” (Lawrence 1998: 201). They expressed hope that electoral reforms “would encourage appeals to voters on the basis of reason” (Rix 2008: 93). By contrast, many Conservatives opposed the new legislation. In a private letter to Salisbury, W. H. Smith, a former secretary of the trea­sury, expressed this opposition most forcefully, arguing that “to stave off the evil day as long as pos­si­ble, I should wish to retain the power of fighting elections by paid agency if necessary” (Feuchtwanger 1968: 159). Henry Cecil Raikes, a former speaker of the House of Commons, also opposed legislators’ intention to “proscribe the comparatively legitimate influence of wealth at elections” (1883: 204). A second indicator of Conservatives’ higher tolerance for the illicit practice was the indictment of prominent Conservative politicians for illicit electoral practices in the years prior to the adoption of the Corrupt Practices Act (Hanham 1959: 277). The House of Commons censured several Conservative Party whips for corrupt electoral practices. Th ­ ese included William Beresford, a former Whip and secretary of war, and Forbes Mackenzie, the Conservative Party Whip. Philip Rose, the Conservative Party agent, also retired in 1859 as a result of damaging revelations regarding his involvement in elections (ibid.). The central goal of the reform considered by British lawmakers was to curtail election expenditures. The solution considered at the time was placing ceilings on the amounts that could be spent and imposing sanctions on ­those who exceeded this ceiling. The proposal to impose strict limits on campaign expenditures originated with resource-­constrained Liberal politicians. In several interventions on the floor of the House of Commons between 1881 to 1883, Henry James, the attorney general and architect of the bill, formulated two arguments justifying t­ hese limits on expenditures. The proposed legislation, James argued, could not only reduce excessive spending and electoral corruption but also lay the foundation for more far-­ reaching changes that could ultimately transform po­liti­cal repre­sen­ta­tion: In the first place, this large expenditure deprived the House of the ser­ vices of many men who would be useful in the House; secondly, it threw an unnecessary burden upon ­those who ­were ­there; and, thirdly, the

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 135

expenditure itself led to corrupt practices. It was almost impossible to draw a line between the expenditure that was lavish and extravagant in a g­ reat degree, and that which involved the commission of corrupt practices. (H.C. Debates, 24 April 1884) James’s argument restated a recurring theme of Liberal politicians of the period who argued that excessive spending biased po­liti­cal repre­sen­ta­tion by restricting the ability to run for office to the very rich candidates. During an 1877 intervention, Gladstone expressed his fears that the good effects of the vast increase in the electorate might be neutralized by a corresponding increase in the costs of electioneering, which would keep out of Parliament all but the very rich: “The two circumstances which strike me most forcibly and most painfully, are, first, the rapid and constant advance of the money power; secondly, the reduction almost to zero, of the chances of entrance into Parliament for men who have nothing to rely upon but their talent and character” (Gladstone 1877: 557). Other Liberals argued that limits on campaign expenditures could “encourage men of ability and character, but of moderate means, to enter public life and devote themselves to the ser­ vice of the country” and that “if the law could get rid of the influence of wealth or diminish the chances of success of a rich man against a poor man . . . ​, they would obtain a much truer repre­sen­ta­tion in the House of Commons and would return to the House ­those who would represent the majority of the ­people and not so much the wealth of the country” (H.C. Debates, 25 April 1882). While such changes in repre­sen­ta­tion ­were desirable in and of themselves, Liberals w ­ ere clearly aware that they might result in a partisan advantage: a higher number of “men of g­ reat ability and high character” who ­were likely to enter politics would compete for the Liberal Party (ibid.). By contrast, Conservatives opposed limits on campaign expenditures. They argued that ­these limits ­were too rigid and did not take into account the heterogeneity among constituencies. Lord George Hamilton expressed this opposition, arguing that “the Bill practically put all the constituencies of ­England in swaddling clothes ­because a certain number of small boroughs had shown themselves unable to walk uprightly without assistance” (H.C. Debates, 25 April 1882). Other Conservative MPs called for a more flexible mea­sure that could accommodate “the interests of dif­fer­ent constituencies, entirely distinct in size, differing in the density of their population, and varying in their circumstances” (ibid.). This required defining “more elastic limits” than did the current law that could be fixed with reference to the ­will, character, and circumstances of constituencies. Other critics pointed

136 C H A P TER 4

out that the rigid definition of spending limits created strong incentives and possibilities for evasion, which would defeat the purpose of the law. Conservative MPs also argued that the law established unrealistic limits for spending (H.C. Debates, 25 April 1882), which severely hampered the ability of candidates to hire professionals to manage electoral campaigns. Conservatives complained that u ­ nder the current spending limits, it would be impossible for candidates to hire solicitors or other specialized personnel (ibid., 4 June 1883). Even the limits on printing costs ­were considered too low. It was not pos­si­ble to adhere to such a narrow limit ­because, “in the heat of the campaign, parties w ­ ere driven into issuing posters and circulars,” the cost of which exceeded the limit. However, not all Conservatives w ­ ere opposed to spending limits. MPs from more populous constituencies w ­ ere relatively more supportive of spending limits as compared to legislators representing smaller constituencies. As one such MP argued, the best way to reform corruption was to increase the size of the constituency up to a point where bribery would simply be too costly. MPs representing smaller constituencies opposed the proposed spending scales defined in the law, considering the limits too constraining. One Conservative MP criticized the scale of the bill for “allowing only a niggardly amount for [an] intermediately sized borough” (H.C. Debates, 4 June 1883). However, Liberals ­were unwilling to compromise on this issue and introduced further differentiation of the allowed expenses for constituencies of dif­fer­ent sizes. As Henry James argued, “the Bill dealt with no one unfairly but left the same and equal weapons to all parties” (ibid., 24 April 1883). In addition to imposing limits on campaign expenditures, the law attempted to reduce campaign expenses that w ­ ere considered wasteful. One such expenditure targeted by reformers was conveyance, incurred by candidates who brought voters to polling places in private carriages. While conveyance expenses made up a large share of election costs, parties differed significantly in terms of the use of this strategy. Conservatives owned and made use of “private carriages and motor cars, blazoned with ancient armorial bearings,” and retained a resource advantage relative to Liberal candidates in accessing the stock of such vehicles (Lowell 1908, 1:226). Lacking such resources, Liberal candidates often had to hire conveyers, which increased the costs of campaigning. The resource asymmetries among parties also informed disagreement over proposals to reduce campaign expenditures. On the one hand, resource-­constrained Liberals recommended that arrangements to bring

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 137

voters to the polls be considered illegal electoral practices. While illegal practices ­were sanctioned with lower penalties than ­were corrupt practices, they nevertheless imposed significant punishments on candidates. If such an offense could be proven, the candidate was prohibited from representing the constituency for seven years. If the irregularity had been committed by the agent, the candidate was prohibited from serving in the current parliament (Seymour 1915: 443). By contrast, Conservatives opposed such proposals, arguing that they would disenfranchise voters. In the words of one MP: ­ ere w Th ­ ere a large number of gentlemen possessing property in the counties who lived considerable distances from the polling booths, who would prob­ably not come at all if they ­were compelled to come at their own expense. Many of ­those persons ­were small proprietors, who would exercise their right to vote without undue influence; but u ­ nder this Bill they would be obliged to travel a long distance at their own expense, which it was very unlikely they would do. The Bill creates injustice on a very respectable and worthy class of voters. However, few MPs w ­ ere willing to accept this argument. As a result, a number of amendments formulated by Conservative MPs seeking to define special exemptions allowing candidates to bring voters to the polls ­were defeated. The argument that ultimately prevailed was that voters w ­ ere quite willing to come to the polls without any external support and that demand for conveyance was low. The final provision that was included in the law prohibited candidates from transporting voters to the polls using private carriages (Lowell 1908, 1:226, 46–47; Vict. c51 paras. 7, 14). Treating was extremely pervasive in British elections. During the second half of the nineteenth ­century, this practice became even more impor­tant. Reforming treating was, thus, a crucial ele­ment of the Corrupt Practices Act. One initial proposal to limit treating was to close public ­houses on election day. Proponents of this solution pointed to cases where local officials had enforced such provisions, which led to a reduction in the number of riots and incidents of disorderly conduct (H.C. Debates, 24 April 1882). This proposal was rejected, however, as excessive. The alternative ave­nue of reform that was pursued by lawmakers sought to eliminate the differential punishment imposed on treating relative to bribery. Lawmakers proposed counting treating as an electoral offense that was liable to the same electoral punishment as bribery. If a candidate had been found guilty of treating, he could not return to the House of Commons as a representative of the same constituency. By contrast, the law imposed a less severe punishment

138 C H A P TER 4

if the agents hired by the candidate w ­ ere responsible for treating. In this case, the candidate could not return to the same constituency for a period of seven years (Seymour 1915: 445). We also find a pronounced partisan disagreement over the reform of treating during the deliberations in the House of Commons. Prominent Conservative politicians opposed strict penalties for treating, arguing that it was difficult to distinguish between hospitality and treating. Liberal MPs, by contrast, objected to proposals to punish voters who engaged in treating and worried that the poor ­were being punished excessively. Consider the remarks of Henry Samuelson, a politician representing Frome: To punish a very poor person, who was tempted by ­those who ­ought to know a g­ reat deal better than to seek to obtain his vote by bribery and treating, seemed to him to be punishing the wrong man. The technical offence of treating, by accepting meat, drink, entertainment, and provision, so far as elections ­were concerned, was an entirely new one; it had not existed before, but had been created by this Bill; and he thought it would be found difficult to make an average En­glishman of the poorer classes understand that he was committing a sin against the law when he accepted a glass of beer, or some other refreshment, from persons whom he would naturally consider to be in ­every way capable of giving him moral as well as material guidance. (H.C. Debates, 25 June 1883) Despite ­these objections, legislators de­cided to push ahead and impose a strict punishment on treating. A majority of MPs who supported the decision argued that the provision was necessary to limit “indiscriminate treatment” that resulted from candidates’ decisions to keep public ­houses open throughout the period of the campaign (H.C. Debates, 27 April 1883). Fi­nally, the Corrupt and Illegal Practices Act addressed the prob­lem of the overemployment of brokers. As discussed above, such practices of “colorful employment” ­were a pervasive prob­lem in nineteenth-­century British elections and contributed to the rise of electoral costs. One egregious example of overemployment, which was invoked repeatedly during the parliamentary deliberations, was a candidate who had employed as many as eight hundred canvassers and poll clerks in a constituency that had only a few thousand voters (H.C. Debates, 24 April 1882). Prominent Liberal politicians, such as Henry James, viewed ­these issues as an opportunity to push for an organ­ization of campaigns that relied entirely on volunteers. Defending this proposal, James argued that if candidates “could enlist the ser­vices of volunteers, instead of ­these paid agents, they would get rid of the class

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 139

who now sets an example of corruption, and ­were the means of corrupting ­others; while, at the same time, they would obtain a much juster repre­sen­ta­ tion” (ibid.). Other Liberal MPs praised campaigns that enlisted volunteers as examples of elections “animated by enthusiasm.” ­These speakers contrasted two contests. In the first contest, the candidate “had paid agents all round him who imposed blackmail upon him at e­ very stage of the contest.” In the second race, candidates relied on tenant farmers as volunteers. ­These volunteers “brought their gigs and their dog-­carts, and carried the voters to the poll, and they carried the election for the smallest conceivable amount of expenditure.” Liberals advocated for reforms that could allow candidates to rely on volunteers rather than paid agents. Prominent Liberal politicians expressed the hope that such a change could, in the long run, transform po­liti­cal repre­sen­ta­tion and “could get rid of the influence of wealth, diminish the chances of the success of a rich man as against a poor man and . . . ​ obtain a much truer repre­sen­ta­tion in the House of Commons” (ibid.). Prominent Conservative MPs objected, however, to this proposal to substitute paid brokers with volunteers. While “sympathizing with the desire to make elections cheaper,” Balfour argued that organ­izations such as caucuses ­were “immoral associations” that w ­ ere as evil as corruption (H.C. Debates, 24 April 1882). The opposition to the organ­ization of a campaign based on volunteers was rooted in the perception among prominent Conservative politicians that their party was at a disadvantage in establishing such organ­ izations relative to Liberals. As W. H. Smith, a prominent Conservative politician, wrote to Salisbury in 1881, “I do not like the Corrupt Practices Bill and I do not think it ­will work well for our friends if it becomes law. You know quite as much as I do of the habits and customs of the Tory, but I have not found them to be e­ ager volunteers in canvassing organ­ization. . . . ​The Radicals have the Trade Unions, the Dissenting Chapels and e­ very society for the abolition of property and morality working for them” (Feuchtwanger 1968: 159). Smith also argued that Conservatives could not match the resources of Liberals in terms of volunteers and warned of Conservative defeat: “The result ­will be I am afraid that we ­shall not poll anything approaching our strength” (ibid.). With few exceptions, prominent Conservative politicians opposed any mea­sures that allowed candidates to rely on support from a local party organ­ization or a caucus. Conservatives’ concern about their inability to develop effective party organ­izations turned out to be excessively pessimistic. Only one and a half de­cades a­ fter the passage of the Corrupt and Illegal Practices Act, Conservatives reor­ga­nized their party around voluntary associations such as the Primrose League.

140 C H A P TER 4

Liberals’ proposals to limit the number of agents ­were successful. The a­ dopted reforms restricted the number of brokers that could be hired at elections to one polling agent at each booth. The law also l­ imited the number of messengers and clerks in proportion to the number of electors (Seymour 1915: 443). Yet, perhaps the most influential innovation proposed by the law required each candidate to hire a single election agent who was responsible for the disbursement of election expenses. The Corrupt and Illegal Practices Act mandated that election agents manage all expenses incurred by a candidate, except his personal expenses (paras. 28, 31, 27). This concentration of responsibilities in the hands of the election agents was an impor­tant innovation that enhanced compliance with the new law, ensuring its success in limiting campaign expenditures. The deliberations about parliamentary reforms illustrate the strong partisan disagreement with regard to all components of electoral reform. Proposals to limit campaign expenditures, to eliminate unnecessary campaign expenditures, and to replace paid workers with volunteers w ­ ere highly contested. I turn now to an analy­sis of the roll-­call votes taken during the deliberations of this bill; this enables me to examine the demand for electoral reforms using individual data. In my analy­sis, I consider twenty votes over vari­ous dimensions of the reform taken in the British lower ­house between 1881 and 1883. The hypotheses presented in chapter 2 explain demand for reform as a function of access to resources, on the one hand, and economic and electoral constraints, on the other. I expect legislators and parties that enjoy a relative resource advantage to support the status quo and legislators who enjoy a relative resource disadvantage to support reforms. I mea­sure ­these resource asymmetries using both individual-­and partisan-­level proxies. At the individual level, I proxy the resource advantage using a variable mea­sur­ ing the wealth of the legislator. Information on wealth is obtained by coding the probates of all legislators elected to the House of Commons in 1880 (National Probate Calendar, vari­ous years). An e­ arlier discussion illustrated that Liberal politicians faced a relative resource disadvantage in financing vote-­buying expenses relative to Conservatives and that this resource disadvantage was concentrated in constituencies that experienced economic development. If resource asymmetries are salient, I expect resource-­ constrained Liberals to support electoral reforms and resource-­endowed Conservatives to oppose them. Reforms to limit vote-­buying are a crucial case for modernization theories of democ­ratization (Stokes et al. 2015). In an influential formulation of

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 141

the modernization hypothesis, Susan Stokes and her coauthors argued that economic development “killed vote-­buying in ­England.” They conjectured that the rising incomes during the pro­cess of modernization, the increase in population associated with the pro­cess of urbanization, and changes in the occupational makeup of districts increased the costs of vote-­buying for politicians. I use a variety of economic variables to mea­sure the consequences of ­these developments for demand for electoral reforms. Th ­ ese include the wealth of the constituency, the size of the electorate, population growth, and the share of manufacturing employment. In chapter 2, I conjectured that elite splits may increase the “electoral costs” associated with certain non-­programmatic strategies and contribute to a po­liti­cal realignment around electoral reforms. The most impor­tant elite split of the period took place in 1885, several years ­after the passage of the Corrupt and Illegal Practices Act (Lubenow 1983). At the time, Gladstone’s decision to embrace Home Rule led to the defection of Liberal Unionists from the Liberal Party. Iain McLean has referred to this development as “the most impor­tant Victorian realignment in electoral politics” (2001: 84–86). I examine the consequences of a less dramatic elite split—­which is associated with the programmatic differentiation of candidates—­for legislators’ preferences for reform. I mea­sure the programmatic orientation of a candidate using a variable mea­sur­ing legislators’ support for the social policies considered at the time. The available vote during this legislative period is one on legislation providing better protection for ­children (Divisions of the House of Commons 1880: 117). The results presented in panel (a) of figure 4.5 consider the importance of resource asymmetries in explaining support for electoral reforms. Individual-­level resource differences—as mea­sured by candidate wealth—do not explain differences in support for reforms. We do find, however, strong differences across parties in their positions ­toward reform that are consistent with expectations following from their relative resource advantage. Demand for electoral reforms originated with resource-­constrained Liberals. By contrast, resource-­endowed Conservatives opposed ­these electoral reforms. The models presented in panel (b) examine the importance of economic conditions in explaining support for the vari­ous proposals of the Corrupt and Illegal Practices Act. The results lend mixed support to some predictions of the modernization perspective. While MPs elected in constituencies experiencing rapid population growth ­were more supportive of electoral reform, the relationship does not reach statistical significance at conventional levels. At the same time, we see no relationship between other mea­sures of

142 C H A P TER 4

(a) Individual and partisan resources

(b) Economic constraints Individual resources

Economic constraints

Wealth of candidate

Landholding inequality Wealth of constituency

Partisan resources

–2

–1

0

1

2

3

4

Conservative

Voting population (ln)

Liberal

Manufacturing employment

Lib/Lab

Population growth

5

–2

–1

0

1

2

(d) Full model

(c) Electoral constraints

Partisan resources Conservative Liberal Lib/Lab

Electoral constraints

Economic constraints Landholding inequality Wealth of constituency Voting population (ln) Manufacturing employment Population growth Electoral constraints Programmatic orientation

Programmatic orientation

Individual resources Wealth of candidate –2

–1

0

1

2

–2

0

2

4

6

The determinants of demand for reforms to limit campaign expenditures. Notes: Results pre­sent correlates of support for the Corrupt and Illegal Practices Act. All models are estimated using multilevel mixed-­models regression. See the appendix for the sources used to construct the variables used in the analy­sis. FIGURE 4.5.

economic development, such as the wealth of the constituency and the size of the voting population, and support for electoral reforms. MPs who supported ­these reforms also ­were elected in districts with a lower share of manufacturing employment. In panel (c), I consider the relationship between the programmatic commitment of a candidate and support for reforms. While the bivariate relationship between ­these variables is statistically significant, the variable loses statistical significance in the fully specified model. The results reported in panel (d) bring together all variables shaping demand for electoral reforms. In this panel, variables mea­sur­ing economic and electoral constraints lose their statistical significance. The cleavage surrounding the introduction of the Corrupt and Illegal Practices Act falls entirely along partisan lines, with resource-­constrained Liberals supporting all aspects of the reform. As discussed above, the urgency with which Liberals took up the issue of electoral reforms can be explained by Liberals’ perception of a widening resource asymmetry relative to Conservatives.

Li m iti n g Vote - B u yi n g a n d Tr e ati n g 143

Economic developments magnified ­t hese perceptions of a widening resource gap ­because the Liberal legislators who spent the largest share of their wealth on vote-­buying competed in districts that experienced rapid economic growth. While economic development does not explain the support for reforms of individual legislators, ­these economic conditions ­were impor­tant in influencing partisan perceptions about resource asymmetries. The adoption of the Corrupt and Illegal Practices Act in 1883 had impor­ tant implications for electoral politics in Britain. Recognizing the significance of the bill, The Times noted during the deliberations, “if passed in its pre­sent form, it can scarcely fail to effect something like a revolution in the mode of conducting Parliamentary elections”; it would succeed in eliminating the “foul and noisome weed of electoral corruption” (The Times, 8 January 1881, 10 January 1881). By reducing vote-­buying expenses, ­these reforms significantly reduced the costs of electioneering incurred by candidates. Summarizing ­these developments, Cornelius O’Leary noted that “on balance, that is the conclusion that must stem from the survey of electoral conditions between 1883 and the end of the ­century. Even if other methods w ­ ere available for spending money in politics, the a­ ctual cost of electioneering was reduced by three quarters which was a ­great advantage to poorer candidates” (1962: 206). The adoption of ­these electoral reforms also spurred far-­reaching changes in the organ­ization of parties. By limiting candidates’ expenses on brokers hired in on-­the-­spot markets, the law created incentives for candidates to invest in professional agents. G. W. E. Russell, a Liberal MP, considered the Corrupt and Illegal Practices Act as a “ ‘dividing line’ between the Ancient (part-­time) solicitor agent and the ‘modern’ usually full-­time professional agent” (H.C. Debates, 27 April 1882). Conclusion

During the second half of the nineteenth ­century, several Eu­ro­pean countries ­adopted electoral reforms that sought to limit vote-­buying and treating. Th ­ ese reforms responded to dif­fer­ent challenges. In most continental Eu­ro­pean countries, penal codes already included significant sanctions for bribery. Such sanctions remained, however, only nominal possibilities, as it was extremely difficult to sanction vote-­buying in the courts. The electoral reforms ­adopted during the second half of the nineteenth ­century attempted to remedy this shortcoming and increase the likelihood that candidates or brokers who campaigned by offering bribes to voters ­were

144 C H A P TER 4

sanctioned. In contrast to continental Eu­ro­pean countries, Britain lacked the history of reliance on a penal code to sanction vote-­buying. As a result, electoral reforms pursued a dif­fer­ent goal. H ­ ere, the main objective of reforms was to limit campaign expenditures. Despite differences in the content of the proposals that sought to limit vote-­buying, we find remarkable similarities in the determinants for support across t­ hese cases. Politicians’ preferences regarding the desirability of vote-­buying reforms reflected, primarily, considerations about their relative resource advantage and their ability to maintain this advantage in the face of rapidly changing economic conditions. Resource-­constrained politicians who lacked the means to offer bribes or gifts to voters or who ­were concerned about the impact of changing economic conditions on their ability to sustain vote-­buying practices supported the adoption of electoral reforms. By contrast, resource-­endowed politicians continued to oppose the adoption of reforms. In contrast to other electoral reforms examined in this book, concerns about the electoral costs associated with bribery or treating did not play an impor­tant role in shaping legislators’ preferences for electoral reforms. The primary sources of conflict about the desirability of electoral reforms confirm some of the predictions of the modernization perspective. Economic development ­limited the attractiveness of vote-­buying, ­because the resources deployed by candidates w ­ ere primarily private. In Britain, economic changes and the growth in the size of electoral constituencies amplified the resource constraints of Liberal politicians, fueling demands for electoral reforms. In France, economic development was also a source of cleavage about the desirability of reforms. Legislators representing districts with higher levels of manufacturing employment ­were more supportive of reforms as compared to legislators elected in rural districts. Considered in conjunction with the findings reported in other chapters, t­ hese results suggest that economic developments reconfigure electoral co­ali­tions that support reforms particularly when such developments magnify asymmetries in private resources among candidates.

5 Protecting Voter Autonomy BALLOTS, ENVELOPES, AND ISOLATING SPACES

In previous chapters I examined the variety of illicit strategies used during campaigns. The moment of voting created distinct opportunities for candidates to subvert voters’ electoral autonomy. Even though electoral laws included a commitment to secret voting, ­these guarantees remained only nominal promises. At the time voters approached the electoral urns to cast their ballots, they ­were surrounded by representatives of candidates seeking to pierce the secrecy of the vote. Significant imperfections in voting technology—­such as the design of ballots and urns—­facilitated such violations of the secrecy of the vote. As a report of a commission in the French Senate noted, irregularities that occurred at the moment voters cast their ballots illustrated the “insufficiencies of material guarantees against coercion, intimidation and corruption” ( JORF SenateDoc 1905, no. 62). To take advantage of imperfections in ballot design, candidates chose brokers who could impose harsh punishments on voters who made the “incorrect” choice. In an effort to maximize the punishment on voters who voted incorrectly, candidates most often chose employers as brokers. In nineteenth-­century elections violations of voting secrecy went hand in hand with economic coercion, and the economic sanctions imposed by employers included reductions in wages and, often, layoffs. An early proposal for electoral reform in France submitted by Charles Guillaume Boysset, a 145

146 C H A P TER 5

Republican from Saône-­et-­Loire, described such practices of economic coercion as follows: “It has been forcefully signaled that the pressure exercised on masses of voters for a long time infringes on the freedom of consciousness. The large landowners, the power­ful employers, and, in general terms, the men who hold in their hands employment, in other words, the bread of families have attempted to use all sorts of influence to falsify elections. ­These have included threats, menaces, harshness, intrigues, a surveillance of the vote . . . ​a close control of ballots through their representatives ­either during the moments preceding the vote or at the moment they hand over the ballot” ( JORF ChamberDoc 1899, no. 107). Such practices of electoral surveillance of voters w ­ ere also common in German elections. Employers and their representatives stationed in the voting place handed ballots to voters, “prohibiting them from exchanging ­these or from placing them in their pockets.” In the vicinity of the voting ­table, additional brokers controlled how voters handed in their ballots (StBer DtRt, 15 January 1890). Electoral reforms seeking to protect voter secrecy considered by many first-­wave democracies ­were, at the same time, an effort to reduce the electoral importance of economic coercion. This chapter examines the adoption of ­these electoral reforms. I begin by discussing candidates’ strategies that took advantage of imperfections in the design of ballots. Drawing on electoral commission rec­ords, I discuss the choices of candidates and their brokers in an environment characterized by imperfections in voting technology. Next, I turn to an analy­sis of electoral reforms. The proposals to limit coercion ­were remarkably s­ imple and involved a common set of solutions: the introduction of ballot envelopes and isolating spaces. While ballot envelopes could eliminate exterior signs that made the vote choice identifiable, isolating spaces created opportunities for voters to reflect f­ ree of coercion prior to casting their ballots. Despite the simplicity of ­these solutions, the adoption of legislation to protect voters’ autonomy remained one of the most contested electoral reforms. In France and Germany, the two cases examined h ­ ere, the negotiation of ­these reforms lasted nearly five de­cades. In this chapter, I examine the sources of initial disagreement about the desirability of electoral reforms using a combination of roll-­call votes and parliamentary deliberations. In both countries, an initial electoral co­ali­tion that brought together legislators with ties to power­ful economic actors successfully resisted the introduction of reforms. Next, I document the pro­cess that led to the erosion of this majority and the creation of an alternative co­ali­tion favoring reforms. Over time, the preferences of some resource-­endowed politicians about

P rotec ti n g Vote r Au to n o m y 147

the importance of electoral reforms changed due to rising economic and electoral costs associated with the use of economic coercion. In France, the elite split within the dominant Republican Party created incentives for candidates to invest in programmatic promises. For Radical and some Republican candidates competing on the basis of programmatic promises, economic coercion became incompatible with programmatic promises. Th ­ ese considerations contributed to the formation of a dif­fer­ent parliamentary majority that embraced the adoption of electoral reforms. Similar developments took place in Germany. Following the breakdown of the electoral cartel among parties representing urban and rural elites, National Liberal politicians who ­were dependent on runoffs experienced rising electoral costs associated with illicit strategies. In both France and Germany, support from cross-­pressured politicians such as Radicals or National Liberals was decisive in leading to the formation of a majority that supported reforms. Electoral Coercion in Nineteenth-­Century Eu­rope

­ fter the transition from elections based on open voting, electoral laws in A many nineteenth-­century elections included nominal protections for ballot secrecy. An impor­tant component of this protection was the design of the ballot. The French electoral law required that ballots be printed on white paper and without externally distinguishing signs. It also required voters to fold the ballot before inserting it into the electoral urn ( JORF Lois et Décrets, 15 March 1849, Articles 47 and 48). The German electoral law also specified that ballots had to be “fabricated of white paper and that they ­couldn’t be marked with exterior signs distinguishing them” (Hatschek 1920). Neither the French nor German electoral law mandated uniform ballots. As state officials ­were not required to print ballots, ­these responsibilities ­were delegated to individual candidates. Candidates used this opportunity to systematically vary the design of ballots in order to keep close track of the choices made by voters. A 1907 report of the French Commission of Universal Suffrage enumerated a variety of strategies by which candidates could vary the ballots. ­These included differences in the types of paper used to produce ballots, in the weight and size of the paper, and in the ways in which the names of the candidates ­were printed that could be vis­i­ble a­ fter the ballots ­were folded ( JORF ChamberDoc 1907, no. 641). First, candidates varied the shape of their ballots. Pierre Mielvaque de Lacour, a candidate competing in Brive (Corrèze) during the 1893 election, distributed two dif­fer­ent kinds of ballots to voters. The first tried to copy the

148 C H A P TER 5

shape and makeup of the ballots of his competitor. The second ballot was smaller and printed on transparent paper ( JORF ChamberDeb, 4 December 1893). In Narbonne (Aude), Edmont Bartissol, a centrist Republican, also used ballots of dif­fer­ent shapes. Bartissol designed smaller ballots in such a way to fit inside larger ballots. In the end, one large ballot could include eight to ten smaller ballots (ibid., 7 July 1898). In some cases, candidates printed ballots in the shape of lottery cards that could be inserted into the urn only with ­great difficulty (ibid., 20 December 1901). Fi­nally, candidates used vari­ous materials to create ballots. In Valenciennes in 1899, candidates attempted to differentiate ballots by printing them e­ ither on cardboard or on ­simple paper. In other cases, ballots ­were printed on blotting or water-­marked paper (ibid.). The German electoral law also contained no provisions regulating the design of ballots. As a result, candidates used ballots of dif­fer­ent shapes and sizes. In Hannover’s seventeenth district during the 1881 election, ballots differed in size by a magnitude of one hundred. The smallest ballot was printed on a page the size of a penny, while the largest ballot was as large as a piece of paper. H ­ ere candidates also varied the time at which they distributed ballots of dif­fer­ent sizes (Klein 2003: 223). Another commonly used strategy was to vary the weight of ballots (StBer DtRt 1881, no. 103). Freiherr von Stumm, the owner of a large iron mill who competed in Trier’s sixth district, printed his ballots on very thick cardboard, while his competitor representing the Zentrum used paper ballots (StBer DtRt 1890, encl. no. 346). In Dittersbach, the owner of a steel mill is reported to have “printed additional ballots in his office, which ­were clearly distinguishable based on their shape and quality of the paper on which they w ­ ere printed” (StBer DtRt 1878, encl. no. 1986). In environments where candidates competed by differentiating the shape of their ballots, it was particularly impor­tant to control the timing of their distribution to prevent opponents from copying one’s ballots. The electoral committee of the Kartellbündnis in Saarbrücken issued special ­orders to its brokers to distribute ballots only on election day, so that their shape and paper remained unknown to other candidates (StBer DtRt, 15 January 1890). In Dortmund during the 1879 election, one candidate modified the appearance of ballots ­every hour. One witness said that it was impossible to vote using a ballot that was dif­fer­ent from the ballots that had been handed to them by their supervisor. The shapes of the ballots ­were well-­known to election officials. They received voters with the “exclamation, ‘Aha ! A small ballot!’ which was understood by all ­those pre­sent at the voting place as meaning a ballot for Söler” (StBer DtRt 1881, encl. no. 104).

P rotec ti n g Vote r Au to n o m y 149

Candidates also attempted to differentiate the colors of their ballots. The challengers of Charles Le Myre de Vilers, a Republican candidate competing in Cochinchine, reported that de Vilers used ballots printed on transparent paper ( JORF ChamberDeb, 28 November 1889). During the 1902 election in Tarn, candidates differentiated ballots by using dif­fer­ent colors of ink to print their name (ibid., 9 July 1902). Sometimes, candidates varied only the shade of the paper of the ballot: some ballots ­were printed on shiny paper, while ­others on paper that was opaque or matte (ibid., 14 July 1898). As mayors summoned during the investigations of the electoral commissions testified, ­these strategies met with vari­ous levels of success. Mimicking the ballots of competitors was often in­effec­tive: “In the end, one can distinguish the ballots. We know very well for whom voters vote. While candidates have all means to copy the shape, the color, and the thickness of the ballots of their opponents, they cannot always succeed. One always ends up by knowing every­thing” (ibid.). Such efforts to subvert ballot secrecy by using ballots of dif­fer­ent shapes and colors imparted a colorful atmosphere to nineteenth-­century elections. ­These practices had, however, pernicious consequences for voters’ autonomy. Candidates tried to exploit imperfections in ballot design with the aim of closely monitoring the electoral choices of their voters and punishing ­those who voted the “wrong” way. Voters’ choices could be ascertained at two dif­fer­ent moments. Brokers could glimpse voters’ intentions prior to the vote, based on information about the design of ballots carried by voters into the voting place. In this chapter, I ­will discuss brokers’ strategies to infer voters’ choices at the moment they cast their ballots. A second moment during which candidates could infer voters’ choices was the time when the ballots w ­ ere counted. H ­ ere, actors seeking to pierce voting secrecy could rely on “internal signs” that differentiated the ballots. I w ­ ill discuss such irregularities that occurred at the moment ballots w ­ ere counted in the next chapter, alongside proposed remedies to the prob­lem of ballots with “internally distinguishable signs.” Brokers who w ­ ere stationed in the vicinity of electoral urns attempted to ascertain voters’ intentions but also sometimes tried to change ­these at the last moment. Brokers who could credibly threaten harsh post-­electoral punishments could also exert the greatest influence on voters’ choices. ­Because of their control over significant long-­term streams of income, employers w ­ ere extremely well-­positioned to impose post-­electoral punishments on voters. This explains why violations of voter secrecy went hand in hand with economic coercion. A report on electoral events occurring

150 C H A P TER 5

in Bochum-­Gelsenkirchen during the 1884 election commented on the pressure exerted by employers as follows: “Whoever notes about the state of economic dependence in which workers in ­these companies find themselves, whoever knows about the high economic pressure [Hochdruck] that supervisors can exert on workers, he ­will have no doubt that the presence of ­these employers at the voting place ­will unfavorably affect voters’ electoral freedom” (StBer DtRt 1884, encl. no. 320). Employers’ electoral involvement began, however, prior to the moment of the vote. They exploited opportunities to distribute ballots with externally distinguishing signs and accompany their employees to the voting place. In Landes, a department in southwestern France, economic activity was dominated by a small number of rural employers, who employed sharecroppers. At the time of elections, ­owners distributed marked ballots to sharecroppers and threatened that “it ­will be easy to recognize the ballot of each and ­every one of them.” ­After this, employers “escorted share-­croppers to polls like criminals” ( JORF ChamberDeb, 12 December 1885). An influential landowner in Saint-­Amand, Cher, distributed marked ballots supporting a Monarchist deputy, Count François-­Charles de Monsaulnin. He also searched the dwellings of peasants to remove ballots for the opposition candidates (ibid., 23 November 1899). Such practices ­were common among not only rural but also urban employers. In Belfort during the 1889 election, employers supporting Émile Keller, a Monarchist candidate, ­were reported to have developed “an odious system of externally marked ballots.” Some ­were handwritten by employers themselves and then distributed to voters (ibid., 1 February 1886). During the 1898 election in Lille, employers of large enterprises summoned their employees to the city center where they distributed ballots for Albert Masurel, an in­de­pen­dent candidate, and asked foremen to accompany workers, arranged in groups of five, to the urns. As one witness recalled, “the workers ­were escorted to the vote in squadrons, ­under the surveillance of foremen and at the voting place, employers surrounded the urns” (ibid., 8 November 1898). Following the distribution of marked ballots, employers stationed supervisors in the vicinity of the urns to monitor voters’ choices. Commissions investigating electoral irregularities in France and Germany repeatedly reported such examples of monitoring. A report in France commented on the “sad monotony of the repetition” of such instances of monitoring, election a­ fter election ( JORF ChamberDeb, 26 June 1906). Elections in Alais (­today’s Alès), a district in Gard, illustrate this monotonous repetition of the monitoring strategies. ­Here, employers’ intimidation was reported in two

P rotec ti n g Vote r Au to n o m y 151

successive elections, in 1898 and 1902 (ibid., 8 May 1898, 16 June 1902). In both instances, electoral reports allege that employers of the local mining com­pany placed supervisors in close vicinity of the urns during the duration of the vote, recording the votes of their employees. During the 1902 elections, employer representatives located next to the urn recorded employees’ votes in blue or red notebooks, depending on the shape of the ballot they held in their hand (ibid., 16 June 1902). A strategy commonly used by employers was to position a ­table with ballots and other electoral information in the vicinity of the urns. German election reports refer to such t­ ables as “agitation t­ ables” (StBer DtRt 1890, encl. no. 181). In the 1881 election in Ottweiler (Trier), officials from the large mine of the locality observed the voting from such an agitation t­ able and recorded voting decisions on lists with the names of employees (StBer DtRt 1881, Anlage no. 103). In Arnsberg’s fifth district, employers placed this agitation ­table six feet away from the urn. Representatives of the main employers of the district asked voters their name and tried to ensure that they carried a “Haarmann ballot,” a ballot for the National Liberal candidate in the district (StBer DtRt 1890, encl. no. 181). In a neighboring locality in the same district, representatives of the firm used more coercive mea­sures to ensure that voters approached the urns equipped with the correct ballot. ­Here, they “ripped the ballots for Schorlemer, the Zentrum candidate, from voters’ hands and gave them a ballot for Haarmann, urging them to vote for the latter candidate” (ibid.). To monitor voters, employers located at ­these ­tables used “parallel lists” that recorded voters’ choices in the order in which they inserted ballots into the urn. The reports of an election in Dortmund document the sophisticated signaling system set in place by employers to keep track of voters’ choices: “Supervisors from the mine stood at the entrance to the room, who distributed ballots for Möller. If a person that had received a ballot for Möller took another ballot from his pocket, then Schmidtmann pointed to his own bag, which gave the supervisors from the mine a clear signal that the respective voter had not voted for Möller” (StBer DtRt 1893, encl. nos. 255, 354). During the 1884 election in Ottweiler, officials of the large mine of the locality used alternative lists to rec­ord the ballot choices of each voter (StBer DtRt 1884, encl. no. 103). A witness recalls that “during the runoff election that took place on November 10, the overman employed by Taeglichsbeck [a local mine] stood in front of the open door of the polling station during the duration of the vote. ­Here, foremen pushed ballots in the hand of ­every employee of the mine” (ibid.). Similar incidents of monitoring ­were reported in Reichenbach-­Neurode, a district in Breslau, during the 1881 election.

152 C H A P TER 5

­ ere, the owner of a local mill distributed ballots at the entrance of the H voting place. This tight supervision made it impossible to replace t­ hese with other ballots (StBer DtRt 1881, encl. no. 104). Sometimes employers’ efforts to monitor the types of ballots held by voters w ­ ere viewed favorably by election officials with similar po­liti­cal sympathies. During the 1893 election, one National Liberal candidate attempted to differentiate his ballot by folding it multiple times (StBer DtRt 1893, encl. no. 354). B ­ ecause local employers ­were stationed in the immediate vicinity of the urn, election officials used this opportunity to publicly display each ballot. Election officials raised each ballot in the air, inspecting it for a long period of time. As the ballots of opposition candidates ­were folded only one time, employers had the opportunity to rec­ord in their parallel lists the names of voters who did not vote National Liberal. Employers used the information gleaned from the inspection of ballots at the moment of voting to punish voters who disobeyed their instructions. In Corsica, the owner of the w ­ ater com­pany in Orezza laid off employees who refused to take the ballots that w ­ ere distributed by foremen at the door of the polling station ( JORF ChamberDeb, 2 December 1893). In Compiègne, Oise, the owner of a sugar refinery laid off employees who refused to take marked ballots for Col­o­nel Bougon, a Republican candidate (ibid., 26 June 1902). Such layoffs ­were often executed swiftly. An agricultural owner in Lodève laid off employees suspected of voting the “wrong way” as soon as the ballot counting ended. As one witness recalled, “at seven thirty at night, this employer announced to his employees that they can search for employment elsewhere” (ibid., 2 July 1906). At the moment this witness’s statement was read in the Chamber of Deputies, legislators on the left aisle of the Chamber exclaimed: “this exemplifies the freedom of the vote and the freedom of employment” (ibid.). Such targeted layoffs ­were also a common practice in German elections. Dr. Ludwig Gross, a National Liberal candidate, deployed employers and foremen of the Badische Anilin und Sodafabrik in Ludwigshafen as brokers. Following the election, employers laid off dissenting employees, both skilled and unskilled. Georg Keller, a day laborer in Ludwigshafen, was laid off for exchanging the Gross ballot with a Social Demo­cratic ballot. Georg Martin, a mechanic within this enterprise, also reported that he had been offered a National Liberal ballot. Martin turned down this offer, indicating that he already had a Social Demo­cratic ballot. Martin was laid off on the day of the vote at noon before the results ­were officially announced (StBer DtRt 1881, Anlage no. 116).

P rotec ti n g Vote r Au to n o m y 153

Election reports in both France and Germany also include accounts of non-­targeted punishment. During the 1893 election in Bressuire, a constituency in Deux-­Sèvres, local employers, supporting de la Rochejaquelin, a Monarchist candidate, laid off all workers who ­were suspected to have supported le Roux, the Republican candidate. Several days ­later, ­these workers ­were also evicted from their homes ( JORF ChamberDeb, 9 December 1893). Such systematic layoffs a­ fter the publication of election results w ­ ere also reported in Germany (StBer DtRt 1887, encl. no. 105). This discussion has documented the phenomenon of economic coercion in nineteenth-­century elections and has discussed the strategies used by candidates and brokers to pierce the secrecy of the vote. The main resource asymmetry in this illicit strategy resulted from access to employers as brokers. This resource asymmetry was likely to shape conflict about the desirability of electoral reforms. Reforming Voter Secrecy in France

The per­sis­tence of employer intimidation undermined voters’ autonomy while also limiting the ability of elections to select politicians who reflected voters’ choices. A ­ fter a long discussion of vari­ous strategies of employer intimidation, a legislator in the Chamber of Deputies posed this question to all other deputies: “What and whom do we represent ­here?” This rhetorical question met with the reply, “The pressure of employers.” In both France and Germany, the proposals for reform endeavored to remedy two imperfections in the voting pro­cess. First, reforms attempted to “remove any constraints in the ability of voters to choose the ballot reflecting their desired po­liti­cal choice” ( JORF SenateDoc 1905, no. 62). Legislators in both countries a­ dopted ballot envelopes in order to allow voters to protect their electoral choice from manipulation by the actors pre­sent in the voting space. Second, reforms attempted to remove “any constraints in the ability of voters to choose the ballot that expressed their desire.” The solution to this second prob­lem was the isolating space. In France, proponents of electoral reforms regarded ballot envelopes as a remedy to the prob­lem of coercion. Proposals to use ballot envelopes ­were introduced at the first legislature of the Third Republic. In 1877, François Le Pommelec, a centrist Republican from Ille-­et-­Villaine, and Ferdinand Malézieux, a Republican from Aine, recommended that uniform ballot envelopes be used in elections. According to this proposal, the state had the responsibility to print ­these envelopes and distribute them to voters,

154 C H A P TER 5

but ­these could also be sold in city halls, at the price of one centime ( JORF ChamberDeb 1877, no. 992). Another proposal, formulated in 1882 by Alfred Girard, a Republican from Nord, envisioned ballot envelopes as metal holsters protecting ballots. The proposal also recommended that t­ hese holsters be complemented with urns constructed from glass ( JORF ChamberDoc 1880, no. 2628). In chapter 2, I conjectured that demand for such reforms originates with legislators who ­either lack the resources to engage in illicit strategies or face high electoral costs if they pursue such strategies. In France, the earliest proposals to introduce ballot envelopes ­were formulated by Socialist and Radical deputies. In 1890, Gaston Laporte, a Radical from Nièvre, submitted a proposal to adopt “opaque and uniform envelopes” ( JORF ChamberDoc 1890, no. 363). During the mid-1890s, Félix Defontaine, a Radical, took up the idea of ballot envelopes. His proposal recommended the implementation of uniform ballot envelopes, stamped with a seal of the state for each electoral college ( JORF ChamberDoc 1902, no. 181, Article 5). Defontaine represented Avesnes, a district in Nord, one of the most highly industrialized districts at the time where practices of electoral coercion ­were common. The proposal’s goal was to “defend workers who are likely to suffer” and “who can be thrown into misery” a­ fter an election ( JORF ChamberDeb, 16 December 1901). The Chamber of Deputies took up Defontaine’s proposals in its deliberations in 1898, 1901, and 1904. By 1901, a strong po­liti­cal consensus had developed around the introduction of ballot envelopes. The article of the law recommending isolating spaces met with nearly unan­i­mous approval. Another source of Radical influence can be attributed to the prominence of ­these legislators on the Commission of Universal Suffrage of the Chamber. This commission, established in 1901, was tasked with bringing t­ hese proposals to the floor of the Chamber of Deputies. At the time of its creation, the commission was led by Edmond Guyot-­Dessaigne, a Radical politician from Puy-­de-­Dôme (Archives Nationales C 7305). Other Radical deputies on the commission played a significant role in drafting numerous, often lengthy reports advocating for the need for electoral reforms. The Radical politician who played the most decisive role in shepherding the bill through the legislative pro­cess was Joseph Ruau, a deputy of Haute Garonne. Introducing the proposal of the commission in the Chamber of Deputies, Ruau eloquently expressed the position shared by other Radical politicians: “Nobody can deny that the secret of the vote is not assured at the moment. ­There is no doubt that with certain ballots one can infer the name

P rotec ti n g Vote r Au to n o m y 155

of a candidate by touching the ballot” ( JORF ChamberDoc 1902, no. 181). In his report, Ruau stressed that the duty of legislators was “to protect the interests of French citizens whose resources have unjustly decreased ­after an election. The mea­sures that we demand include a po­liti­cal idea of protection, and it is always the same idea, to protect wage earners against ­those who own capital and have the means to oblige workers to commit to something his conscience objects to” (ibid.). In his intervention in the Chamber, Ruau argued that ­these mea­sures allowed for the restoration of the “Republican truth” by “raising the moral and po­liti­cal level of universal suffrage and by purging elections of the blight of coercion. Electoral freedom is illusory, if the vote is observed by an influential supporter of a candidate or by a ­little tyrant of a village.” Ruau admitted that in Paris and in large urban centers it was more difficult to monitor voters’ choices. But, “in over half of the rural communes of France, voters are terrorized by the fact that some influential person in the locality ­will observe their vote and this fear is often decisive for the outcomes of elections” ( JORF ChamberDeb, 16 December 1901). Socialist deputies in the Chamber supported the introduction of ballot envelopes. Arthur Lamendin, a Socialist from Pas de Calais, and one of the few legislators of working-­class origin, invoked the electoral plight of Socialist voters, who ­were often the target of the coercive strategies of employers. In his intervention in the Chamber, Lamendin voiced the concerns of voters from low-­income backgrounds. Such voters “demand to place their ballots in an envelope, so that employer pressure can no longer take place. They no longer want to see employers or supervisors leading groups of voters to the urn. ­Under the current system, the voter is not ­free and often loses his bread if he does not insert in the urn the ballot for the candidate who is desirable to his employer” ( JORF ChamberDeb, 17 December 1901). Opponents rejected demands for the introduction of ballot envelopes by denying the pervasiveness of coercion in French elections. One opponent of ballot envelopes argued that “the law dangerously infers from the par­ tic­u­lar to the general and considers exceptions as constant and customary practices” ( JORF ChamberDeb, 17 December 1901). Albert de Benoist, a centrist Republican from Main et Loir, also questioned the frequency of cases of coercion: “I affirm that in our villages, the voter is f­ ree to vote as he desires and to change the ballot that was handed to him in advance; despite efforts to dominate him and force upon a marked ballot, the rural voter is sufficiently cunning to have in his pocket . . . ​the ballot that he wants to insert in the urn” (ibid.).

156 C H A P TER 5

An additional argument raised by opponents of ballot envelopes invoked the traditionalism of French voters. “We enact this law for French voters,” Gérard Aubourg De Boury, a Republican from Eure, asserted. “French voters have been for a long time familiar with universal suffrage . . . ​and they would not accept to change their traditional habits altogether” ( JORF ChamberDeb, 16 December 1901). Régis de l’Estourbeillon, a Monarchist from Morbihan, echoed ­these considerations. Rural voters w ­ ere not ready for this innovation ­because of their “character, their habits, and their electoral temperament. E ­ ither as a result of routine, or inexperience, t­ hese voters love ­s imple ­things and loathe novelties, complications, which frighten them, arouse their mistrust, and appear useless to them” (ibid., 18 December 1901). Fi­nally, opponents of ballot envelopes also objected to the “excessive formalism” of the new mea­sure. The proposal mired a basic right, such as the right to vote, in myriad forms and prescriptions, reducing its efficiency and value (ibid., 1 April 1898). Critics argued that this excessive formalism could lead to delays and abstentions (ibid., 18 December 1901). ­These arguments ­were easily rebutted during the deliberations. Supporters of ballot envelopes such as Émile Chavin, a Radical Socialist from Seine-­et-­Marne, argued that the requirement to place a ballot in an envelope “was unlikely to prevent workers in the fields and ­those in factories from exercising their rights and duties” (ibid., 17 December 1901). Despite the objections, a po­liti­cal majority in support of ballot envelopes formed with relative ease. In contrast to recommendations for creating isolating spaces (discussed below), proposals to introduce ballot envelopes met with very l­ ittle po­liti­cal opposition. In an intervention in the Chamber, Pierre Waldeck-­Rousseau, the prime minister at the time, announced the government’s support for ­these reforms. Waldeck-­Rousseau argued that the “use of ballot envelopes remedies all pos­si­ble abuses” ( JORF ChamberDeb, 19 December 1901). A consensus around ballot envelopes formed with g­ reat rapidity. In a vote on 19 December 1901, deputies in the Chamber overwhelmingly supported the introduction of ballot envelopes. In a second vote in the Chamber, on 27 October 1904, a proposal to introduce ballot envelopes received nearly unan­i­mous support. An additional amendment to Defontaine’s early proposal ­adopted by the Chamber attempted to eliminate the ability of voting place presidents to ascertain voters’ choices. This amendment responded to worries that ballot envelopes did not fully solve the prob­lem. Presidents of voting places could still infer voters’ intentions by weighing ballot envelopes in their hands or

P rotec ti n g Vote r Au to n o m y 157

by attempting to open them. To minimize this possibility, an amendment supported by Socialist deputies recommended that voters place their own ballots in the urn ( JORF ChamberDeb, 16 December 1901). The proposal met with widespread support from both the Chamber of Deputies and the Senate. Even Conservative legislators agreed with this recommendation, noting that it reduced opportunities for voting bureau presidents to mark ballot envelopes when they ­were inserted in the urn and thereby invalidate ­these votes (ibid., 17 December 1901). The Senate also applauded this provision, noting that “the provision that voters themselves place the ballot in the urn is of ­great importance. Without this prescription, the exterior signs made by officials to the ballot ­will be simply translated to the envelope and the lard to grease fin­gers, the stain of ink, or the color u ­ nder the fingernail w ­ ill continue to have their day” ( JORF SenateDoc 1905). This amendment remained part of the final version of the legislation ­adopted in France in 1913. However, proponents of reform argued that ballot envelopes by themselves could not adequately protect voters against economic coercion. The limitation of reforms that introduced only ballot envelopes was that voters could still be observed by p ­ eople pre­sent in the voting place. A number of deputies argued that the adoption of ballot envelopes could increase electoral coercion. Voters who picked up ballot envelopes before casting their vote lacked the opportunity to exchange ballots that had been handed to them by brokers pre­sent at the entrance of the voting place. To reduce the possibility of coercion, reformers advocated for the introduction of isolating spaces as a complement to ballot envelopes. While the po­liti­cal majority in support of ballot envelopes formed with remarkable speed, proposals to introduce isolating spaces met with significant po­liti­cal re­sis­tance. Proposals to introduce isolating spaces ­were put forward during the first de­cades of the French Republic. In 1882, Léonard Correntin Guyho, a legislator from Finistère, recommended the introduction of ballot envelopes together with isolating spaces ( JORF SenateDoc 1912, no. 30). At the time, the Chamber refused to discuss this proposal. In 1885, Gaston Laporte, a Radical from Nièvre, formulated a proposal to introduce a “basket of envelopes” (corbeille d’envelopes) to collect all ballots. Laporte recommended placing this basket in a corner of the voting space in order to allow voters to “turn their back to the election bureau while casting their ballots” ( JORF ChamberDoc 1885, no. 737). Around 1900, a flurry of additional proposals recommended the introduction of isolating spaces. ­These proposals ­were sponsored ­either by Radical Socialist deputies, such as Félix Defontaine, or

158 C H A P TER 5

by Socialists, such as Rochilde Pastre ( JORF ChamberDoc 1898, no. 2966; 1902, no. 113). In the early stages of ­these reforms, we find profound uncertainty about the appropriate term to define the new space inside the voting place. In his meeting with the members of the Commission of Universal Suffrage, Pierre Waldeck-­Rousseau referred to the new provision as an isolating mechanism (mécanisme d’isolement) (Archives Nationales C 5651). Paul Constans, a Socialist deputy from Allier, referred to ­these spaces as “isolating cabins” (cabines d’isolement) ( JORF ChamberDoc 1902, no. 113). The proposal submitted to the Chamber in 1904 used a rather long-­w inded phrase for isolating spaces: “device allowing voters to isolate themselves to place the ballot inside the envelope” ( JORF ChamberDeb, 27 October 1904). The term used in Belgium, isoloir, made its entry into French discourse a­ fter 1900. While this term was eventually a­ dopted to describe this protection, as late as 1907, the Commission of Universal Suffrage critiqued it as a “rather barbarian term” (un terme plutôt barbare) ( JORF ChamberDoc 1907, no. 641). Socialist legislators argued that isolating spaces ­were needed as a remedy against economic coercion. Franc Arnal, a Socialist from Var, decried the absence of po­liti­cal freedom: “­Today as well as on prior occasions voters experience a very heavy social de­pen­dency, as well as a po­liti­cal de­pen­ den­cy . . . ​and, unfortunately, they have to choose between their po­liti­cal consciousness and the interests of their material existence” ( JORF ChamberDeb, 27 October 1904). Constans argued that isolating spaces sheltered voters and allowed them to change the ballots that had been handed to them at the urn (ibid.). In addition to Socialists, Radicals endorsed the recommendation for isolating spaces very early. The proposal for electoral reform formulated by Félix Defontaine recommended introducing isolating spaces in addition to ballot envelopes ( JORF ChamberDoc 1902, no. 181). The Commission of Universal Suffrage incorporated Defontaine’s recommendations in the proposal submitted to the Chamber in 1902. Joseph Ruau justified the need for isolating spaces as follows: “Alongside intimidation, ­there is coercion. This is why the envelope has to be completed by another means, such as a secret corridor or isolating space. This w ­ ill allow voters to escape the gaze of ­those who force voters to hold open ballots in their hand” (ibid., no. 181, p. 5). Radicals argued that isolating spaces ­were a necessary complement to ballot envelopes ­because they counteracted a dif­fer­ent opportunity for coercion. Ballot envelopes sought to reduce the possibility that voters’ electoral choices ­were recognized at the time they cast their ballots. By contrast,

P rotec ti n g Vote r Au to n o m y 159

isolating spaces “­were a remedy against a much more dangerous type of fraud,” which resulted from the surveillance of voters by their supervisors prior to the moment they entered the voting space (Archives Nationales C 7375, 22 November 1902). Opposition to the concept of isolating spaces came from legislators affiliated with Conservative and Monarchists parties but also from centrist Republicans who had the resources to enlist brokers to coerce voters. Conservative candidates defended the status quo as a bulwark against their Socialist and Republican competitors. Michel Robert de Pomereu, a deputy affiliated with the Action Libérale, attacked isolating spaces as a mea­sure that could lead to “the imposition of the preferences and intentions of government on voters during elections.” Isolating spaces had to be rejected, as it was impor­tant for candidates to retain the means to resist governmental pressure ( JORF ChamberDeb, 21 December 1901). Other Conservative legislators decried the increase in polarization that had taken place in several industrial centers, believing that this was fueled by the false promises of Socialists or “collectivists.” The status quo allowed opponents to resist ­these mea­sures (ibid.). A significant number of Republican legislators joined Conservatives in opposing isolating spaces. Rhetorically, Republicans w ­ ere more constrained from opposing isolating spaces as compared to Conservatives, due to their nominal commitment to electoral freedom. Thus, while Monarchists and other Conservatives denied the existence of electoral coercion altogether, Republicans invoked practical difficulties associated with the introduction of isolating spaces. One such argument focused on the costs of ­these mea­ sures. In his meeting with the Commission of Universal Suffrage, Waldeck-­ Rousseau objected to isolating spaces, arguing that t­ hese imposed too high a fiscal burden on communes (Archives Nationales, Commission du Suffrage Universel C 5651; JORF ChamberDeb, 17 December 1901). Republicans also invoked administrative difficulties associated with the introduction of ­these reforms. Many French communes lacked city halls and w ­ ere likely to face im­mense difficulties in providing voters with isolating spaces ( JORF ChamberDeb, 23 December 1901). Given the difficulties in creating uniform isolating spaces across all 36,000 French communes, many mayors w ­ ere likely to respond to this requirement by establishing installations that ­were extremely precarious. Consequently, ­these mayors could l­ater face accusations of trying to undermine the elections. In concluding his intervention in the Commission of Universal Suffrage, Waldeck-­Rousseau confirmed that his government endorsed the concept of ballot envelopes. By contrast,

160 C H A P TER 5

he opposed the introduction of isolating spaces, considering that the latter “had not been sufficiently well-­studied” ( JORF ChamberDoc 1903, no. 1170). A third argument invoked by Republicans was that isolating spaces w ­ ere likely to slow down the voting pro­cess and lead to an increase in abstentions. During the discussion in December 1901, Republicans referred to isolating spaces as “an instrument of systematic obstruction of the vote” or as a “reward for electoral obstruction” (prime à l’obstruction) ( JORF ChamberDeb, 16 December 1901). On the one hand, such slowdowns could result from sudden bursts of voting whereby many voters de­cided to cast their ballot at the same time. On the other hand, the obstruction could result from individual voters’ malicious decisions to block the voting pro­cess (Archives Nationales, C 7375, 22 November 1902). Th ­ ese risks to the voting pro­cess ­were invoked on repeated occasions. “How could one find the means to prevent ill-­intentioned voters from spending an excessive amount of time in the cabine?” ( JORF ChamberDoc 1907, no. 641). Other opponents ­were concerned about disorder at the voting place: “you create a tumult of voters, some hurry to the bureau to take the envelopes and then go to the isoloirs, while ­others w ­ ill be at the bureau to depose their ballots: the congestion is inevitable” ( JORF ChamberDeb, 27 October 1904). ­Those who supported the idea of isolating spaces attempted to minimize Republican opposition by downplaying the scope of the reforms and by stressing the minimalist nature of the intervention. The isoloir was not a “monument that had to be erected in the voting place” ( JORF ChamberDeb, 27 May 1907). Proponents of reforms w ­ ere ready to accept a wooden construction surrounded by an opaque cloth, or even a screen. O ­ thers considered that a tringle or a piece of cloth would suffice. If this solution was ­adopted, it was impor­tant to raise this protective material fifty centimeters above the ground to allow observers to establish w ­ hether the voting place was occupied (ibid., 27 October 1904). Reformers w ­ ere also willing to compromise on plans to mandate the introduction of isolating spaces in all communes. In the Commission of Universal Suffrage of the Chamber, Joseph Ruau argued that “one could give up the princi­ple of uniformity of the installation and even suppress the excessively specific term ‘cabine’ and replace it with a more general expression, implying the momentary isolation of the voter” (Archives Nationales, C 7375, 22 November 1902). ­After a lengthy deliberation, the French Chamber of Deputies voted on a proposal to introduce isolating spaces on 23 December 1901. On this occasion, 204 deputies voted in support of this mea­sure, while 325 deputies opposed it. This vote gives us an opportunity to examine the sources

P rotec ti n g Vote r Au to n o m y 161

of po­liti­cal disagreement over ­these reforms during the early stages of negotiations. The models presented in figure 5.1 examine the relative importance of relative resource advantages, economic costs, and electoral costs in shaping demand for electoral reforms. In the case of this electoral irregularity, the beneficiaries of the status quo w ­ ere candidates who could rely on the support of employers as brokers. An early proposal to introduce isolating spaces recognized this resource asymmetry, noting that the beneficiaries of the status quo ­were employers ( JORF ChamberDoc 1889, no. 107). One can proxy the resource asymmetry using both legislator-­level variables and partisan variables. At the individual level, I proxy the relative resource advantage using a variable that takes the value 1 if the legislator is an urban or rural employer (Chambre des députés 1898). We also find differences in resource asymmetries across parties. Conservatives and Republicans had a higher share of employers as legislators. While 40 ­percent of Conservative legislators ­were ­either urban or rural employers, this share was 25 ­percent for centrist Republicans. By contrast, both Radicals and Socialists had a lower share of legislators with this occupational background. One expects ­these differences in access to affect the partisan disagreements over the desirability of reforms. The results presented in panel (a) of figure 5.1 strongly support this conjecture. First, individual-­level attributes of legislators are significant predictors of their preferences about ­these electoral reforms. Legislators with a background as employers opposed the introduction of isolating spaces, favoring the status quo. The partisan disagreement also mirrors the relative resource asymmetries among parties in accessing brokers who could engage in electoral repression. We find a pronounced left-­r ight cleavage over the introduction of ­these reforms. Parties on the right of the po­liti­cal spectrum—­such as Conservatives and centrist Republicans—­opposed the introduction of isolating spaces. By contrast, both Radicals and Socialists supported ­these reforms. The results reported in panel (b) of figure 5.1 illustrate the importance of economic conditions in shaping demand for electoral reforms. As conjectured in chapter 2, I expect economic constraints to be particularly salient in the case of this electoral reform. As candidates relied on private brokers, economic changes w ­ ere likely to reduce the attractiveness of ­these strategies. We find support for t­ hese hypotheses derived from modernization theories. Legislators competing in more developed districts—as mea­sured by a smaller share of agricultural employment, higher levels of urbanization,

162 C H A P TER 5

(a) Individual and partisan resources

(b) Economic constraints Individual resources Owner

Economic constraints Rural employment

Partisan resources Nonpartisan

Urbanization

Socialist Landholding inequality

Radical Republican

Voting population (ln)

Right –5

0

5

–5

0

(c) Electoral constraints

5

10

(d) Full model

Individual resources Owner Partisan resources Nonpartisan Socialist Radical Republican Right

Electoral constraints

Electoral constraints Programmatic promises Runoff Competitiveness Programmatic promises

–1

0

1

2

3

4

Economic constraints Rural employment Urbanization Landholding inequality Voting population (ln) –5

0

5

10

15

Support for the introduction of the isoloir during early deliberations. Notes: Results pre­sent correlates of support for the adoption of the isoloir during a vote taken in the French Chamber of Deputies on 23 December 1907. All models are estimated using Firth logistic regression. See the appendix for the sources used to construct the variables used in the analy­sis. FIGURE 5.1.

and higher population—­were more supportive of the introduction of isolating spaces. Demand for reform is higher in regions with higher levels of rural in­equality, a finding that disconfirms predictions of redistributive theories of democ­ratization. Panel (c) of figure 5.1 shows the relationship between the programmatic commitment of a candidate and support for electoral reforms. As discussed in chapter 3, I mea­sure the programmatic commitment of a candidate using a variable that takes the value 1 if the candidate pledged in his profession de foi to support the introduction of a progressive income tax. Consistent with the theoretical predictions formulated in chapter 2, we find a positive relationship between a candidate’s support of programmatic policies and his support of isolating spaces. I interpret this relationship as driven by electoral considerations: as programmatic politicians are likely to face steep sanctions from voters for using illicit strategies, they support electoral reforms in

P rotec ti n g Vote r Au to n o m y 163

order to close off opportunities for coercion to competitors. Th ­ ese findings contrast with the results reported in previous chapters. In the case of votes on electoral reforms to limit the use of state resources that took place during the same period (reported in chapter 3), the relationship between programmatic commitment and support for electoral reforms was negative and Radical legislators voted overwhelmingly in opposition to reform. In the case of strategies that politicized state resources, Radicals faced a steeper trade-­off between their resource advantage and electoral costs. By contrast, Radicals’ trade-­off between available resources and programmatic commitments was less pronounced in the case of reforms to protect voter integrity. In line with theoretical expectations, we find that both relative resource availability and electoral and economic constraints explain demand for electoral reforms. Radicals and Socialists who lacked access to economic brokers and who faced higher electoral costs strongly supported this policy. By contrast, Conservative candidates who enjoyed a relative resource advantage and faced low electoral costs opposed the reforms to limit coercion. In 1901, centrist Republicans ­were split over the desirability of reforms, with a significant number of resource-­endowed deputies voting in opposition of isolating spaces. This re­sis­tance contributed to the defeat of the proposal on this occasion. A parliamentary majority in f­avor of isolating spaces emerged only ­after several years of additional negotiations. During this period, centrist Republican politicians’ opposition to isolating spaces subsided. One impor­ tant trend that facilitated this realignment was the growing salience of programmatic competition. As documented in chapter 3, competition on the basis of programmatic proposals became increasingly salient, due to the importance of questions attached to issues of fiscal reform in the period ­after 1900. Both Radical and Republican politicians took up the issue of taxation in their campaign promises. During the 1906 election, over 27 ­percent of Republicans included programmatic promises in their professions de foi. By increasing the electoral costs associated with illicit strategies, the programmatic re­orientation of centrist Republicans contributed to a change in the composition of the electoral co­ali­tion supporting reforms. The discussion of the proposal for isolating spaces in the Senate in February 1906 gave an impetus to the formation of a po­liti­cal majority favoring reforms. Fernand Dubief, the minister of interior at the time, announced that the government supported the introduction of isolating spaces. In justifying this decision, Dubief recalled an idea considered during ­earlier

164 C H A P TER 5

deliberations: “the introduction of ballot envelopes poses more dangers for the sincerity of electoral operations if no guarantees for the isolation of voters are put in place” ( JORF SenateDeb, 6 February 1906). Dubief argued that intimidation or coercion could occur even ­after the introduction of ballot envelopes if a voter entered the voting place surrounded by “a double row of vigilant controllers who supervise him, who examine his lesser movements and who follow him ­until the moment when he places the ballot in the hands of the president of the voting place” (ibid.). The envelope and isolating spaces “­were connected in an absolute manner”; reforms that introduced ballot envelopes alone ­were insufficient. Dubief concluded that “to guarantee the sincerity of the vote, it was necessary to have both envelopes and isolating spaces” (ibid.). The Senate had exercised an impor­tant veto position during the deliberations of ­these reforms. ­Here, a loose alliance between Conservatives and centrist Republicans successfully blocked several proposals to introduce isolating spaces. In the Senate, Eugène Lintilhac, a Radical Socialist representing Cantal, led a special commission appointed to study the vari­ous proposals for electoral reforms developed by the Chamber. In a report submitted to the Senate in 1905, Lintilhac’s commission endorsed the introduction of isolating spaces. “Isolating spaces,” Lintilhac argued, “liberated the voter during the moments preceding the vote so that he could exercise his moral freedom” ( JORF SenateDoc 1905, no. 62). Following the pre­sen­ta­ tion of Lintilhac’s report and Dubief ’s intervention, the Senate voted by a narrow majority to support the introduction of isolating spaces ( JORF SenateDeb, 6 February 1906). This positive vote in the Senate established stronger po­liti­cal preconditions for a successful resolution of the question of isolating spaces in the Chamber of Deputies. At the time the bill was reintroduced in the Chamber, in March 1907, Charles Benoist, the president of the Commission of Universal Suffrage, reached out, in a conciliatory tone, to opponents of isolating spaces. According to Benoist, the differences among the vari­ous positions about t­ hese reforms ­were primarily about means, not ends. “At this point in time,” Benoist argued, “no ­matter which side of the Chamber we are on, we agree on the po­liti­cal and moral necessity to transform the theoretical secrecy of the vote into real­ity and to ensure the freedom of the voter” ( JORF ChamberDeb, 21 May 1907). Other Republican legislators in the Chamber echoed this recommendation. Joseph Paul Amédée Chailley, a centrist Republican from Vendée, urged legislators to conclude the negotiations and adopt isolating spaces. Invoking examples of voters who ­were coerced to vote using

P rotec ti n g Vote r Au to n o m y 165

the ballot handed to them by power­ful economic actors, Chailley argued that not all French voters enjoyed the same po­liti­cal freedom. In districts where electoral coercion was still pervasive, envelopes would be insufficient to protect the freedom of the vote (ibid.). On 21 May 1907, Paul Constans, a Socialist deputy from Allier, reintroduced to the Chamber an amendment recommending the introduction of isolating spaces. Fifty-­two other Socialist deputies cosponsored this proposal. The goal of the amendment was to harmonize the provisions of the legislation across both Chambers of parliament. As such, the amendment included the same definition of isolating spaces as that in the bill that had been ­adopted in the Senate, recommending that all mayors introduce in each voting place “one or several devices allowing voters to isolate themselves in order to place ballots in an envelope.” On the same day, the French Chamber voted again on another amendment recommending the introduction of isolating spaces. On this occasion, 357 legislators voted in support of isolating spaces, while 157 opposed the policy ( JORF ChamberDeb, 21 May 1907). The approval of this provision by both Chambers concluded the long disagreement over the adoption of isolating spaces. This final vote gives us an opportunity to analyze the composition of the electoral co­ali­tion that facilitated the adoption of isolating spaces in France. Figure 5.2 pre­sents the results of four models that isolate the relative importance of considerations about relative resource advantage, economic costs, and electoral costs in shaping the preferences of legislators for ­these reforms. The results reported in panel (d) bring together all variables that are conjectured to impact the preferences of legislators. The composition of the electoral co­ali­tion favoring the introduction of isolating spaces changed dramatically relative to 1901. While in 1901 both moderate Republicans and Conservatives opposed the introduction of isolating spaces, by 1910 large numbers of legislators affiliated with all parties supported the introduction of ­these electoral reforms. The importance of economic conditions in a district in explaining support for reforms also recedes in importance over this period. Thus, while hypotheses derived from the modernization perspective may explain the initial cleavages over the adoption of isolating spaces, they cannot explain the composition of the decisive co­ali­tion favoring reforms. The results presented h ­ ere support the prediction that concerns about electoral costs are an impor­tant determinant of support for electoral reforms. As documented in chapter 3, the share of Republican politicians competing on the basis of programmatic promises changed over this period. During the

166 C H A P TER 5

(a) Individual and partisan resources

(b) Economic constraints Individual resources

Economic constraints

Owner Nonpartisan

Agricultural employment

Socialist

Urbanization

Radical

Landholding inequality

Republican

Population (ln)

Right –2

0

2

4

6

–6

8

–4

–2

0

2

4

(d) Full model

(c) Electoral constraints

Individual resources Owner Nonpartisan Socialist Radical Republican Right

Electoral constraints

Electoral constraints Programmatic promises Runoff Margin of Victory Programmatic promises

–2

–1

0

1

2

3

4

Economic constraints Agricultural employment Urbanization Landholding inequality Population (ln) –10

–5

0

5

10

The composition of the electoral co­ali­tion supporting the isoloir in 1907. Notes: Results pre­sent correlates of support for the adoption of the isoloir during a vote taken in the French Chamber of Deputies on 21 May 1907. All models are estimated using Firth logistic regression. See the appendix for the sources used to construct the variables used in the analy­sis. FIGURE 5.2.

1898 election, 6 ­percent of centrist Republicans competed on the basis of programmatic promises (see ­table 3.2). By 1906, this share reached nearly 30 ­percent. Such legislators w ­ ere “cross-­pressured” between the desire to maintain their resource advantage and considerations about electoral costs. Programmatic appeals w ­ ere incompatible with strategies of economic coercion. A po­liti­cal majority formed ­after a large number of cross-­pressured politicians shifted their position and embraced reforms. Reforming Ballot Secrecy in Germany

Violations of voter secrecy w ­ ere also a common occurrence in German elections. The reports of the commission investigating elections revealed and publicized ample irregularities occurring at the time of voting in German national elections. Reports of voters entering the voting area holding ballots over their heads, of employers stationed at voting places and monitoring the type of ballots held by voters, and of ample post-­electoral

P rotec ti n g Vote r Au to n o m y 167

punishments raised general awareness about the insufficient protection of voters at the time they cast their ballots. Reform proposals that attempted to remedy ­these shortcomings date to the first elections in Imperial Germany. By protecting the secrecy of the vote, ­these proposals attempted to “ensure that the vote cast represented the true intention of the voter” (StBer DtRt, 29 January 1896). The proposals for reform considered in the German Reichstag included recommendations similar to ­those considered in France, namely ballot envelopes and isolating spaces. Ballot envelopes sought to prevent candidates from manipulating the color and shape of ballots in order to monitor voters’ choices. Isolating spaces complemented ballot envelopes by giving voters “that are pressured from all directions the opportunity to reflect in silence before casting their votes” (StBer DtRt, 1 February 1899). In chapter 2, I conjectured that the initial conflict line about the desirability of electoral reforms can be explained by asymmetries among legislators in their ability to access resources that allowed them to undermine the electoral law. Such resources could be ­either public or private. In the case of violations of voter secrecy, the most impor­tant brokers deployed by candidates w ­ ere rural and urban employers. The initial proposals for electoral reforms to protect voter secrecy originated with legislators from parties lacking access to such employers. In Imperial Germany, this informal electoral co­ali­tion included Social Demo­crats, the Zentrum, parties representing ethnic minorities, such as Poles, and ­Free Liberals. By contrast, the electoral co­ali­tion opposing reforms included Conservative parties, representing rural landowners, and National Liberals, who ­were tightly connected with urban employers. The Social Demo­cratic del­e­ga­tion in the Reichstag initiated proposals recommending better protection of voter secrecy. In 1878, a group of Social Demo­cratic legislators in the Reichstag formulated a proposal to modify the electoral law and introduce ballot envelopes and uniform ballots (StBer DtRt 1878, encl. no. 66). This was an effort to protect likely Social Demo­cratic voters against threats of power­ful economic sanctions. August Bebel argued that “in industrial centers one cannot speak of ­free elections.” ­Here, if voters voted their genuine preference, they could lose their job but also the chance of finding any employment in the district (StBer DtRt, 4 March 1879). Paul Singer, a Social Demo­cratic deputy representing Berlin’s suburbs, argued that the law could limit “the threatening ghost of layoffs which immediately appears b ­ ehind workers, leading to unemployment if they do not follow the po­liti­cal command of their employers” (ibid., 15 January 1890). Speaking

168 C H A P TER 5

on behalf of the Social Demo­cratic representatives pre­sent in the Reichstag in 1902, Ignaz Auer, a Social Demo­crat from Glauchau-­Meerane, expressed similar considerations in support of ­these proposals. ­Under German electoral rules, the power­ful actors exploited their social predominance to coerce the socially weaker to vote against their conviction. The adoption of ballot envelopes, Auer argued, could remedy the current situation whereby profound economic inequalities affected po­liti­cal choices (ibid., 29 January 1902). While the first proposals to provide better protection for ballot secrecy originated with Social Demo­cratic politicians, two other parties—­Free Liberals and Catholics—­took up ­these ideas and shepherded the proposals through the legislative pro­cess. Heinrich Rickert, a F ­ ree Liberal legislator elected in Danzig, was the most tireless advocate of ­these proposals. In his passionate defense of better protections for ballot secrecy, Rickert invoked the moral deficiencies of the status quo. F ­ ree Liberals argued that “the manipulation of ballots which are used to coerce eco­nom­ically dependent voters to vote against their consciousness for a party they do not want to support are morally highly reprehensible, . . . ​a violation of consciousness” (StBer DtRt, 29 January 1902). Without t­ hese additional guarantees, “public life was a hy­poc­risy, based on pressure and force” (ibid., 15 January 1890). In the discussions in the Reichstag, ­Free Liberals articulated the strongest normative arguments in support of ballot secrecy. By contrast, representatives of the Zentrum emphasized practical considerations. Adolf Gröber, a Zentrum politician representing Ehingen, a district in Württemberg, played a decisive role in drafting vari­ous proposals to protect electoral secrecy. He was instrumental in drafting the law introducing ballot envelopes that was ­adopted in Württemberg in 1883. In the discussions in the Reichstag, Gröber defended ballot envelopes as critical in limiting the ability of brokers situated in the vicinity of urns to infer voters’ choices. He argued that “once envelopes are introduced, questions w ­ hether ballots are white or not white, bluish or grayish or reddish or yellowish w ­ ill diminish in importance” (StBer DtRt, 29 January 1902). At the same time, Gröber argued that ballot envelopes ­were by themselves insufficient to protect voter secrecy. It was impor­tant to supplement envelopes with an additional device that would “give voters the opportunity to be alone for a moment so that he can insert in the envelope the ballot which expresses his conviction” (ibid., 15 January 1890). During the first de­cade in which the Reichstag considered ­these reforms, both Conservative and National Liberal deputies opposed t­ hese proposals. Iw ­ ill discuss the shift in the position of National Liberals from opposing to endorsing t­ hese proposals below. Conservatives used the opportunity

P rotec ti n g Vote r Au to n o m y 169

created by ­these debates to restate their opposition to secret voting. Von Puttkamer, Prus­sia’s interior minister, argued that secret voting was “a danger for the fatherland, the crown, and the monarchy” (StBer DtRt, 15 January 1890). Conservative legislators such as Carl von Stumm and Julius von Mirbach argued that secret voting was the source of a larger moral evil than open voting. Secret voting created incentives for voters to misrepresent their preferences, while also creating the opportunity for a “chaotic and poisonous agitation” (ibid., 8 February 1896). Von Mirbach, a landowner from Ortelsburg, volunteered to “lend a hand” to eliminate secret voting altogether, considering that such a decision “would be welcomed with euphoria in all rural districts and beyond” (ibid.). Similar to politicians in France, German Conservatives ridiculed isolating spaces, coining the term Klosettraum when referring to the bill (Das Klosettgesetz, Deutsche Tageszeitung, 7 March 1903; Nochmals das Klosettgesetz, Deutsche Tageszeitung, 13 March 1903). Von Mirbach argued that the proposal had a “comical aftertaste” and referred to isolating spaces as “camera obscura” (StBer DtRt, 15 May 1895). National Liberals joined Conservatives in mocking isolating spaces. The Kölnische Zeitung, a prominent National Liberal publication, pejoratively labeled isolating spaces as “chambers of fear” (Angstkammer), where voters ­were held in darkness so nobody could observe their fear (ibid., 15 January 1890). Conservative politicians opposed both the introduction of ballot envelopes and isolating spaces. The strategy they pursued during the parliamentary deliberations over ­these issues was to invoke myriad procedural difficulties that could occur in the implementation of such new mea­sures. In opposing ballot envelopes, Conservatives argued that just like individual ballots, envelopes could also be marked with exterior signs and that they could be pressed together (eingekniffen) to allow election officials to infer voters’ choices. Another practical objection was that voters might be unable to place such ballots in envelopes, which was likely to slow down the voting pro­cess and increase the number of contested elections (StBer DtRt, 15 January 1890). In the discussion of isolating spaces, Conservatives argued that they lacked the means to establish the provisions that would be required by law (StBer DtRt, 15 May 1895). Georg Müller, a Conservative deputy representing Marienwerder, a rural district in East Prus­sia, noted that Belgium and Britain—­two countries with high population densities—­had the necessary preconditions to introduce isolating spaces. By contrast, no single rural district in East Prus­sia could meet the requirements of the new law. It was impossible to establish isolating spaces ­because most voting places on the

170 C H A P TER 5

“flat land” lacked separate rooms (ibid., 15 January 1890). It was relatively easy to dismiss such arguments as an exaggeration. One politician countered: One can find carpenters everywhere, and if ­there are no carpenters, then one can find cabinet makers. If ­there are no cabinet makers, then one can find o ­ thers who can build with very l­ ittle effort a structure in the form of a large fire screen and place a curtain around so that the voting ­table is protected and nobody can peek inside. Objections of this kind cannot be taken seriously. (ibid., 29 January 1902) The composition of the electoral co­ali­tions in f­ avor of electoral reforms changed, however, during the negotiations of ­these proposals. National Liberals w ­ ere the pivotal group of legislators whose position on the desirability of electoral reforms shifted during this period. As discussed above, during the 1880s, National Liberals voted together with Conservatives to oppose the introduction of ballot envelopes and isolating spaces. During the 1890s, however, the position of National Liberals changed, facilitating the creation of an electoral majority in support of reforms and contributing to the passage of laws protecting voter secrecy in 1903. In chapter 3 I documented a similar shift in the position of National Liberals ­toward electoral strategies that deployed state resources during campaigns. I have also shown how the breakdown of the Kartellbündnis between National Liberals and Conservatives changed electoral competition and increased the importance of runoffs. In this po­liti­cal context, clientelistic or coercive strategies became electorally costly for resource-­endowed legislators, as they could limit co­ali­tional opportunities during runoffs. Similar considerations are also at work in explaining the shift in the position of National Liberals about the desirability of reforms to protect voter secrecy. To maintain the possibility of forming electoral co­ali­tions with F ­ ree Liberals or Zentrum politicians during runoffs, National Liberals had to forgo electoral strategies that deployed employers as brokers. National Liberal objections to the reform reflected the arguments of Conservatives. Gustav Struckmann, a National Liberal from Hildesheim, argued that concerns about violations of voter secrecy had been strongly exaggerated (StBer DtRt, 15 January 1890). The proposals ­under consideration did not provide an improvement but only moved the situation “from Scylla to Charybdis.” The main prob­lem of the proposals to introduce ballot envelopes or isolating spaces was their excessive formalism: “You pile up form ­after form, you torment the ­people, you torment the voters and at the end, you achieve nothing.” Struckmann also derided the practical provisions of

P rotec ti n g Vote r Au to n o m y 171

the legislation, such as the proposal that voters close the envelope and hand ­these over to the president of the voting place. ­These rules ­were “ludicrous” and likely to lead to more protests: “The more forms you create, the more likely is it that voters ­will choose not to follow ­these formalisms and the more likely it ­will be that the election ­will be contested” (ibid.). Similar to Conservatives, National Liberals attempted to kill the proposal using myriad practical objections. It was unclear, Struckmann argued, ­whether voters had to glue the envelopes. Invoking the experience of Württemberg, a state that had introduced ballot envelopes in 1883, Struckmann argued that such ambiguities led to countless protests and calls for invalidation. National Liberals also objected to the requirement that voters place ballots in envelopes: “The hands of our workers are used to other activities than the ­handling of paper. It is not at all easy to go ­behind the curtain and place papers quickly in the envelope” (StBer DtRt, 15 January 1890). During the following years, however, National Liberals’ position on the desirability of electoral reforms changed. Representing the National Liberal del­e­ga­tion in the Reichstag on the question of both ballot envelopes and isolating spaces, Ernst Bassermann noted that in 1894, National Liberals ­were split. Bassermann commented that while some National Liberal groups “cannot yet endorse the proposal, a large part of our fraction ­will support the proposal” (StBer DtRt, 17 April 1894). At the time, National Liberals had greater reservations about isolating spaces, which w ­ ere considered “impracticable for cities and for flat land” (ibid.). In 1896, Bassermann reported that the National Liberal del­e­ga­tion supported electoral reforms (StBer DtRt, 29 January 1896). National Liberal support for this legislation remained unchanged ­until the adoption of the legislation in 1903. In his 1896 intervention in the Reichstag, Bassermann justified National Liberal support for protecting ballot secrecy using two dif­fer­ent considerations. On the one hand, Bassermann invoked the electoral commission’s discoveries of violations of voter secrecy. For many deputies, the long pro­cess of investigation created profound uncertainty about the state of their mandate. On the other hand, Bassermann argued that by limiting the ability of brokers to monitor voters’ choices, the new law created significant electoral advantages for deputies who faced Social Demo­cratic challengers in their districts. This latter statement was not only a rhetorical strategy to equate the electoral interventions of Social Demo­crats and the economic coercion of many candidates on the right. Rather, Bassermann implied ­here that at this par­tic­u­lar moment in time—­where electoral coercion was both eco­nom­ically and electorally costly—it was more efficient to

172 C H A P TER 5

rely on a law to limit campaigning by Social Demo­crats rather than deploy the electoral strategies used in previous elections (ibid.). Other parties supporting electoral reforms welcomed this shift in the National Liberal position. Commenting on this, Heinrich Rickert remarked that “I am very happy that a part of our neighbors has withdrawn their previous objections. This w ­ ill strengthen the majority in support of reform and ­will significantly add to the weight of our recommendation” (StBer DtRt, 17 April 1894). During the 1896 deliberations on the law, Franz Schädler, a Zentrum deputy from Eichstatt, noted that initially the proposal for reform had not received a friendly reception. Over time, however, the proposal “had earned more friends,” benefiting from a parliamentary majority (ibid., 29 January 1896). Ultimately, the electoral majority in the Reichstag that supported the adoption of ballot envelopes and isolating spaces included ­Free Liberals, Catholics, Social Demo­crats, and National Liberals. Violations of voter secrecy persisted, however, even ­after the introduction of ballot envelopes in 1903. The source for the per­sis­tence of ­these practices was the design of ballot containers or urns. The German electoral law did not include mandates for standardized urns, leaving their shape and design in the hands of local election officials. As a result, a variety of objects ­were used to collect ballots. Th ­ ese included soup bowls, “cigar boxes, drawers, suitcases, hat boxes, cooking pots, earthen bowls, beer mugs, plates and washtubs” (Siegfried 1903). Commenting on this situation, Ignaz Auer, a Social Demo­cratic deputy, noted that while in Elsass, urns of proper proportions dated to the period of French occupation, Germans had to content themselves with cigar boxes, soup tureens, and white beer glasses as electoral urns (StBer DtRt, 29 January 1902). The small size of ballot containers led to the stacking of ballots, which facilitated violations of voter secrecy. As historian Brett Fairbairn described this prob­lem, “ballots would fall in such a way that they lay directly on top of each other in the exact order in which they w ­ ere dropped in. This would allow a vengeful official to compare the stack with a list of the order in which ­people had voted and arranged for a punishment for t­ hose that had voted in the wrong way” (1997: 218). A report on electoral irregularities in Nienburg, a district in Hannover, described the urn as follows: “In Elze bei Biffendorf, the urn was constructed as to allow each ballot to glide so that each envelope was stacked on the previous one. This allowed officials to determine each voter’s choice. In addition, urns had knot-­holes [Astloch] which allowed election officials to control w ­ hether ballots fell on top of each other” (StBer DtRt, encl. no. 702).

P rotec ti n g Vote r Au to n o m y 173

The prob­lem of stacking ballots persisted even ­after the introduction of isolating spaces and envelopes. Disagreement existed, however, about the magnitude of this prob­lem. A number of politicians argued that the Rickert law had, in fact, diminished voting secrecy b ­ ecause some ballot containers ­were too narrow to accommodate envelopes (StBer DtRt, 25 February 1907). By contrast, Conservative politicians and bureaucratic officials contested the severity of the prob­lem. The Prus­sian Ministry of Interior conducted a comprehensive survey of the types of urns used during the 1903 and 1907 elections. The study identified 156 complaints about the stacking of ballots but concluded that ­these irregularities did not have a significant impact on the outcome of elections (Bundesarchiv Berlin Lichtenberg Wahlurnen R1501/114475). The standardization of urns was the last step of the reform package to protect ballot secrecy. The Reichstag ­adopted legislation mandating uniform urns in 1914. ­Table 5.1 lists all legislative proposals submitted to the Reichstag to better protect voting secrecy. Zentrum and F ­ ree Liberal politicians who recommended the introduction of ballot envelopes and isolating spaces also formulated proposals to standardize the design of the electoral urn. The ­Free Liberal proposals included additional provisions for election officials to verify that urns w ­ ere covered and emptied before voting began (StBer DtRt 1899, encl. no. 22). A proposal for reform submitted by Zentrum deputies in 1900 recommended modifying the electoral law, mandating that urns remained closed during the duration of voting and that election officials shake urns before opening them to count the ballots (StBer DtRt 1907, encl. no. 112). This additional step in the reform mea­sure to protect ballot secrecy was also supported by the National Liberal del­e­ga­tion in the Reichstag. Prominent National Liberal politicians, such as Gustav Streseman and Ernst Bassermann, cosigned a proposal to introduce standardized urns, which was submitted to the Reichstag in 1909 (StBer DtRt 1909, encl. no. 91). To standardize the design of urns, National Liberals recommended the following amendment to the electoral law: “A covered container (electoral urn) to collect ballots ­will be placed on this ­table. The urn ­will be provided by the Reich to all electoral wards at no cost. The urn has to be manufactured in such a way so that it cannot be established who placed the ballots inside the urn. Before the vote, the electoral committee has to ascertain that the urn is empty” (ibid.). Bassermann, the leader of the National Liberal Fraktion in the Reichstag, intervened on subsequent occasions in support of ­these proposals. On 29 February 1912, Bassermann addressed representatives of the government

174 C H A P TER 5

­TABLE 5.1.

Proposals for better protection of voting secrecy considered by the German Reichstag

Year

Document no.

Proposal for reform

Initiating politician

1878

66

Ballot envelopes

Blos, Most (SPD)

1878

119

Ballot envelopes

Liebknecht (SPD)

1881

66

Ballot design

Woelfel (NL)

1889

26

Ballot design and ballot envelopes

Barth (DFP) and Rickert (FVP)

1890

139

Ballot design

Gröber

1892

30

Ballot envelopes

Barth (DFP) and Rickert (FVP)

1892

35

Ballot envelopes

Gröber

1894

20

Ballot envelopes and secret urn

Gröber, von Heereman, Lieber, Rinteln, Schädler, Spahn, Wenzel

1894

21

Ballot envelopes

Rickert (FVP)

1895

25

Ballot design and secret urn

Rickert (FVP)

1899

22

Ballot design and secret urn

Rickert (FVP)

1900

33

Ballot envelopes and secret urn

Gröber (Z)

1907

112

Electoral urn

Hompesch, Schädler, Spahn (Z)

1909

47

Protection of electoral secrecy

Ablass (FVP)

1909

91

Electoral urn

Bassermann (NL)

1910

214

Electoral urn

von Hertling (Z)

1911

816

Electoral urn

Ablass (FVP)

Source: StBer DtRt, vari­ous years.

regarding the absence of standardized urns: “Why do we have to experience year a­ fter year complaints about soup tureens and cigar boxes?” (StBer DtRt, 29 February 1912). The following year, Bassermann urged the Prus­sian Ministry of Interior to approve proposals for uniform urns. According to Bassermann, delays in the adoption of urns could be attributed to the “shameful dependence of the government on the Prus­sian Ministry of Interior” and to the re­sis­tance of Minister of Interior von Dallwitz to the introduction of urns (ibid., 16 April 1913). ­These interventions by prominent National Liberal politicians in support of standardized urns demonstrate that co­ali­tional realignment had taken place in Germany. On the eve of World War I, the electoral co­ali­ tion supporting electoral reforms also included resource-­endowed legislators representing parties on the right, such as National Liberals. I now turn to a quantitative analy­sis of the reform proposals. The German Reichstag did not rec­ord the individual votes taken on ­these vari­ous

P rotec ti n g Vote r Au to n o m y 175

proposals (Mares 2015). In the absence of ­these votes, I proxy demand for reform using information about the cosponsorship of vari­ous legislative proposals supporting electoral reforms. In related publications, I have examined ­whether economic theories of democ­ratization can explain support for ­these reforms (ibid.; Kasara and Mares 2016). ­Here, I expand ­these analyses by exploring the importance of economic and electoral considerations in explaining support for reforms. Throughout this book, I have proxied demand for reforms as a function of relative resources, on the one hand, and economic and electoral constraints, on the other hand. While resource-­endowed politicians are likely to resist reforms, I have conjectured that rising economic and electoral costs may reduce the attractiveness of the status quo among such politicians. This change in preferences, in turn, facilitates a co­ali­tional realignment in support of reforms. The data set constructed to examine ­these propositions brings together economic and po­liti­cal variables with information about the occupations of individual legislators. Information about the sources used to construct ­these variables as well as additional descriptive statistics are included in the appendix. The models presented in figure 5.3 examine the correlates of support for ­these electoral reforms. In panel (a), I examine the consequences of resource asymmetries for demand for electoral reforms. As imperfections in ballot design opened up opportunities for economic coercion, employers enjoyed a resource advantage and ­were likely to oppose electoral reforms. The results confirm this hypothesis. Consistent with their resource advantage, both rural and urban employers opposed the introduction of electoral reforms. Consider now partisan disagreement about the desirability of electoral reforms. Resource-­constrained F ­ ree Liberals, Catholics, and Social Demo­crats supported the reforms. National Liberals also supported legislation protecting voter secrecy. As discussed above, the position of National Liberals with regard to t­ hese electoral reforms shifted over the course of the period ­under study, from opposition to an active endorsement of proposals for reform. This shift in the position of National Liberals (who discounted their relative resource advantage) illustrates the co­ali­tional realignment that took place during this period. The main conjecture of the modernization perspective is that economic development spurs demand for electoral reforms. Increases in population and urbanization are expected to increase demand for electoral reforms by raising the economic costs of vari­ous illicit strategies (Stokes et al. 2015). We find mixed support for this hypothesis in explaining demand for ballot

176 C H A P TER 5

(a) Individual and Partisan Resources

(b) Economic constraints

Individual resources Employer

Landholding inequality

Partisan resources Other

Population (logged)

Conservatives Economic development

National Liberal Free Liberal Zentrum

–2

–1

0

1

2

Social Democrat

Urbanization –5

–4

–3

–2

–1

0

1

2

(d) Full model

(c) Electoral constraints

Individual resources Employer Partisan resources Other Conservatives National Liberal Free Liberal Zentrum Social Democrat

Electoral constraints

Economic constraints Landholding inequality Population (logged) Economic development Urbanization

Elected in runoff

Electoral constraints Elected in runoff –1

0

1

–3

–2

–1

0

1

2

Support for reforms to protect voter secrecy in Germany. Notes: Results pre­sent correlates of support for vari­ous proposals to better protect voter secrecy. All models are estimated using multilevel mixed-­models regression. See the appendix for the sources used to construct the variables used in the analy­sis. FIGURE 5.3.

secrecy in Germany. Contrary to the predictions of modernization theories, proposals for electoral reforms originated with legislators from less populous districts. Moreover, t­ here is no significant relationship between the urbanization of a district and support for the introduction of ballot secrecy. ­There is, however, a positive relationship between the level of economic development in a district (as mea­sured by the percentage of nonagricultural employment) and support for legislation to protect ballot secrecy. In the model presented in the lower left half of figure 5.3, I examine ­whether considerations about electoral costs explain demand for electoral reforms. As discussed in chapters 2 and 3, the breakdown of the electoral cartel between National Liberals and Conservatives and the associated increase in fragmentation among right-­wing parties increased the electoral costs associated with illicit strategies. I conjectured that the use of coercion is costly in ­these fragmented po­liti­cal environments, as it may foreclose opportunities for electoral alliances during runoffs. We find some initial support for this

P rotec ti n g Vote r Au to n o m y 177

conjecture. The bivariate relationship between runoffs and cosponsorship of proposals for ballot secrecy is positive. However, the relationship between runoffs and support for electoral reforms does not reach significance at conventional levels in the full model reported in panel (d) of figure 5.3. The results presented in figure 5.3 support several of the theoretical conjectures formulated in chapter 2. First, the analy­sis demonstrates the importance of resource asymmetries in explaining support for or opposition to electoral reforms. The initial demand for electoral reforms originated with resource-­constrained politicians. I have also documented that this resource asymmetry had both individual and partisan determinants, with employers resisting electoral reforms. However, other considerations in addition to the relative resource advantage explain some of the partisan positions about the desirability of reforms. Despite their similar access to economic brokers, the position of Conservatives and National Liberals with regard to the desirability of electoral reforms diverged during this period. The shift in the position of National Liberals—­announced by Ernst Bassermann in 1896—­can be attributed to economic and electoral concerns. This shift facilitated the formation of an encompassing electoral co­ali­tion that made the adoption of legislation protecting ballot secrecy pos­si­ble. Conclusion

The adoption of electoral reforms to provide better protection of ballot secrecy is one of the most significant achievements in the pro­cess of democ­ ratization following the extension of suffrage. The adoption of ballot envelopes and isolating spaces protected voters’ autonomy at the moment they cast their ballots and l­ imited the ability of candidates and economic brokers to monitor voters’ choices. Over time, such reforms ­limited economic coercion, a pervasive practice in nineteenth-­century elections. This chapter has examined the formation of po­liti­cal co­ali­tions supporting ­these reforms. As with the other electoral changes examined in the book, demand for reform originated with resource-­constrained legislators who lacked the means to engage in coercion. In France, this initial co­ali­tion included Socialists and some Radicals. In Germany, it included Social Demo­ crats, F ­ ree Liberals, and the Catholic Zentrum. However, an encompassing electoral majority that supported reforms came about only a­ fter resource-­ endowed legislators changed their preferences about the desirability of the status quo and embraced electoral reforms. I have documented the most impor­tant po­liti­cal and economic changes that facilitated this co­ali­tional

178 C H A P TER 5

realignment and some of the calculations of resource-­endowed legislators who faced increasing economic and electoral costs associated with illicit strategies. In France, the split between Republicans and Radicals and the rise in the importance of programmatic competition decreased Republican opposition to ­these electoral reforms and facilitated the formation of an encompassing majority in support of change. In Germany, National Liberal politicians’ shift made a parliamentary majority in f­ avor of reforms pos­si­ble. In both cases, concerns about electoral costs explain this co­ali­tional realignment and support for reforms.

6 Curbing Electoral Fraud

In addition to electoral irregularities that took place at the time of voting, electoral malfeasance could also occur a­ fter ballots w ­ ere cast. At this time, voting bureau presidents and members could alter the number of ballots in the electoral urns, thus modifying the results of the popu­lar ­will. A more colloquial term for ­t hese practices, encountered in French electoral reports, was “garnishing the urn.” Penal codes and electoral laws ­adopted during the first de­cade of the nineteenth ­century referred to this irregularity as fraud. In this chapter, I examine the adoption of legislation to limit the incidence of fraud in first-­wave democracies. Similar to the vote-­buying reforms examined in chapter 5, the adoption of reforms to limit fraud proceeded in two stages. The earliest efforts to limit fraud originated with the Napoleonic Code. The code mandated strong punishments for anyone found guilty of modifying election results ­after ballots ­were cast, sanctioning such illicit strategies with prison, monetary fines, and the loss of po­liti­cal rights. Aided by the Napoleonic invasion of Eu­rope, the code diffused rapidly across several continental Eu­ro­pean countries. Electoral laws enacted during the first de­cades of the nineteenth ­century incorporated the provisions of the Napoleonic Code sanctioning fraud. A second wave of electoral reforms to limit fraud took place during the second half of the nineteenth c­ entury. The objective of ­these reforms was to impose further constraints on the ability of voting place presidents to modify electoral outcomes ­after ballots ­were cast. The proposed solution to this 179

180 C H A P TER 6

prob­lem was to allow all candidates to appoint representatives to supervise electoral operations. I ­will examine the early adoption of reforms allowing candidate representatives to supervise electoral operations in Britain and Belgium. In both countries, the low resource asymmetry among mainstream parties facilitated the formation of a po­liti­cal compromise in support of ­these reforms. Next, I examine po­liti­cal developments in Imperial Germany. The German electoral law of 1870 already included provisions upholding the publicity (Öffentlichkeit) of the vote. This provision facilitated changes in jurisprudence that enabled candidates to nominate their representatives in the voting place. By 1900, elections that did not allow candidate representatives to be pre­sent in voting places w ­ ere routinely voided. In the final part of the chapter, I ­will examine po­liti­cal developments in France. ­Here, deep disagreement about the po­liti­cal consequences of candidate representatives constrained the formation of a compromise on this issue. Using roll-­call votes and parliamentary deliberations, I seek to document the most impor­tant arguments advanced by legislators who supported reforms and how a parliamentary majority supporting the adoption of candidate representatives emerged. The Penalization of Fraud in Nineteenth-­Century Eu­ro­pean Countries

­ egal efforts to sanction fraud originated with the Napoleonic Code of L 1810. Articles 111 and 112 defined electoral fraud and enumerated the sanctions against t­ hose who engaged in this strategy. Article 111 of the code defined acts of fraud as including the falsification of ballots, the subtraction or addition of ballots at the time they w ­ ere counted, and the modification of the name of a candidate on a ballot (de Kéguelin 1904: 46). Article 112 of the code listed sanctions for t­ hose who committed t­ hese electoral irregularities. They included imprisonment for six months to at least two years and the loss of po­liti­cal rights—­such as right to vote or the right to be elected—­for a period of five to ten years. ­Legal experts of the period argued that the punishments defined by the Napoleonic Code w ­ ere not severe enough. The classic treaty on penal law at the time—­the Théorie du Code Pénal (1852) by Adolphe Chauveau and Faustin Hélié—­criticized the Napoleonic Code for sanctioning vio­lence much more harshly than fraud. Chauveau and Hélié noted that it is “strange that voiding of voters’ rights, an act of usurpation and tyranny which, by its very nature can disturb society in its entirety, should be less severely punished

C u r bi n g E lec to r a l Fr au d 181

than inconsequential acts of vio­lence against a local policeman and a bailiff ” (ibid., 59). Yet despite t­ hese shortcomings, the Napoleonic Code proposed a clear ­legal pathway to sanction fraud. This stood in contrast to vote-­buying, another electoral irregularity that was sanctioned in the code. As discussed in chapter 5, it was nearly impossible for courts at the time to sanction acts of vote-­buying, as plaintiffs had to offer evidence of the “corrupt intent” of the agent who had made the illicit offer to a voter. This was an insurmountable hurdle that was not pre­sent in the case of electoral fraud. The provisions of the Napoleonic Code diffused to neighboring countries and ­were incorporated in penal codes ­adopted during the first half of the nineteenth ­century. The Belgian Penal Code, which was ­adopted ­under French occupation, used the same definition of fraud as the French code (Nypels 1868). A subsequent penal code, ­adopted a­ fter Belgium gained in­de­pen­dence, set out even stronger punishments for fraud. The addition, subtraction, or modification of ballots as well as the reading of dif­fer­ent names than ones written on ballots w ­ ere sanctioned by prison terms of up to two years and by high fines (ibid., 330). The provisions sanctioning fraud enacted by the Napoleonic Code ­were also incorporated in the penal code of the German Empire. Prus­sia a­ dopted ­these provisions in its 1851 penal code. Paragraph 51 of the code sanctioned fraud with a prison term of one to three years (Strafgesetzbuch Preussen 1851). From ­here, ­these sanctions made their way to the draft of the penal code of the North German Confederation, which became the Penal Code of the German Empire (Reichsstrafgesetzbuch) in 1870. ­Here, the definition of electoral fraud included announcing incorrect results and falsifying election outcomes. Such acts w ­ ere also sanctioned by prison terms and the loss of po­liti­cal rights. The sanctions w ­ ere more severe if the perpetrator was an employee of the state (Hatschek 1915: 209). Electoral laws ­adopted in Eu­rope during the first part of the nineteenth ­century incorporated some of the provisions already pre­sent in the penal code, thus establishing a double protection against fraud. In France, the electoral law of 15 March 1849 sanctioned vari­ous efforts to modify elections. ­These included the “removal of urns containing ballots that had not been counted yet” and the “modification of the votes made e­ ither by the members of the voting bureau or by authorities designated to guard the ballots” (Article 114). The electoral law already imposed punishments for ­these illicit strategies. Fraud was punished by a prison term of one to five years and by a fine that could be as high as five thousand francs. In addition, Article 116 of the electoral law of 1849 sanctioned ­those who committed fraud with the

182 C H A P TER 6

“loss of the right to elect and be elected” for a period of up to five years. A subsequent electoral law, the Organic Decree of 1852, reiterated t­ hese provisions. Commenting on t­ hese provisions, de Kéguelin, a professor of penal law, noted that the 1849 law and the 1852 decree “represented a genuine code of electoral fraud whose provisions went well beyond what had been accomplished by the 1810 Code” (1904: 72). The Introduction of Candidate Representatives in Britain

The early reliance on the penal code remained an effective strategy to sanction fraud. During the second part of the nineteenth ­century, most Eu­ro­ pean countries ­adopted further reforms that attempted to place additional constraints on the ability of voting section presidents to modify the outcome of a race a­ fter votes ­were counted. A common theme among ­these reforms was to allow all candidates to appoint representatives who could supervise electoral operations and ballot counting. This idea originated in Britain, which ­adopted the ballot only in the mid-­nineteenth ­century. ­Because Britain did not have a penal code, it was imperative to design safeguards for ballot counting. Belgium ­adopted a similar reform in 1877, and by 1914, a large number of first-­wave democracies had modified their electoral law or jurisprudence to allow candidate representatives at voting places. In Britain, open voting prevailed for a much longer period of time than in other first-­wave democracies. The adoption of the Secret Ballot Act in 1872 marked the transition from open to secret voting. Prior to 1872, all actors pre­sent at the time of voting could observe all votes. The tallying of raised hands presented ­limited opportunities for fraud. While the rigging of voters’ lists, coercion, and intimidation remained pervasive irregularities in British elections during the period of open voting, fraud was virtually non­ex­is­tent. Britain introduced candidate representatives si­mul­ta­neously with secret voting. During the parliamentary deliberations of the time, the discussion regarding the pos­si­ble illicit strategies that could occur at the time ballots ­were counted was, however, remarkably brief. This reflected legislators’ lack of experience with the practicalities associated with collecting and counting paper ballots. Few politicians ­were wary of the risk of fraud. In the discussions of the time, t­ hese politicians drew on examples in neighboring countries to raise awareness about the likely incidence of ­these illicit strategies. Invoking his experience with elections in France, one politician recalled that “twenty years ago when the vote was taken on a new system of government it was almost universally believed that unfair returns ­were made; the

C u r bi n g E lec to r a l Fr au d 183

officials, indeed, being so zealous for their new master that they returned a larger number in his ­favor than that of the ­whole adult male population of France.” In France, acts of fraud ­were less likely in populous districts such as Paris or Lyon and w ­ ere more frequent “in many rural departments where duplicate locks or seals ­were used, by means of which the voting papers ­were manipulated overnight.” ­Others argued, however, that the prob­lems “encountered in neighboring countries ­were unlikely to arise ­here” (H.C. Debates, 9 May 1872). A consensus that it was impor­tant to “take precautions against the ballot box being tampered with by the Returning Officer” emerged with relative ease (H.C. Debates, 9 May 1872). To limit the risk of fraud, W. E. Forster, a liberal politician, proposed that candidates be allowed to appoint their own “agents” to supervise electoral operations. Forster was one of the main architects of the Secret Ballot Act and Gladstone’s most impor­tant ally in the enactment of this law. The proposal to introduce candidate representatives was introduced by Forster as an amendment to the electoral law during the committee deliberations. A more stringent recommendation considered at the time was for candidates’ agents to supervise ballot boxes during the night between the day of the vote and the day ballots ­were counted. However, this proposal was considered impractical. It was impossible to guarantee that candidate representatives would not fall asleep and lose sight of the ballot boxes. Other legislators objected that such arrangements would add unnecessarily to the expense of elections. At the time of the adoption of secret voting, British reformers established rules regarding the time at which ballots ­were counted. One proposal recommended that vote counting should occur immediately ­after the last ballot was cast. “Packets of votes [would be conveyed] to the Returning Officer so that the number of votes may be ascertained before the presiding officer of each station or the agents (if any) of the candidates ­shall have lost sight of the ballot box” (H.C. Debates, 8 May 1872). This recommendation met with considerable opposition. Opponents of the mea­sure contended that “it would be unnecessarily tedious to place all 100 of presiding officers in one room and count the votes in the presence of the Returning Officer and of the agents. . . . ​In a large borough or county, sleep would close the eyes of some of the persons who ­were to watch or take part in it” (ibid.). An alternative proposal recommended counting the ballots in a dif­fer­ent location than the voting place. The solution had two advantages. On the one hand, it allowed for a rapid centralization of all ballots of a constituency. On the other hand, it further constrained voting place presidents—­known as

184 C H A P TER 6

presiding officers—­from tampering with ballots. Opponents of this solution preferred “to count the ballots before they left the polling booth” and not “lose sight of ballots ­until the results had been ascertained.” A majority of lawmakers argued, however, that it was impossible to finish counting ballots immediately ­after they w ­ ere cast. The solution that ultimately prevailed introduced a differentiation between the “voting place” and the “polling place,” the latter referring to the place where the votes w ­ ere counted. To minimize the risk of ballot tampering before the votes ­were counted, the electoral law included additional provisions requiring ballots to be collected in separate packets at the closing of the polls. Presiding officers had to seal ­these packets in the presence of candidates’ representatives. Next, presiding officers w ­ ere required to deliver such packages to Returning Officers, who could count the ballots only in the presence of candidate representatives. If it was not pos­si­ble to finish the counting in one day, Returning Officers w ­ ere required to seal the ballots again while candidate representatives supervised ­these operations. In short, the solution ­adopted in Britain at the time laid out specific instructions and responsibilities for candidate representatives at all moments of time between when ballots ­were cast and results ­were announced. The Ballot Act (1872) included two novel provisions that attempted to limit the incidence of fraud. The first was the introduction of candidate representatives at the voting place. The second was the separation between voting places and polling places. In Britain, both reforms ­were ­adopted with very ­little po­liti­cal disagreement. Concerns that such rules could asymmetrically help politicians from one party w ­ ere remarkably absent in Britain. In the absence of such fears, both Liberal and Conservative politicians endorsed the adoption of candidate representatives as an impor­tant constraint on polling place presidents. The Introduction of Candidate Representatives in Belgium

The provision in the British Ballot Act to introduce candidate representatives served as a blueprint for the reform proposals considered in Belgium in 1877. Gustave Tack, a Catholic legislator, praised Britain’s electoral reform as a “fortunate innovation which provides novel guarantees for the secrecy of the vote” (Chambre des représentants de Belgique, Annales Parlementaires, 23 May 1877, p. 819). Jules Malou, a Catholic politician and Belgium’s prime minister from 1874 to 1878, embraced proposals to introduce candidate

C u r bi n g E lec to r a l Fr au d 185

representatives as part of an electoral reform initiated in 1877. Taking the name of its sponsor, the reform was also known as the loi Malou. Article 4 of the new electoral law allowed candidates to appoint representatives—­called witnesses (témoins)—to supervise all electoral operations. Candidates could designate one representative for each voting precinct. The electoral law also gave candidates the option to supervise both electoral operations and the counting of ballots themselves. Candidate representatives w ­ ere allowed to be pre­sent during the entire vote and sit in close proximity to the president of the voting bureau (Article 23). Before the vote began, candidate representatives ­were required to take the oath “I pledge to guard the secret of the vote” (Article 24) (Chambre des représentants de Belgique, 1876, Documents Parlementaires, no. 84). During the discussions in parliament, proposals to introduce candidate representatives met with widespread po­liti­cal support. ­These provisions ­were ­adopted without extensive discussion due to the low levels of opposition. The only question that generated disagreement on the floor of the Chamber was the allocation of candidate representatives among established parties (such as Catholics and Liberals) and in­de­pen­dent candidates. The law mandated that in single-­member districts, both candidates representing parties and in­de­pen­dents could appoint one representative. However, the law established some constraints in multimember districts. To prevent overcrowding in the voting place in multimember districts, the presidents of voting places ­were allowed to use a lottery to reduce the number of representatives appointed by in­de­pen­dent candidates. In such cases, the maximum number of representatives of in­de­pen­dent candidates was three (Chambre des représentants de Belgique, Annales Parlementaires, 23 May 1877). A report on the law clarified this provision: “Based on the observations made in the committee and as confirmed by the explanations made by the government, it follows that this article gives both complete and incomplete lists of candidates the right to appoint representatives. By contrast, in­de­pen­dent candidates, no ­matter what their number, cannot appoint more than three candidates in a polling place” (Chambre des représentants de Belgique, 1876, Documents Parlementaires, no. 84). Opponents of the proposal objected to this restriction, which favored established parties but was detrimental for in­de­pen­dent candidates. Multimember districts, such as Brussels, elected thirteen representatives. In a likely pattern of electoral competition for Brussels, Liberal and Catholic candidates could encounter Radicals and a variety of in­de­pen­dent candidates. Some of ­these in­de­pen­dents could choose to compete u ­ nder a party

186 C H A P TER 6

list, while o ­ thers w ­ ere likely to reject such an umbrella affiliation. In such situations, the provision allowing candidate representatives was nothing more than a bonus for large parties. Borrowing from Britain’s Ballot Act, the Belgian electoral law also proposed designating separate voting and polling places. Article 30 of the legislation mandated that “ballots cannot be counted by the bureau that has received them.” Belgian lawmakers a­ dopted an additional innovation that attempted to limit the ability of voting station presidents to modify election outcomes. They chose to rely on lotteries to assign ballots of dif­fer­ent voting places to dif­fer­ent bureaus where t­ hese ballots could be counted. The provision read as follows: If ­there are three sections, the main bureau counts the ballots of the two other bureaus. One of the smaller bureaus, chosen by lot, counts the ballots of the main bureau. If ­there are more than three sections, the main bureau which is responsible for the counting of ballots is chosen by lot. . . . ​The draw of the lot takes place in the main bureau before the vote. (Article 50) The newly introduced distinction between voting and polling places also necessitated a rule about the transportation of ballots between t­ hese sites. A short debate about how to design such rules ensued. The initial proposal formulated by the commission recommended that the president of the voting place and one of the scrutineers seal the urn carry­ing the ballots. One of the scrutineers and one candidate representative w ­ ere to carry the urn to the bureau that had been randomly chosen to count the ballots. Liberal politicians argued that ­these provisions w ­ ere “too ­simple” and “too primitive” and likely to lead to “severe incon­ve­niences.” The worry expressed at the time was that the scrutineer and the witness transporting the urn could be members of the same party. If this was the case, the law could not guarantee that ballots ­were protected against fraud. In defense of this proposal, Jules Malou argued that sufficient guarantees ­were in place. Scrutineers and witnesses who tampered with the urn “risked placing themselves in the ­middle of the penal code” (Chambre des représentants de Belgique, Annales Parlementaires, 23 May 1877). Malou’s re­sis­tance was unsuccessful in defeating the Liberal proposal. The compromise solution ­adopted by the Belgian Chamber of Deputies included the Liberal amendment which required that representatives of all parties ­were pre­sent during the transportation of ballots (Chambre des représentants de Belgique, Annales Parlementaires, 24 May 1877). In addition, the

C u r bi n g E lec to r a l Fr au d 187

Belgian legislation gave candidate representatives the right to challenge the results once ballots ­were counted. ­After the president announced the results of a vote, candidate representatives could submit their complaints. Fi­nally, the law introduced sanctions for candidate representatives who engaged in fraud by “subtracting or adding ballots or by fraudulently altering ballots in order to void them” (ibid.). Such irregularities ­were punished with a prison term of three months to two years and a fine of fifty to one thousand francs. This provision extended sanctions already in effect u ­ nder the Belgian Penal Code. In 1894, the Catholic government led by prime minister Jules de Burlet made additional changes to the electoral law concerning the se­lection of candidate representatives. The loi Malou mandated that candidate representatives should be selected via a lottery. Practical experience led to a reconsideration of the usefulness of lotteries in such conditions. Expressing his reservation on lotteries, Burlet remarked that the role of candidate representatives at the voting place was to “observe the good functioning of the operations and the identity of voters. Given that the vote takes place in communes, ­there cannot be a more efficient or rational position . . . ​than to appoint a representative who is from the respective commune. Th ­ ere is no reason to maintain t­ hese rules, but t­ here are excellent reasons to abolish lotteries for the se­lection of candidate representatives” (Chambre des représentants de Belgique, Annales Parlementaires, 29 May 1894). Burlet’s position prevailed, leading to a simplification of the se­lection of candidate representatives at the time of voting. During the 1894 deliberations, several legislators continued to express concerns about the imperfect protection of the results of the vote ­after the ballots ­were counted. They argued that separate polling and voting places could actually facilitate irregularities rather than prevent them. Lawmakers objected to the rules regarding the transportation of urns, arguing that ­these could generate “very severe incon­ve­niences.” One such worry was that “the transportation of ballots ­will take place in complete darkness if the elections ­will take place at the end of October. It is thus not too difficult to foresee that it could be very easy to commit fraud. Even acts of vio­lence against ­those who transported the urns could take place” (Chambre des représentants de Belgique, Annales Parlementaires, 30 May 1894). Other legislators worried that transporting urns built from heavy wood opened up additional opportunities for fraud. An amendment proposed at the time recommended that ballots be transported in “recipients that ­were more malleable” (ibid.). The amendment was, however, derided as a proposal

188 C H A P TER 6

for an “accordion container” and rejected. Delbeke, the sponsor of the law, opposed this proposal, affirming that the punishments for fraud included in the existing bill ­were sufficient. Most lawmakers concurred that the penal code sanctions of fraud w ­ ere sufficient to deter irregularities that could occur during the transportation of the ballots. Between 1877 and 1895, Belgium introduced a number of electoral reforms seeking to limit fraud. Th ­ ese reforms, inspired by the British Ballot Act, separated the voting and polling places and allowed candidates to nominate “witnesses” to supervise electoral operations. Th ­ ese changes to the electoral law ­were ­adopted swiftly and without significant po­liti­cal disagreements. The low resource asymmetry among Catholics and Liberals, the two largest parties at the time, facilitated the po­liti­cal consensus. Legislators from both parties concurred that t­ hese additional reforms strengthened the preexisting provisions of the penal code safeguarding against fraud. However, the politics surrounding the introduction of candidate representatives changed dramatically once t­ hese ideas for reform made their way to France. ­Here, legislators’ fear that t­ hese reforms would enhance the resource asymmetries among candidates severely hampered the possibility of reaching a po­liti­cal consensus. I w ­ ill turn to the discussion of the French case ­later in the chapter. Limiting Fraud in Imperial Germany

In Imperial Germany, fraud was an electoral offense sanctioned by the penal code. Any party participating in an election could submit a complaint about fraud to the electoral commission. Beginning with the first national elections of the empire, the electoral commission of the Reichstag upheld the provisions of the penal law. One of the first contested elections involving fraud was the 1874 election in Oletzko, a district in East Prus­sia (StBer DtRt, encl. no. 82). Robert von Puttkamer, a candidate of the Imperial Party, won the district by a very narrow majority. Several complaints documented fraud in the counting of ballots in multiple precincts. The electoral commission of the Reichstag recommended voiding the election and asked the courts to further examine ­these irregularities (ibid.). The 1877 election in Lübbecke, which had been won by a candidate of the Imperial Party, was also voided ­because of incidents of fraud. The electoral commission determined that electoral urns had been insufficiently supervised in several precincts of the district. As a result, e­ ither officials supervising the election or third persons had fraudulently inserted a number of ballots in the urns (StBer DtRt 1877,

C u r bi n g E lec to r a l Fr au d 189

Incidents of Fraud in German National Elections 100

Number of incidents

90 80 70 60 50 40 30 20 4 1885

1890

1895 Year

1900

1905

Violations Articles 105-109 Penal Code Convictions Articles 105-109 Penal Code Convictions Article 108 Penal Code Incidents of fraud in German national elections. Sources: Freudenthal 1895: 38; Mayer 1910: 11.

FIGURE 6.1.

encl. no. 187; Klein 2003: 161). Th ­ ese considerations resulted in the voiding of this race. Historians of electoral politics in Imperial Germany agree that fraud was a rare phenomenon in German national elections. Summarizing this consensus, Arsenscheck notes that “acts of fraud committed by electoral bureaus ­were a prob­lem of secondary importance in Imperial Germany” (2014: 348). Instances of fraud where regional or national party organ­izations sought to coordinate their strategies across districts are notably absent in German elections. If pre­sent, fraud resulted from localized initiatives of individual voting bureau presidents. In figure 6.1, I draw on two available data sources to document the ­limited incidence of fraud at the time. The first of ­these sources was collected by Berthold Freudenthal, a ­legal scholar (Freudenthal 1895). Freudenthal’s time series covers the period between 1883 and 1892 and reports convictions of violations of Articles 105–9 of the penal code, in other words, all articles of the penal code addressing electoral irregularities. As the data encompasses other irregularities besides fraud (such as vote-­buying and vio­lence), it should be read only as an upper bound for the incidence of fraud. The second time series was compiled by Max Ernst Mayer, a professor of law and expert

190 C H A P TER 6

on the penal code. Mayer’s time series, which covers the period between 1902 and 1910, reports only cases of fraud, in other words, violations of paragraph 108 of the penal code (Mayer 1910: 11). Together, the two sources illustrate the l­ imited incidence of fraud in German national elections. During the period between 1870 and 1914, we find no significant legislative initiatives to impose additional sanctions on fraud. However, even though the Reichstag did not enact new legislation, changes in electoral jurisprudence during this time affirmed the importance of candidate representatives as safeguards against irregularities. As a result of ­these changes in jurisprudence, Imperial Germany’s policies largely reflected ­those ­adopted by other Eu­ro­pean countries at the time. The idea of candidate representatives that was pioneered in Britain in 1872 was not available as a blueprint at the time the German electoral law was ­adopted. However, Germany’s electoral law did include an impor­tant protection against fraud, which came in the form of its commitment to the “princi­ ple of publicity of elections” (Prinzip der Öffentlichkeit). Article 9 mandated that both voting and the determination of results w ­ ere public acts (Hatschek 1915: 167). Julius Hatschek has noted that the princi­ple of publicity was “a necessary condition of the validity of elections. Publicity guarantees that the wishes of voters are respected correctly” (ibid., 172). The electoral law was vague about the ways in which such election publicity could be ensured. In ­these conditions, the commission verifying electoral irregularities (Wahlprüfungskommission) of the Reichstag became an impor­tant arbiter that could adjudicate w ­ hether par­tic­u­lar electoral activities v­ iolated this princi­ple of publicity. The electoral commission endorsed a very expansive definition of Article 9 of the electoral law. The decisions of the commission spelled out the violations of the princi­ple of the public nature of elections. Such sanctions resulted in the systematic voiding of elections. On repeated occasions, the electoral commission of the Reichstag was asked to weigh in on w ­ hether the eviction of representatives of a par­tic­u­lar candidate from polling stations v­ iolated Article 9 of the electoral law. The commission de­cided that the “exclusion of candidate representatives [Vertrauensmänner] of a par­tic­u­lar party limits the publicity of elections and vitiates elections” (Hatschek 1915: 174). One early decision that was influential for subsequent electoral jurisprudence concerned the 1898 election in Lebus, a district in Frankfurt (StBer DtRt 1898, no. 706/707). The election was won by Gustav Haake-­Letschin, a landowner representing the German Imperial Party, who defeated Metzner-­Berlin, a Social Demo­cratic candidate, in a runoff. Even though representatives of the Social Demo­cratic candidate

C u r bi n g E lec to r a l Fr au d 191

presented letters of invitation from local election committees, they w ­ ere not allowed to be pre­s ent in the voting place and at the time ballots ­were counted. In this case, the commission ruled that “reprimanding and expelling representatives of individual parties v­ iolated the princi­ple of publicity of elections” (ibid.). In 1900, a decision of the electoral commission concerning similar incidents led to the voiding of the election. In 1898, Friedrich von Loebell, a Conservative candidate in Potsdam’s eighth district, defeated Heinrich Peuss, a Social Demo­cratic candidate, by a narrow margin of 166 votes. The Landrat of this district gave explicit ­orders to expel candidate representatives from polling places, which led to the exclusion of ­these observers from eleven precincts in the district (StBer DtRt 1898, encl. nos. 542, 662). During the discussion of this case in the Reichstag, Ignaz Auer, a Social Demo­cratic legislator, accused Conservatives of “creating a system aimed to destroy the publicity of electoral operations.” Auer argued that the electoral commission and the Reichstag had the responsibility to “ensure that the publicity of elections remains unconstrained, as prescribed by the law” (StBer DtRt, 1 May 1900). The majority of members of the Reichstag agreed with this position. In a vote taken on 1 May 1900, the Reichstag de­cided to void von Loebell’s election. The electoral commission reaffirmed this position regarding the exclusion of candidate representatives in a number of subsequent decisions (StBer DtRt 1912, encl. nos. 21, 728, 892, 1061, 1159, 1433). In all ­these decisions, the commission determined that candidate representatives provided the most immediate protection against pos­si­ble fraud. Julius Hatschek argued that if one considers the rulings of the electoral commission, the situation in Germany was remarkably similar to that in other Eu­ro­pean countries that had mandated the princi­ple of candidate representatives (Hatschek 1915: 174). The only difference between Germany and neighboring Eu­ro­pean countries was that this additional protection against fraud developed slowly as a result of changes in electoral jurisprudence rather than through legislative choices made by the Reichstag. The Introduction of Candidate Representatives in Third Republic France

To reduce the incidence of irregularities during the counting of ballots, the electoral law of the Third Republic retained several provisions of legislation enacted during the Second Empire (Pierre 1919: 148, JORF Lois et Décrets,

192 C H A P TER 6

2 February 1852). Articles 27, 28, and 32 of the 1852 electoral law defined the steps that had to be taken at the time ballots ­were counted and laid out punishments for t­ hose who transgressed t­ hese rules. Th ­ ese provisions w ­ ere clarified by additional administrative regulations. Similar to Germany’s law, the French electoral law included a commitment to the transparency of elections. Additional provisions that could guarantee transparency included the requirement that both voting and the counting of ballots take place in public municipal buildings, in a highly vis­i­ble manner. To minimize illicit activities, administrative decrees recommended that elections take place in city halls or other public buildings situated in the center of localities. Th ­ ese recommendations ­were followed in practice. Elections w ­ ere routinely held in a variety of public places, including schools, post offices, courts, barracks, and amphitheaters (Tanchoux 2004: 524). To limit the incidence of fraud, the electoral law and subsequent administrative decrees included extremely detailed regulations about the organ­ization of the voting place at the time ballots w ­ ere counted. For example, voting places had to be sparsely furnished. Article 29 of the 1852 electoral law mandated that the t­ ables and furniture necessary for the emptying of urns had to be arranged in such a way as to allow participants to move around them without any interference (Pierre 1919: 148). Only a few benches along the walls ­were allowed inside the voting place (Tanchoux 2004: 526). Commenting on ­these regulations, Philippe Tanchoux observed that “the soberness of the premises accentuates the visibility of the operations. Any incident during the duration of the vote which violates this visibility of the space ­will lead to the cancellation of the results only if the urn cannot always be observed by the public or by a representative of the voting bureau” (ibid., 527). The electoral law also included extremely detailed regulations for the pro­cess by which urns ­were to be emptied. Electoral bureau presidents ­were required to select scrutineers (scrutateurs) who could read and write ( JORF Lois et Décrets, 2 February 1852, Article 28; Charnay 1965: 638). E ­ ither the president of the voting office or scrutineers ­were asked to open each ballot and read out each name in a loud voice (Charnay 1965: 620). Scrutineers ­were not allowed to read the observations written by voters on ballots a­ fter the name of each candidate, and such remarks could also not be mentioned on the final reports that summarized the results of a race. If scrutineers found one ballot that was hidden inside another ballot, both ballots ­were voided. However, in cases in which both ballots ­were absolutely identical, one of the votes was counted while the other ballot had to be annexed to the election report (Pierre 1919: 149).

C u r bi n g E lec to r a l Fr au d 193

In practice, ­these requirements to keep voting public ­were often ­violated. Mayors took advantage of the fact that voters ­were largely unaware of the provisions of the electoral law and issued o ­ rders to restrict public access at the time ballots ­were counted. For example, during the 1878 election, the mayor of Roaix posted a sign at the entry of the voting place that prohibited voters from remaining in the voting place e­ ither before or a­ fter the vote. This decision was justified as a mea­sure that “ensured the freedom of the vote” ( JORF ChamberDeb, 22 October 1878). Instances where mayors expelled opposition voters from the polling place are also reported (Pilenco 1930: 257). During the 1906 election, in a commune in Tarn, the president of the voting bureau ­limited the presence of voters at the time ballots ­were counted, calling such supervision an act of “spying” ( JORF ChamberDeb 1906, 2115). In Nîmes during the 1914 election, the mayor expelled even the incumbent candidate from the voting place. In a commune in Arriège in 1910, the mayor expelled the municipal councillors who ­were part of the electoral bureau from the voting place and accused them of tampering with the urn ( JORF ChamberDeb, 9 June 1910). In a locality in Guadaloupe, the mayor granted special permits to some voters and allowed them to participate in the counting of ballots but restricted the participation of o ­ thers (ibid., 17 June 1910). In a ruling on electoral irregularities, the Conseil d’État noted that the most common strategy used to tamper with votes was to count ballots ­behind closed doors or only in front of a l­imited group of supporters (Conseil d’État 1909: 1257). The requirement to read out the names on the ballots in a loud voice was also circumvented in practice. A report of elections held in 1876 in Sartène recounted that the mayor hid in a dark corner of the voting place to open ballots. Taking advantage of the darkness, the mayor read out dif­ fer­ent names than ­those inscribed on the ballots ( JORF ChamberDeb, 22 November 1876). During subsequent elections, mayors diversified the strategies used to obfuscate the counting of ballots, including playing loud ­music at the time the names of the ballots ­were read (Marcault 1910). Despite the strict regulations in the French electoral law, electoral fraud continued throughout the elections of the Third Republic. Voting station presidents modified the outcome of the “popu­lar ­will” of voters in their constituencies using a variety of strategies. Some mayors chose to ignore the ballots found in the urn altogether, fabricating a dif­fer­ent po­liti­cal majority. This outright neglect of the votes cast was more common in the colonies. In the Indes Françaises in 1907, the mayor drafted an electoral report without emptying the urns ( JORF ChamberDeb, 4 March 1907). To placate

194 C H A P TER 6

his opponents and to avoid an outcry, the mayor assigned a fictitious number of votes to the candidate from the opposition party (ibid.). Alexandre Pilenco, an astute observer of French electoral practices, suggests that such strategies can explain the high volatility across elections in vote shares received by candidates competing in the colonies: “The consequence of such manoeuvers is that the majorities recorded in the colonial reports are always overwhelming. In 1893, Pierre Alype obtained over 30,000 votes in the Indes Françaises; in the following elections he lost the sympathy of ­those who controlled the urn and obtained only nine votes. During the 1902 election, only two ballots ­were credited to him” (1930: 280). However, ignoring all ballots found in the urn was an excessively blatant manipulation that risked immediate punishment. A less con­spic­u­ous form of electoral fraud involved adding or subtracting individual ballots and replacing opponents’ ballots with ­those of preferred candidates (Conseil d’État 1905: 684). In Olmiccia, a commune in Sartène (Corsica), Marc Caïtucoli, a Republican candidate, received 412 votes, despite the fact that only 108 voters w ­ ere registered and only 49 persons voted ( JORF ChamberDeb, 7 June 1910). In the colonies, the practice of adding ballots to the urn a­ fter voting ended was colloquially referred to as “garnishing the urn” (garnir l’urne). In Guadaloupe, mayors did not count individual ballots but packets containing multiple ballots, which ­were known as “maman cochon.” During the 1910 election, the president of the voting place added five hundred additional ballots for a candidate to the urn only a few minutes before the urn was emptied. As a result, “thirty votes cast in his ­favor of a candidate ­were counted as five hundred” (ibid., 17 June 1910). Alternatively, presidents of voting place could destroy ballots. During the 1881 election in a commune in Quimperlé (Finistère), the mayor de­cided to burn the ballots as soon as the urns w ­ ere emptied but before official results w ­ ere announced (ibid., 16 November 1881). In Lyon’s sixth district during the 1910 election, a bag containing one hundred ballots dis­appeared during the counting of the votes (ibid., 2 June 1910). A police report from Vieux-­Fort, a locality in Guadaloupe, described the disappearance of the ballots as follows: “We lighted matchsticks and searched for the urn, but we could not find it, we could only see a heap of fragments of wood and paper. We asked ­whether Mr. Dubureau, the president of the election bureau, had returned to his place, and ­were told that he ­hadn’t. His brother-­in-­law, Olivier Edmée, informed us: I have seen the president escape through the win­dow, he passed the woods ­behind the city hall and nobody followed him” (ibid., 17 June 1910). Gilbert Dumas, an in­de­pen­dent legislator in the Chamber, commented on the

C u r bi n g E lec to r a l Fr au d 195

incident, “In Guadaloupe, it is not necessary to have voters; it is enough to have ballots. In light of the most recent facts, I recognize that it is not even necessary to have urns” (ibid.). Fi­nally, the announcement of the results created additional opportunities for electoral fraud. The electoral law required presidents of the electoral bureau to announce the results of an election as soon as the ballots ­were counted. In practice, this provision was sometimes v­ iolated. As a strategy of insurance, mayors acting as presidents of voting places delayed the announcement of the results in tight races. By refusing to acknowledge the winner, t­ hese mayors expected an intervention by the government that could shift the balance of power among the candidates. In some instances, the public announcements of the winner took place several days ­after the urns ­were closed ( JORF ChamberDeb, 12 February 1881). Consider the incidents reported in Haute Garonne during the 1906 election. ­Here, the difference between the winner and the runner-up was less than one hundred votes. The mayor of Saint André, a locality with 135 registered voters, refused to open the urns and count the ballots, fearing that his would be the pivotal locality that could change the outcome of the race. Instead of counting the ballots, the mayor sealed the urn and sent it to the general prosecutor, delegating the risky decision to determine the outcome of the race (ibid., 22 June 1906). In Molifao, a locality in Corsica, the mayor refused to announce the results in his locality before hearing the results from the district. His hope was to retain discretion to modify the results in case of a discrepancy between the outcome in his locality and the district-­level outcome (ibid., 4 July 1898). Such strategies of delay ­were not uncommon. An election report of 1914 characterized the dilemma faced by many mayors as follows: “Mayors waited for the moment when other results are communicated, to be able to modify by several dozens of votes their election report, if the general situation required such a radical intervention” (ibid., 11 June 1914). As ­these examples demonstrate, penal sanctions had not entirely eliminated electoral fraud. To establish additional protections against fraud, French reformers embraced the reforms ­adopted in Belgium and Britain. Referring to the Belgian reforms, Alphonse Gourd, a Republican legislator, noted that the “introduction of this policy had fulfilled all the hopes of its supporters, pacifying po­liti­cal conflicts” ( JORF ChamberDeb, 30 May 1907). The French parliament began to consider proposals recommending candidate representatives beginning in the early 1890s. In 1894, Jules Guesde, a Socialist deputy from Roubaix (Nord), submitted an amendment recommending the introduction of candidate representatives in each electoral bureau. Guesde’s

196 C H A P TER 6

proposal drew extensively on the Belgian legislation, also borrowing the term “witness” (témoin) to refer to candidate representatives ( JORF ChamberDoc 1894, no. 336). The draft bill for electoral reform submitted by Félix Defontaine in 1895 also included recommendations for candidate representatives, in addition to ballot envelopes and isolating spaces. In Defontaine’s proposal, candidate representatives ­were to monitor elections and “rec­ord all irregularities in the minutes” ( JORF ChamberDoc 1895, no. 1373). A third proposal originating with Gabriel Dev­ille, a Socialist deputy, in 1898 read as follows. In the bureau of each college, a representative of each candidate should be allowed. This representative should be admitted on the basis of a letter from the candidate; he should have only a consultative role, but could mention his observations in the electoral report and sign ­these. ( JORF ChamberDeb, 1 April 1898) During the early deliberations of ­these reforms, Socialist deputies w ­ ere the strongest advocates of ­these proposals. The Commission of Universal Suffrage of the Chamber of Deputies discussed proposals to limit fraud in one of its earliest meetings, held on 6 November 1901 (Archives Nationales, C 5651). At the time, Emile Goussot, a Socialist representative from Saint-­Denis, and Louis Jourdain, a Radical from Somme, recommended authorizing candidate representatives to supervise the counting of ballots. However, the majority in the commission objected to t­ hese proposals, considering them burdensome and difficult to implement. Consequently, the commission rejected ­these ideas at this stage. The Chamber of Deputies discussed recommendations for candidate representatives for the first time in December 1901 as part of its deliberations on the law assuring the freedom and secrecy of the vote and the sincerity of electoral operations. At the time, Socialist deputies introduced an amendment recommending the introduction of candidate representatives. Ernest Ferroul, a Socialist representative from Narbonne, recalled the numerous irregularities in the counting of ballots that had taken place in his own district during the 1898 election. ­Here, Ferroul recalled, ballots ­were counted in packs of hundreds at a time and presidents of voting places “stuffed bags carry­ing the ballots with ballots with the name of my adversary” ( JORF ChamberDeb, 16 December 1901). “You w ­ ill not obtain any useful result,” Ferroul argued, “if you content yourself with the isolating space: it is imperative that the electoral act is supervised continuously by a representative of a party or of a list of deputies. This is the price to pay to ensure the sincerity of

C u r bi n g E lec to r a l Fr au d 197

the vote” (ibid.). Ferroul invoked Belgium’s successful experience: “The only way to prevent fraud committed by the members of the electoral bureau was to give each candidate the right to have one witness observe all the electoral operations; that’s the way it is done in Belgium” (ibid.). Conservative and Monarchist legislators also supported the introduction of candidate representatives, regarding it as a means to guarantee the sincerity of electoral operations. Michel de Pomereau, a deputy of the Conservative Action Libérale, argued that this amendment was “the only truly serious” addition to the legislation. In this proposed amendment to the law, Pomereau demanded that candidate representatives be given a broader set of prerogatives. Invoking the example of legislation that had been enacted in Canada, Pomereau requested that candidate representatives be allowed to “take all desired mea­sures advancing the interests of their principal,” including the right to exercise “police power” in the voting place ( JORF ChamberDeb, 16 December 1901). The Republican del­e­ga­tion in the Chamber was deeply divided in its assessment of the desirability of ­these reforms. In many French communes and even in larger cities, mayors often served as voting bureau presidents. Legislators’ preferences about the desirability of reforms ­were informed largely by expectations of having a co-­partisan mayor. Legislators representing départements with strong Republican sympathies strongly opposed proposals to introduce candidate representatives in elections. While benefiting from a relative resource advantage, ­these legislators favored an outcome that maintained a high level of mayoral discretion in the determination of election results. By contrast, legislators who faced higher uncertainty about the identity of the mayor in the district ­were more likely to support candidate representatives. During the early stages of deliberations, Republicans w ­ ere extremely vocal in expressing their opposition to the introduction of candidate representatives. In opposing ­these reforms, Republicans depicted candidate representatives as a “gratuitous insult of all French mayors” ( JORF ChamberDeb, 18 December 1901). Republican legislators’ most common objection was that candidate representatives ­were likely to cause disorder and turmoil in the voting place. In a meeting with the Commission of Universal Suffrage, Pierre Waldeck-­Rousseau, who was prime minister at the time, strongly opposed the proposal to introduce candidate representatives (Archives Nationales C 5651, 16 February 1902). Waldeck-­Rousseau argued that allowing candidates to station their representatives in the vicinity of urns amounted to the “organ­ization of a b ­ attle at the voting place” (ibid.). During

198 C H A P TER 6

subsequent deliberations, other Republican deputies took up the theme of disorder, tumult, and potential vio­lence at the voting place. In a Chamber discussion that took place in 1907, Alexandre Bérard, a Republican from Ain, worried that ­these representatives would engage in a variety of delaying tactics, which would make it very difficult to count ballots. The provision was nothing but a “guarantee for disorder” (garantie du gâchis) ( JORF ChamberDeb, 27 May 1907). To offset the possibility of disorder at the voting place, Republicans demanded that the authority of voting place presidents be increased so that they could expel representatives who engaged in disruption (ibid.). Louis d’Iriart d’Etchepare expressed concern that the proposed legislation allowed candidates to bring in observers from outside the locality: “In my own region, it would be extremely objectionable for delegates to be elected outside the commune and to come and start a conflict with the mayor. Given the spirit of particularism that exists in the communes, brawls would start in all voting places in less than a quarter of an hour” ( JORF ChamberDeb, 21 May 1907). In objecting to t­ hese proposals, Republicans invoked differences in electoral competition between Belgium and France. The success of the Belgian electoral legislation was predicated on the small number of po­liti­cal parties. By contrast, the existence of multiple parties in France could lead to situations of conflict resembling the “tower of Babel in the voting place” ( JORF ChamberDeb, 27 October 1904). Prominent Republican legislators on the Commission of Universal Suffrage embraced this par­tic­u­lar objection as the main justification for the rejection of the proposal. Joseph Ruau, the rapporteur of the law in 1904, argued that the Belgian model was inadequate and risky for France, creating opportunities for electoral ambush (guet-­apens electoral) (ibid.). An additional objection raised by Republican legislators was that the provisions about the candidate representatives ­were incompletely integrated with the existing legislation. It was unclear ­whether existing electoral regulations that governed the organ­ization of national elections, such as the 1852 law, prohibited the presence of candidate representatives. Alexandre Bérard argued that “in a majority of French communes, a large number of voters are huddled b ­ ehind the voting t­ able, observing how the ballots are counted.” Given this situation, the additional provision was no longer necessary ( JORF ChamberDeb, 30 May 1907). Another objection was that the proposals did not clearly define the means by which candidate representatives could verify the results of an election. It was unclear, for example, ­whether candidate representatives retained police power at elections.

C u r bi n g E lec to r a l Fr au d 199

The newly gained responsibilities of ­these representatives w ­ ere also not harmonized with the electoral responsibilities of the other actors, such as presidents of voting places. Importantly, however, not all Republicans w ­ ere opposed to the proposal to introduce candidate representatives. Some Republican legislators understood that t­ hese electoral reforms provided impor­tant po­liti­cal advantage in regions of France where they could not count on the presence of co-­partisan mayors in the voting place. Th ­ ese included primarily regions of western France, where hostility to Republican ideals persisted several de­cades ­after the fall of the Second Empire. In such regions, the presence of candidate representatives could constrain oppositional mayors—in most cases supporting Conservative or Monarchist candidates—­from falsifying the outcomes of elections. Representing Sables d’Olone, a district in the Vendée, Joseph Chailley was one of the few Republican legislators elected in western France. He was elected in 1906, ­after experiencing narrow defeats in previous attempts to gain this seat. Recalling the destruction of ballots witnessed recurrently in previous elections, Chailley endorsed ­these proposals for reform ( JORF ChamberDeb, 27 May 1907). Gustave de Kerguézec—­ who was elected on a Republican platform in Côtes-­du-­Nord in 1906—­also advocated for the introduction of candidate representatives. Disagreeing with other Republicans who opposed this mea­sure, Kerguézec urged his co-­partisans to support candidate representatives who could “safeguard the sincerity of the vote” (ibid.). Radical Socialist legislators had been pivotal in the electoral co­ali­tions supporting the other reforms discussed in this book: t­ hose having to do with curbing vote-­buying, limiting governmental pressure, and protecting voter secrecy. However, Radical Socialists w ­ ere much more ambivalent about reforms allowing candidate representatives at the voting place. Although prominent Radical politicians such as Félix Defontaine sponsored one of the earliest legislative proposals for candidate representatives, most Radical Socialists did not share his enthusiasm for this mea­sure. Urbain Sénac, a Radical Socialist from Tarn-­et-­Garonne, referred to candidate representatives as “masters of intrigues” ( JORF ChamberDeb, 27 May 1907). Commenting on this intraparty disagreement, Sénac noted that “we are not unified in our Radical and Radical Socialist party” (ibid.). Legislators from other parties also noted this division and the stark contrast in the position of Radical Socialists t­ oward candidate representatives as compared to other electoral reforms. Commenting on this contrast, Joseph Lasies, a Republican from Gers, noted, “On other occasions, this electoral reform

200 C H A P TER 6

concerning the sincerity of the vote was one of the main princi­ples of the Radical party. T ­ oday, what do we see? Only the Radical party opposes t­ hese reforms” (ibid.). Considerations about relative resource advantages explain some of this intraparty disagreement. Legislators elected in départements of Radical strongholds opposed candidate representatives. By contrast, the new provision was extremely attractive to Radicals competing in fragmented (and implicitly more competitive) districts where the presence of a co-­partisan mayor as president of the polling station could not be taken for granted. An additional worry about this mea­sure was that it was more advantageous to wealthier candidates, the candidatures d’argent, while requiring other candidates “to increase their electoral expenses” ( JORF ChamberDeb, 27 May 1902). Addressing Monarchist legislators in the Chamber, Urbain Sénac noted, “Gentlemen on the right, you are sufficiently rich to pay hired representatives. ­These representatives often cause energetical protests and voice demands to void the election. In our ranks, we cannot do this. As a result, we suffer severe consequences” (ibid., 30 May 1907). Sénac argued that it was very likely that agents of candidates who had offered voters eight hundred or one thousand francs prior to the vote would be the same persons appointed as representatives supervising the counting of ballots. Alexandre Bérard expressed a similar concern: “Once ­these delegates ­will be introduced in all communes, the influential owner of a manor [châtelain] w ­ ill appoint as representative his employees, the employer ­will appoint his foreman” (ibid.). What explains this difference in Radicals’ support for the introduction of candidate representatives versus the other reforms examined in this study? In chapter 2, I explained individual-­level preferences for electoral reforms as a function of three f­ actors: relative resource advantage, economic costs, and electoral costs. One difference, discussed above, concerned resources. Radicals feared that the introduction of candidate representatives could benefit wealthier candidates and that reforms could bring a resource disadvantage. At the same time, considerations about electoral costs w ­ ere absent in the case of reforms to limit fraud. Given that fraud occurred ­after voting had ended, candidates who engaged in this irregularity incurred no electoral costs. The electoral costs associated with dif­fer­ent illicit strategies can account for the variations in the positions of Radicals across ­these reform dimensions. As electoral costs ­were minimal in the case of fraud, we do not find the co­ali­tional realignment encountered in other cases. The debates in the Senate revealed similar cleavage lines about the desirability of reforms as in the Chamber. Conservative senators supported

C u r bi n g E lec to r a l Fr au d 201

proposals to introduce candidate representatives. In support of the proposal, Dominique Delahaye, a Royalist from Maine-­et-­Loire, vividly described the predicament of non-­Republican candidates competing in many districts in southern France. ­Here, Republican mayors did not allow candidate representatives of other parties to be pre­sent in the voting place both during the time of vote and at the time ballots w ­ ere counted ( JORF SenateDeb, 17 February 1910). By contrast, both Republicans and Radicals w ­ ere split over the introduction of this mea­sure. Eugène Lintilhac, a Radical from Cantal and the rapporteur of the commission of the Senate, strongly endorsed the proposal to introduce candidate representatives. Alexandre Bérard, a Republican, restated the familiar argument that the proposal would ­favor wealthy candidates: “In ­every village where the landowner is not mayor, he ­will be pre­sent in the voting place. The new article grants official responsibilities to persons who have an excessive influence. . . . ​­There is only one category of candidates who ­will be able to afford candidate representatives in all communes: the wealthy” (ibid., 18 February 1910). Lucien Cornet, a Radical senator from Yonne who was also opposed to the proposal, reiterated and expanded on Waldeck-­Rousseau’s ironical remark: “The presence of candidate representatives ­will not lead to a ­battle in the voting place, but it ­will lead to an open combat” (ibid.). A po­liti­cal compromise on the proposal to introduce candidate representatives was forged only in 1908. A positive vote in the Senate on 20 December 1907 set in motion additional discussion in the Chamber of Deputies. In a meeting of the Commission of Universal Suffrage held on 16 January 1908, the representative of the government expressed the support of the administration for candidate representatives. “The government is de­c ided to succeed,” the government representative stated on this occasion, urging the commission to reintroduce the proposal in the Chamber (Archives Nationales C 5651, 16 January 1908). Following the endorsement by the commission, the Chamber of Deputies voted in support of candidate representatives on 12 June 1908. This remained a decisive vote that was not overturned by the Chamber in subsequent discussion, leading to the introduction of candidate representatives in 1913. I now turn to an analy­sis of this vote to unpack how considerations about po­liti­cal resources, on the one hand, and economic and electoral constraints, on the other hand, impact legislators’ support for electoral reforms. We need to construct dif­fer­ent variables to mea­sure the relative resource advantage of dif­fer­ent candidates. The “production” of fraud required access to the mayors of a commune. Mayors w ­ ere often appointed as presidents

202 C H A P TER 6

of the voting station and had the opportunity to modify the outcomes of elections ­after the votes ­were counted. Candidates in districts with co-­ partisan mayors enjoyed a relative resource advantage, while candidates lacking mayoral support ­were disadvantaged and ­were, thus, more likely to support electoral reforms. Mea­sur­ing control is more difficult in this case owing to the unavailability of electoral results for local elections. Mayors w ­ ere not elected directly but by representatives of the city councils of their localities. I use information collected by historians of the Third Republic about the partisan strongholds across dif­fer­ent départements to proxy po­liti­cal strength in dif­fer­ent départements. Odile Rudelle (1982) provides information about the distribution of partisan strength for Republicans and Conservatives, while Jacques Kayser (1962) provides this information for Radicals. As is well known to students of the Third Republic, areas of Conservative electoral strength include départements in western France, such as Finistère, Morbihan, Calvados, and Vendée. A second area of Conservative strength was found in southwest France and included departments such as Gers, Landes, and Aveyron (Rudelle 1982: 127). The region of Radical electoral strength stretched from central to southeastern France and included départements such as Cher, Nièvre, Loir, Drôme, Var, Aude, Gard, and Hérault, among ­others (Kayser 1962: appendix 1). Fi­nally, areas of Republican strength ­were located primarily in central and eastern France, from Dordogne to Savoie and from Eure-­et-­Loire to Doubs (Rudelle 1982: 135). Figure 6.2 provides information on the coding of the dif­fer­ent areas of partisan strength. I used this information to create proxies for the partisan resource advantage. I created the variable Stronghold that takes the value 1 if the legislator was elected in a département of partisan strength. Legislators elected in such partisan strongholds are expected to enjoy a relative resource advantage and resist electoral reforms. By contrast, legislators elected in regions that are not their partisan stronghold are likely to oppose the adoption of ­these reforms. Figure 6.3 pre­sents the results. Similar to the analy­sis presented in previous chapters, I estimate the vote using Firth logistic regression. In panel (a), I examine w ­ hether access to resources allowing candidates to engage in irregularities correlates with support for reform. I used several variables to proxy the relative resource advantage. The first variable (Stronghold) takes the value 1 if the candidate was elected in a partisan stronghold. I also proxy the resource advantage using two individual attributes of legislators. Using information on the prior experience of legislators, I coded ­whether

C u r bi n g E lec to r a l Fr au d 203

Departments of Radical strength Departments of Conservative strength Departments of Republican strength Partisan strongholds in the Third Republic. Sources: Rudelle 1982; Kayser 1962: appendix 1. FIGURE 6.2.

the politician was a mayor or owner prior to their election to the Assemblée Nationale. ­These individual-­level resources have dif­fer­ent implications for legislators’ views about the desirability of electoral reforms. Legislators with a previous mayoral experience find the status quo attractive and oppose the introduction of candidate representatives. By contrast, legislators with a previous experience as ­owners are more supportive of electoral reforms. As the above discussion has pointed out, private economic resources ­were likely to create a resource advantage in appointing candidate representatives

204 C H A P TER 6

(a) Individual resources

(b) Economic constraints Economic constraints

Individual resources

Agricultural employment Urbanization

Partisan electoral stronghold

Landholding inequality

Mayor

–1.5

–1

–0.5

0

0.5

Population (ln) –10

(c) Electoral constraints

–5

0

5

10

(d) Full model

Individual resources Partisan electoral stronghold Mayor Owner

Electoral constraints

Partisan resources Nonpartisan Socialist Radical Republican Right Electoral constraints Programmatic promises

Programmatic promises

–0.5

0

0.5

1

Economic constraints Agricultural employment Urbanization Landholding inequality Population (ln) –5

0

5

10

The determinants of support for reforms to limit fraud. Notes: Results pre­sent correlates of support for the introduction of candidate representatives during a vote on 12 June 1908. All models are estimated using Firth logistic regression. See the appendix for the sources used to construct the variables used in the analy­sis. FIGURE 6.3.

at the voting place. By contrast, the partisan electoral stronghold does not predict support for electoral reforms. Thus, relative resource asymmetries ­matter in explaining legislators’ preferences ­toward the proposal to introduce candidate representatives. The models reported in panels (b) and (c) examine the importance of economic and electoral constraints in influencing legislative demand for t­ hese reforms. Economic developments—­such as urbanization and industrialization—­ increase the scale of electoral operations and may constrain the ability of presidents of voting places to engage in fraud. If this logic is at work, one expects a positive relationship between this variable and support for electoral reforms. We find some support for ­these modernization hypotheses. As the results in panel (b) indicate, we find a rural-­urban cleavage among legislators about the desirability of electoral reforms. Politicians elected in urban areas found the status quo less attractive than did politicians elected in rural districts. Other mea­sures of economic development, such as urbanization

C u r bi n g E lec to r a l Fr au d 205

and landholding in­equality, do not explain cleavages over electoral reforms. Panel (c) illustrates the importance of electoral costs in explaining demand for reform. Similar to previous analyses, I proxy electoral costs using a mea­ sure of programmatic promises made by the candidates. As discussed in chapter 2, I expect electoral costs to play no role in shaping demand for reforms to limit fraud. As fraud occurs a­ fter ballots are cast, the possibility that voters w ­ ill sanction candidates who engaged in this form of malfeasance is low. It follows that concerns about electoral costs are unlikely to motivate politicians’ demand for electoral reforms. The results reported in panel (c) support ­these hypotheses. While electoral costs w ­ ere significant in predicting support for other electoral reforms examined in this study, they remain unimportant in explaining demand for the reforms to limit fraud. Panel (d) pre­sents the results of the full model. In broad strokes, we can characterize the co­ali­tion supporting the adoption of electoral reforms to limit fraud as a co­ali­tion of extremes against the center. Representatives of po­liti­cal parties who found themselves at the two extremes of the French po­liti­cal spectrum—­Socialists and Conservative politicians on the right—­ spearheaded demand for the introduction of candidate representatives. As discussed above, the centrist Republicans supported t­ hese reforms belatedly (beginning in 1907). By contrast, the large Radical del­e­ga­tion in the Assemblée found itself divided about the desirability of electoral reforms. While Radicals w ­ ere a pivotal legislative group supporting the other electoral reforms examined in this study, Radicals ­were less supportive of reforms to limit fraud. This difference can be attributed to two ­factors. First, considerations about electoral costs ­were highly salient in the case of irregularities that ­were vis­ib ­ le and that could be sanctioned by voters. Such considerations about electoral costs w ­ ere absent in this case. Second, Radicals’ reluctance to endorse ­these electoral reforms may have been motivated by their fear that reforms privileged wealthier candidates. In this case, the Radical opposition can be explained as a concern that the reform would establish a ­future resource asymmetry. Conclusion

In this chapter, I have examined the adoption of reforms to limit the incidence of fraud in first-­wave democracies. In the first-­wave democracies included in this study, fraud was not a significant irregularity in nineteenth-­ century elections. For the continental Eu­ro­pean countries, the low incidence of fraud can be attributed to two f­ actors. The Napoleonic Code a­ dopted

206 C H A P TER 6

in 1810 that diffused across Eu­ro­pean countries introduced strong punishments for fraud. At the same time, the weak development of centralized party organ­izations l­imited the ability of party leaders to coordinate the production of fraud across localities and districts. Instances of fraud w ­ ere mainly localized events that resulted from the decentralized initiatives of presidents of voting stations. The most significant reforms to limit fraud ­adopted by first-­wave democracies sought to give candidate representatives the right to supervise the counting of ballots. Both Britain and Belgium a­ dopted such electoral reforms swiftly and with ­limited disagreement. ­Here, the low resource asymmetry among parties facilitated the formation of a compromise in support of ­these reforms. Imperial Germany modified its electoral jurisprudence ­toward an ac­cep­tance of candidate representatives. In contrast to Belgium and Britain, France enacted electoral reforms to limit the incidence of fraud only with ­great difficulty. ­Here, the composition of the co­ali­tion supporting reforms differed from that of other reforms examined in the book. The adoption of candidate representatives reflected the influence of a co­ali­tion of the extremes against the po­liti­cal center. Demand for reform originated with Conservative and Socialist legislators, who ­were joined at a ­later stage by resource-­constrained legislators from the Republican and Radical parties.

7 Conclusion

First-­wave democracies experienced a strong conflict between the demo­ cratic ideals promised by their electoral laws and po­liti­cal real­ity. While electoral laws upheld lofty princi­ples, such as the sovereignty of the general ­will, electoral practices presented a dif­fer­ent real­ity, characterized by extensive bribery, intimidation, coercion, and fraud. Writing in 1899, André Daniel, a historian of the Third Republic, remarked on the general disaffection of French voters with the demo­cratic proj­ect: “The Republic is, undoubtedly, sufficiently strong, sufficiently uncontested, but she does not inspire any longer this enthusiasm and she does not light this sacred fire one encountered previously” (1899: 4). And in neighboring Germany, reformist politicians noted that “nothing is more contemptible and in conflict with the constitution than existing efforts to influence the dependent [abhängige] voters and to make it difficult for them to cast their ­free vote” (StBer DtRt, 15 May 1895). Recent democracies have experienced the same disappointment with the demo­cratic proj­ect observed by André Daniel a ­century ago. The transition from authoritarian regimes ushered in an initial period of euphoria and optimism about the role of electoral pro­cesses in guaranteeing the existence of an accountable po­liti­cal class through po­liti­cal turnover. Over time, this initial euphoria gave way to deep pessimism. This pessimism is grounded in the fact that po­liti­cal developments fail to deliver many expected po­liti­cal outcomes, such as the presence of high-­quality representatives. At the same time, the dissatisfaction with demo­cratic institutions reflects a growing frustration 207

208 C H A P TER 7

with the messiness and imperfection of electoral practices. Electoral rules in place in many recent democracies have proved ill-­equipped to limit the incidence of dif­fer­ent forms of electoral malfeasance. Such irregularities include electoral intimidation and coercion, making access to policy benefits dependent po­liti­cal support, vote-­buying, the violation of voter secrecy, and so on (Mares and Young 2016, 2018b). Such electoral imperfections contribute to the per­sis­tence of venal and unaccountable po­liti­cal elites. Demo­cratic theorists such as Pierre Rosanvallon have described the gap in expectations between normative demo­cratic ideals and everyday po­liti­ cal real­ity as the “perpetual disenchantment of democracies” (2018: 216). Rosanvallon argues that democracies w ­ ill inevitably correct the m ­ istakes that have given rise to such disappointed expectations. But demo­cratic theorists provide us with no explanation of the ways in which this “perpetual disenchantment” can be resolved. Acknowledging this limitation of normative theorists, Rosanvallon noted that “the moment when universal suffrage represents an irreversible acquis is hard to locate temporally” (1992: 587). How do democracies close the gap between the lofty promises of their founding elections and the grim subsequent electoral real­ity? Who are the most impor­tant actors pushing for ­these reforms? How do ­these actors succeed in overcoming the opposition of politicians who benefit electorally from the status quo? In this book, I have presented an account of this demo­cratic transition by examining the adoption of reforms to limit electoral malfeasance in first-­ wave democracies. Charles Benoist, the president of the Commission of Universal Suffrage of the French Chamber of Deputies during the 1890s, referred to t­ hese changes in electoral rules as the effort to “pull universal suffrage from its rudimentary, inorganic state” ( JORF ChamberDeb, 12 June 1908). In providing an explanation of this transition, I have characterized the motivations of legislators who supported ­these electoral reforms and described the composition of the co­ali­tions that endorsed changes in electoral legislation. I have documented that the composition of the co­ali­ tions that supported the reduction of electoral malfeasance varied across reform dimensions and across countries. This concluding chapter has two objectives. I review the implications of this study for existing theories of democ­ratization and for efforts to reduce electoral malfeasance in recent democracies. The explanation presented in this study differs both from redistributivist theories of democ­ratization and from recent explanations proposed by Ansell and Samuels that stress the importance of conflict over property rights among rural and urban elites

Co n c lu s i o n 209

(Acemoglu and Robinson 2000, 2006; Ansell and Samuels 2014). This study stresses the importance of electoral motivations in explaining the willingness of elites to embrace reforms. As the motivations of legislators to support reforms differ across policy areas, I also document that the composition of the electoral co­ali­tions supporting reforms varies across reform dimensions. Existing theories of democ­ratization that highlight the importance of elites’ economic motivations cannot account for this variation in support for reforms. In the concluding part of the chapter, I discuss some of the implications of the historical experience of first-­wave democracies for the adoption of reforms to limit electoral malfeasance in third-­wave democracies. I highlight the importance of messages and informational campaigns that increase legislators’ electoral costs associated with illicit strategies as a potentially impor­tant intervention for the formation of po­liti­cal majorities supporting reforms. Implications for Theories of Democ­ratization

This study has provided an empirical and a theoretical contribution to the lit­er­a­ture on democ­ratization. In empirical terms, I documented that electoral integrity reforms have been an impor­tant but overlooked dimension of the pro­cess of democ­ratization. In Britain, such reforms ­were ­adopted alongside the expansion of suffrage. In other cases, such as Third Republic France and Imperial Germany, reforms to protect electoral integrity ­were put in place ­after the introduction of manhood suffrage. How did countries adopt electoral reforms protecting voter integrity? What ­were the most impor­tant motivations of elites who embraced such electoral reforms? And how did encompassing electoral majorities supporting po­liti­cal change emerge? In explaining legislative demand for electoral reforms, this study has highlighted the importance of access to resources and concerns about rising economic or electoral costs associated with vari­ous forms of electoral malfeasance. A large part of the empirical material presented in this study documented the formation of encompassing co­ali­tions that supported electoral reforms in countries with dominant po­liti­cal parties, such as Third Republic France and Imperial Germany. In ­these po­liti­cal environments, access to the resources of the state that can be politicized during elections was unevenly distributed across parties. Legislators affiliated with the governing majority ­were able to politicize employees of the state administration and deploy them during elections. By contrast, legislators affiliated with opposition parties lacked such access to state resources.

210 C H A P TER 7

In the absence of exogenous changes, politicians who possess this electoral advantage are unlikely to give up their access to state resources. In such environments, elite splits of the incumbent co­ali­tion served as the catalyst for reform. Th ­ ese splits increased po­liti­cal fragmentation and created incentives for programmatic differentiation. In electoral systems with runoffs, elite splits also increased the importance of coordination during the second round. Altogether, t­ hese po­liti­cal developments increased the associated electoral costs incurred by politicians who enjoyed a relative resource advantage and made the use dif­fer­ent forms of electoral malfeasance po­liti­cally costly. Legislators enjoying a relative resource advantage but who encountered electoral costs ­were po­liti­cally pivotal during the negotiation of reforms. If electoral costs following elite splits ­were low, such legislators w ­ ere likely to continue to block the adoption of electoral reforms. By contrast, if electoral costs ­were high, resource-­endowed legislators could forgo their resource advantage and endorse changes to electoral rules that sanctioned malfeasance. In this second scenario, a po­liti­cal majority supporting reforms emerged even in countries with hegemonic po­liti­cal parties. The dynamics of reform w ­ ere dif­fer­ent in countries with duopolistic competition. In ­these countries, electoral reforms enjoyed the support of a bipartisan co­ali­tion that included legislators of the two significant parties. The enfranchisement of new voters or economic developments that increase the costs of illicit strategies created disruptions to the status quo. In ­these cases, legislators of both parties endorsed an electoral reform as an effort to maintain the duopolistic competition. The adoption of candidate representatives in Belgium and Britain illustrates this po­liti­cal logic. H ­ ere, both parties viewed such reforms as necessary safeguards that imposed constraints on each other. The adoption of the Corrupt and Illegal Practices Act in Britain is also consistent with this po­liti­cal logic. In this case, changing economic conditions affected Liberals more than Conservatives. While Liberals ­under Gladstone initiated proposals to adopt stronger limits on campaign expenditures, the reform met with bipartisan support. The explanation proposed in this study contrasts with dominant theoretical perspectives on democ­ratization that have stressed the importance of economic considerations in shaping elites’ decisions to support demo­cratic reforms. Redistributivist theories of democ­ratization argue that elites’ incentives to adopt demo­cratic reforms reflect considerations about ­future levels of taxes (Acemoglu and Robinson 2000; Boix 2003). By contrast, Ansell and Samuels view demo­cratic transitions as resulting from an intra-­elite conflict

Co n c lu s i o n 211

between “a group that controls the state and other relatively wealthy groups that do not” (2014: 11). In this account, rising bourgeois elites introduce demo­cratic electoral reforms as a means to limit the power of autocratic elites to expropriate their income and assets (ibid., 7). Neither perspective can account for the composition of electoral co­ali­ tions supporting reforms documented in this study. Consider first the predictions of ­these theories about the position held by legislators connected to the interests of large landowners. Both the Acemoglu/Robinson and the Ansell/Samuels approaches predict that Monarchist legislators in France and Conservatives in Germany ­will oppose electoral reforms. The position of German Conservatives with regard to electoral reforms is consistent with the expectations of ­these theories. In Germany, Conservative legislators staunchly opposed electoral reforms throughout the period. However, ­these theories cannot accommodate the position held by other parties on the right during the vari­ous episodes of reform examined in this study. In France, Monarchists and other non-­Republican parties on the right opposed proposals to limit vote-­buying but endorsed efforts to limit the pression gouvernementale as well as the introduction of candidate representatives. The findings of this book do not support the prediction of the Ansell and Samuels framework that legislators affiliated with the rising bourgeoisie w ­ ill invariably support demo­cratic reforms. In Germany, the position of the National Liberals ­toward electoral reform varied over time. National Liberals’ opposition to legislation protecting ballot secrecy or limiting the use of state resources during elections during the last de­cades of the nineteenth c­ entury disconfirms the prediction of a pro-­democratic bourgeoisie by Ansell and Samuels. National Liberals’ support for the protection of voter secrecy and for constraints on the politicization of state resources occurred only during the elections prior to World War I and was triggered by the breakdown of the electoral alliance with Conservatives, the Kartellbündnis. In France, the Republican Party represented a mix of rural and industrial interests. However, this geo­graph­i­cal cleavage was not salient for many of the electoral reforms examined in the book and we do not find evidence of a strong defense of electoral integrity by Republican legislators in urban, eco­nom­ically developed districts. When it came to questions about the deployment of state resources during campaigns, Republicans elected in both rural and industrial districts opposed the adoption of electoral reforms. In chapter 1, I discussed one theoretical weakness of redistributivist accounts of democ­ratization: its long and untested causal chain between economic conditions, such as in­equality, and the adoption of demo­cratic

212 C H A P TER 7

reforms. Consider, for instance, the logical chain presupposed by ­these theories to explain the decisions of o ­ wners with immobile assets to support the expansion of suffrage. This logic presupposes that the new electoral majority ­will enact higher levels of taxes that out­going elites associated with the old regime ­will be unable to block. It also presupposes that ­these elites ­will be unable to change their ability to move assets a­ fter reforms. A third assumption is that landowning elites ­will be unable to change the collection or administration of taxes, in other words, that their re­sis­tance to taxation following the extension of suffrage w ­ ill be entirely in­effec­tive. While ­these assumptions may be more or less plausible, redistributivist theories provide no evidence that tests the intermediate steps of their explanation. In contrast to redistributivist theories, the account presented in this book does not require such a long causal chain to derive predictions for reforms. The salient ­factors that explain cleavages over electoral reforms involve the relative distribution of assets that may be politicized at elections and the economic and electoral costs associated with the use of illicit strategies. In short, the explanation proposed in this study is superior to alternative explanations b ­ ecause it can account for the existence of heterogeneous electoral co­ali­tions whose composition differed across reform dimensions without relying on a heavy baggage of untestable assumptions. Implications for Recent Democracies

Recent democracies also experience Rosanvallon’s “perpetual disenchantment.” One encounters similar imperfections in the organ­ization of elections as in first-­wave democracies prior to the adoption of reforms to protect electoral integrity. For example, throughout Africa, 48 ­percent of voters surveyed in thirty-­three countries during the fifth round of the Afrobarometer reported fearing vio­lence during elections, and 16 ­percent reported being offered money or goods in exchange for their vote (Mares and Young 2016). Across Latin Amer­i­ca, 15 ­percent of voters surveyed during the 2010 and 2012 rounds of the Amer­i­cas Barometer reported that they had been offered gifts and ­favors in exchange for their vote. Such illicit electoral exchanges are also common in Eastern Eu­ro­pean countries, such as Hungary, Romania, and Bulgaria, that are recent member states of the Eu­ro­pean Union (Mares, Muntean, and Petrova 2016; Mares and Young 2018a, 2018b). I conclude this study by discussing several implications of the reform experience in first-­wave democracies for con­temporary democracies. The first concerns the design of legislation to protect electoral integrity. To

Co n c lu s i o n 213

avoid some of the m ­ istakes and false starts of first-­wave democracies, recent democracies need to design legislation that minimizes opportunities for substitution across dif­fer­ent illicit strategies. This is often accomplished through the imposition of calibrated sanctions that take into account opportunities for substitution across illicit strategies both on election day and prior to elections. A second implication of this study concerns the construction of po­liti­ cal majorities supporting electoral reforms. Actors who promote reforms to limit electoral malfeasance need to acknowledge that the electoral co­ali­tions supporting reforms are highly heterogeneous and that they w ­ ill inevitably bring together politicians with dif­fer­ent resource endowments and dif­fer­ent constraints. At the same time, actors who promote reforms need to exploit politicians’ concerns about the electoral costs associated with vari­ous forms of malfeasance. To this end, campaigns that disseminate information about the normative undesirability of vari­ous forms of electoral malpractice play an impor­tant role in increasing the salience of electoral costs for politicians. In ­doing so, such campaigns may cement legislative majorities supporting electoral reforms. ­Viable electoral reforms protecting electoral integrity need to minimize opportunities for substitution across dif­fer­ent forms of electoral malfeasance. A situation that disproportionately sanctions one form of po­liti­cal malfeasance creates incentives for politicians and their brokers to substitute the strategies that are punished less harshly. The case of the nineteenth-­ century German electoral law illustrates this dynamic. The law imposed a much harsher punishment on vote-­buying, leaving electoral strategies that politicized private brokers as unsanctioned. While the former exchanges ­were offenses that fell ­under the penal law and ­were sanctioned with prison and high fines, the latter strategies ­were unpunished ­under German electoral law (Mares 2015). While this punishment structure reduced the incidence of vote-­buying in German elections, it contributed to an unpre­ce­dented reliance by candidates on employers as brokers. In the words of Margaret Lavinia Anderson, a historian of the period, “election intimidation by bread lords” was perceived as “universal, unvarnished and enduring” (2000: 227). Didi Kuo and Jan Teorell (2017) illustrate a dif­fer­ent form of substitution in the United States. ­Here, the adoption of ballot secrecy reduced bribery but increased voter registration fraud. We find similar developments at work in recent democracies. In their effort to limit illicit electoral practices, con­temporary anticorruption agencies often intervene by imposing unequal punishments on dif­fer­ent forms of electoral malfeasance. Consider the recent initiatives of the Romanian

214 C H A P TER 7

anticorruption agency, one of the strongest institutions fighting electoral malfeasance. Prosecutors in this country have targeted vote-­buying with high levels of intensity and imposed severe sanctions on politicians who offer money or bribes to voters. Such heavy-­handed targeting of this illicit strategy contributed to a reduction of vote-­buying offers in recent elections. Yet such intervention created opportunities for candidates to target voters using private brokers or to use clientelistic strategies that politicized state resources (Mares and Young 2018a, 2018b). As this example illustrates, reform strategies that impose extremely stringent sanctions on one strategy and leave other strategies unpunished are in­effec­tive in reducing electoral malfeasance. Such interventions are likely to change the mix of strategies and may often result in a net increase in the levels of electoral malfeasance. To minimize such opportunities for substitution, electoral reformers need to avoid the excessive sanctioning of par­tic­u­lar forms of malfeasance. Singling out one strategy as abhorrent may allow parties to signal their commitment to an anticorruption cause. However, such strategies may be counterproductive in the long run, as they create opportunities for substitution. Surprisingly, many actors and organ­izations promoting electoral reform overlook this ­simple recommendation. Election monitors and other organ­izations that promote electoral integrity often focus on identifying vote-­buying or fraud associated with vote tallying but leave additional forms of malfeasance unpunished. Efforts to detect election fraud focus mostly on election day developments but fail to consider the variety of strategies deployed by candidates between elections, such as granting access to policy benefits in exchange for support (Mares and Young 2016). Thus, actors who pursue electoral reforms need to maintain a very expansive menu of pos­si­ble electoral irregularities that occur both during campaigns and at the time of voting in order to minimize opportunities for substitution. This study has shown that the electoral co­ali­tions supporting reforms to limit vari­ous forms of electoral malfeasance are highly heterogeneous. Politicians join legislative majorities for electoral reforms depending on their access to resources that can be monetized at elections but also on the electoral and economic costs of d ­ oing so. Legislators who can easily access the bureaucratic apparatus of the state are likely to oppose reforms limiting the use of state resources but are likely to endorse electoral reforms to limit strategies that politicize private resources, such as vote-­buying reforms. By contrast, legislators who can access private resources with g­ reat ease are

Co n c lu s i o n 215

likely to oppose reforms to limit vote-­buying but support reforms that constrain the politicization of state resources. This finding is likely applicable to recent democracies. Efforts to construct po­liti­cal majorities need to take into account t­ hese differential motivations of legislators to accept changes in electoral rules. Po­liti­cal majorities need to be constructed in a piecemeal fashion, recognizing that the mix of resources and constraints faced by legislators varies across reform dimension. Such heterogeneous co­ali­tions may adopt electoral reforms that are po­liti­cally more robust. The resulting electoral reforms are less likely to meet with “ex post” violations. The po­liti­cal pro­cess by which first-­wave democracies ­adopted reforms to limit malfeasance has implications for recent democracies. One of the lessons of this study is that elite splits create opportunities for co­ali­tional realignment. Following elite splits, the pivotal political actors are resource-­ endowed politicians who encounter salient electoral costs. A po­liti­cal majority supporting reforms emerges only if ­these legislators are willing to forgo their resource advantage in the face of concern with electoral costs. Yet while elite splits open up opportunities for co­ali­tional realignment, they do not always result in the formation of a legislative majority supporting reforms. In ­these uncertain moments following elite splits, resource-­endowed legislators may choose to discount such electoral costs and block the adoption of further reforms. This finding of my study has implications for the formation of electoral majorities supporting reforms. International organ­izations or actors who promote electoral reforms to reduce malfeasance need to enlist the support of ­these pivotal groups of legislators facing high electoral costs. In their efforts to create a po­liti­cal majority supporting reforms, such organ­izations need to harness and amplify considerations about electoral costs that are salient for ­these legislators. As such, po­liti­cal interventions and campaigns that increase public awareness of the normative undesirability of illicit strategies may induce legislators who encounter electoral costs to join a majority supporting reforms. To this end, activities that disseminate information about illicit electoral practices that are sanctioned by voters—­such as coercion, bribery, and vote-­buying—­may contribute to the formation of an electoral majority supporting reforms by convincing resource-­endowed legislators to forgo their initial resource advantage. To conclude, the dissemination of information about electoral malfeasance is a highly promising area of intervention by domestic nongovernmental organ­izations or by international organ­izations engaged in democracy

216 C H A P TER 7

promotion. If the content of ­these interventions resonates with legislators’ concerns about rising electoral costs, such efforts may be well powered to induce legislators to reconsider the attractiveness of vari­ous forms of electoral malfeasance. Such interventions may increase the size of legislative co­ali­tions favoring electoral reforms. Both nongovernmental organ­izations combating electoral irregularities and international organ­izations promoting demo­cratic assistance should invest in such informational campaigns.

APPENDIX

Sources for Quantitative Historical Analy­sis This appendix details the sources used for the quantitative analy­s is of this book. ­TABLE A.1

France Variable

Description

Source

Partisanship of legislator

Variables code affiliation of dif­fer­ent legislators with the dif­fer­ent po­liti­cal parties of the Third Republic

­Angles D’Auriac n.d.; Jolly 1960

Occupation of legislator

Variables mea­sure w ­ hether candidate is employer or owner, and ­whether he was a mayor at the time of the election

Chambre des députés (vari­ous years)

Programmatic promises

Variable mea­sures programmatic commitment expressed by candidate during campaign

JORF 1899, 1907

Landholding in­equality

In­equality in the size of agricultural holdings in a county as mea­sured by the 1896 agricultural census

Statistique Agricole de la France 1897

Voting population

Number of voters in district

Chambre des députés (vari­ous years)

Rural employment

Percentage of population in a department employed in agriculture based on 1901 census

Statistique Générale de la France 1903

Urbanization

Takes value 1 if urban population is over 100,000 based on 1901 census

Statistique Générale de la France 1903

0,2:57.96,3:23.34

Germany Variable

Description

Source

Partisanship

Partisanship of the legislator elected to the Reichstag

ICPSR 1984

Occupation of legislator

Takes value 1 if legislator is landowner or industrial employer

Reichstagsprotokolle.de

Runoff

Takes value 1 if legislator is elected in runoff

ICPSR 1984 Continued on next page

217

218 App e n d ix

­TABLE A.1

(continued) Germany

Variable

Description

Source

Landholding in­equality

In­equality in the size of agricultural holdings in a county as mea­sured by the 1896 agricultural census

Ziblatt 2009

Population

Population in a district

ICPSR 1984

Economic development

Percentage of population employed in manufacturing

Mares 2015

Urbanization

Percentage of population in localities over 100,000

Mares 2015

0,2:57.96,3:23.34

­Great Britain Variable

Description

Source

Partisanship of legislator

Partisanship of legislator

Craig 1977

Wealth of candidate

Value of estate of candidate at time of death

National Probate Calendar (vari­ous years)

Programmatic orientation

Support for social policy legislation during the period

Divisions of the House of Commons 1880

Margin of victory

Difference between the vote share of a sitting MP and the vote share of the first losing candidate

Craig 1977

Landholding in­equality

In­equality in the size of agricultural holdings in a county as mea­sured by the 1896 agricultural census

Board of Agriculture 1896

Wealth of constituency

Per capita income in a county at the time of the 1881 census

Hechter 1976

Manufacturing employment

Proportion of a population in a county employed in manufacturing at the time of the 1881 census

Hechter 1976

Population growth

Growth of the population of a county at the time of the 1881 census

Hechter 1976

Size of voting population

Number of voters on the electoral register of a constituency at the time of the election

Craig 1977

REFERENCES

Archival Sources BELGIUM Chambre des représentants de Belgique. Annales Parlementaires. Chambre des représentants de Belgique. Documents parlementaires. Chambre des représentants de Belgique. T ­ able décennale et alphabétique des pièces imprimées par ordre de la Chambre des représentants, 1891–1892 à 1900–1901; précédée d’une liste des membres qui ont siegé pendant cette période parlementaire. Hymans, Louis. Histoire parlementaire de la Belgique de 1831 à 1880. Brussels. Lauwaert, R., R. Depoortere, and L. Vandeweyer. Inventaris van dossiers betreffende de wetgevende verkiezingen 1884–1961. Archives de l’État en Belgique. Pagnol, A. M., C. Vleeschouwers, and C. Bourmanne. 2000. Ministère de l’Intérieur. Affaires Électorales. Brussels: Archives Générales du Royaume.

FRANCE Archives Nationales, Série C Assemblées C 5470 Législation électorale: liberté et secrét du vote. C 5574 Législation électorale: repression de la corruption, vote obligatoire, liberté de ré­union, suppression de la candidature officielle. C 5651 Législation électorale: revision, affichage, liberté de vote. C 5652 Repression de la corruption, liberté de vote, election des Sénateurs et députés, abrogation des candidatures multiples, repre­sen­ta­tion proportionelle. C 7305 Législation électorale: secret du vote. C 7375 Commission du suffrage universel. C 7447 Commission du suffrage universel. C 7448 Commission du suffrage universel. C 7449 Commission du suffrage universel: élections. C 7721 Commission du suffrage universel.

GERMANY Bundesarchiv Berlin Lichtenberg R 1501/114455 Sitzung des Königlichen Staatsministeriums am 17. Januar 1903

219

220 R e fe r e n c e s

R 1501/114456 Zur Sicherung des Wahlgeheimnisses Zentrum und Wahlrecht

R 1501/114460 Zur Einführung einheitlicher Wahlurnen Reichswahlreform

R 1501/114470 Das Norwegische Wahlgesetz vom 24. Februar 1900 Anschrift des Ministerium des Innern betreffend den Antrag Hompesch für eine Strafbestimmung gegen die Verletzung des Wahlgeheimnisses Reichsamt des Innern. 8. Januar 1909. Einführung selbsttätiger Wahlurnen Wahlurnen. Aufzeichnung von Gallenkamp

R 1501/114475 Aufzeichnung über die am 27. Juni 1910 stattgehabte kommissarische Beratung betreffend Abänderungen des Wahlreglements Minister des Innern. Auf die Schreiben vom 5. Marz und vom 21. Dezember betreffend die Einführung selbsttätiger Wahlurnen Eingabe wegen amtlicher Wahlurnen Wahlurnen (Dr. Gallenkamp und Dr. Schulze) Urteile über Wahlfählschungen Äußerung des Ministers des Innern zu der Frage der Beschaffung einheitlicher Wahlurnen für die Reichstagswahl Votum des Staatsministers Dr. Delbruck zur Frage der Einführung von Normativbestimmungen über die Beschaffenheit der Wahlurnen bei den Reichstagswahlen

R 1501/114476 Votum zur Frage der Einführung von Normativbestimmungen über die Beschaffenheit der Wahlurnen bei den Reichstagswahlen Entwurf einer Bekanntmachung über die Abänderung des Wahlreglements vom 28. Mai 1870

R 1501/114696 Die Kruppsche Arbeiterschaft und die Reichstagswahl in Essen Die verschwiegene Wahlurne

GEHEIMES STAATSARCHIV BERLIN. PREUSSISCHER BESITZ I HA Rep 90A, no. 112 Kommissarische Beratung betreffend Abänderung des Wahlreglements Sitzung des Königlichen Staatsministerium vom 21. Oktober 1910 Beratung der Einführung einheitlicher Wahlurnen Normativbestimmungen für die Wahlurnen bei der Reichstagswahl

I HA Rep 90A, no. 307

R e fe r e n c e s 221

Legislative and Statistical Resources Anglès D’Auriac, Dominique. N.d. “Les politiciens de la Troisième République.” Unpublished manuscript. Paris. Board of Agriculture. 1896. Agricultural Returns for ­Great Britain with Agricultural Statistics for the United Kingdom, British Possessions and Foreign Countries. London: HMSO. Chambre des députés. Vari­ous years. Tableaux des élections à la Chambre des députés adressés aux archives de la Chambre des députés (1876–1942). Paris: Imprimerie de la Chambre. Code Électoral de la Belgique présentant la codification de toutes les lois éléctorales. Brussels: Tarlier. Conseil d’État. Vari­ous years. Recueil des arrêts du Conseil d’État. Paris: Librairie du Recueil Général des lois et des décrets. Deutsches Reichsgesetzblatt. 1871. Gesetz betreffend die Verfassung des Deutschen Reichs 16:63–85. Hechter, K. 1976. UK County Data, 1851–1966. UK Data Archive, SN: 430. ICPSR. 1984. German Reichstag Election Data, 1871–1914. Interuniversity Consortium for Po­liti­cal and Social Research. JORF. 1899. Professions de foi et engagements electoraux de 1898. Paris: Imprimerie Nationale. JORF. 1907. Rapport fait au nom de la commission chargée de réunir, de publier et d’analyser les programmes et engagements électoraux par M. Camille Fouquet. Paris: Imprimerie Nationale. JORF. Lois et Décrets. 1914. Loi reprimant les actes de corruption dans les operations électorales. Paris: Imprimerie du Journal Officiel de la République Française. Recueil Lebon. Vari­ous years. Recueil des arrêts du Conseil d’État statuant au contentieux des decisions du tribunal des conflits et de la cour des comptes. Paris: Librairie Du Recueil Général des Lois et des arrêts. Statistique Agricole de la France (Algérie et Colonies). 1897. Résultats généraux de l’enquête décennale de 1892. Nancy: Berger-­Levrault. Statistique Générale de la France. 1903. Résultats statistiques du recensement général de la population effectué le 24 mars 1901. Paris: Imprimerie Nationale.

Secondary Sources Acemoglu, Daron, and James Robinson. 2000. “Why Did the West Extend the Suffrage? Growth, In­equality and Democracy in Historical Perspective.” Quarterly Journal of Economics 115: 1167–99. —­—­—. 2006. Economic Origins of Dictatorship and Democracy. New York: Cambridge University Press. Agulhon, M., et al. 1986. Les maires en France du Consulat à nos jours. Paris: Publications de la Sorbonne. Aidt, Toke, and Raphael Franck. 2015. “Democ­ratization u ­ nder the Threat of Revolution: Evidence from the ­Great Reform Act of 1832.” Econometrica 83(2): 505–47. Albertus, Michael, and Victor Menaldo. 2018. Authoritarianism and the Elite Origins of Democracy. New York: Cambridge University Press. Alevy, D. 1930. La fin des notables. Paris: Grasset. Alexander, M. 2000. Die Freikonservative Partei 1890–1918. Düsseldorf: Droste. Anderson, Margaret Lavinia. 2000. Practicing Democracy: Elections and Po­liti­cal Culture in Imperial Germany. Prince­ton: Prince­ton University Press. Anglès D’Auriac, Dominique. 2015. “Les députats de la Chambre des Députés pendant la IIIème République.” Unpublished database. Paris.

222 R e fe r e n c e s

Ansell, Ben, and David Samuels. 2014. In­equality and Democ­ratization: An Elite-­Competition Approach. New York: Cambridge University Press. Ardanaz, Martin, and Isabela Mares. 2014. “­Labor Shortages, Rural In­equality and Democ­ ratization.” Comparative Po­liti­cal Studies 47(12): 1637–69. Arsenscheck, Robert. 2003. Der Kampf um die Wahlfreiheit im Kaiserreich: Zur parlamentarischen Wahlprüfung und politischen Realität der Reichstagswahlen 1871–1914. Düsseldorf: Droste. Aubert, J., et al. 1978. Les préfets en France 1800–1940. Geneva: Droz. Autran, Aimé. 1958. Un siècle de politique en Vaucluse: Les élections, les événements, les partis et les candidats de 1848 à 1958. Avignon: Rullière. Bateman, David. 2018. Disenfranchising Democracy: Constructing the Electorate in the United States, the United Kingdom and France. New York: Cambridge University Press. Bateman, David, and Dawn Langan Teele. 2019. “A Developmental Approach to Historical Causal Inference.” Public Choice 185: 253–79. Becker, Jean-­Jacques, and Gilles Candar. 2004. Histoire des Gauches en France. L’Héritage du XIXème siècle. Paris: La Découverte. Bendix, Reinhard, and Stein Rokkan. 1968. “The Extension of National Citizenship to the Lower Classes: A Comparative Perspective.” In State and Society: A Reader in Comparative Po­liti­cal Sociology, ed. Reinhard Bendix and Stein Rokkan, 233–56. Berkeley: University of California Press. Bergounjoux, Alain. 1992. “Socialisme et République avant 1914.” In Serge Berstein and Odile Rudelle, Le modèle républicain, 117–28. Paris: Presses Universitaires de France. Bernard, Paul. 1992. Le préfet de la République; le chêne et l’olivier. Paris: Economica. Berstein, Serge. 1992. “La Politique Sociale des Républicains.” In Serge Bern­stein and Odile Rudelle, Le modèle républicain, 159–72. Paris: Presses Universitaires de France. —­—­—. 2006. “Les radicaux.” In Jean-­Jacques Becker and Gilles Candar, Histoire des gauches en France, 2:10–26. Paris: Découverte. Berstein, Serge, and Odile Rudelle. 1992. Le modèle républicain. Paris: Presses Universitaires de France. Birch, Sarah. 2011. Electoral Malpractice. New York: Oxford University Press. —­—­—. 2020. Electoral Vio­lence, Corruption and Po­liti­cal Order. Prince­ton: Prince­ton University Press. Blake, Robert. 1985. The Conservative Party from Peel to Churchill. London: Eyre and Spottiswoode. Block, Maurice. 1880. Dictionnaire Général de la Politique. Paris: Lorenz. Boix, Carles. 2003. Democracy and Re­distribution. New York: Cambridge University Press. Boix, Carles, and Susan Stokes. 2003. “Endogenous Democ­ratization.” World Politics 55(4): 517–49. Bonnefous, Eduard. 1973. Histoire Politique de la Troisième République. Paris: Presses Universitaires de France. Buchstein, Hubertus. 2000. Öffentliche und geheime Stimmabgabe: Eine wahlrechtshistorische und ideengeschichtliche Studie. Baden-­Baden: Nomos. Candar, Gilles. 2004. “La gauche en République (1871–1899).” In Jean-­Jacques Becker and Gilles Candar, Histoire des gauches en France, 2:113–31. Paris: La Découverte. Caramani, Daniele. 2000. Elections in Western Eu­rope since 1815: Electoral Results by Constituencies. London: Macmillan. Challeton, Félix. 1891. Cent ans d’élections: Histoire électorale et parlementaire de la France de 1789 à 1890. Paris: Saumaitre. Chandernagor, André. 1993. Les maires en France: Histoire et sociologie d’une fonction. Paris: Fayard. Charnay, Jean-­Paul. 1965. Les scrutins politiques en France de 1815 à 1962, contestations et annulations. Paris: Armand Colin.

R e fe r e n c e s 223

Chauveau, Adolphe, and Faustin Hélié. 1852. Théorie du Code Pénal. Paris: Cosse. Imprimerie and Librairie Générale de Jurisprudence. Claris, Edmond, ed. 1904. “Jean Jaurès.” Discours Parlementaires. Paris: E. Cornély. Code Pénal. 1815. Code Pénal, Edition conforme à l’édition originale du Bulletin des Lois; précédé de l’exposé des motifs par les orateurs du Conseil d’Etat, sur chacune des lois qui composent ce code. Paris: Garnéry. Cox, Gary. 1987. Making Votes Count: Strategic Coordination in the World’s Electoral Systems. New York: Cambridge University Press. Craig, Fred. 1977. British Parliamentary Election Results, 1832–1885. 2nd ed. London: Macmillan. Dahl, Robert. 1971. Polyarchy: Participation and Opposition. New Haven: Yale University Press. Daniel, André. Vari­ous years. L’Année politique. Paris: Charpentier. Day, Samuel. 1894. Election cases in 1892 and 1893: Being a collection of the points of law and practice arising out of the parliamentary election petitions in ­those years, together with reports of the judgments. London: Stevens. de Kéguelin, Georges. 1904. Crimes et délits électoraux. Paris: Rousseau. Delalande, Nicolas. 2011. Les batailles de l’impôt: Consentement et résistances de 1789 à nos jours. Paris: Seuil. Deloye, Yves. 1993. “La civilité électorale: Vote et forclusion de la vio­lence en France.” Cultures et Conflits 9–10: 65–75. —­—­—. 1994. École et citoyenneté: L’individualisme républicain de J. Ferry à Vichy. Controverses. Paris: Presses de la Fondation Nationale des Sciences Politiques. De Meaux, Camille. 1905. Souvenirs politiques 1871–1877. Paris: Plon. De Meeus, Adrien. 1962. History of the Belgians. New York: Praeger. Divisions of the House of Commons. 1880. London: H.M. Stationery Office. Ducatel, Paul. 1975. Histoire de la IIIème République à travers l’imagerie populaire et la presse satirique. Paris: Grassin. Duplantier, Raymond. 1901. De la pratique du suffrage: Essai d’histoire éléctorale et de législation comparée. Poitiers: A. Masson. Evans, Eric. 2000. Parliamentary Reform, 1770–1918. Harlow: Longman. Ewald, François. 1992. “La Politique Sociale des Opportunistes.” In Serge Bern­stein and Odile Rudelle, Le modèle républicain, 189–208. Paris: Presses Universitaires de France. Fairbairn, Brett. 1997. Democracy in the Undemo­cratic State: The German Reichstag Elections of 1898 and 1903. Toronto: University of Toronto Press. Fenske, Hans. 1981. Der liberale Südwesten: Freiheitliche und demokratische Traditionen in Baden und Württemberg 1790–1933. Stuttgart: Kohlhammer. Ferry, Jules. 1893. Discours et opinions de Jules Ferry. Paris: A. Colin. Feuchtwanger, Edgar. 1968. Disraeli, Democracy and the Tory Party. Oxford: Clarendon Press. Flaubert, Gustave. 2004. Sentimental Education. 2 vols. London: Penguin Books. Freudenthal, Berthold. 1895. Die Wahlbestechung. Breslau: Schletter. Fromes, Armand. 1908. Code Électoral Belge ou Recueil complet de toutes les lois actuellement en vigueur en Belgique relatives aux élections législatives, provinciales et communales. Brussels: Veuve Ferdinand Larcier. Furet, François, and Mona Ozouf. 1993. Le siècle de l’avènement républicain. Paris: Gallimard. Gagel, Walter. 1958. Die Wahlrechtsfrage in der Geschichte der deutschen liberalen Parteien, 1848– 1918. Düsseldorf: Droste Verlag. Gambetta, Léon. 1881. Discours et plaidoyers politiques de M. Gambetta. 11 vols. Paris: Charpentier. Garrard, John. 2002. Demo­cratisation in Britain: Elites, Civil Society and Reform since 1800. Houndmills: Palgrave. Garrigou. A. 1988. “Le secret de l’isoloir.” Actes de la recherche en sciences sociales 71–72: 22–45.

224 R e fe r e n c e s

George, Jacelyne. 1989. Histoire des maires de 1789 à 1940. Paris: Plon. Gladstone, William Ewart. 1877. “The County Franchise and Mr. Lowe Thereon.” Nineteenth ­Century 2(9): 527–65. Grün, Alphonse. 1842. Jurisprudence parlementaire. Paris: Charles Hingray. Guiral, Pierre, and Guy Thuillier. 1980. La vie quotidienne des députés français de 1871 à 1914. Paris: Hachette. Gurwitsch, Girsch. 1910. Der strafrechtliche Schutz des Wahlrechts. Leipzig: Robert Poske. Gwyn, William. 1962. Democracy and the Cost of Politics in Britain. London: Athlone. Haggard, Stephan, and Robert Kaufman. 2016. Dictators and Demo­crats: Masses, Elites and Regime Change. Prince­ton: Prince­ton University Press. Hanham, Harold John. 1959. Elections and Party Management: Politics in the Time of Disraeli and Gladstone. London: Longmans. Hanson, Stephen. 2010. “The Founding of the French Third Republic.” Comparative Po­liti­cal Studies 43(8/9): 1023–58. Hatschek, Julius. 1915. Das Parlamentsrecht des Deutschen Reiches. Berlin: Göschen. —­—­—. 1920. Kommentar zum Wahlgesetz und zur Wahlordnung im deutschen Kaiserreich. Berlin: de Gruyter. Hicken, Allen. 2011. “Clientelism.” Annual Review of Po­liti­cal Science 14: 289–310. Hill, George. 1879. Electoral History of the Borough of Lambeth since Its Enfranchisement in 1832. London: Stanford and Co. Houle, Christian. 2009. “In­equality and Democracy: Why In­equality Harms Consolidation but Does Not Affect Democ­ratization.” World Politics 61: 589–622. Huard, Raymond. 1982. Le movement Républicain en Bas-­Languedoc. Paris: Presses de la Fondation Nationale des Sciences Politiques. —­—­—. 1992. Le suffrage universel en France 1848–1946. Paris: Aubier. —­—­—. 1996. La naissance du parti politique en France. Paris: Presses de Sciences Po. Huber, Ernst Rudolf. 1978. Deutsche Verfassungeschichte seit 1789. Vol. 2. Stuttgart: Kohlhammer. Hymans, Louis. 1880. Histoire parlementaire de la Belgique de 1831 à 1880. Brussels: Bruylant-​Christophe. Ihl, Olivier. 1998. “Les Fraudes Éléctorales: Problèmes de definition juridique et politique.” In How Did They Become Voters? The History of Franchise in Modern Eu­ro­pean Repre­sen­ta­tion, ed. Raffele Romanelli, 77–110. The Hague: Kluwer. Jolly, Jean. 1960. Dictionnaire des parlementaires français: Notices biographiques sur les ministres, députés et sénateurs français de 1889 à 1940. Paris: Presses Universitaires de France. Kalyvas, Stathis. 1996. The Rise of Christian Democracy in Eu­rope. Ithaca: Cornell University Press. Kam, Christopher. 2017. “The Secret Ballot and the Market for Votes in Nineteenth ­Century British Elections.” Comparative Po­liti­cal Studies 50(5): 594–635. Kasara, Kimuli, and Isabela Mares. 2016. “Unfinished Business: The Democ­ratization of Electoral Practices in Britain and Germany.” Comparative Po­liti­cal Studies 50(3): 536–64. Kayser, Jacques. 1962. Les grandes batailles du radicalisme. Paris: Marcel Rivière and Compagnie. Kehr, Helen. 1965. Der Primat der Innenpolitik. Berlin: de Gruyter. Keyssar, Alexander. 2009. The Right to Vote: The Contested History of Democracy in the United States. New York: Basic Books. Kitschelt, Herbert. 2000. “Linkages between Citizens and Politicians in Demo­cratic Polities.” Comparative Po­liti­cal Studies 33(6): 845–79. Kitschelt, Herbert, and Steven Wilkinson. 2007. Patrons, Clients, and Policies: Patterns of Demo­ cratic Accountability and Po­liti­cal Competition. New York: Cambridge University Press. Kinzer, Bruce. 1982. The Ballot Question in Nineteenth-­Century En­glish Politics. New York: Garland. Klein, Thomas. 2003. Gültig—­ungültig: Die Wahlprüfungsverfahren des Deutschen Reichstages. Marburg: Elwert.

R e fe r e n c e s 225

Kramon, Eric. 2016. “Electoral Handouts as Information: Explaining Unmonitored Vote-­Buying.” World Politics 68(3): 454–98. Kühne, Thomas. 1994. Dreiklassenwahlrecht und Wahlkultur in Preussen 1867–1914: Landtagswahlen zwischen korporativer Tradition und politischem Massenmarkt. Düsseldorf: Droste. —­—­—. 1995. Handbuch der Wahlen zum preussichen Abgeordetenhaus 1867–1918. Düsseldorf: Droste. Kuo, Didi, and Jan Teorell. 2017. “Illicit Tactics as Substitutes: Election Fraud, Ballot Reform, and Contested Congressional Elections in the United States, 1860–1930.” Comparative Po­liti­cal Studies 50(5): 665–96. Law, Edward Lord Ellenborough. 1881. A Po­liti­cal Diary. 2 vols. London: Richard Bentley and Sons. Lawrence, Jon. 1998. Speaking for the ­People: Party, Language, and Popu­lar Politics in ­England, 1867–1914. Cambridge: Cambridge University Press. Lefranc, Pierre. 1868. Le Rastell Electoral. Paris: Le Chevalier. Legrand, Maxime. 1877. Doctrines des conservateurs en matière d’élections. Paris: Imprimerie de Jouaust. Lipset, Seymour Martin. 1959. “Some Social Requisites of Democracy: Economic Development and Po­liti­cal Legitimacy.” American Po­liti­cal Science Review 53: 69–115. —­—­—. 1994. “The Social Prerequisites of Democracy Revisited.” American So­cio­log­i­cal Review 59(2): 1–22. Lizzeri, Alessandro, and Nicola Persico. 2004. “Why Did the Elites Extend the Suffrage? Democracy and the Scope of Government, with an Application to Britain’s Age of Reform.” Quarterly Journal of Economics 119(2): 707–65. Llavador, Humbertus, and Robert Oxoby. 2005. “Partisan Competition, Growth and the Franchise.” Quarterly Journal of Economics 120(3): 1155–89. Logre, Bernard. 1986. Jozon Meunier: Récit d’une campagne en 1876 dans l’arrondissement de Meaux. Paris: Amatteis. Lowell, Lawrence A. 1908. The Government of ­England. 2 vols. New York: Macmillan. Lubenow, W. C. 1983. “Irish Home Rule and the G ­ reat Separation of the Liberal Party in 1886: The Dimensions of Parliamentary Liberalism.” Victorian Studies 26(2): 161–80. Luebbert, Gregory. 1991. Liberalism, Fascism, or Social Democracy: Social Classes and the Po­liti­cal Origins of Regimes in Interwar Eu­rope. New York: Oxford University Press. Madrid, Raúl. 2019a. “Opposition Parties and the Origin of Democracy in Latin Amer­i­ca.” Comparative Politics 51(2): 157–75. —­—­—. 2019b. “The Partisan Path to Democracy: Argentina in Comparative Perspective.” Comparative Po­liti­cal Studies 52(10): 1535–69. Magaloni, Beatriz. 2006. Voting for Autocracy: Hegemonic Party Survival and Its Demise in Mexico. New York: Cambridge University Press. Maier, Hans. 1992. L’Église et le démocratie: Une histoire de l’Eu­rope politique. Paris: Criterion. Marcault, Abbé Charles. 1910. L’art de tromper, d’intimider et de corrompre l’électeur. Paris: Blond. Mares, Isabela. 2015. From Open Secrets to Secret Voting: Demo­cratic Electoral Reforms and Voter Autonomy. New York: Cambridge University Press. —­—­—. 2018. “Menus of Electoral Irregularities: German National Elections in Comparative Perspective.” Geschichte und Gesellschaft 44(3): 399–415. Mares, Isabela, Aurelian Muntean, and Tsveta Petrova. 2016. “Economic Intimidation in Con­ temporary Elections: Experimental Evidence from Romania and Bulgaria.” Government and Opposition 53(3): 1–32. —­—­—. 2017. “Pressure, Favours, and Vote-­buying: Experimental Evidence from Romania and Bulgaria.” Eu­rope Asia Studies 69(6): 940–60.

226 R e fe r e n c e s

Mares, Isabela, and Giancarlo Visconti. 2020. “Voting for the Lesser Evil: Evidence from a Conjoint Experiment in Romania.” Po­liti­cal Science Research and Methods 8(2): 315–28. Mares, Isabela, and Lauren Young. 2016. “Buying, Stealing or Expropriating Votes.” Annual Review of Po­liti­cal Science 19: 267–88. —­—­—. 2018a. Conditionality and Coercion: Electoral Clientelism in Eastern Eu­rope. Oxford: Oxford University Press. —­—­—. 2018b. “The Core Voters’ Curse: Clientelistic Threats and Promises in Hungarian Elections.” Comparative Po­liti­cal Studies 51(11): 1441–71. —­—­—. 2019. “Va­ri­e­ties of Clientelism in Hungarian Elections.” Comparative Politics 51(3): 441–90. Mares, Isabela, and Boliang Zhu. 2015. “Economic and Po­liti­cal Determinants of Electoral Intimidation.” Comparative Politics 48(1): 1–24. Mayer, Georg. 1901. Das Parlamentarische Wahlrecht. Berlin: Hæring. Mayer, Max Ernst. 1906. “Verbrechen und Vergehen in Beziehung auf die Ausübung strafrechtlicher Rechte.” In Vergleichende Darstellung des deutschen und ausländischen Strafrechts: Vorarbeiten zur deutschen Strafrechtsreform, ed. Fritz von Calker et al., 257–347. Berlin: Liebmann. —­—­—. 1910. “Bekämpfung der Wahlumtriebe durch das Strafrecht.” Zeitschrift für Politik 3: 10–29. Mayeur, Jean Marie. 1973. Les débuts de la III-­ème République 1871–1898. Paris: Seuil. McLean, Iain. 2001. Rational Choice and British Politics: An Analy­sis of Rhe­toric and Manipulation from Peel to Blair. Oxford: Oxford University Press. Milatz, Alfred. 1967. “Reichstagswahlen und Mandatsverteilung 1871 bis 1918.” In Gesellschaft, Parlament und Regierung: Zur Geschichte des Parlamentarismus in Deutschland, ed. Gerhard Ritter, 207–23. Düsseldorf: Droste. Montesquieu, Charles. 2011. The Spirit of Laws. Holmes Beach: Gaunt. Moore, Barrington. 1966. Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of the Modern World. Boston: Beacon. Müller, K. 1963. “Politische Strömungen in den rechtsrheinischen Kreisen des Regierungsbezirkes Köln von 1879 bis 1900.” PhD diss., University of Bonn. National Probate Calendar. Vari­ous years. ­England and Wales, National Probate Calendar, Index of ­Wills and Administrations, 1858–1966. Ancestry​.­co​.­uk. Nichter, Simeon. 2008. “Vote-­buying or Turnout Buying? Machine Politics and the Secret Ballot.” American Po­liti­cal Science Review 102(1): 19–31. Nicolet, Claude. 1982. L’idée républicaine en France 1789–1924. Paris: Gallimard. Nipperdey, Thomas. 1961. Die Organisation der deutschen Parteien vor 1918. Düsseldorf: Droste. Nypels, S. G. 1868. Le droit penal belge interpreté principalement au point de vue de la pratique. Brussels: Bruylant. O’Donnell, Guillermo, and Philippe Schmitter. 1986. Transitions from Authoritarian Rule: Tentative Conclusions about Uncertain Democracies. Baltimore: Johns Hopkins University Press. Offerlé, Michel. 1998a. “La Nationalisation de la citoyenneté en France à la fin du XIXème siècle.” In How Did They Become Voters? The History of Franchise in Modern Eu­ro­pean Repre­sen­ta­tion, ed. Raffele Romanelli, 53–77. The Hague: Kluwer. —­—­—. 1988b. “Le nombre de vois: Électeurs, partis et électorat socialistes à la fin du XIXème siècle en France.” Actes de la recherché en sciences sociales 72(1): 5–21. O’Leary, Cornelius. 1962. The Elimination of Corrupt Practices in British Elections, 1868–1911. Oxford: Oxford University Press. O’Malley, Edward Loughlin, and Henry Hardcastle. 1893. Reports of the Decisions of the Judges for the Trial of Election Petitions in G ­ reat Britain and Ireland, Pursuant to the Parliamentary Elections Act. 5 vols. London: Stevens and Haynes. Ormond, Thomas. 1994. Richterwürde und Regierungstreue: Dienstrecht, politische Betätigung und Disziplinierung der Richter in Preussen, Baden und Hessen 1866–1918. Frankfurt: Klosterman.

R e fe r e n c e s 227

Parry, Jonathan. 1993. The Rise and Fall of Liberal Government in Victorian Britain. New Haven: Yale University Press. Paul, Johann. 1987. Alfred Krupp und die Arbeiterbewegung. Düsseldorf: Schwann. Phélippeau, Eric. 1993. “La fabrication administrative des opinions politiques: Votes, declarations de candidats, et verdicts des préfets 1852–1914.” Revue Française de Science Politique, no. 4, 587–613. Pierre, Eugène. 1893. Traité de droit politique électoral et parlementaire. Paris: Loysel. —­—­—. 1919. Traité de droit politique, électoral et parlementaire. Paris: Librairie Imprimerie Réunies. Pilenco, Alexandre. 1928. Les Mœurs électorales: Régime censitaire. Paris: Editions du Monde Moderne. —­—­—. 1930. Les Mœurs du Suffrage Universel en France 1848–1928. Paris: Editions de la Revue Mondiale. Pinto-­Duschinsky, Michael. 1981. British Po­liti­cal Finance, 1830–1980. Washington, DC: American Enterprise Institute. Pourcher, Yves. 1987. Les maîtres de granit: Les notables de Lozère du XVIIIe siècle à nos jours. Paris: O. Orban. Prost, Antoine. 1974. Vocabulaire des proclamations électorales de 1881, 1885, 1889. Paris: Presses Universitaires de France. Przeworski, Adam. 1991. Democracy and the Market: Po­liti­cal and Economic Reforms in Eastern Eu­rope and Latin Amer­i­ca. New York: Oxford University Press. —­—­—. 2009. “Conquered or Granted? A History of Suffrage Extensions.” British Journal of Po­liti­ cal Science 39(2): 291–321. Przeworski, Adam, Michael Alvarez, Jose Cheibub, and Fernando Limongi. 2000. Democracy and Development. New York: Cambridge University Press. Przeworski, Adam, and Fernando Limongi. 1997. “Modernization: Theories and Facts.” World Politics 49(1): 155–83. Puhle, Hans-­Jürgen. 1967. “Agrarische Interessenpolitik und preussischer Konservatismus im wilhelminischen Reich 1893–1914.” In Das Kaiserliche Deutschland, ed. M. Stürmer, 340–77. Düsseldorf: Droste. Raikes, Henry. 1883. “The New Law of Elections.” National Review, 204–5. Randall, Vicki. 1988. Po­liti­cal Parties in the Third World. London: Sage. Rejewski, Harro Jürgen. 1973. Die Pflicht zur politischen Treue im preussischen Beamtenrecht 1850– 1918: Eine rechtshistorische Untersuchung anhand von Ministerialakten aus dem geheimen Staatsarchiv der Stiftung preussischer Kulturbesitz. Berlin: Duncker & Humblot. Rémond, René. 1954. La droite en France de 1815 à nos jours: Continuité et diversité d’une tradition politique. Paris: Aubier. Rendu, Ambroise. 1885. Code Électoral ou manuel pratique des élections municipales départamentales et politiques a l’usage des conseillers municipaux, des maires, des délégués sénatoriaux, des conseillers généraux et d’arrondissement, des deputés et des sénateurs. Paris: Durand and Pedone Lauriel. Retallack, James. 1988. Notables of the Right: The Conservative Party and Po­liti­cal Mobilization in Germany, 1876–1918. London: Unwin Hyman. —­—­—. 1990. “ ‘What Is to Be Done?’ The Red Specter, Franchise Questions, and the Crisis of Conservative Hegemony in Saxony, 1896–1909.” Central Eu­ro­pean History 23(4): 271–312. Rix, Kathryn. 2006. “Hidden Workers of the Party: The Professional Liberal Agents, 1885–1910.” Journal of Liberal History 52: 4–13. —­—­—. 2008. “ ‘The Elimination of Corrupt Practices in British Elections’? Reassessing the Impact of the 1883 Corrupt Practices Act.” En­glish Historical Review 123: 65–97.

228 R e fe r e n c e s

—­—­—. 2017. “The Second Reform Act and the Prob­lem of Electoral Corruption.” Parliamentary History 36(1): 64–81. Rix, Kathryn, and Philip Salmon. 2018. “Who Should Have the Vote? What Electoral Rights Did Britons Have in the ­Century before 1918?” History ­Today (August): 24–35. Rogers, Francis. 1880. Rogers on Elections. 3 vols. London: Stevens & Sons. Rogers, Francis, and Charles Willoughby. 1906. Parliamentary Elections and Petitions with Appendices of Statutes and Forms. 2 vols. London: Stevens and Sons. Romanelli, Raffele, ed. 1998. How Did They Become Voters? The History of Franchise in Modern Eu­ro­pean Repre­sen­ta­tion. The Hague: Kluwer. Rosanvallon, Pierre. 1992. Le sacré du citoyen. Paris: Gallimard. —­—­—. 1998. Le peuple introuvable: Histoire de la Représentation Démocratique en France. Paris: Gallimard. —­—­—. 2000. La Démocratie inachevée: Histoire de la souveraineté du peuple en France. Paris: Gallimard. —­—­—. 2018. Good Government: Democracy beyond Elections. Cambridge, MA: Harvard University Press. Rudelle, Odile. 1982. La République absolue 1870–1889. Paris: Publications de la Sorbonne. —­—­—. 1992. “De Jules Ferry à Raymond Poincaré, ou l’échec du constitutionalisme républicain.” In Serge Bern­stein and Odile Rudelle, Le modèle républicain. Paris: Presses Universitaires de France. Rueschemeyer, Dietrich, Evelyne Huber Stephens, and John Stephens. 1992. Cap­i­tal­ist Development and Democracy. New York: Cambridge University Press. Schattschneider, Elmer Eric. 1960. The Semi-­Sovereign P ­ eople: A Realist’s View of Democracy in Amer­i­ca. New York: Holt. Scheyven, C., and P. Holvoet. 1894. Code électoral belge expliqué d’après les travaux parlementaires et les circulaires ministérielles. Brussels: Bruylant. Seymour, Charles. 1915. Electoral Reform in E ­ ngland and Wales: The Development and Operation of the Parliamentary Franchise, 1832–1885. New Haven: Yale University Press. Seymour, Charles, and Donald Paige Frary. 1918. How the World Votes: The Story of Demo­cratic Development in Elections. Springfield, MA: Nichols. Shannon, Richard. 1996. The Age of Salisbury, 1881–1902: Unionism and Empire. New York: Longman. Siegfried, André. 1913. Tableau politique de la France de l’ouest sous la Troisième République. Paris: Colin. Siegfried, R. 1903. “Die verschwiegene Wahlurne.” Annalen des Deutschen Reichs für Gesetzgebung, Verwaltung und Volkswirtschaft 36. Simon, Klaus. 1969. Die württembergischen Demokraten; ihre Stellung und Arbeit im Parteien-­und Verfassungssystem in Württemberg und im Deutschen Reich 1890–1920. Stuttgart: Kohlhammer. Simpser, Alberto. 2013. Why Governments and Parties Manipulate Elections: Theory, Practice and Implications. Cambridge: Cambridge University Press. Sirinelli, Jean-­François, ed. 1994. Dictionnaire historique de la vie politique française of XXème siècle. Paris: Presses Universitaires de France. Sperber, Jonathan. 1997. The Kaiser’s Voters: Electors and Elections in Imperial Germany. New York: Cambridge University Press. Stasavage, David. 2020. The Decline and Rise of Democracy: A Global History from Antiquity to ­Today. Prince­ton: Prince­ton University Press. Stokes, Susan, et al. 2015. Brokers, Voters and Clientelism: The Puzzle of Distributive Politics. Cambridge: Cambridge University Press. Suval, Stanley. 1985. Electoral Politics in Wilhelmine Germany. Chapel Hill: University of North Carolina Press.

R e fe r e n c e s 229

Tanchoux, Philippe. 2004. Les procédures électorales en France de la fin de l’Ancien Régime à la première guerre mondiale. Paris: Comité des travaux historiques et scientifiques. Teele, Dawn. 2018. Forging the Franchise: The Po­liti­cal Origins of the W ­ omen’s Vote. Prince­ton: Prince­ton University Press. Tourneur, A. 1904. Pages d’histoire. Gambetta en 1869. Belle­ville et Marseille. Lettres et documents inédits. Lille: A. Devos. Uzé, Charles. 1896 [2012]. De la nullité en matière d’élections politiques. Paris: Hachette. Valelly, Richard. 2004. The Two Reconstructions: The Strug­gle for Black Enfranchisement. Chicago: University of Chicago Press. Villey, Edmond. 1900. Législation électorale comparée des principaux pays d’Eu­rope. Paris: A. Pedone. Voilliot, Christophe. 2005. La candidature officielle: Une pratique d’État de la Restauration à la Troisième République. Rennes: Presses Universitaires de Rennes. Weil, Georges-­Denis. 1895. Les Élections législatives depuis 1789: Histoire de la législation et des mœurs. Paris: Félix Alcan. Weitz-­Shapiro, Rebecca. 2012. “What Wins Votes: Why Some Politicians Opt Out of Clientelism.” American Po­liti­cal Science Review 56(3): 568–83. —­—­—. 2014. Curbing Clientelism in Argentina: Politics, Poverty and Policy. New York: Cambridge University Press. Weitz-­Shapiro, Rebecca, and Matthew Winters. 2013. “Lacking Information or Condoning Corruption: When Do Voters Support Corrupt Politicians?” Comparative Politics 45(4): 418–36. White, Dan S. 1976. The Splintered Party: National Liberalism in Hessen and the Reich. Cambridge, MA: Harvard University Press. Wilkinson, R. 1884. “New Bribery Act.” Nineteenth ­Century 15: 129–31. Williamson, Jeffrey. 2006. Globalization and the Poor Periphery before 1950. Cambridge, MA: MIT Press. Woronoff, Denis. Histoire de l’industrie en France du XVIème siècle à nos jours. Paris: Seuil. Wulff, Kurt. 1922. “Die Deutschkonservativen und die Preussische Wahlrechtsfrage.” PhD diss., University of Greifswald. Ziblatt, Daniel. 2009. “Shaping Demo­cratic Practice and the C ­ auses of Electoral Fraud: The Case of Nineteenth-­Century Germany.” American Po­liti­cal Science Review 103(1): 1–21. —­—­—. 2017. Conservative Parties and the Birth of Modern Democracy in Eu­rope. New York: Cambridge University Press.

INDEX

Page numbers in italics indicate figures and ­tables. Acemoglu, Daron: democ­ratization, 17, 211 Action Libérale, 64, 86, 90; 120, 122, 197; on isolating spaces, 159; Le Mun and, 46 Alype, Pierre, electoral irregularities in Indes Françaises, 194 amtliche Wahlbeeinflussung, use of state employees in German elections 30, 32 Anderson, Margaret Lavinia: bribery in German elections, 115; economic coercion, 213 Anglican Church, 44 Ansell, Ben: electoral reforms, 211; elite splits 14 Appel au Peuple, 56, 84 Argentina, PAN party, 14 aristocratic electoral practices, vote-­buying as, 120 Arnal, Franc, economic coercion in French elections, 158 Arsenscheck, Robert, electoral use of state resources in German elections, 74 Auer, Ignaz: ballot envelopes, 168; ballot irregularities, 33; on electoral urns, 172; publicity of elections, 191 Australia, adoption of isolating spaces, 5 Austrian Civil Code, vote-­buying and, 114 Bär, Karl, opposition to reforms limiting electoral use of state resources, 100 Badische Anilin und Sodafabrik (BASF), economic coercion, 152 ballot(s): and ballot envelopes, 5, 61, 153–57, 167–68, 171; brokers’ efforts to infer voters’ intentions, 149–50; colors of, 149, 167, 168; counting, 34–36; 192–95; design of, 145–46; distribution by candidates of Kartellbündnis, 148; distribution of, 34, 150–52; falsification of, 34; French

electoral law on ballot design, 147–49; German electoral law on ballot design, 148; manipulation of, 32–34; parallel lists recording voters’ choices, 151–52; position of ­table holding, 33–34, 151; reforming secrecy of, in Germany, 166–77; secrecy proposals 174, 176; shapes of, 148, 149, 167; use of small containers for collection of, 157, 172–74 ballot envelopes, 5, 61, 153–57, 167–68, 171 Ballot Act (1872), Britain, 5, 23, 65, 133 Banks, Edward, 100 Barodet, Désiré, collection of parliamentary manifestoes, 93 Bartissol, Edmont, ballot design, 148 Bassermann, Ernst: support for the adoption of ballot envelopes in 1896, 63; support for the adoption of isolating spaces, 171–72; standardized urns, 173–74 Bavarian Criminal Code, 114 Bérard, Alexandre: candidate representatives f­ avor wealthier candidates, 200, 201; opposition to candidate representatives, 198 Bebel, August, economic coercion in German industrial centers, 167 Belgian Penal Code, 181 Belgium: British Ballot Act inspiring reforms, 184, 186, 188; candidates’ electoral strategies in, 74; elections with runoffs, 10; electoral law, 114; electoral reforms (1877), 5, 6, 23, 65; Frère-­Orban government, 44; introduction of candidate representatives in, 184–88; loi Malou, 5, 65, 185, 187; party system of, 42; reforms limiting fraud, 53; secrecy of vote, 43; suffrage restrictions, 43 Benoist, Albert de, electoral coercion, 155

231

232 I NDEX

Benoist, Charles: contrast between universal suffrage and po­liti­cal real­ity, 3; electoral rule changes, 208; isolating spaces, 164; support for reforms protecting voting secrecy, 164 Beresford, William, indicted for corrupt electoral practices, 134 Berliner Volksblatt (newspaper), 32 Bertrand, Paul, electoral reforms limiting use of state resources, 90 Bienvenu-­Martin, Jean: Commission of Universal Suffrage, 122; electoral reforms limiting use of state resources, 98 Bischoffsheim, Raphaël Louis, vote-­buying in French elections, 29 Bismarck, Otto von: partisanship in elections, 55; politicization of state employees, 49 bistrouille, treating in France, 29 Blanc, Edmond, invalidation of election due to vote-­buying, 120 Blanc, Louis, universal suffrage, 2 Blödau, Herrmann von, use of state employees during elections, 82 Blum, Léon, on relationship between Republicanism and Socialism in France, 46 Boisset, Charles Guillaume, economic coercion, 145–46 Boix, Carles, extension of suffrage, 17 Bonapartists, 45 Bougon, Col­on ­ el, economic coercion and violation of voting secrecy, 152 Bourbon dynasty, 45 Briand, Aristide, on candidates’ ­limited agency to constrain brokers, 123 bribery, 24; convictions and acquittals in German elections, 118; in French Civil Code, 111–13; French Penal Code, 57; modernization and reduction of electoral, 16; negotiating electoral reforms, 133–43; punishments for, 4; treating relative to, 128–29 bribes: average prices of, in British elections, 128; economic costs of, 8 Britain: Ballot Act (1872), 5, 23, 65, 182–84; Corn Laws, 44; Corrupt and Illegal Practices Act (1883), 60; First Reform Act (1831), 42; introduction of candidate representatives in, 182–84; partisan asymmetry in vote-­buying, 130, 131; party system of, 42–43; reforming vote-­buying in, 126–31; reforms limiting fraud, 53; resource constraints for Liberal and Conservative candidates, 131, 132; secrecy of

vote, 43; suffrage restrictions, 43; vote-­ buying and treating, 29, 111–13, 127–29 Bulgaria, electoral use of gifts and ­favors in recent elections, 212 Burlet, Jules de, candidate representative, 187 Caïtucoli, Marc, electoral fraud in Corsica, 194 Canada, introduction of ballot envelopes, 5 candidate-­level perspective: consistency of candidate profile, 8, 40–41; demand for electoral reforms, 36–42; economic costs, 8–10, 13, 16, 38, 52–54, 66, electoral costs, 38–40, 41–42, 66; relative resource availability, 38–40 candidate representatives: in Belgium, 184–88; in Britain, 182–84; German electoral law and, 190–91; in Third Republic France, 191–205 candidature impériale, Republicans on candidature officielle during 1877 elections as, 72 candidature officielle: initiatives reducing, 22, 65; proposals to limit, 87, 88; reform of, 125; state power, 55, 56, 80; state resources and campaigns, 30, use by Monarchists during first elections of Third Republic, 72–73 candidature d’argent: electoral expenses and, 200; vote-­buying exchanges in France, 28, 60 Caprivi, Leo von, 102; economic policies by, 49, 51; government of, 105 Cassagnac, Paul de, on subdued electoral intervention during Third Republic as compared to Second Empire, 78, on universal suffrage, 3 Castellane, Boniface Boni de, as official candidate, 76 Cata­logne, Jacques, prefect endorsement of, 78 Catholics, electoral reforms, 63; Zentrum, 62, 70, 80 Chailley, Joseph Paul Amédée: endorsing candidate representatives, 199; on isolating spaces protecting voters against economic coercion, 164–65 Charnay, Jean-­Paul: on violations of voting secrecy 192; on vote-­buying, 29 Chauveau, Adolphe, on Napoleonic Code penalizing fraud, 180 Chavin, Émile, ballot envelopes, 156 Chiché, Albert: amendment limiting candidature officielle, 87; on state resource use during campaigns, 94, 95

I NDEX 233

Clagny, Albert de, on goals of reforms limiting vote-­buying, 121–22 collective corruption, 31, 56; sanctions on, 89; Senate commission of Universal Suffrage on, 89 colorful employment, 138; treating and expansion of, 130 Commission of Universal Suffrage (France), 3, 50, 56, 88, 98, 122; on candidate representatives, 197–98; on electoral reform, 154; on isolating space, 158; report on ballots, 147; in Senate, 62 commission de vérification des pouvoirs, French parliamentary commission investigating complaints of electoral malfeasance, 74 commission verifying election results, Imperial Germany. See Wahlprüfungskommission Conservative Party: Britain, 44; demand for reforms limiting campaign expenditures, 142; electoral reforms and, 134–40; resource constraints for, 131, 132; vote-­ buying and, 130–31, 131 Conservatives (Germany), 10, 43, 46, 105–6; ballot secrecy, 167; candidate representatives, 65, 197, 199; candidature officielle, 22; departments of po­liti­cal strength, 202, 203; electoral alliance, 11, 48, 49, 50, 99; electoral competition ­after cartel breakdown, 102, 103, 104; electoral interference, 102, 104; electoral reforms, 51, 70–71, 92, 211; electoral use of state resources, 59; Germany, 10, 11, 49, 50, 61–62, 70, 211; importance of runoffs for candidates, 103; isolating spaces for voters, 159, 161, 165; po­liti­cal support, 104; resource advantages of, 61–62; secret voting, 168–69; state resources and, 72, 97, 105–6, 107; support for invalidating elections, 106, 107; voter secrecy, 175–77, 176; voters’ electoral freedom, 100 Constans, Paul, resource disadvantage of Socialists, 65; support for isolating spaces 158, 165 Constitution of the German Empire, 43 Cornet, Lucien, opposition to candidate representatives, 201 Corn Laws, Britain and, 44 Corrupt and Illegal Practices Act (1883), 60, 138–41, 142; adoption of, 143, 210 Corrupt Practices Act (1854), 58, 133, 134, 137, 139 Corte, Graf, economic coercion, 82

Cox, Gary, economic development and increasing costs of bribing, 15, 16 Criminal Code of German Empire, 115 Cuverville, Jules de, opposition to vote-­ buying reforms, 121 Dahl, Robert, executive se­lection, 43–44 Daniel, André: disaffection with demo­cratic proj­ect ­after expansion of suffrage, 207; electoral co­ali­tions between Radicals and Socialists in runoffs, 50; Radical challenge to Méline government, 50; Radical opposition to electoral use of state employees, 56 Danish Civil Code, provisions sanctioning vote-­buying, 114 d’Aulan, Arthur de Suarès, use of mayors as brokers during campaign, 79 De Boury, Gérard Aubourg, opposition to ballot envelopes, 156 Defontaine, Félix: on ballot envelopes, 154; on candidate representatives, 196; electoral reform, 56; on isolating spaces, 157, 158 Delahaye, Dominique, on Republican refusal to allow representatives of opposition candidates in voting place, 201 demo­cratic institutions, dissatisfaction with, 207–8 democ­ratization: economic modernization and, 15–17, 52–53; empirical strategy to study, 19–21; implications for recent democracies, 212–16; implications for theories of, 209–12; Przeworski on, puzzle, 6–7; redistributivist theories of, 17–19, 210–11; theories presenting democ­ ratization as big bang, 19 Delalande, Nicolas, Radical support for progressive income tax, 48 de Monsaulnin, François-­Charles, removing opposition ballots, 150 d’Estourbeillon, Régis de: amendment limiting electoral use of state resource use, 94, 95; opposition to ballot envelopes, 156; opposition to promises made by ministers during campaigns, 87; opposition to reforms limiting bribery, 122 d’Etchepare, Louis d’Iriart, candidate representatives, 198 Deutsche Freisinnige Partei, electoral support by Prus­sian bureaucracy in 1907 election, 104 Deutsche Konservative Partei (German Conservative Party). See Conservatives (Germany)

234 I NDEX

de Vilers, Charles Le Myre, use of transparent ballots, 149 Dev­ille, Gabriel, proposal in support of candidate representatives, 196 Disraeli, Benjamin: Britain and, 44; on electoral malfeasance, 133 d’Ornano, Gustave Cuneo: electoral amendment opposing candidature officielle, 84–85; sponsorship of legislation limiting electoral use of state resources, 56 Dubief, Fernand, government support for isolating spaces, 163–64 Dumas, Gilbert, electoral fraud in French colonies, 194 economic coercion, 32–36; economic costs of, 8; electoral fraud, 34–36; freedom of vote and, 25–26 economic development: costs of electoral reforms, 9–10; democ­ratization and, 15–17, 52–53; electoral reforms and, 9, 96 economic modernization. See economic development Edmée, Olivier, incident of electoral fraud in Guadaloupe, 194 electoral clientelism, 42. See also treating; vote-­buying electoral costs: candidate perspective on, 38–40, 41–42; determinants of demand for electoral reforms, 39; elite splits and, 9 electoral fraud, 4; announcement of results, 195; ballot counting, 34, 35–36; determinants of support for reforms, 204; electoral irregularity, 37; incidents in German national elections, 189; introduction of candidate representatives in Britain, 182–84; limiting, in Imperial Germany, 188–91; Napoleonic Code and sanctions on, 63–64, 179, 180–82; partisan resource advantage, 202–5, 204. See also garnir l’urne electoral irregularities: ballot manipulation, 32–34; economic coercion, 32–36; electoral fraud, 34–36; investigating, 89–91; typology of, 37; universe of, 27–32; using state resources during campaigns, 30–32; violations of voting secrecy, 32–36; vote-­ buying, 28–30 electoral reforms, 4–6; adoption of, 4; combatting electoral fraud, 23–24; composition of co­ali­tions, 66; consistency of candidate’s profile, 40–41; countries with duopolistic competition, 210; demand from candidate-­level perspective, 36–42; desirability of, 21; determinants of demand

for, 38, 39; distribution of resources, 7–8; economic conditions shaping, 161–63, 162; economic development and, 96; elite splits and demand for, 47; elite splits as catalyst for, 210; France voting for, 91–98; limiting the politicizing of state resources, 21–22; limiting vote-­buying, 112–13; motivation of elites to embrace, 208–9; negotiation of, 27; pathways to, 9; secrecy of vote, 22–23; vote-­buying limitations, 22 electoral reforms: co­ali­tions for, 54–67; limiting incidence of fraud, 63–67; protecting voters’ autonomy, 60–63; reducing access to state resources during campaigns, 54–57; vote-­buying reforms, 57–60 elite splits: as catalyst for electoral reform, 210; demand for electoral reforms, 47; electoral be­hav­ior and, 50; electoral costs and, 10, 215; electoral reform, 9, 10; France, 11–12, 54, 56, 70–71, 178; Germany, 21, 50, 70, 99; importance of, 14; party formation and, 11–12, 70–71; term, 48; third-­wave democ­ratization and, 14 executive se­lection, Dahl on, 43–44 Fairbairn, Brett: bureaucratic support of Kartell candidates during runoffs in German elections, 104; use of ballot containers of heterogenous shapes in German elections, 172 Federation of Catholic Circles and Conservative Associations, 44 Ferrette, Paul Henri, on validity of election, 90 Ferroul, Ernest, support for candidate representatives, 196–97 Ferry, Jules, on universal suffrage, 1–2 Finnish civil code, vote-­buying and, 114 First Reform Act (1831), Britain, 42 Flaubert’s Sentimental Education, 1 Forster, W. E., on Secret Ballot Act, 183 Fould, Achille, vote-­buying, 28–29 Fournol, Étienne, governmental support during campaign, 77 Frère-­Orban, Walthère, governing Belgium, 44 France: adoption of isolating spaces, 165, 166; campaigns and state resources in nineteenth-­century elections, 71–83; candidature officielle, 55, 56, 65, 72; candidatures d’argent, 28–29, 60; elections with runoffs, 10; electoral use of state resources, 21; elite split between Republicans and Radicals, 12, 48, 70–71; elite

I NDEX 235

splits and electoral be­hav­ior, 50; garnir l’urne (garnish the urn), 36; government mobilizing voters, 2–3; introduction of candidate representatives, 191–205; investigating electoral irregularities in, 26; party system, 13, 43, 45; reforming voter secrecy in, 23, 153–66; sanctioning pression gouvernementale, 83–91; sources of quantitative analy­sis, 217; vérification des pouvoirs pro­cess, 25; vote-­buying incidence, 29, 115, 116, 117; vote-­buying limitations, 22; voting for electoral reforms, 91–98 Frary, Donald, on electoral importance of French prefects, 31 fraud: term, 34. See also electoral fraud ­Free Conservatives (Germany), 11, 50, 70; electoral alliance, 11, 48, 50; electoral reforms, 70; po­liti­cal support, 104 ­Free Liberal Party, Germany, 47, 51, 70, 71; ballot envelopes, 173; ballot secrecy, 167, 168; electoral alliance, 99; electoral reforms, 51, 63, 70; state resources and, 106, 107; support for invalidating elections, 106, 107; voter secrecy reforms, 175, 176; voters’ electoral freedom, 100 ­Free Liberals (Germany) see ­Free Liberal Party, Germany French Civil Code, sanctions for bribery in, 111–14 French Penal Code, 57 French Socialist Party, 46. See also Section Française de l’Internationale Ouvrière (SFIO) Freudenthal, Berthold, on incidence of electoral fraud in German elections, 189 Garnir l’urne (garnishing the urn), 36 Gambetta, Léon, on universal suffrage, 1, 2 Gauthier, Louis, on use state resources during campaigns, 94, 95 General Dictionary of Politics, 4 German Federation, 43 German Penal Code, 115, 117 German ­People’s Party, 47 Germany: amtliche Wahlbeeinflussung, 30, 32; Bismarck and, 46; breakdown of Kartellbündnis, 11–12, 23, 27, 49, 50–51; elections with runoffs, 10; electoral operations in, 81–83; electoral use of state resources, 21–22, 71–83; 50; elite splits and parties, 11–12, 43, 27, 50, 71, 109; Kartellbündnis, 48–49; limiting fraud in Imperial, 188–91; party system of, 46–48,

70, 71; proposals for protecting voting secrecy, 174; quiet shift in policy in Imperial, 98–109; reforming ballot secrecy, 26, 43, 166–77; Social Demo­cratic Party (SPD), 46–47; sources of quantitative analy­sis, 217–­18; support for reforms protecting voter secrecy, 176; vote-­buying incidence, 115, 116, 117 Girard, Alfred, ballot envelopes, 154 Gourd, Alphonse: candidate representatives, 195; pression officielle, 86; proposal limiting candidature officielle, 87; state employees in French elections, 30 Goussot, Emile, candidate representatives, 196 Graevenitz, Hugo von, electoral announcements for, 101 ­Great Britain: sources of quantitative analy­ sis, 218. See also Britain Gröber, Adolf: ballot envelopes, 61; ballot secrecy, 168 Gross, Ludwig, economic coercion and violation of voter secrecy, 152 Guesde, Jules, proposal to introduce candidate representatives, 195 Guyho, Léonard Correntin, proposal to introduce ballot envelopes and isolating space, 157 Guyot-­Dessaigne, Edmond, electoral reforms, 154 Haake-­Letschin, Gustav, preventing representatives of opposition candidates from being pre­s ent during the counting of ballots, 190 Haggard, Stephan, demo­cratic transitions, 20 Hamilton, Lord George, opposition to limits on campaign expenditures, 135 Hanham, H. J.: candidate expenditures on vote-­buying, 28; indictment of prominent Conservative politicians for corrupt electoral practices, 134 Hatschek, Julius: change in electoral jurisprudence to sanction use of state employees during elections 57, interventions of German state employees, 73–74; legislation introducing candidate representatives 64, princi­ple of publicity of elections, 190; rulings of electoral commission of Reichstag, 191, ­silent democ­ratization of electoral practices in Germany before World War 1, 71 Hélié, Faustin, Napoleonic Code penalizing fraud, 180

236 I NDEX

Hoesch, Felix, invalidation of election of, 106 Home Rule, 141 Hughes, Edwin, treating by, 128 Hungary, electoral use of gifts and ­favors in recent elections, 212 Imperial Party (Deutsche Reichspartei), candidates elected in runoffs, electoral fraud, 188 in­equality: and expansion of suffrage, 17; long causal chain linking economic in­equality and demand for electoral reform, 18, 211; and preferences for electoral reform, 67. See also landholding in­equality isolating spaces/isoloir: adoption in France, 157–59, 165, 166;: as “camera obscura”, 169; as “chambers of fear”, 169; introduction in Germany, 166–76; opposition to, 159–61, 168–69; support for 162, 168; term, 158 James, Henry, 136; on electoral corruption, 134–35; on volunteers and corruption, 138–39 Jaurès, Jean: policy proposals, 46; Socialists’ lack of access to state resources during campaigns, 85; on vote-­buying cases in French elections, 120 Jourdain, Louis, support for candidate representatives, 196 Kalyvas, Stathis, origin of religious cleavage, 44 Kam, Christopher: adoption of secret voting and bribing, 127; turnout-­buying, 129–30 Kartellbündnis (electoral alliance): ballot distribution, 148; breakdown of, 11–12, 23, 27, 49, 50–51, 99, 170; electoral competition ­after breakdown of, 102, 103, 104; electoral strategy of, 49; formation of, 48–49 Kartellkandidaten, coordination around a common candidate, 49. See Kartellbündnis Kasara, Kimuli, economic theories of democ­ ratization and vote-­buying reforms, 175 Kaufman, Robert, demo­cratic transitions, 20 Kayser, Jacques: elite split and origin of Radical Party, 48; departments of Radical strength, 202; Radicals and progressive income tax 48 Kerguézec, Gustave de, support of candidate representatives, 199 Kitschelt, Herbert, voters sanctioning candidates combining programmatic and clientelistic appeals 40–41

Kölnische Zeitung (National Liberal publication), 169 Krupp, Alfred, economic coercion, 60 Kühne, Thomas: role of Prus­sian Ministry of Interior in coordinating lower-­level employees during elections, 31, 49; Puttkamer’s electoral strategy, 79 Kulturkampf, attack on Germany’s Catholic minorities, 46 Kuo, Didi, 213 Lacour, Pierre Mielvaque de, variety of ballots used in French elections, 147–48 Lamarzelle, Gustave de, opposition to candidature officielle, 88–89 Lamendin, Arthur, economic coercion of low-­income voters, 155 landholding in­equality: and deference voting, 131; mea­sure­ment (France), 217; mea­sure­ment (Germany), 218; mea­ sure­ment (Britain), 218; and resource constraints for Liberal and Conservative candidates in Britain, 131–32; and support for elections limiting vote-­buying in France, 124; and support for the introduction of isoloir, 162, 166; and support for invalidation of elections politicizing state resourced in Germany, 107, and support for proposals limiting electoral clientelism in France, 97; and support for reforms limiting fraud, 204; and support for reforms limiting use of state resources in France, 95, 96; and support for reforms protecting voter secrecy in Germany, 176 Landräte: distribution of election material in official envelopes, 108; endorsement of candidates by, 101, 106; expelling representatives of candidates from voting place, 191; intimidation of ballot con­ vey­ors of opposition candidates, 31–32; mobilization of teachers and policemen by Landrat, 80; noiseless mobilization in 1903 election, 51; responsibilities of, 2, 79–80, 99–100, 105 Laporte, Gaston: ballot envelopes, 154; proposal for ballot containers, 157 Lasies, Joseph: comments on opposition of Radicals to reforms limiting fraud, 199; opposition to vote-­buying reforms, 121, 123 Law Repressing Acts of Corruption in the Organ­ization of Elections, 89 Leatham, W. H., on bribing in British elections, 28

I NDEX 237

Le Mun, Alfred, establishment of Action Libérale 46 Le Pommelec, François, proposal for adoption of uniform ballot envelopes, 153 l’Estourbeillon, Régis de: on bribery or treating, 122; electoral promises by ministers, 87–88; proposal limiting candidature officielle, 87; opposition to ballot envelopes invoking traditionalism of rural voters, 156; use of state resources during campaigns, 94, 95 Liberal Party, Britain: demand for reforms limiting campaign expenditures, 141, 142; electoral fraud, 185, 186, 188; electoral reforms and, 134–40; resource constraints for, 131, 132; vote-­buying and, 130, 131, 131 Lintilhac, Eugène: on collective corruption, 89; support for candidate representatives, 201; support for isolating spaces, 164 Lipset, Seymour, modernization theories of democ­ratization, 15 loi Malou (1877), Belgium, 5, 65, 185, 187 Luebbert, Gregory, regime outcomes in interwar Eu­rope, 19 Luxembourg, 5, 44; electoral law of, 114 Mackenzie, Forbes, corrupt electoral practices, 134 McLean, Iain, electoral realignment in Britain, 141 Madrid, Raúl, elite splits and electoral reforms, 14 Malézieux, Ferdinand, uniform ballot envelopes, 153 Malou, Jules, candidate representatives, 184–85, 186–87 Mares, Isabela: economic theories of democ­ ratization and vote-­buying reforms, 175; limitation of electoral reforms in recent democracies, 208; microhistorical institutionalism as research strategy, 20; per­sis­tence of electoral clientelism in recent democracies in Eastern Eu­rope, 212; treating in recent elections in Eastern Eu­rope, 30; use of state resources in recent elections in Eastern Eu­rope, 214 Masurel, Albert, economic coercion, 150 Mayer, Ernst: diffusion of Napoleonic code across Eu­rope, 112, 114,115; incidents of bribery sanctioned by German Penal Code, 115, 116; incidence of fraud in German elections, 189–90 Méline, Jules, downfall of government of, 50

Meltzer-­Richard model, in­equality and re­distribution, 18 micro-­historical institutionalism, research strategy, 20 Millevoye, Lucien, opposition to judicial investigation of electoral corruption, 90 modernization, economic, and demo­cratic reform, 15–17 Monarchists: candidate representatives, 197, 199; candidature officielle, 22; electoral reforms, 211; electoral use of state resources, 59, 62, 72; isolating spaces for voters, 159; legislators, 211; private resources of, 47; profession de foi including progressive income tax, 94; vote-­ buying and, 120, 122–23 Montesquieu, The Spirit of Laws, 3 Moore, Barrington, regime transitions, 19 Müller, Georg, opposition to isolating spaces, 169 Napoleonic Code, 4, 34, 57, 205; earliest efforts to limit electoral fraud, 63–64, 179, 180–81; efforts to sanction fraud, 63–64; provisions to sanction vote-­buying, 113–26 National Liberals, 10, 13, 46, 49, 50; ballots and, 152; ballot secrecy, 167; electoral alliance, 11, 12, 48, 50–51, 54, 99; electoral consequences of cartel breakdown, 102, 103, 104, 106; electoral interference, 102, 104; electoral reforms, 51–52, 63, 70–71, 170–71, 211; electoral urns, 173–74; importance of runoffs for candidates, 103; resource advantages of, 61–62; secret voting, 168–69; state resources and, 57, 106, 107; status quo and, 110; support for invalidating elections, 106, 107; voter secrecy, 175–77, 176; voters’ electoral freedom, 100; World War I and, 51–52, 57 National Union of Conservative Associations, 44 Netherlands, 5; elections with runoffs, 10 neutrality of state employees in elections, implicit rejection in French elections 55, 71, noiseless intervention, Caprivi policy, 51, 102, 104 North German Federation, 43 Norway, introduction of ballot envelopes, 5 Odilon-­Barrot, Jean: electoral reforms, 121; electoral use of state resource, 85; investigating irregularities, 90

238 I NDEX

O’Donnell, Guillermo, elite splits as catalyst of democ­ratization, 14 O’Leary, Cornelius, survey of electoral conditions, 143 Ollivier, Louis, amendment to investigate electoral irregularities by jury, 90 Opportunist Republicans: France, 48; Radicals and, 48 Organic Decree of 2 February 1852, Article 38 of, 73 partisan resource advantage, electoral reform, 202–5, 204 party formation, elite splits and, 11 Pastre, Rochilde, proposal for introduction of isolating spaces, 158 Périer, Casimir, expectation of electoral involvement of state employees, 71 Penal Code of German Empire, 181 Pierre, Eugène: organ­ization of voting place in French elections, 192; rules about voiding of ballots in French elections, 192; treatise on electoral law, 84 Pilenco, Alexandre: ballot counting, 194; candidates and state resources, 55 policy ­favors, 30 polling place, 184 Pomereau, Michel de: candidate representatives, 197; isolating spaces, 159 presiding officers, voting place and, 183–84 pression gouvernementale: reforms limiting, 47, 67, 83–91, 211. See also candidature officielle Primrose League, 139 prefects: announcements of candidates’ connections to ministers, 31; collective promises to localities, 31, 83; discouragement of opposition candidates by, 78; electoral deployment of, 31, 55, 76, 78; endorsement of candidates by, 78; expectation of electoral loyalty of, 55; mobilization of mayors by, 31, 76; proposals for ­legal sanctions on interventions by, 86; punishment of disloyal mayors by, 78–79; Socialists and lack of electoral support by, 85 princi­ple of publicity of elections: Germany’s electoral law, 190; France, 35 professions de foi: Barodet, 93; candidates and parties demanding introduction of progressive income tax, 93–94, 94; promises of candidates in, 125, 162, 162–63 progressive income tax, professions de foi including, 93–94, 94

Prus­sia: Landräte, 2; electoral intervention, 104–5; po­liti­cal developments in, 52 Prus­sian Criminal Code, 114 Prus­sian Ministry of Interior, 80, 104, 105; electoral responsibilities, 30 Przeworski, Adam: democ­ratization puzzle, 6–7, 12; elite splits, 14; modernization theory of democ­ratization, 15 Puttkamer, Robert von: electoral strategy of Kartellbündnis, 49; fraud complaints in election, 188; politicizing elections, 79; state officials in elections, 55 Puttkammer-­Plauth, Bernhard von, 80 Radical Party: France, 12, 48; profession de foi including progressive income tax, 94, 199–201. See also Radicals Radicals: ballot envelopes and isolating spaces, 154–58, 161, 163; candidate representatives, 199–200; departments of po­liti­cal strength, 202, 203; determinants for support of reforms limiting fraud, 185, 204; electoral reforms limiting candidature officielle, 54, 62, 64–65, 92, 96–98; elite split, 11–12, 23, 27, 48, 54, 56, 70–71; vote-­buying reforms, 124, 125 Raikes, Henry Cecil, electoral reforms, 134 Randall, Vicki, voters sanctioning candidates that combine clientelistic and programmatic appeals, 40–41 rastells, vote-­buying in France, 29. See also treating Rémond, René, po­liti­cal right in France, 45 redistributive theories of democ­ratization, 15, 17–19; assumptions about elite preferences for electoral reforms, 66–68, theoretical limitations of, 18 Republican Party: candidate representatives, 197–99; criticism of electoral malfeasance, 1, 3; departments of po­liti­cal strength, 202, 203; electoral reforms limiting candidature officielle, 88, 91–92, 96–98, 211; electoral use of state resources 47, 72–78, 86; elite split, 11, 12, 54, 56, 70–71, 178; governments led by, 45; hy­poc­r isy ­towards use of state resources, 55; isolating spaces, 62, 154–56, 159–61, 163, 165; profession de foi including progressive income tax, 94; reforms limiting fraud, 198–99, 204; vote-­ buying reforms and, 120–23, 124, 125 Republicans. See Republican Party Returning Officers, 184

I NDEX 239

Richard, Pierre, amendment proposing judicial investigation of use of state resources, 90 Richter, Eugen, description of use of state resources in German elections, 82 Richtlinien, rules for electoral intervention of state employees, 104 Rickert, Heinrich: ballot secrecy, 168; economic intimidation, 33–34; electoral reforms, 172, 173; freedom of the vote, 25–26 Roberts, Crompton, collection of information on voters’ needs, 126 Robinson, James, redistributivist theory of democ­ratization, 17, 211 Roche, Jules Antoine, arbitrariness of judges, 90 Romania: adoption of isolating spaces in 1899, 5; anticorruption agency, 213–14; vote-­buying in recent elections, 212 Rosanvallon, Pierre: conflict between demo­ cratic ideals and po­liti­cal real­ity, 208; perpetual disenchantment of democracies, 212 Rose, Philip, corrupt electoral practices, 134 Ruau, Joseph: Belgian electoral reforms, 198; electoral reform bills, 50, 56, 154–55; recommendation for adoption of isolating spaces, 158, 160 Rudelle, Odile, departments of Republican strength, 202 runoffs: consequences for demand for electoral reforms demand for electoral reforms, 47; in German elections, 103 Rueschemeyer, Dietrich, working-­class mobilization and democ­ratization, 19 Russell, G. W. E., Corrupt and Illegal Practices Act, 143 Samuels, David: electoral reforms, 211; elite splits, 14 Samuelson, Henry, opposition to penalties on voters accepting drinks or meals, 138 Schädler, Franz, Zentrum proposals for the adoption of ballot envelopes and changes in the design of the electoral urn, 174 Schmitter, Philippe, elite splits as catalyst for democ­ratization, 14 scrutineers, role in counting of ballots, 192 Sénac, Urbain: candidate representatives, 199; on hired representatives, 200 Second Empire France: continuity of electoral strategies politicizing state resources to Third Republic, 71, 73; electoral deployment of prefects, 56, 78

secrecy of vote: adoption of nominal voting secrecy, 43; candidate representatives in Britain and, 182–84; commitment to, 145–47; electoral irregularity, 37; electoral reforms for, 22–23; reforming, in Germany, 166–77; violations of, 32–36. See also voter autonomy Secret Ballot Act (1872), Britain, 43, 182–84 Section Française de l’Internationale Ouvrière (SFIO): creation, 46; po­liti­cal demands, 46. See also Socialist Party, France Sentimental Education (Flaubert), 1 Seydewitz, Carl Friedrich von, election of, 100 Seymour, Charles: on Corrupt Practices Act, 58, 133, 137; growth of electorate and rise of collective corruption, 129; on importance of French prefects, 31; sanctions on illegal versus corrupt practices in British electoral law 137, 138; treating in British elections, 29 SFIO. See Section Française de l’Interna­ tionale Ouvrière (SFIO) Smith, W. H.: candidate, 28; on Corrupt Practices Bill, 139; on electoral reforms, 134 Social Demo­cratic Party (SPD), Germany, 46–47, 70, 71; candidates, 104; electoral reforms, 70, 167, 171–72; intimidation of ballot con­vey­ors, 82, 83, 92, 101; proposals for reforms protecting voter secrecy, 167–68, 174, 175–77; support for invalidating elections of candidates who politicized state resources, 106, 107; targeted layoffs of voters supporting, 152; voters’ electoral freedom, 100; working-­class mobilization, 46–47 Socialist Party, France, 46; candidate representatives, 65, 195–97; determinants for support of fraud reforms, 204; electoral reforms, 63, 92; isolating spaces for voters, 158, 161, 163; profession de foi including progressive income tax, 94; vote-­buying and, 124 Spirit of Laws (Montesquieu), 3 state coercion, 30 state employees: elections in France, 80–81; elections in Germany, 81–83; influencing elections, 31–32 state f­ avors and pressure, 30; electoral irregularity, 37; electoral reform from candidate perspective, 38–40

240 I NDEX

state resources: amendments seeking to limit use of, 94, 95; campaigns in nineteenth-­ century elections, 71–83; elections in Second Empire France, 71–73; electoral commission refusing to sanction interventions, 107–8; electoral strategies politicizing, 75; electoral use in Imperial Germany, 73–74; limits on politicizing use of, 21–22; policy shift in Imperial Germany, 98–109; reducing access to, during campaigns, 54–57; support for invalidating elections, 106, 107; support for proposals limiting, 96–97, 97; as symbols of state power, 57; unequal distribution of, 67–68; using, during campaigns, 30–32 Stephens, Evelyne Huber, working-­class mobilization and democ­ratization, 19 Stephens, John, working-­class mobilization and democ­ratization, 19 Stokes, Susan: economic development and vote-­buying reforms, 124–25, 141; on financing campaigns, 52; modernization increased difficulties of candidates to monitor brokers, 16 Streseman, Gustav, proposal to introduce standardized urns, 173 Struckmann, Gustav, opposition to proposal to adopt ballot envelopes, 170–71 suffrage: expansion of, 2; incumbent po­liti­cal elites and, 17; universal manhood, 1–2, 43 Surchamp, Abel, electoral use of disaster relief, 77 Swedish civil code, vote-­buying and, 114 Switzerland, introduction of ballot envelopes, 5 Tack, Gustave, praise of Britain’s electoral reform and adoption of candidate representatives, 184 Tanchoux, Philippe, practical ele­ments facilitating transparency of elections in France, 192 Teele, Dawn, electoral motivations explaining extension of suffrage, 19 Teorell, Jan, consequences of ballot secrecy for bribery, 213 Théorie du Code Pénal (Chauveau and Hélié), 180 Third Republic (France), 1, 13; elite splits, 27; first elections in, 3; po­liti­cal system of, 45–46 Tielsch, Egmon, layoff threat of voters who supported opposition candidate, 34

treating: in British elections, 111–13, 128–29; 137–38; in French elections, 29; over time change in importance, 128–29; as signal of electoral strength and competitiveness, 29, 30; vote-­buying and, 29, 143–44. See also rastell Treating Act (1677), 58, 133 turnout-­buying: increase in importance in British elections following adoption of secret voting, 130; oversized network of brokers to facilitate turnout-­buying, 129–30 universal manhood suffrage, 1–2, 6; France and Germany, 43 Viviani, René: cosponsorship of legislation sanctioning use of state resources during elections, 56, 85, 96; proposal to change pro­cess investigating electoral irregularities, 90; proposal to limit vote-­buying, 121, 123 Voilliot, Christophe: break in documentation of electoral intervention of prefects in French elections, 78; electoral reliance on state employees, 73 von Eulenburg, Friedrich Graf, electoral obligations of state employees, 72 von Loebell, Friedrich, election of, 191 von Manteuffel, Otto, expectations of electoral involvement by state employees, 72 von Mirbach, Julius: isolating spaces, 169; secret voting, 169 von Stumm, Carl, opposition to secret voting, 169 von Stumm, Freiherr, ballots printed on cardboard paper to monitor voters’ choices, 148 vote-­buying: Britain, 111–13; decisions in French courts, 118–19; definition of corruption, 122–23; economic modernization limiting, 15–16; electoral irregularities, 28–30, 37; electoral reform from candidate perspective, 37–38; electoral reforms limiting, 22; inequalities among candidates and parties, 119–20; Napoleonic Code and, 113–26; partisan asymmetry in expenses, 130–31, 131; penalization of, 119; reforming, 57–60; reforming, in Britain, 126–31; support for reforms in French Chamber of Deputies, 124; treating and, 143–44

I NDEX 241

voter autonomy, 4; brokers and intention of, 149–50; commitment to secret voting, 145–47; electoral surveillance of, 146; introduction of isolating spaces for, 157–59; isoloir, 160; opposition to isolating spaces, 159–61; parallel lists recording choices of, 151–52; placing ballots in urn, 157; reforming voter secrecy in France, 153–66; reforms protecting, 60–63; targeted layoffs, 152–53 voter secrecy: brokers violating, 149–50; reforming, in France, 153–66; reforming, in Germany, 176 voting: extension of rights, 1; polling place and voting place, 184; reforming pro­cess, 4–5 Wahlprüfungskommission: electoral irregularities, 74; Germany’s verification pro­ cess, 25, 190; policy shift within, 99 Waldeck-­Rousseau, Pierre: on ballot envelopes, 156; on candidate representatives, 197–98, 201; on government’s position in debates, 88; on isolating spaces, 158; opposing isolating spaces, 159, 160; on

proving corrupt intention, 117; supporting official candidate, 76 Williamson, Jeffrey, 20 World War I, 71; electoral reforms and, 21–22; National Liberals, 51–52, 57, 99; National Liberals prior to, 108–9, 110 Young, Lauren: limitation of electoral reforms in recent democracies, 208; per­ sis­tence of electoral clientelism in recent democracies in Eastern Eu­rope, 212; treating in recent elections in Eastern Eu­rope, 30; use of state resources in recent elections in Eastern Eu­rope, 214 Zentrum: ballot envelopes, 173; ballot secrecy, 167, 168; candidates, 62, 70, 80, 104–5; Germany, 46; importance of electoral freedom, 100; paper ballots, 148; po­liti­cal support, 104–5; support for invalidating elections, 106, 107; voter secrecy, 175–77, 176 Zuydwyk, Clemens Heereman von, importance of electoral freedom, 100

A NOTE ON THE T YPE

This book has been composed in Adobe Text and Gotham. Adobe Text, designed by Robert Slimbach for Adobe, bridges the gap between fifteenth- and sixteenth-century calligraphic and eighteenth-century Modern styles. Gotham, inspired by New York street signs, was designed by Tobias Frere-Jones for Hoefler & Co.