Spine 13mm

C4
Royal
Demy
B-format
POST-WAR ERA
THE ORIGIN OF ELECTORAL SYSTEMS IN THE
The Origin of Electoral
Systems in the Post-war
Era
A worldwide approach

Krister Lundell

Krister Lundell

I S B N 978-0-415-47714-7

9 780415 477147 Routledge Research in Comparative Politics

www.routledge.com • an informa business

Bottom edge
Spine 13mm
The Origin of Electoral Systems in the
Post-­war Era

This book explores and presents the influence of contextual factors on the choice
of electoral systems for parliamentary elections in both democracies and non-­
democracies around the world.
Taking a macroscopic approach, the author focuses on structural explana-
tions, with an emphasis on general patterns rather than country specific explana-
tions. Drawn from a wealth of data, the book presents the frequency of the
adoption of each electoral formula and system in the post-­war era and is fol-
lowed by a theoretical elaboration of electoral system choice. The author then
draws on rational, cultural/historical and institutional theories which are system-
atically analysed by means of sophisticated bivariate and multivariate tech-
niques. Lundell demonstrates that few electoral systems have been chosen from
rational considerations and the impact of the cultural and historical setting is tre-
mendous; colonial legacy, regional influence and temporal trends largely explain
the cross-­national variation in electoral systems.
This book will be of interest to students and scholars of political science,
electoral politics and comparative politics.

Krister Lundell is a post-­doctoral researcher at the Department of Political

Science at Åbo Akademi University, Finland. His previous publications include
Determinants of Candidate Selection: The Degree of Centralization in Compar-
ative Perspective.
Routledge research in comparative politics

1 Democracy and Post-­Communism

Political change in the post-­communist world
Graeme Gill

2 Sub-­State Nationalism
A comparative analysis of institutional design
Edited by Helena Catt and Michael Murphy

3 Reward for High Public Office

Asian and Pacific Rim states
Edited by Christopher Hood and B. Guy Peters

4 Social Democracy and Labour Market Policy

Developments in Britain and Germany
Knut Roder

5 Democratic Revolutions
Asia and Eastern Europe
Mark R. Thompson

6 Democratization
A comparative analysis of 170 countries
Tatu Vanhanen

7 Determinants of the Death Penalty

A comparative study of the world
Carsten Anckar

8 How Political Parties Respond to Voters

Interest aggregation revisited
Edited by Kay Lawson and Thomas Poguntke
 9 Women, Quotas and Politics
Edited by Drude Dahlerup

10 Citizenship and Ethnic Conflict

Challenging the nation-­state
Haldun Gülalp

11 The Politics of Women’s Interests

New comparative and international perspectives
Edited by Louise Chappell and Lisa Hill

12 Political Disaffection in Contemporary Democracies

Social capital, institutions and politics
Edited by Mariano Torcal and José Ramón Montero

13 Representing Women in Parliament

A comparative study
Edited by Marian Sawer, Manon Tremblay and Linda Trimble

14 Democracy and Political Culture in Eastern Europe

Edited by Hans-­Dieter Klingemann, Dieter Fuchs and Jan Zielonka

15 Social Capital and Associations in European Democracies

A comparative analysis
Edited by William A. Maloney and Sigrid Roßteutscher

16 Citizenship and Involvement in European Democracies

A comparative analysis
Edited by Jan van Deth, José Ramón Montero and Anders Westholm

17 The Politics of Foundations

A comparative analysis
Edited by Helmut K. Anheier and Siobhan Daly

18 Party Policy in Modern Democracies

Kenneth Benoit and Michael Laver

19 Semi-­Presidentialism Outside Europe

A comparative study
Edited by Robert Elgie and Sophia Moestrup

20 Comparative Politics
The principal-­agent perspective
Jan-­Erik Lane
21 The Political Power of Business
Structure and information in public policymaking
Patrick Bernhagen

22 Women’s Movements
Flourishing or in abeyance?
Edited by Marian Sawer and Sandra Grey

23 Consociational Theory
McGarry and O’Leary and the Northern Ireland conflict
Edited by Rupert Taylor

24 The International Politics of Democratization

Comparative perspectives
Edited by Nuno Severiano Teixeira

25 Post-­communist Regime Change

A comparative study
Jørgen Møller

26 Social Democracy in Power

The capacity to reform
Wolfgang Merkel, Alexander Petring, Christian Henkes and Christoph Egle

27 The Origin of Electoral Systems in the Post-­war Era

A worldwide approach
Krister Lundell
The Origin of Electoral
Systems in the Post-­war
Era
A worldwide approach

Krister Lundell
First published 2010
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Simultaneously published in the USA and Canada
by Routledge
270 Madison Ave, New York, NY 10016
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2010 Krister Lundell
Typeset in Times by Wearset Ltd, Boldon, Tyne and Wear
Printed and bound in Great Britain by TJI Digital, Padstow, Cornwall
All rights reserved. No part of this book may be reprinted or reproduced or
utilized in any form or by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying and recording, or in
any information storage or retrieval system, without permission in writing
from the publishers.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging in Publication Data
The origin of electoral systems in the Post-­war era: a worldwide approach/
Krister Lundell.
p. cm. – (Routledge research in comparative politics; 27)
Includes bibliographical references and index.
1. Elections. 2. Representative government and representation. 3.
Comparative government. I. Title.
JF1001.L87 2009
324.609′045–dc22 2009008119

ISBN10: 0-415-47714-X (hbk)

ISBN10: 0-203-87184-7 (ebk)
ISBN13: 978-0-415-47714-7 (hbk)
ISBN13: 978-0-203-87184-3 (ebk)
This book is dedicated to Gun-­Britt and Bengt Lundell
Page Intentionally Left Blank
Contents

List of illustrations xi
Preface xiii
Acknowedgements xv

1 Introduction 1
Objectives and scope 1
Does electoral system choice matter? 3

2 Research design 5
2.1 Methodological considerations 5
2.2 The research population 6
2.3 Previous attempts at broad theoretical explanations 11
2.4 A frame of reference 16

3 Electoral systems 17
3.1 Electoral formulas and systems – a presentation 18
3.2 Classifying electoral systems 23
3.3 An introductory survey of electoral system choice 26

4 The rational perspective 32

4.1 Structurally generated determinants 34
4.2 Ethnic and cultural diversity 34
4.3 Country size 41
4.4 Actor-­related determinants 45
4.5 Party system structure 47
4.6 Party system transformation 53
4.7 The rational perspective: summary 57
x   Contents
5 History and culture: patterns of diffusion 59
5.1 Colonial diffusion 62
5.2 Regional diffusion 69
5.3 Temporal diffusion 73
5.4 Patterns of diffusion: summary 86

6 The institutional perspective 88

6.1 Form of government 90
6.2 Territorial organization 93
6.3 Chamber structure 95
6.4 The institutional perspective: summary 98

7 Multivariate patterns 99

8 Conclusion 110
Are electoral systems permanent? 110
Summary of findings 111
What next? 114

Appendix 116
Notes 122
Bibliography 124
Index 130
Illustrations

Figures
3.1 Frequencies of electoral system choice 1945–2007 31
3.2 Frequencies of electoral system choice in democracies 1945–2007 31
5.1 Frequencies of first electoral system choices 1945–2007 76
5.2 Frequencies of electoral system change 1945–2007 76
5.3 Frequencies of first electoral system choices in democracies
1945–2007 77
5.4 Frequencies of electoral system change in democracies 1945–2007 78

Tables
3.1 A typology of electoral systems 25
3.2 Electoral system choice 1945–2007 27
3.3 Electoral formulas and systems 2007 28
3.4 Electoral system choice in democracies 1945–2007 29
3.5 Electoral formulas and systems in democracies 2007 30
4.1 Frequencies of ethnic and cultural diversity 38
4.2 Associations between ethnic/cultural diversity and electoral
system choice 39
4.3 Frequencies of ethnic and cultural diversity in democracies 40
4.4 Associations between ethnic/cultural diversity and electoral
system choice in democracies 40
4.5 Association between country size and electoral system choice 43
4.6 Association between country size and electoral system choice
in democracies 44
4.7 Party system fragmentation in plurality systems 51
4.8 Party system fragmentation in proportional systems 52
4.9 Party system fragmentation in democracies with plurality systems 52
4.10 Party system fragmentation in democracies with proportional
systems 53
5.1 Colonial powers, colonies and colonial legacy 66
5.2 Associations between colonial diffusion and electoral system choice 68
xii   Illustrations
5.3 Colonial powers, colonies and colonial legacy in democracies 68
5.4 Regional diffusion and electoral system choice 73
5.5 Regional diffusion and electoral system choice in democracies 73
5.6 Temporal diffusion and electoral system choice 77
5.7 Temporal diffusion and electoral system choice in democracies 79
6.1 Associations between form of government and electoral
system choice 92
6.2 Associations between territorial organization and electoral
system choice 95
6.3 Associations between chamber structure and electoral system
choice 97
7.1 Patterns of diffusion and electoral system choice 101
7.2 Patterns of diffusion, the rational perspective and electoral
system choice I 102
7.3 Patterns of diffusion, the rational perspective and electoral
system choice II 103
7.4 Three explanatory perspectives and electoral system choice I 104
7.5 Three explanatory perspectives and electoral system choice II 105
7.6 Patterns of diffusion and electoral system choice in democracies 106
7.7 Three explanatory perspectives and electoral system choice in
democracies 108
A.1 Electoral system choice by country 1945–2007 116
Preface

The study of electoral systems was long dominated by efforts to establish their
consequences. In the mid-­1980s, one of the most prominent experts in the field
pointed out that there was a shortage of research on causes of electoral systems.
Some progress has certainly been made but there is still a sizeable gap with regard
to understanding the basis of electoral system choice. The purpose of this book is
to fill a part of this gap in electoral system research – more exactly, the study deals
with contextual determinants of electoral system choice. In order to break new
ground, a broad design is applied. A survey of the literature on the origin of polit-
ical institutions in general and electoral institutions in particular results in a frame-
work that encompasses rational, historical/cultural and institutional explanations. It
should be observed that the book is not about deciding which is the best electoral
system. In fact, I venture to say that there is no universally correct, optimal or most
democratic electoral system. Thence, no arguments for or against particular
systems are expressed within the frames of this book. The ambition is to identify
general patterns of electoral system choice from a contextual perspective.
The present book is based on my doctoral thesis completed a few years ago.
Needless to say, I am indebted to many people. First of all, I would like to thank
Professor Emeritus Dag Anckar at the Åbo Akademi University – my chief
source of inspiration and encouragement in the academic world – for valuable
assistance and insightful comments. I am deeply thankful to Professor Lauri Kar-
vonen and Professor Carsten Anckar for providing guidance during various
phases of the project. I also wish to thank all the other colleagues at the Depart-
ment of Political Science at Åbo Akademi for making Hus Lindman such a stim-
ulating and inspiring research environment. Furthermore, I want to express my
gratitude to Professor David M. Farrell, who as opponent at the disputation criti-
cally, yet constructively examined the thesis – due to his advice, comments and
expert knowledge, I was able to improve the original work.
In addition to intellectual stimulus, I have received financial support from a
number of instances while working on the project, most notably from the Finnish
national graduate school VAKAVA, the Research Centre at the Åbo Akademi
University Foundation, and the research project Democracy: A Citizen Perspec-
tive. I am very thankful for all the financial support that I have received and for
having had the opportunity to focus entirely on writing this book.
xiv   Preface
Last, I would like to express my thankfulness to Tarja for providing mental
strength during the completion of the book, and to my parents for encourage-
ment and support in every aspect of life; thereby, among other things, facilitating
an academic career.

Krister Lundell
Acknowledgements

The publisher and author would like to thank the following for permission to
reprint material in The Origin of Electoral Systems in the Post-­war Era:

Harvard University Press for permission to reprint excerpt from Public Policy:
reprinted by permission of the publishers from Public Policy, vol. 17, edited by
John D. Montgomery and Albert O. Hirschman, Cambridge, MA: Harvard Uni-
versity Press, (1968), 273, copyright © 1968 by the President and Fellows of
Harvard College.

Oxford University Press for permission to reprint excerpts from their journals.
Excerpt from Mixed-­Member Electoral Systems by M.S. Shugart and M.P. Wat-
tenberg (2001), 16 and excerpt from Democracy and Elections by R.S. Katz
(1997), 109. Online, available at: www.oup.com.

Palgrave Macmillan for permission to reprint excerpt, Comparing Electoral

Systems by David M. Farrell (1998), 5.

Åbo Akademi University Press for permission to reprint material, now revised,
which was previously published in Contextual Determinants of Electroral
System Choice: A Macro-­Comparative Study 1945–2003 by Krister Lundell
(2005), ISBN 951-765-298-4.

Universitetsforlaget/Aschehoug Agency for permission to reprint Citizens, Elec-

tions, Parties by Stein Rokkan, New York: David McCay (1970), 77 and 157.
Page Intentionally Left Blank
1 Introduction

Objectives and scope

The traditional way of studying electoral systems is to focus on the consequences
of different electoral rules. Electoral system research in the twentieth century
also encompasses strategic effects of electoral methods, that is, the nature and
extent of choice given to voters under different systems. Since the fall of the Iron
Curtain and the breakdown of Communism, a growing trend in the literature has
been to understand the basis for electoral system choice rather than the con-
sequences of electoral laws. The wave of democratization in Third World coun-
tries and the debate on electoral reform in advanced democracies since the early
1990s have also contributed to the trend of studying electoral systems as
explanandum instead of explanans. The present book belongs to this field of
electoral system research.
Until the late 1950s, there was a strong belief that political results were
almost completely determined by the institutional framework. Maurice Duver-
ger, one of the foremost theorists within this approach, emphasized the impor-
tance of electoral rules in explaining the nature of a country’s party system
(1951). The growing influence of sociological methods and theories in political
science in the 1950s called this view into question by emphasizing the role of
social phenomena as the main determinants of party systems. Party Systems and
Voter Alignments: Cross-­national Perspectives, written by Seymour Martin
Lipset and Stein Rokkan (1967) is perhaps the most central work within the soci-
ological approach. The authors maintained that the emergence of party systems
in the industrialized world is explained by four lines of social cleavage: centre/
periphery, state/church, land/industry and owner/worker. Modern party systems
are seen as a result of the kinds of cleavages that were present in society when
political parties were organized, and how conflicts were managed. Electoral
systems only modified socially determined party systems.
Although major historical events and ethnic, religious and economical cleav-
ages cannot be neglected in the study of electoral institutions, several research
efforts since the 1960s have confirmed the importance of electoral rules in
shaping politics. Both approaches bear relevance to this book. The contextual
environment provides the basic theoretical framework for explaining electoral
2   Introduction
system choice. However, the third wave of democratization has given rise to a
similar third wave of electoral system choice, in which deliberate electoral
system design has played a central role. The virtues of different electoral systems
have been widely debated on the basis of research on political consequences of
electoral rules. The sociological and the institutional approach are actually
related to each other: the former regards the electoral system as an expression of
social cleavages and party systems, whereas the latter regards the electoral
system as a manipulative instrument for shaping politics in accordance with
societal requirements. Thus, both represent a rational perspective on the role of
electoral systems. However, it should be observed that this study is not con-
cerned with rational choice analysis. In this context, the term rational means that
the structural traits of a society constitute problems that need to be solved by
appropriate electoral arrangements.
Another approach maintains that electoral systems, as well as other constitu-
tional and institutional features, reflect the historical and cultural contexts to
which they belong. It considers electoral system choices as the consequence of
various kinds of diffusion. According to this perspective, the choice of political
institutions in a society is the result of influence from other societies rather than
a response to sociopolitical requirements. The rational and the historical/cultural
approaches are not opposed to each other: an institutional choice may be both
rational and imitated. Furthermore, the constitutional setting may determine the
choice of particular institutions. This approach maintains that some institutional
choices follow logically upon others, whereas some combinations tend to be
inefficient and inappropriate. In a certain institutional setting, some electoral
systems may be more attractive than others. The influence of other political insti-
tutions constitutes the third basic approach to the study of electoral system
choice.
Consequently, the purpose of the study is to explore the influence of contex-
tual factors on the choice of electoral systems for (direct) parliamentary elections
during the post-­war era. Because of the macroscopic approach, I shall not pay
attention to the specific factors that led to the adoption of the electoral law in
each country. Rather, my intention is to discover general patterns and create an
explanatory framework of the adoption of electoral systems in the post-­war era.
Electoral system choice in democracies is separately analysed. A combination of
induction and deduction is applied, since there is no established theory in this
field of electoral system research. Associations between contextual factors and
electoral system choice are analysed from three competing but simultaneously
overlapping theoretical perspectives: a rational, a historical and cultural, and an
institutional perspective.
Four variables are analysed within the rational perspective: ethnic/cultural
diversity, country size, party system structure, and party system transformation.
The former two are pure structural traits, whereas the latter two also include an
actor-­related dimension. To be sure, the book is concerned with contextual deter-
minants but the role of political actors is not totally disregarded. More specifi-
cally, the study does not focus on actors but on the political context in which
 Introduction   3
actors make their decisions on electoral provisions. This matter becomes particu-
larly apparent when the relationship between party systems and electoral systems
is analysed. Electoral system choice as a consequence of the party system struc-
ture is foremost an actor-­related matter but has a structural dimension as well.
As to the rational perspective, I consequently distinguish between structurally
generated determinants and actor-­related determinants.
The historical and cultural perspective consists of the following variables:
colonial diffusion, regional diffusion and temporal diffusion. Although both
colonial legacy and regional influence are concerned with imitating institutional
choices of prestigious countries, I follow the common approach of treating colo-
nial legacy as a distinct kind of diffusion. In addition to the hierarchical and
spatial dimensions of imitating institutional choices embodied by colonial and
regional diffusion, a temporal dimension is introduced by including a variable
that regards electoral system choice as an epoch phenomenon. Within the insti-
tutional perspective, three variables are analysed: form of government, territorial
organization and chamber structure. This perspective certainly rests upon a
weaker theoretical basis than the other two; yet it is justified with respect to the
interest of analysing constitution-­building and the way countries tend to combine
different institutions.
There are several reasons for choosing the year 1945 as starting point. All
national elections prior to the twentieth century were held under either the plu-
rality or the majority rule; thus, providing little variation of the dependent varia-
ble. Concerning electoral system choice during the first few decades of the
twentieth century, the most essential feature was the change from majoritarian to
proportional systems in several Western European countries. These changes are
well documented in the literature and explanations have been provided. The
rational perspective of this study is to a large extent based upon these findings.
The frequency of newly independent countries increased rapidly as of 1945,
having previously been quite low. Generally speaking, the end of the Second
World War meant the beginning of a new era in practically every field. Taking
these circumstances into consideration, it appears natural to focus on the post-­
war era, thereby following a rather common approach in macro-­comparative
political science.

Does electoral system choice matter?

Popular elections are the defining institution of modern democracy. The purpose
of parliamentary elections is, first, to decide who will represent each constitu-
ency in the legislature, and, second, what the overall composition of the legislat-
ure by political party will be. By translating votes into seats, these decisions are
managed by the particular electoral system used for each single election. Hence,
the electoral method is a key factor in the political process: it largely determines
who gets what, when and how. Electoral rules are commonly assumed to con-
dition the chances of success of political parties and candidates in competitive
elections.
4   Introduction
Within the range of democratic institutions, several experts assert that the
choice of electoral system is the most important one. Arend Lijphart (1994: 1),
for instance, has described the electoral system as the most fundamental element
of representative democracy. Furthermore, the electoral system is regarded by
many as the principal key to reforming the political system. Electoral rules are
easier to change than most other features of the political system, not least
because the electoral law is quite rarely included in the constitution. Referring to
Giovanni Sartori, the electoral system is ‘the most specific manipulative instru-
ment of politics’ (1968: 273). To be sure, most of the theory of electoral systems
presupposes a democratic form of government – however, the choice of electoral
systems in non-­democracies matters as well. For one thing, all countries were
non-­democratic when the first electoral provisions were introduced. Further-
more, the electoral method plays an important role in consolidating a democratic
system. The institutional framework of a non-­democratic nation may have a
decisive impact on the development towards democracy.
As mentioned, the bulk of the electoral system research has dealt with polit-
ical consequences of elections conducted under different electoral rules, while
much less attention has been given to the understanding of cross-­national vari-
ation in electoral systems. In a critical survey of electoral system research in the
mid-­1980s, Lijphart (1985b) maintained that there was a shortage of studies of
electoral system choice – in a volume published two decades later, Matthew S.
Shugart (2005: 51) points out that this is still the case today, although some
progress has been made. The purpose of writing this book is to fill part of this
gap in the field of electoral system research. A great deal of the scientific discus-
sion on the advantages and disadvantages of electoral systems has been con-
cerned with the question of which electoral system should be chosen in each
given context. It emphasizes the political outcomes of electoral system design as
the principal criterion for choosing between different systems. If we reformulate
the question and ask which system is likely to be chosen, we recognize that
electoral systems are chosen in the context of existing institutional and cultural
settings. The contextual framework shapes the preferences of political actors as
to which electoral system is to be chosen, and constrains their negotiating strat-
egies over alternative rule configurations. Obviously, both actors and structures
matter in the process of choosing electoral systems. Explanations based on stra-
tegic calculations of political actors have been provided by, for instance, Carles
Boix (1999), Josep M. Colomer (2004a) and Lijphart (1992a). The present study,
in contrast, aims at explaining electoral system choice from a contextual
perspective.
2 Research design

2.1 Methodological considerations

Basically, as the title suggests, this is a comparative study. Explanations of elect-
oral system choice are sought for by comparing countries with different electoral
systems. More precisely, associations between variables, and the mutual strength
of explanatory variables, are studied by means of statistical analysis. There are
two fundamental strategies of arriving at law-­like generalizations about the rela-
tionship between two or several social phenomena: deduction and induction.
Deduction means that we depart from a theoretical framework in the research
field. On the basis of the theory, we derive hypotheses which are tested and
either confirmed or rejected. As a result, the theory is developed, rejected or
modified. When applying the inductive strategy, in contrast, we aim at detecting
causal mechanisms by observing the empirical reality. The research process
begins with charting the variation of the dependent variable, whereas the deduc-
tive strategy starts with selecting cases that show a variation in the explanatory
variables.
When choosing between a deductive and an inductive approach, the amount
of previous research in the area is decisive. If a comprehensive theory already
exists, the deductive strategy should be applied. The inductive approach is pref-
erable when such a theory is yet to be formulated. However, empirical studies
are rarely totally deductive or inductive. In case of the latter, there is always an
element of deduction involved. More exactly, although we lack a well-­founded
theoretical framework, we nevertheless have an idea of which settings to look
into in order to find potential determinants of the phenomenon that we are study-
ing. Accordingly, in the lack of a comprehensive theory of electoral system
choice, this is primarily an inductive study. However, the independent variables
are derived by using deduction. To elucidate, in the search of relevant variables,
the present study makes use of previous attempts at, theoretically or empirically,
explaining electoral system choice. The independent variables emanate from
three broad settings: a rational, a historical and cultural, and an institutional
context.
Two different operationalizations of the dependent variable are constructed in
the following chapter: a detailed categorization of 15 electoral formulas and a
6   Research design
broad typology of four categories of electoral systems (plurality, majority, pro-
portional and mixed systems). The latter is subject to statistical analysis. With
regard to proportionality, the typology resembles an ordinal classification: mixed
systems are more proportional than plurality/majority systems, and proportional
systems are more proportional than mixed systems. In a strict sense, however,
the classification is not ordinal but categorical. For one thing, every plurality/
majority system does not produce disproportional results, and all proportional
systems do not result in proportional seat allocation. Furthermore, the inherent
proportionality of PR systems and non-­proportional nature of majoritarian
systems should be regarded as effects rather than characteristics of electoral
systems. As a consequence, despite the common classification of electoral
systems approximate to a scale of different degrees of proportionality, the four-­
scale classification is nonetheless categorical. Furthermore, definitions of the cat-
egories do not meet the criteria for a pure ordinal scale. Logistic regression is
therefore preferred to OLS (ordinary least squares) regression. Since the depend-
ent variable consists of four values, multinomial logistic regression is primarily
applied. The multiple analyses are preceded by descriptive and bivariate analy-
ses in which simpler statistical methods are used. Due to the different nature of
the explanatory variables, the classification and identification of electoral
systems varies from case to case; some independent variables require a more
detailed categorization, whereas the four-­scale classification constitutes the
logical basis of comparison in other contexts.
Data on electoral systems and the explanatory variables is collected from a
variety of sources. Regarding the dependent variable, the main sources are Elec-
tions in Africa (Nohlen et al. 1999), both volumes of Elections in Asia and the
Pacific (Nohlen et al. 2001a; 2001b), Handbuch der Wahldaten Lateinamerikas
und der Karibik (Nohlen 1993b), ‘Mixed Electoral Systems: A Conceptual and
Empirical Survey’ (Massicotte and Blais 1999) and different volumes in the
series Chronicle of Parliamentary Elections, provided by the Inter-­Parliamentary
Union, as well as the organization’s website, Parline Database on National
Parliaments.
In addition to the sources above, data on the independent variables has been
gathered from The International Almanac of Electoral History (Mackie and Rose
1974), Britannica Book of the Year (1999), Svensk uppslagsbok (different edi-
tions 1947–1955), The World in Figures (editions 1981 and 1987), The World
Factbook (CIA website), Statsskick: att bygga demokrati (Karvonen 2003) and
Political Systems of the World (Derbyshire and Derbyshire 1989; 1999).

2.2 The research population

The unit of analysis is electoral system choice. All choices of electoral systems
in independent countries between 1945 and 2007 are observed, irrespective of
the way in which they have been adopted. This comprises the first choice of
electoral systems in countries that have declared independence during the rele-
vant period as well as every major electoral system change in countries that pre-
 Research design   7
viously have had electoral laws as independent nations. There are, however,
several countries in which electoral provisions existed and elections had been
held before independence. A common feature of these countries is that they have
been under some kind of foreign rule. It means that the electoral system had
already been chosen when the country became independent. The study consists
of independent countries only – autonomous territories that may become inde-
pendent in the future are not paid attention to.
The apparent question is whether I should observe the year of electoral
system choice for each country irrespective of whether independence had been
attained at that point in time or not? Considering that there are several autono-
mous regions with electoral institutions in the world, the answer is no. Several of
these territories may become independent in the future but they are, nonetheless,
excluded from the study since it deals, as mentioned, exclusively with independ-
ent countries. However, an electoral system choice that was made a few years
before independence is not less important than a choice that was made shortly
after independence was proclaimed. Therefore, electoral system choices prior to
independence cannot be disregarded.
The dilemma is to decide the timing of electoral system choice in countries
that already had electoral provisions when independence was declared. It seems
reasonable to consider the year of independence as the year of electoral system
choice. In most cases, the introduction of pre-­independence electoral provisions
formed a part of the independence process. However, there are some exceptions
to this rule in the Caribbean and the Pacific. Antigua and Barbuda, St Lucia, and
St Kitts and Nevis, for example, held elections several decades before finally
receiving independence.
The first general elections in Antigua and Barbuda took place in 1951; that is,
30 years before the archipelago became independent (Hillebrands 1993: 22).1
Yet, the coded year of electoral system choice in this study is 1981. If 1951 is
considered as the official year of choosing an electoral system, we may ask why
other non-­independent territories in which general elections have been held are
excluded from the analysis. The obvious answer is, as already stated, that the
study deals with electoral systems in independent countries. When it comes
down to it, the country in question had the opportunity to reject the old electoral
system and adopt another when independence was achieved. As for countries
without pre-­independence electoral institutions, the timing of the relevant elect-
oral laws determines the coding of electoral system choice. If information on the
date of the electoral law is missing, the year of the first parliamentary elections
under the new electoral system is decisive. Another reason to consider the timing
of independence as the year of electoral system choice is that information on the
first pre-­independence elections is often inadequate. Besides, if the first elections
were decisive, only cases of electoral system change would fulfil the criteria of
democracy stipulated below.
Election results in democratic countries are usually regarded as more import-
ant than electoral outcomes in non-­democratic countries. Therefore, electoral
system choice is of greater scientific interest in countries that allow multi-­party
8   Research design
competition. If all candidates in an election belong to the same political party,
seat allocation is only concerned with who is elected and who is not. A propor-
tional system is not needed, since no seat allocation between political parties
takes place. However, all non-­democratic countries are not authoritarian one-­
party states – several countries that do not hold democratic standards allow some
competition. Consequently, majoritarian systems must not be regarded as uni-
versally prevailing in non-­democratic countries.
In order to distinguish between democratic and non-­democratic countries, we
need a measurement of the level of democracy. Democracy is usually regarded
as a more-­or-less rather than an either–or phenomenon. Democracy is associated
with several elements, or requirements, which a country may fulfil to varying
extent. Consequently, no absolute dividing line between democracy and non-­
democracy exists. However, a classification based on two categories is far from
inconceivable.
The most widely used source for determining the level of democracy is
Freedom House’s yearly survey of political rights and civil liberties. These two
general sets of characteristics constitute the overall concept of freedom in the
world. The description of political rights and civil liberties also represents the
overall definition of democracy in this study. For each of the two dimensions, a
scale ranging from 1 to 7 is applied, with 1 representing the most free and 7 the
least free. The concept of political rights implies that the people are allowed to
participate freely in the political process. Countries with low level of political
rights also score low points on civil liberties, and vice versa. Indicators of civil
liberties are: freedom of opinion, freedom to develop institutions and personal
autonomy without state involvement. Needless to say, free and fair political
competition cannot exist without a substantial degree of freedom and human
rights.
Within the framework of the present study, the dimension of political rights is
more relevant, because it deals more with elections than the other dimension.
Scores on political rights are therefore decisive in distinguishing democracies
from non-­democracies. Countries that score 1 or 2 must without doubt be
included in the democratic sample. The exclusion of countries with the values 6
and 7 is also self-­evident. Hence, the borderline between democracy and non-­
democracy is in the interval 3–5. According to the explanations of political rights
in the 2000–2001 Freedom House volume, damaging elements in countries
receiving these values are, for instance, civil war, heavy military involvement in
politics and unfair elections. On the other hand, some democratic elements such
as the right to organize non-­governmental parties and quasi-­political groups,
partly free referenda, or other forms of citizen influence on the government may
exist (Freedom in the World 2001: 652).
In order to minimize the risk of including countries with serious damaging
elements, I set the limit between 3 and 4. A further criterion for inclusion in the
democratic sample is that the value 3 or lower has been received every year
since the last elections prior to the observed year, that is, the year in which the
electoral system was adopted. This criterion guarantees that the electoral system
 Research design   9
has been chosen by a democratically elected leadership. For example, the intro-
duction of list PR in Poland in 1991 did not take place under democratic circum-
stances, because the previous elections in 1989 resulted in the value 4 on
political rights in Freedom House’s ratings. In contrast, the adoption of list PR in
Bulgaria in 1991 is regarded as a democratic choice, since the country had
already reached the required level of democracy as a consequence of the parlia-
mentary elections held the previous year.
However, Freedom House’s ratings cover the time period from 1972 onwards,
which implies that other sources are needed for a classification of countries
between 1945 and 1972. For this purpose, I shall partly rely on the Polyarchy
Dataset, which was created by Tatu Vanhanen and, at the time of data compila-
tion, was available on-­line (Polyarchy Dataset). The dataset provides indices of
democracy in 187 countries over the period 1810–1998. The index of democracy
is the same as Vanhanen has used in a number of studies of democratization (e.g.
Vanhanen 1984; 1990). The index is based on two components: competition and
participation. The smaller parties’ share of the votes cast, that is, all parties but
the largest one, in parliamentary or presidential elections, or both, represents the
degree of competition. The percentage of the total population who voted in these
elections represents the degree of participation.2 The index of democracy (ID) is
calculated by multiplying these two components and dividing the product by
100. High values on both competition and participation produce high values, and
low values on one of the two, or both, result in a low level of democracy (Van-
hanen 1984: 28–32). The lowest possible value is 0; a maximum value is not
defined. In practice, the most democratic countries exceed ID 40, but not 50.
A prerequisite of using Vanhanen’s index as a complementary measure of
democracy is that it corresponds to the ratings by Freedom House. The later
volume mentioned above, Vanhanen (1990: 25) compares all three indicators
(i.e. competition, participation and ID) with the arithmetic means of political
rights and civil liberties in Freedom House (also called Gastil’s ratings) over the
period 1980–1987. Vanhanen’s indices turn out to be strongly correlated with
Freedom House’s combined measure of democracy. The weakest and strongest
correlation coefficients are –0.811 and –0.902 respectively.3 In my study,
however, only the dimension of political rights is paid attention to. Nonetheless,
a comparison between Vanhanen’s ID and Freedom House’s political rights
from 1980–1988 produces similar results: the correlation coefficients vary
between –0.836 and –0.865.
In order to separate democracies from non-­democracies in the Polyarchy
Dataset, a threshold value is needed. Vanhanen selects 5.0 index points as a
threshold value. In addition, the author stipulates two further criteria: the value of
participation must not be lower than 10 per cent, whereas the threshold of com-
petition is 30 per cent. Accordingly, countries that fulfil these criteria are regarded
as crossing the threshold of democracy. Do these criteria correspond to the thresh-
old value of 3 on political rights? The question can easily be explored by con-
structing diagrams in which the values of both scales are plotted for each year
between 1980 and 1988. An analysis of such diagrams results in the following
10   Research design
conclusions. First, the threshold of ID 5.0 is too lenient compared to the thresh-
old of 3 on political rights. Each year, between five and 15 countries receive
higher ID values than 5.0, although they do not meet the criterion of 3 or lower
on political rights. Second, we cannot be completely sure that countries scoring
lower than 5.0 on Vanhanen’s scale also get higher values than 3 (which indicate
a low level of democracy) on political rights. Every year from 1980 to 1988, a
few countries verify this conclusion.
Despite high correlation coefficients between Vanhanen’s ID and Freedom
House’s political rights, we cannot rely on any specific threshold value when
distinguishing democracies from non-­democracies before 1972. However, the
Polyarchy Dataset is not useless. It can be used as a guideline for evaluating
whether a country has been democratic or not at the relevant point in time. Fur-
thermore, in order to get a better understanding of the politics and the state of
democracy in every single case, I have used the historical overviews in Hand-
buch der Wahldaten Lateinamerikas und der Karibik (Nohlen 1993b) and the
volumes of the Elections Worldwide series (Nohlen et al. 1999; Nohlen et al.
2001a; 2001b). The colonial background also matters in this respect, since we
know that Britain established democratic institutions in large parts of its colonial
empire, whereas other colonial powers paid less attention to indigenous repre-
sentative institutions (Diamond and Lipset 1995).
A further dilemma is to decide whether a country that has chosen its electoral
system before independence is democratic or non-­democratic. Neither Freedom
House nor Vanhanen’s dataset provides information on the level of democracy
prior to independence. Again, I have to make a general estimation of the con-
ditions in each case by means of the sources mentioned above. I shall use the
cases of Mauritius and Antigua and Barbuda to illustrate this dilemma. Mauritius
became independent in 1968, one year after the constitution and the electoral law
was enacted and the first general elections were held (Krennerich 1999: 603–22).
This implies that the decision on the electoral system (BV) was taken under non-­
democratic circumstances, although the elections fulfilled democratic standards
and Mauritius was democratic when independence was received. Therefore,
Mauritius’ choice of electoral system in 1968 is coded as non-­democratic.
However, Antigua and Barbuda’s electoral system choice in 1981 (as coded) is
considered democratic, even though the polity could not possibly be democratic
when the plurality system was introduced prior to the first elections in 1951. The
critical point is under which circumstances, in democratic terms, the electoral
law and/or constitution of the newly independent country is written. When the
electoral law and the constitution of independent Antigua and Barbuda were
adopted, the country already had some experience of democracy. Mauritius, by
contrast was not democratic when its electoral law and constitution that paved
the way for independence were introduced. Countries that were democratic
before the Second World War and continued functioning as democracies imme-
diately after that are regarded as democratic, despite the possible absence of
democratic politics during the war.
 Research design   11
2.3 Previous attempts at broad theoretical explanations
There is no comprehensive theoretical framework of how countries adopt elect-
oral systems. Yet, some explanatory models and several possible determinants
have been suggested. Sub-­sets of theories and assumptions are dealt with in con-
nection to each independent variable of the present study. The purpose of this
section is to present some theoretical models and broad empirical studies of
determinants of electoral system choice, and to discuss some dubious explana-
tory factors put forward in the literature.
Ben Reilly and Andrew Reynolds (1999) maintain that there are four basic
ways in which electoral systems are introduced: through conscious design, via
colonial heritage, by external imposition and through accidental evolution. From
a historical perspective, they claim, it is quite unusual that electoral systems are
consciously and deliberately chosen. Too often, the choice of electoral system is
a consequence of a passing trend, a quirk of history or a combination of circum-
stances. The earliest attempts at electoral engineering occurred in continental
Europe a century ago when majoritarian systems were replaced by proportional
ones in order meet the demand for fair citizen representation in culturally diverse
societies. In contrast, the decades of decolonization after the Second World War
witnessed few cases of deliberate electoral engineering. Reilly and Reynolds
point out that a large number of ethnically plural countries in Africa and Asia
inherited misplaced electoral systems from the colonial powers, thereby contrib-
uting to the second ‘reverse wave’ of democracy. In fact, the authors regard
colonial inheritance as probably the most common way in which democratizing
countries adopt electoral provisions. However, a revival of electoral system
design has taken place during the third wave of democratization. Several coun-
tries with deep ethnic, linguistic and religious cleavages have been subject to
electoral engineering as a means of mitigating conflict. A few electoral systems
have been deliberately designed and imposed by external powers, the mixed-­
member proportional system in West Germany after the Second World War
being the foremost example (Reilly and Reynolds 1999; see also Reynolds and
Reilly 1997; Reynolds et al. 2005).
In a study of electoral system design in post-­Communist Europe, Sarah Birch
et al. (2002) present four different approaches to explaining electoral system
choice: historical factors, foreign influences, contextual factors and interest-­
based calculations. The object of the study is one specific type of electoral
system choice, namely electoral reform, and their frame of reference is, con-
sequently, determined by these circumstances. Notwithstanding, their theoretical
framework also bears relevance to the study of electoral system choice in
general.
The author team maintains that the extent of influence of history largely
depends on three factors. First, if the nature of historical experience is positive,
it may well be reflected in contemporary institutional design. Second, if the his-
torical legacy is not very distant, it is more likely to be embraced. Third, if the
decision-­makers need to make a decision rapidly, a previous model might turn
12   Research design
out to be the easiest acceptable alternative. The extent of foreign influence also
depends on several circumstances. For instance, if there are few domestic experts
on electoral system design, there may be a greater temptation to look at electoral
arrangements elsewhere. Cultural affinity is another factor that makes countries
more prone to adopt similar electoral systems. As to contextual factors, the
authors intend that electoral systems should be designed in accordance with the
social and political context in which they operate. Ethnic diversity, for instance,
calls for proportional representation in order to secure minority interests. The
overall institutional framework is another factor that may influence electoral
system design. Economic considerations are also regarded as relevant to elect-
oral system choice in some circumstances.
The authors pay most attention to interest-­based calculations. Political actors
are often assumed to have self-­interested goals in addition to their obligation
towards promoting the collective good. By referring to Rokkan (1970), they
maintain that major electoral reforms occur when an alternative system is likely
to produce a larger seat share for one or several parties that have the power to
implement an electoral reform. A difference is made between strategic decision-­
making at the founding stage and post-­zero-stage bargaining. A large degree of
uncertainty about electoral outcomes prevails when a new state emerges or when
a transitional period is in progress. After the first elections under the new elect-
oral system, the strategic context is altered: uncertainty decreases, actors become
better aware of how to pursue their interests, and successful contestants become
institutionally embedded in the legislative structures. Another electoral reform
may well take place between these two stages, but thereafter, electoral systems
are expected to become ‘sticky’ (Birch et al. 2002: 9–22).
In an article titled ‘Electoral Formulas: A Macroscopic Perspective’, André
Blais and Louis Massicotte (1997) examine which formulas are the most wide-
spread, and whether the prevalence of a formula correlates with geographical, his-
torical, economic and political factors. Their analysis covers all the sovereign
countries of the world that had a functioning, directly elected parliament in Decem-
ber 1995. Freedom House’s ratings of political rights are used for determining the
level of democracy. In addition to a separate analysis of democracies, the authors
distinguish between old and new democracies. The criterion for being regarded as
a democracy is to score 1 or 2 on political rights in 1994. In order to qualify as an
old democracy, a country had to meet this criterion ten years in a row.
In addition to the level of democracy, the following independent variables are
analysed: continent, size, the level of economic development and colonial legacy.
As for economic conditions, no direct link between economic development and the
choice of an electoral formula is expected. Among the colonial powers, only
Britain and France are included, because, first, these countries possessed consider-
ably more colonies than other colonizers, and, second, both Britain and France
established representative institutions in their colonies prior to independence. Con-
cerning the dependent variable, the authors distinguish between plurality, majority,
PR and mixed electoral systems. Plurality and PR systems are approximately
equally popular; 70 per cent of all countries use one or the other. Majority and
 Research design   13
mixed systems are both applied in 25 countries. A slightly different pattern
emerges when only democratic countries are studied: PR is more frequently used
than plurality systems, and mixed systems surpass majority ones.
The analysis indicates that four sets of factors influence electoral system
choice. British colonial background is the most important explanation: their data
suggest that if there had been no British influence, only 36 instead of 59 coun-
tries would apply the plurality system. By contrast, the influence of French colo-
nial rule is weak. The majority system, which France has applied most
frequently, is not significantly more common than other systems among former
French colonies.
Continent is also of considerable importance. Three groups of continents are
differentiated: first, North America, Africa, Asia and Oceania, where plurality
systems prevail; second, Europe, where PR is most frequent; and third, South
America, where PR is predominant. The degree of democracy is the third factor
that seems to be associated with the choice of an electoral formula: a high level
of democracy correlates with the use of PR. Blais and Massicotte suggest that
the idea of proportional seat allocation among parties often corresponds to the
idea of equally weighed vote of all citizens. The last factor related to electoral
arrangements is territory size: the larger the country, the more likely it is to have
single-­member districts. They maintain that a vast country needs single-­member
districts in order to provide close contact between representatives and their con-
stituents. As expected, no association is observed between the level of economic
development and the dependent variable. Neither is there any difference between
old and new democracies. These results are based on an analysis of all countries.
Rather similar patterns emerge when only democratic countries are analysed
(Blais and Massicotte 1997: 107–29).
In an article titled ‘Trade and Democratic Institutions’, Ronald Rogowski
(1987: 203–25) analyses the influence of economic conditions on the choice of
electoral systems. The author says that there is a natural affinity between trade
and proportional electoral systems: the more an economically advanced state
relies on external trade, the more it will be inclined to adopt a proportional
system with large districts. In addition, trade-­dependent states are likely to intro-
duce a parliamentary system rather than presidentialism. Insulation, autonomy
and stability constitute intermediate variables in Rogowski’s explanatory model.
Advanced economies that are largely dependent on international trade are likely
to experience strong pressures for democratic participation. Therefore, it is
important to adopt democratic institutions that maximize the country’s insula-
tion, autonomy and stability. These factors are, for a number of reasons, best
promoted by proportional representation. First, insulation from regional and sec-
toral pressure is most easily achieved with large districts. Second, autonomy is
best achieved by strong parties, which are furthered by list PR. Third, stability is
best promoted by PR, large districts and a parliamentary system.
Based on an empirical analysis of 24 countries belonging to the Organisation
for Economic Co-­operation and Development (OECD), Rogowski concludes
that trade-­dependent states tend to have both PR and large districts. However,
14   Research design
the author concedes that the association does not reflect dynamic and historical
processes; PR was not adopted in European countries because of increased
dependence on trade. Still, Rogowski asserts that in an age of increasing depend-
ence on trade, states will be increasingly constrained to adopt institutions condu-
cive to openness and to effective competition on world markets. In other words,
trade-­dependence is proposed as a fundamental factor for an increased domi-
nance of PR in the future. However, partly because Blais and Massicotte (1997)
concluded that the level of economic development does not affect the choice of
electoral systems, I shall not include economic conditions in my analysis.
Besides, they did not even expect any particular relationship between economic
development and electoral system choice.
Some authors maintain that there is a link between the level of democracy
and the choice of an electoral system. In Blais and Massicotte’s study, for
example, the level of democracy is one of five independent variables. The
hypothesis that PR is associated with a higher level of democracy than other
systems is confirmed. They argue that less democratic countries are less con-
cerned about the representation of minority groups, and therefore less willing to
adopt PR systems. Proportional representation coincides with the idea of demo-
cracy in contrast to the winner-­take-all feature of majoritarian systems (Blais and
Massicotte 1997).
However, the higher frequency of PR systems among democratic countries
should not come as a surprise at all, since authoritarian one-­party states do not
need proportional representation. There are, hence, natural explanations of the
discovered pattern. Most authoritarian states that conduct general elections apply
majoritarian systems, thereby contributing to the positive correlation between
democracy and PR. There are other reasons, too, for rejecting the level of demo-
cracy as a possible determinant of electoral system choice. The direction of cau-
sality, for instance, constitutes a problem. In order to reach a certain level of
democracy, general elections need to be held, which in turn requires electoral
provisions. In other words, electoral system choice precedes the level of demo-
cracy. To be sure, in cases of electoral system change, there is a measurable level
of democracy. Notwithstanding, the presence/absence of universal suffrage and
the right for/prohibition on opposition parties to contest elections represent
primary conditions that may determine the choice of a particular system. If the
level of democracy bears some relevance to the choice of electoral systems, dif-
ferences should be explained by these two factors. A century ago, extension of
universal suffrage contributed to electoral reforms in favour of proportional
arrangements, whereas introduction of competitive politics is assumed to be a
relevant factor during the latter half of the twentieth century. Consequently, I am
inclined to assert that the extension of political rights, of which introduction of
universal suffrage and competitive politics are two cornerstones, captures the
essence of the causality between democracy and electoral system choice.
Birch et al. (2002: 14) also regard PR as a more appropriate system than plu-
rality or majority in a democratizing country. Open debate and a concern for
women’s representation are further reasons why democratic countries are likely
 Research design   15
to adopt inclusive electoral arrangements. When the constitutional basis for a
new regime is established, they say, electoral provisions that provide representa-
tion of as many groups as possible are preferable – democratization is better fur-
thered by PR than majoritarian systems. However, their argumentation actually
reflects the viewpoint that democratization is a consequence rather than a deter-
minant of electoral rules.
The role of democracy in choosing electoral systems is better understood in
an article written by Blais, Agnieszka Dobrzynska and Indridi H. Indridason
(2005). The authors maintain that two factors were particularly important in
influencing the adoption of PR at the turn of the twentieth century: the spread of
democratic ideas and the presence of a majority system, in contrast to a plurality
system. They argue that PR had come to be regarded as the most ‘democratic’
electoral system during the analysed period (1865–1938), and, therefore, suggest
that the push for PR was stronger in an environment in which democratic norms
were more widespread. However, the normative appeal of PR was not enough.
Another crucial factor is whether the choice was between plurality and PR or
majority and PR. There are several factors that make PR more attractive when
elections are conducted according to the majority rule. First, the incentives to
vote strategically are smaller in a majority system, because it requires more
complex information on, for example, the identity of the third runner-­up and
how the voters of the unsuccessful candidates will split their votes between the
front-­runners. As a consequence, the number of parties tend to be higher in
majority systems than in plurality systems, which in turn leads to a higher fre-
quency of multi-­party governments. The smaller coalition parties are likely to
advocate a change to proportional elections, because they are electorally disad-
vantaged by the disproportional electoral system. The major parties are likely to
resist these efforts to a lesser extent in a majority system than in a plurality
system, since they already face the necessity of forming government coalitions.
All these circumstances, by reinforcing one another, suggest that PR is more
likely to be adopted if majority elections have previously been held.
The unit of analysis is each legislative term during which PR could have been
introduced. Naturally, the former electoral system must have been held under
some other electoral system than PR, and furthermore under democratic circum-
stances. The sample consists of 18 countries and 183 legislative terms. With a
brief reference to some previous attempts to explain electoral system choice, the
authors also include in their model socialist threat (the interactive term of the
strength of socialism and the effective number of non-­socialist parties), the log
of population, and the log of the number of years since the earliest observation.
The findings are in line with the assumptions. Both substantive variables, a
democratic environment and a majority system, have a statistically significant
effect on the propensity of PR reform. Concerning majority systems, however,
the results are significant only at the 0.1 level. The adoption of PR was also more
plausible in smaller countries. The hypothesis that PR was more frequently
adopted in countries where the right-­wing parties faced a serious socialist threat
is not confirmed. The variables of interest in the article are not included in my
16   Research design
study, because the author team focuses exclusively on the adoption of PR, which
in turn requires a different research design than the one applied in this book
(Blais et al. 2005: 182–90).

2.4 A frame of reference

In an essay that aims at explaining why some countries resort to some basic
forms of constitutional amendment methods, whereas other countries apply other
types, Dag Anckar and Lauri Karvonen (2004) use a theoretical frame of refer-
ence consisting of three different perspectives. The first approach regards polit-
ical institutions as problem-­solvers, thereby constituting a rational point of
departure. It maintains that countries adopt institutions that correspond to the
particular needs of their societies. An institution is not an end in itself – there is
a specific purpose for its existence. Societies with different sociopolitical con-
ditions have different needs; therefore, they adopt different political institutions.
In other words, the specific structure of a society is associated with certain fea-
tures and problems, which the adopted institutions should meet and solve.
The second perspective sees institutional features as reflections of the cultural
and historical context of which they are part. This perspective embodies colonial
legacy as well as other patterns of imitation. In addition, the timing of independ-
ence of countries is regarded as a possible determinant. Institutional choice may,
in other words, be an epoch phenomenon, associated with certain periods and
eras, in the same way as the spread of democracy is said to evolve in ‘waves’
(Huntington 1991). The third approach focuses on the overall constitutional
design, that is, the influence of other political institutions on institutional choice.
It maintains that institutions are dependent on each other; the choice of one
device leads to the choice of another device.
The theoretical frame of reference in this book follows the same structure as
that of Anckar and Karvonen (2004). First, rational explanations are dealt with.
Second, electoral system choice is studied from a cultural and historical perspec-
tive. Third, the influence of other political institutions is examined. It should be
observed that the explanatory variables are of varying dignity, largely because of
varying distance between independent and dependent variables. The temporal
diffusion variable will shed some light on which systems have been the most
popular at various points in time. As an explanation of electoral system choice,
however, trend is unquestionably inferior to, for example, colonial legacy, which
constitutes a more profound explanation of the adoption of a particular electoral
system. Variables that belong to the institutional perspective are likewise of less
importance than, for instance, those that belong to the rational perspective.
Several institutional choices are often made at the same time, put in concrete
form by adopting a new constitution, thereby making the distance between the
choice of electoral system and the choice of other institutions practically non-­
existent. In this context, however, the choice of a particular system is regarded
as dependent on the presence or the adoption of other specific political
institutions.
3 Electoral systems

The third chapter deals with electoral systems and the classification of different
voting procedures in legislative elections. To begin with, a description of all rel-
evant electoral formulas is provided, and some classifications in the literature on
electoral systems are presented. Thereafter, the typology of electoral systems in
this study is constructed. Finally, a descriptive survey of electoral system choice
during the relevant time period is conducted.
Before describing different voting procedures, we need to distinguish between
electoral laws and electoral systems. Electoral laws are the set of rules governing
the whole process of elections, including the official calling of elections, suf-
frage and registration requirements, candidate selection, campaigning, voting
procedures and the final determining of election results. Among these electoral
laws, the electoral system is concerned with the final stage of the election
process: how citizens vote, the style of the ballot paper, constituency structure,
the method of counting, and the final determination of who is elected. David M.
Farrell presents the following definition: ‘Electoral systems determine the means
by which votes are translated into seats in the process of electing politicians into
office’ (1998: 5). Blais defines electoral systems as ‘those rules which govern
the processes by which preferences are articulated as votes and by which these
votes are translated into the election of decision-­makers’ (1988: 100).
Some words concerning the use of the terms electoral formula and electoral
systems also need to be said. In the literature on electoral systems, the term elect-
oral system is frequently used when only one of the three basic elements of an
electoral system, namely the electoral formula, is intended. The incorrect usage
of the term electoral system is perhaps due to the common classification of elect-
oral systems on the basis of formula. Another reason might be that some formu-
las, particularly the single-­member plurality formula, are widely used in
comparison with other formulas of the same electoral system family, and tend to
represent the entire electoral system in question. In this study, the term electoral
formula is used with reference to those electoral methods that are numbered in
Table 3.1. Regarding the broad classification into four categories, electoral
system is the appropriate term. Electoral system is also used when the relation-
ship between vote and seat shares is in focus, that is, when a distinction between
plurality/majority, semi-­proportional and proportional systems is made. Plurality
18   Electoral systems
and majority systems are, moreover, referred to as majoritarian systems in cases
where no principal difference between these two is intended. Plurality and
majority systems are often regarded as the main alternatives to proportional
systems, and, as a consequence, it comes natural to use a common term for the
former two.
There are two exceptions to the application of formula and system described
above. First, the category of list PR contains a large number of countries with
different forms of proportional formulas and combinations of other elements,
and no list PR systems are exactly alike. List PR is therefore referred to as one
form of electoral system rather than a distinct formula. Second, mixed systems
are per definition combinations of different formulas, each mixture representing
an electoral system in itself.

3.1 Electoral formulas and systems – a presentation

The electoral formula determines how votes are translated into seats. There are,
broadly speaking, three methods by which seats can be allocated to candidates or
political parties. Seats can be allocated to a candidate or a party winning a plu-
rality of the vote, a majority of the vote, or proportionally among the competi-
tors. Thus, we can distinguish between plurality formulas, majority formulas and
proportional formulas. In addition, some electoral systems combine the propor-
tional formula with the plurality or the majority formula in one election. These
are called mixed electoral systems. Despite the large range of voting procedures,
all electoral formulas constitute subtypes of these four categories. This study
makes use of the traditional approach in classifying electoral systems, that is,
according to electoral formula.
In a plurality system, the candidate or party with the largest number of votes
wins, irrespective of the proportion of the total vote. In contrast to majority
systems, it is possible to win an election without winning a majority of the votes.
The most common plurality formula is the single-­member plurality formula, also
called, among other things, simple majority, relative majority and first-­past-the-­
post. In the present study, the term single-­member plurality (SMP) is used,
because the term comprises the main features of the formula, and distinguishes it
from other plurality formulas. Only single-­member constituencies are used,
voters are given one vote each, and the candidate or party with a plurality of the
total vote is elected.
Block vote (BV) is simply the application of the plurality formula in multi-­
member districts. The electors vote for individual candidates (or parties), the
number of votes per elector corresponds to the district magnitude, and the
winners are those candidates that get a plurality of the votes in each district.
Some countries use both single-­member and multi-­member districts in plurality
elections – this method is referred to as SMP–BV. Yet, most authors ignore the
distinction between plurality in single-­member districts and the combination of
plurality rule in single- and multi-­member districts; simply regarding them all as
plurality systems. However, I consider it necessary to treat the mixture of plural-
 Electoral systems   19
ity formulas as a category of its own. The detailed classification of electoral for-
mulas and systems serves partially an informative and a descriptive purpose. BV
is therefore separated from the very popular SMP, which is frequently praised
and criticized for its merits and demerits. Because plurality elections in single-­
member and multi-­member districts are separated from each other, elections that
use both fit into neither of them. A further criterion of SMP–BV is that at least
10 per cent of the representatives are elected under a formula different from the
one used for electing the other representatives.
The limited vote (LV) is a plurality formula that uses multi-­member districts,
in which the voter has fewer votes than there are representatives to be elected.
When each voter has only one vote, this method is called the single non-­
transferable vote (SNTV). During the second half of the last century, no national
legislative (lower) chamber has been elected by the limited vote with more than
one vote per elector. SNTV, by contrast, has been applied in a few countries.
Because LV most often results in a higher degree of proportionality than the
majority systems and the plurality formulas presented above, it is usually
referred to as a semi-­proportional electoral method. Notwithstanding, it belongs
to the family of plurality systems, since the candidates with the most votes are
elected. The cumulative vote (CV) is another plurality formula considered as
semi-­proportional. CV implies that the voters are granted as many votes as there
are members to be elected, but in contrast to BV, the voters are allowed to cast
two or more votes for a single candidate. This book does not include any empiri-
cal cases of this formula. Certainly, the first legislature of independent Barbados
(1966–1971) was elected under CV but the elections were conducted 27 days
prior to the declaration of independence, and on the independence day, SMP was
adopted. Consequently, in a strict sense, CV has not been adopted by any nation
during the second half of the twentieth century.
In order to be elected under a majority system, a candidate must obtain at
least 50 per cent of the total vote. There are basically two kinds of majority elec-
tions: the alternative vote and the double-­ballot, also called majority run-­off,
two-­ballot, the second ballot and the two-­round system. In a few countries, the
electoral law requires a majority for victory, although no second round is con-
ducted. Hence, it is obvious that this method is only applied in countries with
non-­competitive elections. In Cuba, for instance, a candidate must receive more
than 50 per cent of the valid votes cast in a constituency. If none of the candid-
ates receives a majority, the seat in question remains vacant. In practice,
however, there is only one candidate in each constituency. Within the framework
of the present study, the distinction between majority elections in one round and
two rounds is of no considerable importance but, rather, noticed on grounds of
preciseness; a one-­round majority election is not a run-­off election, and defi-
nitely not an alternative vote election. For the purpose of using a consequent ter-
minology, I call this formula one-­ballot majority (OBM), and, accordingly, use
the term two-­ballot majority (TBM) for majority elections in two rounds.
The two-­ballot formula implies that if a candidate gets a majority of the votes
in the first round, he or she is elected. If no one attains 50 per cent of the total
20   Electoral systems
vote, a second round of voting is conducted, usually within one or two weeks.
Single-­member districts are normally used, but multi-­member districts have also
been applied. In the latter case, electors have several votes at their disposal.
There are basically two variants of TBM: the majority run-­off formula and the
majority-­plurality method. In majority run-­off elections, the second ballot is held
between the two candidates who got the highest number of votes in the first
round. The majority-­plurality formula, on the other hand, does not require such a
drastic reduction in the number of contestants on the second ballot – the winner
is the candidate who gets a plurality in the second round of voting. However, a
threshold is usually imposed for candidates to stand at the second ballot.
Under the alternative vote (AV), also known as preferential voting, the electors
are asked to place the candidates in order of precedence on the ballot. Usually,
single-­member districts are used. In order to win, a candidate must obtain at least
50 per cent of the votes. If no candidate has received a majority of first prefer-
ences, the candidate with the lowest number of first preferences is eliminated. The
votes of those who ranked that candidate highest are then transferred to the second
preferences expressed by the voters. Those ballot papers on which no other can-
didates were given ranking are set aside. This process is repeated until a candidate
receives majority support. Both single- and multi-­member districts may be used.
This formula is usually associated with Australia. In Australian elections, the voter
has to rank order all candidates; otherwise the ballot is rejected.
The principle of proportional electoral systems, often simply called PR, is
that the share of seats awarded to each party should be equal to the share of
votes. The preferences of the electorate should be mirrored in the legislature.
There are two basic types of proportional elections: list PR and the single trans-
ferable vote (STV). Multi-­member districts are applied in both kinds of elec-
tions. In list systems, each party presents a list of candidates in the constituencies.
Electors vote for a party, and the proportion of votes for each party determines
the seat allocation. Winning candidates are normally taken from the lists in order
of their position on the lists. In addition to the party vote, some list systems also
include an element of preferential voting, that is, voters are allowed to vote for
both candidates and parties. Two main seat allocation formulas have been used:
the highest average and the largest remainder. The former determines the seat
allocation by division, whereas the latter does so by subtraction. In addition, the
list PR systems greatly vary as to district magnitude, electoral threshold, adjust-
ment seats, district level and ballot structure. In fact, all list PR countries differ
from each other with regard to one or several of these elements. Therefore, list
PR choices are treated as a single category.
STV is a preferential voting method. Electors may vote for as many or as few
candidates on the ballot paper as they wish. Candidates are elected if they
receive a specific quota (called Droop-­quota) of the total vote in a constituency.
This quota is calculated by dividing the total number of valid votes by the
number of seats to be allocated plus one, and rounding upwards or adding one.
The first stage of the counting process involves the sorting of all ballot papers
according to the first preferences of the voters. Any candidate who has attained a
 Electoral systems   21
number of votes equal to or greater than the quota is immediately elected. At the
next stage, all surplus votes of those candidates already elected, that is, the
number of votes by which the quota is exceeded, are transferred to the remaining
candidates according to the second preference on the ballot paper. If no candi-
date has attained the quota, the candidate with the lowest number of first prefer-
ences is eliminated, and his or her second preferences are redistributed to the
other candidates. At each stage, every candidate that attains the quota is declared
elected. The transfer of the surplus votes of winning candidates and the votes of
eliminated candidates, based on second, third, fourth preferences and so on, con-
tinues until all seats in the constituency are filled.
In addition to plurality, majority and proportional formulas and systems, there
are countries that combine either the plurality or the majority rule with propor-
tional seat allocation in one electoral system. In their detailed survey of all mixed
electoral systems in the world since the beginning of the twentieth century, Mas-
sicotte and Blais (1999) identify five different ways of combining PR with either
plurality or majority. These are called coexistence, superposition, fusion, condi-
tional and corrective. Furthermore, there are a few countries that apply super-
mixed electoral systems, implying that formulas are combined in more than one
of the five basic ways.
Before describing the functioning of these mixed systems, I shall shortly
discuss another comprehensive study of mixed electoral systems, conducted by
Shugart and Martin P. Wattenberg (2001). In the volume Mixed-­member Elect-
oral Systems: The Best of Both Worlds, all major countries that have adopted
mixed systems are examined. To begin with, they argue that Massicotte and
Blais’ definition of a mixed system is too broad, mainly because it includes
countries that employ majoritarian formulas in some parts of the country while
using PR in other areas. Their preference is for a narrower definition in which
mixed systems are seen as a subset of the two-­tier electoral systems used in some
PR countries. The distinguishing feature of mixed systems is that one tier must
allocate seats nominally, whereas the other tier must allocate seats by lists.
Nominal voting means that electors vote for candidates by name, and seats are
distributed to individual candidates on the basis of the votes they receive. Seat
allocation by lists corresponds to the voting procedure in list PR systems.
The combination of a nominal and a list tier implies a combination of the
majoritarian and the PR principles in one electoral system. In this respect,
Shugart and Wattenberg’s notion of a mixed system does not differ from that of
Massicotte and Blais. They also agree that the most basic distinction between
different mixed systems is the presence or absence of a linkage between the tiers,
or in Massicotte and Blais’ terminology, whether the application of one formula
is dependent on the outcome produced by the other formula or not. However,
their classifications of subtypes differ somewhat from each other. Shugart and
Wattenberg point out that most mixed systems tend to lean towards either a
majoritarian or a PR system in their overall effects, and the basic subtypes are
therefore called mixed-­member majoritarian (MMM) and mixed-­member pro-
portional (MMP). In an MMM system, there is no linkage between the nominal
22   Electoral systems
and the list tier – in other words, the seat allocation in the list tier is not depend-
ent on the results in the nominal tier. In an MMP system, however, the purpose
of the list tier is to compensate for any disproportionality in the seat allocation
between the parties in the nominal tier so that the final seat distribution becomes
proportional.
In Massicotte and Blais’ classification, the independent mixture of formulas
consists of three subtypes (coexistence, superposition and fusion), whereas two
forms of dependent combination applies (conditional and corrective). In a coex-
istence system, one part of the seats is filled by plurality or majority elections,
while the other part is filled by PR elections. The division into two parts usually
corresponds to single- and multi-­member districting, and seats are allocated in a
single tier. The superposition mixed system corresponds to the MMM system in
Shugart and Wattenberg’s typology. In contrast to the coexistence system, both
majoritarian and PR elections apply throughout the territory to all voters in sepa-
rate tiers. In a fusion system, both formulas are used within each single district:
some seats are allocated by the majoritarian rule and the remaining seats are pro-
portionally distributed. Massicotte and Blais point out that this system currently
exists only for French municipal elections (1999: 352). If a party list receives a
majority of the vote, it gets one-­half of the seats. The other half is allotted under
PR among all parties, including the leading party. If no list obtains a majority, a
second round is conducted. This time, one-­half of the seats are awarded to the
party with a plurality of the vote, and the remaining seats are proportionally dis-
tributed among all parties.
Regarding conditional systems, the use of one formula is dependent on the
vote distribution or the outcome produced by the other formula. The second
formula comes into play only if the outcome of the first formula, as specified by
the electoral law, does not meet a certain condition, or if the most successful
party (or cartel) fails to get a certain vote share. The corrective system is quite
similar to the MMP system; the seat distribution in the list (PR) tier is dependent
on the seat distribution in the nominal (majoritarian) tier. However, the main dif-
ference is that Massicotte and Blais in the mixed-­corrective category also include
systems with partial compensation for under-­representation caused by the major-
itarian elections. Some mixed systems fit into none of these five basic types,
because plurality/majority and PR are combined in more than one way. These
systems are called supermixed.
Considering the purpose of the study, I regard the typology by Massicotte and
Blais as more appropriate than that of Shugart and Wattenberg. The distinction
between MMM and MMP systems to a large extent depends on one specific
feature, namely the consequences of different mixed systems in terms of dispro-
portionality. This book, on the other hand, is concerned with the question of why
countries have chosen different electoral systems. Another reason for preferring
Massicotte and Blais’ typology is that they also take into consideration the com-
bination of plurality/majority and proportional elections in a single tier. I main-
tain that these systems necessarily must be classified as mixed, simply because
two different election principles are applied, provided that both methods are used
 Electoral systems   23
to some considerable extent. A classification that disregards a system by which
part of the legislature is elected is imperfect. The extent to which two or several
formulas are applied in mixed systems naturally varies – therefore, a threshold
must be introduced. If a diminutive number of representatives is elected accord-
ing to a formula different from the predominant one(s), the system can hardly be
regarded as a mixed system. In order to qualify as a mixed system, I require that
at least 10 per cent of the seats are allocated by a formula different from the
other formula(s).

3.2 Classifying electoral systems

Most of the existing classifications of electoral systems are primarily based on
the electoral formula, only taking secondary account of the other components of
electoral systems (see e.g. Blais and Massicotte 1996; Bogdanor 1983; Farrell
2001; Lakeman and Lambert 1946; Lijphart 1999). Classifications that give
equal attention to all three main dimensions of electoral systems have also been
presented but they tend to be rather impractical in a methodological sense. Blais
(1988: 99–110), for instance, proposes a more sophisticated classification than
the existing ones. He distinguishes three components of the ballot structure (the
object of the vote, the number of votes and the type of vote), two components of
the constituency structure (nature and magnitude), and three electoral formulas
(majority, plurality and proportionality). This classification, the author says, clar-
ifies and refines existing distinctions, and takes superiority over earlier typolo-
gies. Although it may be a more appropriate typology in a theoretical sense, it is
not very useful in empirical analyses of causes and effects.
Classifications of electoral formulas and systems are often based on their out-
comes in terms of vote–seat relationship. Those systems that have proportional
outcomes are usually distinguished from those with non-­proportional outcomes.
The essence of proportional electoral systems is that seats in the legislature
should be allocated to the political parties in the same proportion, or as closely
as possible, as support for those parties is divided among the electorate. In non-­
proportional systems, by contrast, greater importance is attached to strong and
effective government by ensuring, or at least promoting, that the winning party
receives a majority in the legislature. In addition, there are different degrees of
proportionality and many scholars, consequently, identify an in-­between cat-
egory called semi-­proportional systems. However, Richard Katz points out that

This approach is both circular and self-­contradictory – circular in that if the

defining characteristic of an electoral formula is its actual degree of propor-
tionality in the distribution of seats, then electoral formula cannot explain
that distribution; self-­contradictory in that . . . single-­member systems are
not uniformly less proportional than all semiproportional systems, which in
turn are not uniformly less proportional than all systems using proportional
representation.
(Katz 1997: 109)
24   Electoral systems
Although Katz’s remark is true, classifications of electoral systems often distin-
guish between proportional systems, non-­proportional systems and an intermedi-
ate category of formulas that fit into none of the former two. Lijphart, for
example, argues that these are the main types of electoral systems, each one con-
taining a large number of subtypes. In Electoral Systems and Party Systems
(1994), the author classifies the plurality formula, two-­ballot systems and AV as
the main subtypes of majoritarian systems. Proportional systems are further
divided into largest remainders, highest averages and STV. The category of
semi-­proportional systems consists of CV and LV. The same classification,
including some additional subtypes, is applied in Patterns of Democracy
(Lijphart 1999). In The International IDEA Handbook of Electoral System
Design (1997), Reynolds and Reilly classify all existing electoral formulas and
systems according to the same scheme as above, called ‘electoral system famil-
ies’. In one of the first comparative analyses of electoral systems, Voting in
Democracies, written by Enid Lakeman and James D. Lambert (1946), four
main electoral systems are identified: the first-­past-the-­post system, semi-­
proportional systems, proportional list systems and STV.1
In Electoral Systems: A Comparative Introduction, Farrell (2001) distin-
guishes between plurality, majority and proportional electoral systems. In addi-
tion, a category of electoral systems that combines different electoral formulas is
created. A similar classification is applied by Blais and Massicotte (1997) in a
cross-­sectional analysis of electoral system choice in the world, presented in
Chapter 2 of this volume. In a study of the adoption of PR at the turn of the
twentieth century, likewise presented above, Blais et al. (2005) distinguish
between plurality, majority and proportional systems. For the purpose of analys-
ing electoral system change, Colomer (2004b) distinguishes between indirect
elections, majority rule (which includes plurality), mixed systems and propor-
tional representation.
To sum up, many classifications distinguish between plurality/majority, semi-­
proportional and proportional systems. Another common ground of classification
is to divide formulas into plurality, majority, proportional and mixed electoral
systems. Most classifications follow one of these two principles, sometimes with
small modifications. If electoral formulas are classified according to degrees of
proportionality, plurality and majority formulas are usually brought together into
a single majoritarian category.
The degree of proportionality also varies between different types of mixed
systems. If one-­half of the seats are allocated according to the majoritarian rule
and the other half is proportionally distributed, the final vote–seat ratio should
theoretically be somewhere between that of pure majoritarian and PR elections.
However, if the PR tier is aimed at compensating for the disproportionality in
the nominal tier, the mixed system becomes a proportional system in terms of
effects. There are, furthermore, mixed systems that only partially correct the dis-
proportionality caused by the majoritarian elections, which renders the classifi-
cation of mixed systems according to the three-­scale typology difficult. Another
disadvantage of this classification is that it is by definition based on a particular
 Electoral systems   25
feature of electoral systems, namely the degree of proportionality. Since the
present study is not concerned with effects of electoral systems, a classification
based on proportionality is inappropriate.
Considering the nature of the study, the other common ground of classifying
electoral systems appears more attractive. As mentioned earlier, seats can prin-
cipally be allocated in three different ways: by the plurality rule, by the majority
rule, and proportionally among parties according to a specific quota or divisor. It
seems appropriate to regard these methods as separate categories of electoral
systems. Plurality and majority formulas certainly share some common features,
especially in comparison with proportional systems, but the plurality and the
majority rule have, notwithstanding, always been regarded as different electoral
systems. As a matter of fact, before the introduction of PR, these methods were
the main alternatives. All electoral systems that combine either the plurality or
the majority rule with proportional allocation of seats constitute the fourth cat-
egory of mixed electoral systems. This implies that, for example, the German
system, in spite of proportional seat distribution, does not belong to the propor-
tional category. I regard the characteristic of combining two different electoral
systems in one election as more fundamental than the fact that elections result in
proportional seat distribution between parties. Besides, the number of countries
with mixed-­member proportional systems is very small. Likewise, only a few
countries have adopted the semi-­proportional SNTV during the second half of
the twentieth century.
The typology of electoral systems in this book is presented in Table 3.1. The
broad classification distinguishes between plurality, majority, proportional and
mixed systems, whereas the detailed classification consists of a total of 15 sub-­
categories of these four systems. Cumulative vote is left out of the typology,
because no independent country has adopted this formula since 1945. The broad

Table 3.1 A typology of electoral systems

Electoral systems Electoral formulas

A. Plurality systems   1. Single-member plurality (SMP)

  2. Block vote (BV)
  3. Single-member plurality – Block vote (SMP–BV)
  4. Limited vote (LV): Single non-transferable vote (SNTV)
B. Majority systems   5. One-ballot majority (OBM)
  6. Two-ballot majority (TBM)
  7. Alternative vote (AV)
C. Proportional systems   8. List PR
  9. Single transferable vote (STV)
D. Mixed systems 10. Coexistence
11. Superposition
12. Fusion
13. Conditional
14. Corrective
15. Supermixed
26   Electoral systems
classification takes precedence over the detailed one. It would be of great value
to provide explanations of the adoption of electoral formulas, that is, to conduct
empirical analysis on the basis of the detailed typology. For practical reasons,
however, this is an impossible task. In order to conduct large-­scale analyses, a
generalization is inevitable. Consequently, the multivariate analyses are based
on the broad classification. This holds true for a large part of the bivariate analy-
ses, too. The detailed classification serves partly a descriptive and informative
purpose. Given the explorative nature of the study, it is of general interest to
map out the spread of electoral formulas around the world. Concerning some
particular independent variables, it serves as a tool for constructing an appropri-
ate categorization in analysing bivariate associations.

3.3 An introductory survey of electoral system choice

The typology of electoral formulas and systems in Table 3.1 constitutes the basis
for the dependent variable of the study, that is, electoral system choice. To begin
with, we need to define what is meant by electoral system choice. The analysis
covers the time period from 1945 onwards, which implies that all existing elect-
oral systems at the beginning of the year 1945 are excluded from the analysis.
The study encompasses all electoral systems for parliamentary elections (to the
lower house) that have been introduced between January 1945 and December
2007, major electoral system reforms included. However, the extent to which
electoral systems are altered varies greatly. For example, an electoral reform
might involve a change from closed party lists in a proportional list system to
preferential voting within party lists, or the introduction of an electoral thresh-
old. These kinds of reforms do not change the electoral system as a whole; there-
fore, I do not regard them as electoral system changes. I only include such
reforms that involve a change from one of the major categories – plurality,
majority, proportional or mixed – to another category.
In the following, the frequency of adopted electoral formulas and systems
around the world is presented. Separate information on electoral system choice
in democracies is provided. In addition to all choices of electoral systems
between 1945 and 2007, a survey of the electoral formulas and systems in the
contemporary world, that is, at the end of 2007 is provided. This is not relevant
to the purpose of the study; it is, rather, presented on grounds of general interest.
The total number of electoral system choice is 278. As we see in Table 3.2, list
PR is the most frequent electoral method with 78 cases. The other proportional
formula, STV, has been chosen only once (in Malta) after 1945. The second most
frequent formula is SMP of which 52 cases are identified, followed by BV with
35 cases. Of the majority formulas, TBM has been chosen 34 times, whereas only
three countries (Nauru, Fiji and Papua New Guinea) have adopted AV since
1945. OBM has been applied in North Korea since 1948 and in Cuba since 1992.
This formula was also used in Burmese legislative elections from 1978 to 1985.
There are six cases of SNTV. Vanuatu has applied SNTV since independ-
ence, Japan applied it until 1993, and Afghanistan introduced the formula in
 Electoral systems   27
Table 3.2 Electoral system choice 1945–2007

Electoral formula and system N %

Single-member plurality 52 18.7

Block vote 35 12.6
SMP–BV 12 4.3
Limited vote: SNTV 6 2.2
One-ballot majority 3 1.1
Two-ballot majority 34 12.2
Alternative vote 3 1.1
List PR 78 28.1
Single transferable vote 1 0.4
Mixed-coexistence 7 2.5
Mixed-superposition 28 10.1
Mixed-fusion 2 0.7
Mixed-conditional 2 0.7
Mixed-corrective 9 3.2
Supermixed 6 2.2
N 278 100.0
Plurality systems 105 37.9
Majority systems 40 14.4
Proportional systems 79 28.4
Mixed systems 54 19.4
N 278 100.0

2004. In addition, it has been applied in Taiwan, Bahrain and the Maldives. A
mixture of SMP and BV has been adopted 12 times, for instance, in India 1950,
Western Samoa 1963, Ivory Coast 1985 and Palau 1994.
Superposition is the most popular mixed system, having been adopted 28
times – all the other mixed systems are rather infrequent. Most of the superposi-
tion systems have been introduced in the 1990s, for example in Guinea, Croatia,
Azerbaijan, Andorra and Thailand. Senegal and Egypt adopted this system in
1983 and 1986 respectively. The earliest example of superposition during the
relevant time period is South Korea 1963. A total of seven choices of mixed-­
coexistence are observed: Guatemala 1946, Greece 1956, Suriname 1975, Zim-
babwe 1980, Panama 1980, and Madagascar 1983 and again 1998. In their
survey of mixed electoral systems, Massicotte and Blais (1999) did not find any
national instances of the mixed-­fusion system. However, in Paraguayan legisla-
tive elections from 1968 to 1989, the winning list was awarded two-­thirds of the
seats, whereas one-­third was distributed proportionally among the other parties
(León-Roesch 1993: 638). A similar system was applied in Nicaragua for legis-
lative elections between 1962 and 1974 (Krennerich 1993: 583–4). These
systems correspond to Massicotte and Blais’ description of the mixed-­fusion
system (1999: 352–3). Mixed-­conditional systems have been adopted only in
France 1951 and Italy 1953. Mixed-­corrective systems have been chosen
28   Electoral systems
eight times: in West Germany 1949, Mexico 1963, Albania 1992, Italy 1993,
Venezuela 1993, Tunisia 1993, New Zealand 1996 and Bolivia 1997. Super-
mixed systems have been introduced in six countries: Madagascar 1960, Ecuador
1978, Hungary 1989, Cameroon 1992, Chad 1996 and Monaco 2002.
Lower down in Table 3.2, electoral formulas are merged into electoral system
families. Plurality systems have been most frequently adopted since 1945; they
represent 38 per cent of the population. Majority systems constitute the smallest
category with less than 15 per cent of all cases. Considering that mixed systems
have, at least until the 1990s, been given quite little attention, they are far from
rare: one-­fifth of all adopted electoral systems are combinations of majoritarian
and proportional elections. Of all cases, 28 per cent are proportional systems.
All electoral formulas in the contemporary world are presented in Table 3.3.
Accordingly, this table also includes electoral formulas adopted before 1945. At
the end of 2007, 12 of all 194 independent states lacked electoral provisions and/
or a functioning directly elected national legislative body. Saudi Arabia, Oman,
Qatar, the United Arab Emirates, Eritrea, and the Vatican City have never held
parliamentary elections. The Chinese and the Bhutanese governments have never
allowed their citizens to directly choose their representatives, and the only legis-
lative elections in Brunei were held, indirectly, in 1962 and 1965, that is, two

Table 3.3 Electoral formulas and systems 2007

Electoral formula and system N %

Single-member plurality 41 22.5

Block vote 10 5.5
SMP–BV 2 1.1
Limited vote: SNTV 3 1.6
One-ballot majority 2 1.1
Two-ballot majority 16 8.8
Alternative vote 4 2.2
List PR 73 40.1
Single transferable vote 2 1.1
Mixed-coexistence 1 0.5
Mixed-superposition 16 8.8
Mixed-fusion 0 0.0
Mixed-conditional 0 0.0
Mixed-corrective 7 3.8
Supermixed 5 2.7
N 182 100.0
Plurality systems 56 30.8
Majority systems 22 12.1
Proportional systems 75 41.2
Mixed systems 29 15.9
N 182 100.0
 Electoral systems   29
decades before independence. The last popular elections in Libya were held in
1965. Burma and Somalia had no electoral institutions at the end of 2007.
Table 3.3 shows that list PR is by far the most frequently used electoral
formula today. Seventy-three countries use proportional list systems, whereas
SMP is applied in 41 countries. STV is used in two countries only, Ireland and
Malta. BV is not very popular anymore; ten countries apply this formula today.
The two-­ballot majority formula, which is often referred to as the French system,
is used in 15 other countries as well. SNTV is applied in Vanuatu, Jordan and
Afghanistan. Among the mixed systems, superposition is by far the most popular
formula, being applied in 16 countries. Mixed-­fusion and mixed-­conditional are
not applied anywhere. Proportional and plurality systems are much more popular
than majority and mixed systems.
In Table 3.4, all choices of electoral formulas and systems from 1945 to 2007
in democracies are given – the sample consists of 66 cases. Hence, the great major-
ity (76 per cent) of all electoral system choices has occurred in non-­democratic cir-
cumstances. Both SMP and list PR have been adopted 18 times; together they
constitute a good one-­half of the sample. The dispersion of other formulas is more
even among democracies. More than 40 per cent of all electoral system choices
under democratic circumstances are plurality systems, whereas the majority

Table 3.4 Electoral system choice in democracies 1945–2007

Electoral formula and system N %

Single-member plurality 18 27.3

Block vote 5 7.6
SMP–BV 3 4.5
Limited vote: SNTV 1 1.5
One-ballot majority 0 0.0
Two-ballot majority 3 4.5
Alternative vote 2 3.0
List PR 18 27.3
Single transferable vote 1 1.5
Mixed-coexistence 3 4.5
Mixed-superposition 5 7.6
Mixed-fusion 0 0.0
Mixed-conditional 2 3.0
Mixed-corrective 4 6.1
Supermixed 1 1.5
N 66 100.0
Plurality systems 27 40.9
Majority systems 5 7.6
Proportional systems 19 28.8
Mixed systems 15 22.7
N 66 100.0
30   Electoral systems
Table 3.5 Electoral formulas and systems in democracies 2007

Electoral formula and system N %

Single-member plurality 23 20.7

Block vote 2 1.8
SMP–BV 1 0.9
Limited vote: SNTV 1 0.9
One-ballot majority 0 0.0
Two-ballot majority 3 2.7
Alternative vote 3 2.7
List PR 58 52.3
Single transferable vote 2 1.8
Mixed-coexistence 1 0.9
Mixed-superposition 10 9.0
Mixed-fusion 0 0.0
Mixed-conditional 0 0.0
Mixed-corrective 5 4.5
Supermixed 2 1.8
N 111 100.0
Plurality systems 27 23.4
Majority systems 6 6.3
Proportional systems 60 54.1
Mixed systems 18 16.2
N 111 100.0

systems constitute a very modest part of the sample. Surprisingly, mixed systems
are almost as frequent as proportional systems among democracies.
The spread of electoral formulas and systems among democracies in the con-
temporary world is given in Table 3.5. At present, list PR systems constitute
more than one-­half of all electoral systems in democratic countries. SMP is used
by 20 per cent of the democratic regime. Mixed-­superposition, which is the third
most applied system today, exists in approximately one-­tenth of all democracies.
The most striking difference in comparison with all countries in the world is
that TBM constitutes merely 3 per cent of all electoral systems; that is, only
three democracies (France, Mali and Kiribati) apply this formula, which after all
is regarded as one of the main electoral formulas. Concerning the broad typol-
ogy, the proportion of plurality and majority systems is smaller among democra-
cies than among all countries, mostly at the expense of proportional systems.
The frequency of electoral system choice by each year from 1945 to 2007 is
illustrated in Figure 3.1. During the first year, only two electoral systems were
adopted: Brazil and France introduced pure list PR for the first time. In 1946, eight
electoral systems were adopted. Thereafter, the frequency of electoral system
choice remained rather low until 1960 when 16 systems were adopted, mostly due
to the fact that many African countries achieved independence that year.
Electoral systems were adopted quite frequently during the 1960s. In 1969,
 Electoral systems   31

20
18
16
14
12
10
8
6
4
2
0
Electoral
Figure 3.1 Frequencies of system
electoral choice
system 1945–2007
choice 1945–2007.

0
Electoral
Figure 3.2 Frequencies system choice
of electoral systeminchoice
democracies 1945–2007
in democracies 1945–2007.

however, no electoral system was introduced. The frequency of adopted systems

was quite low during the 1970s until a new peak was reached in 1978 with nine
choices. Few choices of electoral systems were implemented between 1981 and
1989. However, the changes in the Communist world resulted in several new cases
in the 1990s, 1993 being the culmination with 19 cases of electoral system choice.
The temporal spread of electoral system choice under democratic circum-
stances is given in Figure 3.2. A similar pattern as in the previous figure is dis-
cernible but the pattern is not as clear as in the total population. Between 1945
and 1956, only ten electoral systems were adopted in democratic countries.
Thereafter, electoral system choice became somewhat more popular: 12 cases
emerged from 1958 to 1966. Between 1967 and 1973, only Nauru and Fiji
adopted electoral systems. A frequent period of electoral system choice began in
1974; seven years later, 16 additional choices had taken place. Only three demo-
cratic cases are documented between 1982 and 1990, two of which occurred in
France. A revival of electoral system choice among democracies started in the
early 1990s – 23 cases are observed between 1991 and 2007.
4 The rational perspective

The rational approach to electoral system choice maintains that electoral systems
serve certain purposes. It is largely concerned with deliberate design of electoral
rules on the basis of the sociopolitical conditions of a society. According to this
view, the structural differences between countries explain variations in electoral
system design. To begin with, we need to specify what is meant by electoral
system design, also called electoral engineering. In a broader sense, we may
speak of institutional design or constitutional engineering. There is an important
distinction between institutional choice and constitutional engineering. Political
actors in a newly independent country may adopt certain institutions, because
they expect them to maximize their gain in the short term. This is an institutional
choice based on rational considerations regarding the short-­term consequences
of institutional arrangements. By contrast, the discipline of constitutional engin-
eering rests upon the assumption that long-­term stability is the primary goal of a
country, and that the most appropriate institutions in this respect may not be the
same as those that the negotiating actors will benefit from. Institutional design is,
hence, concerned with engineering political outcomes through the choice of
institutional structures (Reilly and Reynolds 1999: 5; Reynolds 1999: 12–13).
This field was given contemporary prominence by Sartori (1968) who urged
political scientists to take up the challenge of becoming participants in the
shaping of political institutions via constitutional engineering.
Deliberate electoral engineering occurred for the first time at the end of the
nineteenth century and the beginning of the twentieth century. For instance, the
alternative vote in Australia was introduced (in 1918) in order to solve the prob-
lems of conservative forces ‘splitting’ their vote opposed by a rising Labor Party
(Reilly and Reynolds 1999: 24). Proportional list systems were initially adopted
in several culturally diverse countries in Western Europe for the purpose of
ensuring balanced inter-­ethnic representation (Rokkan 1970). An electoral
system may also be a disastrous solution for a country. As Jean Laponce and
Bernard Saint-­Jacques point out; political institutions may be problem-­solvers as
well as problem-­creators (1997: 233). Indeed, too often electoral system choices
are the result of accidents and miscommunications leading to several unintended
consequences. This way of adopting electoral laws to a great extent coincides
with colonial heritage. Most of the accidental electoral system choices with dis-
 The rational perspective   33
astrous consequences have been inherited from former colonial powers. Yet,
there are other cases of accidentally adopted electoral rules as well. In 1993,
Jordan rejected SMP–BV and introduced the more proportional SNTV on the
personal initiative of King Hussein, thereby facilitating the election of a quite
large share of Islamic fundamentalists to the legislature (Reynolds and Elklit
1997: 53–4). However, not every accidental choice has led to tragic outcomes.
The Papua New Guinean heritage (AV) from Australia is a case in point. An
electoral system that encourages vote trading between competing candidates and
different communal groups was exceptionally appropriate for the ethnically frag-
mented colony.1 Notwithstanding, the poor experience of many accidental
choices, inherited or not, has emphasized the importance of designing electoral
rules for the specific sociopolitical conditions in a society rather than automati-
cally assuming that a design that is successful in one context will work well in
other social, political and economic circumstances, too.
The rational perspective contains two dimensions. An electoral system choice
may be a rational solution with regard to the societal needs that emerge from
structural traits. However, the adoption of an electoral system may also be a
rational solution from the actors’ point of view. Every politician in the negotia-
tion process most likely prefers an electoral system that he or his party will
benefit from, the adopted system thereby being a rational choice for the potential
winners. This book is concerned with contextual determinants, but the role of
political actors cannot be completely disregarded. More specifically, the study
does not focus on actors but on the political context in which actors make their
decisions on electoral arrangements. This matter becomes particularly apparent
when the association between party systems and electoral system choice is dis-
cussed. Electoral system choice as a consequence of party system structure or
party system transformation is foremost an actor-­related matter but has a struc-
tural dimension as well. As for the rational perspective, I shall distinguish
between ‘structurally generated determinants’ and ‘actor-­related determinants’.
There are some important differences between these two. Structural theories
are principally deterministic; they regard the occurrence of certain political phe-
nomena as the consequence of certain structural traits. However, structural
explanations do not comprise the dynamics of the political process. The actor-­
related approach, on the other hand, is concerned with the process in which
political actors implement decisions on the basis of the existing structural set-
tings. Another noteworthy difference concerns the distance between the inde-
pendent and the dependent variable. Structural theories usually meet the
scientific criterion that the explanatory factor should precede the explained phe-
nomena. Concerning actor-­related explanations, however, the analyst runs the
risk of explaining a phenomenon by the phenomenon itself. Yet both approaches
are important and may be seen as mutually supporting each other: one deals with
the structural prerequisites, whereas the other deals with the dynamic of the
political process within these prerequisites.
34   The rational perspective
4.1 Structurally generated determinants
The adoption of list PR coincided with the extension of suffrage and the devel-
opment of mass parties in Western Europe at the end of the nineteenth and the
beginning of the twentieth century. Before the introduction of the d’Hondt
formula for parliamentary elections in Belgium in 1899, all legislative elections
had been held either under the plurality or the majority rule. Proportional repre-
sentation of the electorate was of secondary importance, if any. The right to vote
was restricted: the will of a part of the citizens was taken to express the will of
all citizens.
The majoritarian electoral systems came under heavy attack as representative
democracy developed. The extension of the suffrage facilitated the organization
of strong lower-­class parties but the electoral rules rendered the entry of such
parties into national politics difficult. The majoritarian principle kept small
parties outside the legislature, or at least restricted their representation to a few
seats. The earliest pressures towards electoral reform were felt in Belgium and
Switzerland, both of which were divided societies in terms of ethnic, linguistic
and religious composition. So called ‘electoral reform societies’ which
demanded an electoral system that would equalize the representation of the dif-
ferent communities involved were established in several democracies.
The adoption of list PR in Belgium was followed by similar electoral reforms
in Finland in 1906 and in Sweden in 1907. By the mid-­1920s, most countries in
continental Europe had switched to list PR. France and Iceland adopted pure PR
systems in 1945 and 1959 respectively, whereas the collapse of the authoritarian
regimes in Spain and Portugal in the 1970s paved the way for proportional list
systems at the Iberian Peninsula. Ireland chose another kind of proportional
system, the single transferable vote, in 1921. The United Kingdom, Monaco and
Andorra are the only Western European countries that never adopted propor-
tional representation for national elections.
Two phases of the introduction of proportional systems can be distinguished:
the ‘minority protection’ phase before World War I, and the ‘anti-­socialist’
phase immediately after the war. The first phase was concerned with the problem
of ethnic and religious minorities. In divided societies, the majoritarian system
could clearly threaten the stability of the political system. PR was consequently
designed to facilitate minority representation, thereby counteracting potential
threats to national unity and political stability. ‘It was no surprise’, Rokkan
argues, ‘that the earliest moves toward proportional representation (PR) came in
the ethnically most heterogeneous countries’ (1970: 157). The ‘anti-­socialist’
phase was concerned with the dynamic of the democratization process, which is
dealt with in section 4.6.

4.2 Ethnic and cultural diversity

The possibility of achieving political stability and mitigating ethnic conflicts in
plural societies through deliberate electoral system design has been given much
 The rational perspective   35
attention in recent times. Most experts on divided societies agree that deep ethnic
and other societal cleavages cause serious problems for democratization and that
the establishment and maintenance of democracy is more difficult in fragmented
than in homogeneous societies. The discipline of constitutional engineering rec-
ognizes that all societies are to some degree characterized by conflict, and that
democracy is a means of managing and processing conflict, rather than abolish-
ing it. Conflicts under democracy are never solved for good; they are temporarily
accommodated and reformulated time after time (Reilly 2001: 5). The challenge
for divided societies is to mitigate political conflict, provide incentives for coop-
eration, and introduce accommodative elements into the political game through
constitutional engineering. To be sure, institutional design also matters in homo-
geneous countries, but democratic stability in these countries is not to the same
extent dependent on electoral arrangements as in societies with profound ethnic,
religious and/or linguistic cleavages. In divided societies, the electoral system
may systematically favour or discriminate against certain ethnic, religious and
national groups.
Democratic competition between political parties in divided societies is inher-
ently difficult because of the strong tendency towards politicization of ethnic
interests, which in turn tends to produce zero-­sum, winner-­take-all politics.
Campaigning along ethnic lines rather than ideological lines provides often a
more effective means of mobilizing voter support (Reilly 2001: 4). The tendency
towards establishing ethnic parties is cumulative since other groups tend to
follow. Furthermore, the tendency to organize parties along ethnic lines is
particularly strong in those divided societies where a few major ethnic groups
compete for political power. There is no reason for ethnic parties to defuse the
ethnic conflict, because the more members of a party’s ethnic base that vote the
greater the electoral success for that party (Horowitz 1985).
There are several structural techniques to reduce ethnic conflict in divided
societies. Federalism and regional autonomy are two institutional means of miti-
gating tensions in countries with deep ethnic cleavages. Electoral system design
is the third basic way of promoting ethnic accommodation. Donald Horowitz
(1985) presents five possible aims for an electoral system in a divided society.
First, the electoral system should fragment the support of the majority group in
order to prevent it from permanent domination. Second, the electoral system
should promote moderate behaviour of the majority ethnic group towards other
groups and interethnic bargaining. Third, the electoral method should encourage
the formation of multiethnic coalitions. Fourth, the electoral system should
invoke incentives for multipolar balance among several groups to prevent per-
manent exclusion of certain minority groups. Fifth, disproportional electoral
results in favour of the largest group should be avoided. The author points out
that no electoral reform can work magic on ethnic conflict – yet, several con-
structive measures are possible in deeply divided societies by means of appropri-
ate electoral arrangements (Horowitz 1985: 601–51).
Consociationalism – a concept developed by Lijphart – is one of the most dis-
cussed prescriptions for plural societies. Consociational democracy consists of
36   The rational perspective
four basic elements. The first principle prescribes executive power-­sharing
among the representatives of all significant groups. Consociational democracy
is, in this respect, the opposite of majoritarian democracy according to the West-
minster model. The second principle, a high degree of internal autonomy for
each of the segments, is a complement to the power-­sharing principle: decisions
on issues of common interests should be made jointly; on all other issues each
segment should be allowed to decide for itself. Proportional representation is the
third element of consociationalism. A proportional electoral system is needed in
order to eliminate the sharp distinction between winners and losers evident under
majoritarian democracy, which normally uses the plurality system. By means of
proportional seat allocation, Lijphart argues, both majorities and minorities can
be ‘winners’ in the sense that the number of representatives corresponds to their
societal strength respectively. The fourth principle prescribes a minority veto on
the most vital issues (Lijphart 1985a: 6–9; see also Lijphart 1977; 2006: 42–55).
The great value of consociationalism lies in its powerful conflict-­solving
capacity in divided societies that have no hope of generating interethnic political
accommodation. However, if a divided society shows potential for voting
beyond ethnic lines, consociational democracy provides few incentives for
accommodative political behaviour. In these societies, the centripetalist approach
provides more appropriate arrangements (Reilly and Reynolds 1999: 31–2).
Referring to Timothy D. Sisk (1995: 19), centripetalism is understood as institu-
tions and policies that encourage cooperation and centrist policies, while coun-
tering extremism and conflict behaviour. The goal is to engineer a centripetal
spin to the political system by pulling the parties towards moderate, compromis-
ing policies and by discovering and reinforcing the centre of a deeply divided
political spectrum. At the heart of centripetalism is the need to create incentives
for accommodation between competing interests in societies with deep ethnic
cleavages, and to process centripetal outcomes from divergent interests and pref-
erences (Reilly 2001: 7–8; see also Reilly 2006: 27–41).
The electoral system is regarded as the chief agent of interethnic accommoda-
tion. In deeply divided societies, the thought of voting for a member of a rival
group is usually very far-­fetched. However, the alternative vote formula, which
permits voters to declare not only their first choice, but also their second, third
and subsequent choices, has the ability of inducing interethnic bargaining and
promoting accommodative behaviour by means of ‘vote-­pooling’ and ‘prefer-
ence swapping’. Since the first preferences received from voters of one’s own
ethnic group may not constitute a majority, politicians are encouraged to cam-
paign for the votes of members of rival groups. Thereby, political actors from
different groups have an incentive to bargain and negotiate in the search for
cross-­partisan and cross-­ethnic vote-­pooling deals. In a situation where no candi-
date can be assured of an outright majority of support, those candidates who suc-
cessfully ‘pool’ both their own first preferences and the second preferences of
others will be elected. In order to attract the support of other groups, candidates
are encouraged to, first, move to the centre on policy issues, and, second, to
focus on cross-­cutting issues that unite rather than divide different ethnic groups.
 The rational perspective   37
The more groups in a given district, the more likely it is that meaningful vote
pooling will occur.
The main difference between consociationalism and centripetalism is that the
former replicates existing ethnic divisions in the legislature, whereas the latter
tries to break down the salience of ethnicity by encouraging cooperation and
accommodation between rival groups before elections. Preferential systems have
the capacity of transcending the political salience of ethnicity by promoting
accommodation and bargaining across group lines (Horowitz 1991; Reilly 2001:
21). According to the theoretical arguments put forward here, ethnically and cul-
turally fragmented countries should use proportional systems or AV in order to
mitigate conflict and promote a peaceful collaboration between societal groups.
It should be observed that the theory does not assert that this pattern is likely to
be found; it is first and foremost intended as a device for plural societies.
The concept of ethnicity has often been broadened to encompass several
characteristics such as language, culture and religion. Narrowly defined,
however, ethnicity is similar to race, thereby constituting a particular type of
cleavage on the side of language, religion and other cultural features. Still, the
term ethnicity is more frequently used than language or religion when countries’
cultural features are described. In the empirical analysis, I shall use two values
of fragmentation: one that measures ethnic fragmentation and one that observes
the cleavage – ethnic, linguistic or religious – that returns the highest level of
fragmentation. Another reason for this distinction is that information on ethnic
affiliation is in most cases more accurate than information on other cultural
characteristics. The sources used determine which differences between cultural
groups are relevant. An ethnic, linguistic or religious group is considered distinct
in every case where it is reported as a separate group. Concerning religious frag-
mentation, the following groups are distinguished: Christians, Muslims, Jews,
Buddhists, Hindus, Shintoists and Taoists. In addition, relevant cleavages within
these groups are also paid attention to. There are four categories of fragmenta-
tion that apply: low, bipolar, medium-­sized and high level of ethnic and cultural
fragmentation. The categories are as follows:

1 low level of fragmentation: largest group >79.9 per cent;

2 bipolar fragmentation: two largest groups >89.9 per cent, largest group
<70.1 per cent;
3 medium-­sized level of fragmentation: largest group between 50 per cent and
79.9 per cent;
4 high level of fragmentation: largest group <50 per cent, two largest groups
<90 per cent.

The index runs from 1 to 4, the former representing a low level of fragmentation.
For example, a country in which the largest ethnic and linguistic group consti-
tutes 95 per cent and the largest religious group 45 per cent of the population
scores 1 and 4 on the index of ethnic and cultural fragmentation respectively.
For countries with medium-­sized and high level of fragmentation, proportional
38   The rational perspective
systems and AV are considered more appropriate than other systems. Similarly,
majoritarian (except for AV) and mixed systems are better adapted for countries
with bipolar or low level of fragmentation.
According to the categorization, medium-­sized level of fragmentation denotes
higher level of fractionalization than bipolar fragmentation. Calculated by
Douglas W. Rae’s and Michael J. Taylor’s index of fractionalization (1970),
however, several constellations of medium-­sized fragmentation score lower than
bipolar ones. In a fictive example, the ethnic affiliation (percentage) in two coun-
tries, A and B, is 45–45–5–5 and 70–15–10–5 respectively. According to my
categorization, country A is bipolar, whereas country B is characterized by
medium-­sized fragmentation. Yet, Rae’s and Taylor’s index denotes that the
level of fragmentation is higher in country A (0.59 compared to 0.475). Notwith-
standing, a majoritarian system is more appropriate in country A with two large
ethnic groups than in country B, in which a majoritarian system would discrimi-
nate against the three smaller groups at the expense of the large one. Therefore,
bipolar fragmentation is regarded as lower than medium-­sized level of
fragmentation.

Empirical evidence
We are now ready to empirically study the relationship between diversity and
electoral system choice. To begin with, some descriptive information is pro-
vided. The dispersion of all cases is given in Table 4.1. A low level of ethnic
fragmentation characterizes approximately 40 per cent of the sample. The group
with high level of fragmentation is also considerably larger than the categories
with bipolar and medium-­sized fragmentation. When the level of cultural frag-
mentation is observed, the number of low-­fragmented cases is reduced to 57, that
is, slightly more than one-­fifth of the entire population. The group of high frag-
mentation is naturally the largest since the cultural diversity variable observes
the cleavage that returns the highest level of fragmentation.
In the following, chi-­square tests are conducted in order to determine whether
there is a relationship between ethnic/cultural diversity and electoral system
choice. If countries have taken ethnic and cultural traits into consideration, the
share of list PR, STV and AV should increase as we move towards a higher level
of fragmentation. It may be pointed out that the number of countries that have

Table 4.1 Frequencies of ethnic and cultural diversity

Level of fragmentation Ethnic (N) % Cultural (N) %

Low 109   39.4   57   20.6

Bipolar   38   13.7   27    9.7
Medium-sized   48   17.3   75   27.1
High   82   29.6 118   42.6
N 277 100.0 277 100.0
 The rational perspective   39
Table 4.2 Associations between ethnic/cultural diversity and electoral system choice

Level of fragmentation List PR, STV, AV (N) % Other systems (N) %

Ethnic fragmentation
   Low 32 29.4   77 70.6
   Bipolar 14 36.8   24 63.2
   Medium-sized 15 31.3   33 68.8
   High 20 24.4   62 75.6
N 81 29.2 196 70.8
Cultural fragmentation
   Low 21 36.8   36 63.2
   Bipolar 10 37.0   17 63.0
   Medium-sized 20 26.7   55 73.3
   High 30 25.4   88 74.6
N 81 29.2 196 70.8

Notes
Ethnic fragmentation: chi-square value 2.088, sig. 0.554.
Cultural fragmentation: chi-square value 3.456, sig. 0.327.

adopted proportional systems or AV is much smaller than the number of ­countries

that have chosen other kinds of electoral systems. Results are given in Table 4.2.
The small chi-­square value concerning ethnic fragmentation and electoral
system choice indicates that there is no association between these variables. We
notice that the share of proportional systems (and AV) slightly increases from
low to bipolar fragmentation, but decreases from bipolar to medium-­sized frag-
mentation as well as from medium-­sized to high level of fragmentation. Among
those countries that are most in need of conflict-­mitigating electoral systems, the
share of appropriate systems is less than 25 per cent.
Neither is there a significant relationship between cultural fragmentation and
the dependent variable. The pattern resembles that of ethnic diversity. Again, the
smallest share of proportional systems (and AV) is observed among countries
with a high level of fragmentation. This category consists of such countries as
the Philippines, the Federated States of Micronesia, the Gambia, Tanzania and
Zambia which all apply SMP. Hence, the findings in both samples contradict the
theoretical assumptions.
Next, the pattern of ethnic/cultural diversity and electoral system choice in
democracies is dealt with. The frequencies of electoral system choice in the four
categories are given in Table 4.3. Regarding ethnic fragmentation, one-­half of all
choices belong to the category of low fragmentation, whereas only seven cases
are observed in the category with bipolar fragmentation. As for the variable that
measures cultural diversity, the largest group is the one with high level of frag-
mentation; two cases of five belong here.
In Table 4.4, bivariate patterns are analysed. Again, other systems than PR and
AV have been chosen much more frequently. Concerning ethnic fragmentation, a
pattern that contradicts the theoretical outlines is observed. Majoritarian and
40   The rational perspective
Table 4.3 Frequencies of ethnic and cultural diversity in democracies

Level of fragmentation Ethnic (N) % Cultural (N) %

Low 34   51.5 19   28.8

Bipolar   7   10.6   4    6.1
Medium-sized   9   13.6 17   25.8
High 16   24.2 26   39.4
N 66 100.0 66 100.0

Table 4.4 Associations between ethnic/cultural diversity and electoral system choice in

democracies

Level of fragmentation List PR, STV, AV (N) % Other systems (N) %

Ethnic fragmentation
   Low 13 38.2 21 61.8
   Bipolar   2 28.6   5 71.4
   Medium-sized   3 33.3   6 66.7
   High   3 18.8 13 81.2
N 21 31.8 45 68.2
Cultural fragmentation
   Low   9 47.4 10 52.6
   Bipolar   1 25.0   3 75.0
   Medium-sized   4 23.5 13 76.5
   High   7 26.9 19 73.1
N 21 31.8 45 68.2

Notes
Ethnic fragmentation: chi-square value 1.948, sig. 0.583.
Cultural fragmentation: chi-square value 3.029, sig. 0.387.

mixed systems are particularly frequent in the category with a high level of frag-
mentation. PR and AV are more evenly distributed among the four categories
when cultural fragmentation is observed but the tendencies still contradict the
theoretical assumption. Both patterns are much too vague in order to return a
significant relationship. Not even among democratic countries are electoral
systems chosen as the theory of electoral systems and cultural diversity suggests.
Despite the fact that proportional systems and AV are considered as effective
tools for mitigating conflict in divided societies, such countries have not adopted
these electoral provisions to any greater extent. All culturally fragmented coun-
tries are naturally not marred by tensions between different groups, but even in
peaceful societies these electoral systems may, to a larger extent than other
systems, integrate different segments of the society and promote cooperation.
However, regarding the ethnic component of countries’ societal structure, elect-
oral systems have not been rationally chosen. On the contrary, plural countries
have to an overwhelming extent adopted inappropriate systems with regard to
the ethnic, linguistic and religious composition.
 The rational perspective   41
4.3 Country size
Country size is a structural factor that affects political institutions and practices
in a multitude of ways (see e.g. C. Anckar 2000; Dahl and Tufte 1973).
Rogowski (1987), among others, asserts that small countries are more likely to
adopt PR than countries with a large population. The association is indirect:
small countries tend to use PR because they are more dependent on trade. In
addition, small countries tend to use large electoral districts. The effect of popu-
lation is, again, through trade; its direct effect on constituency size is small.
A direct link between country size and electoral system choice has also been
suggested. Rokkan (1970: 76–7) observed that the small democracies in Europe
applied PR, whereas the larger ones had either rejected it or called its mainte-
nance in question. By 1922, all 11 smaller democracies in Europe, except for
Iceland, had replaced majoritarian systems with proportional systems. Britain
retained the plurality system, France switched back and forth between PR and
majoritarian elections, and the PR arrangements in Germany and Italy were met
with resistance and controversies. With reference to contemporary theoretical
literature, Rokkan suggested that ‘PR is tolerable in the smaller units because
they face lesser loads of decision-­making’ (1970: 77). In large countries, which
are characterized by heavier burdens of responsibility, PR might be a disastrous
solution.
Peter J. Katzenstein (1985) deals with the association between size and elect-
oral systems from another perspective. He argues that the more frequent adop-
tion of PR in small countries reflects a search for consensus and compromise.
Strive for unity is stronger in small countries than in large ones, which are char-
acterized by zero-­sum politics. When universal suffrage was introduced, the
small European countries opted for proportional systems because of their will-
ingness to share power among disparate political actors.
Blais and Massicotte (1997) provide still another explanation for the sug-
gested link between size and electoral systems. According to them, there might
be practical reasons for avoiding PR in large territories: a plurality system with
single-­member districts enables representatives to keep closer contact with those
who have elected them. Large multi-­member districts in large countries would
seriously restrict the capacity of the candidates to reach the electorate during
electoral campaigns and weaken the interaction between representatives and
voters during the term of office. Their study of determinants of electoral formu-
las indicates that the larger the territory, the more likely it is to have single-­
member districts. Population size, in contrast, is of minor importance. They find
no association between size and electoral systems with regard to the dichotomy
between plurality/majority and PR.
D. Anckar (2002) maintains that the suggested link between diminutiveness
and PR is incorrect. By looking at the electoral arrangements in microstates, the
author concludes that smallness leads to plurality rather than PR. Among 40
countries with a population less than one million, the vast majority has adopted
majoritarian systems.2 Hence, the theoretical arguments put forward by Rokkan,
42   The rational perspective
Katzenstein and Rogowski do not stand an empirical test that comprises all small
countries of the world. However, Anckar says that the prevailing pattern of plu-
rality systems in microstates reflects colonial legacy and a process of diffusion
rather than a consequence of size itself.
Both population and area characterize the size of a country – furthermore,
they are highly interrelated.3 All theoretical assumptions, except for that of Blais
and Massicotte (1997), are based on population instead of territory size. Besides,
Blais and Massicotte’s finding on causality between territory size and electoral
system choice is concerned with district magnitude rather than the electoral
system. Consequently, I prefer population to area as an indicator of size.
It is reasonable to assume that the potential effect of size on political phenom-
ena depends on the relative increase of the former. In small entities, changes in
size probably have a greater effect than in large entities. For example, an
increase in population size from 100,000 to 200,000 is likely to have a more
visible effect than an equally large increase from ten million to 10,100,000
people. Hardly anyone notices the difference in the latter case, whereas the
former makes a considerable difference. Therefore, in the statistical analyses, I
shall use the logarithmized version of population size.
Two methodological strategies are applied: correlation analysis and chi-­square
analysis based on threshold values. The purpose of the second approach is to
examine whether there is a non-­linear relationship between country size and elect-
oral system choice. The categorization is made by means of the logarithmic values:
by treating each integer as one category, six categories are created. In both cases,
the broad category of four basic electoral systems is used. It should be observed
that correlation analysis is not the optimal technique for measuring the relationship
between a categorical and a continuous variable. However, the dependent variable
of the study is reminiscent of an ordinal variable: majoritarian and PR systems are
poles apart and mixed systems constitute an intermediate category.

Empirical evidence
The empirical analysis begins with the large sample. The correlation coefficient
(Spearman’s rho) for population size and electoral system choice is 0.189 and
significant at the 0.01 level. Accordingly, large countries apply proportional
systems to a greater extent than small countries. The chi-­square test in Table 4.5
also returns a significant association between population size and the dependent
variable. The division between countries with less than one million citizens into
three categories – one with a population below 10,000, a second with no more
than 100,000 people, and a third with a population ranging from 100,000 to one
million – may seem unnecessary. All countries with less than one million cit-
izens can be regarded as microstates, thereby in need of the same kind of elect-
oral system according to theoretical considerations. Yet, the analysis returns a
quite clear pattern. Plurality systems constitute a majority of all electoral system
choices (36 cases of 59) in microstates, three cases of four are majoritarian ones,
and proportional systems are more frequent in larger entities.
 The rational perspective   43
Table 4.5 Association between country size and electoral system choice

Population Plurality % Majority % Mixed % PR %

(N) (N) (N) (N)

0–10,000    1 50.0   1 50.0   0   0.0   0   0.0

10,000–100,000    8 53.3   2 13.3   4 26.7   1   6.7
100,000–1 million   27 64.3   4   9.5   1   2.4 10 23.8
1–10 million   41 32.5 19 15.1 27 21.4 39 31.0
10–100 million   26 29.5 14 15.9 20 22.7 28 31.8
100–1 billion    1 25.0   0   0   2 50.0   1 25.0
N 104 37.5 40 14.4 54 19.5 79 28.5

Note
Chi-square 28.54, sig. 0.018.

The fourth category with population size ranging from one to ten million is
the largest one, consisting of 126 cases. The share of proportional systems in this
category is somewhat larger than the share of proportional systems in the total
data sample. However, majoritarian systems, especially plurality ones, are still
more frequent considering that countries of this size are, at least according to
Rokkan’s reasoning, still quite small. The category of countries with a popula-
tion between ten million and 100 million consists of 88 cases. The share of plu-
rality, majority, mixed and proportional systems is similar to that of the previous
category.
The fourth category consists of almost one-­half of all cases in the data
sample. It may be suggested that a possible threshold value is present among this
group of countries, but a division between countries with one million to five
million, and five million to ten million people does not substantially alter the
picture. The only difference is that the share of mixed systems increases by 12
per cent at the expense of proportional and plurality systems in the larger group.
The fifth group is also quite large, and therefore a split between countries with
less and more than 50 million citizens may provide some further information.
Among countries with ten million to 50 million people, the dispersion of elect-
oral systems is approximately the same as in the category with one million to ten
million people. Among countries with 50 million to 100 million people,
however, the share of mixed systems grows at the expense of proportional
systems. Only two countries in this sub-­category – Indonesia (1953) and France
(1986) – have adopted proportional systems, whereas mixed systems have been
chosen five times, plurality systems four times, and majority systems three
times.
Accordingly, proportional systems are rare among very large countries –
instead, there seems to be a need for single-­member constituencies in countries
with a huge population. The findings are also in accordance with D. Anckar’s
(2002) conclusion regarding the possible link between smallness and electoral
system choice. Majoritarian electoral systems, rather than PR, dominate among
small countries. This association is strong enough to bring about a significant
44   The rational perspective
correlation between size and electoral system choice. Hence, these findings con-
tradict the theoretical arguments put forward by most scholars.
Turning to the democratic sample, the bivariate analysis returns a strong cor-
relation (0.447), significant at the 0.001 level, suggesting that plurality systems
are related to smallness. Table 4.6 shows that a great majority of electoral system
choices in democracies with less than one million people are indeed plurality
ones. Mixed and proportional systems are frequent among larger countries.
Nauru and Tuvalu are the only democratic countries in the sample with less
than 10,000 citizens (when electoral systems were chosen). Nauru adopted AV in
1968 and Tuvalu chose SMP–BV ten years later. In the second category ten elect-
oral system choices are observed, eight of which are majoritarian. Seychelles and
Monaco adopted mixed systems in 1976 and 2002 respectively. Among countries
with a population ranging from 100,000 to one million, five PR cases are identi-
fied: Iceland 1959, Malta 1964, Guyana 1966, Cyprus 1981 and Montenegro
2006. Suriname adopted mixed-­coexistence in 1975; all other cases in this cat-
egory are plurality systems. Among countries with more than one but less than
ten million people, six cases of proportional as well as plurality systems of a total
of 17 are observed. The proportion of plurality systems in comparison with PR
further decreases in the fifth category, plurality having been adopted only in the
Philippines (1946 and 2001). France introduced TBM in 1958 and again in 1988.
Mixed systems are almost as popular as list PR, having been chosen in, for
instance, Venezuela 1993, Thailand 1997 and Madagascar 1998.
In the fifth category six of a total of eight PR choices are observed in coun-
tries with a population between ten million and 50 million people. In other
words, there are only two proportional choices (France in 1986 and Italy in
2005) among those seven democratic cases with a population ranging from 50
million to 100 million. A total of three (Italy 1993, Thailand 1997 and the Phil-
ippines 1998) cases are mixed systems, whereas France and the Philippines
adopted majoritarian systems in 1988 and 2001 respectively. The sole case in the
sixth category – the Japanese electoral reform in 1994 – belongs to the family of
mixed systems.

Table 4.6 Association between country size and electoral system choice in democracies

Population Plurality % Majority % Mixed % PR %

(N) (N) (N) (N)

0–10,000   1 50.0 1 50.0   0    0.0   0   0.0

10,000–100,000   7 70.0 1 10.0   2   20.0   0   0.0
100,000–1 million 11 64.7 0   0.0   1    5.9   5 29.4
1–10 million   6 35.3 1   5.9   4   23.5   6 35.3
10–100 million   2 10.5 2 10.5   7   36.8   8 42.1
100–1 billion   0   0.0 0   0.0   1 100.0   0   0.0
N 27 40.9 5   7.6 15   22.7 19 28.8

Note
Chi-square 27.838, sig. 0.023.
 The rational perspective   45
Thus, we may conclude that mixed electoral systems are popular among very
large democracies. The overall pattern that was observed in the entire population
is even clearer among democratic cases. Plurality systems dominate among small
countries, whereas PR is more frequent in medium-­sized and large countries.
Among very large countries, however, mixed systems are more common than
proportional ones. By using majoritarian or mixed systems with single-­member
districts, an element of geographical accountability is maintained in populous
countries. As for the rest, however, it seems dubious that country size really
explains electoral system choice, since rational motives for the observed pattern
are lacking.

4.4 Actor-­related determinants

A general study of electoral system choice cannot completely ignore the role of
political actors. The sociopolitical, historical and geographical settings constitute
a framework for choosing an electoral system but the choice itself is imple-
mented by political actors. Irrespective of why a given electoral system is
adopted, the decision is taken by politicians who pursue their individual inter-
ests. Their interests may be centred on furthering their political careers, or they
may be focused on voters of their own districts or the country as a whole. Never-
theless, each participant in the process of choosing an electoral system pursues
specific policy goals. Hence, the final choice reflects the interests of individual
political actors and negotiated settlements of political conflicts over the institu-
tional design (Mozaffar 1998: 81; see also Gallagher 2005: 538–41).
However, the self-­interest of politicians is difficult to define. It cannot be
defined solely as winning the next election, since this goal may be against long-­
term interests. An underlying interest of every responsible politician in democra-
cies is preservation of stability, which may conflict with preferences based on
individual success. Tradition and familiar examples from abroad also affect the
minds of decision-­makers. Nevertheless, some conclusions on the likely prefer-
ences of political actors can be drawn. The preferences of self-­interested politi-
cians depend on their roles, which societal interests their parties represent, and
whether their parties are rising or declining relative to others. Politicians and
parties prefer electoral rules that they think will maximize their legislative repre-
sentation. Members of small parties, declining parties and parties with an uncer-
tain future most likely prefer proportional representation. Large parties and
established parties with a firm voter support, on the other hand, benefit from a
majoritarian system that over-­represents the largest parties.
On the basis of a study of electoral system choice and parliamentary elections
in South Korea in 1988, David Brady and Jongryn Mo (1992) regard the choice
of electoral systems as the outcome of political competition among the partici-
pating parties. The goal of political parties is to maximize their seat share
through the choice of electoral rules. In order to avoid a collapse of the political
system, however, the opposition must be given a fair chance to succeed under
the new rules. The majority party or coalition cannot inconsiderately force its
46   The rational perspective
first preference through without seeking acceptance from at least some minority
parties. The final choice should, in other words, be viewed as legitimate.
However, the range of electoral systems is constrained by local political con-
ditions and traditions. Certain systems might be omitted from the menu for
several reasons. A country may have had bad experience of a particular system,
and other systems may not be relevant options simply because they are unfamil-
iar. A third condition that influences the bargaining over electoral systems is
uncertainty of the outcome of elections under the new system. In addition, the
outcome may be affected by exogenous factors such as campaign resources and
the emergence of new issues. Brady and Mo conclude that the final agreement
on the new electoral system in South Korea (mixed-­superposition) was a com-
promise between the desire of the largest party (DJP) to create a system that
would secure them a majority and the constraints placed on them by, first, public
opinion and, second, the need for an electoral system that other participants
would accept (Brady and Mo 1992: 405–29). Thus, the new electoral system was
a rational choice with regard to both the largest party and to the stability of the
political system.
A decade ago, Boix (1999) pointed out that there are yet only two seminal
works on what causes the high degree of cross-­national variations in electoral
laws: Citizens, Elections, Parties: Approaches to the Comparative Study of the
Processes of Development (Rokkan 1970) and ‘Trade and Democratic Institu-
tions’ (Rogowski 1987). Boix emphasizes the different strategies that political
actors pursue, depending on certain conditions and constellations. The author’s
theory of the selection of electoral systems is largely based on the ‘Rokkan
hypothesis’. However, Boix argues that Rokkan’s argument is under-­specified.
Rokkan does not indicate the conditions under which policy-­makers will feel
sufficiently threatened to change the current electoral rules. Why, for instance,
did Britain introduce universal suffrage without adopting PR (as did Denmark
and Sweden)? Why have certain countries shifted back and forth between major-
itarian and PR systems during the last century? A third problem, Boix claims, is
that Rokkan’s explanation is too historically bounded. The adoption of PR in
Europe a century ago was associated with the rise of socialism – this factor is
not relevant to electoral system choice in post-­authoritarian Latin America, post-­
Communist Eastern Europe and democratizing countries today.
In his study of electoral system choice in 23 developed democracies, Boix
shows that electoral systems derive from the decisions that the ruling parties
make to maximize their representation. As long as the electoral arena remains
unchanged and the ruling parties are favoured by the current rules, the electoral
system is not altered. As the electoral arena changes, electoral reform becomes
desirable for the ruling parties. The type of electoral reform that is preferred
depends on the strength of the new parties and the coordinating capacities of the
old parties. When the new parties are weak, the majoritarian system is main-
tained. When the new entrants are strong, a PR system is adopted if no old party
enjoys a dominant position. However, if there is a dominant old party, an elect-
oral system change is not likely to take place (Boix 1999: 609–24). Thus, elect-
 The rational perspective   47
oral system choice is determined by strategic calculations of the old-­established
parties. These calculations are, in turn, based on suffrage extension, the electoral
arena and power constellations between the old parties internally as well as
between old and new parties.
This work is not concerned with strategic calculations of political actors or
decision-­making concerning the adoption of electoral systems. Yet, the prevail-
ing conditions in the bargaining process are related to a structural dimension,
namely the party system, which is relevant to the present study. The aforesaid
suggests that the result of a negotiating process with two main competitors will
most likely be a majoritarian system. Likewise, we may expect a proportional
system to be adopted if many participants – that is, several political parties –
take part in the decision-­making process. In other words, the structure of the
party system may be regarded as an explanatory factor. Accordingly, the party
system is both an actor-­related and a structurally generated factor. It is primarily
an actor-­related matter and secondary a structural phenomenon, because the
party system consists of political parties that make decisions on the electoral
rules, and the decision depends on the number of participants and their mutual
strength. Of course, the decision depends on other factors, too – the level of
uncertainty for instance – but concerning the structural dimension, the party
system is decisive.

4.5 Party system structure

The influence of electoral systems on party systems has been given much atten-
tion. Fragmented party systems are regarded as a consequence of PR, whereas
two-­party systems are considered as the product of plurality systems. Duverger
(1951) was the first to present theoretically founded arguments for the relation-
ship between plurality electoral systems and two-­party systems. In the seminal
work Les Partis Politiques, Duverger also maintained that the two-­ballot major-
ity formula favours multi-­party systems, yet admitted that the relationship was
rather difficult to define (1964: 239). Later research has shown that the effect of
the two-­ballot formula on party system structure is more similar to the effect of
plurality systems than that of proportional ones. The defractionalizing effect of
plurality systems and fractionalizing effect of PR on party systems has been fre-
quently confirmed (e.g. Grofman and Lijphart 1986; Lijphart 1994; Rae 1967;
Taagepera and Shugart 1989).
However, there are those who maintain that the direction of causality is
reversed. The choice of electoral system is determined by the party system,
which, in turn, is shaped by the social cleavages within a society. This insight
belongs to the sociological approach as opposed to the institutional approach that
stresses the influence of electoral structures on party competition. Arguments for
reversed causality between electoral systems and party systems were also put
forward before Lipset and Rokkan (1967) published their classical work on party
systems and social cleavages. In opposition to Duverger’s propositions concern-
ing the effects of electoral systems, John Grumm maintained that proportional
48   The rational perspective
systems are a consequence rather than a determinant of the party system (1958:
375).
In the 1980s, Bogdanor asserted that any theory that emphasizes the electoral
system as a fundamental determinant in shaping party systems cannot be sus-
tained (1983: 254). The dynamics of electoral systems and party systems consti-
tute a complex and reciprocal entity – a point to which I shall return shortly.
Referring to a thesis by Dieter Nohlen (1993a: 27), Gary W. Cox writes that the
number of social cleavages affects not only the number of parties but also the
kind of electoral system (majoritarian or proportional) a given country chooses
(1997: 19). Following Grumm (1958), Cox argues that the greater the social
fragmentation, the more likely the emergence of a multi-­party system and the
adoption of a proportional electoral system. The greater the social homogeneity,
the more likely the choice of a plurality system and the emergence of a two-­
party system or, at the most, limited party pluralism (Cox 1997: 19; see also
Colomer 2004b: 3–5). In the concluding chapter of a rich volume on the elect-
oral systems in 22 democracies, Michael Gallagher (2005: 538) points out that in
several Western European countries multi-­party systems already existed when
majoritarian electoral systems were replaced by list PR, and that the multi-­party
system, on its part, was often a consequence of ethnic, linguistic and religious
heterogeneity.
Stanislaw Gebethner (1996) also supports the view that party systems precede
electoral systems. Where a multi-­party system already exists, Gebethner argues,
PR is likely to be chosen in order to maintain political stability by avoiding large
vote–seat discrepancies and preventing winner-­take-all politics, or saving minor-
ity parties from eternal exclusion from power. In a study of electoral system
choice in Poland after the collapse of Communism, the author points out that
there is no possibility of creating a well-­functioning two-­party system in post-­
Communist countries; it could eventually lead to an authoritarian system of gov-
ernment. PR is the only way to incorporate the majority of the various political
groups and create political stability (Gebethner 1996).
Colomer (2005) argues that the number of parties explains the choice of elect-
oral systems rather than the other way around, thereby turning Duverger’s laws
upside down. He presents the concept of ‘behavioural–institutional equilibrium’
concerning the relation between party systems and electoral systems. The behav-
ioural approach suggests that political actors adapt to the prevailing electoral
rules by making new efforts of collective action, splitting from parties or enter-
ing previously existing parties in order to obtain electoral success. The institu-
tional approach is concerned with the ability of political parties to bring about a
change in the electoral rules, the primary alternatives being a majoritarian or a
proportional system. Colomer’s dataset comprises 219 elections since the nine-
teenth century in 87 countries with some democratic experience. By comparing
the level of party system fragmentation in 37 countries prior to electoral system
change from plurality/majority to PR or mixed systems with 33 countries that
use majoritarian systems in periods of electoral system stability, the author finds
proof of the hypothesis that party systems determine electoral system choice.
 The rational perspective   49
The ‘effective number of parties’ is 3.9 and 2.8 respectively. The analysis shows
that there is a statistically significant relationship between the number of parties
and the probability of an electoral system change from plurality/majority to
mixed/PR.
For each country in the comparison group of ‘stable’ electoral systems,
Colomer has chosen the earliest initial electoral experience under the majoritar-
ian rule. However, I find the decision to regard the first experience of an elect-
oral system as a stable environment rather dubious. As mentioned earlier,
electoral reform is likely to occur after the first elections but less likely after the
second elections, which means that the first elections are often held in unstable
circumstances regarding electoral institutions. Moreover, several subsequent
democratic elections are usually required to provide a stable institutional setting
for the purpose of studying effects of electoral systems. Colomer also provides
data on the latest elections in all countries that are included in the study. The
average ‘effective number of parties’ among majoritarian systems is 3.6, which
the author regards as an indicator of party systems becoming more fragmented
over time. However, this sample might also be more appropriate for comparing
the level of party system fragmentation (3.6) among majoritarian systems with
the level of fragmentation (3.9) among cases of electoral system change. Such a
comparison would suggest that the influence of party systems on electoral
system choice is rather insignificant. Notwithstanding, the assessment that pro-
portional systems will be even more popular in the future is strongly supported
by the empirical data (Colomer 2005: 9–18).
Several experts have emphasized that social and electoral structures interact
in the process of party formation and maintenance. Nohlen (1996: 45), for
example, asserts that there is a circular causal relationship between the electoral
system and the party system. Although electoral rules affect the party system,
the adoption of an electoral system to a great extent depends on the party system
as it exists when decisions on electoral system choice are made. Lijphart (1991:
73) points out that plurality elections certainly favour the maintenance of a two-­
party system, but an existing two-­party system also favours the maintenance of
plurality elections, which gives the two main parties great advantages that they
are unlikely to abandon. Peter Ordeshook and Olga Shvetsova (1994) have found
that the number of parties in a country increases with both the fragmentation of
the social structure and with the proportionality of the electoral system, but also
that these factors interact. Increased proportionality produces more fragmented
party systems in heterogeneous societies but not in homogeneous ones. In the
same way, increasing the cultural heterogeneity in a majoritarian system does
not affect the party system, whereas it does so in a proportional system. Both
social cleavages and electoral structures obviously matter.
According to the theory of party systems and their effect on electoral systems,
countries that adopt proportional systems should have fragmented party systems
and countries that adopt plurality systems should have few relevant political
parties. However, the degree of party system fragmentation can only be accu-
rately measured in countries that have already held elections, which implies that
50   The rational perspective
cases of electoral system change must be analysed. Consequently, the expected
causality is not likely to emerge, since electoral rules affect the party system
structure, that is, countries that have previously held elections under PR are
likely to have fragmented party systems and vice versa. Therefore, the following
assumption is made: countries that replace plurality with proportional or mixed
systems have more fragmented party systems than countries that retain plurality
systems; and countries that replace PR with plurality or mixed systems have less
fragmented party systems than countries that stick to proportional elections.4
For a long time, classification of party systems according to the number of
parties was almost universally prevailing. Nowadays, the common way of classi-
fying party systems is to pay attention not only to the number of parties but also
to their relative strength, that is, the vote (or seat) shares for each party. In 1967,
Rae constructed an index of party system fractionalization, ranging from 0 to 1.
Low values indicate a low degree of party system fragmentation and vice versa.
Rae’s index has later on been modified. ‘The effective number of parties’,
created by Markku Laakso and Taagepera (1979), is perhaps the most widely
used measure of party system fragmentation. The effective number of parties is
calculated according to the following formula: divide one by the sum of the
squared vote (or seat) share of each party in the previous elections.
The smallest possible value is 1; no upper limit exists. The effective number
of parties can be calculated on the basis of either vote shares or seat shares. In
this case, vote shares are more appropriate. If the party system is assumed to
affect the choice of electoral system, the electoral support reflected in the vote
share of each party is the crucial factor. If the effective number of parties is cal-
culated on the basis of seat shares, each value is influenced by the electoral
system used in the previous elections, because the electoral rules determine the
allocation of seats. In the theoretical survey of party systems and electoral
systems, it was argued that both variables affect each other. A fragmented party
system might be the consequence of a proportional electoral system as well as
the initial reason for choosing that kind of system. The application of vote shares
rather than seat shares decreases the influence of the former electoral system on
the party system structure.
According to the theory of electoral systems and party systems, the defrac-
tionalized party system in plurality systems is partly a consequence of the dis-
proportional seat allocation between parties, which, in turn, is caused by the use
of single-­member districts. For the same reason, a high degree of disproportion-
ality is also associated with majority systems. However, a number of studies
indicate that the effective number of parties is higher among majority systems
than among plurality systems (C. Anckar 2002; Blais et al. 2005; Lijphart 1994;
Rae 1967). Therefore, majority systems are not included.
The criterion for inclusion in the comparison groups is that the same electoral
system has been applied in at least three subsequent legislative elections. Party
system fragmentation in the countries of the world at the end of 2003, or follow-
ing the last available election results, is observed. In order to get a higher level
of party system stability in the categories of electoral system change, two elec-
 The rational perspective   51
tions prior to each reform are included. The association between party system
fragmentation and electoral system choice is analysed in the total data sample as
well as in democracies. Usually, however, this matter is studied in democracies
only, because the range of choice is limited in non-­democracies, which affects
the party system structure. Consequently, the democratic sample should be con-
sidered as more important than the larger sample.

Empirical evidence
The number of elections in the category of changes from plurality to PR is 15,
conducted in nine countries. Needless to say, the number of countries that have
changed electoral systems in that direction is much larger, but several countries
are disqualified because one-­party rule has prevailed before the electoral reform.
Furthermore, some countries are excluded because data has not been accessible.
These circumstances also apply to countries that have changed from PR to less
proportional systems.
Results are given in Table 4.7. As hypothesized, the degree of party system
fragmentation is higher in the category of electoral system change from plurality
to PR than in the category of ‘retained’ plurality systems, but the difference is
rather small and not statistically significant. The most fragmented party system
in the group of electoral system change is observed in Morocco: the effective
number of parties was 10.71 following the last plurality elections in 1997. The
category also consists of several low-­fragmented party systems, for instance in
Argentina, El Salvador and Cape Verde. The category of electoral system change
from plurality to mixed or proportional systems consists of 33 cases. The level
of party system fragmentation in the enlarged category, likewise presented in
Table 4.7, is somewhat higher than in the category of change from plurality to
PR – yet, the difference to the group of plurality systems is not statistically
significant.
In Table 4.8, countries that have replaced PR with plurality systems (and
mixed systems) are compared to ‘retained’ proportional systems in the countries
of the world. There is a significant difference between the groups in the first
comparison but the relationship contradicts the assumption, that is, the category

Table 4.7 Party system fragmentation in plurality systems

Electoral system Party system N Std. deviation

fragmentation

Change from plurality to PR 3.28 15 2.84

Plurality systems 2.79 27 1.10
Change from plurality to mixed and PR 3.43 33 2.22
Plurality systems 2.79 27 1.10

Notes
First comparison: eta squared 0.016, sig. 0.429.
Second comparison: eta squared 0.031, sig. 0.178.
52   The rational perspective
Table 4.8 Party system fragmentation in proportional systems

Electoral system Party system N Std. deviation

fragmentation

Change from PR to plurality 7.26   4 2.08

Proportional systems 4.38 53 2.01
Change from PR to plurality and mixed 4.73 11 2.55
Proportional systems 4.38 53 2.01

Notes
First comparison: eta squared 0.123, sig. 0.007.
Second comparison: eta squared 0.004, sig. 0.602.

Table 4.9 Party system fragmentation in democracies with plurality systems

Electoral system Party system N Std. deviation

fragmentation

Change from plurality to PR 2.52   5 0.65

Plurality systems 2.60 17 0.68
Change from plurality to mixed and PR 3.66 14 1.41
Plurality systems 2.60 17 0.68

Notes
First comparison: eta squared 0.003, sig. 0.820.
Second comparison: eta squared 0.205, sig. 0.011.

of change from PR to plurality returns a higher degree of party system frag-

mentation than the comparison group. However, the category of electoral system
change consists of elections in two countries only: Greece and Congo (–Kin-
shasa). Likewise, the level of party system fragmentation in the category of
countries that have replaced PR with plurality or mixed systems is higher than
among proportional systems in general. The results of these comparisons suggest
that party system structure does not influence electoral system choice.
Next, the democratic sample is examined. There is practically no difference
between the groups in the first comparison in Table 4.9. However, only three
democratic countries (Turkey, Cyprus and Sri Lanka) have replaced plurality
systems with proportional systems. The small standard deviation values indicate
that there are only small discrepancies from the mean values of the groups. The
second comparison that includes mixed systems supports the thesis that party
system structure influences electoral system choice: countries that introduce
electoral reforms implying a higher degree of proportionality have more frag-
mented party systems than countries with stable plurality systems. The differ-
ence between the two groups is almost significant at the 0.01 level. The effective
number of parties was almost five in both Thailand and the Philippines when
plurality systems were replaced with mixed-­superposition systems in 1997 and
1998 respectively. Also, New Zealand had a much more fragmented party
 The rational perspective   53
Table 4.10 Party system fragmentation in democracies with proportional systems

Electoral system Party system N Std. deviation

fragmentation

Change from PR to plurality 6.08   2 2.73

Proportional systems 4.11 44 1.75
Change from PR to plurality and mixed 5.18   6 1.83
Proportional systems 4.11 44 1.75

Notes
First comparison: eta squared 0.051, sig. 0.131.
Second comparison: eta squared 0.039, sig. 0.167.

system than countries with plurality systems in general when a mixed-­corrective

system was adopted in 1993.
Referring to Table 4.10, the category of change from PR to plurality has a
higher degree of party system fragmentation than the comparison group.
However, the former category consists of two Greek elections only, which
makes the comparison as good as futile. The Greek legislative elections in 1950
and 1951 returned the values of 8.01 and 4.15 respectively. The inclusion of
changes from PR to mixed systems does not alter the picture; the level of party
system fragmentation is still higher in the category of electoral system change
than in the PR category.
A total of eight comparisons have been conducted. The thesis of party system
structure influencing electoral system choice is supported to a certain extent.
Countries that replace plurality systems with mixed or proportional systems have
more fragmented party systems than countries with plurality systems in general.
The difference is significant among democracies only, which nonetheless must
be considered as the more important sample. A relationship between party
system fragmentation and electoral system choice does not exist when changes
from PR to plurality systems are compared to proportional systems in general.
The expanded samples with changes to mixed systems confirm these tendencies.
The analysis of change from PR to less proportional systems is marred by a
small number of cases, which largely makes the comparison meaningless. Since
these comparisons are based on cases of electoral system change, party system
structure cannot be included in a multiple analysis. The analyses above confirm
that the party system is not one of the main determinants of electoral system
choice. However, it should be emphasized that the level of party system frag-
mentation seems to have some influence on changes from plurality to more pro-
portional systems.

4.6 Party system transformation

The introduction of PR in continental Europe comprised a ‘minority protection’
phase as well as an ‘anti-­socialist’ phase. The former was concerned with
­providing minority representation in multicultural societies as a means of
54   The rational perspective
c­ ounteracting potential threats to national unity and political stability. As a con-
sequence of extended suffrage, demands for proportional elections were also
heard in homogeneous countries. The introduction of PR came about as

A convergence of pressures from below and from above. The rising working
class wanted to lower the thresholds of representation in order to gain access
to the legislatures, and the most threatened of the old-­established parties
demanded PR to protect their position against the new waves of mobilized
voters created by universal suffrage.
(Rokkan 1970: 157)

Anxious about a complete socialist takeover, a switch from all-­or-nothing major-

itarianism to PR would at least deliver a modest return for the old parties.
Andrew Carstairs (1980: 215) points out that wherever PR was recognized as a
practical possibility, the social democrats and the socialists tended at first to be
advocates of an electoral reform. However, when the social democrats realized
that the introduction of universal manhood suffrage would most likely give them
the status of a major or dominant party, they turned against such a reform. The
bourgeois and non-­socialist parties became now the main advocates of propor-
tional representation. Thus, PR constituted a rational solution with regard to both
the preferences of political actors and the requirements in culturally hetero-
geneous societies.
In a study of constitutional choices in some Eastern European countries after
the breakdown of Communism, Lijphart (1992a: 207–23) finds that the ‘Rokkan
hypothesis’ of demand for PR as a consequence of the democratization process
still applies. Both the old Communist parties and their challengers in Czecho­
slovakia, Poland and Hungary needed PR to protect their interests when demo-
cracy evolved in the beginning of the 1990s. The post-­Communist transition
gave representatives of the old regime reason to consent to PR reform once they
realized that their position was under serious threat.
The demand for PR at the turn of the twentieth century was partly a con-
sequence of the extended suffrage. In Eastern Europe, however, universal suf-
frage had, at the time of the Communist collapse, prevailed for a long time.
Instead, the demand for proportional elections was a result of another crucial
element of the democratization process, namely the right to establish political
parties and participate in popular elections. During the Communist era, the use
of majoritarian systems was self-­evident, because elections were non-­
competitive. The democratization process resulted in a large number of parties
and, as a consequence thereof, a switch from majoritarian to proportional elec-
tions. More specifically, we may speak of party system transformation from a
one-­party system to a multi-­party system.5 The crucial feature is the end of one-­
party rule and the right for citizens to establish political parties. During the time
period that this work covers, universal suffrage has been a common feature for
most of the independent countries in which parliamentary elections have been
held. Consequently, this issue is not included in the empirical analysis. Rather,
 The rational perspective   55
transformation from a one-­party system to a competitive party system represents
the dynamic of the democratization process.
Party system transformation is foremost an actor-­related matter but has a
structural dimension too, which justifies the inclusion of this variable in the anal-
ysis. As long as all seats are allocated to candidates from a single political group,
there is no need to manipulate the electoral system in order to maximize the
electoral support – majoritarian elections prevail. When the ban on political
parties is abolished and competitive elections are introduced, the electoral
system becomes an effective manipulative instrument for the political actors,
especially the ruling party, which is concerned about retaining as much as pos-
sible of its former power position. Simultaneously, party system transformation
is a structural phenomenon: absolute one-­party rule is replaced by a multi-­party
system and a competitive political arena.
This dimension is a cornerstone of the democratization process, albeit it by no
means guarantees the realization of democratic politics. In several one-­party
states, elections have been held despite the fact that no competition has been
allowed. These elections have, with a few exceptions,6 been held under a majori-
tarian system, since there is no need for proportional seat allocation in a party
system with only one party. Countries that have lifted their ban on political
parties (other than the ruling organization) are, according to this hypothesis,
expected to adopt mixed or proportional systems. This is primarily regarded as a
consequence of the (former) ruling party fearing the loss of power if a majoritar-
ian winner-­take-all system is maintained. In addition, the new parties may feel
insecure about their possibilities of electoral success, and as a consequence,
prefer proportional seat allocation.
The units of analysis are all cases of transformation from a one-­party to a multi-­
party system between 1945 and 2003. Cases where the multi-­party system is pre-
ceded by military rule or other kinds of dictatorship without elections are also
included, provided that elections have previously been held and the former elect-
oral system was majoritarian. A prerequisite is, however, that the former ruling
organization is still present at the political arena when the party system is trans-
formed. Some former Soviet states are also included, although the transformation
of the party system took place shortly before independence was proclaimed. Yet,
the introduction of multi-­party systems in these countries was part of the independ-
ence process, and they are, furthermore, important cases of bargaining between the
old elite and the new contestants. The prerequisite of former elections under the
majoritarian rule is fulfilled, because pre-­independence elections were held in all
these countries. Systems characterized by some degree of competition between
factions within the sole party are regarded as one-­party systems.
The introduction of multi-­party politics is assumed to influence the choice of
electoral system. In some cases, however, the distance between the independent
and the dependent variable may be diminutive, since both the transformation and
the electoral reform are regarded as part of the democratization process. Both
reforms may actually be introduced simultaneously as part of a new constitution.
Notwithstanding, I argue that the choice of electoral system is a consequence of
56   The rational perspective
the fundamental changes in the party system structure. The electoral system is a
means of dealing with a new environment consisting of several political parties.
From a democratic point of view, the introduction of competitive elections takes
precedence over the choice of electoral provisions – the former is more import-
ant than the latter. In addition, from a logical point of view, the regime hardly
decides on the adoption of a proportional system before it decides to lift the ban
on political parties.
Party system transformation is closely related to the variable that measures
the level of party system fragmentation. A crucial difference is that party system
fragmentation is concerned with countries that have previously held competitive
or at least semi-­competitive elections, whereas the transformation variable pays
attention to countries that introduce multi-­party politics. In addition, party
system transformation is, as mentioned earlier, particularly concerned with the
democratization process. Nevertheless, both assume that the party system influ-
ences the choice of electoral system.

Empirical evidence
There are a total of 62 cases of party system transformation. Note that countries,
in which the former regime no longer exists at the political arena when the party
system is transformed, are excluded. A proportional or mixed electoral system
has been adopted simultaneously as the transformation of the party system took
place, or between the transformation and the next elections, in the following
countries:

Bulgaria (1990), Burkina Faso (1970), Burundi (1992), Cambodia (1993), Cam-
eroon (1992), Chad (1991), Czechoslovakia (1989), East Germany (1990), Egypt
(1981), Equatorial Guinea (1989), Estonia (1988), Georgia (1990), Guinea
(1992), Guinea-­Bissau (1991), Hungary (1989), Latvia (1988), Lithuania (1988),
Moldavia (1993), Mozambique (1994), Niger (1993), Panama (1980), Romania
(1990), Russia (1993), Sao Tome and Principe (1990), Senegal (1978), Sey-
chelles (1992), Sierra Leone (1996), Slovenia (1991), Yugoslavia (1992).

The year of party system transformation is given in brackets. In the following

countries, the party system transformation was not followed by electoral system
reform as hypothesized:

Albania (1990), Algeria (1989), Armenia (1990), Azerbaijan (1990), Cape Verde
(1990), Central African Republic (1993), Comoros (1990), Congo (– Brazza-
ville) (1991), Côte d’Ivoire (1990), Djibouti (1992), Gabon (1990), Ghana
(1992), Jordan (1992), Kazakhstan (1994), Kenya (1991), Kyrgyzstan (1995),
Lesotho (1993), Malawi (1993), Mauretania (1991), Mongolia (1990), Myanmar
(1990), Nepal (1991), Pakistan (1988), Poland (1989), Tajikistan (1991), Tanza-
nia (1992), Thailand (1975), Togo (1994), Turkey (1946), Ukraine (1993),
Yemen (1993), Zambia (1991), Zimbabwe (1990).
 The rational perspective   57
Almost one-­half (29 of 62) of the countries that have experienced party
system transformation have introduced electoral reforms as expected. Majoritar-
ian systems have been replaced with proportional systems in 19 countries,
whereas ten have adopted mixed electoral provisions. At a first glance, this
might seem like a meagre result. However, one can hardly expect that every
country holding multi-­party elections for the first time also adopts a proportional
(or a mixed) electoral system. Of course, there are several reasons for using a
plurality or a majority system rather than a proportional or a mixed one.
Of the 278 electoral system choices in the study, 138 are cases of electoral
system change. The number of changes from plurality/majority to mixed/propor-
tional systems is 67. Consequently, a large part of these changes have been pre-
ceded by a transformation from a one-­party to a multi-­party system.7 The
frequency of electoral system change as a consequence of party system trans-
formation should therefore be considered as rather high. In addition, nine of the
33 countries that did not introduce mixed or PR systems introduced electoral
reforms as hypothesized prior to the second parliamentary elections after the
introduction of multi-­party politics.8 Compared to the total share of changes from
majoritarian to mixed or proportional systems (24 per cent), the share of similar
reforms as a consequence of party system transformation (47 per cent) is also
quite striking. Party system transformation has preceded about one-­fifth of all
mixed and PR cases in the total data sample. The listing of countries that have
introduced multi-­party politics also shows that party system transformation is a
phenomenon of the 1990s. Of 62 transformations, 47 have occurred between
1990 and 1996. Only four transformations took place before 1980.
All in all, the findings indicate that there is a tendency for countries to intro-
duce mixed or proportional systems when the ban on political parties is lifted
and electoral competition is allowed. The analysis suggests that party system
transformation bears relevance to explaining electoral system choice. However,
this variable cannot be included in a multivariate analysis, because the unit of
analysis is party system transformation instead of electoral system choice. As a
consequence thereof, we cannot be sure that the association would remain in a
control for other determinants. However, according to the theoretical arguments,
party system transformation is a profound explanation to the adoption of mixed
and proportional systems. The decision to adopt a more inclusive electoral
system is often a rational solution to both the former ruling party and the new-­
comers at the parliamentary arena.

4.7 The rational perspective: summary

Four variables within the rational perspective have been analysed: ethnic/cultural
diversity, population size, party system structure and party system transforma-
tion. Concerning ethnic and cultural diversity, a pattern contrary to the expecta-
tions is found. The number of proportional systems and AV, which are regarded
as appropriate for divided societies, is very small among countries with a high
degree of ethnic and cultural fragmentation. In contrast, the frequency of PR and
58   The rational perspective
AV among countries with a low degree of fragmentation is relatively high.
However, the pattern is rather weak and no significant relationship prevails.
Considering the importance attached to electoral system design in divided socie-
ties, the small share of designed electoral systems according to ethnic and other
cleavages is astonishing.
As to population size, a bivariate association between smallness and plurality
systems is observed. The relationship is stronger among democracies than in the
total data sample. This finding contradicts the theoretical arguments put forward
by several authors. Apparently, there are other explanations of the observed
pattern. Another interesting finding is that despite the positive correlation
between a large population and PR, countries with more than 50 million people
tend to use majoritarian or mixed systems. There seems to be a need for single-­
member districts in very large countries in order to maintain a link between
voters and representatives.
The impact of party system structure on electoral system choice has been
studied by comparing cases of electoral system change to samples of plurality
systems and proportional systems. The expected pattern according to the theory
of party system structure and electoral system choice prevails among countries
that replace plurality systems with mixed or proportional systems. The discov-
ered pattern is more evident and statistically significant in democracies.
However, the analysis of electoral system changes in the opposite direction does
not support the thesis of party system affecting the choice of electoral system.
Consequently, on the basis of this study, the effect of party system structure on
electoral system choice in general must be considered as rather limited. A larger
dataset would probably have shed more light on the relationship between party
system fragmentation and electoral system choice.
A total of 62 cases of party system transformation are identified. In almost
one-­half of them, the expected electoral reforms have taken place. The share of
electoral system change might be considered quite large for three reasons. First,
almost one-­half of all major electoral reforms towards mixed and proportional
systems have been preceded by party system transformation. Second, the share
of changes from majoritarian to PR/mixed systems in connection to or directly
after party system transformation is practically twice as large as the share of
similar reforms in the total data sample. Third, approximately one-­fifth of all
mixed and PR choices in this study is associated with this variable. Party system
transformation is, in other words, a factor that determines electoral system
choice, thereby increasing the importance of the party system in general as well,
although no striking pattern was observed with regard to party system structure.
5 History and culture
Patterns of diffusion

The second approach maintains that electoral systems, as well as other constitu-
tional and institutional features, reflect the cultural and historical contexts to
which they belong. This point of view has been stressed, among others, by
Lijphart (1991; 1992b) and Lipset (1992). In Lijphart’s introduction to Parlia-
mentary versus Presidential Government (Lijphart 1992b; 1992c), the congru-
ency between particular constitutional regimes and sharply delineated
cultural-­regional settings is emphasized. In the same volume, Lipset (1992:
207–11) brings out the importance of political culture with respect to democratic
tradition. With reference to Belize and Canada as the only states in the mainland
region of the Americas having other executive systems than the (limited) presi-
dential, D. Anckar maintains that regime diffusion can also be a consequence of
a colonial past (2004: 207). A bird’s eye view of the map of the world suggests
that electoral system choices indeed seem to be dependent on the cultural, histor-
ical and regional context. Proportional systems dominate in South America and
Europe, whereas majoritarian systems prevail in the Caribbean, the Pacific and
large parts of Asia.
The concept of diffusion is a complex matter, and the way it is approached in
this study is derived from a greater context. Diffusion is concerned with a famil-
iar problem in the social sciences, usually referred to as ‘Galton’s problem’,
which is the methodological problem of sorting out emulation of social and
political phenomena from other causes of variation in social systems (Peters
1998: 42). In trying to explain the spread of democracy and its considerable shal-
lowness, Larry Diamond (1999) presents five models of diffusion. The first one
is called the power model: the most powerful democracies of today press the
formal model of electoral democracy on the weaker states over which they have
influence. The second model rests on more comprehensive efforts on the part of
established democracies as well as international organizations. The third model
of diffusion is concerned with imitation of political models that are perceived as
highly successful, powerful and prestigious. The fourth model rests upon norm-
ative judgements on what is intrinsically good, right and desirable. The last
model partly encompasses the other four but includes a trend towards global
standardization too, that is, cognitive models that have produced increasingly
similar state structures (Diamond 1999: 56–8).
60   History and culture
Of course, diffusion of electoral systems is not the same as diffusion of demo-
cracy. Democracy is a process, which is established and consolidated often
during a long time, whereas an electoral system is a set of formal rules of how to
manage elections. Nonetheless, the frame of reference above is useful in deter-
mining what electoral system choice emanating from diffusion implies. In order
to broaden our knowledge of diffusion as an explanatory factor of political phe-
nomena, I shall present another theoretical frame of reference that partly over-
laps with Diamond’s framework.
In an article on diffusion of public policy, Kurt Weyland (2005: 262–95)
presents a theoretical framework of policy diffusion that consists of four altern-
ative models: foreign pressures, symbolic and normative imitation, rational
learning and cognitive heuristics. More exactly, the article aims at exploring
which of these approaches drives waves of diffusion the most. The explanatory
framework can be summarized in the following way. A policy innovation ema-
nates either from external pressure or domestic initiative. The latter may take
two forms: a quest for legitimacy or a pursuit of interests. Rational learning and
cognitive heuristics are two alternative types of pursuing interests. These models
are theoretically as well as empirically analysed, the empirical evidence coming
from the area of social security reform in Latin America. Nevertheless, the ana-
lytical framework bears relevance to other areas of diffusion as well – institu-
tional choice is one of them.
Weyland points out that processes of diffusion have three major character-
istics: diffusion occurs in a distinct wave, it tends to have a clear geographical
concentration, and it produces commonality in diversity. The wavelike process
yields an S-­shaped pattern by starting slowly, then gathering speed, and finally
tapering off. In spatial terms, it initially spreads through the region in which the
innovation was born, and thereafter reaches other parts of the world. In substan-
tive terms, diffusion entails the adoption of the same innovation in diverse
settings.
The external pressure model is concerned with the influence of powerful
external actors, particularly international organizations, on single countries by
means of incentives, sanctions and coercion. While acknowledging the signifi-
cance of external pressures in some contexts, the author maintains that the capac-
ity of this model to explain all basic features of diffusion is rather limited. The
normative imitation approach regards the import of advanced innovations as an
attempt by the decision-­makers to gain international legitimacy as well as an
ambition to demonstrate the country’s modernity and accommodation to interna-
tional standards. The rapid spread of innovations is partly driven by a desire to
impress global public opinion. Countries tend to keep up with new trends before
they have enough information on the likely effects of the reform. This model
succeeds rather well in explaining different aspects of policy diffusion, particu-
larly the puzzle of commonality amid diversity. The rational learning model sees
political action as a goal-­oriented choice driven by decision-­makers’ self-­
regarding interests. If a promising foreign model is identified and its superiority
over established policies is confirmed by means of cost–benefit analysis, the
 History and culture   61
model is likely to be adopted. However, the model confronts difficulties in
explaining both commonality in diversity and the geographical clustering of
diffusion.
Deriving from cognitive-­psychological theory, the cognitive heuristics model
also regards diffusion as a consequence of goal-­oriented activities driven by self-
­interests of political actors. However, in contrast to the rational learning
approach, the cognitive heuristics framework maintains that external models are
attractive because of their apparent promise to solve problems rather than their
demonstrated success. This approach stands out as superior to the other three
models in explaining the three main features of diffusion, that is, temporal
sequence, geographical clustering and the substantive nature of spreading com-
monality amid diversity. Finally, Weyland maintains that the cognitive heuristics
approach to explaining diffusion has wider theoretical significance. It sheds light
on political phenomena in the real world and broadens our understanding of
political decision-­making.
It should be pointed out that no analysis of diffusion in its proper sense is
provided in the present study. A case study of each process of diffusion is out of
the scope of the book. Consequently, I shall not compare these four explanatory
models and make conclusions on how well they are capable of explaining exist-
ing patterns in the world of electoral systems. The empirical analysis is restricted
to observing whether processes of diffusion have occurred or not. Nevertheless,
the theoretical discussion above is important to the understanding of diffusion as
a general explanation of institutional choices. Furthermore, it lays down the
general outlines for an operationalization of diffusion.
Three variables within the cultural and historical approach are examined:
colonial diffusion, regional diffusion and temporal diffusion. Timing of inde-
pendence, mentioned by Anckar and Karvonen (2004: 8) as the third factor
within the cultural and historical perspective, is of secondary importance in this
context. To be sure, timing of independence largely coincides with timing of
electoral system choice, because a country usually adopts an electoral system
and holds elections when it becomes independent. However, by paying attention
to the year of independence, we do not observe electoral system changes. There-
fore, timing of electoral system choice as an indicator of temporal diffusion is
more appropriate. Colonial, regional and temporal diffusion represent different
forms of imitation that together comprise the three main characteristics of policy
diffusion presented above.
Colonial diffusion of electoral systems shows similarities to the external pres-
sure model in Weyland’s theoretical framework of policy diffusion. Electoral
systems are transferred from a colonial power to countries that do not necessarily
share the same characteristics that otherwise would explain the adoption of
similar institutions, thereby dealing with commonality amid diversity. Regional
diffusion is concerned with the geographical clustering of similar reforms. Tem-
poral diffusion comprises the wavelike spread of policies and institutions.
Returning to the analytical framework by Diamond (1999), some models are
obviously more appropriate for the purpose of this study. The transfer of an
62   History and culture
electoral system from the colonial power to its colony resembles the power
model: powerful countries have introduced their electoral systems to countries
over which they hold sway. Regional diffusion corresponds to the third model:
the institutional choices made by successful and powerful countries tend to be
attractive to small and less prestigious countries. The second, fourth and fifth
model described by Diamond are to a much smaller extent concerned with elect-
oral system choice as a consequence of diffusion. The distinction between colo-
nial legacy on the one hand and regional influence on the other appears, in the
light of Diamond’s frame of reference, natural and justified.

5.1 Colonial diffusion

Colonial legacy is often emphasized as a central determinant of electoral system
choice. To a varying extent the colonial powers left behind political traditions
and political institutions that reflected their own constitutional models. The
United Kingdom in particular, which along with France possessed the largest
colonial empires, is said to have implemented a political tradition and a political
culture that has been beneficial for its colonies in the post-­colonial period. Other
foreign rulers were Spain, Portugal, the Netherlands, Germany, Italy, Australia,
New Zealand, the United States and Japan. In several colonies, the electoral
system was introduced by the metropolitan powers, and at the time of independ-
ence chosen as the only familiar option. In some colonies, the electoral system
choice was a result of imitation. The training of local elites in the metropolis
brought them into contact with a specific set of electoral arrangements that they
were later inclined to emulate in their native country.
However, not all former colonies have adopted the electoral system of their
colonial rulers. According to the literature on electoral systems, it is notably the
electoral systems of many former British territories that are regarded as a colo-
nial legacy. Several electoral system choices in former French colonies are also
considered as a product of French colonial rule. Reilly and Reynolds (1999:
23–4) point out that most of the former French territories use either the con-
temporary French TBM system, or list PR, which France has applied in parlia-
mentary elections 1945, 1946 and 1986, and widely in municipal elections.
Furthermore, Reilly and Reynolds observe a clear pattern of proportional
systems in Spanish-­speaking and Lusophone countries. Furthermore, they main-
tain that TBM, which was used in the Soviet Union, is dominant among the
former Soviet Republics of the Commonwealth of Independent States. In the fol-
lowing, I shall shortly deal with the different colonial empires and the matter of
colonial legacy as a determinant of electoral system choice.
Different colonial powers administered their societies differently. For
instance, the impact of British colonial rule was commonly regarded as positive
and developmental, whereas colonial rule by the Spanish and Portuguese was
considered intrusive and exploitative. Patterns of administration by the same
colonial power also varied between different territories, depending on what the
colony had to offer. While most British colonies were indirectly ruled, Burma
 History and culture   63
was governed directly by the British and has been more authoritarian since inde-
pendence than most of the other former British colonies. Likewise, Spanish and
Portuguese control in colonies with small slave markets and poor mineral
resources was less penetrating and authoritarian than Iberian colonial rule in
general (Diamond and Lipset 1995: 263). Rudolf von Albertini (1982: 489–90)
points out that the difference between French centralization and direct rule, on
the one hand, and British decentralization and indirect rule, on the other, should
not be exaggerated. In practice, the boundaries between the two were fluid. At
various times and places, both systems were employed by all colonial powers. In
Cambodia and Laos, for instance, France practised indirect rule.
However, the concepts of direct or indirect rule are important as indications
of divergent aims of the colonizers. British colonial administration was to a large
extent based on indirect rule. They established the rule of law through effective
(and indigenous) bureaucratic and judicial institutions, and provisions for some
system of representation and election, which gave native elites experience in
political leadership and limited governance. As a consequence, democratic pol-
itics has managed to evolve and survive in large parts of the former British
Empire. The resulting legacy was, according to Myron Weiner (1987), effective
political institutions and an enduring cultural commitment to the procedures of
democratic politics and governance, as well as to the rule of law as a constraint
on government. Especially the early British colonization of Asia and the Carib-
bean, along with some other islands such as Mauritius, has been advantageous to
these countries by giving them long and deep contact with British values and
institutions as well as more time for the gradual emergence of indigenous repre-
sentative institutions. In continental Africa, on the other hand, British coloniza-
tion occurred later and its withdrawal was more hurried. Therefore, British
influence on post-­colonial government was weaker in these countries (Diamond
and Lipset 1995: 263). The slow movement from British supervision to local
leadership that occurred in most parts of the British colonial empire promoted
the development of parliamentary democracy.
French colonial rule also resulted in democratic traditions and values in some
countries but not to the same degree as British colonial rule. Contrary to British
colonial practices, France governed its colonies directly from the capital. Tradi-
tional rulers were usually undermined and replaced by ‘straw chiefs’ that served
at the bottom of the colonial administrative hierarchy. The colonies were
strongly tied to France, and little attention was paid to the development of indi-
genous political and administrative abilities. Not until 1956 were broadly
elected indigenous assemblies established in its African territories. Despite the
differences in colonial policies, both Britain and France left behind political
institutions that reflected their own constitutional models. Diamond and Lipset
(1995: 266) point out that the British transferred their parliamentary system and
their electoral system of single-­member districts, whereas the French transferred
the presidentialism of the Fifth Republic and provided for assembly elections
with party lists on winner-­take-all basis. Furthermore, British colonial rule
resulted in constitutions with emphasis on dispersed power and autonomous
64   History and culture
local government, whereas France colonial rule led to centralist and unitary
constitutions.
The legacy of Spanish and Portuguese colonization in Latin America is quite
difficult to assess, since almost two centuries of independence have seen the
colonial heritage recede in importance. Political instability and lack of demo-
cratic politics have characterized most of the Latin American countries. Those
countries (Chile, Costa Rica and Uruguay) in which Spanish colonial presence
and influence were weakest have been most successful in maintaining demo-
cratic politics. Portugal was the most repressive of all colonial powers in Africa.
When the Portuguese were forced to withdraw from its colonies in the mid-­
1970s, practically no preparation for independence had been made. Neither was
Belgian Congo prepared for self-­government in 1960 when Belgium had to give
up its intention of holding on to the colony indefinitely. All former Portuguese
and Belgian colonies in Africa have, as a consequence of intrusive and exploita-
tive colonial politics, been characterized by extensive repression, ethnic turmoil,
and political instability during the post-­colonial era.
The Australian and American colonial rule resembles the British pattern,
especially in Papua New Guinea and the Philippines. The Americans put a lot of
effort into training the Filipino people in democratic citizenship. They left
behind some important developmental and institutional legacies, which unfortu-
nately had to struggle against the autocratic legacy of four centuries under
Spanish colonial rule (Diamond and Lipset 1995: 263–7). The other American
colonies had slightly more than one decade of experience of representative pol-
itics when full sovereignty was achieved in the early 1990s. By using AV in
three pre-­independence elections, the highly fragmented society of Papua New
Guinea was well consolidated when independence was attained in 1975. In
Nauru, legislative and executive organs were established shortly before the small
island became independent in 1968.
If colonial background is decisive in explaining the choice of electoral
systems, former colonies are expected to adopt the electoral system that the colo-
nial power applied when independence was received, or the electoral system that
the colonial power established in the colonies. Electoral reform in the former
colonial power later on is not assumed to influence its former colonies. The
influence of colonial powers is probably less significant if the decolonization
took place long before electoral institutions were adopted. During the second
half of the twentieth century, however, electoral provisions were enacted almost
immediately after independence was achieved. In several colonies, representa-
tive institutions, and consequently electoral laws, already existed before the dec-
laration of independence. Those countries that did not adopt electoral provisions
before or directly after independence are, with a few exceptions, those who
never have adopted representative institutions, such as Qatar and the United
Arab Emirates. As a matter of precaution, I shall introduce a time limit of ten
years, within which the influence of colonial powers on former colonies is con-
sidered relevant. In colonies where several colonial powers have been present,
only the electoral system of the last one is regarded as inheritable. In cases of
 History and culture   65
shared colonial rule, both colonial powers are considered as potential sources of
legacy.
As to the classification of electoral systems, a distinction between plurality,
majority, proportional and mixed systems is obviously too broad in this respect.
If the colonial power applies AV and its former colony adopts TBM, the choice
cannot be regarded as a colonial legacy. Similarly, list PR and STV are not con-
sidered as equivalent systems despite the fact that they both are proportional
systems. Former British colonies are expected to apply SMP. Plurality elections
in multi-­member districts (BV) are usually regarded as a British colonial legacy
as well. I follow the same practice. Concerning former French colonies, different
patterns are expected, depending on when the colony received independence.
France used three different electoral systems during the decolonization era: list
PR until 1946, mixed-­coexistence in the 1951 and 1956 parliamentary elections,
and TBM as of 1958. In addition, BV is regarded as a colonial legacy, because
the French introduced plurality elections with party lists in several territories
before independence was granted (Diamond and Lipset 1995: 266). It might
appear controversial to regard both SMP and BV in former British colonies as a
legacy of British colonization, but only the block vote version of plurality
systems as a French colonial legacy. However, the crucial point is that France
has never used the pure SMP formula, which implies that SMP cannot be con-
sidered a legacy of French colonial rule. The same logic applies to former colo-
nies that use a mixture of single- and multi-­member districts: the SMP–BV
formula may be inherited from the United Kingdom but not from France.
Portugal used nationwide list PR for the first time in April 1975 (Flora et al.
2000: 783–4). The Portuguese colonies in Africa received independence the
same year, except for Guinea-­Bissau which declared independence in 1973 and
was recognized by Portugal the following year. However, no former Portuguese
colony in Africa enacted any electoral law before the adoption of PR in Portugal
(Nohlen et al. 1999). The occurrence of list PR is consequently regarded as a
colonial legacy from Portugal. The only country that has been a Spanish colony
during the post-­war period, Equatorial Guinea, became independent in 1968. No
electoral law existed in Spain between 1936 and 1977, which implies that Equa-
torial Guinea could not have inherited its electoral system from Spain.
The occurrence of list PR in former Belgian colonies is regarded as a legacy of
Belgian colonial rule. List PR is also associated with South African, Italian,
Dutch and Indonesian colonial rule. The United States and New Zealand are
assumed to have transferred SMP to their former colonies. If Australian colonial
rule matters to electoral system choice, AV should be applied in its former territ-
ories. Soviet Union and Yugoslavia are also regarded as colonial powers,
although their territories did not have colonial status. Nevertheless, the electoral
system in the nations that emerged from the breakdown of the Soviet Union and
Yugoslavia may be regarded as a legacy from pre-­independence times. Elections
in the Soviet Union were conducted under the TBM formula. SMP, mixed-­
superposition and list PR are considered as a heritage from Yugoslavia, depend-
ing on when the countries declared independence and adopted electoral systems.1
66   History and culture
Empirical evidence
Of a total of 278 cases in the post-­war era, 140 are first choices of electoral
systems, whereas 138 are cases of electoral system change. Of the 140 coun-
tries that have introduced electoral provisions for the first time, some are dis-
qualified because they were never colonies. Some others are excluded because
the first electoral system was chosen more than ten years after the end of the
colonial period. Some 107 countries meet the requirements of having been a
colony during the relevant time period and having adopted an electoral system
within ten years after receiving independence. Of these countries, 69 inherited
the first electoral system from their colonial masters. Concerning the Somali
Republic, both Britain and Italy are regarded as colonial powers, because the
Somalian territory comprises the former British protectorate of Somaliland as
well as the former Italian-­administered UN Trust Territory. Vanuatu, which was
jointly administered by Britain and France before proclaiming independence in
1980, is also related to two colonial powers. Cameroon was administered by
both France and Britain, but only France is relevant to Cameroon’s first choice
of electoral system in 1960, because the British section joined the Republic of
Cameroon in 1961.2
In Table 5.1, the colonial powers are listed together with the number and
share of colonies that inherited their first electoral system from the colonial
power. Almost 80 per cent of the British colonies inherited the plurality formula:
of a total of 35, 26 chose SMP. BV was introduced in Cyprus, Fiji, Kuwait,
Mauritius and Swaziland, whereas India, Jordan, Burma and Tuvalu adopted a
mixture of SMP and BV. A total of ten former British colonies did not inherit
the British system: Kiribati preferred TBM; the Maldives, Bahrain and Vanuatu
adopted SNTV; Malta introduced STV, which nevertheless has its origin in the

Table 5.1 Colonial powers, colonies and colonial legacy

Colonial power Colonies Electoral system % Electoral system %

(N) inherited not inherited

United Kingdom 45 35 77.8 10 22.2

France 26 16 61.5 10 38.5
Portugal 2 0 0.0 2 100.0
Netherlands 2 1 5.0 1 50.0
Belgium 3 2 66.7 1 33.3
Italy 2 1 50.0 1 50.0
United States 4 4 100.0 0 0.0
Australia 2 1 50.0 1 50.0
New Zealand 1 1 100.0 0 0.0
South Africa 1 0 0.0 1 100.0
Indonesia 1 1 100.0 0 0.0
Soviet Union 14 6 42.9 8 57.1
Yugoslavia 4 1 25.0 3 75.0
N 107 69 64.5 38 35.5
 History and culture   67
United Kingdom, but since STV is a proportional formula, I do not regard the
choice of electoral system in independent Malta as a legacy from the United
Kingdom; mixed electoral systems were adopted in Seychelles and Zimbabwe;
Israel, Somalia and Guyana introduced list PR.
Of 26 former French colonies, 16 adopted BV or TBM after this formula was
introduced for French elections in 1958. Since no former French colony adopted
list PR or mixed systems between 1945 and 1958, only BV and TBM have been
inherited from France. Merely one former colony, Comoros, inherited the latter
mentioned system (in 1978) from France. Syria adopted TBM in 1947, but
France used list PR at that time. Some countries (Cambodia, Laos, South
Vietnam and Togo) preferred SMP, Cameroon and Laos introduced SMP–BV,
and Madagascar adopted a mixed electoral system. Morocco was the only former
French colony that introduced a proportional system (list PR). As noted, Vanuatu
adopted SNTV.
The only Portuguese colonies that qualify for this sample – Cape Verde and
Guinea-­Bissau – chose BV instead of list PR, which Portugal had recently
adopted. The other Portuguese colonies in Africa – Mozambique, Angola and
Sao Tome and Principe – did not adopt electoral provisions until the 1990s.
Indonesia inherited list PR from the Netherlands; Suriname, however, preferred
a mixed electoral system. All colonies of the United States, that is, the Philip-
pines, the Federated States of Micronesia, Marshall Islands and Palau, inherited
the plurality formula when independence was attained. The Americans adminis-
tered their colonies much in the same way as the British, which probably
explains the pattern. Western Samoa inherited the plurality method from New
Zealand. Belgium transferred list PR to Congo (– Kinshasa) and Rwanda –
Burundi, the third Belgian colony, adopted SMP. Somalia inherited list PR from
Italy, whereas Libya preferred SMP. Nauru inherited AV from Australia – Papua
New Guinea, by contrast, rejected AV and adopted SMP when independence
was declared. Namibia, the only colony of South Africa, chose list PR a few
years before South Africa introduced the same system. East Timor adopted the
same electoral system as Indonesia applies, namely list PR.
More than one-­half of the former Soviet states abandoned TBM when inde-
pendence was achieved. Notwithstanding, it is noteworthy that Belarus, Kyr-
gyzstan, Tajikistan, Turkmenistan, Ukraine and Uzbekistan retained the old
electoral system as independent states. Only one of the countries that declared
independence from the Federal Republic of Yugoslavia in the 1990s, Bosnia-­
Herzegovina, adopted the same electoral system (in 1996) that simultaneously
existed in Yugoslavia, that is, list PR.
In Table 5.2, all former British, French and Soviet colonies that adopted their
first electoral systems within ten years after the end of the colonial rule are sepa-
rately compared to all other cases. A fourth variable encompassing all former
colonies is also included. There are significant, positive associations between
British and French colonial legacy, on the one hand, and electoral system choice, on
the other, indicating that the former colonies are associated with majoritarian
systems, plurality in particular. Soviet legacy does not correlate with the dependent
68   History and culture
Table 5.2 Associations between colonial diffusion and electoral system choice

Colonial legacy Electoral system choice

British 0.376**
French 0.286**
Soviet –0.073
All cases of colonial legacy 0.499**
N 278

Note
** Significance at the p < 0.01 level.

Table 5.3 Colonial powers, colonies and colonial legacy in democracies

Colonial power Colonies Electoral system % Electoral system %

(N) inherited not inherited

United Kingdom 23 18 78.3 5 21.7

France 2 1 50.0 1 50.0
Netherlands 1 0 0.0 1 100.0
United States 4 4 100.0 0 0.0
Australia 2 1 50.0 1 50.0
Soviet Union 3 0 0.0 3 100.0
N 35 24 68.6 11 31.4

variable. The variable that observes all former colonies returns the strongest cor-
relation. Yet, this result does not come as a surprise, since two-­thirds of all colo-
nial cases are either British or French colonies.
Of the 107 countries in the previous sample, only 34 were democratic when
they adopted their first electoral systems as independent nations. Of these, 24
inherited their electoral systems from the colonial powers, that is, approximately
the same share of inherited systems as in the sample of both democracies and
non-­democracies. The number of colonies with regard to each colonial power is
given in Table 5.3.3 Roughly two-­thirds of all democratic cases are former
British colonies. All four American colonies were democratic when independ-
ence was declared. The former French colonies show the opposite pattern: only
two (Gabon and Vanuatu) of the 26 countries were democratic when the first
electoral system was chosen. Neither were the former colonies of Portugal,
Belgium, Yugoslavia, Indonesia and New Zealand democratic when electoral
provisions were adopted. All but two (Gabon and Nauru) of the inherited elect-
oral systems among democracies originate from the United Kingdom or the
United States.
Bivariate analysis of British colonial legacy and electoral system choice
returns the correlation coefficient 0.535, significant at the 0.001 level. It may be
repeated that correlation analysis is a rather unwieldy technique for studying
relationships between categorical variables. Notwithstanding, on the basis of
 History and culture   69
these initial analyses, British colonial legacy appears to be a central determinant
of electoral system choice. This conclusion applies to the total population as well
as to democracies. French colonial legacy seems to have some impact on the
dependent variable when all cases of electoral system choice are studied. In the
democratic sample, however, a French legacy does not exist.

5.2 Regional diffusion

In social science, the idea of diffusion means that properties exist in a society
because of influence from other societies rather than as a consequence of some
internal circumstances. In other words, institutional choices are made by imitat-
ing institutional choices of other countries. Transfer of institutions naturally
takes place from other countries than colonial powers, too. Whether a certain
country may serve as a model to other countries in the process of constitution-­
building depends on several factors. Countries with a similar culture and nearly
located countries are often imitated. Samuel Huntington writes about demonstra-
tion effects, also called ‘snowballing’: countries that are geographically proxi-
mate and culturally similar are most likely to imitate each other (1991: 102).
Powerful countries have also served as models to newly independent countries.
In some countries, the electoral system has been imposed by foreign powers.
Countries may, furthermore, make the same institutional choices as other coun-
tries with similar sociopolitical conditions. In this respect, diffusion is not the
opposite of rationality, even though they constitute two different explanatory
models of institutional choice.
When political institutions are adopted, it is natural to look at institutional
arrangements in neighbouring countries. Most countries cooperate with their
neighbours in several ways, particularly in the political and the economic fields.
Moreover, adjacent countries often share a similar culture. Therefore, in the dif-
ficult process of constitution-­building, countries tend to imitate the institutional
choices of nearly located countries. A further reason for this tendency is that the
knowledge of electoral rules is often limited. Some options may not be seriously
considered because of unfamiliarity and uncertainty regarding their con-
sequences. The institutional arrangements in adjacent countries are, con-
sequently, likely to be imitated. In addition to geographic and cultural factors, a
psychological effect may be present. Not every neighbour is expected to serve as
a model in the constitution-­building process. Some countries are more prestig-
ious than others are; those countries that have political, economic and military
power are those that will most likely be imitated.
The political systems in Latin American countries are usually regarded as a
result of diffusion. When the South and Central American countries became
independent in the early nineteenth century, the presidential model of the United
States was practically the only model available (Lijphart 1992b: 25). The adop-
tion of PR during the first half of the twentieth century was carried through in
collaboration with European electoral engineers. In both Europe and South
America, proportional representation was considered the most democratic
70   History and culture
e­ lectoral system; it represented the wave of the democratic future (Lijphart 1991:
75). Similarly, the evolution of electoral systems in Western Europe may be seen
as a product of diffusion. Before the electoral reform in Belgium 1899, all
Western European countries applied plurality or majority systems. By the mid-­
1920s, most countries had introduced proportional systems. In addition, the
adoption of PR in those countries that initially applied plurality systems was,
with a few exceptions, preceded by an intermediate phase of majority elections.
There are certainly other explanations for the introduction of PR in Europe but a
tendency of imitation was surely present as well. When some countries switched
to PR, others tended to follow.
External imposition of electoral systems by foreign powers occurred in some
countries after the Second World War. The West German mixed electoral system,
which combines constituency representativeness with proportional representation,
is probably the foremost example of electoral system design resulting from exter-
nal imposition. It was designed by the allied powers to avoid the apparent mis-
takes of the Weimar period. External imposition also occurred in Austria, Japan
and South Korea (Reilly and Reynolds 1999: 9). The introduction of list PR in
Namibia in the late 1980s is likewise regarded as a case of external imposition.
The rationale for a nationwide list system, as a means of trying to calm warring
factions, came initially from the United Nations, and South Africa played later on
an important role in designing Namibia’s electoral system (1999: 26).
However, this kind of electoral system choice is quite rare. It occurs only in
extraordinary circumstances when the previous system has caused serious
damages inside as well as outside the country borders. On the one hand, external
imposition of electoral systems is a process of diffusion. A model of electoral
arrangements is transferred from one part (or several parts) to another. On the
other hand, the transferred system need not necessarily exist in another country,
which is one of three criteria of regional diffusion stipulated below. The German
system, for instance, did not exist elsewhere when it was externally imposed in
1949. It was deliberately designed by foreign powers with particular goals in
mind. In this sense, external imposition is closer to a rational explanation than a
process of diffusion. Accordingly, I shall not regard external imposition of elect-
oral systems by foreign powers as a process of regional diffusion.
Regional diffusion of electoral systems implies that the choice of an electoral
system is a consequence of influence of other countries within the same region.
Some further criteria are also needed. First of all, a sufficient amount of similar-
ity is required between the influencing and the imitating country with regard to
the phenomenon that is assumed to be subject to diffusion. In practice, the imi-
tating country must adopt the same electoral system as the imitated country
applies. Similarities with regard to the electoral formula are also required. There
are, however, a few exceptions to this rule. No distinction between SMP, BV
and SMP–BV is made; that is, the adoption of BV may be considered a transfer
from a country that uses SMP and vice versa. The same applies to TBM and
OBM. The various mixed electoral systems combine the majoritarian and the
proportional rule in different ways, each one of them thereby constituting a
 History and culture   71
‘system’ of its own. In order to establish a process of diffusion between two
countries with mixed electoral systems, the combination of plurality/majority
and PR must not vary much between the two. Each case needs to be pondered
over carefully.
Furthermore, we need an apparatus for proving that a process of regional dif-
fusion has taken place. Similarity between two objects may come into existence
in four different ways. First, it may be a consequence of diffusion from A to B.
Second, it may be a process of diffusion from a third part to both A and B. Third,
the similarity may have been developed independently in both contexts. Fourth,
it may have come into existence randomly (Karvonen 1981: 81–2). Only the
former two are cases of diffusion.
In the following, three criteria for determining whether similarities are a con-
sequence of regional diffusion are laid down. There are basically two dimen-
sions of diffusion: spatial and hierarchical diffusion. First, a ‘transfer’ of an
electoral system from one country to another must occur within the same sub-­
region in order to qualify as a process of regional diffusion. This is the spatial
dimension. Second, a ‘transfer’ of an electoral system from country A to country
B does not qualify as a process of diffusion if country B is politically, economi-
cally and militarily more powerful than country A. This criterion is concerned
with the hierarchical dimension of diffusion. Countries do not tend to imitate
countries lower down the hierarchy of power and prestige. Country size also
matters in this regard. For instance, the introduction of mixed-­superposition in
Russia does not qualify as a process of diffusion from Georgia, since Russia is
superior to Georgia in every respect mentioned above. However, it is often diffi-
cult to assess whether country A is more powerful than country B. Country A
might be small but economically well off, while country B might be the largest
one in the sub-­region but relatively poor. As a rule of thumb, I shall consider
some kind of equilibrium between the countries in question as a satisfying
criterion.
Third, an electoral system must exist in country A before country B adopts
the same electoral system. If country B chooses its electoral system before
country A, the electoral system choice in country B is not a consequence of dif-
fusion. The adoption of mixed-­superposition in Georgia in 1992 was not a result
of diffusion from Russia, because Russia did not adopt the same electoral system
until 1993. To introduce cultural similarities as a necessary criterion does not
seem appropriate. Diffusion and imitation are certainly concerned with culture
but there are examples of diffusion between culturally different societies, too –
the transfer of the presidential system from the United States to South America
is a case in point. Yet, the three criteria presented above are regarded as suffi-
cient in order to establish whether a process of diffusion has taken place or not.
The regional division below comprises the cultural dimension. Countries within
a given sub-­region often share similar cultural features. Since imitation of former
colonial powers is separated from imitating other influential or culturally similar
countries, an electoral system choice that is regarded as a colonial legacy does
not qualify as a case of regional diffusion.
72   History and culture
By way of introduction, countries are classified according to the following
continents: America, Europe, Africa, Asia and the South Pacific. However,
regional influence may also occur in certain parts of a continent. Accordingly,
the division is too broad, and might overlook some tendencies of imitation in
certain smaller regions. A common categorization of sub-­regions is applied.
America is divided into North America, Central America (including Mexico),
the Caribbean and South America. Europe is split up in two: Western and
Eastern Europe. There are five distinct African regions identified: North, West,
Central, East and Southern Africa. Concerning Asia, the categorization in the
double volume Elections in Asia and the Pacific: A Data Handbook (Nohlen et
al. 2001a; 2001b) is applied: Middle East, Caucasus and Central Asia, South
Asia, Southern East Asia and East Asia. Russia (and the Soviet Union) is
regarded as part of Eastern Europe as well as Caucasus and Central Asia. A
further specification is needed. If an influential country is situated on the fringe
of a sub-­region, its electoral system may naturally be imitated by neighbouring
countries in the adjacent sub-­region. These cases may also be regarded as cases
of regional diffusion if the established criteria are fulfilled.
To sum up, the criteria of regional diffusion are: (a) transferring of an elect-
oral system within the same sub-­region; (b) superiority of the imitated country,
or at least a state of equilibrium in relation to the imitating country in political,
economic or military respects; and (c) occurrence of the electoral system in the
imitated country before occurrence of the same electoral system in the imitating
country. Furthermore, similarities in electoral rules must prevail. Cases that are
coded as colonial diffusion are not coded as cases of regional diffusion. This
decision is taken in order to avoid over-­determination of the cultural and histor-
ical perspective.

Empirical evidence
The frequency of electoral systems as a consequence of diffusion is given in
Table 5.4. Some 36 per cent of all cases meet the criteria of regional diffusion.
In order to provide some further information, the frequency of electoral system
choice in each continent is also presented. Only in Europe is a majority of all
choices associated with regional diffusion. Approximately one-­third in America
and Asia, one-­quarter in Africa, and merely one of seven electoral system
choices in the South Pacific are identified as processes of diffusion. However, it
should be observed that most countries in Africa and the South Pacific have been
colonies in the post-­war era, and many of them have inherited their electoral
systems from the former colonial powers.
There is a striking pattern with regard to the adoption of mixed electoral
systems in former Communist countries. Hungary was the first country in
Eastern Europe that introduced a mixed system, namely supermixed in 1989.
Lithuania, Croatia and Georgia adopted mixed-­superposition in 1992, followed
by Russia in 1993. Armenia and Azerbaijan joined the company of former Soviet
states applying mixed systems two years later. Kazakhstan, Kyrgyzstan and
 History and culture   73
Table 5.4 Regional diffusion and electoral system choice

Electoral system N Regional % No regional %

choice diffusion (N) diffusion (N)

Europe 57 33 57.9 24 42.1

America 40 13 32.5 27 67.5
Africa 94 25 26.6 69 73.4
Asia 73 27 37.0 46 63.0
South Pacific 14 2 14.3 12 85.7
N 278 100 36.0 178 64.0

Table 5.5 Regional diffusion and electoral system choice in democracies

Electoral system N Regional % No regional %

choice diffusion (N) diffusion (N)

Europe 25 11 44.0 14 56.0

America 15 2 13.3 13 86.7
Africa 3 0 0.0 3 100.0
Asia 11 2 18.2 9 81.8
South Pacific 12 1 8.3 11 91.7
N 66 16 24.2 50 75.8

Tajikistan replaced their TBM formulas with mixed-­superposition in 1999. The

same system has also been applied in Ukraine from 1998 to 2002, in Bulgaria
1990, in Yugoslavia 1992 and in Macedonia 1998. As observed in the previous
section, six former Soviet states of 14 adopted the Soviet TBM formula as inde-
pendent states. A decade later, nine of these 14 countries applied mixed electoral
systems. Interestingly, two years after Russia adopted list PR (2005), Kazakhstan
and Kyrgyzstan followed.
In Table 5.5, democratic cases are dealt with. The pattern among democracies
resembles that of the total population: processes of regional diffusion are most
frequent in Europe and least frequent in Africa and the South Pacific. Correlation
analysis of diffusion and electoral system choice provides the following coeffi-
cients: 0.263 in the whole research population and 0.524 in the democratic
sample. Both associations are significant at the 0.01 level, suggesting that pro-
portional systems are related to regional diffusion.

5.3 Temporal diffusion

The third way of dealing with electoral system choice within the cultural and his-
torical context is to regard it as an epoch phenomenon and look for possible trends
concerning the adoption of electoral systems. Basic constitutional patterns are
often products of early statehood (Anckar and Karvonen 2004: 8). Consequently,
the adoption of certain electoral systems may be linked to certain periods and eras.
74   History and culture
Reilly and Reynolds (1999: 8–9) have observed that the pattern of electoral
system choice to a great extent corresponds to the three waves of democrat-
ization, as described by Huntington (1991). During the first wave from 1828 to
1926, multi-­party competition and democratic institutional structures began to
evolve in the United States, Canada, the United Kingdom, Australia, New
Zealand and across much of continental Europe. Electoral systems emerged
gradually, very much in line with the evolution of the democracies themselves.
The Anglo-­American societies were relatively homogeneous, based around a
single partisan cleavage and a two-­party system. Hence, the SMP formula was a
natural choice for these countries. A different pattern occurred in Western
Europe, which was characterized by plural societies and the lack of a single
dominant group. The debate over electoral systems in Scandinavia and contin-
ental Europe was concerned with the perceived trade-­offs between accountabil-
ity and representativeness, on the one hand, and geographic representativeness
and proportional representation of parties in the legislature, on the other. Accord-
ingly, a change from SMP to list PR – often via the TBM formula – took place
(Carstairs 1980). Referring to Reilly and Reynolds (1999), Farrell (2001: 177)
maintains that electoral system choice in the first wave was partly a result of
conscious design, partly an accidental evolution.
The second wave of democratization which began in 1943 and ended in 1962
was characterized by the re-­establishment of some democracies and the process
of decolonization. Electoral systems were either inherited from former colonial
powers or imposed by outside powers in order to meet a specific purpose. The
post-­war decolonization resulted in many new independent countries in Africa
and Asia, many of which introduced, or continued using, the electoral system of
their colonial masters. As mentioned earlier, external imposition of electoral
systems occurred in West Germany, Austria, Japan and South Korea (Reilly and
Reynolds 1999: 9).
Reilly and Reynolds assert that the central feature of the third wave of demo-
cratization, which began with the overthrow of the Salazar dictatorship in Portu-
gal in 1974, has been deliberate electoral system design (1999: 9). An extensive
debate on the merits of different electoral systems has been witnessed in several
cases of democratic transition as well as in established democracies. Due to the
rareness of electoral reform in established democracies, Farrell (2001: 179)
regards the growing number of electoral system reforms as a new fourth wave of
electoral system choice.
Electoral system choice as an epoch phenomenon is analysed by means of a
variable called temporal diffusion. A particular electoral system may be fre-
quently adopted at a specific point in time and rarely chosen at other times.
Accordingly, countries are to a great extent assumed to adopt electoral systems
that many other countries have recently chosen. The cognitive heuristics model
underlies the wavelike pattern of diffusion: a modest start, sudden rise, and even-
tual decline. At certain points in time, there is a need for developing new institu-
tions. Usually a particular system is more popular than other systems, and serves
as a model for countries in the midst of a constitution-­building process. The con-
 History and culture   75
stitution is often a child of its time; the single parts of the constitutional frame-
work can be perceived in the same way. However, it is difficult to predict which
institutions will be in fashion at a particular point in time.
Three matters are of interest in connection to the temporal diffusion variable.
First, the timing of each electoral system choice is observed. Are there periods
during which few countries have adopted electoral systems, and other periods
during which electoral system choice has been frequent? Of course, this matter
is largely related to timing of independence – countries usually adopt electoral as
well as other institutions when they obtain the power of self-­determination.
However, by separating all cases of electoral system change from first choices of
electoral systems, some interesting patterns may emerge. On the whole, a pres-
entation of the timing of all choices of electoral systems during the relevant time
period is of basic interest concerning the topic of the book.
Second, I shall examine whether particular periods are associated with the
adoption of particular electoral systems. Which electoral systems have been the
most popular at which point in time? Are some electoral systems attached to
some given periods? The third matter to be dealt with is to examine whether
some of the other explanatory factors are valid during certain eras of electoral
system choice and less important during other periods. Which are the main deter-
minants in various epochs? Certainly, this is an excursion from the general logic
of the study – yet, I find the exercise appropriate given the low level of dignity
of temporal diffusion as a general determinant of choosing political institutions.
This variable is rather similar to the concept of colonial and regional diffu-
sion; countries are expected to imitate recent institutional choices of other coun-
tries. There are, however, three differences between temporal diffusion and other
types of diffusion. First, according to the definition of regional diffusion in the
last section, diffusion may occur within the same sub-­region only. Timing of
electoral system choice is not concerned with the spatial dimension of politics.
Second, in contrast to colonial and regional diffusion, temporal diffusion deals
exclusively with timing. Third, temporal diffusion is partly analysed in relation
to other explanatory variables.

Empirical evidence
Figure 5.1 shows frequencies by year of all 140 first electoral system choices. A
large share of electoral systems was adopted in the 1960s; a total of 15 new cases
emerged already during the first year of the decade. During the 1980s, only six
initial choices were made, since the number of independent nations increased
only modestly that decade. The breakdown of the Soviet Union in the early
1990s resulted in several new independent countries and, as a consequence, lots
of electoral system choices.
In Figure 5.2, cases of electoral system reform are illustrated. Electoral
reform did not become a frequent phenomenon until 1990 when eight changes
took place. More than one-­half of all 138 cases of electoral system change have
occurred between 1990 and 2007. The number of electoral reforms peaked in
76   History and culture

16
14
12
10
8
6
4
2
0
Figure 5.1 FrequenciesFirst
of first electoral
electoral system
system choices
choice 1945–2007.
1945–2007

12

10

0
Figure 5.2 FrequenciesElectoral
of electoral system
system change
change 1945–2007.
1945–2007

1992 when 11 changes were made. Most often, however, the number of changes
has been one or two per year.
On the basis of Figure 3.1 in the Chapter 3, which shows the frequencies by
year of all cases of electoral system choice, four specific periods are distin-
guished: 1945–1955, 1956–1969, 1970–1989 and 1990–2007. At the end of each
period, the frequency of electoral system choice has been rather low. In the fol-
lowing, the frequencies of electoral systems in each of these periods are pre-
sented. The total number of electoral system choice in each category is 36, 73,
61 and 108 respectively.
Some distinct tendencies are discernible in Table 5.6. Almost one-­half of all
electoral system choices during the first period were plurality systems. During
the second period, when a large part of the African countries became independ-
ent, plurality constituted more than 60 per cent of all adopted electoral systems.
The frequency of plurality system choices remained high in the 1970s and the
 History and culture   77
Table 5.6 Temporal diffusion and electoral system choice

Time period Plurality % Majority % PR % Mixed %

(N) (N) (N) (N)

1945–1955 17 47.2 5 13.9 10 27.8 4 11.1

1956–1969 45 61.6 6 8.2 16 21.9 6 8.2
1970–1989 33 54.1 8 13.1 10 16.4 10 16.4
1990–2007 10 9.3 21 19.4 43 39.8 34 31.5
N 105 37.9 40 14.4 79 28.4 54 19.4

Note
Chi-square value 66.843, sig. 0.000.

0
First electoral system choice in democracies 1945–2007
Figure 5.3 Frequencies of first electoral system choices in democracies 1945–2007.

1980s, but decreased to merely one-­tenth of all choices during the fourth period.
In contrast, the number of majority system choices since 1990 is twice as high as
plurality ones. Proportional systems have gained renewed popularity since 1990,
representing 40 per cent of all adopted electoral systems. After having been
rather unattractive before 1990, mixed electoral systems represent nearly one-­
third of all choices during the fourth period. Accordingly, there have been some
drastic changes regarding the popularity of different electoral systems during the
post-­war era. The chi-­square value is significant at the 0.001 level.
There are 38 first choices and 28 cases of electoral system change in the
democratic sample. First choices of electoral systems among democracies are
illustrated in Figure 5.3: four distinct epochs appear. The first period encom-
passes the years between 1945 and 1957, the distinct feature being almost com-
plete absence of electoral system choice. Only the Philippines, Israel and Sri
Lanka chose electoral systems for the first time before 1958. The second epoch
ranges from 1958 to 1970, when ten democracies adopted electoral systems for
the first time as independent countries. During the third period, from 1974 to
1983, 15 new cases are observed. Between 1984 and 1990, no first choices of
78   History and culture
electoral systems in democracies took place. A fourth, brief period occurred
between 1991 and 1994, eight cases being observed.
Electoral system change among democracies is presented in Figure 5.4. Only
France implemented a major electoral reform before 1951. A period of electoral
system change took place between 1951 and 1961, ten such cases being
observed: seven of these electoral reforms occurred in France, Italy and Greece.
Iceland and Turkey introduced PR reforms, whereas Lebanon replaced TBM
with BV. Thereafter, a 17-year period without electoral reforms followed. Sri
Lanka and Cyprus replaced plurality systems with list PR in 1978 and 1981
respectively. Over the second half of the 1980s, France started a period of elect-
oral system change. Between 1991 and 2002, 13 democratic countries changed
their electoral systems. Bulgaria adopted list PR in 1991. Italy and Venezuela
introduced mixed-­corrective systems in 1993 after having used list PR. Italy
reintroduced list PR in 2005. Mixed electoral systems were also chosen in Japan,
New Zealand, Bolivia, Thailand, Madagascar, the Philippines and Monaco. The
Philippines resumed SMP in 2001 after merely one election under mixed-­
superposition. Papua New Guinea reintroduced AV in 2002 after having used
SMP since 1975.
On the basis of Figure 3.2 in Chapter 3, which illustrates all cases of electoral
system choice in democracies, four specific periods are distinguished: 1945–
1956, 1958–1970, 1974–1988 and 1991–2007. Characteristic of all four periods
is that electoral systems have been scarcely adopted at the end of the period. The
frequency of electoral systems in each of these periods is presented and analysed
by means of a chi-­square test in Table 5.7. The total number of electoral system
choice under democratic circumstances in each category is ten 14, 19 and 23
respectively.
As in the previous table, there is a considerable variation with regard to the
popularity of different electoral systems in different times. A majority of adopted
systems during the second and the third period are plurality systems. Since 1991,
however, plurality systems constitute less than one-­fifth of all choices. Majority

0
Electoral system
Figure 5.4 Frequencies change
of electoral in democracies
system change in 1945–2007
democracies 1945–2007.
 History and culture   79
Table 5.7 Temporal diffusion and electoral system choice in democracies

Time period Plurality % Majority % PR % Mixed %

(N) (N) (N) (N)

1945–1956 4 40.0 0 0.0 3 30.0 3 30.0

1958–1970 7 50.0 2 14.3 5 35.7 0 0.0
1974–1988 12 63.2 2 10.5 3 15.8 2 10.5
1991–2007 4 17.4 1 4.3 8 34.8 10 43.5
N 27 40.9 5 7.6 19 28.8 15 22.7

Note
Chi-square value 18.475, sig. 0.030.

systems have been scarcely adopted during the entire post-­war era. Proportional
systems were attractive during the first and the second period, and enjoyed some
renewed popularity in the 1990s after having been scarcely adopted in the 1970s
and 1980s. Although mixed systems enjoyed some popularity during the first
era, no democratic country adopted such a system in the second period. After
Suriname and Seychelles chose mixed systems in the 1970s, another ten demo-
cratic countries adopted mixed systems during the fourth epoch. On the whole,
plurality systems have lost popularity during the fourth period at the expense of
proportional and, in particular, mixed electoral systems. The chi-­square value is
considerably smaller among democracies than in the total data sample, yet signi-
ficant at the 0.05 level.

On the correlates
In the following, the relevance of the other explanatory variables during each of
the four periods is explored. Due to rather small samples, the analyses are
descriptive and not presented in table format. Also, with a few exceptions, no
controlling for competing factors is undertaken. It should be observed that the
analysis in this section represents an excursion from the general logic of the
study. In addition, the discussion to some extent anticipates the multiple analy-
ses. However, because of the modest role of temporal diffusion as a general
determinant of institutional choices, it seems appropriate to probe into some of
its correlates more closely. It is reasonable to assume that the importance of dif-
ferent explanatory factors varies over time. Consequently, the analysis may
provide some important, additional information on how electoral systems have
been chosen in the post-­war era. Furthermore, it might facilitate, or contribute to
the task of predicting future electoral system choices.
To begin with, patterns during the first period (36 cases) are analysed. Plural-
ity systems were most frequent between 1945 and 1955, 17 cases being
observed, followed by ten proportional systems. The number of adopted major-
ity and mixed systems were five and four respectively. In comparison with the
whole time period 1945–2007, plurality systems are over-­represented and mixed
80   History and culture
systems are under-­represented. No pattern emerges concerning ethnic and cul-
tural diversity on the one hand and electoral system choice on the other. By and
large, the same conclusion applies to population size: the share of plurality and
proportional systems is rather similar among small as well as large countries. It
should be observed that three of a total of four mixed systems were adopted in
countries with more than 40 million people (France, Italy and West Germany).
However, the mixed systems (conditional) in France and Italy did not involve
single-­member districts, which constitute a rational solution for large countries.
Party system structure is not relevant to electoral system choice in the total
research population, and few party system transformations took place during the
first two periods.
Several British and French colonies in Asia were granted independence
between 1945 and 1955. India, Sri Lanka, Burma, Israel and Jordan became
independent from the United Kingdom, whereas Vietnam, Cambodia, Laos and
Syria became independent from France. In addition, the Philippines, Indonesia
and Libya were granted independence from their respective colonial powers. Of
these 12 countries, seven inherited their electoral systems from the former colo-
nial power; Cambodia, Laos, Syria, Israel and Libya preferred other systems.
Accordingly, colonial legacy is not a very relevant factor in explaining electoral
system choice during the first period. Of 36 cases, 12 qualify as processes of
regional diffusion, that is, close to the share (36 per cent) of all cases. Only four
of the 17 plurality systems are connected to regional diffusion, whereas one-­half
of the proportional systems have been imitated. Considering that seven electoral
systems were inherited from former colonial powers, approximately one-­half of
all cases during the first period are explained by colonial and regional diffusion.
As for the institutional perspective, form of government is not analysed in
this context, because it concerns democracies only. No pattern is discernible
with regard to territorial organization and electoral system choice. There are five
cases of federalism observed: Brazil and Venezuela adopted proportional
systems, West Germany introduced a mixed system, and India as well as Libya
preferred plurality systems. Concerning chamber structure, however, some dif-
ferences appear. The combination of two-­chambers and plurality was the most
common during the first period, ten cases belonging to this category. It should be
observed that the British parliamentary system does not explain this pattern,
because the decolonization process had only started. Of these ten bicameral
countries that adopted plurality systems seven had not been British colonies. In
contrast, all majority systems were adopted in unicameral countries (Syria, North
Korea, Mongolia, Egypt and Iraq). Mixed systems were adopted in foremost
bicameral countries, whereas proportional systems were rather evenly divided
between one- and two-­chamber entities.
The second period (73 cases) began in 1956 and ended in 1969. A strong
dominance of plurality systems characterizes this epoch, the number being 45,
compared to six majority systems, 16 proportional systems, and six mixed
systems. The category of mixed systems is the most under-­represented in com-
parison with the frequencies of other systems between 1945 and 2007. There is a
 History and culture   81
strong relationship between a high level of cultural fragmentation and electoral
systems inappropriate for plural societies. During the second period one-­half of
all cases belong to the category with a high level of fragmentation and other
electoral systems than list PR, STV and AV. A great majority of those countries
that are characterized by medium-­sized fragmentation also adopted inappropriate
systems. However, among countries with low or bipolar levels of fragmentation,
proportional systems (and AV) are more frequent than other systems. The chi-­
square value is significant at the 0.001 level.
The average population size among countries that chose electoral systems
between 1956 and 1969 is rather small: 19 countries, of which all but Iceland
adopted electoral systems for the first time, had less than one million people.
Another 40 cases are characterized by a population size between one million and
ten million. The share of proportional systems is appreciably larger among coun-
tries with more than ten million citizens than among smaller countries, but plu-
rality systems are, notwithstanding, most frequent in all categories. Party system
transformation is not relevant to electoral system choice during the second
period. The most striking feature concerning the rational perspective during the
analysed period is that almost every country with a high level of cultural frag-
mentation adopted an inappropriate electoral system. However, explanations of
this pattern are to be found elsewhere.
Between 1956 and 1969, two-­thirds of all electoral system choices (49 of 73)
were introduced in former colonies. Some 24 countries became independent
from the United Kingdom and 19 declared independence from France. A total of
39 of the former colonies inherited their electoral system from the colonial
powers. In addition to 20 cases of British legacy and 13 cases of French legacy,
three systems were inherited from Belgium, and one each from Italy, New
Zealand and Australia. The association between colonial legacy and electoral
system choice also explains the correlation between cultural diversity and elect-
oral system choice: former colonies are considerably more fragmented than other
countries during the analysed period. The correlation coefficient between former
colonies and cultural diversity is 0.485, with 19 cases of electoral system choice
related to regional diffusion, which leaves only 15 cases not comprised by colo-
nial or regional diffusion during the second period. The category of regional dif-
fusion consists of eight plurality systems and eight proportional systems each.
No mixed system was introduced as a consequence of regional diffusion. Of
those cases that are not affected by colonial legacy or regional influence two-­
thirds are either mixed or proportional systems, which is remarkable during an
epoch dominated by the adoption of plurality systems. No pattern emerges when
electoral systems are analysed with regard to territorial organization and chamber
structure.
The third period (from 1970 to 1989) consists of 61 cases. Plurality systems
were still prevailing, whereas the other systems were under-­represented, PR in
particular, in comparison with the dispersion of electoral systems since 1945.
During the third period, 33 plurality, eight majority, ten proportional and ten
mixed systems were adopted. There is no pattern whatsoever concerning ethnic/
82   History and culture
cultural diversity and electoral system choice. In contrast, population size corre-
lates with the dependent variable: plurality is associated with smallness. Popula-
tion size is below one million in 23 cases, between one and ten million in 19
cases, and larger than ten million in likewise 19 cases. Plurality systems were
adopted by 17 of the microstates, whereas the number of plurality systems in the
other two categories is eight each. The frequency of other systems only margin-
ally increases when population size increases. Plurality systems are, in other
words, frequent in all categories of population size, but exceedingly popular
among small countries. Again, diffusion may explain the observed pattern.
Between 1970 and 1989, nine party system transformations took place. Of these,
seven resulted in the adoption of mixed or proportional systems.
During the third period 22 colonies declared independence– 15 of these coun-
tries had been under British rule. Electoral systems were inherited from the colo-
nial powers by 13 former colonies. The British exported their electoral system to
11 newly independent countries. All countries that inherited their electoral
systems between 1970 and 1989 are microstates; thus, explaining the relation-
ship between plurality and smallness. A total of 19 electoral system choices are
processes of regional diffusion, which constitutes a somewhat smaller share
compared to the total data sample. Only five cases are PR, whereas all the others
are plurality systems. The chi-­square test returns a significant relationship at the
0.01 level between regional diffusion and electoral system choice. All in all,
more than one-­half of all cases during the third epoch of electoral system choice
are related to colonial or regional diffusion. Again, no pattern is discernible
regarding territorial organization and chamber structure on the one hand and
electoral system choice on the other.
The fourth period (1990–2007) consists of 108 cases, of which ten are plural-
ity, 21 majority, 43 proportional, and 34 mixed systems. Thus, plurality systems
are under-­represented, whereas proportional systems and mixed systems are
over-­represented in comparison with the total research population. No associ-
ation exists between ethnic/cultural diversity and electoral system choice. The
analysis of population size and electoral system choice during the fourth period
returns some interesting findings. Countries that have chosen proportional
systems are, in accordance with the theoretical arguments by Rokkan (1970),
Katzenstein (1985) and Rogowski (1987), smaller than countries with other
systems on average. Moreover, seven countries with more than 50 million people
adopted mixed electoral systems, thereby lending support to the thesis that large
countries need small (that is, single-­member) districts in order to maintain geo-
graphical accountability. A large number of party system transformations took
place during the fourth period. Of a total of 54 transformations, 24 resulted in
the adoption of mixed or proportional systems. Consequently, party system
transformation is an important factor in explaining electoral system choice
during the fourth epoch.
By contrast with earlier periods, colonial legacy has not been a prominent
factor since 1990. During the analysed period, 24 countries achieved independ-
ence from foreign rule, 11 of which inherited their electoral systems. A total of
 History and culture   83
six electoral systems were inherited from the Soviet Union, three from the
United States, and one each from Yugoslavia and Indonesia. A total of 51 of 108
cases meet the criteria of regional diffusion. An overwhelming majority of the
regionally influenced cases are proportional (28) and mixed (15) systems. The
association between regional diffusion and electoral system choice during the
fourth period is significant at the 0.01 level. During the fourth period, nine
federal countries adopted electoral systems– no particular pattern is observed.
The same conclusion applies to electoral system choice with regard to chamber
structure.
To sum up, colonial legacy and regional influence explain approximately one-
­half of all electoral system choices during the first period. A total of seven elect-
oral systems were inherited, whereas 12 electoral systems were the result of
imitating neighbouring countries. Proportional systems in particular were sub-
jected to regional diffusion. The rational perspective does not explain electoral
system choice during the first two epochs. Concerning cultural diversity,
however, there is a clear pattern during the second period: almost all countries
characterized by medium-­sized or high level of fragmentation adopted inappro-
priate electoral systems. Yet, colonial legacy explains this relationship. More
than one-­half of all systems were inherited from the colonial powers, mainly
Britain and France. Another 19 choices were associated with regional diffusion.
Almost every adopted plurality system was a consequence of either colonial or
regional diffusion.
Party system transformation preceded the adoption of mixed or proportional
systems in seven cases during the third period. In other respects, diffusion was
the prominent factor. A total of 13 electoral systems were inherited from the
colonial powers, mostly the United Kingdom. Some 19 choices, foremost plural-
ity systems, were influenced by neighbouring countries. The rational perspective
became a relevant explanatory model during the fourth period. Mixed systems
were adopted in several large countries, the use of single-­member districts being
regarded as the crucial factor. A total of 24 party system transformations were
succeeded by changes from majoritarian to mixed or proportional systems. Colo-
nial legacy no longer played a prominent role. Regional diffusion, by contrast,
was a major determinant, characterizing almost one-­half of all choices. Most of
the imitated systems were non-­majoritarian. The institutional perspective does
not explain patterns of electoral system choice on the basis of these analyses.

Next, electoral system choice in democracies is dealt with. The first period
(1945–1956) consists of ten cases: four plurality systems, three proportional
systems, and likewise three mixed systems. All countries that adopted propor-
tional or mixed systems, except for Israel, were ethnically, linguistically or reli-
giously homogeneous, whereas all countries that introduced plurality systems,
except for Greece, were culturally heterogeneous. Thus, cultural features were
not taken into consideration when electoral systems were chosen. Those coun-
tries that adopted plurality systems were much smaller than the others, the
average population size being 8.7 million people in comparison with 30.5 million
84   History and culture
among countries that chose PR and 32.3 million among countries that introduced
mixed systems. These results contradict the theoretical arguments put forward by
Rokkan (1970) and others. In the previous section, I concluded that the relation-
ship between smallness and plurality is a consequence of colonial legacy. In this
sample, however, the connection between population size and electoral system
choice is only partly affected by colonial heritage. The Philippines and Sri Lanka
inherited SMP from the United States and the United Kingdom respectively, but
Israel, formerly under British administration, preferred list PR. Deviating level of
party system fragmentation did not exist before the electoral system changes in
Greece and Italy. Party system transformation is not relevant to explaining elect-
oral system choice in democracies, because cases of transformation are per defini-
tion non-­democratic. A total of two cases out of ten meet the criteria of regional
diffusion. Accordingly, patterns of diffusion apply to four cases of electoral
system choice. None of the variables of the institutional perspective produces any
pattern in connection to the dependent variable.
During the second period (1958–1970), one-­half of the 14 cases are plurality
systems. Nauru and France adopted majority systems, whereas the remaining
five cases are proportional systems. Regarding cultural diversity, five cases
belong to the category of high level of fragmentation combined with plurality
systems, whereas four cases combine a low level of fragmentation with PR.
Among the culturally diverse countries, only Nauru and Guyana adopted appro-
priate electoral systems. Those countries that adopted plurality systems were
considerably smaller than other countries. A majority of the democratic coun-
tries that chose electoral systems during the second epoch were former colonies,
which might explain the pattern. Party system structure did not influence any
electoral system change between 1958 and 1970. Of those ten countries that
recently had been under foreign rule, only Malta and Guyana preferred other
systems than that of the colonial power. A total of six countries inherited the
British electoral system. A total of three cases qualify as processes of regional
diffusion: Greece, Iceland and Guyana. Consequently, diffusion is related to 11
of a total of 14 cases.
Despite some discernible tendencies, no definitive pattern emerges concern-
ing form of government and electoral system choice. The sample consists of two
presidential regimes – both Cyprus and Gabon adopted plurality systems. A total
of two out of four semi-­presidential regimes, France and Nauru, introduced
majority systems, whereas the other two, Iceland and Guyana, chose propor-
tional systems. The parliamentary regimes in the sample are divided between
five plurality systems and three proportional systems. All countries that com-
bined plurality with parliamentarism were former British colonies. Malaysia is
the only federal country in this sample. Regarding chamber structure, the follow-
ing pattern is observed: a majority of the unicameral countries adopted propor-
tional systems, whereas most bicameral polities introduced plurality systems.
However, all countries that combined plurality systems with two-­chamber legis-
latures (Malaysia, Jamaica, Fiji, Barbados and Trinidad and Tobago) were
former British colonies.
 History and culture   85
Plurality systems dominate the third period (1974–1988) at the expense of
proportional and mixed systems. Of the 19 cases, 12 are plurality, two are major-
ity, three are proportional, and two are mixed systems. The frequency of cultur-
ally diverse countries with plurality systems is high, whereas only two countries
(Sri Lanka and France) with medium-­sized or high level of fragmentation
adopted PR. Almost every country in the sample had less than one million cit-
izens when electoral systems were chosen. Party system structure is not relevant
to electoral system choice in this sample. Once again, colonial legacy explains
the frequent adoption of plurality systems in culturally diverse societies. As
many as 13 of these 18 countries (France changed their electoral system twice
during the third period) are former British colonies – all but three adopted SMP.
Regional diffusion was not relevant to electoral system choice between 1974 and
1988. Merely one case (Cyprus 1981) fulfils the criteria of regional diffusion.
There is a clear pattern with regard to form of government: all parliamentary
countries introduced plurality systems, whereas all other forms of government
are combined with other systems. However, all parliamentary regimes with plu-
rality systems, except for Papua New Guinea, are former British colonies. No
federal system is represented in this sample of electoral system choice. A total of
seven cases are bicameral; two of which are related to the French electoral
reforms in 1986 and 1988. The remaining five belong to the British Common-
wealth. Surprisingly enough, eight of 13 former British colonies preferred uni-
cameral assemblies.
The fourth period (1991–2007) consists of 23 cases: four plurality systems,
one majority system, eight proportional, and ten mixed systems. No pattern
appears with regard to ethnic and cultural diversity. Inappropriate choices in
heterogeneous countries were still common. However, examples of the opposite
were also present: the highly fragmented Papua New Guinea, for instance,
adopted AV, which is regarded as a conflict-­mitigating system. Population size
varies considerably in the four categories of electoral systems. The six largest
countries in the sample adopted mixed systems. However, Italy reintroduced list
PR in 2005 and the Philippines reintroduced SMP three years after having
chosen mixed-­corrective.4 The remaining three plurality systems were chosen in
microstates, all of which are former American colonies. The average population
size among countries that adopted proportional systems is way below the average
of the remaining cases, the mean values being 12 million and 30 million people
respectively. Concerning the level of party system fragmentation, a significant
difference was observed between democracies that have changed from plurality
to mixed or PR systems and democracies with plurality systems in general. The
fragmented party systems in Thailand, the Philippines and Japan prior to elect-
oral system change contribute largely to this pattern. However, these countries
are also very large in terms of population. This leads to the question whether size
or party system structure is the decisive factor. Both factors may, of course, have
been relevant to the choice of electoral systems.
In addition to the Marshall Islands, the Federated States of Micronesia and
Palau, which all inherited the plurality system from the United States, another
86   History and culture
three countries in this sample chose their first electoral system after having been
under foreign rule. Estonia, Latvia and Lithuania did not, however, adopt the
Soviet electoral system. A strong relationship between regional diffusion and
electoral system choice is observed. All but one (Italy 2005) proportional system
have been adopted as a consequence of imitation, whereas only two mixed, one
majority, and no plurality system qualify as processes of regional diffusion. The
chi-­square value is significant at the 0.01 level. Patterns of diffusion comprise 13
of 23 electoral systems adopted by democracies during the fourth epoch. No
pattern appears concerning form of government and electoral system choice.
The findings among democracies resemble those in the total research popu-
lation. None of the independent variables is of great importance during the first
period. All in all, however, diffusion is related to four of ten cases. Almost
every electoral system choice during the second period is related to some kind
of diffusion. A total of eight cases were inherited, mostly from the United
Kingdom, and three choices, all proportional ones, were inspired by neighbour-
ing countries. The British influence is also visible with regard to other institu-
tions: former colonies did not only inherit the plurality system but also the
parliamentary system with a bicameral legislature. The frequent adoption of
plurality systems in culturally fragmented countries during the third period is,
once again, largely a consequence of British colonial rule. The same explana-
tion applies to the common combination of plurality and parliamentarism. Con-
trary to the findings in the second period, plurality systems are more associated
with unicameralism than bicameralism. In accordance with the theoretical argu-
ments, countries that adopted PR were considerably smaller than countries that
preferred majoritarian systems during the fourth epoch. Mixed systems, in par-
ticular, were popular in countries with a large population. Colonial legacy was
of little importance but regional diffusion was still relevant to the adoption of
proportional systems.

5.4 Patterns of diffusion: summary

Three variables within the cultural and historical perspective have been ana-
lysed: colonial diffusion, regional diffusion and temporal diffusion. On the
whole, this approach has been much more fruitful than the rational approach in
explaining variations of electoral system choice. All three variables seem to be
relevant explanatory factors.
Beginning with colonial diffusion, there is a bivariate association between
British as well as French legacy and the adoption of majoritarian systems, plu-
rality in particular. Among democracies, however, there is no French legacy
since only two of the former French colonies were democratic when electoral
systems were chosen. Of all countries that have been former colonies and have
adopted electoral systems within ten years after the end of the colonial rule, two-­
thirds have inherited their electoral systems from the colonial powers. Indeed,
the strongest correlation is observed between the variable that observes all
former colonies and majoritarian systems.
 History and culture   87
The descriptive analysis suggests that regional diffusion is of appreciable
importance in explaining electoral system choice. More than one-­third of all
choices of electoral systems since 1945 are identified as processes of regional
diffusion. In the democratic sample, however, the share is merely 24 per cent.
Europe is largely characterized by regional diffusion – in other parts of the
world, the frequency is considerably smaller. The pattern is to a great extent a
consequence of the fact that large parts of Africa, Asia and the South Pacific
were under colonial rule at the beginning of the analysed time period. Con-
sequently, many countries in these regions have inherited their electoral systems
from the colonial powers. The analyses of both samples return significant corre-
lations, indicating that there is a link between regional diffusion and mixed/pro-
portional systems.
Concerning temporal diffusion, some specific trends are discernible. Introduc-
tion of electoral provisions were particularly frequent in the 1960s and 1990s,
whereas the number of cases was relatively small in the 1950s and 1980s. To be
sure, this pattern coincides with the frequency of countries having declared inde-
pendence at various points in time, but the illustration of electoral system change
over time suggests that there may also be a trend effect irrelevant of the inde-
pendence factor. Major electoral reforms did not become a typical phenomenon
until the 1990s. There are four epochs of electoral system choice distinguishable
in the total as well as the democratic sample. The cut-­points between periods
slightly differ but both categorizations indicate the same trend: after having been
the most popular electoral method between 1945 and 1990, the plurality system
lost attraction in the 1990s at the expense of mixed and proportional systems. In
comparison with previous decades, the share of adopted mixed systems since
1990 is conspicuous.
The analyses of each period show that the significance of explanatory varia-
bles varies at different points in time. Patterns of diffusion are noticeable but not
salient during the first period. Regional diffusion is somewhat more important
than colonial legacy. The influence of the cultural and historical setting is over-
whelming during the second and the third epoch. British colonial legacy, in par-
ticular, is of great importance. Contrary to the findings among cases in the first
and the second period, regional diffusion is more associated with the adoption of
plurality systems than PR during the third epoch. Party system transformation is
a major determinant during the fourth period. The tendency of adopting mixed
systems in foremost large countries strengthens the importance of rational expla-
nations in recent times. A large part of the mixed and proportional systems is a
consequence of regional diffusion. The findings correspond rather well to the
pattern that Reilly and Reynolds (1999) have observed, particularly with regard
to colonial influence during the first few post-­war decades. The authors maintain
that deliberate electoral system design is the distinguishing feature in recent
times. Electoral system design has definitely been present during the last two
decades, but not as prominently as the authors suggest.
6 The institutional perspective

The third approach maintains that institutional choices are explained by the overall
constitutional design. Pippa Norris, among others, emphasizes the importance of
the constitutional structure in setting the context for many aspects of political
behaviour (2004: 41). Anckar and Karvonen point out that institutional choices
may be interrelated, the choice of one device being a logical consequence of
another device (2004: 8). For example, most federal systems are characterized by a
bicameral chamber structure; two chambers have even been considered as neces-
sary in federal countries (Roskin 1986: 9). Lijphart (1991: 72–84) maintains that
architects of new democratic constitutions are confronted with two fundamental
choices, namely those between plurality and PR and between parliamentary and
presidential forms of government. By analysing all four possible combinations,
that is, the four basic types of democracy (presidential–PR, presidential–plurality,
parliamentary–PR, parliamentary–plurality), Lijphart concludes that the combina-
tion of parliamentarism and PR is the most preferable. It performs better than the
other types with respect to, among other things, voter participation, accommoda-
tion of ethnic differences and control of unemployment (1991: 81–3).
Electoral and governmental institutions interact in a number of ways. Coun-
tries with plurality systems often have two-­party systems, or at least a low degree
of party system fragmentation, which in turn produces one-­party governments.
Such executives are usually dominant in relation to the legislature. In contrast,
PR countries are likely to have multi-­party systems, coalition governments, and
more balanced relations between executive and legislative power. These two
models are known as the majoritarian (also called Westminster) model of demo-
cracy and the consensus model of democracy (Lijphart 1984; 1991; 1999). Since
the electoral system has consequences with regard to the governmental system,
it is reasonable to assume that institutional choices are to some extent interre-
lated. It should be observed that the causality is reversed in the outline above –
the present study deals with electoral system choice as a dependent variable.
Notwithstanding, we may assume that the existing governmental system influ-
ences the outcome of an electoral reform, because the chosen electoral method is
likely to affect the strength of the executive.
Furthermore, Lijphart (1991: 73) points out that a presidential form of gov-
ernment has majoritarian effects on the party system. Presidentialism is likely to
 The institutional perspective   89
create a two-­party system, because only the largest parties have a realistic chance
to win presidential elections. This advantage tends to carry over into parliament-
ary elections, particularly if both elections are held at the same time (see also
Shugart 1988; Shugart and Carey 1992). Thus, electoral and governmental insti-
tutions interact in shaping the party system. Accordingly, if constitutional engi-
neers are aware of these tendencies, we may expect the choice of one institution
to be relevant to the choice of another institution.
Returning to the models of democracy, Lijphart presents a number of charac-
teristics that emphasize either the majoritarian or the consensus model of gov-
ernment. The performance of these regime types are systematically analysed in
two volumes, Democracies (1984) and Patterns of Democracy (1999), the latter
being an updated and expanded edition of the former. The Westminster model
prescribes: (a) concentration of executive power in single-­party majority cabi-
nets; (b) cabinet control of parliament and a fusion of executive–legislative
authority; (c) assymetrical bicameralism and legislative dominance by the lower
house (or unicameral legislature); (d) a two-­party system; (e) a majoritarian
electoral system; (f) interest group pluralism; (g) unitary and centralized govern-
ment; (h) constitutional flexibility; (i) legislatures having the final word on the
constitutionality of their own legislation; and (j) central banks that are dependent
on the executive (Lijphart 1999). These institutions and devices create a logical
entity; an ideal regime type called majoritarian democracy. Majoritarian demo-
cracy means ‘government by the majority of the people’: majorities should
govern and minorities should oppose (1999: 31).
The elements that together make up the consensus model are: (a) executive
power-­sharing in broad multi-­party coalitions; (b) separation of executive and
legislative authority; (c) division of legislative power between two equally strong
but differently constituted houses; (d) multi-­party system; (e) proportional repre-
sentation; (f) corporatist interest group system; (g) federal and decentralized
government; (h) rigid constitutions that can be changed only by extraordinary
majorities; (i) judicial review by supreme or constitutional courts; and (j) inde-
pendent central banks (1999). In the same way as the elements of the Westmin-
ster model, these characteristics logically belong together, representing an
alternative model called consensus democracy. Contrasting the majoritarian
model, consensus democracy aims at maximizing the ruling majority instead of
being satisfied with a bare majority. It emphasizes consensus instead of opposi-
tion, inclusion rather than exclusion (Lijphart 1999: 33).
However, these are ideal regime types and only a handful of countries closely
approximate the Westminster and the consensus model. The United Kingdom,
New Zealand (prior to the electoral reform in 1996) and Barbados represent the
former, whereas Switzerland and Belgium are prototypes of the latter (Lijphart
1999: 10, 33). Still, democracies tend to lean towards one or the other govern-
ment types. The concepts of majoritarian and consensus democracy constitute a
framework for getting an idea of which institutional devices might correlate with
majoritarian alternatively proportional electoral systems. However, the discus-
sion on performance of countries with different institutional frameworks is
90   The institutional perspective
c­ oncerned with democracies only, whereas this book consists of considerably
more non-­democracies than democracies. Therefore, the assumed relationships
are not theoretically developed. Nonetheless, within the framework of the
present study, it is of interest to explore how electoral systems are chosen in
relation to other institutional choices. It may be repeated that the choice of polit-
ical institutions is not necessarily unimportant in non-­democracies. The task of
choosing appropriate institutions may in fact be very important in countries
taking their initial steps towards democratization.
In a volume that deals with institutional dimensions of constitutions, Karvo-
nen (2003) analyses origin, occurrence and consequences of the following insti-
tutions: territorial organization, executive power, electoral system, legislative
power and constitutionalism. Broadly speaking, these are traditional topics that
political scientists have dealt with as long as political science has existed. All of
these, except for constitutionalism, are institutions that most often take one of
two forms. A country is organized on either unitary or federal basis, it has either
a parliamentary or presidential government, the electoral system is either majori-
tarian or proportional, and the legislature consists of either one or two chambers.
Despite the existence of intermediate forms in each instance, they are basically
dichotomous phenomena. Constitutionalism, by contrast, is primarily something
of which there is more or less. The concept of constitutionalism is appreciably
more difficult to handle than the other more ‘visible’ institutions. Certainly, there
are comparative data on this subject but they reflect present constitutions – inclu-
sion of this variable in the analysis would require data from 1945 onwards. The
information needed is only obtained by studying each constitution throughout
the relevant period, which is beyond the scope of my ambition here. Therefore,
constitutionalism is not included in the analysis.
Three properties of constitutions are dealt with as possible determinants of
electoral system choice: form of government, territorial organization and
chamber structure. Although the theoretical basis is rather weak, the institutional
approach is justified with regard to the interest of analysing constitution-­building
and the performance of different institutional frameworks. On the basis of the
theory of majoritarian and consensus democracy, the assumptions are as follows:
a parliamentary form government, unitary territorial organization, and unicam-
eral legislature should be related to majoritarian systems in general and plurality
systems in particular, whereas presidentialism, federal government, and bicam-
eralism should be associated with proportional systems.

6.1 Form of government

Lijphart (1992b: 1) asserts that the relationship between the executive and the
legislature is probably the most important institutional difference among demo-
cracies. Presidentialism and parliamentarism are the two dominant forms of
democratic governmental systems. Presidentialism is characterized by separation
of powers, whereas parliamentarism represents a fusion between executive and
legislative power. In presidential regimes, the head of the government (that is,
 The institutional perspective   91
the president) is elected for a fixed, constitutionally prescribed term and cannot
normally be dismissed by a parliamentary vote of no confidence. In parliament-
ary regimes, the head of the government – most often called prime minister –
and his or her cabinet are dependent on the confidence of the legislature and may
be forced to resign from office by a legislative vote of no confidence. A second
crucial distinction, according to Lijphart’s definitions (1992b: 2–3), is that presi-
dential heads of government are popularly elected and parliamentary heads of
government are selected by the legislature. The third fundamental difference is
that parliamentary regimes have collective or collegial executives, whereas pres-
idential regimes have one-­person, non-­collegial executives. In contrast to a rela-
tively high degree of collegiality in parliamentary decision-­making, the members
of presidential cabinets are mere advisers and subordinates of the president.
Although some scholars argue that these criteria of presidentialism and parlia-
mentarism are not sufficient, most experts agree that they are fundamental
dimensions of the basic regime types (e.g. Sartori 1994: 83–4; Shugart and
Carey 1992: 19; Verney 1959).
A number of countries combine features of presidentialism and parliamen-
tarism, thereby constituting hybrid, usually called semi-­presidential forms of
government. According to Sartori (1994: 131–2), a governmental system is
semi-­presidential if the following characteristics apply. First, the president is
elected directly or indirectly by popular vote for a fixed term of office. Second,
the president shares the executive power with a prime minister. Third, the pres-
ident works independently of the legislature but he is not entitled to govern
alone. Fourth, the prime minister and his cabinet are dependent on the confi-
dence of the legislature. Fifth, the dual authority structure allows for different
balances and shifting prevalence of power within the executive. Carsten Anckar
prefers a minimal definition, prescribing that ‘there is a dual authority structure’
and that ‘executive powers are shared by a president . . . and a prime minister,
who is responsible to parliament’ (2003: 4). Shugart and Carey (1992: 23–5)
have suggested a separation of semi-­presidential systems into two categories,
premier-­presidentialism and president-­parliamentarism, depending on the power
balance between the prime minister and the president. Sartori (1994: 133) rejects
this proposal since the premier-­presidential category is very wide, leaving the
president-­parliamentary category with only a few, rather dubious cases. My clas-
sification of semi-­presidential regimes builds upon the definition by Sartori.
These are the democratic forms of government. Authoritarian regimes cannot
be classified according to this categorization. The parliamentary, semi-­
presidential and presidential models cannot work properly if basic democratic
elements are lacking. In the Derbyshire handbooks (1999: 46), one-­party regimes
are subdivided into Communist, nationalistic socialist, authoritarian nationalist
and military authoritarian. The authors write that the most common form of
executive among these countries is presidential. States with individual execu-
tives exercising virtually unbridled power are classified as absolutist regimes.
The study certainly consists of non-­democracies as well but it is of little scient-
ific interest to examine how military and other authoritarian regimes combine
92   The institutional perspective
governmental and electoral institutions. If elections are held, they probably use
the majoritarian rule. What matters is the distinction between the three primary
forms of democratic government and to what extent electoral system choice is
dependent on them. Consequently, only democratic countries are included in the
analysis.
There are, furthermore, some other forms of democratic governmental
systems that cannot be classified according to this three-­grade scheme. The
seven-­member Swiss Federal Council, for example, is elected by the legislature
for a fixed four-­year term and cannot be dismissed by a vote of no confidence. A
president, chosen by the legislature, is the head of government. The term of
office is one year and the president lacks significant powers. In Botswana,
Estonia and Sri Lanka, the president is elected by parliament. In Kiribati, Lithua-
nia, Namibia, Botswana and Sri Lanka, the president is not independent of the
legislature, because he may be dismissed by a parliamentary vote of no confi-
dence (Karvonen 2003: 59–60). These countries are excluded from the analysis.

Empirical evidence
Bivariate associations are studied by means of chi-­square tests. The number and
frequencies of electoral system choice within each category of the independent
variable are given in Table 6.1. The sample consists of 62 cases: 40 of these are
associated with a parliamentary form of government; one-­half of the parliament-
ary regimes are combined with plurality systems, one with majority, 11 with
proportional, and eight are combined with mixed systems. Most of the parlia-
mentary–plurality countries are former British colonies. The frequency of differ-
ent electoral systems among parliamentary democracies corresponds rather well
to the frequency of systems in the whole research population. A different pattern
is observed among semi-­presidential countries. Only Lebanon has combined
semi-­presidentialism with plurality, whereas three of a total of four majority
systems have been chosen by semi-­presidential regimes. Yet, two of these three
systems have been adopted in France. Countries that have chosen a semi-­
presidential form of government and mixed electoral systems are Suriname
(1975), Seychelles (1976) and Madagascar (1998). Almost one-­half of the semi-­

Table 6.1 Associations between form of government and electoral system choice

Form of Plurality % Majority % PR % Mixed %

government (N) (N) (N) (N)

Parliamentary 20 50.0 1 2.5 11 27.5 8 20.0

Semi-presidential 1 8.3 3 25.0 5 41.7 3 25.0
Presidential 6 60.0 0 0.0 1 10.0 3 30.0
N 27 43.5 4 6.5 17 27.4 14 22.6

Note
Chi-square value 14.821, sig. 0.022.
 The institutional perspective   93
presidential countries have adopted PR. Plurality systems are frequently repre-
sented in presidential regimes; only Cyprus has combined presidentialism with
PR. A total of three cases of presidentialism combined with mixed systems are
identified: Venezuela (1993), Bolivia (1997) and the Philippines (1998).
To some extent, the findings correspond with the expectations and the pattern
is significant at the 0.05 level. The majoritarian model is well represented – one-­
third of the cases combine a parliamentary form of government with plurality
elections. However, a majority of the presidential regimes are also related to the
plurality system, whereas the consensus model is represented by one country
only. Even among PR countries, the parliamentary executive model is the most
common.

6.2 Territorial organization

A nation-­state is one that claims sovereignty over the whole of its territory. This
means that everyone within its boundaries is subject to its laws. Even if the state
authority is sovereign within its boundaries, the territorial organization of the
state varies between countries. In most countries, the central state authority exer-
cises primarily unlimited power over the complete territory. In a number of
countries, however, significant parts of the state authority are located at the
regional level, where representatives and authorities within each region respec-
tively have autonomous decision-­making and executive power. This kind of ter-
ritorial organization is called federalism, whereas the former is called unitarism.
The fact that a state is unitary does not necessary mean that it is highly central-
ized. In the Scandinavian countries, for example, the political power is rather
decentralized because of a high degree of municipal autonomy. However, this is
not a federal arrangement since the central state authorities have unilaterally
decided to transfer power to the municipalities. Neither is a country federal if
one or a few regions are self-­governing, whereas all other regions are subordi-
nate to the state authorities (Karvonen 2003: 22–3).
Several definitions of federalism have been presented, largely due to the
varying degree of federal arrangements in the world. Authors that regard the
American federal organization as the prototype of federalism usually list several
necessary defining characteristics of federalism. The Encyclopedia of Demo-
cracy enumerates six fundamental principles of federalism. Federal systems are:
(a) non-­centralized; (b) they are strongly predisposed towards democracy; (c)
they rest on a system of checks and balances; (d) they operate through a process
of open bargaining among institutions and their representatives; (e) they have a
written constitution; and (f) they have constitutionally demarcated the fixed units
of power within the polity (Elazar 1995: 476–8). However, this kind of defini-
tions runs the risk of mixing a phenomenon with its preresquities and/or con-
sequences. To require that democracy is a necessary characteristic of federalism
is to assert that constitutional provisions only matter if democracy works. Sim-
ilarly, a written constitution may be a consequence of the federal structure rather
than a defining feature of federalism (Karvonen 2003: 24–5).
94   The institutional perspective
Definitions should be strict and analytical instead of detailed and empirical.
Karvonen (2003: 25) emphasizes the following three defining characteristics of
federalism as the most important:

1 The main part of the territory is divided into self-­governing regions.

Regional autonomy is the basic principle of territorial organization. Only
small deviations from this principle may occur.
2 Political power is divided between the central and the regional level. Both
must have autonomous decision-­making power over some important polit-
ical matters. Concerning these matters, the central and the regional level
must not be subordinate to any other possible administration level.
3 Equality between the central and the regional level must prevail. The distri-
bution of power must not be altered against the will of the self-­governing
regions.

Accordingly, this definition prescribes the contemporaneous occurrence of

federal territorial organization, distribution of power between the central and
regional level, and guarantees for the continuance of the federation. Countries
that do not meet any of these criteria are called unitary states. In addition to
federal and unitary states, however, there are countries that retain a unitary
nature but still have for a variety of regional and political reasons decentralized
features. In some countries, the regional autonomy is significant but concerns
only part of the territory. In other countries, there is a regional division in the
whole territory but the regional autonomy is rather weak. Denmark, for instance,
is a unitary country with federal features, Faroe Islands and Greenland being
separately administrated. France with several elected regional councils and Italy
with 20 regions are examples of countries with regional autonomy not strong
enough to be classified as federations. In Political Systems of the World, 36
countries belong to the intermediate category (Derbyshire and Derbyshire 1999:
23). However, concerning the relationship between territorial organization and
electoral systems, the distinction between strictly unitary states and unitary states
with federal features is not relevant. Since an electoral system is applied in the
whole territory, the classification of territorial organization must also be based
on principles that apply to the whole territory. Countries are therefore classified
as either federal or unitary states. Deriving from the theory of majoritarian and
consensus democracy (Lijphart 1984; 1999), I assume that federalism is related
to PR elections, and that unitary arrangements are associated with plurality and
majority systems.

Empirical evidence
Contrary to the forms of government that were analysed in the previous section,
federalism and unitarism are applied in both democracies and non-­democracies,
even though federal arrangements are more meaningful when institutions work
according to democratic principles. There are merely 21 cases of federalism in
 The institutional perspective   95
Table 6.2 Associations between territorial organization and electoral system choice

Territorial Plurality % Majority % PR % Mixed %

organization (N) (N) (N) (N)

All cases
Unitary 96 37.5 39 15.2 72 28.1 49 19.1
Federal 9 40.9 1 4.5 7 31.8 5 22.7
N 105 37.8 40 14.4 79 28.4 54 19.4
Democratic cases
Unitary 24 40.7 5 8.5 16 27.1 14 23.7
Federal 3 75.0 0 0 0.0 0.0 1 25.0
N 27 42.9 5 7.9 16 25.4 15 23.8

Notes
First comparison: chi-square value 1.902, sig. 0.593.
Second comparison: chi-square value 2.456, sig. 0.483.

the total data sample. A total of nine of the federal regimes are related to the
adoption of plurality systems. Countries that have had federal arrangements or
introduced them at the same time as plurality systems have been chosen are India
(1950), Libya (1951), Malaysia (1958), Nigeria (1960), Kenya (1963), Equato-
rial Guinea (1968), Pakistan (1970), the Federated States of Micronesia (1991)
and Palau (1994). Comoros is the only country that has combined federalism
with a majority system. A total of seven federal states have adopted proportional
electoral systems: Brazil (1945), Venezuela (1946), Congo (–Kinshasa) (1960),
Czechoslovakia (1990), Yugoslavia (1993), Bosnia-­Herzegovina (1996) and
Russia (2005). The five remaining cases are combinations of federalism and
mixed systems: West Germany (1949), Mexico (1963), Yugoslavia (1992), Ven-
ezuela (1993) and Russia (1993). The chi-­square test in Table 6.2 indicates that
territorial organization is not related to electoral system choice.
Turning to democracies, only four countries have introduced or have already
had federal arrangements when electoral systems were chosen: Malaysia, the
Federated States of Micronesia and Palau adopted plurality systems, whereas
Venezuela introduced a mixed system. Despite the tendency of combining feder-
alism with plurality systems, the number of democratic federal countries is too
small to bring about a significant relationship. Therefore, no far-­reaching conclu-
sions can be drawn. As a whole, territorial organization is not related to electoral
system choice.

6.3 Chamber structure

In a majority of the countries in the world, the representative assembly is organ-
ized in one single chamber. These regimes are called unicameral. In some coun-
tries, however, two parallel assemblies exercise legislative power – this
phenomenon is called bicameralism. Several different names of the bicameral
96   The institutional perspective
legislature are used. In the Netherlands, for example, the politically more import-
ant and directly elected part of the legislature is called the second chamber
(Tweede Kamer), whereas the first chamber (Eerste Kamer) is elected by
members of the provincial councils and is less powerful. Usually, however, the
opposite terminology is used: the ‘additional’ chamber beside the directly elected
and politically more important legislative body is called the second chamber. In
the Derbyshire volumes, the terms upper house and lower house are frequently
used (1999: 69–74). Henceforth, these terms will be applied.
Karvonen (2003: 121) suggests the following definition of bicameralism: a
country’s legislative assembly is bicameral if it is organized as two separate
chambers. This definition is clarified by an examination of two borderline cases:
Norway and the Federated States of Micronesia. In Norway, Stortinget is elected
as one single representative assembly, but after the elections, it is organized as
two separate chambers, Lagtinget and Odelstinget. In Micronesia, four of the 14
members of the federal Congress are elected by the assemblies in each state.
These members are called senators but they nonetheless work side by side with
the directly elected members in a single chamber. Karvonen (2003: 122) main-
tains that these legislatures do not qualify as bicameral. A separate election to
both chambers is a prerequisite of bicameralism. What matters is that representa-
tives stand for the legislative office and are elected (or appointed) to one of the
two chambers.
There is, however, a great deal of variety among bicameral systems. In some
bicameral countries, the upper house is politically insignificant and almost power-
less, whereas in others, both chambers are equally powerful. Differences are hence
largely manifested in symmetrical and asymmetrical forms of bicameralism. The
power constellation between the chambers is dependent on several factors. First,
the pattern of recruitment varies between upper houses. If the upper house is
elected at-­large in national elections according to the ‘one man, one vote’ prin-
ciple, both chambers represent the same interests and cannot be expected to behave
very differently from each other. Asymmetrical bicameralism prevails. In several
countries, however, the upper house represents regions instead of citizens. Indeed,
one of the main reasons for having an upper house is to help resolve regional dif-
ferences in countries with a vast territory and/or a socially and culturally diverse
population. The upper house is rather often deliberately over-­represented in order
to reduce the threat of ‘tyranny by the majority’. The most obvious example is the
United States where each state, irrespective of size, has two senators. In addition to
region, representation in the upper house may be based on ethnicity, language, reli-
gion or occupational group. Moreover, bicameral countries differ with respect to
legislative procedures. In some countries, both chambers may initiate legislation,
whereas in others, this right is reserved for the lower house only. In case of conflict
between the chambers, two principal solutions apply: either the lower house has
decisive decision-­making power or both chambers have mutual veto that forces
them to some kind of adjustment procedures.
The apparent question is whether these differences should be taken into con-
sideration in the analysis of possible patterns between chamber structure and
 The institutional perspective   97
electoral system choice. Can we expect countries with symmetrical bicameral-
ism to be in need of other electoral provisions than countries with asymmetrical
bicameralism? If chamber structure is relevant to electoral system choice, the
most fundamental dividing line is that between bicameral and unicameral coun-
tries, that is, the occurrence of an ‘additional’ chamber. In this respect, then, I
prefer an ‘either-­or’ to a ‘more-­or-less’ conception of chamber structure. On the
basis of the majoritarian and consensus model of democracy, bicameralism
should be more associated with proportional systems than majoritarian elections.

Empirical evidence
The empirical analysis is presented in Table 6.3. About two-­thirds of all cases
are associated with unicameralism – the number of one-­chamber parliaments is
175 compared to 100 cases of bicameralism.1 No clear pattern appears in the
large sample. Bicameralism is, relatively speaking, more common in democra-
cies than in non-­democracies. The number of unicameral and bicameral assem-
blies is 36 and 30 respectively. Contrary to the pattern in the total research
population, plurality systems are more associated with bicameral regimes, and
proportional systems are more related to unicameral regimes.
The large share of bicameral countries with plurality systems in the demo-
cratic sample is partly a consequence of British colonial legacy. Several former
British colonies have not only inherited their electoral systems but also the
chamber structure from the United Kingdom. Examples of such countries are
Jamaica, Trinidad and Tobago, Bahamas, Grenada, Saint Lucia and Belize.
However, there are also former British colonies with a single legislative
chamber; among others, Saint Kitts and Nevis, Dominica, Solomon Islands and
Tuvalu. Japan and Thailand represent countries with two legislative chambers
and mixed electoral systems. Turkey and the Czech Republic combine bicamer-
alism with proportional systems.

Table 6.3 Associations between chamber structure and electoral system choice

Chamber Plurality % Majority % PR % Mixed %

structure (N) (N) (N) (N)

All cases
Unicameral 70 40.0 26 14.9 49 28.0 30 17.1
Bicameral 34 34.0 13 13.0 29 29.0 24 24.0
N 104 37.8 39 14.2 78 28.4 54 19.6
Democratic cases
Unicameral 13 36.1 3 8.3 13 36.1 7 19.4
Bicameral 14 46.7 2 6.7 6 20.0 8 26.7
N 27 40.9 5 7.6 19 28.8 15 22.7

Notes
First comparison: chi-square value 2.307, sig. 0.511.
Second comparison: chi-square value 2.357, sig. 0.502
98   The institutional perspective
6.4 The institutional perspective: summary
Three variables within the institutional perspective have been analysed: form of
government, territorial organization and chamber structure. Only form of gov-
ernment is significantly related to the choice of electoral systems. However, the
pattern is far from clear-­cut. Both parliamentary and presidential regimes tend to
favour plurality systems, whereas the combination of semi-­presidentialism and
plurality is rare. Another striking feature is that proportional systems are rarely
combined with presidentialism. This conclusion might be surprising, considering
the well-­known fact that several South American countries have presidential
forms of government and proportional electoral systems. However, only one (the
semi-­presidential Guyana) of those South American countries that apply list PR
was democratic when electoral system was chosen. All other South American
countries that have adopted electoral systems in the post-­war era (under demo-
cratic circumstances) use mixed electoral systems. The frequent changes of
electoral rules in France contribute to the third distinct feature concerning regime
types and electoral system choice: majority systems are foremost associated with
semi-­presidentialism.
To all intents and purposes, no relationship exists between territorial organ-
ization and electoral system choice; neither is there a connection between
chamber structure and the dependent variable. The only discernible pattern
regarding territorial organization is that three federal democracies of a total of
four have chosen plurality systems; two of them have probably inherited not
only the electoral system but also the federal system from the United States. As
for chamber structure, there is a diminutive tendency towards a combination of
bicameralism and proportional/mixed systems as well as unicameralism and plu-
rality systems in the total data sample, and an opposite tendency to combine
bicameralism with plurality, and unicameralism with PR in democracies.
However, the connection is weak. The influence of political institutions on elect-
oral system choice must be considered very limited. Although some political
institutions may be dependent on other institutions within the constitutional
framework, the choice of electoral system does not seem to be affected by other
institutional choices.
7 Multivariate patterns

In this chapter, the strength of the explanatory variables with regard to the
dependent variable is analysed in order to distinguish the relevant determinants
of electoral system choice. Since the dependent variable is categorical, logistic
regression is applied.1 The primary method for dealing with a nominal depend-
ent variable consisting of three or more categories is multinomial logistic regres-
sion, whereas binary logistic regression is designed for dummy dependent
variables. Multinomial logistic regression selects one category as the baseline
and estimates sets of coefficients for contrasts with the other three variables
respectively. Each set of coefficients represents the effects of a one-unit change
in the explanatory variables on the logged odds of belonging to each category
relative to the reference category. In order to make a complete estimation includ-
ing all four categories of the dependent variable, another three comparisons must
be made. In each comparison, the coefficients refer only to the subset of cases
falling into the two categories used in a contrast.
A categorical dependent variable that has more than two values might also
lend itself to a set of separate binary logistic regressions. It is possible that four
categories might involve four logistic regressions with four dummy variables:
the first category versus all others, the second category versus all others, the
third category versus all others, and the fourth category versus all others.
However, the combination of several categories in making a comparison with
one category implies that precise contrast between two categories cannot be iso-
lated. The advantage of running several separate binary regressions instead of
multinomial regression is that the number of cases is larger, which means that
the results are more reliable. Therefore, the democratic sample is subject to
binary regression, while multinomial regression is preferred when the total
research population is analysed.
Since variables of three competing perspectives are included in the study,
three different regression models are presented. Party system structure and party
system transformation cannot be included in the regression analyses, which
means that the rational perspective is represented by two variables only: popula-
tion size and ethnic/cultural diversity. Either ethnic or cultural diversity must be
excluded, because they to a great extent correlate with each other.2 Cultural
diversity is preferred to ethnic diversity, since the former takes several aspects
100   Multivariate patterns
into consideration. The alternative vote is considered appropriate for ethnically
and culturally diverse countries in contrast to other majority formulas – there-
fore, the formula is excluded from those analyses that include cultural diversity.
Since only two variables of the rational perspective are analysed, I shall begin
with an analysis of patterns of diffusion. Thereafter, cultural diversity and popu-
lation size are added. In the last analysis of the entire data sample, variables of
all three perspectives are included. Another reason to begin with the cultural and
historical perspective instead of the rational perspective is to keep the number of
models at a minimum. Form of government is not analysed in the large sample,
because the variable concerns democracies only. To begin with, a decision must
be made on which variables of the historical and cultural perspective should be
included in the logistic regression. As to colonial diffusion, four variables are
conceivable: British legacy, French legacy, Soviet legacy and a fourth variable
that observes all former colonies. Legacy of other colonial powers than these are
not considered, because they possessed only a few colonies.
We noticed earlier that Soviet legacy does not correlate with electoral system
choice. Accordingly, this variable is not included in the regression analysis. A
control for multicollinearity shows that the variable that observes all cases of
colonial diffusion correlates with both British legacy and French legacy, particu-
larly with the first mentioned. Therefore, the variable that concerns all colonies
is omitted from the analysis. This variable also correlates with regional diffusion
and temporal diffusion, which further justifies the exclusion. Both regional and
temporal diffusion are operationalized as dummy variables. The latter is repre-
sented by four dummies, each variable denoting a distinct period. The first
period, which consists of considerably fewer cases than the latter periods, is not
included in the analysis.
Variables of the cultural and historical perspective are analysed in Table 7.1.
Beginning with the comparison of plurality systems with other systems, we
notice that British legacy stands out as the most important factor. French legacy
also affects the adoption of plurality systems but not to the same extent as British
legacy. The rareness of plurality choices in recent times results in a significant
effect of the fourth period when plurality systems are compared to the other
electoral system categories. The association is particularly strong when mixed
systems are compared to plurality systems, suggesting that mixed systems have
been very popular during the fourth period. Distinct trends of electoral system
choice did not exist during the second and the third period. Regional diffusion
plays a significant role in explaining the introduction of proportional systems in
comparison with majority and mixed systems. Double-­checking with four sepa-
rate binary logistic regression models confirms these results.
Cultural diversity and population size are added to the regression model in
Table 7.2 and Table 7.3. Comparisons between plurality systems and all other
systems are presented in the former table, whereas coefficients of the three
remaining comparisons are given in the latter. The chi-­square and the Nagelkerke
values are slightly higher but the −2 log likelihood value also exceeds the one in
Table 7.1, suggesting that the explanatory model is not considerably improved
Table 7.1 Patterns of diffusion and electoral system choice

Electoral systems Determinants B Std. error Prob.

Majority–plurality British legacy –3.117 1.086 **

French legacy –1.746 0.873 *
Regional diffusion –0.710 0.464
Period II (1956–1969) –0.002 0.742
Period III (1970–1989) –0.072 0.683
Period IV (1990–2007) 1.533 0.669 *
Mixed–plurality British legacy –2.522 0.819 **
French legacy –2.494 1.122 *
Regional diffusion –0.772 0.440
Period II (1956–1969) 0.277 0.773
Period III (1970–1989) 0.345 0.701
Period IV (1990–2007) 2.236 0.693 ***
PR–plurality British legacy –2.057 0.641 ***
French legacy –3.145 1.084 **
Regional diffusion 0.518 0.388
Period II (1956–1969) 0.272 0.575
Period III (1970–1989) –0.670 0.583
Period IV (1990–2007) 1.372 0.566 *
Mixed–majority British legacy 0.595 1.299
French legacy –0.748 1.338
Regional diffusion –0.063 0.463
Period II (1956–1969) 0.279 0.920
Period III (1970–1989) 0.417 0.823
Period IV (1990–2007) 0.703 0.731
PR–majority British legacy 1.060 1.192
French legacy –1.399 1.313
Regional diffusion 1.227 0.424 **
Period II (1956–1969) 0.274 0.771
Period III (1970–1989) –0.598 0.739
Period IV (1990–2007) –0.162 0.627
PR–mixed British legacy 0.465 0.956
French legacy –0.651 1.490
Regional diffusion 1.290 0.387 ***
Period II (1956–1969) –0.006 0.798
Period III (1970–1989) –1.015 0.757
Period IV (1990–2007) –0.864 0.652
Model chi-square 135.616 ***
–2 Log likelihood 135.975
Nagelkerke R square 0.416

Notes
N = 278.
* significance at the p < 0.05 level, ** significance at the p < 0.01 level, *** significance at the
p < 0.001 level.
102   Multivariate patterns
Table 7.2 Patterns of diffusion, the rational perspective and electoral system choice I

Electoral systems Determinants B Std. error Prob.

Majority–plurality British legacy –3.034 1.138 **

French legacy –1.536 0.893
Regional diffusion –0.928 0.499
Period II (1956–1969) 0.037 0.788
Period III (1970–1989) 0.168 0.706
Period IV (1990–2007) 1.717 0.706 *
Cultural diversity –0.361 0.193
Population size (log) 0.010 0.279
Mixed–plurality British legacy –2.220 0.873 *
French legacy –2.323 1.139 *
Regional diffusion –0.828 0.459
Period II (1956–1969) 0.621 0.795
Period III (1970–1989) 0.666 0.725
Period IV (1990–2007) 2.605 0.727 ***
Cultural diversity –0.370 0.181 *
Population size (log) 0.272 0.267
PR–plurality British legacy –1.816 0.678 **
French legacy –2.930 1.094 **
Regional diffusion 0.512 0.406
Period II (1956–1969) 0.496 0.590
Period III (1970–1989) –0.475 0.594
Period IV (1990–2007) 1.609 0.585 **
Cultural diversity –0.324 0.164 *
Population size (log) 0.777 1.325
Model chi-square 141.244 ***
–2 Log likelihood 572.285
Nagelkerke R square 0.435

Notes
N = 273.
For key to the signs, see Table 7.1.

by including the two variables of the rational perspective. Surprisingly enough,

cultural diversity has a significant effect on the adoption of plurality systems in
two comparisons – however, there is reason to believe that this relationship is
spurious, since it runs contrary to the expectations. Concerning patterns of diffu-
sion, the tendencies observed in the former model remain. When these findings
are double-­checked with four separate binary logistic regressions, the difference
between plurality systems and other systems is once again the most prominent of
all, whereas the comparison of majority systems with other systems returns little
to explain. Mixed systems are linked to the fourth period, and proportional
systems are associated with regional diffusion.
Variables of all three perspectives are included in Table 7.4 and Table 7.5.
Again, plurality systems are compared to the other systems in one table, while
the remaining three comparisons are given in the other. The institutional per-
spective is represented by territorial organization and chamber structure. It has
 Multivariate patterns   103
Table 7.3 Patterns of diffusion, the rational perspective and electoral system choice II

Electoral systems Determinants B Std. error Prob.

Mixed–majority British legacy 0.815 1.339

French legacy –0.786 1.346
Regional diffusion 0.100 0.492
Period II (1956–1969) 0.584 0.953
Period III (1970–1989) 0.498 0.844
Period IV (1990–2007) 0.888 0.761
Cultural diversity –0.008 0.190
Population size (log) 0.262 0.275
PR–majority British legacy 1.218 1.231
French legacy –1.394 1.322
Regional diffusion 1.440 0.456 **
Period II (1956–1969) 0.460 0.808
Period III (1970–1989) –0.643 0.758
Period IV (1990–2007) –0.108 0.656
Cultural diversity 0.037 0.181
Population size (log) 0.147 0.266
PR–mixed British legacy 0.403 0.989
French legacy –0.607 1.494
Regional diffusion 1.341 0.399 ***
Period II (1956–1969) –0.125 0.808
Period III (1970–1989) –1.141 0.772
Period IV (1990–2007) –0.997 0.674
Cultural diversity 0.045 0.162
Population size (log) –0.115 0.252
Model chi-square 141.244 ***
–2 Log likelihood 572.285
Nagelkerke R square 0.435

Notes
N = 273.
For key to the signs, see Table 7.1.

been emphasized that these two variables are connected to each other. More spe-
cifically, bicameralism is often considered necessary in federal systems. A test
of covariance in this data sample shows that they correlate to some extent but the
coefficient is smaller than 0.3 – hence, both variables can be included in the
regression model.
The model chi-­square and the Nagelkerke values are somewhat higher,
whereas the −2 log likelihood value is modestly lower than in the previous
model, indicating that the model is slightly improved by adding variables of the
institutional perspective. Still, territorial organization and chamber structure
have no influence on electoral system choice in any comparison. Again, the
adoption of plurality systems is associated with British and, to a smaller extent,
French colonial legacy. The adoption of mixed systems and the unpopularity of
plurality systems is strongly related to temporal diffusion during the fourth
104   Multivariate patterns
Table 7.4 Three explanatory perspectives and electoral system choice I

Electoral systems Determinants B Std. error Prob.

Majority–plurality British legacy –3.031 1.160 **

French legacy –1.449 0.928
Regional diffusion –1.143 0.516 *
Period II (1956–1969) –0.471 0.830
Period III (1970–1989) –0.122 0.727
Period IV (1990–2007) 1.536 0.716 *
Cultural diversity –0.327 0.199
Population size (log) 0.004 0.301
Territorial organization –1.651 1.158
Chamber structure –0.086 0.522
Mixed–plurality British legacy –2.287 0.878 **
French legacy –2.358 1.148 *
Regional diffusion –0.901 0.464
Period II (1956–1969) 0.496 0.800
Period III (1970–1989) 0.534 0.743
Period IV (1990–2007) 2.496 0.734 ***
Cultural diversity –0.374 0.185 *
Population size (log) 0.252 0.275
Territorial organization –0.340 0.730
Chamber structure 0.077 0.468
PR–plurality British legacy –1.858 0.681 **
French legacy –3.044 1.103 **
Regional diffusion 0.425 0.412
Period II (1956–1969) 0.514 0.606
Period III (1970–1989) –0.560 0.623
Period IV (1990–2007) 1.592 0.601 **
Cultural diversity –0.340 0.167 *
Population size (log) 0.203 0.249
Territorial organization –0.104 0.650
Chamber structure –0.230 0.424
Model chi-square 145.780 ***
–2 Log likelihood 559.299
Nagelkerke R square 0.450

Notes
N = 270.
For key to the signs, see Table 7.1.

period. The model also emphasizes the importance of regional diffusion in

explaining the adoption of proportional systems. As in previous models, no vari-
able explains anything when majority and mixed systems are compared with
each other. Double-­checking with four binary logistic regressions confirms these
findings.
Normally, a final model that constitutes a parsimonious compromise is pre-
sented. In this case, the most parsimonious model would consist of British
legacy, French legacy, regional diffusion and the fourth period, that is, a model
similar to the first model. When the second and the third period are excluded
 Multivariate patterns   105
Table 7.5 Three explanatory perspectives and electoral system choice II

Electoral systems Determinants B Std. error Prob.

Mixed–majority British legacy 0.743 1.354

French legacy –0.910 1.376
Regional diffusion 0.242 0.505
Period II (1956–1969) 0.967 0.993
Period III (1970–1989) 0.656 0.859
Period IV (1990–2007) 0.961 0.768
Cultural diversity –0.047 0.194
Population size (log) 0.248 0.300
Territorial organization 1.311 1.164
Chamber structure 0.163 0.507
PR–majority British legacy 1.172 1.254
French legacy –1.596 1.348
Regional diffusion 1.569 0.472 ***
Period II (1956–1969) 0.984 0.863
Period III (1970–1989) –0.438 0.781
Period IV (1990–2007) 0.057 0.671
Cultural diversity –0.013 0.186
Population size (log) 0.199 0.291
Territorial organization 1.547 1.144
Chamber structure –0.144 0.491
PR–mixed British legacy 0.429 0.994
French legacy –0.686 1.500
Regional diffusion 1.327 0.401 ***
Period II (1956–1969) 0.017 0.816
Period III (1970–1989) –1.094 0.791
Period IV (1990–2007) –0.904 0.685
Cultural diversity 0.034 0.165
Population size (log) –0.049 0.263
Territorial organization 0.236 0.686
Chamber structure –0.307 0.423
Model chi-square 145.780 ***
–2 Log likelihood 559.299
Nagelkerke R square 0.450

Notes
N = 270.
For key to the signs, see Table 7.1.

from the first model, the model chi-­square value (130.7) and the Nagelkerke
value (0.408) are only modestly affected, whereas the −2 log likelihood value is
considerably improved from 136 to 80.1. In comparison with Table 7.1, the most
parsimonious model strengthens the significance of temporal diffusion in recent
times. British and French legacy are related to the introduction of plurality
systems, whereas regional diffusion is associated with the adoption of propor-
tional systems.
As a whole, these four variables explain electoral system choice in the total
research population. In general, the effect of French legacy is weaker than the
106   Multivariate patterns
effect of British legacy, regional diffusion and temporal diffusion. The impor-
tance of temporal diffusion is confined to the frequent adoption of mixed elect-
oral systems and the decline of plurality system choices between 1990 and 2007.
In the following, the democratic sample is analysed. Majority systems are not
compared to the other electoral systems, because there are only five democratic
cases of majority systems. There is, in other words, no variance to explain con-
cerning majority systems in relation to other electoral systems. Again, patterns
of diffusion are initially analysed. French legacy is not included in the model,
because there are only two (Gabon and Vanuatu) former French colonies in the
democratic sample, none of which inherited electoral provisions from France.
Coefficients for British legacy, regional diffusion and temporal diffusion in three
separate regression models are presented in Table 7.6. British legacy stands out
as the main determinant of choosing plurality systems in comparison with other
electoral systems. Contrary to the findings in the total research population, tem-
poral diffusion does not explain the adoption of mixed systems in this model. No
variable has a significant effect when mixed systems are compared to other

Table 7.6 Patterns of diffusion and electoral system choice in democracies

Electoral system Determinants B Std. error Prob.

Plurality system British legacy 2.295 0.827 **

Regional diffusion –2.125 1.141
Period II (1958–1970) –0.416 1.080
Period III (1974–1988) –0.306 1.033
Period IV (1991–2007) –0.407 0.960
Model chi-square 26.862 ***
–2 Log likelihood 62.439
Nagelkerke R square 0.451
Proportional system British legacy –1.186 0.909
Regional diffusion 2.420 0.769 **
Period II (1958–1970) 0.755 1.102
Period III (1974–1988) 0.053 1.126
Period IV (1991–2007) –0.625 0.999
Model chi-square 18.252 **
–2 Log likelihood 60.980
Nagelkerke R square 0.346
Mixed system British legacy –1.598 1.293
Regional diffusion –1.314 0.819
Period II (1958–1970) –19.892 10267.1
Period III (1974–1988) –0.813 1.132
Period IV (1991–2007) 0.650 0.878
Model chi-square 18.414 **
–2 Log likelihood 52.333
Nagelkerke R square 0.370

Notes
N = 66.
For key to the signs, see Table 7.1.
 Multivariate patterns   107
systems. Regional diffusion explains the adoption of proportional systems rela-
tive to other electoral systems. The highest chi-­square value is observed in the
comparison between plurality systems and other systems.
Regarding the democratic sample, population size and cultural diversity
cannot be included in the same regression analysis with British legacy, because
all three variables are interrelated. First, the relationship between population size
and electoral system choice is contaminated by the peculiarities of microstates.
In contrast to the statements by Rokkan (1970), Katzenstein (1985) and
Rogowski (1987), D. Anckar (2002) asserts that plurality systems are related to
smallness. The pattern is particularly evident among diminutive microstates.
After a discussion on other potential explanations, D. Anckar (2002: 16) con-
cludes that the thesis of diffusion cannot be evaded. British influence, rather than
population size or other rational explanations, is the critical factor in this respect.
The covariance between these two variables is considerably smaller in the total
data sample. This matter is more evident among democracies, because a large
part of the former British colonies have had a high level of democracy since
independence.
Second, in the large sample, we found a significant, positive association
between cultural diversity and the adoption of plurality systems. However, there
is good reason to assume that this relationship is spurious. Certainly, the possi-
bility of affirmative gerrymandering having taken place in some countries should
not be rejected. By drawing constituency borders according to ethnic lines and
concentrating minority groups to a few electoral districts, it is possible to provide
representation of several minorities in plurality elections. In addition, many
highly fragmented countries are agrarian societies, which, according to Joel D.
Barkan (1998), are better served by plurality systems than PR. He argues that the
basic needs of the local community take precedence over ethnic cleavages when
it comes to elections in agrarian societies.
Still, the main theoretical arguments contradict the observed pattern. If the
ethnic, linguistic and religious composition of society matters to electoral system
choice, fragmented countries should, contrary to the empirical findings, choose
proportional systems. By looking around, further light is shed on the complex of
problems. Among those 25 countries that are characterized by a high level of
fragmentation, 14 have adopted plurality systems, whereas only five have
adopted proportional systems. Most of the plurality systems are inherited from
colonial powers. The Philippines, Palau and the Federated States of Micronesia
inherited their systems from the United States, whereas the United Kingdom
exported SMP to Malaysia, Jamaica, Trinidad and Tobago, Fiji, Dominica, and
the Solomon Islands. In addition, 13 of a total of 17 cases with low fragmenta-
tion are related to mixed or proportional systems. Thence, low-­fragmented coun-
tries with mixed or proportional systems cause a great deal of the correlation
between cultural diversity and electoral system choice. Accordingly, cultural
diversity and population size are not included in the second model.
In Table 7.7, variables of the institutional perspective are added to the first
model. The explanatory model is somewhat improved, the model chi-­square and
108   Multivariate patterns
Table 7.7 Three explanatory perspectives and electoral system choice in democracies

Electoral system Determinants B Std. error Prob.

Plurality system British legacy 2.855 1.007 **

Regional diffusion –2.274 1.279
Period II (1958–1970) –0.960 1.188
Period III (1974–1988) –0.217 1.131
Period IV (1991–2007) –0.973 1.136
Form of government 0.725 0.581
Territorial organization 1.246 1.556
Chamber structure –0.162 0.720
Model chi-square 32.178 ***
–2 Log likelihood 52.738
Nagelkerke R square 0.543
Proportional system British legacy –1.330 1.043
Regional diffusion 2.650 0.951 **
Period II (1958–1970) 1.234 1.212
Period III (1974–1988) –0.032 1.286
Period IV (1991–2007) –0.439 1.044
Form of government –0.985 0.625
Territorial organization –18.718 18742.6
Chamber structure –0.207 0.740
Model chi-square 21.150 **
–2 Log likelihood 51.685
Nagelkerke R square 0.418
Mixed system British legacy –1.477 1.358
Regional diffusion –1.428 0.952
Period II (1958–1970) –20.015 10060.6
Period III (1974–1988) –0.710 1.201
Period IV (1991–2007) 0.808 0.945
Form of government 0.374 0.573
Territorial organization –1.763 1.689
Chamber structure 0.224 0.748
Model chi-square 17.767 *
–2 Log likelihood 48.469
Nagelkerke R square 0.380

Notes
N = 62.
For key to the signs, see Table 7.1.

the Nagelkerke values now being larger, and the −2 log likelihood values being
smaller than in the previous table. Yet, no variable of the institutional perspec-
tive has any significant influence on the dependent variable. In accordance with
the previous regression models, the adoption of plurality systems is associated
with British colonial legacy, proportional systems are related to regional diffu-
sion, and no variable has any significant effect when mixed systems are com-
pared to other systems.
Again, the most parsimonious model is rather similar to the model in Table
 Multivariate patterns   109
7.6. Electoral system choice in democracies is mainly explained by British colo-
nial legacy and regional diffusion. Temporal diffusion is of no importance in the
democratic sample. Neither is French colonial legacy of any relevance in
explaining electoral system choice among democracies. The findings among
democracies resemble those in the total research population.
8 Conclusion

Are electoral systems permanent?

In the introduction, I pointed out that the electoral system is easier to change
than other political institutions. However, it has been claimed that the likelihood
of electoral reform should not be overestimated (Taagepera and Shugart 1989:
2). According to the academic literature, electoral system change has been quite
rare in the democratic world. In the mid-­1980s, Nohlen (1984: 218) maintained
that electoral reforms occur in extraordinary historical situations only. In the
same volume, Lijphart and Bernard Grofman addressed the following question:
how likely is electoral reform? Their answer to the question was that electoral
system changes are possible but they may not be highly probable (1984b:
11–12). Many other scholars have characterized electoral institutions as ‘sticky’,
implying that they are difficult to change (e.g. Birch et al. 2002: 1; Geddes 1996:
31). Several reasons for this stickiness have been given. An apparent reason why
major changes are unlikely is that those who are capable of reforming the elect-
oral system are those who have been successful under the current system.
Conventional wisdom among political scientists has stressed that electoral
institutions reflect deep-­rooted aspects of society and political life across demo-
cracies – therefore, countries change their electoral rules very rarely, if at all.
Electoral systems reflect the politics of the time of their creation and are only
altered when radical political change occurs; a change that makes the existing
electoral system too restrictive (Taagepera and Shugart 1989: 234). However,
major reforms in several established democracies since the early 1990s have
seriously challenged the notion that electoral systems are permanent. In addition,
electoral system change among democracies before the 1990s was not as rare as
many authors intimate. And what is more, little attention, if any, has been given
to electoral reforms in non-­democratic countries.
This book covers all major electoral system choices for parliamentary elec-
tions in the world during the post-­war era. A total of 278 cases of electoral
system choice were indentified. Of these, 140 are first choices in (for the most
part) newly independent countries, whereas the remaining 138 observations are
cases of major electoral system change. Considering that several authors have
emphasized the stickiness of electoral systems, in the sense that major changes
 Conclusion   111
of electoral rules are rare, the observed frequency of electoral system reform is
quite astonishing. To be sure, the permanent nature of electoral systems concerns
democracies – however, the findings in this study show that electoral reforms are
almost as frequent in democracies as in non-­democracies. Of a total of 66 cases
in the democratic sample, 28 constitute major electoral reforms, that is, a change
from one of the four main systems to another. Of these, 13 have taken place in
the 1990s and 2000s. However, before 1960, France, Italy and Greece changed
their electoral systems several times. Furthermore, major electoral reforms
undertaken by a democratic regime have taken place in Iceland (1959), Lebanon
(1953) and Turkey (1961). Also, the electoral reforms in Sri Lanka (1978) and
Cyprus (1981) qualify as democratic cases. A great majority of the long-­standing
democracies have certainly retained their electoral systems for nearly a century
(or more) but there is plenty of evidence for dismissing the notion that electoral
rules are very difficult to change.

Summary of findings
The purpose of this book is to explain why countries have chosen different elect-
oral systems. A broad design including explanations from three competing but
simultaneously overlapping theoretical perspectives has been applied. The analy-
ses show that the explanatory power of contextual factors varies considerably
with regard to both general tendencies as well as patterns at different points in
time.
The rational perspective maintains that electoral systems serve certain pur-
poses. Countries with different sociopolitical circumstances need different elect-
oral systems in order to deal successfully with these traits. As a whole, the
rational perspective provides a rather weak explanatory model. Despite the fact
that a lot of attention has been given to electoral system design in divided socie-
ties, electoral systems have rarely been chosen by taking ethnic and other cleav-
ages into consideration. Proportional systems are considered appropriate for
countries with deep cleavages by securing representation of minority groups.
The alternative vote is likewise regarded as a conflict-­mitigating system, because
it promotes accommodative behaviour by means of vote-­pooling. The share of
electoral systems considered inappropriate is, nevertheless, much larger than the
share of proportional systems and the alternative vote in plural societies.
The theoretical arguments concerning the influence of country size on elect-
oral system choice suggest that small countries should apply proportional
systems. Several reasons have been given; for instance, that the adoption of PR
in small countries reflects a search for consensus and compromise, and that PR
is inappropriate in large entities because of heavier burdens of responsibility.
The opposite pattern is observed, however. There is a significant correlation
between plurality systems and smallness – the relationship is even stronger in
the democratic sample. The frequent adoption of plurality systems in small coun-
tries is a consequence of the British colonial empire, which comprised lots of
small island states. Yet, very large countries deviate from the general pattern by
112   Conclusion
having chosen mostly majoritarian and mixed electoral systems. This tendency
is most distinct during the fourth period. An advantage of these systems is that
the use of single-­member districts provides geographic accountability through
closer bonds between constituents and their representatives. This matter becomes
increasingly important in large countries. However, it should be observed that it
concerns only a few cases in the research population. In general, population size
offers little explanation of electoral system choice.
The impact of party system structure was analysed by observing cases of
electoral system change. Countries that have replaced plurality systems with PR
were compared to countries that have been applying plurality systems in several
subsequent elections. A similar comparison was made between changes from PR
to plurality and PR systems that have been applied for a long time. Extended
samples including cases of change to mixed systems were also analysed. The
only significant relationship in the expected direction was observed in the com-
parison between plurality systems and changes to mixed or proportional systems
among democracies. In spite of the negative findings in all other comparisons,
this association is worthy of attention. When dealing with party system structure,
more attention should be paid to patterns among democratic countries. The com-
parisons between changes from PR to plurality (or mixed systems) and long-­
standing PR systems are as good as futile, because there are not enough
empirical cases of the former. The significant association should, in other words,
be regarded as an indication that the party system structure has at least a minor
influence on the choice of electoral systems. All cases that contribute to the
significant relationship belong to the fourth period, thereby contributing to the
importance of the rational perspective during the last few decades.
Party system transformation from absolute one-­party rule to competitive pol-
itics in countries that previously have held elections under the majoritarian rule
was assumed to result in the adoption of mixed systems or PR. This assumption
builds upon the theory of the dynamic of the democratization process. A total of
62 cases of party system transformation were observed during the post-­war era.
In almost one-­half of them, mixed or proportional systems were adopted prior to
the first multi-­party elections. In another nine countries, electoral reforms in the
expected direction took place between the first and the second competitive
parliamentary elections. Nearly all electoral system changes have occurred
during the last two decades. Accordingly, party system transformation has been
a main determinant of electoral system choice in recent times.
The historical and cultural setting has a great impact on electoral system
choice – all three variables are of considerable importance. Numerous countries
have inherited their electoral systems from the former colonial powers. The
United Kingdom is associated with the largest share of inherited systems. The
impact of British colonial legacy is confirmed in the multiple analyses, even
though British legacy is by no means superior to other relevant explanatory
factors. More exactly, the United Kingdom has exported the single-­member plu-
rality formula and the block vote to several former colonies. French colonial
legacy is also of some importance in explaining electoral system choice. Among
 Conclusion   113
democracies, however, a French legacy does not exist, because only two French
colonies were democratic when independence was proclaimed. The fact that
block vote is not a popular electoral formula anymore may be seen as an indica-
tor of French legacy not being as deep-­rooted and strong as British legacy. The
impact of British colonial rule has been regarded as positive in several ways,
whereas the rule by several other colonial powers has been considered intrusive
and exploitative. Possibly, these circumstances explain why several former colo-
nies have preferred other systems than those of their former masters. In this con-
nection, it is important to note that all four countries previously under American
colonial rule, which is considered as reminiscent of British colonialism, adopted
plurality systems.
Transfer of institutions has taken place from other countries than colonial
powers, too. In the process of constitution-­building, it is natural to look at
arrangements in culturally akin countries, and make similar decisions as they
have. The empirical analyses provide support for the theoretical arguments.
Regional diffusion is primarily associated with the adoption of proportional
systems, mostly in Europe and Latin America. Another finding is that several
countries that were formerly part of the Soviet Union adopted similar electoral
systems as independent countries, namely mixed-­superposition. Possibly, there
has been a perception in the region that this system is superior to other systems.
The findings concerning regional influence emphasize the role of cognitive heu-
ristics as the logic behind policy adoptions in general and institutional choices in
particular. The unique availability of electoral reforms implemented in the neigh-
bourhood often puts similar reforms on the agenda in other countries nearby.
Temporal diffusion is the third variable within the historical and cultural per-
spective. The exposition of electoral system choice over time results in a classifi-
cation into four distinct periods during the post-­war era, each one characterized
by rather few adopted systems at the end of the period. The second phase of this
analysis consisted of examining which systems have been the most attractive
during each period. Plurality systems have been frequently adopted during the
first three periods, that is, from 1945 to 1989, but have lost popularity during the
fourth period. Mixed systems show the opposite pattern: after having been rather
unpopular before 1990, they represent almost one-­third of all choices during the
fourth period. Proportional systems have also been frequently adopted in recent
times. The findings among democracies resemble those in the total data sample.
The multivariate analyses show that temporal diffusion has a significant effect
on electoral system choice. More exactly, temporal diffusion is related to the fre-
quent adoption of mixed systems during the fourth period.
However, the fact that a given electoral system is very frequently adopted
during a certain epoch does not sufficiently explain why this electoral system is
chosen. The third step in the analysis of temporal diffusion was therefore to
examine whether different explanations apply to different epochs of electoral
system choice. Regional diffusion is the most important variable during the first
period from 1945 to 1955. The impact of colonial and regional diffusion is
immense during the second period between 1956 and 1969. Almost every
114   Conclusion
adopted plurality system at that time was a consequence of either colonial legacy
or regional influence. Party system transformation was of some importance
between 1970 and 1989. Nonetheless, colonial and regional diffusion are the
main determinants also during the third period. Surprisingly enough, regional
diffusion is related to the adoption of plurality systems rather than PR. The
rational perspective becomes a relevant explanatory model during the fourth
period from 1990 to 2007. Party system transformation, in particular, is related
to the adoption of mixed and proportional systems. Population size and party
system structure may also explain the introduction of mixed electoral systems in
some countries. Only a few colonies became independent during the fourth
period. Regional diffusion, in contrast, is still an important variable, related to
mixed and proportional systems.
The third perspective sees the adoption of electoral institutions as a con-
sequence of other characteristics of the constitutional framework. Some institu-
tional arrangements follow logically upon others and together they constitute a
coherent whole. However, the empirical findings are negative. To be sure, some
tendencies concerning the adoption of other institutions and electoral system
choice are discernible at various points in time but they are to a great extent a
consequence of British colonialism.

What next?
I predict that that the trend of growing importance of rational explanations and
declining relevance of colonial legacy that we have witnessed during the last two
decades will continue in the future. A majority of the former British colonies
will probably continue applying the single-­member plurality formula for a long
time to come, thereby keeping the British colonial legacy alive, but its relative
strength is likely to decline. It is worth emphasizing that the importance of
British legacy relative to other determinants is not as predominant as the literat-
ure on the origin of electoral systems often maintains. Likewise, on the basis of
the findings here, deliberate electoral system design has not been as prominent
during the third wave of democratization as some authors intimate. Notwith-
standing, the extent to which electoral systems are carefully chosen on the basis
of an extensive debate will probably increase in years to come. Factors related to
party systems and ethnic diversity are likely to influence electoral system choice
in the future. An element of regional diffusion will most likely be present in the
future as well; there has always been a tendency to imitate choices of adjacent
countries and successful actors.
Proportional systems will continue to be popular, not least because they
provide a rather stable parliamentary arena for the parties involved. Future adop-
tion of the single-­member plurality formula depends partly on whether several
autonomous British and American territories will become independent, partly on
whether more states will become democratic. In democratizing countries that
implement electoral reforms in the future, SMP will hardly be considered a real
option. A two-­ballot majority formula will probably not gain much attraction
 Conclusion   115
either. Most of the existing cases are non-­democracies in Africa, Caucasus and
Central Asia. Future democratization processes in these countries are likely to
result in the replacement of two-­ballot majority with more inclusive electoral
provisions. The alternative vote will not get a prominent position – it has always
been regionally restricted to the South Pacific. Mixed systems in general, and
superposition, in particular, have been frequently adopted since the early 1990s.
The apparent question is whether the popularity of mixed systems with an inde-
pendent combination of plurality/majority and PR will persist. The attraction
might continue for a while, but a plausible scenario is that more and more coun-
tries will come to agree with Sartori (1994: 75) who has described these mixed
systems as a bastard-­producing hybrid that combines the defects of majoritarian
and PR systems. A logical consequence of this insight is to introduce a correc-
tive mechanism to the mixed system, or, more likely, to replace the hybrid with
a proportional list system.
Appendix

Table A.1 Electoral system choice by country 1945–2007

Country Democracy Electoral system

Afghanistan (2004) no SNTV

Albania (1992) no Mixed-corrective
Algeria (1962) no Block vote
Algeria (1991) no Two-ballot majority
Algeria (1997) no List PR
Andorra (1993) no Mixed-superposition
Angola (1991) no List PR
Antigua and Barbuda (1981) yes Single-member plurality
Argentina (1957) no List PR
Armenia (1995) no Mixed-superposition
Azerbaijan (1995) no Mixed-superposition
Bahamas (1974) yes Single-member plurality
Bahrain (1973) no SNTV
Bahrain (2002) no Two-ballot majority
Bangladesh (1972) no Single-member plurality
Barbados (1966) yes Single-member plurality
Belarus (1991) no Two-ballot majority
Belau (1994) yes SMP-BV
Belize (1981) yes Single-member plurality
Benin (1960) no Block vote
Benin (1964) no List PR
Bolivia (1956) no List PR
Bolivia (1997) yes Mixed-corrective
Bosnia-Herzegovina (1996) no List PR
Botswana (1966) no Single-member plurality
Brazil (1945) no List PR
Bulgaria (1953) no List PR
Bulgaria (1956) no Two-ballot majority
Bulgaria (1990) no Mixed-superposition
Bulgaria (1991) yes List PR
Burkina Faso (1965) no Block vote
Burkina Faso (1970) no List PR
Burma (1948) no SMP–BV

continued
 Appendix   117
Table A.1 Continued

Country Democracy Electoral system

Burma (1974) no One-ballot majority

Burma (1989) no Single-member plurality
Burundi (1965) no Single-member plurality
Burundi (1993) no List PR
Cambodia (1953) no Single-member plurality
Cambodia (1992) no List PR
Cameroon (1960) no SMP–BV
Cameroon (1992) no Supermixed
Cape Verde (1975) no Block vote
Cape Verde (1992) no List PR
Central African Republic (1960) no Block vote
Central African Republic (1998) no Two-ballot majority
Chad (1960) no Block vote
Chad (1996) no Supermixed
Comoros (1978) no Two-ballot majority
Congo (Brazzaville) (1960) no Block vote
Congo (Brazzaville) (1992) no Two-ballot majority
Congo (Kinshasa) (1960) no List PR
Congo (Kinshasa) (1967) no Block vote
Congo (Kinshasa) (2006) no Mixed-superposition
Côte d’Ivoire (1960) no Block vote
Côte d’Ivoire (1980) no Two-ballot majority
Côte d’Ivoire (1985) no Block vote
Croatia (1991) no Mixed-superposition
Croatia (1999) no List PR
Cuba (1992) no One-ballot majority
Cyprus (1960) yes Block vote
Cyprus (1981) yes List PR
Czechoslovakia (1990) no List PR
Czech Republic (1993) yes List PR
Djibouti (1977) no Block vote
Dominica (1978) yes Single-member plurality
Dominican Republic (1962) no List PR
East Timor (2002) no List PR
Ecuador (1946) no List PR
Ecuador (1978) no Supermixed
Egypt (1950) no Two-ballot majority
Egypt (1983) no List PR
Egypt (1986) no Mixed-superposition
Egypt (1990) no Two-ballot majority
El Salvador (1963) no List PR
Equatorial Guinea (1968) no Single-member plurality
Equatorial Guinea (1993) no List PR
Estonia (1992) yes List PR
Ethiopia (1955) no Block vote
Fiji (1970) yes Block vote
Fiji (1999) no Alternative vote
France (1945) yes List PR
France (1951) yes Mixed-conditional
France (1958) yes Two-ballot majority
118   Appendix
Table A.1 Continued

Country Democracy Electoral system

France (1986) yes List PR

France (1988) yes Two-ballot majority
Gabon (1960) yes Block vote
Gabon (1990) no Two-ballot majority
Gambia (1965) no Single-member plurality
Georgia (1992) no Mixed-superposition
Germany (East) (1949) no Block vote
Germany (East) (1990) no List PR
Germany (West) (1949) no Mixed-corrective
Ghana (1957) no Single-member plurality
Greece (1952) yes Single-member plurality
Greece (1956) yes Mixed-coexistence
Greece (1958) yes List PR
Grenada (1974) yes Single-member plurality
Guatemala (1946) no Mixed-coexistence
Guatemala (1956) no List PR
Guinea (1963) no Block vote
Guinea (1990) no Mixed-superposition
Guinea-Bissau (1976) no Block vote
Guinea-Bissau (1994) no List PR
Guyana (1966) yes List PR
Haiti (1995) no Two-ballot majority
Honduras (1957) no List PR
Hungary (1966) no Two-ballot majority
Hungary (1989) no Supermixed
Iceland (1959) yes List PR
India (1950) no SMP–BV
Indonesia (1953) no List PR
Iran (1963) no SMP–BV
Iran (1979) no Two-ballot majority
Iraq (1952) no Two-ballot majority
Iraq (1995) no Block vote
Iraq (2005) no List PR
Israel (1948) yes List PR
Italy (1953) yes Mixed-conditional
Italy (1954) yes List PR
Italy (1993) yes Mixed-corrective
Italy (2005) yes List PR
Jamaica (1962) yes Single-member plurality
Japan (1946) no SNTV
Japan (1993) yes Mixed-superposition
Jordan (1947) no SMP–BV
Kazakhstan (1993) no Single-member plurality
Kazakhstan (1995) no Two-ballot majority
Kazakhstan (1999) no Mixed-superposition
Kazakhstan (2007) no List PR
Kenya (1963) no Single-member plurality
Kiribati (1978) yes Two-ballot majority
Korea (North) (1948) no One-ballot majority
Korea (South) (1948) no Single-member plurality
 Appendix   119
Table A.1 Continued

Country Democracy Electoral system

Korea (South) (1963) no Mixed-superposition

Korea (South) (1973) no Block vote
Korea (South) (1980) no Mixed-superposition
Kuwait (1962) no Block vote
Kyrgyzstan (1994) no Two-ballot majority
Kyrgyzstan (1999) no Mixed-superposition
Kyrgyzstan (2007) no List PR
Laos (1953) no SMP–BV
Laos (1960) no Two-ballot majority
Laos (1965) no SMP–BV
Latvia (1993) yes List PR
Lebanon (1953) yes Block vote
Lesotho (1966) no Single-member plurality
Lesotho (2002) no Mixed-corrective
Liberia (1985) no Two-ballot majority
Liberia (1997) no List PR
Libya (1951) no Single-member plurality
Lithuania (1992) yes Mixed-superposition
Macedonia (1994) no Two-ballot majority
Macedonia (1998) no Mixed-superposition
Macedonia (2002) no List PR
Madagascar (1960) no Supermixed
Madagascar (1977) no Block vote
Madagascar (1983) no Mixed-coexistence
Madagascar (1993) no List PR
Madagascar (1998) yes Mixed-coexistence
Madagascar (2007) no SMP
Malawi (1964) no Single-member plurality
Malaysia (1958) yes Single-member plurality
Maldives (1968) no SNTV
Mali (1960) no Block vote
Mali (1992) no Two-ballot majority
Malta (1964) yes STV
Marshall Islands (1991) yes Single-member plurality
Mauritania (1960) no Single-member plurality
Mauritania (1992) no Two-ballot majority
Mauritania (2006) no Mixed-superposition
Mauritius (1968) no Block vote
Mexico (1963) no Mixed-corrective
Micronesia (1991) yes Single-member plurality
Moldova (1993) no List PR
Monaco (1958) no Two-ballot majority
Monaco (2002) yes Supermixed
Mongolia (1949) no Two-ballot majority
Mongolia (1992) no Block vote
Montenegro (2006) yes List PR
Morocco (1962) no List PR
Morocco (1970) no Single-member plurality
Morocco (1997) no List PR
Mozambique (1994) no List PR
120   Appendix
Table A.1 Continued

Country Democracy Electoral system

Namibia (1990) no List PR

Nauru (1968) yes Alternative vote
Nepal (1959) no Single-member plurality
Nepal (2007) no Mixed-superposition
New Zealand (1996) yes Mixed-corrective
Nicaragua (1962) no Mixed-fusion
Nicaragua (1984) no List PR
Niger (1960) no Block vote
Niger (1993) no List PR
Nigeria (1960) no Single-member plurality
Pakistan (1970) no Single-member plurality
Panama (1972) no Single-member plurality
Panama (1980) no Mixed-coexistence
Papua New Guinea (1975) yes Single-member plurality
Papua New Guinea (2002) yes Alternative vote
Paraguay (1947) no List PR
Paraguay (1967) no Mixed-fusion
Paraguay (1990) no List PR
Philippines (1946) yes Block vote
Philippines (1998) yes Mixed-superposition
Philippines (2001) yes Single-member plurality
Poland (1991) no List PR
Portugal (1975) no List PR
Romania (1990) no List PR
Russia (1993) no Mixed-superposition
Rwanda (1962) no List PR
Rwanda (1978) no Block vote
Rwanda (2003) no List PR
St Kitts and Nevis (1983) yes Single-member plurality
St Lucia (1979) yes Single-member plurality
St Vincent and Grenadines (1979) yes Single-member plurality
Samoa (Western) (1963) no SMP–BV
San Marino (1951) no List PR
Sao Tome and Principe (1991) no List PR
Senegal (1963) no Block vote
Senegal (1978) no List PR
Senegal (1983) no Mixed-superposition
Serbia (2006) yes List PR
Seychelles (1976) yes Mixed-superposition
Seychelles (1979) no Single-member plurality
Seychelles (1993) no Mixed-superposition
Sierra Leone (1961) no Single-member plurality
Sierra Leone (1996) no List PR
Singapore (1965) no Single-member plurality
Slovakia (1993) yes List PR
Slovenia (1991) no List PR
Solomon Islands (1978) yes Single-member plurality
Somalia (1964) no List PR
Somalia (1979) no Block vote
South Africa (1993) no List PR
 Appendix   121
Table A.1 Continued

Country Democracy Electoral system

Spain (1977) no List PR

Sri Lanka (1948) yes Single-member plurality
Sri Lanka (1978) yes List PR
Sudan (1956) no Single-member plurality
Suriname (1975) yes Mixed-coexistence
Suriname (1987) no List PR
Swaziland (1968) no Block vote
Syria (1947) no Two-ballot majority
Syria (1973) no Block vote
Taiwan (1946) no SNTV
Taiwan (1991) no Mixed-superposition
Tajikistan (1994) no Two-ballot majority
Tajikistan (1999) no Mixed-superposition
Tajikistan (2005) no Two-ballot majority
Tanzania (1964) no Single-member plurality
Thailand (1997) yes Mixed-superposition
Togo (1960) no Single-member plurality
Togo (1992) no Two-ballot majority
Togo (2007) no List PR
Trinidad and Tobago (1962) yes Single-member plurality
Tunisia (1959) no Block vote
Tunisia (1997) no Mixed-corrective
Turkey (1946) no Block vote
Turkey (1961) yes List PR
Turkmenistan (1994) no Two-ballot majority
Tuvalu (1978) yes Single-member plurality
Uganda (1980) no Single-member plurality
Ukraine (1993) no Two-ballot majority
Ukraine (1997) no Mixed-superposition
Ukraine (2004) no List PR
Uzbekistan (1992) no Two-ballot majority
Vanuatu (1980) yes SNTV
Venezuela (1946) no List PR
Venezuela (1993) yes Mixed-corrective
Vietnam (1946) no Block vote
Vietnam (1976) no Two-ballot majority
Vietnam (South) (1956) no Single-member plurality
Yemen (1993) no Single-member plurality
Yemen (North) (1971) no Single-member plurality
Yemen (South) (1978) no Block vote
Yugoslavia (1992) no Mixed-superposition
Yugoslavia (1993) no List PR
Zambia (1964) no Single-member plurality
Zimbabwe (1980) no Mixed-coexistence
Zimbabwe (1985) no Single-member plurality
Notes

2 Research design
1 A plurality system has been used ever since the first elections.
2 Note that this value is not the same as turnout, since turnout in elections is based on the
size of the electorate instead of the total population.
3 Correlation coefficients are negative, because Freedom House ratings rise with the
decline of democratization, whereas Vanhanen’s indices rise with the increase of
democratization.

3 Electoral systems
1 The reason why STV is treated as a basic type of electoral system may be attributed to
the pro-­STV nature of the book.

4 The rational perspective

1 When Papua New Guinea became independent in 1975, an electoral reform replaced
AV with SMP, which resulted in a highly fragmented party system and less accommo-
dating political behaviour. AV was reintroduced in 2002.
2 Plurality and majority systems are treated as a single category of electoral systems.
Mixed systems and proportional systems constitute the other major types of electoral
systems. Among the majoritarian systems, the plurality formula is considerably more
frequent than the majority formula.
3 The following correlation values have been reported: 0.69 (Dahl and Tufte 1973: 17),
0.83 (Hadenius 1992: 126), and 0.87 (C. Anckar 1998: 56).
4 Mixed systems are included since the number of electoral system changes is rather
small, particularly changes from PR to plurality systems and changes among democra-
cies in general.
5 In this connection, the term multi-­party system implies the existence of several parties
and parliamentary elections being contested by more than one party, that is, the oppos-
ite of one-­party system. It is more common to conceive multi-­party system as the main
alternative to two-­party system in democracies.
6 In Benin, for example, list PR has officially been in force since the 1964 elections,
although multi-­party elections did not take place until 1991 (Hartmann 1999: 82–3).
7 It should be observed that the analysis in this section also consists of some electoral
system changes that are not observed as reforms in the main database, since the
changes in question took place prior to independence – the Baltic countries are cases in
point. This matter concerns seven of the 62 cases above.
8 These countries are Albania, Algeria, Armenia, Azerbaijan, Cape Verde, Kyrgyzstan,
Poland, Tajikistan and Ukraine.
 Notes   123
5 History and culture: patterns of diffusion
1 In May 1992, Yugoslavia adopted mixed-­superposition after having previously applied
SMP. In the 1993 and 1996 parliamentary elections, list PR was applied.
2 Actually, there were two British sections. The Northern section voted for integration
into Nigeria, whereas the Southern section decided to join the Republic of Cameroon
(Mehler 1999: 167).
3 The total number of cases is 35, because Vanuatu is coded as both a British and a
French colony.
4 According to the electoral law, a mixed system was still applied after the reform in
2001. However, the magnitude of the nationwide constituency was reduced from 52 to
merely five seats prior to the 2001 elections, which means that the Filipino electoral
system is, to all intents and purposes, a plurality system (at the end of 2007). The
number of single-­member districts is 209.

6 The institutional perspective

1 Information on three cases is lacking: East Germany 1949 and Bulgaria 1953 as well as
1956.

7 Multivariate patterns
1 There are basically three different values to estimate in logistic regression: probabili-
ties, odds and logged odds. Logistic regression is basically a matter of estimating prob-
abilities, that is, the likelihood of an occurrence relative to the likelihood of a
non-­occurrence. However, the likelihood of an occurrence depends on the level of the
independent variable, which means that the relationships between independent vari­
ables and probabilities cannot be fully expressed by a single coefficient. By dividing
the probability value (which ranges from 0 to 1) with one minus the probability value,
the odds of having an event are received. The final step in the process is to take the
natural logarithm of the odds, thereby obtaining the logged odds, which are represented
by the B-­coefficients in the SPSS output. The B-­coefficient indicates the change in the
logged odds of experiencing an event or having a characteristic for a one-­unit change
in the independent variable.
   In addition to the logged odds and the odds, the statistic program used provides the
standard error, the Wald value, and the significance value for each independent varia-
ble. The Wald value is the square of the ratio of the coefficient divided by its standard
error. Contrary to the logged odds, the Wald values of the independent variables are
comparable with each other irrespective of the measure scale. However, Fred Pampel
(2000: 30) maintains that with a large absolute value for the logistic regression coeffi-
cient, the estimated standard error might lack precision because of rounding error,
thereby providing an incorrect test of significance. Because statistical significance is
dependent on sample size, he continues, the significance value provides little informa-
tion on the strength of the relationship between independent and dependent variables.
Particularly large samples can produce significant values for otherwise inconsiderable
effects. In the present analysis, the logged odds (B-­coefficient), standard error and sig-
nificance value are presented.
2 Spearman’s rho is 0.68 in the total research population, and 0.72 in the sample of
democracies.
Bibliography

Albertini, R. von (1982) European Colonial Rule, 1880–1940: The Impact of the West on
India, Southeast Asia, and Africa, Oxford: Clio Press.
Anckar, C. (1998) Storlek och partisystem: en studie av 77 stater, Åbo: Åbo Akademis
förlag.
Anckar, C. (2002) Effekter av valsystem: en studie av 80 stater, Stockholm: SNS förlag.
Anckar, C. (2003) ‘From Fusion of Powers to Separation of Powers: A Quantitative
Measure of Form of Government’, paper presented at the nineteenth IPSA World Con-
gress, Durban, 29 June–4 July 2003.
Anckar, D. (2002) ‘Valsystem i lilleputtar: pluralitet och diffusion’, Politiikka, 1: 4–17.
Anckar, D. (2004) ‘Regime Choices in Microstates: The Cultural Constraint’, Common-
wealth and Comparative Politics, 42: 206–23.
Anckar, D. and Karvonen, L. (2004) ‘Constitutional Amendment Methods in the Demo-
cracies of the World’, unpublished manuscript, Åbo Akademi University.
Barkan, J.D. (1998) ‘Rethinking the Applicability of Proportional Representation for
Africa’, in T. Sisk and A. Reynolds (eds), Elections and Conflict Management in
Africa, Washington, DC: United States Institute of Peace Press.
Birch, S., Millard, F., Popescu, M. and Williams, K. (2002) Embodying Democracy:
Electoral System Design in Post-­Communist Europe, New York: Palgrave Macmillan.
Blais, A. (1988) ‘The Classification of Electoral Systems’, European Journal of Political
Research, 16: 99–110.
Blais, A., Dobrzynska, A. and Indridason, I.H. (2005) ‘To Adopt or Not to Adopt Propor-
tional Representation: The Politics of Institutional Choice’, British Journal of Political
Science, 35: 182–90.
Blais, A. and Massicotte, L. (1996) ‘Electoral Systems’, in L. LeDuc, R.G. Niemi and P.
Norris (eds), Comparative Democratic Elections, Thousand Oaks: Sage.
Blais, A. and Massicotte, L. (1997) ‘Electoral Formulas: A Macroscopic Perspective’,
European Journal of Political Research 32: 107–29.
Bogdanor, V. (1983) ‘Introduction’, in V. Bogdanor and D. Butler (eds), Democracy and
Elections: Electoral Systems and their Political Consequences, Cambridge: Cambridge
University Press.
Bogdanor, V. and Butler, D. (eds) (1983) Democracy and Elections: Electoral Systems
and their Political Consequences, Cambridge: Cambridge University Press.
Boix, C. (1999) ‘Setting the Rules of the Game: The Choice of Electoral Systems in
Advanced Democracies’, American Political Science Review, 93: 609–24.
Brady, D. and Mo, J. (1992) ‘Electoral Systems and Institutional Choice: A Case Study
of the 1988 Korean Elections’, Comparative Political Studies, 24: 405–29.
 Bibliography   125
Britannica Book of the Year (1999) London: Encyclopedia Britannica, Inc.
Carstairs, A.M. (1980) A Short History of Electoral Systems in Western Europe, London:
Allen & Unwin.
Chronicle of Parliamentary Elections (1967–2008), Geneva: Inter-­Parliamentary Union.
Colomer, J.M. (ed.) (2004a) Handbook of Electoral System Choice, New York: Palgrave
Macmillan.
Colomer, J.M. (2004b) ‘The Strategy and History of Electoral System Choice’, in J.M.
Colomer (ed.), Handbook of Electoral System Choice, New York: Palgrave Macmillan.
Colomer, J.M. (2005) ‘It’s Parties That Choose Electoral Systems (or, Duverger’s Laws
Upside Down)’, Political Studies, 53: 1–21.
Cox, G.W. (1997) Making Votes Count: Strategic Coordination in the World’s Electoral
Systems, Cambridge: Cambridge University Press.
Dahl, R.A. and Tufte, E.R. (1973) Size and Democracy, Stanford: Stanford University Press.
Derbyshire, J.D. and Derbyshire, I.D. (1989) Political Systems of the World, Edinburgh:
Chambers.
Derbyshire, J.D. and Derbyshire, I.D. (1999) Political Systems of the World, vol. I and
vol. II, Oxford: Helicon Publishing.
Diamond, L. (1999) Developing Democracy: Toward Consolidation, Baltimore: John
Hopkins University Press.
Diamond, L. and Lipset, S.M. (1995) ‘Colonialism’, in S.M. Lipset (ed.), The Encyclope-
dia of Democracy, vol. I, London: Routledge.
Diamond, L. and Plattner, M.F. (eds) (2006) Electoral Systems and Democracy, Balti-
more: Johns Hopkins University Press.
Duverger, M. (1951) Les Partis Politiques, Paris: Armond Colin.
Duverger, M. (1964) Political Parties: Their Organization and Activity in the Modern
State, reprinted version, London: Methuen.
Elazar, D.J. (1995) ‘Federalism’, in S.M. Lipset (ed.), The Encyclopedia of Democracy,
vol. II, London: Routledge.
Farrell, D.M. (1998) Comparing Electoral Systems, London: Macmillan Press.
Farrell, D.M. (2001) Electoral Systems: A Comparative Introduction, New York: Pal-
grave.
Flora, P., Kraus, F. and Rothenbacher, F. (eds) (2000) The Societies of Europe: Elections
in Western Europe since 1815, London: Macmillan.
Freedom House. Online, available at: www.freedomhouse.org/.
Freedom in the World: The Annual Survey of Political Rights and Civil Liberties, 2000–
2001 (2001), New York: Freedom House.
Gallagher, M. (2005) ‘Conclusion’, in M. Gallagher and P. Mitchell (eds), The Politics of
Electoral Systems, Oxford: Oxford University Press.
Gallagher, M. and Mitchell, P. (eds) (2005) The Politics of Electoral Systems, Oxford:
Oxford University Press.
Gebethner, S. (1996) ‘Proportional Representation versus Majoritarian Systems: Free
Elections and Political Parties in Poland, 1989–1991’, in A. Lijphart and C.H. Waisman
(eds), Institutional Design in New Democracies: Eastern Europe and Latin America,
Boulder: Westview Press.
Geddes, B. (1996), ‘Initiation of New Democratic Institutions in Eastern Europe and
Latin America’, in A. Lijphart and C.H. Waisman (eds), Institutional Design in New
Democracies: Eastern Europe and Latin America, Boulder: Westview Press.
Grofman, B. and Lijphart, A. (eds) (1986), Electoral Laws and their Political Con-
sequences, New York: Agathon Press.
126   Bibliography
Grumm, J.G. (1958) ‘Theories of Electoral Systems’, Midwest Journal of Political
Science, 2: 357–76.
Hadenius, A. (1992) Democracy and Development, Cambridge: Cambridge University
Press.
Hartmann, C. (1999) ‘Benin’, in D. Nohlen, M. Krennerich and B. Thibaut (eds), Elec-
tions in Africa: A Data Handbook, Oxford: Oxford University Press.
Hillebrands, B. (1993) ‘Antigua und Barbuda’, in D. Nohlen (ed.), Handbuch der Wahl-
daten Lateinamerikas und der Karibik, Opladen: Leske & Budrich.
Horowitz, D.L. (1985) Ethnic Groups in Conflict, Berkeley: University of California Press.
Horowitz, D.L. (1991) A Democratic South Africa? Constitutional Engineering in a
Divided Society, Berkeley: University of California Press.
Huntington, S. (1991) The Third Wave: Democratization in the Late Twentieth Century,
Norman: University of Oklahoma Press.
Karvonen, L. (1981) Med vårt västra grannland som förebild: en undersökning av policy-
diffusion från Sverige till Finland, Åbo: Åbo Akademis förlag.
Karvonen, L. (2003) Statsskick: Att bygga demokrati, Stockholm: SNS förlag.
Katz, R.S. (1997) Democracy and Elections, Oxford: Oxford University Press.
Katzenstein, P.J. (1985) Small States in World Markets: Industrial Policy in Europe,
Ithaca: Cornell University Press.
Krennerich, M. (1993) ‘Nicaragua’, in D. Nohlen (ed.), Handbuch der Wahldaten Latein-
amerikas und der Karibik, Opladen: Leske & Budrich.
Krennerich, M. (1999) ‘Mauritius’, in D. Nohlen, M. Krennerich and B. Thibaut (eds),
Elections in Africa: A Data Handbook, Oxford: Oxford University Press.
Laakso, M. and Taagepera, R. (1979) ‘ “Effective” Number of Parties: A Measure with
Application to West Europe’, Comparative Political Studies, 12: 3–27.
Lakeman, E. and Lambert, J.D. (1946, Voting in Democracies: A Study of Majority and
Proportional Electoral Systems, London: Faber and Faber.
Laponce, J. and Saint-­Jacques, B. (1997) ‘Introduction: Institutions as Problem-­solvers’,
International Political Science Review, 18: 233–6.
León-Roesch, M. (1993) ‘Paraguay’, in D. Nohlen (ed.), Handbuch der Wahldaten
Lateinamerikas und der Karibik, Opladen: Leske & Budrich.
Lijphart, A. (1977) Democracy in Plural Societies: A Comparative Exploration, New
Haven: Yale University Press.
Lijphart, A. (1984) Democracies: Patterns of Majoritarian and Consensus Government
in Twenty-­one Countries, New Haven: Yale University Press.
Lijphart, A. (1985a) Power Sharing in South Africa, Berkeley: Institute of International
Studies.
Lijphart, A. (1985b) ‘The Field of Electoral System Research’, Electoral Studies, 1:
3–14.
Lijphart, A. (1991) ‘Constitutional Choices for New Democracies’, Journal of Demo-
cracy, 2: 72–84.
Lijphart, A. (1992a) ‘Democratization and Constitutional Choices in Czecho-­Slovakia,
Hungary and Poland 1989–91’, Journal of Theoretical Politics, 4: 207–23.
Lijphart, A. (1992b) ‘Introduction’, in A. Lijphart (ed.), Parliamentary versus Presiden-
tial Government, Oxford: Oxford University Press.
Lijphart, A. (ed.) (1992c) Parliamentary versus Presidential Government, Oxford:
Oxford University Press.
Lijphart, A. (1994) Electoral Systems and Party Systems: A Study of Twenty-­seven Demo-
cracies 1945–90, Oxford: Oxford University Press.
 Bibliography   127
Lijphart, A. (1999) Patterns of Democracy: Government Forms and Performance in
Thirty-­six Countries, New Haven: Yale University Press.
Lijphart, A. (2006) ‘The Case for Power Sharing’, in L. Diamond and M.F. Plattner (eds),
Electoral Systems and Democracy, Baltimore: Johns Hopkins University Press.
Lijphart, A. and Grofman, B. (eds) (1984a) Choosing an Electoral System: Issues and
Alternatives, New York: Praeger.
Lijphart, A. and Grofman, B. (1984b) ‘Introduction: Choosing an Electoral System’, in
A. Lijphart and B. Grofman (eds), Choosing an Electoral System: Issues and Altern-
atives, New York: Praeger.
Lijphart, A. and Waisman, C.H. (eds) (1996) Institutional Design in New Democracies:
Eastern Europe and Latin America, Boulder: Westview Press.
Lipset, S.M. (1992) ‘The Centrality of Political Culture’, in A. Lijphart (ed.), Parliament-
ary versus Presidential Government, Oxford: Oxford University Press.
Lipset, S.M. and Rokkan, S. (eds) (1967) Party Systems and Voter Alignments: Cross-­
national Perspectives, New York: The Free Press.
Mackie, T. and Rose, R. (1974) The International Almanac of Electoral History, London:
Macmillan.
Massicotte, L. and Blais, A. (1999) ‘Mixed Electoral Systems: A Conceptual and Empiri-
cal Survey’, Electoral Studies, 18: 341–66.
Mehler, A. (1999) ‘Cameroon’, in D. Nohlen, M. Krennerich and B. Thibaut (eds), Elec-
tions in Africa: A Data Handbook, Oxford: Oxford University Press.
Mozaffar, S. (1998) ‘Electoral Systems and Conflict Management in Africa: A Twenty-­
eight-state Comparison’, in T.D. Sisk and A. Reynolds (eds), Elections and Conflict
Management in Africa, Washington, DC: United States Institute of Peace Press.
Nohlen, D. (1984) ‘Two Incompatible Principles of Representation’, in A. Lijphart and B.
Grofman (eds), Choosing an Electoral System: Issues and Alternatives, New York:
Praeger.
Nohlen, D. (ed.) (1993a) Enciclopedia Electoral Latinoamericana y del Caribe, San José:
Instituto Interamericano de Derechos Humanos.
Nohlen, D. (ed.) (1993b) Handbuch der Wahldaten Lateinamerikas und der Karibik,
Opladen: Leske & Budrich.
Nohlen, D. (1996) ‘Electoral Systems and Electoral Reform in Latin America’, in A.
Lijphart and C.H. Waisman (eds), Institutional Design in New Democracies: Eastern
Europe and Latin America, Boulder: Westview Press.
Nohlen, D., Grotz, F. and Hartmann, C. (eds) (2001a) Elections in Asia and the Pacific: A
Data Handbook, vol. I, Middle East, Central Asia and South Asia, Oxford: Oxford
University Press.
Nohlen, D., Grotz, F. and Hartmann, C. (eds) (2001b) Elections in Asia and the Pacific: A
Data Handbook, vol. II, South East Asia, East Asia and the South Pacific, Oxford:
Oxford University Press.
Nohlen D., Krennerich, M. and Thibaut, B. (eds) (1999) Elections in Africa: A Data
Handbook, Oxford: Oxford University Press.
Norris, P. (2004) Electoral Engineering: Voting Rules and Political Behavior, Cam-
bridge: Cambridge University Press.
Ordeshook, P. and Shvetsova, O. (1994) ‘Ethnic Heterogeneity, District Magnitude, and
the Number of Parties’, American Journal of Political Science, 38: 100–23.
Pampel, F. (2000) Logistic Regression: A Primer, Iowa City: Sage Publications.
Parline Database. Online, available at: www.ipu.org/parline-­e/parlinesearch.asp.
Peters, G.B. (1998) Comparative Politics: Theory and Methods, Basingstoke: Macmillan.
128   Bibliography
Polyarchy Dataset. Online, available at: www.svt.ntnu.no/iss/data/vanhanen/.
Rae, D.W. (1967) The Political Consequences of Electoral Laws, New Haven: Yale Uni-
versity Press.
Rae, D.W. and Taylor, M. (1970) The Analysis of Political Cleavages, New Haven: Yale
University Press.
Reilly, B. (2001) Democracy in Divided Societies: Electoral Engineering for Conflict
Management, Cambridge: Cambridge University Press.
Reilly, B. (2006) ‘Dealing with Divided Societies’, in L. Diamond and M.F. Plattner
(eds), Electoral Systems and Democracy, Baltimore: Johns Hopkins University Press.
Reilly, B. and Reynolds, A. (1999) Electoral Systems and Conflict in Divided Societies,
Washington, DC: National Academy Press.
Reynolds, A. (1999) Electoral Systems and Democratization in Southern Africa, Oxford:
Oxford University Press.
Reynolds, A. and Elklit, J. (1997) ‘Electoral System Design in the Arab World’, in A.
Reynolds, and B. Reilly (eds), The International IDEA Handbook of Electoral System
Design, Stockholm: International Institute for Democracy and Electoral Assistance.
Reynolds, A. and Reilly, B. (eds) (1997) The International IDEA Handbook of Electoral
System Design, Stockholm: International Institute for Democracy and Electoral Assist-
ance.
Reynolds, A. and Sisk, T.D. (1998) ‘Elections and Electoral Systems: Implications for
Conflict Management’, in T.D. Sisk and A. Reynolds (eds), Elections and Conflict
Management in Africa, Washington, DC: United States Institute of Peace Press.
Reynolds, A., Reilly, B. and Ellis, A. (eds) (2005) Electoral System Design: The New
International IDEA Handbook, Stockholm: International Institute for Democracy and
Electoral Assistance.
Rogowski, R. (1987) ‘Trade and Democratic Institutions’, International Organization,
41: 203–25.
Rokkan, S. (1970) Citizens, Elections, Parties, New York: David McCay.
Roskin, M.G. (1986) Countries and Concepts, Englewood Cliffs: Prentice-­Hall.
Sartori, G. (1968) ‘Political Development and Political Engineering’, in J.D. Montgomery
and A.O. Hirschman (eds), Public Policy, 17, Cambridge: Harvard University Press.
Sartori, G. (1994) Comparative Constitutional Engineering: An Inquiry into Structures,
Incentives and Outcomes, Basingstoke: Macmillan.
Shugart, M.S. (1988) Duverger’s Rule, District Magnitude, and Presidentialism, doctoral
dissertation, Irvine: University of California.
Shugart, M.S. (2005) ‘Comparative Electoral Systems Research: The Maturation of a
Field and New Challenges Ahead’, in M. Gallagher and P. Mitchell (eds), The Politics
of Electoral Systems, Oxford: Oxford University Press.
Shugart, M.S. and Carey, J.M. (1992) Presidents and Assemblies: Constitutional Design
and Electoral Dynamics, New York: Cambridge University Press.
Shugart, M.S. and Wattenberg, M.P. (eds) (2001) Mixed-­member Electoral Systems: The
Best of Both Worlds?, Oxford: Oxford University Press.
Sisk, T.D. (1995) Democratization in South Africa: The Elusive Social Contract, Prince-
ton: Princeton University Press.
Sisk, T.D. and Reynolds, A. (eds) (1998) Elections and Conflict Management in Africa,
Washington, DC: United States Institute of Peace Press.
Svensk uppslagsbok, (1947–1955), Malmö: Norden.
Taagepera, R. and Shugart, M.S. (1989) Seats and Votes: The Effects and Determinants
of Electoral Systems, New Haven: Yale University Press.
 Bibliography   129
The World Factbook. Online, available at: https://www.cia.gov/library/publications/the-­
world-factbook/.
The World in Figures (1981; 1987), New York: Facts on File.
Vanhanen, T. (1984) The Emergence of Democracy: A Comparative Study of 119 States,
1850–1979, Helsinki: Finnish Society of Sciences and Letters: Academic Bookstore.
Vanhanen, T. (1990) The Process of Democratization: A Comparative Study of 147
States, 1980–88, New York: Crane Russak.
Verney, D. (1959) The Analysis of Political Systems, London: Routledge and Kegan Paul.
Weiner, M. (1987) ‘Empirical Democratic Theory’, in M. Weiner and E. Ozbundun (eds),
Competitive Elections in Developing Countries, Durham: Duke University Press.
Weyland, K. (2005) ‘Theories of Policy Diffusion: Lessons from Latin American Pension
Reform’, World Politics, 57: 262–95.
Index

absolutist regimes 91 Belize 59, 97

accidental evolution 11 Bhutan 28
actor-related determinants 33–4, 45–7 bicameral regimes 95–7
administrative abilities, development of Birch, S. 11, 12, 14–16, 110
62–5 Blais, André 6, 12, 13, 14, 17, 21, 22–3,
Afghanistan 26–7, 29 24, 27, 41, 42, 50
Africa 13, 63, 64, 72–3, 76–7, 87, 115 block vote (BV) 18–19; adoption of 26–32
Albania 28 Bogdanor, V. 23, 48
Albertini, Rudolf von 63 Boix, Carles 4, 46
alternative vote (AV) 20; adoption of Bolivia 28, 78, 93
26–32; in divided societies 35–7, 38–40 book objectives/scope 1–3
America 72–3 Bosnia-Herzegovina 67, 95
Anckar, C. 41, 50, 91 Botswana 92
Anckar, Dag 16, 41, 43, 61, 72–3, 107 Brady, David 45–6
Andorra 27, 34 Brazil 30, 80, 95
Anglo-American societies 74 Britain see UK
Angola 67 Britannica Book of the Year (1999) 6
‘anti-socialist phase’, PR 34, 53–4 British Commonwealth 85
Antigua 7, 10 Brunei 28–9
Argentina 51 Bulgaria 9, 73, 78
Armenia 72 Burma 26, 29, 62–3, 66, 80
Asia 13, 59, 63, 72–3, 80, 87 Burundi 67
asymmetrical bicameralism 96–7
Australia 32, 33, 74; colonial legacy 62–9, Cambodia 63, 67, 80
81 Cameroon 28, 66, 67
Austria 70 Canada 59, 74
authoritarian regimes 91–2 Cape Verde 51, 67
Azerbaijan 27, 72 Carey, J.M. 89, 91
Caribbean 59, 63, 72–3
Bahamas 97 Carstairs, Andrew 54, 74
Bahrain 27, 66 Caucasus 72–3, 115
Barbados 19, 84, 89 Central Africa 72–3
Barbuda 7, 10 Central America 69, 72–3
Barkan, Joel D. 107 Central Asia 72–3, 115
‘behavioural–institutional equilibrium’ centripetalist approach 36, 37
48–9 Chad 28
Belarus 67 chamber structure 80–6, 95–7, 104, 105,
Belgian Congo 64 108
Belgium 34, 70, 89; colonial legacy 62–9, Chile 64
81 China 28
 Index   131
Chronicle of Parliamentary Elections 83–6; frequency of electoral system
(Inter-Parliamentary Union) 6 choice 77–9; frequency of PR systems
Citizens, Elections, Parties: Approaches to 14–15; multivariate patterns of diffusion
the Comparative Study of the Processes 106–9; party system fragmentation 51–3
of Dsevelopment (Rokkan) 46 democratisation: process of 54–7; waves
civil liberties 8 of 74–5
coexistence systems 22; adoption of 26–32 Denmark 46, 94
Colomer, Josep M. 4, 24, 48–9 dependent variable 5–6, 12–13
colonial diffusion 62–9, 80–6; multivariate Derbyshire, J.D. & I.D. 89, 91, 94, 96
patterns 100–9; patterns 86–7; summary Diamond, Larry 10, 59–60, 61–2, 63–4, 65
113–14 diffusion: colonial diffusion 62–9;
colonial heritage 11, 12–13 multivariate patterns of 100–9; patterns
Commonwealth of Independent States 62 of 86–7; regional diffusion 69–73;
Communist regimes 54, 91 temporal diffusion 73–86
Comoros 52, 67, 95 direct rule 62–4
competition, degree of 9–10 divided societies 34–40
competitive elections 55–7 Dobrzynska, Agnieszka 15–16, 24, 50
conditional systems 22; adoption of 26–32 Dominica 97, 107
Congo 67, 95 Droop-quota 20–1
consociational democracy 35–6, 37 Duverger, Maurice 1, 47–8
constitutional design 16, 35; definition of
32; summary 114 East Africa 72–3
contextual factors 11–12 East Asia 72–3
continents 13, 72–3 Eastern Europe 46, 54, 72–3
corrective system 22; adoption of 26–32 economic development 12–14
correlates, temporal diffusion 79–86 Ecuador 28
Costa Rica 64 Egypt 27, 80
country size 41–5, 80–6; summary 102, El Salvador 51
103, 104, 105, 111–12 Elazar, D.J. 93
country, electoral system by 116–21 Elections in Africa (Nohlen et al.) 6
Cox, Gary W. 48 Elections in Asia and the Pacific (Nohlen
Croatia 27, 72 et al.) 6, 72
Cuba 19, 26 Elections Worldwide series (Nohlen et al.)
cultural context 16 10
cultural diversity 34–40, 80–6, 102, 103, electoral formulas: definition of 17–18;
104, 105 presentation of 18–23
cultural perspective: overview 3; summary electoral laws: definition of 17–18;
112–13; see also diffusion knowledge of 69
culturally different countries 71 electoral system choice: actor-related
culturally similar countries 69–73 determinants 45–7; by country 116–21;
cumulative vote (CV) 19 and country size 42–5; definition of 26;
Cyprus 44, 52, 66, 78, 84, 85, 93, 111 and diversity 38–40; importance of 3–4;
Czech Republic 97 introductory survey of 26–31; and party
Czechoslovakia 54, 95 system structure 51–3; and party system
transformation 56–7; structurally
Dahl, R.A. 41 generated determinants of 34; timing/
data sources 6 frequency of 7, 75–9
deductive approach 5 electoral system design: accidental/
Democracies (Lijphart) 89 deliberate 32–3, 34–40; definition of
democracy: level of 12–13, 14; 32
measurement of level of 8–10 Electoral Systems and Party Systems
democratic citizenship, training in 64–5 (Lijphart) 24
democratic countries 7–10; effects of Electoral Systems: A Comparative
country size 44; electoral system choice Iintroduction (Farrell) 24
132   Index
electoral systems: classification of 23–6; government form 80–6, 90–3, 108
definition of 17–18; future of 114–15; Greece 27, 52, 53, 78, 83, 84, 111
permanency of 110–11; presentation of Greenland 94
18–23; stickiness of 110–11; typology Grenada 97
of 25–6 Grofman, Bernard 47, 110
Elklit, J. 33 Grotz, F. 72
Ellis, A. 11 Grumm, John 47–8
empirical evidence: chamber structure 97; Guatemala 27
colonial diffusion 66–9; country size Guinea 27
42–5; cultural/ethnic diversity 38–40; Guinea-Bissau 65, 67
government form 92–3; party system Guyana 44, 67, 84, 98
structure 51–3; party system
transformation 56–7; regional diffusion Handbuch der Wahldaten Lateinamerikas
72–3; temporal diffusion 75–9; und der Karibik (Nohlen) 6, 10
territorial organization 94–5 Hartmann, C. 72
Encyclopedia of Democracy, The 93 hierarchical diffusion 71
Equatorial Guinea 65, 95 Hillebrands, B. 7
Eritrea 28 historical context 16
Estonia 86, 92 historical factors 11–12; overview 3;
ethnic diversity 34–40, 80–6 summary 112–13; see also diffusion
ethnicity, definition of 37 Horowitz, D.L. 35, 37
Europe: electoral systems 13, 14, 46, 59, Hungary 28, 54, 72
69–70; post-Communist 11–12; regional Huntington, Samuel 16, 69, 74
diffusion 72–3, 113
external imposition 11, 70, 74 Iceland 34, 44, 78, 84, 111
independence: democracy level prior to 10;
Faroe Islands 94 timing of 64–5, 75, 80–6
Farrell, David M. 17, 23, 24, 74 independence process 55
federal regimes 35, 93–5 independent countries 6–7
Federated States of Micronesia 39, 67, 85, index of democracy (ID) 9–10
95, 96, 107 India 27, 66, 80, 95
Fiji 26, 31, 66, 84, 107 indirect rule 62–4
Finland 34 Indonesia 43, 80; colonial legacy 62–9
Flora, P. 65 Indridason, Indridi H. 15–16, 24, 50
foreign influences 11–12 inductive approach 5
frame of reference 16 inherited electoral systems 66–9, 80–6
France: colonial legacy 62–9, 80, 81, 83, institutional perspective: chamber structure
86, 100–9; electoral systems 27, 30, 34, 95–7; government form 90–3; overview
43, 44, 78, 111; presidentialism 92, 98; 1–2; summary 98, 104; territorial
regional autonomy 22, 94; temporal organization 93–5
diffusion 31 Inter-Parliamentary Union 6
Freedom House 8, 9–10, 12 interest-based calculations 11–12
Freedom in the World 8 International Almanac of Electoral
fusion system 22; adoption of 26–32 History, The (Mackie/Rose) 6
International IDEA Handbook of Electoral
Gabon 68, 84, 106 System Design, The (Reynolds/Reilly)
Gallagher, Michael 45, 48 24
‘Galton’s problem’ 59 Iraq 80
Gambia 39 Israel 67, 77, 80, 83, 84
Gebethner, Stanislaw 48 Italy 27, 28, 41, 44, 78, 80–6, 94, 111;
Geddes, B. 110 colonial legacy 62–9, 81
geographic proximity 69–73 Ivory Coast 27
Georgia 71, 72
Germany 25, 41; colonial legacy 62–9 Jamaica 84, 97, 107
 Index   133
Japan 26, 44, 70, 85, 97; colonial legacy methodological considerations 5–6
62–9 Mexico 28, 95
Jordan 29, 33, 66, 80 Middle East 72–3
Military authoritarian regimes 91
Karvonen, Lauri 16, 61, 71, 73, 90, 92, Millard, F. 11, 12, 14–16, 110
93–4, 96 ‘minority protection phase’, PR 34, 53–4
Katz, Richard 23–4 Mixed Electoral Systems: A Conceptual
Katzenstein, Peter J. 41–2, 82, 107 and Empirical Survey (Massicotte/
Kazakhstan 72–3 Blais) 6
Kenya 95 mixed electoral systems: adoption of
Kiribati 30, 66, 92 26–32; and chamber structure 97;
Kraus, F. 65 diffusion of 77; in divided societies 40;
Krennerich, M. 6, 10, 27, 65 former colonies adoption of 65–8;
Kuwait 66 frequency of 77, 79–86; and government
Kyrgyzstan 67, 72–3 form 92–3; overview 21–3; and party
system transformation 56–7; regional
Laakso, Markku 50 diffusion 72–3; and territorial
Lakeman, Enid 23 organization 94–5
Lambert, James D. 23 mixed federal/unitary regimes 93–5
Laos 63, 67, 80 Mixed-member Electoral Systems: The
Laponce, Jean 32 Best of Both Worlds (Shugart/
Latin America 46, 60, 64, 69, 113 Wattenberg) 21–2
Latvia 86 mixed-member majoritarian (MMM)
Lebanon 78, 92 system 21–3
León-Roesch, M. 27 mixed-member proportional (MMP)
Libya 29, 67, 80, 95 system 21–3
Lijphart, Arend 4, 5, 23, 24, 35–6, 47, 49, Mo, Jongryn 45–6
50, 54, 60, 69–70, 88–9, 90–1, 94, 110 Monaco 28, 34, 44, 78
limited vote (LV) 19 Mongolia 80
Lipset, Seymour Martin 1, 10, 47, 60, Montenegro 44
63–4, 65 Morocco 51, 67
list proportional representation (PR) 20; Mozaffar, S. 45
adoption of 26–32; in divided societies Mozambique 67
34, 38–40 multi-party competition 7–8
Lithuania 72, 86, 92 multi-party systems 47, 48, 54–7
multivariate patterns 99–109
Macedonia 73
Mackie, T. 6 Namibia 67, 70, 92
Madagascar 27, 28, 44, 78, 92 nationalistic socialist regimes 91
majority systems: adoption of 26–32; and Nauru 26, 31, 44, 67
chamber structure 97; and country size Netherlands 96; colonial legacy 62–9
42–5; in divided societies 39–40; former Nevis 7, 97
colonies adoption of 65–8; frequency of New Zealand 28, 52–3, 74, 78, 89; colonial
78–86; and government form 92–3; legacy 62–9, 81
overview 15–16, 19–20; subtypes of 24; Nicaragua 27
and territorial organization 94–5 Nigeria 95
Malaysia 84, 95, 107 Nohlen, Dieter 6, 10, 48, 49, 65, 72, 110
Maldives 27, 66 non-democratic countries 7–10
Mali 30 non-independent countries 7
Malta 26, 44, 66–7, 84 Norris, Pippa 88
Marshall Islands 67, 85 North Africa 72–3
Massicotte, Louis 6, 12, 13, 14, 21, 22–3, North America 13, 59, 72–3
27, 41, 42 North Korea 26, 80
Mauritius 10, 63, 66 Norway 96
134   Index
Oceania 13 president-parliamentary regimes 91
Oman 28 presidential regimes 90–3
one-ballot majority (OBM) 19; adoption of prestigious countries 69, 72
26–32 Principe 67
one-party regimes 54–7, 91–2 proportional representation (PR) systems:
Ordeshook, Peter 49 adoption of 26–32; allocation of seats
Organization for Economic Co-operation 23–4; and chamber structure 97; and
and Development (OECD) 13–14 country size 42–5; degree of
proportionality 23–5; in divided
Pacfic 59 societies 34, 36–7, 38–40; former
Pakistan 95 colonies adoption of 65–8; frequency of
Palau 27, 67, 85, 95, 107 77, 79–86; and government form 92–3;
Panama 27 overview 14–16, 20–1; and party system
Papua New Guinea 26, 33, 64, 67, 78, 85 fragmentation 51–3; and party system
Paraguay 27 transformation 56–7; subtypes of 24;
parliamentary regimes 90–3 and territorial organization 94–5
Parliamentary versus Presidential
Government (Lijphart) 59 Qatar 28, 64
participation, degree of 9–10
Partis Politiques, Les (Duverger) 47–8 Rae, Douglas W. 38, 47, 50
party numbers 49–50, 51–3; majority rational perspective: actor-related
systems 15 determinants 45–7; country size 41–5;
party system structure 47–53, 80–6; ethnic and cultural diversity 34–40;
summary 112 overview 1–2; party system structure
party system transformation 53–7; 47–53; party system transformation
summary 112, 114 53–7; structurally generated
Party Systems and Voter Alignments: determinants 34; summary 57–8, 111
Cross-national Perspectives (Lipset/ regional autonomy 35
Rokkan) 1 regional diffusion 69–73, 80–6;
Patterns of Democracy (Lijphart) 24, 89 multivariate patterns of 100–9; patterns
Peters, G.B. 59 of 86–7; summary 113–14
Philippines: colonial legacy 64, 67, 84, Reilly, Ben 11, 24, 32, 35, 36, 37, 62, 70,
107; election systems 39, 44, 52, 93, 94; 74, 87
temporal diffusion 77, 78, 85 religious fragmentation 37
plurality systems 18–19; adoption of research design: frame of reference 16;
26–32; and chamber structure 97; and methodological considerations 5–6;
country size 42–5; former colonies previous theoretical explanations 11–16
adoption of 65–8; frequency of 78–86; research population 6–10
and government form 92–3; party research, summary of findings 111–14
system fragmentation 51–3; and Reynolds, Andrew 11, 24, 32, 33, 36, 62,
territorial organization 94–5 70, 74, 87
Poland 9, 48, 54 Rogowski, Ronald 13–14, 41–2, 82, 107
political abilities, development of 62–5 Rokkan, Stein 1, 12, 32, 34, 41–2, 43, 47,
political institutions, as problem-solvers 16 54, 82, 84, 107
political rights 8–10 Rose, R. 6
Political Systems of the World (Derbyshire/ Roskin, M.G. 88
Derbyshire) 94 Rothenbacher, F. 65
Polyarchy Dataset 9–10 Russia 71, 72–3, 75, 95; colonial legacy
Popescu, M. 11, 12, 14–16, 110 62–9
population size see country size Rwanda 67
Portugal 34, 74; colonial legacy 62–9
post-Communist Europe 11–12 Saint-Jacques, Bernard 32
‘preference-swapping’ 36–7 Saint Kitts 7, 97
premier-presidential regimes 91 Saint Lucia 7, 97
 Index   135
Sao Tome 67 Taylor, Michael J. 38
Sartori, G. 32, 91, 115 temporal diffusion 73–86; diffusion
Saudi Arabia 28 patterns 86–7; summary 113–14
Scandinavian countries 74, 93 territorial organization 80–6, 93–5, 104,
semi-presidential regimes 91–3 105, 108
semi-proportional systems 23–4 Thailand 27, 44, 52, 78, 85, 97
Senegal 27 theoretical explanations, previous attempts
Seychelles 44, 67, 79, 92 at 11–16
Shugart, Matthew S. 4, 22, 47, 89, 91, 110 Thibaut, B. 6, 10, 65
Shvetsova, Olga 49 Togo 67
single-member plurality (SMP) 18–19; Trade and Democratic Institutions
adoption of 26–32 (Rogowski) 13–14, 46
single non-transferable vote (SNTV) 19; trade dependent states 13–14
adoption of 26–32 Trinidad and Tobago 84, 97, 107
single transferable vote (STV) 20–1; Tufte, E.R. 41
adoption of 26–32; in divided societies Tunisia 28
38–40 Turkey 52, 78, 97, 111
Sisk, Timothy D. 36 Turkmenistan 67
social fragmentation 47–8 Tuvalu 44, 66, 97
sociological approach 1–2 two-ballot majority (TBM) 19–20;
Solomon Islands 97, 107 adoption of 26–32
Somali Republic 66 two-party systems 47, 48, 49–50
Somalia 29, 67
South Africa 70; colonial legacy 62–9 UK 12–13, 34, 46, 74, 89, 93; colonial
South America 13, 59, 69–70, 71, 98 legacy 62–9, 80, 81, 83, 84, 86, 97,
South Asia 72–3 100–9, 111–13
South Korea 27, 45–6, 70 Ukraine 67, 73
South Pacific 72–3, 87, 115 unicameral regimes 95–7
South Vietnam 67 unitary regimes 93–5
Southern Africa 72–3 United Arab Emirates 28, 64
Southern East Asia 72–3 United Nations (UN) 70
Soviet Union see Russia universal suffrage 14, 34, 41, 54–5
Spain 34; colonial legacy 62–9 Uruguay 64
spatial diffusion 71 US 71, 74, 96; colonial legacy 62–9, 84,
Sri Lanka 52, 77, 78, 80, 84, 85, 92, 111 85, 92
Statsskick: att bygga demokrati Uzbekistan 67
(Karvonen) 6
strategic voting 15 Vanhanen, Tatu 9–10
structurally generated determinants 33–4 Vanuatu 26, 29, 66, 67, 68, 106
supermixed systems, adoption of 26–32 Vatican City 28
superposition mixed system 22; adoption Venezuela 28, 44, 78, 80, 93, 95
of 26–32 Verney, D. 91
Suriname 27, 44, 67, 79, 92 Vietnam 80
Svensk uppslagsbok 6 ‘vote-pooling’ 36–7
Swaziland 66 vote-seat relationship 23–4
Sweden 34, 46 Voting in Democracies (Lakeman/
Switzerland 34, 89, 92 Lambert) 24
symmetrical bicameralism 96–7
Syria 67, 80 Wattenberg, M.P. 22
Weiner, Myrom 63
Taagepera, R. 47, 50, 110 West Africa 72–3
Taiwan 27 West Germany 28, 70, 80, 95
Tajikistan 67, 73 Western Europe 34, 48, 53–4, 70, 72–4
Tanzania 39 Western Samoa 27
136   Index
Weyland, Kurt 60–1 Yugoslavia 73, 95; colonial legacy 62–9
Williams, K. 11, 12, 14–16, 110
World Factbook (CIA) 6 Zambia 39
World in Figures 6 Zimbabwe 27, 67
eBooks – at www.eBookstore.tandf.co.uk

A library at your fingertips!

eBooks are electronic versions of printed books. You can

store them on your PC/laptop or browse them online.
They have advantages for anyone needing rapid access
to a wide variety of published, copyright information.
eBooks can help your research by enabling you to
bookmark chapters, annotate text and use instant searches
to find specific words or phrases. Several eBook files would
fit on even a small laptop or PDA.
NEW: Save money by eSubscribing: cheap, online access
to any eBook for as long as you need it.

Annual subscription packages

We now offer special low-cost bulk subscriptions to
packages of eBooks in certain subject areas. These are
available to libraries or to individuals.
For more information please contact
webmaster.ebooks@tandf.co.uk
We’re continually developing the eBook concept, so
keep up to date by visiting the website.

www.eBookstore.tandf.co.uk