Farrell

U NS W P R E S S
UNSW
PRESS
The Australian Electoral System has
many strengths, chief among them
being the authors ability to combine
detailed analysis of Australias varied
electoral experience with the broader
issues raised by the international
literature on elections and electoral
behaviour. This publication makes a
major contribution to the study of
electoral politics and to Australian
political science.
G. CAMPBELL SHARMAN, DEPARTMENT OF
POLI TI CAL SCI ENCE, UNI VERSI TY OF BRI TI SH
COLUMBI A, VANCOUVER
The Australian Electoral System is an
outstanding exploration of the range
of electoral systems employed in
Australia. It benef its greatly from being
informed by the broader comparative
electoral systems literature and it is a
book that all those interested in electoral
systems generally will want to read.
MI CHAEL GALLAGHER, DEPARTMENT OF
POLI TI CAL SCI ENCE, TRI NI TY COLLEGE, DUBLI N
A key aim of this clearly written
and polemic-free book is to place
the origins, evolution and operation
of preferential election methods in
Australia in comparative context
This volume should be of great
interest to all students of elections and
representation, and to students of
comparative politics more generally
It will have special appeal to those
interested in electoral reform in
Australia.
BERNARD GROFMAN, PROFESSOR OF POLI TI CAL
SCI ENCE, UNI VERSI TY OF CALI FORNI A, I RVI NE
The Austral i an
El ectoral
System
or i gi ns,
var i at i ons
and
consequences
DAV I D M. F A R R E L L I A N Mc A L L I S T E R
FARRELL
McALLISTER
T
h
e

A
u
s
t
r
a
l
i
a
n

E
l
e
c
t
o
r
a
l

S
y
s
t
e
m
AustElecSystrover copy.indd 1 11/10/05 11:11:20 AM
THE
AUSTRALIAN
ELECTORAL
SYSTEM
David M. Farrell is Professor and Head of Politics at the University of
Manchester. He is author of Electoral Systems(Palgrave, 2001) and co-editor
of several volumes, most recently Political Parties in Advanced Industrial
Democracies(Oxford University Press, 2002). He is completing a study (with
Roger Scully) on Representing Europes Citizens? Electoral Institutions and the
Failureof Parliamentary Representation(to be published by Oxford University
Press).
I an McAllister is Professor of Political Science at The Australian National
University, and was previously Professor of Government at the University of
Manchester. He is co-editor of TheCambridgeHandbook of theSocial Sciences
in Australia (Cambridge University Press, 2003), with Steve Dowrick and
Riaz Hassan, and is co-author, with David Lovell, Chandran Kukathas and
William Maley, of TheAustralian Political System(Longman, 1998). He has
written extensively on Australian politics, parties and elections, and has been
a co-director of the Australian Election Study since 1987. He is currently
Chair of the 50-nation Comparative Study of Electoral Systems group.
ANZSOG Program on Government,Politics and Public Management
The Australia and New Zealand School of Government (ANZSOG) is a
network initiative of five jurisdictions (the Australian and New Zealand gov-
ernments, New South Wales, Victoria and Queensland) and nine universities.
Established in 2003, ANZSOG represents a new and exciting prospect for
the development of world-class research and teaching in the public and com-
munity sectors.
ANZSOG has announced an extensive research program that promotes
innovative and cutting-edge research in partnership with academia and the
public sector (<http://www.anzsog.edu.au>). I n association with UNSW
Press, ANZSOG has undertaken to publish a series of books on contempo-
rary issues in Australian government, politics and public management. Titles
in this program will promote high-quality research on topics of interest to a
broad readership (academic, professional, students and general readers) and
will include teaching texts relevant to the ANZSOG consortia in the areas of
government, politics and public management.
Series editors are Professor John Wanna and Professor R.A.W. Rhodes,
Research School of Social Sciences, Australian National University,
Canberra.
Recent titles include:
Terms of Trust: Arguments over ethics in Australian government by John Uhr
Yes, Premier: Labor Leadership in Australias States and Territories, edited by
John Wanna and Paul Williams
Westminster Legacies:Democracy and ResponsibleGovernment in Asia and the
Pacific, edited by Haig Patapan, John Wanna and Patrick Weller.
THE
AUSTRALIAN
ELECTORAL
SYSTEM
ORIGINS, VARIATIONS AND CONSEQUENCES
David M. Farrell and Ian McAllister
UNSW
PRESS
A UNSW Press book
Published by
University of New South Wales Press Ltd
University of New South Wales
Sydney NSW 2052
AUSTRALI A
www.unswpress.com.au
David M. Farrell and I an McAllister 2006
First published 2006
This book is copyright. Apart from any fair dealing for the purpose
of private study, research, criticism or review, as permitted under the
Copyright Act, no part may be reproduced by any process without
written permission. I nquiries should be addressed to the publisher.
National Library of Australia
Cataloguing-in-Publication entry
Farrell, David M., 1960
Australian electoral system: origins, variations and
consequences.
Bibliography.
I ncludes index.
I SBN 0 86840 858 1.
1. Elections - Australia - History. 2. Voting - Australia -
History. 3. Representative government and representation -
Australia - History. 4. Election law - Australia. I .
McAllister, I an, 1950 . I I . Title. (Series: ANZSOG
series).
324.60994
Design Ruth Pidd
Cover Di Quick, based on photo courtesy of Australian Electoral Commission
Print BPA
| v |
Contents
Acknowledgments vii
Abbreviations ix
One Australian electoral system design 1
Australian electoral systems: origins and variation 3
Studying the operation of Australias electoral systems 7
Locating preferential electoral systems in Australia 13
Assessing Australias electoral systems 17
Two Australias electoral systems,190283 21
Early debates over electoral system design 23
The 1902 Commonwealth Electoral Act 29
The adoption of AV for the House of Representatives 36
Reform of the Senate electoral system, 191883 40
Conclusion: accident and intent in electoral system design 45
ThreeAustralias contemporary electoral systems 47
Australias evolving electoral systems 49
The alternative vote and its variants 52
The single transferable vote and its variants 57
Australias experiments with alternatives to AV and STV 68
Conclusion 76
Four Preferential systems and their consequences 78
AV as a non-proportional system 80
STV as a proportional system 83
Preferential systems and the fate of individual candidates 91
Conclusion 100
Five The voter versus the party 102
Party service among election candidates 106
Candidates views of their representative roles 110
Campaign activity by Australian parliamentary candidates 113
Conclusion 119
Six Compulsory voting,party control and the voter 121
Compulsory voting in Australia 122
Preferential voting and the strategic voter 126
Ticket voting and the preferential voter 133
Compulsion, complexity and voter engagement 138
Conclusion 145
Seven Preferential systems and voter satisfaction 147
Ballot structure and democracy 149
Preferential systems and voter satisfaction with democracy 155
Electoral systems and political attitudes 158
Conclusion 164
Eight Preferential voting and electoral system design 166
The politics of electoral system design 167
Variations in electoral system families 169
Electoral and behavioural consequences 171
From innovation to stagnation: electoral system design 174
Where to now? 178
Appendix 180
Notes 184
Bibliography 195
I ndex 207
| vi |
A ck n o wl ed g m en ts
I n researching for and writing this book we have accumulated a very long list
of people whose help should be acknowledged. I n particular, we have relied
heavily on expert advice and assistance from the various electoral offices and
parliamentary libraries across Australia, and we wish to express our gratitude
to the following individuals who assisted us in various ways, either by pro-
viding access to materials, or in responding to our requests (in some instances,
regular requests) for information and guidance: Scott Bennett, Gerry
Newman and Sarah Miskin, Commonwealth Parliament Library; Rosa
Ferranda, Senate Table Office; Michael Maley, Andy Christian, Brian Hallett
and Roger Wills, Australian Electoral Commission; Fay dElmaine,
Australian Electoral Commission librarian; Bogey Musidlak, President of the
Proportional Representation Society of Australia; Guy Dickson, Research
Officer, Parliamentary Library of South Australia; Kirsten Robinson and
Justin Harbord, Western Australian Electoral Commission; T. Jessop, State
Electoral Office, New South Wales; Jenni Newton-Farrelly, Parliamentary
Library of South Australia; Vena Boman and Terry Newman, Parliamentary
Research Service, Tasmania; Gareth Griffith, Parliament Library New South
Wales. Other colleagues we also wish to single out for thanks include: David
Black, Alastair Fischer, Michael Gallagher, Murray Goot, Wolfgang Hirczy-
de-Mino, Colin Hughes, John Lane, Malcolm Mackerras, Narelle
Miragliotta, Jeremy Moon, Joan Rydon, Marian Sawer, Marian Simms,
Bruce Stone, John Uhr, and Paul Wilder.
| vi i |
We are particularly grateful to the following individuals who read and
commented insightfully on some of our draft chapters: Scott Bennett, Vena
Boman, Shaun Bowler, Antony Green, Gareth Griffith, Terry Newman,
Jenni Newton-Farrelly, and Campbell Sharman. I n this context, we also wish
to acknowledge the constructive and detailed comments received from our
two readers, from the series editor, John Wanna, and from our publisher, John
Elliot. We are also grateful to our copy editor, David Brennan, for his effi-
cient and courteous treatment of this manuscript through the production
stage. I n addition, we benefited greatly from feedback from academic col-
leagues at the various forums in which we presented papers based on this
research, most notably at the meetings of the American Political Science
Association, the Elections, Public Opinion and Parties Specialist Group of
the Political Studies Association, the Irish Political Studies Association, and
the Australasian Political Studies Association.
We are also grateful to the staff of the following libraries: John Rylands
Library at the University of Manchester; Lakeman Library for Electoral
Research at the McDougall Trust; Library of the I nstitute of Commonwealth
Studies, London; the libraries of the Australian National University; the
Australian National Library; the Library of the Australian Electoral
Commission; the Commonwealth Parliament Library.
The study benefited from research grants facilitating two research trips to
Australia by David Farrell: in 1994, an Australian Research Council Visiting
Fellowship at the Australian Defence Force Academy, Canberra, and in 2000,
a visiting fellowship at the Research School of Social Sciences, Australian
National University. I an McAllister is grateful to the I nstitute of Governance,
Public Policy and Social Research at Queens University Belfast for two vis-
iting fellowships which facilitated research on the book. We also acknowl-
edge the support of the Australian Research Council for funding the
Australian Election Studies, conducted at each federal election since 1987,
which are used extensively throughout the book, and we express our thanks
to the Comparative Study of Electoral Systems (CSES, based at the Center
for Political Studies, University of Michigan, Ann Arbor) for release of their
data set which is used in Chapter 7.
I t goes without saying that we alone are responsible for any remaining
errors or infelicities.
The final vote of thanks is to our wives, Arlene and Toni, who have put
up with our long periods of absence during our respective overseas visits.
| T he Austral i an El ectoral System |
| vi i i |
A b b rev i a ti o n s
ACS Australian Candidate Study
ACT Australian Capital Territory
AEC Australian Electoral Commission
ALP Australian Labor Party
AV alternative vote
CPD Commonwealth Parliamentary Debates
CSES Comparative Study of Electoral Systems
DLP Democratic Labor Party
IDEA I nstitute for Democracy and Electoral Assistance
JSCEM Joint Select Committee on Electoral Matters
JSCER Joint Select Committee on Electoral Reform
LCL Liberal and Country League
LR Largest Remainder (list system)
MM mixed member
MMP mixed member proportional
MP Member of Parliament
PR proportional representation
SAPD South Australian Parliamentary Debates
SMP single member plurality
SNTV single non-transferable vote
SSD single-seat district
STV single transferable vote
VAP voting age population
| i x |
o n e
A u stra l i a n el ecto ra l
sy stem d esi g n
I n the pantheon of representative democracy, Australia has its name stamped
on many of the major advances in electoral system design as well as on the
steps towards democratising electoral laws. Many of its contributions are well
known and have already been closely documented (most recently, see Sawer
2001). I t was a trailblazer in terms of the opening up of the electoral process;
half a century ago Louise Overacker (1952: 15) commented that no modern
democracy has shown greater readiness to experiment with various electoral
methods than Australia. As early as 1859 the bulk of the Australian colonies
had established systems of parliamentary government with adult male suf-
frage. I n 1894, South Australia was second only to New Zealand in extending
voting rights to women for its lower house elections. The secret ballot (known
to this day as the Australian ballot ) was also an Australian invention, first
used in South Australia and Victoria in the mid-1850s. By the criterion of
adult suffrage, Australia was the first truly democratic state, achieving that
status in 1903, with most of the other contemporary liberal democracies fol-
lowing suit in the years after World War I (Aitkin and Castles 1989: 208).
Australia is also the home of a particular set of electoral systems that are
generally known by the generic title of preferential systems. The character-
istic feature of these systems is that they permit a greater degree of flexibility
in the voting act. I nstead of voters expressing a categorical vote for one can-
didate or party list, under preferential systems, voters are to varying degrees
able to vote ordinally, by rank-ordering candidates on the ballot paper, and
| 1 |
can therefore express their vote across several or more candidates. Above all
else, ballot structure the rules regarding how a voter votes is what distin-
guishes these systems from other electoral systems.
Preferential systems occur in various forms, of which the most prominent
include cumulative vote, limited vote, alternative vote (AV)
1
, open list,
panachage, and single transferable vote (STV)
2
(Bowler et al. 2003; Bowler
and Grofman 2000; Grofman et al. 1999). These systems share a ballot struc-
ture in which voters are able to express more than a simple, categorical choice
between the competing candidates or parties. The voters are given much
greater freedom in completing the ballot paper, either in terms of making
multiple marks on the ballot paper against several candidates, or in some
cases being able to rank order the candidates. AV and STV, the two systems
dealt with in this book, share the latter feature, and are used for electing
respectively Australias lower and upper houses. I t is this emphasis on voter
choice, specifically with regard to how voters may rank order candidates on
the ballot paper, that sets these systems apart from other electoral systems.
Given Australias prominent role in influencing the development of elec-
toral institutions, it is ironic how few countries actually followed her lead in
the adoption of these two preferential electoral systems. At present, only
Ireland and Malta use STV, and the use of AV is even rarer, raising only spo-
radic interest in the Asia-Pacific region (for which Australia can claim some
credit: see Reilly 1997a, 2001) and having limited use in certain other coun-
tries (such as for electing the Irish president or the London mayor). Despite
the apparent lack of interest in preferential systems on the part of practicing
politicians and electoral engineers, these electoral systems still excite much
interest and debate among theorists, political scientists and electoral
reformers (a point made, among others, by Bowler and Grofman 2000;
Lijphart and Grofman 1984). But even given this scholarly interest in pref-
erential systems, we still need some explanation as to why Australia has ended
up as one of very few countries actually using them. This is one of the goals
of this book.
Quite apart from its Australian focus, this book also has an international
goal: we wish to locate preferential voting in the comparative study of elec-
toral systems. This book therefore addresses two main audiences: those inter-
ested in the comparative study of electoral systems and those interested in the
Australian political system. I n the latter case, we show what the study of elec-
toral systems can contribute to our understanding of the evolution and oper-
ation of a single political system; in the former case, we show what the
| 2 |
particularities of electoral system development in one polity can contribute to
the comparative study of electoral systems.
I n setting the scene for the book, this chapter is arranged in four sections.
We start by establishing two key questions of Australian electoral system
design that this study will be addressing those relating to the origins and
variation of these systems. The next section reviews a third set of questions
underlying this study, relating to the operation of Australias electoral
systems. This does not only cover narrow electoral outcomes; we argue that it
can also be broadened to include such issues as the logical properties of these
systems and the fate of individual candidates, the candidate-centred nature of
party politics and representation, and the theme of voter empowerment
and/or confusion. The third section provides a brief review of the wider insti-
tutional context in which these electoral systems operate in Australia. I t
reveals in-built tensions in the design of the countrys electoral institutions;
these tensions exist between (on the one hand) electoral systems that empha-
sise such features as voter choice and a candidate emphasis in electoral pol-
itics and (on the other) electoral institutions that introduce elements of
compulsion in the voting act and privilege parties over candidates. The last
section gives an outline of the structure of the book.
Australian electoral systems:
origins and variation
The essentials of both AV and STV are very similar, although they are dis-
tinct in terms of certain key electoral outcomes (for general summaries, see
Farrell 2001). AV and STV ballot papers are shown respectively in Figures
1.1 and 1.2. The idea underlying them is that voters are able to rank-order
the candidates on the ballot paper, thereby potentially having a say in the
election of all successful candidates. So, for instance, if a voters preferred can-
didate does not attract sufficient votes to be elected, the voter may still have
an opportunity to determine the fate of the other candidates in the race. AV
is a majoritarian electoral system operating with single-seat constituencies (or
electorates) in which, to be elected, a candidate requires at least 50 percent
of the vote. I f, on the basis of counting the first preferences on the ballot
papers, no candidate achieves an overall majority, the candidate with the least
votes is excluded and those ballot papers are distributed among the remaining
candidates based on the next preferences indicated on the ballot papers.
| 3 |
| Austral i an el ectoral system desi gn |
The process continues until one candidate emerges with an overall majority.
The process is similar for STV, but with one major difference: as a pro-
portional electoral system which means that there is a much closer corre-
spondence between the proportion of
votes a party receives and the number
of seats it ends up with in parliament
STV operates in multi-seat con-
stituencies. A quota is established,
based on the numbers of votes and
seats, and equivalent to the majori-
tarian formula applied in AV. This
quota determines how many votes
each candidate must attract in order
to be elected, and this produces a pro-
portional election result. I n addition,
it adds a further layer of complication
to the count process because now
account must also be taken of those
candidates who are elected with votes
surplus to the quota.
This book is about the origins,
development and operation of these
electoral systems in Australia. We do
not pretend to be the first to cover
such themes. Book length treatments
have included such monographs as
Jack Wright s classic Mirror of the
Nations Mind(1980) or, more recently,
the excellent collections by Marian
Sawer (2001) and Graeme Orr and
colleagues (2003). In addition, impor-
tant chapter-length studies have been
produced by Murray Goot (1985) and
John Uhr (1999), among others. Our
aim is to contribute to this corpus of
knowledge in two main respects, one
specific to Australian political science,
and the other relating to the compara-
| 4 |
FIGURE 1.1 AN ALTERNATIVE VOTE (AV)
BALLOT PAPER
tive study of electoral politics. In the first instance, our objective is to integrate
the findings of existing studies together with further scrutiny of historical doc-
uments and the analysis of aggregate and survey data so as to provide the most
thorough survey of these electoral systems to date, in terms of their origins,
operation, variation and consequences. Secondly, we will be conscious
throughout of the need to locate the Australian case in terms of the extensive
comparative and theoretical electoral systems literature. We see these aims as
entirely complementary: Australian experience of electoral system design has
much to inform the political science community, but equally there are issues
and methodologies applied in comparative research that can provide important
insights into the nature of these systems.
Three main questions are analysed in this book, two of which are reviewed
in this section and the third in the following section. I n the first instance, we
are interested in exploring how Australia ended up with these electoral
systems. We already know from existing Australian studies that a combina-
tion of factors was behind this, notably lessons learned from experiments in
certain pre-federation Australian colonies (especially Queensland and
Tasmania), and the role of key actors such as Catherine Helen Spence, I nglis
Clark and Edward Nanson. There is also excellent coverage in existing studies
(e.g. Graham 1968; Goot 1985; Uhr 1999) about the point at which these
systems were introduced; these studies make it clear that selfish party inter-
ests played at least some role. But such facts, while necessary parts of the
story, fall short of explaining why exactly Australia ended up as the home of
preferential electoral systems.
| 5 |
FIGURE 1.2 A SINGLE TRANSFERABLE VOTE (STV) BALLOT PAPER
I t is undoubtedly true, and has been shown by comparative research,
that the ultimate decision on which electoral system to adopt usually rests
with the party political elite (although, in recent years, there has been
increasing reference to the electorate through a referendum, such as in
I taly and New Zealand), and, therefore, the role of inter-elite bargaining
has tended to attract the attention of scholars. The general consensus is
that, for the most part, short-term self-interest (or, more accurately, party
interest) tends to prevail over altruistic visions of what s best . A good
example of this point is provided by Kathleen Bawns study on the origins
of Germanys post-war electoral system (1993; more generally, see the dis-
cussion in Part I I of Shugart and Wattenberg 2001). This issue has already
received coverage in the Australian literature, in which attention is given
to the so-called Flinders deal in 1918 that resulted in the introduction of
AV for the House, and to the decision by the Chifley government in 1948
to introduce STV for the Senate (see pp. 3644). I n both instances, accu-
sations were made that narrow partisan interests had prevailed over
grander notions of social engineering and well-considered electoral system
design. Regardless of the accuracy or otherwise of such arguments
(Graham 1968; Uhr 1999), the degree to which they are sufficient in
explaining Australias penchant for preferential electoral systems in the
first place needs further exploration.
Furthermore, close scrutiny of the historical evidence reveals important
new insights into the precise details of Australian electoral system design
which, in turn, informs our understanding of the relative significance of con-
tributions by key actors (Clark, Spence and Nanson) and of lesson-learning
at state level. The role of electoral system experts has tended to be overlooked
in the debates (comparative as well as Australian) on electoral system design.
The evidence we present in chapter 2 suggests that they played a very impor-
tant role in this case. Given the complexities in the operation of these
systems, particularly in the case of STV, there seems to have been great scope
for expert advice to have had an important role, in some notable instances
developing features in the electoral systems that may well not have been those
intended by politicians. Electoral systems serve three main functions: first,
setting out how votes are to be cast; second, establishing rules on how the
votes are to be counted; and third, thereby determining who is to be elected.
Arguably, when considering electoral system design, the major interest of the
party political elite is in matters relating to who ultimately gets elected, and
how votes are cast. The issue of how votes are counted moves into slightly
| 6 |
more specialised and esoteric territory, which is very much the domain of the
backroom specialist, particularly when as in the case of STV we are
dealing with an electoral system whose counting rules can become very
complex. As we shall see in chapter 2, Nansons input in 1902 was particu-
larly interesting in this regard.
A second set of questions posed by this study refer to the evolution and
development of Australian electoral system design. I n this instance, our aim is
to build on previous analyses by such scholars as Murray Goot (1985), Scott
Bennett (1992) and Ben Reilly (2001), and on our own earlier research (Farrell
et al. 1996), to provide the most complete coverage to date of the nature of
electoral system design in Australia, at national, state and territory level.
Existing studies provide important insights into the extent of variation in elec-
toral system design, such as Reillys analysis of Queenslands experiments with
the contingent vote (Reilly 1997), Bennett s emphasis (1992) on the contri-
butions made by New South Wales to the origins of Australian ticket-voting,
or our own research on how STV has tended to vary in its detail across
Australia (Farrell et al. 1996). I n this book we integrate these previous studies
together with a re-reading of the relevant primary legislation to provide the
most thorough and complete contemporary mapping of Australias national,
state and territory electoral systems and therefore, by implication, the most
complete mapping of the AV and STV electoral systems to date.
Studying electoral system variation across Australia is not just of intrinsic
interest for its own sake. I t also informs the analysis of electoral system con-
sequences which brings us to our final set of questions: how Australias elec-
toral systems operate in practice.
Studying the operation of
Australias electoral systems
To date, Australian research has tended to focus on the electoral implications
of these systems. I n the case of AV, the question posed is over the degree to
which this majoritarian electoral system produces electoral outcomes that are
significantly different from the simple member plurality (SMP) electoral
system that it replaced (that is, the electoral system that was adopted in the
first Commonwealth Electoral Act of 1902). Joan Rydons seminal analysis
(1956) indicated little significant difference between SMP and AV. More
recent research (e.g. Reilly 2001) has suggested that there are important
| 7 |
differences, and this study will seek to extend and update his analysis.
I n the case of STV there is debate especially in the comparative litera-
ture over whether the system should even be classified as a proportional
electoral system. Table 1.1 provides details on the worlds main electoral
systems, showing where AV and STV are located in comparison to other
systems, and how the various electoral systems fare in terms of their effects
on overall levels of disproportionality (that is, the correspondence between
proportions of votes won by a party and the proportion of seats it is awarded)
and the number of parties that win seats.
3
I t is generally accepted that the list
systems of proportional representation tend to produce the most proportional
results with an average disproportionality of 3.9 (see also Katz 1997; Lijphart
1994; Taagepera and Shugart 1989). Given the small number of cases of
STV, its record is less impressive in this regard, scoring a poor 5.4 on the dis-
proportionality scale, causing some authors to go so far as to suggest that it is
a semi-proportional or quasi-proportional system (Katz 1984; Taagepera
and Shugart 1989). By extending the analysis to include the full set of STV
cases across the world (namely the Australian national and state cases, as well
as the Irish Republic and Malta), this study will provide the first-ever system-
atic test of just how proportional STV is.
This book also broadens the analysis of electoral system consequences to
cover themes that have tended to be neglected or under-emphasised in
Australian research to date. The first theme is the logical properties of these
| 8 |
TALE 1.1 TlE VAlN FAVlLlE3 0F ELECT0RAL 3Y3TEV3
E|eotora| systems Proport|on
of oountr|es [7}
Proport|on of
popu|at|on [7}
Average
d|sprop. [S|}
Average no. of
part|es [N
s
}
Nor-propor||ora| sys|ers
3|rg|e reroer p|ura|||y 18. 52.8 12.3 3.1
A||erra||ve vo|e 1.Z 0. 10.3 2.5
TWo-rourd 3.1 2.3 21.1 2.3
V|xed reroer rajor||ar|ar 20.3 21.9 9.9 1.Z
Propor||ora| sys|ers
L|s| 1Z.5 18.3 3.9 1.5
3|rg|e |rarslerao|e vo|e 1.Z 0.1 5.1 3.2
V|xed reroer propor||ora| .8 3.9 3.8 1.3
N0TE8 0r|y |rc|ud|rg cour|r|es W||| a popu|a||or grea|er ||ar |Wo r||||or (1998 es||ra|es), ard W||| a Freedor louse
score ol 1.0 or |ess (1998-99 rar||rgs).
0l = 0a||ag|er (Leas| 3quares) |rdex ol d|spropor||ora|||y.
N
s
= Laa|so/Taagepera |rdex ol ellec||ve ruroer ol par||arer|ary par||es.
80UR6E Farre|| (2001).
systems and how they impact on the electoral fate of candidates; the second
is the behavioural consequences of preferential systems. The first of these
concerns whether, on all occasions, an electoral system always elects the
correct candidate. This relates to a branch of the theoretical literature on
electoral systems known as social choice theory. Studies in this tradition
adopt an axiomatic approach, establishing a set of properties and measuring
the success of different voting procedures against them. I n this case the focus
is less on measures of proportionality and on how many seats each party wins,
and rather more on the issue of which candidates are actually elected. What
concerns studies in this tradition, in particular, is whether an electoral system
can guarantee that, on all occasions, the candidate with the largest number of
votes is the one who is elected.
I ain McLean and his various collaborators have shown the close relation-
ship between the modern historical antecedents of social choice theory and
the practical development of systems for electing parliamentary representa-
tives in the first wave democracies of Europe, North America and Australasia
(McLean 1996; McLean and Hewitt 1994; McLean and Urken 1995;
McLean et al. 1996). The debates in Britain in the late nineteenth and early
twentieth centuries were particularly significant in this regard, not only
reflecting its general status as a world power, but also its direct influence over
electoral system design in some of its colonies. Given the prominence then
being given in Britain to debates over electoral reform and the potential of
adopting STV (and AV), not least by prominent thinkers such as John Stuart
Mill, and influential pressure groups like the Proportional Representation
Society of the United Kingdom and Ireland (see Hart 1992), it is not sur-
prising to find theoretical interest being shown in such systems, notably in
the work of scholars like C. L. Dodgson (otherwise known as Lewis Carroll)
and Edward Nanson (who features in our next chapter). While finding much
to praise in AV and STV as compared to other electoral systems in operation
or being proposed as alternatives, these scholars were quick to point out some
fundamental problems that meant the systems could produce perverse social
choice outcomes a finding that is given extensive support by more recent
social choice research (for a sample, see Brams and Fishburn 1984; Doron
and Kronick 1977; Dummett 1984; Nurmi 1996/7; Tideman 1995).
4
The shortcomings of preferential systems are incontrovertible and well
documented; however, it is worth noting that social choice criticism is not
directed exclusively at these electoral systems. I ndeed, throughout the debates
in this literature it is observed that similar criticisms can be levelled at all the
| 9 |
other available alternatives, and while theoretically there are electoral systems
that could be designed to address many of these problems (such as, for
instance, Brams and Fishburns [1983] approval voting, or the Borda Count),
it is arguable that no electoral system could ever be perfect (Arrow 1951).
This latter point has led some scholars to suggest that, instead of striving for
electoral system perfection, attention might better be directed towards
assessing the advantages and disadvantages of the different electoral systems
currently on offer, so that instead of assessing the systems in terms of pre-
defined axioms, the judgement should be based on the purported merits of
the systems themselves (Levin and Nalebuff 1995; Sen 1995; also Saari
2001a).
5
This is the procedure that we shall be adopting in this study, in
which we assess the extent to which AV and STV can influence the electoral
fate of individual candidates.
Finally, we examine the behavioural implications of the preferential elec-
toral systems both for Australian politicians in their campaign and repre-
sentative roles (chapter 5) and for voters with regard to how much they
understand of these systems, their strategic use of them, and their general
attitude towards them (chapters 6 and 7). I n the introduction to this chapter
it was observed that the one feature that separates preferential systems like
AV and STV from other electoral systems is the emphasis they place on voter
choice, in the sense that voters are not restricted to voting for a single candi-
date or party but are free to vote for a range of individual candidates on the
ballot paper. This is the main feature that is said to set these systems apart
from their main competitors. According to George Hallett, for instance,
under STV the voter is saved from any worry about wasting his vote. His
preferential vote is insurance against that. He can safely vote his real order of
choice as far as he has any Voting will always be worthwhile. You are
almost sure to be a winner (1984: 11920). Enid Lakeman extols the virtues
of STV for enabling each citizen to take part as freely and as fully as possible
(1974: 111). Even among more sober scholarly treatments, there is stress on
how systems like STV give voters greater choice and makes possible ballot
splitting to express highly differentiated preferences (Bowler and Grofman
2000a: 1).
The question of choice is not only concerned with how much choice is
available to voters; it also concerns the nature of the choice that is offered, in
particular whether the choice is between parties or candidates (Bowler and
Farrell 1993; Farrell 2001). This provides an additional perspective on how
preferential systems can be distinguished from other electoral systems, and it
| 10 |
adds strength to Bowler and Grofmans view that such systems may not
belong to just one family tree but to several, depending on the kinds of ques-
tion we want to ask about electoral systems (2000b: 270).
These two dimensions of voter choice extent and nature provide the
basis for a simple typology of electoral systems as shown in Table 1.2, in
which AV and STV are separated from other electoral systems by their char-
acteristics as both candidate-based and ordinal. The comparative literature
has produced evidence of how variations across these dimensions can affect
politicians both in terms of their campaign styles (Bowler et al. 1996; Katz
1980) and representative roles (Bowler and Farrell 1993; Farrell and Scully
2003). At one extreme, in closed list and simple member plurality systems,
voters must choose one or other party (closed list) or candidate. At the other
extreme are those preferential systems that allow voters to rank-order all the
candidates from all the parties (e.g. Bowler and Grofman 2000b; Farrell and
Gallagher 1998). Ordinal ballot structures and candidate-centred voting
place a greater emphasis on individual politicians and on the need to cultivate
a personal vote (Cain et al. 1987; Searing 1994). This manifests itself in a
greater attention by politicians to the constituency link with their voters,
greater emphasis on personal vote chasing in campaigns and greater stress
outwardly on voters than inwardly on the party.
Then there are the voters, the objects of all this attention by politicians.
According to supporters of these systems, like George Hallett, John
Humphreys, or Enid Lakeman, their central virtue is that they empower the
voters. This is the sense that the voting act faces minimal interference by the
party establishment: voters are able to pick and choose between all the politi-
cians on the ballot paper; in the case of STV the parties are unable to priori-
tise certain politicians over others (there is no such thing as a safe seat );
indeed, at one point it was even believed (and was argued in the parliamen-
tary debates over these systems in Australia see chapter 2) that the parties
| 11 |
TALE 1.2 A TYP0L00Y 0F ELECT0RAL 3Y3TEV3 8A3E0 0N 8ALL0T 3TRuCTuRE ClARACTERl3TlC3
Extent of oho|oe
0aregor|oa| Dro|na|
Nature
of
oho|oe
0ano|oare-oaseo 3|rg|e reroer p|ura|||y (uK) A||erra||ve vo|e, 3|rg|e |rarslerao|e
vo|e
Parry-oaseo C|osed ||s| (3pa|r) 0rdered ||s| (8e|g|ur)
might abolish candidate selection altogether and simply let the voters pick
and choose between a large range of candidates, thereby taking the represen-
tative process about as far away from Schumpeterian, party-dominated
notions as possible.
Two issues stem from this feature of allowing voters to express as many
preferences as they like on the ballot paper. The first is whether the voters are
given free rein to express only a few preferences; the second is whether it
makes the voting act so complex that it confuses them. The first point has a
particular relevance to the Australian case, in which there is a tendency to
require voters to express some minimum number of preferences (see the fol-
lowing section). Some would argue that this is contrary to the spirit of pref-
erential voting, because it replaces choice with compulsion (e.g. Lakeman
1974: 14546). Taken together with the other elements of compulsion in
Australias electoral institutions compulsory registration and voting we
have an interesting nation-based qualification to the electoral systemic-
derived norm of voter choice, and one that warrants attention.
As for the issue of complexity and voter confusion, this is a criticism that
is regularly levelled at STV: for instance, it was a stated reason why, in the
deliberations over possible electoral reform in Britain in the late 1990s STV
was deemed ineligible ( Jenkins 1998). The complexity of preferential systems
also has a bearing on debates over strategic voting. According to the social
choice literature, all electoral systems are subject to voter strategizing, in
which voters are said to vote non-sincerely paying more attention to pre-
ferred outcomes than preferred candidates. This has been extensively studied
in the case of single member plurality elections, in which a voter, knowing
that her preferred candidate has no chance of being elected, votes tactically
for another candidate who appears to stand a better chance of defeating the
candidate least liked by the voter. Strategic voting under simple electoral rules
does appear likely, but once the electoral system becomes more complex there
is a question over the degree to which voters will have the informational and
computational skills required to vote strategically. The general view in the lit-
erature is that the complexity of STV makes it unlikely that voters will be able
to make such simple strategic decisions (Bartholdi and Orlin 1991; Bowler
1996; Levin and Nalebuff 1995). Given that fewer permutations are involved
in the case of AV, it would appear more open to strategic use (Dummett
1997), though even here there are some doubts (Cox 1997: 95).
A distinction needs to be drawn between two issues: whether there is
strategic voting and whether it is effective. Even if there are questions over
| 12 |
the extent to which systemic complexity makes strategic voting infeasible,
this does not preclude the possibility that some voters may still attempt to
vote in a strategic manner. There are studies that have shown evidence of
strategic voting under STV (Bowler and Farrell 1991; Dunleavy et al. 1992;
Jesse 1999; Laver 2000). I ndeed, once we consider the theme of strategic
voting under preferential systems, it brings into sharp relief the notion of
strategic voting as insincere voting, in which voters are said to place a higher
priority on electoral outcome than on preferred candidate, for, as we shall see,
there is scope for a voter to be both strategic and sincere under these electoral
systems, and this suggests a need to give more thought to the conceptualisa-
tion of strategic voting.
Locating preferential
electoral systems in Australia
Given this books focus on Australia, it is important that we clarify the wider
institutional context in which our electoral systems are operating, for as
Grofman has asserted: seemingly identical institutions may not always yield
similar behavior once we recognize contextual factors and constraints (1999:
xi). Australia exemplifies this perfectly, because while there is a strong
emphasis on voter choice in the countrys electoral systems, at the same time
we are dealing with an institutional context in which parties predominate.
The vast majority of voters identify with and vote for one of the major polit-
ical parties; for an election candidate, gaining election to the lower house, at
least at national level, is next to impossible without the benefit of one of three
party labels Liberal, National or Labor. For their part, minor parties or
independents have had little role in shaping the development of the party
system. Within the legislature, party government operates in every sense of
the word, with the enforcement of rigid party discipline among elected rep-
resentatives (McAllister 2002).
Strong party control of the political system has its origins in the countrys
political culture. Louis Hartz (1964) argued that the cultural development of
Anglo-American colonial societies is determined by the values and beliefs
that were dominant during the period in which they split from the host
society, Britain. Since Australias split occurred in the nineteenth century, the
colonial fragment that emerged was imbued with the utilitarian ideas of
Jeremy Bentham and his followers. As a result, Australian political culture has
| 13 |
been avowedly utilitarian in its orientation (Collins 1985). Keith Hancock
(1930: 69) famously characterized the situation as follows: Australian democ-
racy has come to look upon the state as a vast public utility, whose duty it is
to provide the greatest happiness for the greatest number. Examples of this
utilitarian outlook include the extensive use of the law to regulate society and
to ensure proper social conduct, from industrial relations to the system of
compulsory voting.
At a time of large scale electoral change across most developed industrial
societies, the Australian party system has remained remarkably stable. The
parties that compete for electoral support at the beginning of the twenty-first
century can trace their ancestry to the parties that existed at the beginning of
the last century ( Jaensch 1994; McAllister 2002). Moreover, with the excep-
tion of some comparatively minor splits and fissures, parties outside the major
Labor versus Liberal-National division
6
have gained little electoral success.
Of the 39 federal elections that have been held between 1901 and 2004, only
seven have produced a non-major party vote that has exceeded 10 per cent of
the first preference vote
7
and in only one federal election 1990 has a single
minor party gained more than 10 per cent of the first preference vote.
8
This
is a level of party dominance with the electorate which is rare perhaps
unprecedented among the established democracies (Dalton and
Wattenberg 2001).
Underpinning the predominance of party among legislators is, then, the
central role that parties have among voters. I n turn, this has its origins in the
system of compulsory voting (or, more accurately, compulsory attendance; see
chapter 6, pp. 122-26), which ensures that even uninterested or apathetic
voters have to go to the polls. Complementing this is the frequency of elec-
tions at both the state and federal levels. Federal elections must be held every
three years (staggered in the case of senators whose terms of office are six
years). I n most states the same rule applies; the exceptions are New South
Wales, where the period between elections is four years, and South Australia,
which has a four-year fixed term. I n practice, however, elections are held
much more frequently. Of the 22 federal elections that have been held in the
post war years, seven have been conducted less than two years after the pre-
vious election. The net effect is that the average Australian voter must attend
the polls around once every 12 to 18 months, an unprecedented level of
voting among the advanced democracies.
The consequence of almost all of the electorate voting in an election is a
high level of party identification (Figure 1.3), leading Don Aitkin to
| 14 |
comment: the causes of stability are to be found in the adoption, by millions
of Australians then and since, of relatively unchanging feelings of loyalty to
one or other of the Australian parties (1982: 1). Surveys demonstrate that,
with the exception of the late 1960s and the early 2000s, periods of Liberal
dominance and Labor decline, no more than 5 percentage points separated
the proportions identifying with the two major parties. Although the vast
majority of voters identify with one or other of the major parties, there has
been a modest increase in the proportions with no party attachment. I ndeed,
this group has increased threefold in size since 1987, reaching 16 percent in
the 2004 election. The main change occurred in the early 1990s, when the
number with no attachment jumped from eight percent of the electorate in
1990 to 17 percent in 1996.
Party identification is also expressed in strength (Figure 1.4). There has
been a decline in the strength of party identification: those who see them-
selves as very strong identifiers have declined by almost half, from 34 per cent
in 1979 to 20 per cent in 2004. Once again, the major change appears to have
occurred at a specific point in time, in this case between 1979 and 1987. The
| 15 |
0
10
20
30
40
50
60
1967 1969 1979 1987 1990 1993 1996 1998 2001 2004
Labor
Lib-Nat
Other
None
FIGURE 1.3 THE DIRECTION OF PARTY IDENTIFICATION, 19672004 (%)
SOURCE 1967 and 1979 Australian National Political Attitudes surveys; 1987-2004 Australian Election Study
surveys
net increase resulting from this weakening in partisanship has taken place
among those who are not very strong partisans, who now make up around
one-third of all voters. Despite these changes, placed in a comparative context
Australia has one of the highest levels of partisanship in the world (Dalton
and Wattenberg 2001).
The combination of a dominant party political culture and a strong inbuilt
element of compulsion in the sense that, by law, voters are obliged to reg-
ister, to turn out, and (if they wish their vote to count) to complete a
minimum number of preferences on the ballot paper are interfaced with an
electoral system design that supposedly favours high elements of voter choice.
As a consequence, Australian voters (and, by extension, also the politicians)
face tensions between the party-centred nature of Australian electoral poli-
tics, the compulsory aspects of the vote process and the voter choice emphasis
of the electoral systems. We can expect these institutional tensions to have
implications for how both politicians and voters operate in Australia, and this
expectation is tested in the analysis that follows, particularly that part of the
analysis focused on the behavioural consequences of such systems.
| 16 |
0
10
20
30
40
50
60
1967 1969 1979 1987 1990 1993 1996 1998 2001 2004
Very strong
Fairly strong
Not very strong
SOURCE 1967 and 1979 Australian National Political Attitudes surveys; 19872004 Australian Election Study
surveys
FIGURE 1.4 THE STRENGTH OF PARTY IDENTIFICATION, 19672004 (%)
Assessing Australias electoral systems
This book is organised into four main parts, examining Australias preferen-
tial systems in terms of their origins, variations, consequences, and voter reac-
tion to them. We start, in chapter 2, with a review of the origins and historical
evolution of Australias electoral systems at national level.
9
The chapter deals
with the first failed attempts to introduce preferential electoral systems in the
Commonwealth Electoral Act (CEA) of 1902, an episode in Australian
political history that revealed strong signs of deliberate intent in electoral
system design, influence from the states (Queensland and Tasmania in par-
ticular), a prominent role by electoral system experts (notably Edward
Nanson and Catherine Helen Spence), and inter-party machinations at the
highest level. A fascinating feature of the 1902 debate was the nature of the
STV system that was being proposed: bearing the hallmarks of strong influ-
ence by Edward Nanson, one of the founders of social choice theory
(McLean 1996), the system might have gone some way towards dissipating
social choice concerns with STV. Had the government won the debate, this
would have been a unique version of STV, the like of which has never been
tried elsewhere.
I n 1902 the Protectionist party was unsuccessful in its efforts to introduce
preferential electoral systems, but the fact that these electoral systems played
such a prominent role at this critical stage of state-building was to guarantee
their place in future debates over electoral reform in Australia. Chapter 2 also
examines the subsequent efforts to introduce these systems, culminating in
the 1918 legislative reform that brought in AV for House of Representative
elections, a hybrid preferential block system for the Senate a year later, and
ultimately, in 1948, the adoption of STV for Senate elections (and its subse-
quent amendment in 1983). As we shall see, the short-term machinations of
party leaders were crucial at every episode; but looking at the fifty years of
debate as a whole, it is undoubtedly the case that longer-term views on elec-
toral system design were also important in the gradual move towards
adopting preferential systems.
As a federal system, Australia provides plenty of scope for debates over
electoral system design at different levels of government. Chapter 3, in
shifting from a temporal to a spatial focus, considers a second strand of
analysis of the preferential systems, namely their variability. The fact is that
these systems are open to considerable variation particularly STV, which is
probably more accurately described as an evolving family of vote-counting
| 17 |
rules rather than a single rule (Tideman 1995). As Bruce Coram (1996) has
shown in his discussion of second-best theories, small deviations in institu-
tional design can have profound implications for how an institution operates.
Preferential electoral systems lend themselves perfectly to an examination of
this point (indeed, Coram makes just such an observation; Coram 1996:
10001). I s one variant of a preferential system preferable to another? How
does a change in one feature such as the minimum number of preferences
required of voters, or how surplus votes are transferred under STV affect
the overall operation of the electoral system?
Chapter 3 elaborates on the main variations in the AV and STV systems,
thereby increasing our observations and providing the raw material for the
answers to these questions in the chapters that follow. The chapter explores
the rich range of variation in the versions of AV and STV being employed at
state, territory and national level, and the reasons for such variation.
Attention is also given to the use of other electoral systems in Australia (such
as the variants of PR-list once used in South Australia and the ACT). The
chapter provides a bridge between our analysis of the origins and the conse-
quences of Australias preferential systems. Variations in the design of the
preferential electoral systems play a significant role in our understanding of
electoral system consequences such as who gets elected (thereby potentially
affecting system proportionality), how the count proceeds (speaking to a
wider social choice agenda), and the strategic behaviour of politicians and
voters. Moreover, the very fact that there are notable differences demonstrates
the importance of taking account of the embedded nature of electoral system
design, of locating the systems in their wider political and historical context
(Grofman et al. 1999a), and of enabling greater understanding of the origins
of these electoral systems, why they take a particular shape, and how they
affect the wider political system and its relevant actors.
The third part of the book considers the consequences of these preferen-
tial systems and does so primarily in the context of how these systems operate
in Australia at national, state and territory level. The two systemic implica-
tions of AV and STV, both in terms of aggregate proportionality and also
with regard to social choice debates over voting paradoxes, are considered in
chapter 4. The issue of proportionality is of particular relevance to STV. By
including data on trends in all post-war Australian elections (national, state
and territory), and also including data from the other two prominent STV
cases, Ireland and Malta, the chapter provides the most complete investiga-
tion to date of how this electoral system fares in terms of aggregate measures
| 18 |
of proportionality, how it compares with other proportional systems, and how
the variants of STV (as set out in chapter 3) compare in this regard. This
scope for variation in these systems also sets the scene for an analysis of how
they fare in terms of the electoral fates of individual candidates speaking to
the concerns of social choice theorists.
We saw above how the nature of electoral system design can have a
bearing on the way in which politicians implement their representative role.
This theme is taken up in chapter 5, where we make use of the Australian
Candidate Study to examine how Australian politicians operate in the cam-
paign environment, and their conceptions of their representative role.
Particular attention is paid to the tensions in Australias electoral systems
between the voter-oriented elements of preferential voting and party-ori-
ented elements of compulsory expression of preferences, how to vote cards
and ticket voting.
Chapter 6 considers a final set of implications of these electoral systems,
relating to voter engagement with them, specifically in the context of
Australian electoral politics. I n the light of prominent elements of party
control and voter compulsion that percolate through the Australian electoral
process, the chapter starts with a review of these features. Adding to the cov-
erage of the compulsory elements of Australias electoral systems dealt with
in chapter 3, in this chapter we trace the origins and practice of compulsory
voting at the national level. The chapter also assesses the efforts of the parties
to control the preference voting activities of their supporters (which, in turn,
affects the form we might expect strategic voting to take in such systems).
Between them, these elements of voter compulsion and party control are at
odds with the much-touted supposed virtue of AV and STV in promoting
voter choice. This leads to an analysis of voter reaction to the countrys elec-
toral institutions. Given the elements of voter compulsion and the inherent
complexity of the electoral systems, there are grounds for expecting high
levels of invalid votes and dissatisfaction with the electoral process. The
chapter makes use of survey evidence to assess this.
Since preferential systems are lauded for the degree of choice and flexi-
bility they give to voters, we finish chapter 6 with an examination of how the
voters view the electoral system and trace some aspects of voter reaction in
the Australian context. But given the peculiarities of the Australian case in
juxtaposing the free choice features of preferential voting with a strong
element of voter compulsion, it is important to take the analysis out of a
national context. Accordingly, chapter 7 provides a comparative assessment of
| 19 |
the place of preferential systems in terms of measures of voter satisfaction
with democracy, using data from the Comparative Study of Electoral Systems
(CSES) project.
Finally, the book concludes, in chapter 8, with a discussion of what the
Australian usage of preferential systems contributes to our understanding of
electoral systems in general and preferential systems in particular, and a
review of current debates over Australian electoral reform.
| 20 |
t w o
Australias electoral
systems, 190283
I n 1918, with the introduction of the alternative vote (AV) for elections to
the House of Representatives, Australia became the first country to adopt
preferential voting at the national level. Thirty years later, when the single
transferable vote (STV) was introduced for Senate elections, the shift to pref-
erential voting was completed. At the time, these reforms were seen as
nothing more than short-term political fixes by politicians anxious to cling
onto power. I n 1918, the essence of the Flinders deal (see p. 38) was an
attempt by the Hughes government to prevent the right-of-centre parties
from splitting their support to the benefit of the Labor party. I n 1948, it was
the turn of the Chifley Labor government to replace the Senates preferential
block system with STV, in a manoeuvre to diminish the expected electoral
gains of the new Liberal leader, Robert Menzies.
There can be little dispute over the fact that political expediency played an
important role in both cases, and certainly they were the immediate triggers
for the adoption of these systems. I ndeed, to G.S. Read and Martyn Forrest,
[i]n both cases the reforms became law not as a result of the pursuit of prin-
ciples of electoral justice, but from pragmatic consideration of party gain
(1989: 99; see also, Rydon 1968). But, as others have observed (Graham
1968; Uhr 1999, 2000), such explanations are only partial accounts of the
story. What still needs to be explained is why these electoral systems were
chosen, and why it took so long for them to become established. The purpose
of this chapter is to provide answers to three questions:
| 21 |
What are the origins of Australias long-standing attachment to preferen-
tial voting?
Why were these particular forms of preferential systems chosen?
I f, as we shall discover, the adoption of these systems followed a logical
developmental path, why did it take almost twenty years for AV to be
introduced for House elections and virtually half a century before the
Senate adopted STV for its elections?
1
For answers to the first two questions, we need to look back to events
long before 1918, and in particular to the range of influences that predomi-
nated in the early debates before, and in the years surrounding, Federation.
I n their classic essay on the origins of Western Europes party systems,
Seymour Martin Lipset and Stein Rokkan (1967) wrote of the freezing of
party alternatives around historical cleavages that were prominent at the
time of suffrage extension. I n a similar manner, it can be argued that a
countrys debate over political institutions is in large part historically prede-
termined by the nature of events at a crucial historical juncture, such as
during a time of nation building, or, as in this case, a time of federation. As
we shall see, the debates over Australian electoral system design froze
around two main alternatives that predominated at the turn of the nine-
teenth and twentieth centuries, namely the plurality systems (single member
plurality SMP and block
2
) and the preferential systems (AV and STV).
This issue is explored in the following section, where we assess the range of
factors and significant people who affected elite debates over the electoral
systems for the new Federation.
The subsequent section examines the first attempt to introduce preferen-
tial systems for national elections in 1902. The fact that this was to prove
unsuccessful provides an explanation for why these systems were to take so
long to become established in Australia. Given that, by now, the political
parties were arranged into two camps those favouring plurality systems
versus those pushing for preferential systems the 1902 debate is generally
seen as a crucial first step in a reformist agenda that would finally bear fruit
several decades later when preferential systems were introduced (Graham
1968; Uhr 1999, 2000). There is undoubtedly a lot of truth in this, certainly
with regard to the AV system introduced for House of Representative elec-
tions in 1918 (see page 38 ff.), which bears a very close relationship to the
system proposed in 1902. But, as Table 2.1 indicates, the picture is more
complex when we examine trends in the design of the Senates electoral
| 22 |
system in the second section of this chapter. As we shall see, there is a lot
about the form of STV adopted for Senate elections in 1949 (to say nothing
of its reform in 1983) that is very different from the system proposed origi-
nally in 1902.
Early debates over electoral system design
Electoral systems in liberal democracies around the world went through a
very important developmental phase at the turn of the nineteenth and twen-
tieth centuries. As comparative histories have shown, a number of countries
were experimenting with new forms of electoral systems, and the first wave
of proportional representation (PR) systems date from around this time. I n
1885, an international conference was held in Antwerp to discuss the relative
merits of different proportional systems; Belgium became the first country to
adopt a proportional system in 1899, followed soon after by Finland (1906)
and Sweden (1907) (Carstairs 1980). That the new national Parliament in
Australia should also consider the adoption of some alternative to SMP or
block vote systems would not seem too surprising, therefore; indeed, there is
evidence that Australias first prime minister, Edmund Barton, studied
Belgiums PR system prior to the debate over the 1902 Commonwealth
Electoral Act (Graham 1968).
| 23 |
| Austral i as el ectoral systems, 1902 83 |
TABLE 2.1 HOUSE OF REPRESENTATIVES AND SENATE ELECTORAL SYSTEMS SINCE 1901
State Electoral System
House of Representatives
1901 NSW, Vic, WA Single member plurality (SMP)
Qld Alternative vote (AV using contingent vote)
SA Block vote
Tasmania Single transferable vote (STV)
1903 All states SMP
1918 All states AV
Senate
1901 All states, except Tasmania in 1901 Block vote (STV in Tasmania)
1919 All states Preferential block
1949 All states STV
1983 All states STV with ticket-voting
But where does Australias fixation with preferential systems come from?
I t is evident that a cocktail of three main factors lay behind the countrys
attraction to these systems: electoral reform debates in Britain, the role of
Australian activists and state-level experiments in Australia. First, like all
other British colonies, Australia was heavily influenced by events and debates
in the mother country. This was the period in which creaky British institu-
tions were gearing up for the inevitability of mass suffrage extension, and in
consequence there was considerable debate over the nature of the electoral
system. There was also some small-scale experimentation with alternative
systems, none of it deemed terribly successful. Notable examples are the
limited vote from 1867 to 1884 (as we shall see, at one point this almost
became the system for Australian Senate elections in 1902); and the cumula-
tive vote (for school board elections) in 1870 (Bogdanor 1981; Butler 1963;
Bowler, Donovan and Farrell 2000). But it was the British debates over STV
that had most influence on events in Australia.
STV was for the most part credited as a British invention and therefore
as the British system of PR (Lijphart 1987) though, in truth, it was also
invented separately and at around the same time by Carl Andrae in
Denmark. The merits of STV began to be debated from the 1850s onwards,
most notably after the publication of Thomas Hares Treatise on the Election
of Representatives, Parliamentary and Municipal (1859) and the active and
enthusiastic promotion of his system by John Stuart Mill (Hart 1992). From
the 1880s onwards, the Proportional Representation Society also featured
prominently in efforts to promote STV. I n large part, Hares ideas were dis-
missed by the British establishment, and it was not until the 1917 Speakers
conference that any real effort was made to propose the adoption of STV, in
combination with AV, for British elections. The proposals were largely unsuc-
cessful;
3
what followed, instead, was a period of experimentation in various
parts of the British empire, with STV being introduced for all-Ireland elec-
tions in 1920 (subsequently retained by the Irish Free State in 1922), in Malta
in 1921, and in parts of I ndia in the 1930s (Farrell 2001). Evidently there was
no such direct British government intervention in electoral system design in
the Australian colonies, but the British debates undoubtedly had some influ-
ence on Australian-based actors, and through their actions on developments
in some colonies.
The second factor that is seen to have influenced electoral system design
for Australian national elections was the move in some Australian colonies to
| 24 |
experiment with new electoral systems. I n 1896, STV was introduced in
Tasmanias two urban areas, Hobart and Launceston. Several years earlier, in
1892, AV was brought in for electing Queenslands Legislative Assembly.
There is good reason for assuming that the Tasmanian and Queensland
experiments would have influenced those designing the national electoral
system in 1902 who were searching for alternatives to SMP and block
systems. We deal with the details of the Tasmanian electoral system shortly,
but at this juncture it is important to note a few points about Queenslands
AV system.
The reform of Queenslands electoral system by the Griffith government
was a compromise. The original intention had been to adopt a runoff system
(then quite popular in a number of countries: see Carstairs 1980), but after
long parliamentary debate, the government accepted the argument that it
would be both costly and difficult to administer in two-seat constituencies
(which then predominated), as well as in constituencies comprising large geo-
graphical areas with far-flung populations. I t was proposed, instead, to collapse
the runoff system into a single-shot version known as the contingent vote, in
which voters should vote just once, rank-ordering the candidates (Queensland
Parliamentary Debates July 5, 1892, 59496). I f no candidate achieves a
majority of the vote in a single-seat constituency, the candidates with the
highest and second highest vote tallies remain in the race and the ballot papers
of all the other candidates are transferred to them based on the next preference
expressed by the voters. Since only two candidates remain in the race, the
result, as in runoff elections, is majoritarian. As we shall see, the origins of this
contingent vote version of AV, as a one-shot variant of the runoff system, is
different from other versions of AV (see below and chapter 3).
The third influence was the role of key actors. They provide an important
corrective to the view that Australias contributions to electoral system design
were borrowed from overseas, a view held, for instance by Hancock, who sug-
gests that while Australia does indeed deserve her reputation as a seed-bed
of experiment , these seeds have come from outside (1947: 81). He is not the
only scholar to qualify praise for Australias willingness for institutional
experimentation by suggesting that many ideas were borrowed (e.g. Reeves
1902: 181). There may well be some truth in this, but such an argument runs
the risk of seriously downplaying the highly important role of Australian
actors in influencing the design and launch of these experiments. Nowhere
was this more apparent than in 1902, when the debates were dominated by
| 25 |
references to three key figures: a campaigner, Catherine Helen Spence; a leg-
islator, I nglis Clark; and a theorist, Edward Nanson, each of whom promoted
the adoption of preferential systems.
4
There has already been considerable historical and biographical coverage
of these three individuals (notably, Haward and Warden 1995; Magarey
1985; McLean 1996; Reid and Forrest 1989; Uhr 2002). But many of these
accounts tend to gloss over the precise details of what they were proposing.
Apart from some rare exceptions (McLean 1996), very little has been written
about Nansons views on preferential systems, while in the cases of Spence
and Clark, if anything, the tendency has been to overplay the distinctions
between their respective versions of STV. I n fact, as we shall see, Spence and
Clark were largely in agreement over how STV should operate (indeed, Clark
largely borrows from Spence). Spence and Clark had a huge influence on the
1902 debate, but as the next section reveals, when it came to the specifics of
the actual electoral systems being proposed, their role paled in comparison to
Nansons, for these were largely of his design. Before dealing with the 1902
proposals, it is useful to outline the main features of the Spence and Clark
versions of STV.
I n the parliamentary debates surrounding the 1902 Electoral Bill, there
are many references to the Hare-Spence system of STV, a title that Spence
herself used (Spence 1987: 557). This consisted of three main features. First,
she proposed the use of large multi-member constituencies (or electorates),
ideally electing as many as nine or ten members. Spence is credited as the first
person to propose this important simplification of the Hare scheme.
5
Hare
and Mill were fixated on the idea that the whole country should be one large
constituency. Spence recognised the obvious complications that this would
impose, and in an 1861 pamphlet, suggested as a modification that the
country should form a series of multi-member constituencies. A second
feature of the Hare-Spence system was the use of a Hare quota. According to
Magarey (1985: 157), Spence toyed briefly with the Droop quota
6
in 1892,
but quickly abandoned it in favour of the Hare quota
7
on the quite reasonable
grounds that it is simpler to explain, but also on the mistaken grounds that it
produced a fairer result. Third, Spence proposed the use of the Gregory
method for the transfer of surpluses.
8
Clarks role as a promoter of STV relates to his period as Tasmanias
Attorney General in the 1890s, when he sought to reform the colonys elec-
toral system. He was on record as a supporter of Hares scheme, praising its
qualities in an 1874 article in Quadrilateral, a short-lived journal that he
| 26 |
edited. Twenty years later, he proposed the use of an adapted form of the
Hare system for Tasmanian House of Assembly elections in two urban
areas, Hobart and Launceston, as a means of eradicating the problems
caused by the operation of the single member plurality (SMP) system, in
which manipulation by electoral agents and the running of candidate
tickets had made it very difficult for minorities to win any seats (Bennett
1995: 146). Clark proposed the use of STV in the two urban areas only
because he realised that to attempt a uniform adoption would have met
with serious opposition from rural politicians. For now they would continue
to be elected by SMP, and Clark hoped that in time STV would be
extended across the colony.
The Hare-Clark system was, for a time, called Clark-Hare, including by
Clark himself. I t would evolve later, but at this point it consisted of the fol-
lowing four features (which in some instances resulted from amendments to
Hares 1896 Bill at committee stage): multi-member constituencies (six seats
in Hobart, four in Launceston); the Hare quota; the transfer of surplus votes
entailing the transfer only of votes surplus to quota at full face value (that is,
random selection of ballots rather than the Gregory method); and voters
being required to express preferences up to half the number of seats to be
filled. Apart from not using the Gregory method for transferring surpluses
(which subsequently was introduced in 1907) and the rule requiring voters to
express a minimum number of preferences (an amendment to Clarks original
Bill), there is very little to distinguish this system from the Hare-Spence
system discussed above. I ndeed, from the outset it was clear that Catherine
Helen Spences writings had a lot of influence on the design of this first
Tasmanian system (Bennett 1995: 147; Newman 1992).
The inherent similarities of Hare-Clark and Hare-Spence were to become
even more apparent in subsequent amendments to the original Hare-Clark
system, made after Clark had left the political scene to become a Tasmanian
Supreme Court judge. I n its passage through the Tasmanian Legislative
Council, a clause was inserted into the 1896 Electoral Act limiting its oper-
ation to just one year and therefore, supposedly, for just one election. This was
subsequently extended annually, through to the 1900 election, the govern-
ment claiming that lack of time prevented it from giving the Act proper con-
sideration. The need to have a system in place for the first Commonwealth
elections of 1901 ensured the survival of the 1896 Act for one additional year
(Bennett 1995: 15758; Townsley 1956: 76), after which it was scrapped.
| 27 |
SMP was used for all Tasmanian elections until 1907 when STV was rein-
stated for House of Assembly elections, where it has been used ever since. But
it incorporated two important adaptations: first, the replacement of the Hare
quota by the Droop quota, and second, the use of the Gregory method for
dealing with transfers of surplus votes (see also chapter 3).
According to Townsley (1956: 76), in the light of such amendments, to
call the system Hare-Clark is a misnomer, although other authors are less
concerned about such historical detail (Mackerras 1995; Newman 1992). On
the whole, it is fair to say that Clarks major contribution to Australias con-
version to STV was in the area of policy proposals and the implementation
of the first STV system in Australia (and apart from Andraes Danish
experiments in 1855 [see McLean and Urken 1995: 4647] the first use of
STV in the world). Clarks contributions in the areas of theory and ideas was
rather less than Spences, and certainly far less than that of the last of our
activists, Edward J. Nanson.
Known as the experts expert (Uhr 1999: 27), Nanson was the man of the
moment when federal politicians needed ideas and ideals to incorporate in
the new electoral law (Reid and Forrest 1989: 88), and today he is feted by
social choice historians as one of the founders of social choice theory
(McLean and Urken 1995). He started publishing on the subject of electoral
systems as early as 1880, and in 1899 and 1900 he produced two pamphlets
that sought to influence the impending debate over the electoral systems for
the Commonwealth parliament. Having witnessed the use of the block vote
system for electing delegates to the federal conventions, Nanson was con-
cerned that the same system should not be adopted for the Senate. Neither
this nor the SMP system which he feared would be used to elect the House
of Representatives appealed to him on the grounds that they could not
ensure the election of the correct candidate, not only because of the dispro-
portionality of the result (even more acutely a problem in block vote systems)
but also because of voting rule inconsistencies built into such electoral
systems (which were outlined in chapter 1).
Nanson proposed, instead, the use of STV for the Senate and AV for the
House of Representatives. This was despite the fact that in an earlier paper
(Nanson 1995), he had also criticised these systems voting rule inconsisten-
cies. On the face of it, Nansons ambiguous attitude towards these systems is
curious. McLean accuses him of appearing to ignore his own lesson (1996:
381), and even more strongly, of intellectual failure (p. 380). Perhaps the
simplest explanation is that Nanson was merely acting pragmatically, sug-
| 28 |
gesting that Reid and Forrest s categorisation of him as an idealist (1989: 88)
may require some qualification. Certainly his influence over the original draft
of the 1902 Commonwealth Electoral Bill cannot be overstated. This is
evident from the regular references to him by Senator Richard OConnor,
who led the debate on the government side, as well as by many other parlia-
mentarians. But, as the next section shows, Nansons hand can also be seen in
the details of the proposed systems.
The 1902 Commonwealth Electoral Act
I t might have been expected that the design of Australias electoral system
would have featured in the 189198 convention debates over federation.
After all, here was a perfect opportunity for the founding fathers to insert
clauses laying down the electoral systems for both houses, making it difficult
for politicians subsequently to re-fashion the systems in a form more
favourable for themselves. John Uhr (1999: 29) is critical of what he calls the
framers mysterious consideration (or ignorant neglect) of the electoral
system question, which resulted in the wonderfully permissive s. 9 of the
Constitution allowing the parliament to decide on a uniform electoral system
for itself. According to Galligan and Warden (1986: 110), however, the deci-
sion not to incorporate electoral system design was related to the question of
the franchise, which was complicated by the recent decision of South
Australia to allow women the vote, and so, [n]ot wishing to disenfranchise
any section of the electorate, the Convention resolved to leave [election
matters] for the Commonwealth Parliament to determine. More generally,
the constitutional framers were anxious to avoid questions about uniform
electoral systems that would mean settling on electoral systems that would be
bound to upset some of the states, and thereby threaten the ratification of the
Constitution (Crisp 1949: 65).
The proposed section 9 of the Constitution, that the national parliament
would legislate on a uniform electoral procedure, invited some speculation on
what the system should be. For instance, Uhr observes that the scattered
commentary on the Senate electoral system revealed that the framers con-
sidered PR to be an acceptable, possibly even, the preferred, electoral basis of
the upper house (1999: 2122). There were some delegates who argued in
favour of not having uniform electoral procedures on the (ultimately quite
correct) grounds that a future national parliament might legislate against the
| 29 |
Hare system, and that therefore it would be better to let the states decide for
themselves. One delegate expressed the view that Tasmania has a much
better chance of getting the system from her own parliament than from the
federal one.
9
But the bulk of comments made it clear that many delegates
expected the national parliament to select a PR system for Senate elections.
10
Australias first national election of 1901 was held under a mix of different
electoral procedures across the Commonwealth, in which each state elected
its representatives for the new national parliament according to the electoral
rules used to elect its own representatives at state level (Table 2.1; see also
chapter 3). Party politics had yet to consolidate in the new Australian polit-
ical system, and was not to do so until 1910, when the amalgamation of the
Protectionists and Free Traders into the new Liberal party created a conser-
vative party to oppose the already-established Labor Party ( Jaensch 1994;
Loveday 1977). I n the years prior to 1910, then, party politics at the national
level was a complex three-way conflict dominated by Labor, Free Trade and
Protection.
I n 1901, Sir Edmund Barton became Australias first prime minister. His
Protectionist party lacked a clear majority of seats in either house. I t had only
a plurality of the seats in the House of Representatives: 32 seats against 27
for the Free Traders, who were the main opposition party, and who were
implacably opposed to the introduction of preferential electoral systems. The
Federal Labor Party, which had 16 seats in the House, was more agnostic.
The party had agreed to keep an open mind on the matter, and initially
backed the government; however, it is clear that most Labor members were
unsympathetic to STV for the Senate, and therefore their support was fragile
(Crisp 1978: 21417). The situation for the Barton government was far more
tenuous in the Senate, where the Protectionists had just 11 seats, and Labor
eight, against 17 for the Free Traders. I n large part, this poor result in the
Senate had been caused by the use of the block vote system by most states
(with the exception of Tasmania), which had harshly penalised the
Protectionist teams (Graham 1968: 205).
I n early 1902, the Barton government produced its Commonwealth
Electoral Bill, a complex and wide-ranging piece of legislation, containing
218 clauses and a series of annexes. I t endeavoured to cover all things elec-
toral, from the issuing of election writs, through candidate nomination, ballot
paper design, rules of procedure in polling booths and at electoral counts, and
on to spending restrictions on parties. A core feature of this legislation was
| 30 |
the proposed electoral systems for both houses of the Commonwealth parlia-
ment AV for the House and STV for the Senate.
Under the proposed institutional structure for the Senate, each state
would comprise a constituency.
11
I n contrast to Tasmania, which at that time
used the Hare quota, the Bill proposed using the Droop quota. This adapta-
tion was advised by Nanson, who argued that the Hare quota produced per-
verse results, in particular because candidates elected early on had to cross a
higher threshold than candidates elected late, who could be elected without
reaching the quota. While this point may be demonstratively true in social
choice terms (McLean 1996), this did not prevent it from being attacked on
political grounds, not least by supporters of Catherine Helen Spence, as an
inappropriate modification of the Tasmanian system.
Another important feature of the Bill was the proposal to operate a system
of optional preferences, allowing voters to express as many preferences as they
liked, with a vote counting as valid (or formal) providing the voter had
declared at least one preference. This was consistent with the version of pref-
erential voting being proposed for the House of Representatives. I n addition,
voters could strike out the names of those candidates they did not want to
give any support to. The implication of combining optional preference
ranking and the striking out of candidate names was that there could be some
candidates neither ranked nor struck out. Such candidates would be treated
as having equal preferences which tallied with ideas developed by Nanson
in his earlier theoretical work (1995: 354ff.) and, in the schedule to the Bill,
special arrangements were proposed for dealing with this in the counting
process (Commonwealth Electoral Bill (No. 45), second schedule, rules 58
24 January 1902). To avoid the prospect of candidates being elected without
having reached the quota (which could result from there being too many
exhausted ballots), the Bill included the possibility for a supplementary elec-
tion, entailing the re-counting of ballots taking account only of those prefer-
ences for candidates not already elected or eliminated, and re-numbering the
ballot papers accordingly (Commonwealth Electoral Bill (No. 45), second
schedule, rule 3). Finally, the Bill proposed the use of the Gregory method for
the transfer of surpluses.
Unlike the PR system proposed for the Senate, the intention was that the
House should be elected by a majoritarian system, which entailed single-
member constituencies. Lauded as a modification of the Queensland prac-
tice, and an improvement upon it (OConnor, CPD January 31 1902, 9534),
| 31 |
the 1902 Bill proposed the use of AV in which voters would be able to express
a preference for as many of the candidates on the ballot paper as they liked,
with weakest candidates being eliminated individually and their votes trans-
ferred until such time as one candidate would emerge with an overall
majority. I n essence, this is the same system as is used today for electing the
Australian House of Representatives; the major difference was the proposal
that preferential voting should be optional rather than compulsory. Clearly,
Nanson also played a major role in designing this adaptation of Queenslands
contingent vote system. Whereas the Queensland legislators were influenced
by French experience, Nansons interest in AV stemmed from voting theory.
I n the 1870s, a Harvard professor, W. R. Ware, proposed an adaptation of
Hares STV system to the case of single-seat constituencies. Unlike the con-
tingent vote system, under the Ware method (that is, AV) the voters pref-
erences are taken account of one at a time. Despite holding certain theoretical
objections to this system, it was clear to Nanson that AV had clear advantages
over SMP (Nanson 1995), and he also considered it an improvement on the
Queensland system because it took greater account of all the voters prefer-
ences (McLean 1996: 378).
I t had been the intention of the government to start the proceedings in
the House of Representatives, where the Bill was first introduced in June
1901, five days into the first sitting of the new parliament. The balance of
power between the parties in that chamber suggested the government would
have had more chance of getting its way. But owing to the pressure of busi-
ness (CPD June 5 1902, 13353), the government was forced to take it to the
Senate first.
12
I n consequence, the focus of attention was on the clause
dealing with the proposed electoral system for the Senate.
From the outset the Bill received a rough ride in the Senate, and to a very
large degree this was due to the proposed Senate electoral system. Criticisms
were of three main forms: principled objections to PR, objections to the type
of STV system being proposed (including criticisms from Spences sup-
porters), and concerns over the modalities of including this new system as
part of a larger piece of complicated legislation. After three months of delib-
erations, the Bill finally made it through to the committee stage. But by now
it was obvious that the proposed STV system was under threat. Senator
Richard OConnor, who led the debate on the government side, mounted a
half-hearted defence, and as soon as the vote went against him, he announced
that the only alternative was to use the block vote.
| 32 |
One might have expected the issue to end there, and certainly this is the
consensus of those scholars who have examined the parliamentary debate
(Graham 1968; Reid and Forrest 1989; Uhr 1998, 1999). After all, the oppo-
nents of the proposed STV system had won the day, and the government had
stated that it would now use the block system, which was already the
common system for electing senators in most states.
13
Things were not so
simple, however. The government may have lost its PR system, but
OConnor, despite what he said, was not so ready to abandon the aim of pro-
ducing something other than the block vote system. The following day he
presented the Senate with a new clause for what he called a block vote
system, but there was an important twist, and that was that the voter was to
be given the option of voting for fewer candidates than the number of seats
to be filled. Prima facie, this might seem innocuous: it was consistent with the
government desire at all stages in this Bill to make life as easy as possible for
the voters, not forcing them to express preferences for all candidates; and as
OConnor observed, this was already the practice in Queensland, Western
Australia and South Australia.
The fact was, however, that regardless of what the government may have
called this system, they were actually proposing quite a different electoral
system, one that is more normally referred to as the limited vote (Farrell
2001), a system first used in Britain experimentally in 14 constituencies
between 1867 and 1885 (Bogdanor 1981; Lakeman 1974).
14
I n proposing
and defending this system, OConnor developed a clever albeit somewhat
disingenuous argument about the need to avoid forcing voters to vote for a
full slate of candidates when they might prefer to vote for only one or two
candidates. I f a voter in this situation was faced with having to vote for the
full slate in order for the vote to be counted as valid, this might leave them
with no option but to not vote, or to deliberately spoil their vote.
Undoubtedly what was at the heart of this particular debate was the fact
that having performed badly in the Senate election of 1901, which for the
most part involved the use of block voting, the Protectionist government was
anxious to avoid the system in the future. As is well known, in the case of
non-proportional systems, the larger the district magnitude (that is, the more
seats in a constituency) the more disproportional the result (precisely the
reverse relationship exists for proportional systems), and therefore the easier
it is for larger parties (in this case the Free Trade party) to sweep the board in
a block vote system. This can be countered by using the limited vote, in which
allowing the voters fewer votes than the number of seats to be filled reduces
| 33 |
the chance for a larger party to have a full slate of candidates elected, thereby
increasing the chance for smaller parties to pick up some seats. As the system
becomes more limited, the proportion of votes required to win a seat
decreases, thereby making this a semi-proportional system (Engstrom 1998;
Lijphart et al. 1986).
15
Evidently the opposition was keenly aware of what OConnor was trying
to achieve, and much was made of the dangers of allowing voters the right
to plump for one or a few candidates. After some debate, the Senate forced
an amendment requiring voters to cast a ballot for as many candidates as
there are seats to be filled, in other words the block vote system.
16
The issue
was not yet settled, however, because the Bill still had to pass before the
House of Representatives, and it was soon evident that members there
would not let things go that easily. On 23 July, in the House of
Representatives, Littleton Groom proposed an amendment to the clause on
the Senate electoral system, re-introducing the right for voters to plump, by
implication re-introducing the limited vote system. A long debate ensued,
and although William Lyne (the Home Affairs minister who was leading on
behalf of the government in this debate) cautioned (and voted) against the
amendment, it was carried.
I n the ensuing months the Bill ping-ponged between the two houses,
with the principal focus of debate and disagreement being over the Senate
electoral system. I n effect, there was a standoff between the two houses, with
the House backing the limited vote system and the Senate the block system.
Only when it became clear that the entire Bill was in jeopardy did the House
back down, and the block vote system was adopted, though not before some
members had remarked on the fascinating twists and turns of the government
over the course of this parliamentary standoff.
17
I n contrast to the tortured path of the Senate electoral system clause, the
deliberations on the electoral system for the House of Representatives were
short and sweet. I ndeed, a striking difference from the Senate experience was
how quickly and easily the government caved in on the issue of the proposed
electoral system for the House: it was all over in a matter of minutes. I n intro-
ducing the clause at the committee stage, Lyne advanced the rather spurious
argument that having made some back-of-the-envelope calculations a few
days before, he apparently produced some rather astonishing results, and
consequently his faith in the system was very much shaken (CPD July 23
1902, 14613). I nstead, he proposed replacing the system with SMP.
| 34 |
I t seems incredible that the responsible minister would have produced
such a conclusion so late in the day. More likely, the reason for the govern-
ment reversal was that it had lost the support of the Labor members (whose
caucus had decided to formally oppose the AV system) and it therefore saw
no point in trying to force an issue that it was bound to lose (Crisp 1975: 220;
Reid and Forrest 1989: 105). The basic problem was that, since the Senate
had rejected STV, there were fears that adopting AV for the House of
Representatives could lead to voter confusion, particularly in the event of
both elections occurring on the same day (as was expected to happen in
1903). The block vote system for the Senate meant that the voting act would
consist of placing Xs in the relevant boxes on the ballot paper, while AV in
the House would require voters to rank order candidates. Lynes amendment
was carried, without a vote, and SMP was introduced for House of
Representatives elections.
I t is intriguing to think what might have happened if the Commonwealth
Electoral Bill had been brought first to the House of Representatives rather
than the Senate. Attention might have focused on the AV system that was
being proposed for the House, rather than on the STV system proposed for
the Senate. As we have seen, by the time the House first saw the Bill, the die
was cast: STV was lost and AV was lost with it.
There is a range of reasons that help to explain why this first push for
preferential voting systems failed. One major factor was the balance of party
political forces in the new parliament that were ranged against the
Protectionist government, making it clear from the outset that there was
going to be a tough battle. As Graham comments: the government seemed
unlikely to obtain parliamentary approval for its proposed voting methods
(1968: 206). Bartons party did not have a majority of seats in either House,
and the support for the proposed electoral systems that it initially received
from Labor was half-hearted and fragile. Added to the party-political arith-
metic was the fact that many New South Wales and Victorian legislators
(who, between them, comprised two-thirds of the membership of the House
and one-third of the Senate) were unfamiliar with the AV and STV systems,
and were disinclined to try them (Hughes 2000: 161).
Matters were complicated by two additional factors. First, as we have
seen, since the legislative agenda was already very crowded understandable,
given that this was a brand new political system the government was forced
to introduce the Bill in the Senate where it received a rough handling
(Graham 1968: 206). This was not helped by the fact that, as a parliamentary
| 35 |
chamber, the Senate had not yet succumbed to the discipline of parliamen-
tary parties and in consequence the government faced a considerable burden
in managing the bills through the legislative uncertainties of the first
Parliament (Uhr 2000: 11). Second, Bartons party (and, indeed, cabinet) was
divided on the issue, explaining perhaps the lacklustre parliamentary per-
formance of some of the Bills proponents, who Graham suggests, argued
without conviction (1968: 206).
18
A final factor to mention could be summed up as the battle of the elec-
toral system specialists. This is best illustrated by the comment, towards the
end of the second reading of the Bill, of an evidently exasperated Senator
Styles, who remarked: Mr Droop disagrees with Mr Hare, and somebody
comes along and disagrees with Mr Droop. Miss Spence disagrees with them,
and Professor Nanson disagrees with the lot. Then the Messrs Ashworth
sweep the whole lot aside, and say they are all wrong (CPD March 6 1902,
10690). The fact was, of course, that if the specialists were unable to agree on
one version of STV, this had to raise serious questions in the minds of the
legislators over the appropriateness of this relatively untested electoral
system. I n the words of Senator Pulsford, STV stands rather condemned by
the fact that since it was first mooted some 60 years ago, its advocates have
quarrelled one with another as to the different phases in which it should be
drafted, while when it has been adopted by a Legislature, it has been more or
less speedily discarded (CPD February 28 1902, 10496).
The adoption of AV for
the House of Representatives
Australias use of plurality electoral systems (SMP and block) continued for
almost two decades, but the issue of electoral reform never quite disappeared.
I t emerged in backroom negotiations (such as letter exchanges between the
Protectionist and Labor party leaders in 1906), a royal commission report
(191415), parliamentary motions (1909, 1911, 1914) and a government Bill
(1906). The details of all these debates have been well documented already
(Graham 1968; Reid and Forrest 1989; Uhr 2000).
19
The bulk of the initia-
tives were from anti-Labor forces, and the reasons for this are readily
apparent. Labor benefited from having more cohesive party structures than
its competitors. I n particular, early on Labor developed more effective candi-
date nomination (pre-selection) rules than the other parties (Loveday 1977:
| 36 |
394) enabling it to field the correct number of candidates in the different
constituency races. A particular problem with non-proportional electoral
systems is the danger of running too many candidates and splitting the vote
in triangular contests. As Figure 2.1 shows, more often than not, Labor ben-
efited quite considerably from disproportional tendencies in their favour, par-
ticularly in Senate races using the block vote (notably in 1903, 1910, 1913
and 1914).
A problem for the anti-Labor parties after the 1910 election, the anti-
Labor party was that the parliamentary numbers simply were not sufficient
to be able to push through a major overhaul of the electoral systems against
Labor opposition. With the election of the Nationalist government in 1917,
circumstances changed dramatically. The new party had made great gains in
that election, and on this occasion the electoral systems worked in its favour.
For the first time since 1901, a non-Labor government was in a position to
implement a major overhaul of the national electoral systems.
The Nationalist party, and its immediate predecessor, the Liberals, had
been actively promoting preferential systems since 1914 (Reid and Forrest
1989: 115). During the course of the 1917 election, the party called for the
introduction of AV for House of Representative elections, though it was more
muted on the issue of electoral reform for the Senate. I n part this was because
the focus inevitably was on government formation and, therefore, on the need
to remove the problem of triangular contests for House elections. But it also
reflected divisions within the Nationalist party over the merits of PR, which,
| 37 |
0
20
40
60
80
100
%
Seats
Votes
1
9
4
9
1
9
4
6
1
9
4
3
1
9
4
0
1
9
3
7
1
9
3
4
1
9
3
1
1
9
2
8
1
9
2
5
1
9
2
2
1
9
1
9
1
9
1
7
1
9
1
4
1
9
1
3
1
9
1
0
1
9
0
6
1
9
0
3
1
9
0
1
FIGURE 2.1 VOTES AND SEATS FOR THE ALP, 190149 (%)
it was believed, Senate electoral reform would have involved. Some leading
figures were strongly in favour of STV most notable among them Catherine
Helen Spences old friend, Patrick Glynn, who had pushed for STV in the
189798 federal convention, and again in parliamentary motions in 1909 and
1914, and who was now the government minister charged with bringing
forward electoral reform proposals. As events would show, however, this was
not the majority position.
Glynn introduced his Commonwealth Electoral Bill in October 1918
proposing AV for the House of Representatives. The ongoing world war and
the need to include certain technicalities in the legislation (for example,
postal voting) had caused some delay in the drafting. By the time Glynn
introduced the Bill in the House of Representatives, there was an air of
urgency on the government benches. Fresh elections were expected soon, and
the Nationalists were anxious to have the new AV electoral system on the
statute books. I n almost all respects, the proposed new electoral system was
identical to what the Barton government had proposed in 1902. The one big
difference was the requirement of exhaustive preferences; the 1902 bill had
proposed optional preferences. This was consistent with trends in those states
that had already adopted AV systems for their lower house elections (Victoria
and Western Australia, both of which had started with optional preference
systems; see Table 3.2 on p. 50). I t was also consistent with the aim of trying
to deal with the problem of over-nomination, a problem that would be exac-
erbated in the event of voters plumping for one or a few candidates. (The
properties of the AV system are dealt with in the next chapter.)
I t was obvious to everyone that the issue of electoral reform for the House
had been brought to a head as a result of a by-election in the Victoria con-
stituency of Flinders in May 1918. During the by-election, the Nationalist
party struck a deal (dubbed the Flinders deal) to ensure that it faced no com-
petition from another non-Labor candidate representing the Victorian
Farmers Union; and it promised to bring forward legislation for an AV
system in return for the withdrawal of the Farmers candidate. Flinders was
not the only by-election at that time. As the debate over the Electoral Bill
unfolded, a second by-election occurred in the Western Australian con-
stituency of Swan. On this occasion, the Nationalist party was unable to
prevent other non-Labor candidates being nominated and, in a classic trian-
gular contest, Labor won the seat with just 34 percent of the vote. This result,
and the imminence of another by-election (due in December 1918) in the
| 38 |
Corangamite constituency in Victoria, put pressure on the government to
rush through the legislation to avoid another electoral defeat. The govern-
ment applied a guillotine to rush through those parts of the Bill relating to
AV; in the event, however, the Nationalists still lost Corangamite, on this
occasion to a Country party candidate.
I n introducing the legislation, Glynn did not disguise the fact that a major
aim of the change was to deal with the problem of triangular contests, a
problem that was particularly acute for the Nationalist party. The somewhat
nave argument Glynn put forward was that AV would remove the need for
candidate nomination: The party choice is made under this system, not
through the Caucus, but at the same time as the choice at the poll and the
electors there may express their preferences between men of their own party
(CPD October 4 1918, 6678). Apart from the selfish motivation of party
gain, Glynn was at pains throughout the debate to stress the merits of pref-
erential systems, reminding colleagues of his credentials as a long-standing
supporter of such systems.
The parliamentary debate was ill tempered, and the opposition objected
strongly to the government s guillotine: in the Senate, it provoked a 12-hour
filibuster by Senator Gardner. Not only were there criticisms over the
unseemly haste with which the legislation was being rushed through for the
purposes of partisan advantage in an imminent by-election, there were also
the usual phalanx of complaints, such as that the timing was inappropriate
during wartime while the boys were away, or that this legislation distracted
parliamentary attention from more important matters.
20
Much criticism
focused on the danger of having different electoral systems for the two
houses, one requiring voters to mark an X next to their preferred candidates
(for the Senate) and the other requiring them to rank-order all candidates
(for the House). On each occasion such a point was raised, Glynn made clear
his sympathy for the view that the Senate system should also be changed.
But, as he pointed out, the Nationalists election promise in 1917 had referred
only to the Houses system, and it still had to be determined which system
would be most suitable for the Senate.
21
As the debate unfolded, the govern-
ment position was seen to shift, and towards the end of the committee stage,
it was made clear that, in due course and in time for the next election, fresh
legislation would be introduced to change the Senate system.
With the passing of the 1918 legislation, the chapter on the House of
Representatives electoral systems was largely closed. Apart from some
| 39 |
minor amendments, the system has remained virtually unchanged up to
the present, and there has been relatively little serious consideration given
to changing it. As we shall see next, however, the Senates electoral system
would still have to pass through a number of additional evolutionary
stages.
Reform of the Senate electoral system, 191883
The government was true to its word about reforming the Senate system, and
in 1919 a new Electoral Bill was introduced. On this occasion it was even
more apparent that narrow partisan advantage had taken precedence over
electoral system ideals. I nstead of the much-expected push for STV, the
Nationalists proposed a hybrid AV system, marrying features of AV and
block. This preferential block system attracted much criticism and rightful
derision, its critics dubbing it variously as an electoral abortion, a mongrel
proposal, and a blockhead system.
22
The government s response was that the
motivation simply was to resolve the problem of having different electoral
systems for the two houses; in essence, this was designed to do for AV what
block does for SMP that is, operate a majoritarian electoral system in a
context of multi-member constituencies. As the critics were quick to point
out, such a translation to a multi-seat context would have much the same
effect in terms of overall proportionality, making the system even more dis-
proportional than in the case of single-seat constituencies.
The original Bill included two other features. First, voters would be
required to complete a minimum number of preferences (which normally
should amount to seven preferences in a half-dissolution).
23
Second, in a
move to encourage high intra-party preference voting, the legislation pro-
posed that candidates should be grouped by party according to alphabetical
codes rather than using party labels. I n the face of sustained criticism over
this features implications for encouraging party machines, and clearly
wanting to avoid having this feature threaten the success of the legislation,
the government subsequently withdrew the party-grouping proposal at com-
mittee stage. This issue would be re-visited within three years: in the light of
high numbers of invalid votes in the first outing of the preferential block
system in 1919 (8.6 percent, up from 3.9 percent in 1917), the government
introduced new legislation in July 1922 to give effect to the proposal for the
grouping of candidates.
24
| 40 |
The basic idea behind the preferential block system was simple enough.
I t was designed to meet two objectives: first, to allow voters to rank order the
candidates in the same manner as under AV and STV; and second, to ensure
that each candidate was elected with an overall majority. The similarity with
STV was that the count consisted of a series of stages in which votes, based
on next preferences, were transferred from winning candidates or from the
exclusion of the candidate with the least number of votes. The two impor-
tant differences from STV were, first, the electoral quota (which was a
majority formula)
25
and, second, the transfer of votes at full value.
26
I n effect,
this produced a series of mini-elections, one for the election of each candi-
date. All the votes of the winning candidate were transferred to the
remaining candidates, based on the next preferences, and then there was a
fresh count to see which of the remaining candidates had an overall majority
of the vote. The process continued until the required numbers of candidates
were elected.
Needless to say, the preferential block system had some unusual proper-
ties. As we saw in chapter 1, social choice theorists criticise AV and STV for,
among other things, their inconsistent treatment of preferences (not all are
taken account of, some are taken account of more often than others) and for
giving all preferences the same value (e.g. Dummett 1997). Both points
apply with even greater force in the case of the preferential block system. By
the nature of how it operated (particularly once the grouping of candidates
under party codes was introduced in 1922), this system ensured that, for sup-
porters of minority parties, there was very little likelihood of anything but
the first preference being taken into account; by contrast, supporters of
larger parties (particularly of the large party on a winning streak) saw their
full gamut of preferences being taken into account. And when preferences
were taken into account in the preferential block system, this meant that
they were transferred at their full value: in effect, some voters got to vote
more than once in an election.
This electoral system ensured the institutionalisation of the windscreen
wiper effect, where one party wins all or most Senate seats in one election,
only to lose all or most in the next election. The effect had already been in
evidence under the block system. Elections in the following thirty years were
to produce dramatic shifts in representation in the Senate based on relatively
small moves in voting support (Figure 2.1 shows the trends for the ALP).
Apart from two minor changes, the preferential block system was to
| 41 |
remain in force for thirty years. The first of these changes has already been
mentioned (regarding the grouping of candidates on the ballot paper); the
second was in 1934, when full compulsory (exhaustive) preferences were
introduced, to bring the Senate system into line with the procedures for
House elections. At various stages during that period there were calls to
replace this system with STV.
27
When the change finally came, the principal
motivation for it was the need to take account of the fact that the size of the
Senate was to be increased.
The background to the 1948 reforms has already been covered in some
detail (most notably in Uhr 1999), and so only a brief summary is given here.
At issue was the need, for the first time since federation, to increase the size
of the House of Representatives, which meant that the size of the Senate
would also have to be increased.
28
I t was in this context that the Chifley gov-
ernment also gave thought to changing the Senate system. The government
feared that the windscreen wiper effect referred to above would be even
starker in an enlarged Senate, and might threaten its credibility (Crisp 1978:
219). I nevitably, there was also a good short-term reason for favouring elec-
toral reform. Chifleys term of office was coming to an end and it was
expected that Labor would lose its majority in both houses. By changing to
STV, while at the same time resisting moves for a double dissolution, there
was a good chance that Labor would be able to maintain control of the Senate
(Uhr 1999).
The 1948 Bill proposed a variant of STV that owed rather more to
Catherine Helen Spence (or I nglis Clark) than to Edward Nanson. I n
essence it was a copy of the Tasmanian system (including the use of the
Droop quota), but there were two big differences. First, instead of the
Gregory method for transferring surplus votes, it was proposed to use random
transfers at full value. The government rationalised this on the grounds of
administrative convenience and argued that its statistical effects were likely to
be negligible (CPD, April 16 1948, 966ff ). The second big difference from
Tasmanian practice was the requirement of full expression of preferences,
which was consistent with the 1934 reform and the stated aim of ensuring
close correspondence between the AV and STV systems. Also consistent with
earlier reforms in this case, 1922 was the proposal that candidates should
continue to be grouped on the ballot paper.
Despite opposition objections to these two deviations from Tasmanian
practice, the Bill passed through all its stages without difficulty. But these two
| 42 |
features were to dominate the next stage of the evolution of the Senate system.
John Uhr observes that the 1948 decision was the final stage in a fre-
quently-deferred plan of parliamentary reform that goes back to Federation
(2000: 17) in the sense that it saw the introduction of STV but this was by
no means the final chapter in the evolution of the Senate system.
Apart from some minor administrative changes in the 1960s and 1970s,
and some failed attempts at more ambitious reforms during the Whitlam
period in the early 1970s, the Senate system remained untouched for some 35
years. Change, when it came, was heralded by the election of the Hawke
Labor government in 1983, which promoted a reformist agenda; the change
also coincided, yet again, with the need to increase the size of both houses.
Within months of coming into office, the Labor government established a
Joint Select Committee on Electoral Reform ( JSCER), giving it a brief to
examine a range of matters relating to the administration of elections in
Australia. The JSCERs influential report, published in September 1983, pro-
posed a series of reforms relating to such matters as the funding of parties,
the redrawing of constituency boundaries, the establishment of an electoral
commission, party registration, ballot paper design and the operation of the
Senate STV system.
I n the case of the operation of the system, there had been particular con-
cerns about the high levels of invalid (or informal) votes, due in large part to
the complexities of the Senate ballot paper and the requirement to express
preferences for all candidates, which, in the words of one member of the
House, forced voters to choose between Senate candidates from the flat earth
society and the radical bomb throwers and collectors, or lose his or her entire
vote (Child, CPD, November 9, 1983, 2534). Other contributing factors
included the differences in the state and national electoral systems and the
frequency with which elections were carried out. Under the system of com-
pulsory voting, this meant that some 95 percent of registered voters had to
attend a polling place about once every 18 months. Although levels of invalid
voting had always been high for Senate elections (McAllister and Makkai,
1993), in elections in the 1970s invalid votes constituted about one in ten of
all the votes cast (see also chapter 5, pp. 12526).
The JSCERs report provided a major impetus for change, not least
because for the first time it represented a genuine bipartisan attempt at
implementing far-reaching electoral reform. Within two months of its pub-
lication, the Hawke government produced its ambitious electoral reform
| 43 |
bill. Covering 187 pages and including 156 clauses, this represented the
most ambi ti ous effort at major electoral reform si nce the 1902
Commonwealth Electoral Act, and while there was plenty of debate and
many of the usual accusations of government opportunism, the delibera-
tions of the JSCER that had preceded this legislation and some skilful
exploitation of divisions in the opposition ensured its successful passage
through both houses.
The 1983 legislation introduced four main changes to the national elec-
toral systems, the bulk of these affecting the Senate system only.
29
A major
motivation behind three of these changes was to try to reduce the proportion
of invalid votes by making the ballot paper more voter-friendly and by
amending the rules on the minimum numbers of preferences required. First,
the letter codes for party groups were supplemented by actual party names
(and replaced by them, in further reforms, a year later). Second, the rules
regarding the compulsory expression of preferences were relaxed: from now
on, ballot papers with at least 90 percent of the preferences correctly com-
pleted would be accepted as valid.
The third change to the Senate electoral system was by far the most
radical. Borrowing from the recent experiences in South Australia, which
used an adapted form of PR list system in the 1970s (discussed in chapter
3), the Senate ballot paper was adapted to give voters the option of
expressing just one preference for a party ticket instead of the laborious task
of having to rank-order all the candidates on the ballot paper. The details
and implications of this new ticket voting system are discussed below (see
chapters 3 and 6), but in essence, while undoubtedly it made the act of
voting in Senate elections simpler, it also served the purpose of transforming
the existing STV system into a form of closed-list system, in which the
rank-ordering of the candidates was determined by the political party rather
than by the voters.
The fourth alteration to the STV system related to the procedures for
counting votes, and the most significant change here was the introduction of
a modified version of the Gregory method for transferring surplus votes. The
details of this are discussed in chapter 3 and its implications in chapter 4, but
at this juncture it is useful to note how little attention was given to this reform
in the parliamentary debate: in essence it was simply nodded through without
demur, despite the best efforts of the Proportional Representation Society of
Australia to warn of potential problems with this modified version of the
Gregory method.
| 44 |
Conclusion: accident and
intent in electoral system design
This chapters historical overview of electoral system design in Australia pro-
vides some important examples of the often accidental nature of institutional
design (Goot 1985). Nowhere is this better demonstrated than in 1902, when
AV might have stood a better chance of being introduced if the legislation
had gone before the House first, rather than the Senate. Certainly, there was
far less discussion about it than about STV. Furthermore, the relative ease
with which AV was introduced in 1918 (when, in effect, it was taken first )
tends to support this interpretation. I ndeed, had AV been successfully intro-
duced in 1902, it is interesting to speculate whether STV would also have
been successful then.
Events also play an important role in explaining why the Australian
political establishment has always attached such weight to preferential
systems whenever considering alternatives to the plurality systems. The rel-
evant point here is that preferential systems featured prominently in the
scholarly and activist debates over electoral system design at the turn of the
nineteenth and twentieth centuries, around the time when Australian leg-
islators first began to consider issues. The elite debate froze around plu-
rality and preferential systems as the two main alternatives, and in
consequence, the other families of electoral systems (notably the PR list
systems) tended to be overlooked.
30
Political events and the force of circumstance explain a great deal in the
origins and evolution of Australias electoral systems, but responsibility must
also be given to the deliberate efforts and clear intentions of the designers.
The year 1902 represented a significant milestone in this regard (Graham
1968; Uhr 1999, 2000), though only to a degree. The electoral systems being
proposed for the Commonwealth in 1902, and some of those already adopted
in the states, were ahead of their time; nothing like them had been proposed
anywhere else before nor since, in the case of the STV system then being
proposed for the Australian Senate. I t is certainly true that the AV system
that was adopted in 1918 owed much more to the intentions of the electoral
system designers of 1902 (notably Edward Nanson) than to any other influ-
ence. But the picture is not so clear with regard to STV, for the version finally
adopted in 1948 bore very little resemblance to the version proposed in 1902.
I ndeed, it owed more to experiments in Tasmania in the first half of the
| 45 |
twentieth century than it did to the theoretically informed designs of Edward
Nanson. I n this regard, therefore, it can be argued that Spence and Clark
whose influence over the 1902 debate was less than Nansons had the last
word on Australian STV design (with a final South Australian-inspired twist
in 1983 when ticket voting was introduced).
A final point to note about the 1902 debate, when compared with the sub-
sequent parliamentary debates over electoral system design, was just how
sophisticated was the level of argument and how well informed the leading
politicians in 1902 were compared to later episodes. Not only did electoral
system experts (Clark, Spence, Nanson) play a significant role in this first
debate, it was apparent that many of the legislators had read and assimilated
much of the detail of their arguments. By contrast, in the debates of 191819,
1948 and 1983, there were few if any references to the views of electoral
system experts, and for that matter there was little reference to the 1902
debate. The legislators seemed far less well informed about the details of the
electoral systems being debated over. The one qualification to this point was
the role of the Joint Select Committee on Electoral Reform in 1983. I n this
instance much of the difficult work on electoral system design had been
carried out before the legislation came before the parliament.
The year 1902 marked a very important step in the design of Australias
electoral systems. Once the debate had crystallised around the relative merits
of preferential systems, it was only a matter of time before they would be
introduced for national elections. This episode is central to understanding
both why preferential systems have become such a prominent feature of
Australias electoral institutions and also why they took so long to become
established. Having been knocked back at this critical juncture at a time of
nation and institution building an important opportunity to experiment
with alternative electoral systems was missed, and it would take several gen-
erations before Australian politicians would finally embrace them.
Our historical overview of debates and events at national level has demon-
strated a strong penchant for electoral system experimentation in Australia,
making this a fascinating case in the study of institutional design. So far our
attention has been primarily on developments at national level. Once we
broaden our enquiry to include events at state level, the picture looks even
more interesting.
| 46 |
t h r e e
Australias contemporary
electoral systems
I n line with the idea of the democratic laboratory, Australian politicians
have devoted considerable effort to refining the countrys electoral systems at
all levels of government. The last chapter traced a busy century of electoral
system experimentation in the Australian Commonwealth. I t is now neces-
sary to move away from a chronology of the main events and to provide,
instead, a more systematic analysis of contemporary Australian electoral
systems. At this juncture, it is useful to also widen the scope of our analysis:
up to this point we have been dealing predominantly with the electoral
systems used at the national level. I n this chapter, our focus widens to con-
sider the entire constellation of electoral systems used in Australia at national,
state and territory level.
At least in theory, the current pattern of electoral systems at national and
state/territory level is simple enough. As Table 3.1 shows, parliamentarians in
todays Australia are elected under one of two systems. For the most part
though not always AV is used to elect lower house parliamentarians, while
STV tends to be used for upper house elections. As a result of Victorias
recent reform, the sole exception is Tasmania. Behind this simple fact,
however, lies a rich tapestry of variation in the versions of AV and STV that
are employed. Furthermore, the table does not to do justice to the range of
other electoral systems that have been experimented with over the years.
1
This chapter develops this theme of electoral system variation. We start,
in the first section, with a brief historical overview of how Australias electoral
| 47 |
systems have evolved over the last century, providing some additional state-
wide colour to the history of national-level trends painted in the last chapter.
The second section focuses on the AV system, examining the range of varia-
tion in its different versions. I n similar fashion, the third section explores the
main variants of STV, in this instance widening the discussion to take
account of the two other cases in which STV is used on a national scale
Ireland and Malta.
I t would be a mistake to conclude from this study that Australia has only
ever considered preferential systems as the alternative to plurality voting. The
fourth section reviews those instances of electoral reform in which other
systems, supposedly non-preferential, were considered. However, as we shall
see, even in these cases, aspects of preferential voting still feature prominently,
demonstrating yet again the central importance of this form of voting to rep-
resentative politics in Australia.
I n short, this chapter seeks to answer three sets of questions:
What is the constellation of electoral systems in use, at national state and
territory level, across Australia?
I n what ways do the AV and STV systems vary?
What other electoral systems have featured in Australia?
But the principal function of this chapter is to set the scene for our exami-
nation of the consequences of preferential systems in the chapters that
| 48 |
TABLE 3.1 THE ELECTORAL SYSTEMS OF AUSTRALIA
Lower Houses Upper Houses
AV Commonwealth House of Representatives
NSW Legislative Assembly (LA)
Victoria LA
Queensland LA*
South Australia House of Assembly (HA)
Western Australia LA
Northern Territory LA*
Tasmania Legislative Council (LC)
STV Tasmania HA
ACT LA*
Commonwealth Senate
New South Wales LC
South Australia LC
Victoria LC
Western Australia LC
*Unicameral parliament.
follow. By extending our analysis to include the systems at state and territory
level, we are able to increase our number of cases and to also provide the
basis for examining how variations in the detail of AV and STV systems can
impact on voters, on politicians and on the systemic properties of these
systems. Given the exchange of ideas between all Australian legislatures,
studying state electoral systems provides an outline of the repertoire of elec-
toral changes available to reformers throughout the federation. I nevitably,
some of the material in this chapter strays into the technicalities of electoral
system design, but wherever possible the more esoteric details are consigned
to footnotes.
Australias evolving electoral systems
I f we take two snapshots in time, a hundred years apart, we can see just how
long (and in some cases, winding) a road Australias states have travelled on
their route towards the adoption of preferential electoral systems. I n 1900,
with the exception of Tasmania (which used single member plurality or
SMP), all states that had elections to the upper house used block plurality
systems. (At this stage, all were operating with restricted suffrage; in New
South Wales and Queensland the upper house was appointed.) For lower
house elections, most states used some combination of SMP and block, with
two exceptions: Queensland, which used an AV system known as the contin-
gent vote, and Tasmania, which used a mix of SMP and STV.
2
One hundred
years later, this situation has been transformed: now every state and territory
uses a preferential voting system for both houses and all upper houses are
directly elected (except Queenslands, which was abolished in 1921).
The shifts in electoral systems and the principal dates of change are shown
in Table 3.2. For lower house elections, three main patterns are apparent.
First, there are the cases of Tasmania and the ACT, each of which experi-
mented with a non-preferential system before settling on STV. I n the case of
Tasmania, there was a period of almost twenty years (18901907) in which
the colony/states electoral system evolved from a mixture of plurality systems
towards STV; in the case of the ACT, there was a brief flirtation with a list
electoral system in 19891992 (see pp. 7173), before the territory voted in a
referendum for STV and used it first in 1995.
The second and most common pattern in Table 3.2 is provided, in
chronological order, by Western Australia, Victoria, South Australia and the
| 49 |
| Austral i as contemporary el ectoral systems |
| 50 |
Upper houses
CorrorWea||| NeW 3ou|| wa|es 3ou|| Aus|ra||a
1901 8|oc| (3Tv |r Tasrar|a) Appo|r|ed 8|oc|
1903 8|oc|
1909
1919 Prelerer||a| 8|oc|
1921
1931 lrd|rec| e|ec||or (3Tv)
1919 3Tv
191
19Z5 Vod|led LR-0roop
19Z8 3Tv (0P)
1981 3Tv
1983 3Tv W||| ||c|e| vo||rg
198Z
1988 3Tv W||| ||c|e| vo||rg
1999 0p||ora| prelerer||a| ||c|e|
vo||rg
2003
N0TE8 3VP = 3|rg|e Veroer P|ura|||y, 3Tv = 3|rg|e Trarslerao|e vo|e, Av = A||erra||ve vo|e, Cv = Cor||rger| vo|e,
CP = Corpu|sory Prelererces, 0P = 0p||ora| Prelererces. T|e da|es re|a|e |o W|er ||e re|evar| |eg|s|a||or Was erac|ed,
W||c| |s |yp|ca||y up |o |Wo years oelore ||e reW sys|er |s used |r ar e|ec||or.
a TWo-sea| cors|||uerc|es, W||| ore reroer oe|rg e|ec|ed per cors|||uercy every ||ree years.
Lower houses
CorrorWea||| ACT NeW 3ou|| wa|es 0ueers|ard
1901 3|a|es' sys|ers r.a. 3VP Av (Cv)
1903 3VP
190Z
1910 Vajor||y-ruroll
1911
191
1918 Av (CP) 3Tv
192 Av (Cv)
1928 Av (CP)
193
1912 3VP
192 Av (CP)
1980 Av (0P)
1989 Vod|led d'lord|
1992 Av (0P)
1995 3Tv
TALE 3.2 TlE Ev0LuTl0N 0F ELECT0RAL 3Y3TEV3 ACR033 Au3TRALlA, 1901-2003
| 51 |
Tasrar|a v|c|or|a wes|err Aus|ra||a
3VP 8|oc| 8|oc|
Av (CP)
Av (CP)
3VP
a
3Tv
3Tv
Nor||err Terr||ory 3ou|| Aus|ra||a Tasrar|a v|c|or|a wes|err Aus|ra||a
r.a. 8|oc| 3VP 3VP/8|oc| 3VP
3VP
3Tv Av (0P)
Av (0P) Av (CP)
Av (CP)
Av (CP)
Av (CP)
Northern Territory, all moving from plurality to AV. I n each case, the system
adopted was compulsory preferential voting, in which, as we shall see below,
the voters are required to complete all or most preferences on the ballot paper
though Western Australia and Victoria briefly operated optional preferen-
tial voting, in which the voter was required to express only one preference.
The third pattern is specific to Queensland and New South Wales (see
also Table 3.3). Both states made use of all three main forms of AV: first con-
tingent voting (Queensland 18921942; New South Wales 192628); then
compulsory preferential voting (Queensland 196292; New South Wales
192880); and now optional preferential voting (Queensland since 1992;
New South Wales since 1980).
3
En route, New South Wales also flirted with
a majority-runoff system (1910) and STV (1918), in the latter case being
second only to Tasmania in adopting STV (clearly, Spence and her followers
had a lot of influence here; see Wright 1980: 121).
4
As indicated in Table 3.2, the upper houses were not popularly elected
with a broad franchise until later, in some cases much later (notably New
South Waless Legislative Assembly which was not directly elected until
1978) (Stone 2002). This, plus the fact that there are fewer cases (because the
two territories and Queensland since 1921 are unicameral) mean that the
picture is much simpler. There are basically two patterns: Tasmania uses AV,
whereas in the other four cases (New South Wales, Western Australia, South
Australia, and most recently Victoria) the trend has been towards STV. Here,
as we shall see in the fourth section, the interesting case is South Australia,
which briefly experimented with a list electoral system in the 1970s.
The alternative vote and its variants
The AV system is quintessentially Australian, where the tendency is to refer
to it as preferential voting. I t is used in Ireland (for presidential elections and
parliamentary by-elections) and Sri Lanka (for presidential elections), and at
various occasions at local or regional level in parts of the USA, Canada, and
the UK. I n the USA it is referred to as instant runoff voting; in the UK it is
used to elect the London mayor, and is referred to as the supplementary vote.
The only other significant sign of this system taking root outside of Australia
has been in fledging democracies in the Oceania region, suggesting strong
indications of diffusion of best practice by Australian international non-
government organisations (Reilly 1997a; Reynolds and Reilly 1997). I n
| 52 |
effect, then, despite its entrenched character in Australia, few other nations
have adopted AV.
AV is a non-proportional system: what distinguishes it from plurality
electoral systems is the expectation that, to be elected, a candidate should
have an overall majority of the vote (Farrell 2001). For the most part (and cer-
tainly in all the Australian cases today), the system is characterised by single-
member constituencies (or electorates), but this need not be the case. There
have been various occasions throughout the evolution of Australias electoral
systems when AV was applied in multi-member constituencies the most
bizarre case being the system used to elect the Senate from 1919 to 1948, as
discussed in chapter 2.
| 53 |
TABLE 3.3 VARIANTS OF AV IN AUSTRALIA
Compulsory
Preferential Voting
Optional Preferential
Voting
Contingent Vote
I. Preferences Minimum = # candidates Minimum = 1 Minimum = 1
II. Count procedures
in the event that no
candidate has more than
50% of the vote after the
rst count
Eliminate the candidate
with the fewest votes and
transfer ballots based on
next preference. Continue
process until one
candidate emerges with
overall majority
Eliminate the candidate
with the fewest votes and
transfer ballots based on
next preference. Continue
process until one
candidate emerges with
overall majority
Candidates with the
highest and second
highest vote totals remain
in the race. All other
candidates are eliminated
and their ballots
transferred based on next
preferences
Cases where used in
Australia today
Commonwealth LH 1918
Northern Territory LH
1980
South Australia LH 1936
Tasmania UH 1909
a
Victoria LH 1916
Western Australia LH
1911
a
New South Wales LH
1980
Queensland LH 1992
Cases where used in
Australia in the past
New South Wales LH
192880
Queensland LH 196292
Victoria UH 19212003
Victoria LH 191116
Western Australia LH
190711
New South Wales LH
192628
Queensland LH 1892
1942
NOTES LH = Lower House; UH = Upper House.
a Minimum number of preferences = 3.
SOURCES Table 3.2; Electoral Council of Australia (www.eca.gov.au); relevant electoral ofce and parliamentary web sites,
and the following legislative Acts: Commonwealth Electoral Act 1918; New South Wales Constitution Act 1902,
Parliamentary Electorates and Elections Act 1912; South Australia Electoral Act 1985, Electoral Regulations 1997;
Western Australia Electoral Act 1907; ACT Electoral Act 1992; Tasmania Electoral Act 1985.
The proportional consequences of AV are dealt with in the next chapter.
At this point, it is important to outline the ways in which the systems fea-
tures can be varied, adapting it to particular circumstances. Over the years, as
summarised in Table 3.3, AV has come in three main forms: compulsory pref-
erential voting, optional preferential voting and the now defunct (at least in
Australia) contingent vote.
5
These forms are based on one of two main char-
acteristics.
6
First, AV systems can vary with regard to counting procedures, although
in Australia this point has only historical significance given the demise of the
contingent vote (see p. 25). The compulsory and optional preferential systems
share the practice of eliminating, in step-wise fashion, the candidates with the
fewest votes and transferring their ballot papers to the remaining candidates
according to the next preference expressed by the voters. This process con-
tinues until one candidate emerges with a majority of the vote. The contin-
gent vote, by contrast, collapses the process into two main stages, in which (in
a single-seat constituency) the candidates with the highest and second
highest vote tallies remain in the race and the ballot papers of all the other
candidates are transferred to one or other of them based on the next prefer-
ence expressed by the voters. Since only two candidates remain in the race,
the result must be majoritarian (that is, a candidate wins with at least 50
percent of the vote).
I n effect the contingent vote system is a collapsed version of the runoff
system, as we saw in the previous chapter; it was first introduced into
Queensland in 1892 and was used there until the 1940s. The only other state
to experiment (briefly) with the contingent vote was New South Wales,
which used it once in 1927. I t has since disappeared in Australia: most
recently, it emerged in the UK, where in the 1990s it was given the title the
supplementary vote (its proponents were unaware of the prior existence of
the contingent vote; see Reilly 1997), and was subsequently adopted as the
system for electing the London mayor in 2000 (Farrell 2001a).
The second variation in AV systems, currently more significant, relates to
a distinction between a requirement that voters must express just one prefer-
ence (contingent vote and optional preferential voting) or as many prefer-
ences as there are candidates (compulsory preferential voting).
7
As Table 3.3
shows, there is a tendency for Australian electoral engineers to favour systems
that require voters to complete all (or certainly most) preferences: as we shall
see in the next section, this is consistent with the trend in the Australian STV
systems.
| 54 |
Optional preferential voting is currently in use in just two states, New
South Wales and Queensland. I n the past, it was used briefly by Victoria and
Western Australia. I t was introduced by a Labor government in New South
Wales in 1980 (where it already applied for upper house elections); in
Queensland, its introduction in 1992 was due to the recommendations of a
government commission that expressed strong objections to the compulsory
expression of preferences, which required voters to express preferences for
candidates they neither knew nor supported. But for both of these states, the
likely partisan gains of the governing Labor party was the critical factor, and
the ability to campaign for just vote 1 has the potential to turn optional pref-
erential voting into something close to SMP (Wanna 2003: 96).
Outside of New South Wales and Queensland, the tendency is to favour
the compulsory expression of preferences. This peculiarly Australian prac-
tice has been seen as reflecting a general political culture that favours regu-
lation and efficiency, and an emphasis on citizens duty over and above
individual freedom and choice (McAllister 2002). But the dominant expla-
nation is the partisan gain accruing to the large parties, particularly the
Liberal Party and its predecessors. There is also the view that the compul-
sory expression of preferences is a natural corollary to compulsory voting,
and for both parties, compulsory preferences helps to shore up their final
two-party preferred vote. State Labor governments have rarely controlled
majorities in their upper houses, so the ability to remove preferential voting
has been difficult: it is no accident that Queenslands reversion to SMP and
easy adoption of optional preferential voting has occurred in a unicameral
system. An additional explanation, offered by Reilly and Maley, is that
compulsory expression of preferences reflects a sense on the part of
Australian legislators that it helps to reinforce the system of compulsory
turnout, for [ i] f it were to be conceded that voters have the right to be
indifferent in regard to a subset of candidates, it would seem to follow that
voters have the right to be indifferent in regard to all candidates (Reilly and
Maley 2000: 44; see also pp. 5768).
The compulsory expression of preferences helps to facilitate the virtual
institutionalisation of the Liberal-National coalition, avoiding the dangers
(for them) of vote splitting in three-cornered contests, allowing both parties
to field candidates in the same constituency, and increasing the likelihood
that one or other of them will succeed in having a candidate elected. While
these trends tend to produce a greater sympathy for optional preferential
voting among ALP politicians, this is by no means universal. For instance,
| 55 |
after a new Western Australian Labor government was elected in 2001, the
Liberal Party introduced proposals for a switch to optional preferential voting
for future state elections. This move in part reflected the growing difficulty
the larger parties have in controlling the full spread of preferences at a time
when, as we see in the next chapter, the preferences are counting more and
more towards the final result.
Optional versus compulsory preferential voting has also featured promi-
nently in debates at national level, with attention centred on the effects of
compulsory voting on smaller parties. There was a range of complaints after
the 1998 election over the degree to which compulsory preferential voting is
used by the larger parties as a means of discriminating against smaller parties.
I n particular, attention was focused on the fact that Pauline Hansons One
Nation Party failed to win any seats in the House of Representatives, despite
attracting 8.5 percent of the national vote. I n a wide-ranging review of the
evidence, the Joint Standing Committee on Electoral Matters received a large
number of submissions from One Nation supporters, complaining that the
larger parties had ganged up on One Nation candidates: particular attention
was focused on the result in the Queensland electorate of Blair in which party
leader Pauline Hanson was defeated despite having received a substantial first
preference vote. But the Committee was not persuaded of the arguments in
favour of a shift to optional preferential voting, believing that there is a
strong chance that an optional preferential system will eventually lead to
voters casting only one preference as the realisation sinks in to the voters that
to indicate second and subsequent preferences will decrease the possibility
that their most preferred candidate will win ( Joint Standing Committee on
Electoral Matters 2000: 113).
8
One Nation is not the only critic of compulsory preferential voting, nor,
for that matter, has it been the most prominent. That accolade is reserved for
the Melbourne-based activist, Albert Langer, who launched his own personal
crusade for optional preferential voting in the early 1990s, earning him a
period of imprisonment in the process, as well as the naming of a form of
(mis)voting in his honour Langer-style voting and causing an amend-
ment to the Commonwealth Electoral Act in 1998. The background was the
large-scale review of the Commonwealth Electoral Act in 1983 (as outlined
in chapter 2), which among other things saw the introduction of an amend-
ment to section 270, allowing for non-consecutive numbering errors. I n the
light of trends in the subsequent 1987 and 1990 elections, in which some
voters cast votes in such a way as to result in optional preferential voting (for
| 56 |
example, by casting the votes in the following sequence: 1, 2, 2, 2, 2, etc.),
the Electoral Act was amended again in 1992. A new clause, section 329a,
was introduced, making it an offence to encourage such voting.
I n the subsequent 1993 election, Langer mounted a campaign to exploit
section 270. Efforts by the Australian Electoral Commission to prevent him
promoting his views resulted in a High Court case in which Langer argued
that the 1992 amendment to introduce section 329a was an unconstitutional
move by the established parties to discriminate against the smaller parties.
The matter was finally resolved in 1996, on the eve of the election, with the
court finding the new clause constitutional. Undeterred, Langer continued
his campaign, resulting in his imprisonment and his naming by Amnesty
I nternational as Australias first prisoner of conscience. The outcome was a
recommendation by the Joint Standing Committee on Electoral Matters, in
its 1996 report, to remove section 329 and to insert in its place a new sub-
section stating that for a vote to be valid the numbers must be consecutive,
without the repetition of any number. This amendment came into force just
in time for the 1998 election. The Langer episode shows how important
compulsory preferences is to the main political parties, and how far they will
go to ensure that it remains in place.
The single transferable vote and its variants
Two decades ago, Lijphart and Grofman (1984: 6) referred to the irony that,
despite the fact that STV is such a popular system among many academics
(on this, see most recently Bowler et al. 2005) and PR enthusiasts, it is so
rarely used. This observation was even more starkly borne out a decade later
when, with the temporary exception of Estonia, none of the new democracies
of Central and Eastern Europe and the former Soviet Union adopted STV as
their electoral system; many opted instead for the more fashionable mixed-
member systems (Shugart and Wattenberg 2001). The situation in the early
years of the twenty-first century, therefore, remains little changed from that
of the 1920s: apart from some cases of regional elections (for example
Northern Ireland) or local ones (for example Massachusetts), STV is used for
national-level elections in just three countries: Australia, Ireland and Malta.
Only in the latter two cases is it actually used for lower house elections at the
national level. As a consequence of its lack of widespread use, and the fact
that the three countries where it is used are small in population terms, STV
tends to be sidelined as a somewhat exotic system that, despite its apparent
| 57 |
advantages (though, as we saw in chapter 1, these are open to dispute), is not
perhaps suitable for larger political communities.
STV also has a tendency to be misunderstood: scholars often draw con-
clusions based on a scrutiny of just one system, Ireland. This results in some
distortions in the conclusions about the place of STV among the different
families of electoral systems for instance, scholars refer to its semi-propor-
tional tendencies (Katz 1984; Taagepera and Shugart 1989: 207) and its
potential to promote intra-party factionalism and excessive attention by
politicians to localist, particularistic concerns (Chubb 1963; Sacks 1977). I t
also tends to give the impression that STV is a single, universal system. These
concerns are addressed in later chapters.
The theme of this section is adaptability. I t goes without saying that, to
varying degrees, all electoral systems are adaptable to particular circum-
stances. Plurality and majority systems can vary the size of their electoral con-
stituencies or regions (referred to generally by the term district magnitude).
PR list systems can do the same and can also vary the electoral formulas (that
is, the means of determining which candidates get elected) and the ballot
structure (offering various degrees of flexibility to voters). Mixed-member
systems consisting, as they do, of a mix of plurality and list systems can
offer the same range of choices as list systems, with the additional feature of
direct constituency representation by individual MPs. I n consequence, to
some scholars at least, they represent the best of both worlds (Shugart and
Wattenberg 2001). STV is no different from the other electoral systems in
this regard: it too can vary in its operation (and effects), and arguably its
potential for variation is even greater than for many other electoral systems.
The first point to be clarified is the name of this electoral system. While
here we refer to it as STV, its precise title has varied in the countries where it
is used. I n the early parliamentary debates in the UK it tended to be called
the Hare system, after the English barrister, Thomas Hare (180691), who,
together with Carl Andrae, is credited most with devising it in the 1850s
(Hart 1992). Today in Ireland and the UK it tends to be referred to as STV,
in the USA as choice voting, and in Malta as the Hare-Droop system.
Matters are somewhat more confused in Australia. We have already seen in
chapter 2 how, in the debates at the start of the twentieth century, there were
some differences of opinion not only over the mechanics of the system but
also over its name, with I nglis Clark, Catherine Helen Spence and Edward
Nanson each having some input into the debate. The situation is no simpler
today. I n Tasmania which was the first state to adopt this system it is
| 58 |
referred to as Hare-Clark (the latter after I nglis Clark). The system used for
Senate elections is generally called the Senate system or the PR system, and
in the other places that use it the common practice is to refer to the system
as modified Hare-Clark or STV.
To a degree, the differences of opinion over the name of this electoral
system in early 1900s Australia may have had less to do with real differences
between its variants and more to do with personal pride if not necessarily
on the part of Clark or Spence, then certainly of their supporters. Chapter 2
showed how there were really very few differences between the Spence
system and the original form of the Clark system. Nansons proposals, which
formed the basis for the Senate system promoted in the 1902 Electoral Bill,
were certainly very different, and indeed were arguably of such an ambitious
and complex form that, to date, nobody has sought to adopt them. I n this
sense, therefore, Nansons version of STV can be treated as being of little
more than theoretical interest. Of far more practical interest are the actual
degrees of variation in the versions of STV in use today. For, while these vari-
ants all have certain key characteristics in common notably they are propor-
tional systems in which electors vote preferentially for individual politicians
they also have their differences, and in some cases these are quite distinct.
Table 3.4 outlines the five main dimensions of variation in the forms of STV
used today in Australia:
9
for the sake of full comparison, the table also
includes details on Ireland and Malta.
10
The first, and most obvious, form of variation is with regard to number of
seats in a constituency, or district magnitude (characteristic I ), which, it is
generally accepted, is the most important feature for determining the overall
proportionality of an electoral system (Lijphart 1994). As a general rule of
thumb (see also Taagepera and Shugart 1989), district magnitude needs to be
at least five members to ensure a proportional result. I n this regard, therefore,
Ireland stands out as having a number of constituencies (usually the majority)
electing fewer than five politicians. I n Australia, the only cases to fall below
this informal threshold are the two territories in the Senate voting system,
which elect just two senators each (making voting almost a meaningless
issue). The range of district magnitudes across Australia suggests consider-
able scope for variation in the degree of proportionality that will be produced.
This question is explored in the next chapter.
Second, there are variations in ballot paper design (characteristic I I ) that
have important implications for the extent to which a candidates success at
the polls is determined by elements of chance as opposed to the deliberate
| 59 |
intentions of the electorate.
11
The issue here concerns the shortcuts that
voters take in completing their ballot papers, generally referred to in the lit-
erature as alphabetical or donkey voting (Darcy and McAllister 1990). The
more the voter is taxed by the system, the greater the likelihood that she will
make use of shortcuts, in effect voting for candidates on the basis of list-order
rather than out of real preference. Therefore, in a system requiring compul-
sory turnout and with compulsory preferential voting, it is only to be expected
that donkey voting will feature quite prominently (Darcy and Marsh 1994;
Kelley and McAllister 1984; Mackerras 1970). This problem is compounded
| 60 |
TABLE 3.4 VARIANTS OF STV IN AUSTRALIA, IRELAND AND MALTA
Australias Upper House (UH) Systems
Characteristics Commonwealth UH New South Wales UH
I. District Size Range 6-12
a
21
II. Ballot Paper Design Across = party groupings (by lot)
Down = candidate names (party
choice)
Across = party groupings (by lot)
choice)
III. Surplus Transfer All ballots at fractional value
(inclusive-Gregory method)
Surplus ballots only, at full value
IV. Preferences Minimum = # candidates,
b
or if adopt
ticket, one vote only
Minimum = 15, or if adopt ticket,
minimum = 1
c
V. Casual Vacancies State parliament selects someone
from same party
Joint sitting selects someone from
same party
Australias Lower (LH) House Systems
Characteristics ACT LH Tasmania LH
I. District Size Range 57 5
II. Ballot Paper Design Across = party groupings (by lot)
Down = candidate names (by
rotation)
Down = candidate names (by
rotation)
III. Surplus Transfer The last parcel of ballots at fractional
value (Gregory method)
The last parcel of ballots at fractional
value (Gregory method)
IV. Preferences Minimum = 1 Minimum = #seats
V. Casual Vacancies Count back Count back
NOTES
a The ACT and Northern Territory elect just two Senators each.
b In general a vote for 90 percent of candidates sufces (with no more than three omissions or duplications in sequential
numbering).
c When voting above the line, voters can rank-order the parties (optional preferential ticket-voting). There are no longer any
ticket-deals between the parties.
d The rst STV election is due in 2006.
SOURCES As for Table 3.3; Victoria Constitution (Parliamentary Reform) Act 2003; Ireland Electoral Act 1923; Malta
General Elections Act 1991.
| 61 |
South Australia UH Western Australia UH Victoria UH
d
11 Range 5-7 5
choice)
Left = party groupings (by lot)
Right = candidate names arranged in
columns (party choice)
Left = party groupings (by lot)
Right = candidate names arranged in
columns (party choice)
All ballots at fractional value
Minimum = # candidates, or if adopt
Minimum = # candidates, or if adopt
Minimum = # seats, or if adopt ticket,
one vote only
Joint sitting selects someone from
same party
Count back Joint sitting selects someone from
same party
Ireland and Malta
Characteristics Ireland LH Malta LH
I. District Size Range 35 5
II. Ballot Paper Design Down = candidate names
(alphabetical); includes party labels
Down = candidate names (sorted
alphabetically within party groupings)
III. Surplus Transfer Surplus ballots only, at full value Surplus ballots only, at full value
IV. Preferences Minimum = 1 Minimum = 1
V. Casual Vacancies By-election Count back
in Australia by the presence of a large number of non-English-speaking
migrants, many from countries lacking democratic traditions. With minimal
information about the electoral system, they often utilise shortcuts when
completing the ballot, such as candidate placement on the ballot paper.
There are two ways of dealing with this problem. One is to make full use
of it in a partys strategy, and to list party candidates in order of importance,
or on the basis of some other relevant criterion. This is the procedure fol-
lowed in all Australian upper house cases, in which the rank ordering of can-
didates is set by the party organisations. I t is heavily reinforced by ticket
voting: the vast bulk of voters opt to vote the ticket, thereby ensuring that the
rank-order set by the party organisations remains unchallenged. (This is
further discussed in chapter 6.) Thus, party strategy in this regard again
reflects their strength and discipline within the political system as a whole.
An alternative strategy for dealing with donkey voting is to rotate the
position of the candidates from one ballot paper to the next, thereby giving
each candidate an equal chance of being placed at the top of the ballot paper.
This procedure was adopted in Tasmania in 1980 where it was given the
title Robson rotation, named after the Tasmanian legislator who introduced
the Bill and it appeared to reduce the effect of donkey voting (Darcy and
Mackerras 1993). When it adopted STV in 1995, the ACT also incorporated
this procedure.
12
While Robson rotation appears to be effective in reducing
the incidence of donkey voting, the main disadvantage is the additional
administrative burden that is places on the authorities responsible for
mounting an election. Even in relatively small elections, such as the ACTs,
many hundreds of ballot papers have to be printed and distributed to voters
in approximately equal proportions, in order to achieve the desired outcome.
Ballot paper design can also have important implications for the degree of
leadership control over candidate nomination and, indeed, for leadership
influence over candidate election. The rank ordering of candidates by the
parties and the operation of the ticket system represents the extreme position.
I t has been shown that in the case of the Senate, for instance, candidates need
to pay more attention to the party leadership than to their voters. As Jaensch
(1986: 58) puts it: [I ]n the real world of parties and voters in Australia
position on the ballot paper is crucial to re-election. I n states where there are
two major party lists, a candidate for either of the two major parties placed in
the first or second position on the list is guaranteed election, a third position
is marginal, and a lower position on the list guarantees failure; therefore, the
senatorial campaign is less one of trying to win the hearts and minds of the
Australian voters and more one of trying to win the hearts and minds of party
selectorates. We return to this theme in chapter 5.
I n the ACT, Tasmania, Ireland and Malta, which determine candidate
placement either by lot or by alphabetical placement, the party leadership has
no direct influence.
A third source of variation in the STV system also has implications for
elements of chance in the electoral outcome and is relevant to the social
choice debates reviewed in chapter 1. I t relates to the process for transferring
ballot papers in the case of surplus votes (characteristic I I I ). I n Malta, Ireland
and New South Wales, when transferring a surplus after the election of a can-
didate, only those ballots that are surplus to the quota are transferred, at their
full value, and based solely on the next preference marked on the ballot paper.
This used to be the standard procedure for STV systems: Tasmania used it in
its first version of STV in the late nineteenth century, the Australian Senate
used it prior to 1984, and it remains in use in New South Wales.
13
| 62 |
I n the New South Wales case, the determination of which ballots to actu-
ally transfer is based on random selection, which can have important impli-
cations in later counts (of which there are many given that 21 seats are being
filled), particularly when the results between two candidates are very close.
Depending on which ballot papers were selected from the pile at an earlier
stage in the counting process, in a close finish the fate of a candidate could
be sealed by the particular pattern of preferences that predominated in those
ballot papers. I n other words, there are random effects involved in the
counting process. While some have suggested that it is extremely unlikely
that this would result in the incorrect election of a candidate (Lakeman 1974:
140), statistical analysis has demonstrated conclusively that this procedure of
transferring surpluses permits an element of arbitrariness that can[not]
be ignored with impunity (Gallagher and Unwin 1986: 253; see also, Coakley
and ONeill 1984; Fischer 1978, 1981).
14
A solution to the problem of preference selection is to take account of all
preferences when allocating a surplus: this is done by transferring all ballot
papers received by the candidate (not just a sample of those ballots surplus to
the quota) at a fraction of their count value.
15
This is usually referred to as the
Gregory method, after the Melbourne mathematician, J. B. Gregory, who
devised the scheme in 1880. The Gregory method was first used in Tasmania
in 1909 when the state re-introduced STV, and it has been used there ever
since. As Table 3.4 shows, the ACT also adopted this method. Apart from
these two cases, and New South Wales (which only transfers surplus ballots),
all the other Australian cases use a system that bears close similarities to the
original Gregory method, but with one significant difference.
16
This relates to
the transfer of surpluses at later stages in the count (after a first surplus
transfer has occurred). Under the Gregory method, only the last parcel of
ballot papers is transferred; all the other ballot papers remain undisturbed. By
contrast, under the inclusive Gregory method used for the Senate and in
South Australia and Western Australia, no distinction is made between the
different ballot papers; instead, all of the ballot papers are transferred at a
fractional value.
17
The implications of this variation are discussed in the next
chapter. I n the following paragraphs we briefly outline its origins.
As part of the 1983 reforms reviewed in the last chapter, the Australian
Senates system was changed from the random sampling method to the inclu-
sive Gregory method. The decision to adopt this version of the Gregory
method provides an interesting illustration of the contingent nature of elec-
toral system design that was discussed in chapter 1.
18
A series of close Senate
| 63 |
elections in the 1970s and 1980s raised concerns over whether random sam-
pling always produced the correct result. Perhaps the most celebrated case
was the 1974 double dissolution that threw up evidence of the dangers of
random sampling, as well as of the problems of taking account only of the last
parcel received. I n Queensland the Labor party candidate, Malcolm Colston,
failed by a narrow margin to win the last seat. One of the Liberal candidates
was the incumbent, Neville Bonner, who was placed third on the Liberal
ticket. There was evidence from scrutineers reports that a large proportion of
the first preference votes for Bonner had subsequent preferences for Labor
candidates. However, because Bonner was elected on the basis of votes trans-
ferred from another candidate, only those ballot papers received in the last
parcel were transferred on to the remaining candidates: none of Bonners
other preference votes were transferred. There is good reason to believe that,
had all of Bonners ballot papers been included in the transfer to the
remaining candidates, Colston would have secured the final seat (Colston
1975: 109; Fischer 1981: 59).
Bonner syndrome,
19
as it has been labelled, and the problem of random
sampling more generally, featured in the deliberations of the Joint Select
Committee on Electoral Reform ( JSCER) over reform of the Senate elec-
toral system in 1983. Advice from the Proportional Representation Society of
Australia (particularly by its president, Jack Wright), from academic experts
(such as Alastair Fischer) and from the research staff of the Australian
Electoral Office (the precursor to todays Australian Electoral Commission)
all favoured a shift away from random sampling, but it is apparent that there
was some ambiguity over what exactly was being proposed as an alternative.
Certainly the presumption on the part of some of those providing the JSCER
with advice was that it would opt for the Gregory method.
20
But advice from
the Electoral Office, and the desire of Labor politicians to prevent the reoc-
currence of the 1974 debacle, favoured moves to modify the Gregory method
so that it would include all ballot papers, not just the last parcel received. This
was the recommendation of the JSCER in 1983, ( JSCER First Report,
September 1983: 65) and it was adopted virtually without any parliamen-
tary debate
21
in the Commonwealth Electoral Legislation Amendment Act
of that year.
22
A fourth source of variation in the STV system is over what is expected
of the voters in completing the ballot (characteristic I V) an issue that
refers back to the discussion of the previous section regarding the AV
system. I n the ACT, as in I reland and Malta, there is optional preferential
| 64 |
voting: voters need only mark a first preference on the ballot paper for their
votes to be counted as valid. I n Tasmania, and now Victoria, there is
minimum-length preferential voting: voters have to declare at least as many
preferences as there are seats to be filled. The task of voting, however, is far
more onerous in most of the upper house versions of STV, which operate a
system of compulsory preferential voting in which voters have to declare a
preference for each candidate on the ballot paper (or, at least, for all but one
candidate the blank space is counted as the final preference).
23
Given its
very large district magnitude (21 seats) and accordingly its large ballot
papers (of record-breaking proportions see below), New South Wales
allows some leniency in requiring voters to only declare 15 preferences for
their vote to be counted as valid.
We saw in chapter 2 how the operation of this compulsory preferential
voting system tended to result in high numbers of invalid (or informal) votes,
and this, together with greater party control over the flow of preferences, was
to lead in 1983 to the adoption of the ticket voting system for Australian
Senate elections. State governments with upper houses elected by STV
systems quickly followed suit, allowing voters the easy option of simply
declaring one preference for a party or group above the line (or, in the case
of Western Australia, to the left of the line) and allowing the preferences to
be determined on the basis of prior arrangements between the parties. Such
arrangements reached between the parties are rarely known by voters, and
receive little, if any, publicity. For example, in the 2004 election, Labor
reached an arrangement with the Family First party in Victoria, Tasmania
and South Australia whereby preferences flowed to them before the Greens.
Until after the election, few voters (and especially, few Labor voters) were
aware of this arrangement.
I nevitably, the modalities can vary across Australia in terms of the flow of
preferences. For instance, in the case of Senate and South Australia elections,
but not in Western Australia, it is a requirement that the tickets are displayed
at polling stations for the voters to inspect. Western Australia is also excep-
tional in allowing the parties, groups and candidates to register only one
ticket: in the other cases, there is the option of registering more than one
ticket, in which the votes transfer automatically on a proportional basis as
halves or thirds. A maximum of three tickets can be registered for the Senate;
up to two tickets can be registered in South Australia. The use of ticket voting
gives even more power to the party bosses than they would have under the
most rigid or closed of party lists systems.
24
Not only can the party bosses
| 65 |
determine the order in which their candidates are elected, they can also
decide on how voters later preferences will transfer to other parties.
A recent modification of ticket voting in New South Wales has resulted
in yet another shift of the STV system in the direction of a conventional
closed-list electoral system. The catalyst for change was the 1999 tablecloth
election, which saw 80 parties and groups fielding between them 264 candi-
dates for the 21 seats, resulting in a ballot paper that measured one metre by
70 centimetres.
25
Micro parties took advantage of what transpired to be the
lowest electoral threshold in Australia (see also chapter 4),
26
combined with a
lax party registration and candidate nomination regime. For many of them,
their sole purpose was to act as preference-funnels, soaking up first prefer-
ence votes that would transfer through ticket-voting deals to other parties.
Here was a deliberate, undisguised attempt to use the ticket-voting system to
wrestle some seats from the larger established parties (e.g. Green 1999;
Jackson and Harris 1999).
The strategy seems to have had limited success, as only a few of the groups
that appeared to be at the centre of the preference-funnel arrangements (Morris
1999) benefited in terms of seat gains. Nevertheless, as Table 3.5 shows, there
were some extraordinary results, with micro parties winning seats based on tiny
proportions of the first preference vote. One example was the Outdoor
Recreation Party, which won a seat despite having a mere 7,000 first-preference
votes (0.2 percent of the total). In consequence, the New South Wales result
produced a relatively high level of disproportionality (5.88 percent). Another
example of this phenomenon occurred in the 2004 Senate election in Victoria,
when the Family First party won two percent of the first preference vote but
secured a quota of 14.3 percent as a consequence of pacts with the Liberal and
Labor parties over preferences. These examples demonstrate how an electoral
system can actually be toofair to smaller parties, resulting in a rise in dispropor-
tionality caused by the fact that small parties can win seats on tiny proportions
of the vote backward disproportionality, as it were.
I n the light of this perverse result, the Carr Labor government moved
quickly to introduce legislation reforming the electoral procedures to make
it more difficult for micro-parties to win seats (see Green 2003; Griffith and
Srinivasan 2001: 82). The legislative reforms included a series of steps to
tighten up the party registration and candidate nomination rules.
27
There
was also a change to the rules on ticket voting, designed to prevent further
bouts of preference trading between micro parties. Under the new ticket
voting procedures, voters are now able to rank the party lists above the line,
| 66 |
and there are no ticket preference deals by the parties that lead to the
transfer of preferences.
28
This is an optional preference system. I f they wish,
a voter may express just one preference, voting for just one party and
thereby treating the New South Wales Legislative Council electoral system
as a conventional closed list election.
29
The final set of variations in the STV systems is over the filling of casual
vacancies (characteristic V in Table 3.4). On the one hand, there is the
assumption that the seat belongs to a party and therefore that the filling of
the vacancy should be determined by that party. This is the procedure now
followed in the Senate (since 1977),
30
South Australia, New South Wales,
and Victoria.
31
On the other hand, in the ACT, Tasmania and Western
Australia, as well as in Malta, it is argued that the voters have decided and
thus that the best way to fill the vacancy is to go back to the original ballot
papers and carry out a re-count (or count back), awarding the seat to the next
best-placed candidate. Ireland falls between these two extremes, preferring to
| 67 |
TABLE 3.5 THE 1999 NEW SOUTH WALES LEGISLATIVE COUNCIL ELECTION (THE TABLECLOTH ELECTION)
Party/Group Votes (%) No of seats Seats (%)
Labor 37.27 8 38.10
Liberal/National 27.39 6 28.57
One Nation 6.34 1 4.76
Australian Democrats 4.01 1 4.76
Christian Democratic Party 3.17 1 4.76
The Greens 2.91 1 4.76
John Tingle The Shooters Party 1.67
Progressive Labor Party 1.58
Marijuana Smokers Rights Party 1.24
Reform the Legal System 1.00 1 4.76
Unity 0.98 1 4.76
Country Summit Alliance 0.89
Registered Clubs Party 0.77
Gun Owners & Sporting Hunters 0.71
Country Party 0.56
Whats Doing? Party 0.51
Others
a
9.00 1 4.76
a
The remaining 65 parties and groups with less than 0.5 percent of the vote. The seat was won by the Outdoor Recreation
Party, which secured 0.20 percent of the rst preference vote.
SOURCE Green (2000).
ask the voters to decide again in a by-election (and thereby, in a single-
member election, using AV).
32
Only in Ireland, therefore, does a government
have to contend with a mid-term test. I n this aspect, the Irish system can also
be said to favour the larger parties, which stand a greater chance of having
their candidates elected in a majoritarian by-election.
33
On the basis of this review of the five main dimensions of variation in
forms of STV, we are now in a position to comment on the ongoing debate
over whether or not Hare-Clark as practiced in Tasmania is fundamentally
different from the forms of STV practiced elsewhere. Table 3.4 provides
compelling evidence that there are two main streams of STV: a lower house
variant (Tasmania, ACT, Ireland and Malta) and an upper house variant
(Commonwealth, New South Wales, South Australia, Western Australia, and
now Victoria).
34
At the heart of this is a distinction between a party-con-
trolled form of STV, as practiced in upper houses, and a voter-oriented form
used in lower-houses: in other words, the principal point of distinction is over
the degree of party control of the preferential element. To that extent, there-
fore, scholars are right to point to a particular form of STV which some
prefer to refer to as Hare-Clark but there should be greater recognition of
the fact that this variant is not entirely unique to Tasmania and the ACT.
Of course, Australia has not always resorted to using preferential systems
as the sole alternative to plurality voting. The next section reviews those
instances of electoral reform in which other, ostensibly non-preferential,
systems were considered. However, as we shall see, even in these cases, aspects
of preferential voting still feature quite prominently, demonstrating yet again
the central importance of this form of voting to representative politics in
Australia.
Australias experiments with
alternatives to AV and STV
Although AV and STV predominated for most of the twentieth century,
other electoral systems have emerged from time to time in Australia. I t is cer-
tainly clear from Table 3.2 (p. 50) that Australia has not been entirely com-
mitted to preferential systems. At the state and national level, at various
stages (though, in large part, not since the end of the first half of the last
century), use has been made of plurality electoral systems, usually SMP. On
occasions, this has been combined with the multi-member block vote system.
| 68 |
There has also been some experimentation with the majority-runoff
system (as used, for instance, in France for its presidential elections). I n this
system, if no candidate secures more than 50 percent of the vote, a second
election takes place within a few weeks of the first. This time, only the can-
didates who secured the highest and second highest proportion of votes in the
first round are allowed to run. This system was used in New South Wales for
three elections to its Legislative Assembly between 1910 and 1917, after
which it was replaced briefly by STV (Hogan 1995). The Liberals had intro-
duced the runoff system at a time when they were experiencing some splin-
tering of their support, in the hope that it would help to shore them up
against incursions by Labor. This was to prove unsuccessful. I n general, AV
has been favoured over the majority-runoff system, in part because of the vast
sizes of the Australian electorates and the desire to minimise costs and dis-
ruption in the electoral process. Arguably New South Wales, with its high
degree of urbanisation and population density, could be seen as better suited
than many other parts of Australia to the majority-runoff system. Even so,
the system was removed by the Liberals in 1917, partially on grounds of costs,
but also due to instances of turnout variation across the two rounds, leading
to suspicion of improper campaign practices by the Labor party (Hogan
2001b: 148).
When it comes to considering PR systems, Australia has not entirely
ignored the list systems, which are common to continental Europe. Even
here, however, the Australian experiences with these systems provide plenty
of evidence in support of the law that, when choosing PR, Anglo-Saxon
democracies tend to favour STV over list (Lijphart 1987) a law broken
ultimately by New Zealand in 1993. The list systems of proportional repre-
sentation made two brief appearances: in South Australias upper house in
the mid-1970s and for the first elections to the ACT Legislative Assembly
in the late 1980s and early 1990s. As we shall see, both cases, particularly the
ACT, reveal just how significant a role can be played by the process of com-
promise in altering the proposed system, sometimes drastically, changing it
from the original intentions of the drafters (Coram 1996). We start with
South Australia.
Don Dunstan was premier of South Australia from 1967 to 1968 and
again from 1970 to 1979, following a long period of uninterrupted rule by the
Liberal and Country League (LCL). I t was generally recognised (though see
Goot 1985: 218) that a major factor behind the repeated electoral successes
of the LCL had been the heavy weighting of votes in favour of rural parts of
| 69 |
the state. This very effective malapportionment (known colloquially as a
Playmander, after the LCL premier, Thomas Playford) gave the LCL an in-
built advantage in House of Assembly elections. This advantage was com-
pounded by the restricted franchise for electing members of the Legislative
Council. Although electoral reform began with the then premier, Steele Hall,
in 1969, the pace of reform quickened in the 1970s. One of the hallmarks of
the Dunstan Decade, therefore, was the end of rural weighting and the
introduction of full adult suffrage for Council elections, the latter including a
new electoral system ( Jaensch 1981).
The original intention had been to introduce a conventional list system.
I n the 1973 Council and Electoral Acts Amendment (Council Elections)
Bill, the Labor government proposed a largest remainder system, using the
Droop quota,
35
and including a minimum legal threshold. As part of the
reform, the size of the Council was increased from 20 to 22 members. I t was
proposed that the entire state would be one large electorate, with 11 members
being elected at a time (that is, a district magnitude of 11). I t was this large
district magnitude that determined why the Dunstan government had
selected a list system, as opposed to STV. This was a decade before the intro-
duction of ticket voting in the Commonwealth Senate and, therefore, no easy
solution offered itself as to how to deal with the problem for voters of com-
pleting (compulsory) preferences on large ballot papers. I t was expected that
there would be anywhere between 30 and 50 candidates on the ballot paper.
While the use of a Droop quota is uncommon in LR systems, there are
other cases where it is or has been used (Farrell 2001). The unusual feature
entered in the first draft of the Bill was the means by which the legal
threshold would be determined. The usual practice is to set a minimum pro-
portion of the vote that a party must win before it is awarded any seats (such
as Germanys five percent threshold). I n this case, however, and perhaps
reflecting the greater familiarity of Australian politicians with STV, it was
decided to set the legal threshold at one-half of a quota, which was expected
to translate to about four percent of the vote.
36
I n the original Bill it had been
proposed that voters would only be required to indicate support for one party.
I f they wanted, voters could express more than one preference, but, as
Dunstan observed during the course of the parliamentary debate, expressing
more than one preference would be pointless as it would have no bearing on
the result: the inclusion of this right was simply to avoid confusing voters who
were accustomed to voting by numbers rather than by a simple cross (South
Australian Parliamentary Debates (SAPD) June 19 1973: 20).
| 70 |
Criticisms in the parliamentary debate focused on the legal threshold and
on the lack of preferential voting. The government showed no sign of yielding
on either. A sticking point, which linked both criticisms, was as follows.
Small parties would be treated unfairly by being excluded at the outset for not
passing the threshold, whereas a candidate of a larger party elected last, on the
basis of largest remainders, could well be elected with less than half a quota.
I n addition, their supporters would in effect be disenfranchised because their
votes would be discounted. A solution was offered by the opposition leader in
the Council, Ren DeGaris, who proposed that voters should vote preferen-
tially and that the preferential system be preserved up to the point where all
votes excluded because they do not reach a prescribed number for a candidate
or a group are allocated by preference to a group that has a number of votes
higher than that prescribed number (SAPD June 26 1973: 121).
On the face of it, this was quite an elegant solution, and the Dunstan gov-
ernment eventually agreed to make a concession, albeit with some reluctance
and not without some last minute high drama threatening the success of the
entire Bill ( Jaensch 1981: 230). However, it was proposed that preferential
voting should be voluntary, not compulsory as had been proposed in the
DeGaris amendment.
37
The list system was to prove short-lived, however. After just two elections
(1975 and 1979), the Liberal government passed new legislation in 1981,
replacing it with STV (see Jaensch 1986a). In this case, the reason for dropping
the system as also was the reason for adopting the system in the first place
was expediency, and distrust of a system designed by their partisan opponents.
As we shall see next, the short-lived nature of the ACTs experiment with a list
system had less to do with the proclivities of the politicians and a lot more to
do with the need to replace a disastrously complex and unworkable system.
I n South Australia, the design of electoral systems was the purview of the
state politicians. I n the case of the ACT, the debate was between federal
politicians designing electoral institutions for a new self-governing territory.
The ACTs slow and rather faltering steps towards self-government culmi-
nated, in the mid-1980s, in a debate in the Commonwealth parliament over
what electoral system to adopt for its unicameral Legislative Assembly. The
national ALP government s first proposal, in 1985, was for a majoritarian
system based on single-member constituencies; by contrast, the other parties
favoured STV, then being used for the partially elected ACT Advisory
Council. Given that the ALP was in a minority position in the Senate, it
could not get its way on this proposal, and the Bill faltered, though not before
| 71 |
a last ditch attempt to introduce a mixed-member electoral system entailing
a combination of single-seat members of the legislative assembly (MLAs)
and an ACT-wide top-up of additional MLAs.
I n 1988, the ALP produced a new proposal for a highest average dHondt
system, one of the most common forms of list system currently in use (Farrell
2001). The Legislative Assembly of 17 members would be elected at large,
across the entire territory (that is, a district magnitude of 17). Bearing many
similarities to other dHondt systems, the Bill proposed a semi-open version
of list, in which voters could either mark a preference for a party or indicate
their preferred candidate on the party list, thereby influencing the order in
which a candidate might be elected.
38
The original design was based specifi-
cally on European practice; there is no evidence that any account had been
taken of the South Australian experiments of a decade earlier ( JSCEM
1989). By the time the Bill had been through both houses of the national par-
liament and had been worked over by the Australian Democrats, it was
changed beyond recognition, and the new modified dHondt system that
eventually emerged was a hybrid of dHondt and STV.
The system was characterised by four main features.
39
First, prima facie,
the ballot paper looked much like a Senate ballot paper (as shown in Figure
1.2): parties and independents were arranged left to right, and party candi-
dates were arranged below the relevant party ticket.
A second feature was that voters could vote in a mixture of ways: for party
tickets, for party candidates or for independent candidates (who, for the pur-
poses of the exercise, were treated as parties). The vote was valid (or formal)
as long as a voter expressed at least one preference. The voter could give addi-
tional optional preferences, and this could be in any sort of mix. For instance,
the voter might choose to give some preferences for a number of party tickets,
and then might express some preferences for individual candidates. I n other
words, and unlike a Senate election, the voter could vote both above and
below the line. As a consequence, a complex series of deeming rules were
employed to determine what constituted a valid vote. For instance, a valid
vote did not even have to involve consecutive numbers; all that was required
was that the preferences be ascending and unrepeated (e.g. 1, 6, 20, 35).
Third, a legal threshold was set, a Droop quota, which in the context of a
17-seat legislature, approximated five and a half percent.
40
As in the South
Australian list system, the ballot papers of any excluded parties or independent
candidates were transferred to the remaining parties, based on the second pref-
erence votes of those voters who expressed more than one preference.
| 72 |
Finally, seats were allocated to parties and candidates under dHondt pro-
cedures.
41
I n the cases of those parties that won fewer seats than they had
candidates, individual preferences by voters were taken into account to deter-
mine the final ranking of the candidates of that party, and therefore which of
the candidates were to be elected, and in what order. Senate-style STV
counting rules were deployed at this stage. Again, there was a complex
process for determining the preferences for individual candidates (with
ballots from excluded and non-excluded candidates taken into account at the
various stages in the count process).
Considerable confusion arose from this complex electoral system and
from the fact that 117 candidates stood for election in 1989. (Such a large
number was predictable given the large number of seats and the low level of
deposit required.) The number of candidates meant that the ballot paper
was then the largest ballot paper in Australian history, measuring one metre
by 35 centimetres, and causing major logistical problems in the count
centre. There were also complex electoral rules, in particular a set of
deeming rules that required the scrutiny of each square, both to confirm
whether a vote was valid and to determine the preference order across
parties and candidates. This meant that, across the more than 150,000
ballot papers involved, the scrutiny staff had to check every one of 20
million squares ( JSCEM 1989: xviii). As a consequence the counting
process took more than nine weeks to complete. A subsequent committee
of inquiry recommended scrapping the system ( JSCEM 1989), and in
1992, coinciding with the subsequent ACT Legislative Assembly election,
a referendum was held in which the voters were asked to choose between
the Tasmanian form of STV or AV. The voters voted overwhelmingly for
STV, with 65 percent in favour.
Both the South Australian and the ACT experiments with versions of list
were short-lived, indicating a low level of sympathy among Australian politi-
cians for those forms of PR most commonly associated with continental
Europe. I n this context, however, it is worth noting that these need not have
been the only cases of list systems in Australia. For instance, in 1977 Neville
Wran (New South Waless answer to Don Dunstan) also brought forward
proposals for a list system as part of his drive to reform the process of electing
the New South Wales Legislative Council. I n hindsight, it is apparent that
this was a ploy by Wran to gain the support of the opposition parties: so
opposed were they to the list proposal that when he proposed STV instead,
they quickly fell in line (Chaples 1985; Turner 1985).
| 73 |
We have seen how, both in South Australia and the ACT, the politicians
were loath to move entirely away from features of STV: preferential voting was
allowed for and the Droop quota was used to determine the electoral
threshold. (I n the South Australian case, only half a quota was required, and
the quota was also the means of allocating seats). I n South Australia, this latter
modification of the largest remainder system was a sensible compromise by the
government party to ensure the passage of the legislation; in the event it had
little tangible affect on the operation of the system. Matters were distinctly
different in the case of the ACT, where, in effect, STV was bolted onto a
highest average list system, causing havoc for the counting process and ulti-
mately discrediting the electoral system. This was to result in its swift demise.
The extent of the modifications to the dHondt system also resulted in its
poor performance as a proportional representation system, particularly in its
first outing in the ACT in 1989. As Table 3.6 shows, the level of dispropor-
tionality in that election was 15.28.
42
I f we compare that with a recent
analysis of disproportionality trends in more than 50 of the worlds democra-
cies, this placed the ACT at the bottom of the league, competing with non-
proportional UK, Canada, Mongolia, France and Jamaica for the accolade as
the most disproportional electoral system in the world (Farrell 2001: 15759;
see also Table 1.1 above). However, this was just one election, and, given that
it was a new political system, it inevitably attracted a lot of small esoteric
parties and single-issue candidates, swelling the number of wasted votes. I n
total there were 31 parties and independents. I n the subsequent 1992 elec-
tion, using precisely the same electoral system, but this time with only 17
parties and independents, the level of disproportionality dropped to 9.69 (still
a very high figure for a PR system, especially given the district magnitude).
But it was not only the number of parties and candidates that caused such
a poor PR performance. Two features of the electoral system were specifically
to blame: the legal threshold and the right of voters to vote preferentially. The
legal threshold which was high by international standards attracted a lot
of criticism, and for good reason. As Table 3.6 demonstrates, if we re-calcu-
late the count without the legal threshold, there is a dramatic drop in the
measure of disproportionality. The threshold undoubtedly had the greatest
effect on overall proportionality, but the operation of the preferential vote, at
the first stage of the count, also had a significant effect on the share of seats
won by individual parties. This is caused by the way in which the votes of
excluded parties and candidates transfer to those parties remaining in the
race. No other list system operates such a system; the usual practice is for such
| 74 |
ballots to be excluded from the count altogether. As we see in the second-last
column of Table 3.6, when we remove the modification of preferential voting
whilst keeping the legal threshold in place, there are some distinct shifts in
the share of seats won by certain parties: notably, the ALP share increases
sharply while that of the Residents Rally is more than halved. This re-jigging
of seat shares may produce more or less favourable outcomes for particular
parties, but its overall effect on proportionality is small (the disproportionality
index records a small decline to 14.26).
| 75 |
TABLE 3.6 THE 1989 ELECTION TO THE ACT LEGISLATIVE ASSEMBLY MODIFIED DHONDT VERSUS PURE
DHONDT
Votes (%) Seats (%)
Modied dHondt Pure dHondt
with threshold
Pure dHondt
without threshold
Australian Labor Party 22.82 29.41 35.29 29.41
Liberal Party 14.87 23.53 23.53 17.65
Residents Rally 9.62 23.53 11.77 11.77
No Self Government Party 11.47 17.65 17.65 11.77
Abolish Self-Government Coalition 7.40 5.88 11.77 5.88
Fair Elections Coalition 5.47 5.88
Independent Haslem 4.84 5.88
The ACT Community
Party
4.07 5.88
Bill Mackey (Ind.) 4.01 5.88
Canberra First Party 3.47
Family Team 2.74
Australian Democrats 1.66
National Party 1.37
Harold Hird (Ind.) 1.32
Sun-Ripened Warm Tomato 1.17
Party! Party! Party! 0.69
Christian Alternative
Party
0.60
Socialist Workers Party 0.50
Others
a
1.91
Disproportionality
b
15.28 14.26 6.88
a
Disabled & Redeployed Workers Party; Home Rule OK; A Better Idea; Sleepers Wake; Tony Spagnolo Independent for
Canberra; Surprise Party; Frank Crnkovie (Ind.); Bob Reid (Ind.); Kevin Robert Wise (Ind.); Gary James Pead (Ind.);
Bill Pye (Ind.); John Rocke (Ind.); Lyall L. Gillespie (Ind.).
b
Gallagher index.
SOURCE ACT Electoral Commission
Because the modifications to largest remainder list were far less dramatic
in South Australia, and, more particularly, because of a lower legal threshold
(about 4 percent), there was a far better degree of proportionality than in the
ACT. I n the first election of 1975, overall disproportionality was a creditable
5.53, while in 1979 it fell to 4.23.
Conclusion
This chapter, together with the detailed history of national developments in
the previous chapter, provides much evidence of just how open Australian
governments have been to modifying existing electoral systems and to
embracing far-reaching electoral reform. Over the past hundred years or
more, if we include the brief flirtations with list systems in South Australia
and the ACT, Australia has witnessed almost every conceivable electoral
system. Perhaps the only system that has not been used is the mixed-member
system, although even this, at one point, was mooted for the ACT and was
actively considered in Queensland in 199091. Moreover, the Australian
Democrats are currently pushing for the introduction of MMP at national
level. As we have seen, in due course AV and STV in all their varieties were
settled on for all levels of government, and here the range of variations in
these systems has been impressive.
I f electoral system experimentation is one important theme to emerge
from this chapter, another relates to the dominance in the Australian electoral
process of forms of preferential voting. As the third section showed, even in
the two most significant cases where alternatives to STV were tested, features
of preferential voting still managed to percolate through (with disastrous con-
sequences for the count process in the case of the ACT). Yet another theme
is the formal and informal interventions by which the political parties have
harnessed electoral system design to suit their own purposes, and how they
have evolved complex strategies to use the system to their own best advan-
tage. This is hardly a process unique to Australia, but there are few other
established democracies in which the major political parties have had such
free rein in producing a system that reduces their commitments in time, effort
and material resources.
I t is not solely the distinction between proportional and majority prefer-
ential systems that is of interest, even if this clearly has the largest bearing on
overall proportionality. Also interesting are the varieties in the forms of these
| 76 |
systems. The main features of variation that were traced in this chapter will
form the basis for much of the analysis that follows. I n chapter 4 we examine
the systemic implications of these preferential systems, starting with the issue
of aggregate proportionality here focusing particularly on STV. I n the first
chapter (see Table 1.1) we saw how poorly STV compared with other propor-
tional systems on this dimension. By extending our number of cases to
include Australian state level STV systems, we can provide a far more com-
plete analysis than heretofore of how this system performs. Furthermore, the
range of constituency sizes, or district magnitudes, shown in Table 3.4, pro-
vides scope for considering which STV systems fare best on the scale of pro-
portionality. Chapter 4 also addresses the social choice theme of how
preferential systems perform in terms of electing the correct candidate. Once
again making use of the range of variation in the preferential systems, as sum-
marised in Tables 3.3 and 3.4, we assess how two features of variation the
rules on how many preferences are required for a vote to count as valid and
the transfer of surplus votes under STV may affect the electoral fate of indi-
vidual candidates.
Chapter 5 assesses the inbuilt tension in Australian electoral system
design between a voter-oriented preferential system on the one hand and, on
the other, that systems conditioning by party-oriented elements such as ticket
voting and the compulsory expression of preferences. I n this instance, our
attention will be on the role of the elected representatives in the two national
houses of parliament.
Chapter 6 addresses the theme of strategic voting by voters under condi-
tions of electoral system complexity, and conditioned by the operation of
ticket voting. Given that ticket voting is not used for elections at state or ter-
ritory level in Tasmania and the ACT (Table 3.4), there is the basis for an
interesting examination of what effect ticket voting has on conditioning
voters in the act of voting.
| 77 |
f o u r
Preferential systems and
their consequences
The previous chapters have traced the evolution of Australias electoral
systems and explored their principal variations. I n this and the next three
chapters the attention shifts to an examination of electoral system effects.
This chapter considers two features of electoral system consequences. First,
we examine their aggregate systemic effects as measured by proportionality
and party system indices. Second, we return to the social choice theme raised
in chapter 1, relating to the logical properties of electoral systems: here we
examine whether certain features of AV and STV influence which candidates
actually get elected.
The most common means of distinguishing electoral systems is by the
degree to which they produce proportional results (that is, the degree to
which a partys vote share corresponds with its seat share). From Rae (1967)
onwards, political scientists have devised indices of disproportionality as a
means of distinguishing the different families of electoral systems. There have
been debates (some more heated than others) over which factors in electoral
system design most affect the proportionality of the result, with general
agreement that, of all the potential factors, the two that have the greatest
impact are constituency (or electorate) size, known technically as district
magnitude, and the formula used for translating votes into seats (that is, the
electoral formula; for the most recent overview, see Katz 1997). While Rae
(1967) found no evidence that ballot structure (the rules governing how to
complete the ballot paper) had any influence on proportionality, subsequent
| 78 |
research by Lijphart (1994) found some support for the proposition. I n that
study, Lijphart also showed how the size of the parliamentary assembly could
influence proportionality.
AV is located in the non-proportional group of systems, producing trends
similar to those observed in plurality systems (Lijphart 1994; Blais and
Massicotte 1996; see also Table 1.1). Given that Australia is the only case (at
least, the only case with a respectable long run trend) of AV, there has been
little scope to examine its proportional traits in any great detail. For the most
part, the general assumption has been that, in the aggregate, it produces
results not unlike SMP. Thus, while AV does at least ensure that each con-
stituency politician is elected with 50 percent or more of the votes in her con-
stituency, there is insufficient evidence that this has any real effect on the
overall result in terms of the distribution of seats between the parties.
The treatment of STV in the comparative literature has also tended to be
cursory, in large part reflecting the fact that it is only used in two countries
for national lower-house elections (Ireland and Malta). Given the relatively
small district magnitude in these two cases (see Table 3.4), it is no surprise to
find that this system is regularly referred to as semi-proportional, and cer-
tainly its record in this regard does not bear favourable comparison with other
PR systems (Farrell 2001; Katz 1994; Lijphart 1994; Taagepera and Shurgart
1989; see also Table 1.1). Accordingly, in the first two sections of this chapter
we assess the records of AV and STV with regard to proportionality. I n our
examination of STV, given the larger number of cases that are available, we
are in a position to carry out a definitive test of the degree of proportionality
of the system, showing the extent to which the variations in district magni-
tude (characteristic I of Table 3.4) and assembly size can affect proportion-
ality, and what consequences this can have for the numbers of parties in the
system.
I n the final section of this chapter, our attention shifts from the theme of
proportionality regarding the aggregate effect of electoral systems on parties
to a consideration of how certain properties of the preferential systems may
determine the fate of individual candidates. This relates to the review in
chapter 1 of the social choice debates over electoral systems. Following, in
particular, the argument put forward by Levin and Nalebuff (1995) and Sen
(1995), our interest is in examining certain properties of the preferential
systems and how these may affect the election of some candidates. This
section will focus on two features that were raised in the previous chapter,
namely the rule regarding the compulsory expression of preferences (we will
| 79 |
| Preferenti al systems and thei r consequences |
focus on its application in the case of AV), and the alternative ways of
transferring surpluses under STV.
I n summary, then, this chapter seeks to answer the following three
questions:
in terms of election outcomes, how different is AV from a simple
member plurality system?
how proportional is STV?
to what extent is the electoral fate of candidates affected by minor
differences in electoral system design?
AV as a non-proportional system
Table 4.1 summarises the electoral record of AV in post-war national
House of Representatives elections. Following Lijphart s rule (1994: 13)
that an electoral system change occurs when there is at least a 20 percent
shift in the size of the assembly,
1
the table reports two sets of figures: one
covering the 15 elections from 1949 to 1983, the other reporting trends in
the eight elections that have occurred since the increase in the number of
parliamentary representatives in 1984. The first column of data sum-
marises the disproportionality trends across the period,
2
revealing high
levels of disproportionality, which are comparable with trends among other
non-proportional systems at the same time, such as SMP (Farrell 2001;
Lijphart 1994).
| 80 |
TABLE 4.1 DISPROPORTIONALITY IN HOUSE OF REPRESENTATIVES (AV) ELECTIONS
Period (N elections) Disproportionality (GI) Two-party preferred vote
share minus seat share
a
194983 (15) 8.55 6.95
19842004 (8) 9.94 6.27
NOTES
GI: Gallagher index. According to Lijpharts (1994: 13) convention, electoral system change occurred in 1984 when the size
of the House was increased, resulting in a change of greater than 20 percent in the Assembly Size.
a: two-party preferred votes and seat shares for the ALP.
SOURCES Disproportionality trends supplied by Arend Lijphart, updated from the Australian Electoral Commission (AEC)
website; two-party preferred vote data from the AEC website.
The increase in the size of the index in more recent elections reflects the
growing numbers of parties fielding candidates (and increased numbers of
independents) in recent elections (Sharman et al. 2001). AV has also pro-
duced some notably anomalous results, such as the systematic bias in favour
of the Liberal Party in nine elections (1949, 1955, 1958, 1963, 1975, 1977,
1980, 1996 and 2001) in which it was awarded more seats than Labor
despite having won fewer votes. Even when we exclude the inevitable bias
against smaller parties and simply deal with the share of the vote between
the two largest parties, the trends are not dissimilar. The final column in
Table 4.1 reports the differences between the share of the two-party pre-
ferred vote and the share of the seats in post-war elections,
3
revealing high
levels of distortion between the proportions of the votes won by one of the
two larger political groupings and their share-out of seats: once again, the
trends reveal a systematic bias against Labor.
According to the disproportionality index, AV produces levels of distor-
tion in the electoral result similar to those produced by the SMP system,
but to what extent does it actually produce different election results to those
produced by the simpler, less-demanding system? To put it another way,
| 81 |
TABLE 4.2 NUMBER OF COUNTS BEFORE WINNER DECLARED AND EFFECTIVE NUMBER OF ELECTORAL
PARTIES, HOUSE OF REPRESENTATIVES ELECTIONS, 19832001
1983 1984 1987 1990 1993 1996 1998 2001
Effective number of electoral parties (N
v
)
2.67 2.77 2.90 3.37 2.91 3.21 3.46 3.42
Number of counts (%)
1 73.6 69.6 63.2 39.9 56.5 56.1 33.8 42.0
2 7.2 5.4 8.1 12.8 2.0 2.0 2.0 2.7
3 10.4 12.8 15.6 12.1 4.7 3.4 3.4 5.3
4 7.2 8.1 10.2 12.8 7.4 7.4 5.4 13.3
5 0.8 3.3 2.7 9.4 6.1 12.2 18.2 10.7
6 0.8 0.7 6.0 6.1 7.4 12.8 8.7
7 5.4 9.5 4.1 10.8 8.7
8 1.3 3.4 4.1 4.7 3.3
9 2.0 3.4 4.7 2.0
10 2.0 3.4 2.0
11 0.7 1.3
SOURCES N
v
data supplied by Arend Lijphart, up-dated by authors; for data on numbers of counts, 198393 (calculations by
Shaun Bowler); 19962001 (authors calculations).
what evidence is there that the use of preferential voting actually makes a
difference to the result?
The common perception is that AV behaves in all its particulars like
SMP (Rae 1967: 108), that what difference it makes to the election result is
less than most people suppose (Butler 1973: 96). I n the first large scale
analysis of trends, Joan Rydon (1956) found little evidence of preferential
voting influencing the overall result; for the most part, its only effect tended
to be on the distribution of seats between the National/Country and Liberal
parties. There was also evidence in the 1960s of preferences from the
Democratic Labor Party (DLP) a breakaway from Labor favouring the
Coalition parties (Goot 1985). Apart from these instances, preferences were
seen to matter little, and certainly had little bearing on the overall election
result.
According to Table 4.2 and Figure 4.1, this pattern has changed consid-
erably in recent elections, and there are signs that preferential voting can
make a difference: preferences can matter, and in recent elections they seem
to matter even more (see also Reilly 2001). I n Figure 4.1 we see how over
the post-war period (19492001), there has been a growing number of cases
in which preferences were required before the result was known (or, to put
this another way, a declining number of cases in which the election was
determined, as for SMP, on the first count). The rise has been continuous
over the decades, with a noticeable acceleration in the 1990s: in the 1950s
there was an average of 12 percent of cases where preferences were used to
determine the result; in the 1960s this more than doubled to 27 percent; it
increased gradually in the 1970s (30 percent) and 1980s (31 percent), before
jumping in the 1990s to 54 percent. I n 2001, 58 percent of constituency
counts were decided by preferences.
The sudden increase over the past decade is also reflected, in Table 4.2, by
a rise in the total number of counts (including a few cases in which 10 or
more counts were required before the final result was determined). Clearly,
the increased significance of preferential voting is related to the growing
numbers of parties fielding candidates (as shown, in Table 4.2, by the effec-
tive number of electoral parties; see also Sharman et al. 2001), but it also
reflects a growing willingness of voters to make use of their preferences in
order to vote strategically, as well as growing attention by the parties to the
strategic potential offered by the AV system. We return to this theme of
strategic voting in chapter 6.
| 82 |
Preferences may matter more in the determination of the final result, but
there is no evidence that preferences actually make much difference to the
final outcome. As Figure 4.1 shows, preferential voting tends to make only a
small difference to the final outcome, and this has not been on the increase.
4
For the most part, the proportions remain firmly in single figures.
STV as a proportional system
I nevitably, as a member of the proportional family of electoral systems,
STV has tended to attract more attention than AV in the comparative
debates. However, two main difficulties have been raised in the assessment
of the proportional consequences of STV. First, the relatively low level of
district magnitude (at least as it is applied in I reland, which is the usual
focus of attention) means that STV tends to be labelled as less propor-
tional, or as the phrase goes quasi-proportional (Taagepera and
Shugart 1989: 207; also Katz 1984). One way around this problem is
simply to ignore district magnitude and focus instead on the electoral
formula. This is the approach adopted by Lijphart (1986) and also by
Blondel (1969) who goes so far as to suggest that STV is the most propor-
tional system.
| 83 |
FIGURE 4.1 PROPORTION OF SEATS AFFECTED BY THE DISTRIBUTION OF PREFERENCES, HOUSE
OF REPRESENTATIVES ELECTIONS, 19492001
0
10
20
30
40
50
60
70
(%)
Preferences used
Preferences change result
1
9
4
9
1
9
5
5
1
9
6
3
1
9
7
2
1
9
7
7
1
9
8
4
1
9
9
3
2
0
0
1
SOURCE 194977 (Rydon: 1986); 19802001 (authors calculations).
A second difficulty is that, because STV is quintessentially a candidate-
based system, it causes difficulties for measures of proportionality that are
based on vote and seat shares for parties(Gallagher 1975; Mair and Laver
1975). Scholars have approached this problem in a number of ways. Rae
(1967: 38) makes the admission that, because of this, [i]t is not quite clear
how this arrangement is likely to compare with other PR formulae, and
accordingly he does not attempt an overall assessment of the degree to which
STV is more or less proportional than list systems. I nstead, he merely con-
cludes that in general, the I rish formula behaves like any other sort of pro-
portional representation. I t operates quite proportionally (p. 111).
Lijphart (1986: 175) has suggested two methods of dealing with this
problem. First, it could be assumed that all voters cast a ballot for the candi-
dates of only one party (so that there are no inter-party transfers;
5
alterna-
tively, one could assume that inter-party transfers cancel each other out).
Second, he proposes analysing the votes in the final round of counting (that
is, after all of the transfers have taken place). Under either method, STV rules
become roughly equivalent to largest remainder-Droop, which Lijphart
(1986; 1994) categorises as being intermediate, between the most propor-
tional and the least proportional of the PR systems. Crucially, according to
Lijphart, STV is a proportional system; it is not merely quasi-proportional.
To test the relationship between district magnitude and proportionality,
and the wider implications for the Australian party system (at national, state
and territory level), this chapter adopts two main approaches. First, we carry
out some simple bivariate examinations of trends among the main variants of
STV, employing Lijphart s (1994) methodology of defining an electoral
system as a set of essentially unchanged electoral rules. Second, we follow the
more common practice of treating individual elections as units of analysis
(Katz 1997; Rae 1967), thereby increasing our number of cases and so
enabling a more extensive multivariate analysis. Given that we are dealing
with the same electoral formula throughout, the main independent variables
of note are district magnitude (M) (or its alternative, effective threshold
[T
eff
]), and assembly size (AS). Lijphart s rule is that there should be a 20
percent change in M, T
eff
, or AS to cause a new electoral system. As Table
4.3 indicates, this results in 16 STV cases (see Appendix Table A.1).
The 16 cases are listed in terms of increasing size of M, which is defined,
in the usual manner, as the average number of seats per constituency. Lijphart
(1994) building on Taagepera and Shugart (1989) proposes a refinement
of M, effective threshold (T
eff
), which is designed to take account of the two
| 84 |
| 85 |
TABLE 4.3 STV SYSTEMS, PROPORTIONALITY AND THE NUMBER OF PARTIES, 19462005
Period (N
elections)
District
magnitude
(M)
Effective
threshold
(T
eff
)
Assembly
size
(AS)
Dispropor-
tionality
(GI)
Effective
number
electoral
parties (N
v
)
Effective
number
parliament-
ary parties
(N
s
)
Ireland 19482002
(17)
3.82 15.5 154.41 3.87 3.26 2.88
Australia3 197780
(2)
4.25 14.2 34 4.41 2.73 2.37
Tasmania2 19982002
(2)
5 12.5 25 6.61 2.72 2.25
Malta1 194755
(5)
5 12.5 40 3.79 2.97 2.80
Malta2 196281
(5)
5.10 12.3 57.00 3.73 2.33 2.14
Australia4 19842004
(7)
5.11 12.3 40.86 5.44 3.23 2.58
Australia2 195370
(7)
5.17 12.1 31 4.97 2.53 2.25
ACT 19952004
(4)
5.67 11.2 17 7.41 3.50 2.77
Western
Australia
19892005
(5)
5.67 11.2 34 7.11 3.08 2.63
Tasmania1 194696
(16)
6.69 9.7 33.44 3.99 2.44 2.19
Australia1 1949
(1)
7 9.3 42 3.42 2.19 1.98
Australia
DD2
197587
(3)
8.50 7.9 68.00 3.02 2.56 2.36
Australia
DD1
195174
(2)
10 6.8 60 3.38 2.29 2.07
South
Australia
19822002
(6)
11 6.2 11 6.14 3.04 2.58
New South
Wales1
197891
(5)
15 4.6 15 4.02 2.59 2.37
New South
Wales2
19952003
(3)
21 3.4 21 4.99 3.77 3.43
NOTES Cases ranked by M. By assembly size, we mean the number of legislators being elected in a particular election (i.e. in the
case of staggered elections, such as NSW or SA, this measure does not take account of the actual assembly size). We follow
Lijpharts convention that all integers indicate exact and unchanging numbers, all other numbers indicate averages: because T
eff

is only an approximate value, it is recorded to just one decimal place (Lijphart 1994: 29). An electoral system change occurs
whenever either M or AS shifts by at least 20 percent (p.13). The Australia DD scores refer to double dissolutions.
SOURCES Election returns; Electoral Commission websites; Black (1989); information supplied by Arend Lijphart, Michael
Gallagher, and by parliamentary libraries.
main thresholds in electoral systems: the threshold of inclusion (the
minimum vote a party requires to win a seat) and the threshold of exclusion
(the maximum support a party can attain without winning any seats).
6
The
resulting figure (seen as an estimate and therefore presented as a value to only
one decimal place) is described as a midpoint in a range between no repre-
sentation and full representation.
7
A number of observations can be made about Table 4.3. First, and yet
again, we see real evidence of the extent to which Australian politicians, par-
ticularly at national level, are prone to fiddle with the electoral system. I n par-
ticular, there have been two changes to the national system which, according
to Lijphart s rule, qualify as electoral change.
8
The granting of two seats each
to the ACT and the Northern Territory in 1975 resulted in a large drop in
average M, causing a change to new systems for half-Senate and full-Senate
elections. Then the expansion in the membership of the Senate in 1984
resulted in an increase in average M, which in the case of half-Senate elec-
tions was above the Lijphart 20 percent cut-off. Other systems to undergo
significant shifts over the period have included:
Malta, which increased its AS in the early 1960s
New South Wales which increased its AS in 1991
Tasmania (somewhat controversially), which reduced its average M from
7 to 5 in 1998 in a deliberate strategy to impede the electoral prospects of
small parties.
9
Second, Table 4.3 provides an important corrective to the view that STV is a
small-M system (and hence its accolade as a quasi-proportional electoral
system). For instance, in Lijphart s (1994) study, STV has the lowest M of all
the proportional systems. He argues that this is inevitable because high M
(that is, large constituencies) entails large numbers of candidates and thereby
imposes heavy burdens on the voters who have to rank order the candidates
(Lijphart 1994: 30). The use of ticket voting in the Australian upper house
STV systems has the important consequence of facilitating a larger M; but
the ACT and Tasmania (with Ms of 5.67 and 6.69 respectively [T
eff
= 11.2
and 9.7]) demonstrate that STV systems can be more proportional without
experiencing the strictures of ticket voting. As indicated by the rank orders in
Table 4.3, adding the various Australian cases to the population of STV
systems (which generally are taken to include just Ireland and Malta)
increases average M and decreases average T
eff
, in both cases quite consider-
ably. The mean M for all 16 STV cases is 7.8. This compares favourably with
| 86 |
an average M of 9.0 for the 13 dHondt systems in Lijphart s study, and
average score of 7.7 for the four other non-dHondt systems.
10
A third observation to be made about Table 4.3 relates to the apparent
lack of any relationship between the independent variables (M/T
eff
and AS)
and the dependent variables (disproportionality and the numbers of parties in
the system).
11
Because it is not always the case that systems with large assembly sizes
also have the lowest effective thresholds, we need to take account of the con-
tradictory pressures that these two variables can produce. For instance, as
Table 4.3 shows, I reland has both the largest AS (favouring high propor-
tionality) andthe highest T
eff
(favouring low proportionality). The effects of
AS and T
eff
on disproportionality and the effective numbers of parties (N
v
and N
s
) are explored in Table 4.4.
12
Once T
eff
is taken into account, there is
some evidence of a relationship between AS and disproportionality,
although only in the case of low effective threshold where, apart from New
| 87 |
TALE 4.4 AvERA0E 0l3PR0P0RTl0NALlTY AN0 NuV8ER3 0F PARTlE3, CLA33lFlE0 8Y EFFECTlvE
TlRE3l0L0 AN0 A33EV8LY 3lZE, lN 1 ELECT0RAL 3Y3TEV3
E|eotora| system
& effeot|ve
thresho|d
Assemb|y s|ze
[A8}
0|sproport|on-
a||ty [S|}
Effeot|ve number
e|eot part|es [N
v
}
Effeot|ve number
par| part|es [N
s
}
l|g| T
ell
(11.2-15.5)
lre|ard 151.11 3.8Z 3.2 2.88
Va||a2 5Z 3.Z3 2.33 2.11
Aus|ra||a1 10.8 5.11 3.23 2.58
Va||a1 10 3.Z9 2.9Z 2.80
Aus|ra||a3 31 1.11 2.Z3 2.3Z
wes|err Aus|ra||a 31 Z.11 3.08 2.3
Aus|ra||a2 31 1.9Z 2.53 2.25
Tasrar|a2 25 .1 2.Z2 2.25
ACT 1Z Z.11 3.50 2.ZZ
LoW T
ell
(3.1-9.Z)
Aus|ra||a002 8 3.02 2.5 2.3
Aus|ra||a001 0 3.38 2.29 2.0Z
Aus|ra||a1 12 3.12 2.19 1.98
Tasrar|a1 33.11 3.99 2.11 2.19
NeW 3ou|| wa|es2 21 1.99 3.ZZ 3.13
NeW 3ou|| wa|es1 15 1.02 2.59 2.3Z
3ou|| Aus|ra||a 11 .11 3.01 2.58
80UR6E8 As lor Tao|e 1.3.
South Wales, there is a monotonic rise in disproportionality as the assem-
blies become smaller. I n the case of high effective thresholds, there is little
evidence of a consistent trend. What this table shows is that, although
double-dissolution Australian Senate STV may have the lowest dispropor-
tionality of all the versions of STV, I reland and Malta despite their rela-
tively low M (high T
eff
) tie for the lowest disproportionality among these
high effective threshold systems. I n Table 4.4 there is no evidence of any
relationship between our independent variables and the effective numbers of
parties indices.
We can extend this analysis by using multiple regression to assess the
impacts of effective threshold and assembly size on proportionality. I n this
way, we can estimate which of the nine STV systems being considered is
likely to lead to the most proportional outcome.
13
These estimates are made
by calculating ordinary least squares (OLS) regression equations predicting
proportionality from a range of independent variables. I n the first equation in
Table 4.5, T
eff
and AS only are used, the latter re-estimated as a logarithm, in
line with Lijphart s model (1994). I n equation 2, we include dummy variables
| 88 |
TALE 4.5 EvALuATlN0 PR0P0RTl0NALlTY lN 3Tv 3Y3TEV3 (RE0RE33l0N C0EFFlClENT3)
Equat|on 1 Equat|on 2
o oe|a o oe|a
Ellec||ve ||res|o|d .22 .3 .19 .30
Assero|y s|ze (|og) -1.5 -.58 -.15 -.0
Cour|r|es (Aus00)
- ra -
ACT 3.33 .33
N3w 1.5 .23
3A 2.91 .35
Tas .52 .10
wA 3.11 .35
Aus| .82 .1
Va||a -.38 -.0
lre|ard -.9 -.13
Cors|ar| 8.0Z 2.39
Adj R-sq .15 .19
(N) (90) (90)
, s|a||s||ca||y s|gr|lcar| a| p<.05, p<.10.
N0TE Par||a| ard s|ardard|sed regress|or coellc|er|s pred|c||rg ||e 0a||ag|er |rdex ol d|spropor||ora|||y. lr equa||or 2,
Aus00 |s ||e exc|uded ca|egory. T|e grea|er ||e |rdex ||e grea|er ||e d|spropor||ora|||y.
80UR6E As lor Tao|e 1.3.
for eight of the nine STV variants, with Australian double dissolution elec-
tions forming the reference category. The figures are partial and standardised
regression coefficients; the former show the change in the proportionality
index caused by a change in the independent variable in question; since the
disproportionality indices are scored from low (least disproportional) to high
(most disproportional), a positive coefficient indicates greater disproportion-
ality. The standardised coefficients show the relative importance of a partic-
ular variable in the equation in question.
Based on the simple two-variable equation in Table 4.5, T
eff
emerges as
having a modest influence on disproportionality, a higher threshold resulting
in greater disproportionality. By contrast, AS has a much stronger impact
than T
eff
, with large assembly sizes leading to less disproportional results, as
we would expect since smaller assemblies will lead to more distortion in
translating votes into seats. The stronger effect of the assembly size probably
reflects, at least in part, the fact that we are dealing with a large range
(11165), whereas in the case of the effective threshold the dispersion is far
less (3.416.9). Once the specific STV variants are added to the equation in
terms of dummy variables (equation 2 in Table 4.5), we can see which coun-
tries, by virtue of the size of their elected assemblies, are less disproportional
than others (the effective threshold variable is no longer statistically signifi-
cant). Most of the systems are less proportional than Australian double dis-
solution elections, the excluded category, but otherwise the different
measures of proportionality produce different (albeit modest) effects. Relative
to the excluded category, the ACT is the most disproportional once T
eff
and
AS are taken into account and is closely followed by Western Australia and
South Australia; Malta is the least disproportional.
14
The data also enable us to address the question of the consequences of
these electoral systems for the size of their party systems, measured as the
effective number of parliamentary parties (N
s
). These estimates are made in
Table 4.6. Once again, the results are divided into two equations, the first
excluding the separate STV variants, the second including them. The first
equation shows that larger party systems are a consequence of a larger
assembly, but that neither the proportionality of the system nor the effective
threshold has any significant effect on N
s
, net of other factors. The second
equation shows that none of the core variables has any significant effect,
although assembly size becomes more important (the fact that this larger
coefficient fails to reach statistical significance is a consequence of the
greater number of independent variables in the equation). Taking these
| 89 |
factors into account and judging the coefficients in relation to Australian
double dissolution elections, the results show that all of the systems have
larger party systems than the excluded category, net of other things. The
ACT has the largest party system, and has the only coefficient that reaches
statistical significance, followed by New South Wales, South Australia, and
Western Australia.
Reflecting our smaller number of cases, the analysis of disproportion-
ality has not produced results as definitive as those given by Lijphart
(1994) in his study of a larger number of electoral systems across a wider
range of countries. Two points in conclusion are worth making. First, AS
appears to also have important effects on disproportionality; in this
instance its role is even more significant than suggested by Lijphart (1994:
110). Second, contrary to our expectations, the Australian Senate does not
produce markedly more proportional results than the other systems. I n
| 90 |
TA8LE 1. PRE0lCTlN0 TlE EFFECTlvE NuV8ER 0F PARLlAVENTARY PARTlE3 lN 3Tv 3Y3TEV3
(RE0RE33l0N C0EFFlClENT3)
Equat|on 1 Equat|on 2
o oe|a o oe|a
Ellec||ve ||res|o|d 1.81 .11 .01 .02
Assero|y s|ze (|og) .19 .32 .2Z .11
0|spropor||ora|||y .03 .11 -.01 .0
Cour|r|es (Aus00)
- ra -
ACT .89 .39
N3w .83 .51
3A .Z8 .12
Tas .1 .13
wA .5Z .28
Aus| .33 .28
Va||a .35 .23
lre|ard .1Z .39
Cors|ar| 1.81 1.1Z
Adj R-sq .02 .19
(N) (90) (90)
, s|a||s||ca||y s|gr|lcar| a| p<.05, p<.010.
N0TE Par||a| ard s|ardard|sed regress|or coellc|er|s pred|c||rg ||e ellec||ve ruroer ol par||arer|ary par||es (N
s
).
0|spropor||ora|||y |s reasured oy ||e 0a||ag|er |rdex. lr equa||or 2, Aus00 |s ||e exc|uded ca|egory.
80UR6E As lor Tao|e 1.3.
other words, despite the cost to free and fair choice due to the operation
of ticket voting, the Australian Senate does not reap any particular bene-
fits from greater proportionality. (The theme of voter choice is taken up in
chapter 6.)
Preferential systems and the
fate of individual candidates
The previous chapter showed how both AV and STV vary in a number of
respects. So far in this chapter we have seen how this can influence overall
proportionality (particularly in the case of STV), thereby determining how
many seats each party wins. Some of the variations can also affect the fate of
individual candidates, which, as social choice theorists argue, demonstrates
the logical shortcomings of such systems. Following the work of scholars such
as Levin and Nalebuff (1995), Saari (2001a) and Sen (1995), it is useful to
assess how these systems operate in their own right, rather than the more
usual social choice practice of judging them against pre-defined axioms. I n
doing this, we examine how certain variations in the detail of how the system
operates in some cases apparently innocuous variations may have impor-
tant effects on the electoral outcome. We review two features here: the rules
regarding how many preferences are required under AV, and the procedures
for transferring surpluses under STV.
I n the last chapter (Table 3.3), we examined the three different types of
AV system, distinguishing between compulsory preferential voting, optional
preferential voting and the contingent vote. Like the optional and compul-
sory preferential voting systems, the contingent vote is a majoritarian elec-
toral system, in the sense that a candidate is elected with an overall majority
of the vote; nevertheless, the properties of the system can result in the elec-
tion of a different candidate to one who would have been elected under either
of the other two variants. I n effect, this produces what social choice theorists
would refer to as a non-monotonic result. As Reilly (1997: 100) argues, this
is particularly likely in the case of high candidature because the votes of
lower-placed candidates are more likely to see a candidate initially placed
third or fourth gain enough preferences under [AV] rules to leap-frog one of
the two leaders during the distribution of preferences. For this reason, the
contingent vote is preferred by the larger parties, particularly in a context
(such as in London in 2000) where there are likely to be a large number of
fringe candidates.
| 91 |
A recent example of how the contingent vote can produce a different
result to the other AV variants is provided by the count in the Queensland
constituency of Blair in the 1998 House of Representatives election (Table
4.7). The One Nation leader, Pauline Hanson, won 36 percent of the first
preference votes. However, she was ultimately defeated in the eighth count
by the Liberal Party candidate, Cameron Thompson, who had received 22
percent of the first preference votes. Prima facie, this was an excellent result
for AV enthusiasts, because it demonstrated how the use of preferential
voting could make a difference: had the election been fought under the SMP
system, Hanson would have been elected comfortably. However, a signifi-
cant additional detail is the fact that Thompson not only defeated Hanson
but also beat the Labor candidate, Virginia Clarke, whose first preference
tally had amounted to 25 percent, placing her in second position. Had the
election been fought under contingent vote rules, Thompson and the other
six candidates would have been excluded after the first count and all their
| 92 |
TALE 4.7 TlE l0u3E 0F REPRE3ENTATlvE3 ELECTl0N C0uNT lN TlE 0uEEN3LAN0 0lvl3l0N 0F 8LAlR,
1998
6ount HoKenz|e,
Nea|
[0EH}
wh|te,
rett
[NP}
'wa|ghte,
0wen
[6E6}
6|arke,
V|rg|n|a
[ALP}
6onnors,
L|bby
[SRN}
hanson,
Pau||ne
[hAN}
Thompson,
6ameron
[LP}
Roberts,
Lee
[|N0}
8|oan,
Hark
[A68}
Tota|
1 21Z8 989 199 1Z239 1230 2151 11Z8Z 55 1Z0 811
2 '7 '' 5 '8 '6 40 '9 44 Exo|uoeo '70
2195 Z000 201 1Z25Z 121 2155 1180 00 811
3 26 44 Exo|uoeo 20 '2 74 'J '5 204
2521 Z011 1Z2ZZ 1258 2130 11819 15 811
1 '28 7' 62 ''' 'J6 '07 Exo|uoeo 6'5
219 Z115 1Z339 139 21Z 1192 811
5 590 89 J85 Exo|uoeo '88 ''7 'J69
3239 Z201 1ZZ21 21951 15013 811
Exo|uoeo 926 ''86 J99 728 J2J9
8130 18910 25353 15ZZ1 811
Z Excluded '064 ''72 5894 8'J0
199Z1 2525 215 811
8 Exo|uoeo 524' '47JJ '9974
31Z 3398 811
Elected
80UR6E Aus|ra||ar E|ec|ora| Corr|ss|or C0-Ror
second preferences would have been transferred to the top two candidates
Hanson and Clarke. Undoubtedly, we would have had a different election
result, and it is interesting to speculate over who the victor might have been.
A scrutiny of the transfer patterns in Table 4.7 suggests some potential for
Hanson to have picked up a lot of stray preferences: while we cannot be categor-
ical over whether she could have won the seat under these circumstances, the
trends point quite strongly in her favour. Consider the following three points:
On the first count, Clarkes vote was 30 percent less than Hansons, and
after seven counts (the count preceding Thompsons eventual victory) this
difference was still 25 percent.
On that seventh count, while the bulk (73 percent) of the votes of the
excluded National Party candidate, Brett White, transferred to
Thompson, the remaining transfers were divided evenly between Clarke
and Hanson, and if Thompson had not been in the count at that point,
Hanson would have been the only right-of-centre candidate in the
running.
I n the earlier counts eliminating the independents and small parties, with
the exception of the Australian Democrats, there was a healthy flow of
transfers to Hanson. I n some cases, she picked up a plurality of the trans-
fers available.
The discussion in the preceding paragraphs demonstrates not only how dif-
ferent AV is from other non-proportional systems like SMP, but also provides
some important clues about the importance for electoral outcomes of ballot
structure. While, in the aggregate, the number of cases in which preferential
voting can make a difference to the final election result may be small (as
shown by Figure 4.1), it would be wrong to conclude from this that somehow
preferential voting is irrelevant to electoral outcomes. Allowing voters more
choice on the ballot paper can have an important effect on some key con-
stituency results, as Pauline Hanson found to her cost in 1998,
15
although it
is also important to take into account the absence of any preference arrange-
ments for One Nation, which could have improved her vote. As we shall see
in chapter 5, the parties are very aware of the importance of preferential
voting and make every effort to ensure that it works to their advantage.
Based on the evidence from New South Wales and Queensland the two
states operating optional preferential voting (Table 3.3) there would appear
to be good reason for the politicians in other parts of Australia to prefer to
keep compulsory preferential voting. According to their analysis of the
| 93 |
trends, Reilly and Maley (2000: 46) find evidence that the rate of plumping
(that is, expressing just one preference, and thereby turning an AV election
into an SMP election) is on the increase in both states. Their research is but-
tressed by a more recent analysis of trends in the 2003 New South Wales leg-
islative assembly election, which shows that one in nine (10.9 percent) of the
ballots were exhausted before the end of the count.
16
And, in the case of the
preceding 1999 election, according to the New South Wales Liberal party,
optional preferential voting cost them seats because large proportions of
Liberal and National supporters plumped for the candidates of their respec-
tive parties; the coalition parties lost out from a lack of cross-over support
(Humphries 1999).
The fate of individual candidates is also affected by variations in how
surplus votes are transferred under STV. The case of Labor candidate
Malcolm Colston in the 1974 election, which we reviewed in chapter 3, is
illustrative of one feature of this, namely the problems that can be encoun-
tered at later stages of an election count when only the last parcel received is
made use of (see Fischer 1981). Both under the random selection method,
then in use for Senate elections, and the Gregory method, the only other
alternative used in Tasmania, only those votes received in the last parcel are
available for transfer to the remaining candidates.
I t is possible to imagine scenarios in which the method of transferring
surplus votes could have a significant affect on the election result. Take the
case of candidate A, who is a fundamentalist Christian and who is swept to
victory on a strong anti-abortion manifesto. A certain proportion of As sup-
porters gave their second preferences to candidate B, who, despite holding
more secular views on politics, is also a right-of-centre politician. I t is likely
that a large proportion of Bs supporters, who gave her their first preference,
did so on the basis of her right-of-centre views on the economy; if anything,
many of them may have been put off by As views on abortion. We can expect,
therefore, that the patterns of preferences of those who gave A their number
one vote is likely to show some significant difference to those giving B their
number one. Under the random selection and the Gregory methods, only the
parcel of ballot papers received from A would be available for transfer, and it
is debatable whether they represent an accurate sample of the average sup-
porter of B.
Such a scenario which had featured in the 1974 election weighed in
the minds of some of the advisors involved in drawing up proposals for
amending the Senate system a decade later, and in large part this explains the
| 94 |
reason for adopting the inclusive Gregory method (see pp. 6364 above).
Unlike the Gregory method, in which the transfer value (TV) for the ballots
being transferred is calculated as s/a(where sis the candidates surplus vote
and ais the last parcel of ballot papers received), under the inclusive Gregory
method TV = s/n, where nis the total number of ballot papers in the candi-
dates pile. This removes the problem of Bonner syndrome (see p. 64)
because all ballot papers are included, not just the last parcel received. I n fact,
however, it merely replaces one potential anomaly with another. As critics
point out, under the inclusive Gregory method, it is possible for a ballot
papers TV to increase in later counts, thereby attaching undue weight to
some ballot papers and insufficient weight to others. As ever, and as we shall
see below, such an anomaly brings with it the danger of electing the wrong
candidate.
To avoid this potential pitfall, it has been suggested by, among others,
the Proportional Representation Society of Australia that all ballot papers
should be included in calculating TVs, but in order to do so correctly, the
contention is that all past TVs should be given appropriate weights to reflect
their contribution to previous counts (see Dummett 1997, 129; Meek 1994,
23). This is achieved by taking account of the TVs that were applied in pre-
vious counts. The adaptation that has been recommended by the Proportional
Representation Society, which we shall call the weighted inclusive Gregory
method, produces the following procedure for determining the TV for a can-
didates surplus votes. For those votes that the candidate has received at full
value, TV = s/v, where v is the candidates total vote. For those votes that the
candidate has received from another candidates surplus, TV = (s/v), where
is the TV that was applied in the transfer of the surplus votes to the pre-
vious candidate.
I n recent years, the Proportional Representation Society of Australia has
downplayed this option, favouring instead a strategy of simply trying to
replace the inclusive Gregory method with the Gregory method. But given
that, in certain prominent cases, there has been a shift towards computer-
aided counting for example, the ACT, Western Australia and recent
experiments in I reland there are now grounds for giving the weighted
inclusive Gregory method serious consideration. For that matter, attention
might also be given to even more elaborate revisions of STV counting rules,
such as the Meek method (Hill 1994; Meek 1994) recently adopted by New
Zealand for its district health board elections and possibly also for local
government elections.
17
The Meek method builds a number of innovations
| 95 |
into STV, most notably that an elected candidate should continue to receive
vote transfers until the count is completed (thereby ensuring that the
transfer of votes proceeds in the original order set by the voters on their
ballot papers) and that the quota should be recalculated throughout the
count to take account of non-transferable ballot papers. The implication of
this is that the count process becomes so complex that it can only be done
by computer. Under the Meek method, the transfer of vote surpluses is very
similar to the weighted inclusive Gregory method outlined above, in that it
takes account of all ballot papers and attaches appropriate weights to those
ballots transferred on the basis of previous surpluses (Meek 1994), and for
this reason we shall not include it in the analysis that follows.
18
The difference between the three types of Gregory method, and their
effects on election outcome, is best illustrated by way of a hypothetical
example.
19
Shirley has received 35000 first preference votes, but is still short
of the quota of 50 000. I n the second count, she receives a further 10 000
votes from Toms surplus, representing 100 000 ballot papers at a TV of 0.1.
Ultimately, Shirley is elected in the third count, when she receives 25000
| 96 |
TALE 4.8 vARlATl0N3 lN TlE vALuE 0F 3uRPLu3 8ALL0T3 uN0ER 0lFFERENT 3Tv C0uNTlN0
PR0CE0uRE3: A lYP0TlETlCAL EXAVPLE
8h|r|ey's ba||ot papers
1st oount
8h|r|ey's 35 000 rst
preferenoes
2nd oount
100 000 ba||ot
papers from Tom
3rd oount
25 000 ba||ot papers
from 0|ok
(0regory re||od)
lrcor|rg va|ue 1 0.1 1
0u|go|rg va|ue 0 0 0.8
Cor|r|ou||or |o surp|us () 0 0 100.0
(lrc|us|ve 0regory re||od)
0u|go|rg va|ue 0.125 0.125 0.125
Cor|r|ou||or |o surp|us () 21.9 2.5 15.
(we|g||ed-|rc|us|ve 0regory re||od)
0u|go|rg va|ue 0.28 0.029 0.28
Cor|r|ou||or |o surp|us () 19.9 11.5 35.
N0TE Ca|cu|a||ors are suojec| |o rourd|rg errors.
votes, at full value, from Dick who had been excluded after the second count.
Shirleys vote total of 70 000 ensures her election, with a surplus of 20 000
votes. The fourth count involves the distribution of her surplus. I t is at this
point that the differences between the different forms of Gregory method
become all too apparent.
The differences in outcome are revealed in Table 4.8. I n the case of the
classic form of the Gregory method (as used in the ACT and Tasmania), the
potential for Bonner syndrome is shown by how neither Shirleys 35 000 first
preferences nor the 10000 vote surplus she received from Tom make any con-
tribution to her 20000 vote surplus. I nstead, the only ballot papers to transfer
are those that she received in the third count from Dick. The inclusive
Gregory method (used in the Australian Senate, South Australia and
Western Australia), by contrast, does take account of all of Shirleys ballot
papers, but here the problem is that the transfer value of Toms ballot papers
have actually increased from 0.1 to 0.125. To put it another way, the 100 000
ballot papers that were worth 10 000 when received from Tom are now worth
12 500. I n effect, this gives more than one vote to all those voters involved in
first voting for Tom.
The weighted-inclusive Gregory method avoids both anomalies: all ballot
papers are included, and the transfer values are weighted appropriately. To
date, the Proportional Representation Society of Australia has been unsuc-
cessful in its attempts to persuade the JSCEM to change the method used for
transferring surplus votes. This is despite the fact that the Society has pro-
duced evidence of real-world cases where TV has increased in value.
20
The
most likely reasons why the JSCEM has not been minded to recommend
change is a feeling (certainly one held by its advisors in the Australian
Electoral Commission) that such occasions of where TV increases in value
are likely to be very rare.
Whether or not the types of anomalies identified in the two versions of
the Gregory method currently in use in Australia actually affect real election
results, it is still worth stressing that in each version of the Gregory method,
different values are being given to the ballots papers that are available for
transfer. This is shown by the contribution to surplus entries in Table 4.7.
Compare, for instance, the values for the 25 000 ballot papers that Shirley
received from Dick across the three methods. Under the Gregory method
they are transferred from Shirley with the value of 20000 votes; under the
inclusive Gregory method they are worth 3125 votes; while under the
weighted-inclusive Gregory method, they are worth 7150 votes. Without
| 97 |
having to make any judgement as to which of the three values is most
appropriate, it is apparent that the choice of procedure for allocating surplus
votes is bound to have significant effects on the final election result. I t is
inconceivable that the fate of some candidates will not be influenced by this
decision.
I n 1974, Malcolm Colstons supporters felt they had good reason to argue
that, because account was taken of only the last parcel of ballot papers
received when transferring Neville Bonners surplus votes, Colston had been
robbed of the election. I n 2001, similar accusations flew in the Legislative
Council elections in the Mining and Pastoral region of Western Australia,
except on this occasion the argument was in reverse: the basis of the criticism
now was that only the last parcel of ballot papers should have been taken into
account in determining a transfer.
The relevant details are as follows. John Fischer, a One Nation candidate,
was elected in count 234 as a result of receiving 4397 vote transfers from the
exclusion of Mark Nevill (Independent). Fischers final tally of 11 874 votes
gave him a surplus of 3813 votes to be distributed among the remaining can-
didates. This surplus transfer would determine which of the two remaining
candidates Robin Chapple (Greens) or Greg Smith (Liberal) would be
elected: only 214 votes separated them. In the event, more of Fischers transfers
went to Chapple (1949 votes) than to Smith (1233) and Chapple was elected.
Smiths supporters argued that the wrong candidate had been elected, and
their criticisms (for the most part implicitly) centred on the use of the inclu-
sive Gregory method in Western Australian STV. The criticisms were
twofold. First, it was observed that the weight of Nevills transfers to Fischer
(the bulk of which, under Nevills voting ticket, placed Smith above Chapple)
was greatly reduced when they were being transferred from Fischer, because
of the inclusive Gregory method of taking account of all ballot papers, not
just the last parcel received.
The gist of the argument was that Smith would have won the last seat if
the Gregory method, in which account is taken only of the last parcel received,
had been used instead.
21
This claim which reverses the Colston (1975) crit-
icism of Senate counting rules in 1974 is undoubtedly correct. As can be
seen in Table 4.9, under the Gregory method, only Nevills votes would have
counted and since all but 88 of these placed Smith above Chapple, there can
be no doubt that Smith would have won the seat.
22
However, the point is that
under the existing legislation (the 1907 Electoral Act), the correct procedures
were applied, and therefore Chapple was the correct person elected.
| 98 |
The second criticism made by Smiths supporters and one backed up by
detailed analysis by mathematicians
23
relates to the issue raised above about
the potential for the inclusive Gregory method to inflate the value of a vote.
This centred on 89 ballot papers that Fischer had received (directly and indi-
rectly via other candidates) from the surplus votes of Jon Ford (ALP). As
Table 4.9 shows, Fords transfers were each worth 0.14246115 when received
by Fischer, but their outgoing value increased to 0.26582596. This provides
tangible support for the arguments made above (see Table 4.8) about how the
inclusive Gregory method can result in anomalous situations.
However, as the Western Australian Electoral Commissioner observed
(Evans 2002, iii), given the tiny number of ballot papers involved on this
occasion, there was a negligible effect on the ultimate electoral result. This
is shown by our analysis in Table 4.9 of how the weighted inclusive Gregory
| 99 |
TALE 4.9 vARlATl0N3 lN TlE vALuE 0F 3uRPLu3 8ALL0T3 uN0ER 0lFFERENT 3Tv C0uNTlN0
PR0CE0uRE3: TlE wE3TERN Au3TRALlAN VlNlN0 AN0 PA3T0RAL C0uNT 0F 2001
John F|soher's ba||ot papers
F|soher's
7073 rst
preferenoes
385 ba||ot
papers from
exo|us|ons
24 ba||ot
papers from
8tephens
ba||ot
papers from
Hoore
89 ba||ot
papers from
Ford
4397 from
Nev|||'s
exo|us|on
(0regory re||od)
lrcor|rg
va|ue
1 1 0.5Z1Z2121 0.3829182 0.1121115 1
0u|go|rg
va|ue
0 0 0 0 0 0.Z23902
Cor|r|ou||or |o
surp|us ()
0 0 0 0 0 100
(lrc|us|ve 0regory re||od)
lrcor|rg
va|ue
1 1 0.5Z1Z2121 0.3829182 0.1121115 1
0u|go|rg
va|ue
0.258259 0.258259 0.258259 0.258259 0.258259 0.258259
Cor|r|ou||or |o
surp|us ()
59.1 3.2 0.2 0.0 0.Z 3.Z
(we|g||ed |rc|us|ve 0regory re||od)
lrcor|rg
va|ue
1 1 0.5Z1Z2121 0.3829182 0.1121115 1
0u|go|rg
va|ue
0.2801Z9 0.2801Z9 0.1532590Z1 0.098Z2832 0.038188802 0.2801Z9
Cor|r|ou||or |o
surp|us ()
59.5 3.2 0.1 0.0 0.1 3Z.0
N0TE Ca|cu|a||ors are suojec| |o rourd|rg errors.
method would have performed. The differences between it and the inclu-
sive Gregory method are negligible. I n short, while there may well have
been an anomaly in how the inclusive Gregory method dealt with votes
transferred from earlier vote surpluses, this had no bearing on the eventual
electoral outcome on this occasion. But there is no guarantee that more
serious problems may not emerge in the future.
Conclusion
Variations in the design of AV and STV systems have important conse-
quences for the political system. This chapter has examined their systemic
consequences in terms of two main themes. The first and second sections
considered their proportional consequences in terms of indices of dispropor-
tionality and effective numbers of parties. While there were no surprises in
the case of AV, in that it produces results not unlike other non-proportional
systems, there was evidence nonetheless that preferences matter in the final
determination of the result, and that this influence is on the increase. The
greater range of cases available for STV (especially when we also included
Ireland and Malta) allowed for more detailed and systematic study of the
effects of variations in district magnitude (M) on election outcomes. This
analysis allowed us finally to put to rest the mistaken notion that STV is a
semi-proportional system; in terms of overall proportionality STV compares
very favourably with other PR systems, and certainly far better than implied
in Table 1.1. We were also able to provide evidence that high levels of pro-
portionality do not necessarily require Senate-style ticket voting.
I n the third section, the focus shifted to a consideration of how variations
in the application of preferential systems can affect the fate of individual can-
didates. We saw how the use of an AV system that gives voters a full range of
preferences, as opposed to the contingent vote variant in which voters have
just two preferences, can influence election outcomes in this case examining
the potential for Pauline Hanson to have won the seat of Blair in the 1998
House of Representatives election.
The section also examined how variations in the rules governing the
transfer of vote surpluses under STV can affect the weights given to indi-
vidual ballot papers, which in turn are likely to play a role in influencing the
election of certain candidates. This points to a logical shortcoming in the
operation of the STV system. As we discussed, by varying the type of
Gregory method used, it is possible to influence the weight being attached to
| 100 |
particular vote preferences. I n a close race, this can be crucial in determining
the fate of individual candidates. Such an inconsistency in the treatment of
candidates under the different variants of STV provides support for the the-
oretical criticisms of STV in the social choice literature that were reviewed in
the opening chapter.
A central concern of social choice theory is that an electoral system should
ensure that the correct candidate, the Condorcet winner, is elected. Staunch
critics of STV prominent among them Michael Dummett berate it for
producing quasi-chaotic results, in which the fate of a candidate is deter-
mined by particular features of the counting process, most notably with
regard to how preferences are treated. Given that, as we have seen, the weight
attached to voter preferences can vary dramatically across each of three
Gregory methods reviewed in this chapter, there is strong empirical support
for social choice theoretical criticisms of STV.
The evidence presented in this chapter also provides strong support for
the criticisms levelled, among others, by the Proportional Representation
Society of Australia against the particular variant of Gregory that has been
used for Australian Senate elections (as well as in South Australia and
Western Australia) over the past 20 years. The inclusive Gregory method can
produce a significant anomaly, namely that under certain circumstances a
ballot paper can actually increase in value at later stages of the count. While
the AEC officials may well be correct in the view that the likelihood of such
an occurrence affecting the outcome of an election is very small, the argument
does not end there. The count in the mining and pastoral region of Western
Australia in 2001 provides real evidence that this anomaly occurs. Even if this
had no effect on the election result on this occasion, the fact that it can occur
is surely reason enough to consider changing this procedure. And given the
potential now being offered by computer-aided counting (as used in parts of
Australia), now is an opportune time to re-examine the potential of more
appropriate surplus transfer methods such as offered by the weighted inclu-
sive Gregory method or, indeed, by Meek.
This chapter has reviewed the main systemic consequences of the prefer-
ential systems, or what might be termed their mechanical effects. I n the next
two chapters, the focus shifts to the psychological consequences of electoral
systems in terms of how voters and politicians operate under them and use
them. We start, in chapter 5, with the politicians.
| 101 |
f i v e
The voter versus
the party
The previous chapter examined how Australias preferential electoral systems
can affect the fateof parties and individual candidates. I n this and the next
chapters our attention turns to the impact of these electoral systems on the
behaviour of politicians (this chapter) and voters (chapter 6) the subject of
a large and growing body of literature. I n the case of the politicians, for the
most part, this international research focuses on matters relating to the pro-
portionality of electoral systems questions such as whether PR produces
parliaments as a microcosm, or if there is greater issue congruence between
elected politicians and their voters (e.g. Huber and Powell 1994). I n other
words, according to this literature, what matters is the overall proportionality
of the result, caused by the types of factors dealt with in the previous chapter,
namely district magnitude and electoral formula.
According to this research, those countries using proportional electoral
systems produce a set of parliamentary representatives who are politically
and socially more representative and by implication also attitudinally more
congruent a mirror of the people. This is something closer to the ideal of
representation, classically expounded by John Adams: a parliament that is
an exact portrait, in miniature, of the people at large (quoted in McLean
1991: 173; see also Pitkin 1968: ch. 4). The electoral systems literature has
long been concerned with producing hierarchies of electoral systems based
on how they perform on this dimension (most recently, Farrell 2001;
Lijphart 1999).
| 102 |
Analysis of this sort (of which there is much in the electoral systems
literature) can tell us much about how electoral systems shape the partisan
balance of legislatures (Lijphart 1994; Taagepera and Shugart 1989), and how
representative the body of elected politicians is of the population in terms of
age, social class, gender and other social characteristics what Hannah Pitkin
(1968) refers to as the standing for or descriptive side of representation (e.g.
Lakeman 1982; Phillips 1995; Norris 1985; Rule 1987). Such studies generally
commend proportional electoral systems for promoting more representative
electoral results, which in turn, it is argued, produce a more satisfied electorate.
Throughout this book we are considering another dimension of electoral
systems based principally on a focus on ballot structure relating more to
the theme of how much choicea voter has in determining the fate of individual
candidates. (District magnitude is also relevant here, in terms of the size of
the district the legislator represents.) There is research that suggests that this
can influence the representative style of politicians (e.g. Ames 1995; Bowler
and Farrell 1993; Carey and Shugart 1995; Cox 1990; Farrell and Scully
2003; Katz 1980; Shugart 2001), calling for a greater focus than heretofore
on parliamentary behaviour. I n other words, rather than focusing on legislator
backgroundsand attitudes, our interest is in examining their activities.
1
There are good grounds for arguing that electoral systems can shape the
behaviour of politicians (and, as explored in the next chapter, voters also). For
instance, the vast literature on Duvergers Law is predicated on the assump-
tion that voters and politicians respond to electoral system-induced incen-
tives, such as the tendency for voters and politicians to emphasise larger
parties in non-proportional systems given that these are most likely to win
seats (see especially Cox 1997). And if electoral systems affect what happens
prior to elections, we must assume they have an analogous impact on what
happens subsequently. I n short, therefore, it is implausible to suggest that the
effects of electoral systems on representation stop with the conclusion of the
election itself, and important questions remain concerning how the represen-
tative process operates, and how it is perceived, by both elected representa-
tives and voters.
Considering only the elected representatives for the moment, whom do
they see themselves as representing, how do they (or should they) go about it,
and how are their attitudes and behaviour shaped by the institutional incen-
tives and opportunities that electoral systems generate? Our concerns here
thus encompass the much-debated mandate-independence controversy (cf.
Pitkin 1968) the extent to which representatives consider themselves bound
| 103 |
| T he voter versus the party |
to articulate and behave in accordance with the views of those they are elected
to represent. But our concerns also include the willingness of those who have
an electoral mandate to be responsive to public concerns and to pursue indi-
vidual or collective public grievances what in some contexts (notably the
UK and US) has come to be termed constituency service, a diverse range of
activities generally understood to include the maintenance of personal con-
tacts with, and advancing the concerns and interests of, voters within the geo-
graphical region from which the member is elected.
Scholarly knowledge of constituency representation comes mainly from
research conducted in the US and British contexts, most notably that derived
from the work of Richard Fenno (1978) on elected representatives home
styles. Constituency representation, for Fenno, was not only something that
elected representatives put a great deal of time and effort into; the manner
and forms of that representation could tell scholars much about how elected
representatives perceived their constituency and understood the representa-
tive relationship. Fenno placed great emphasis on the electoral implications
of constituency representation (see also Kuklinksi 1979). The electoral con-
nection dimension of constituency representation was further developed by
Cain et al. (1987). An important part of the stimulus for constituency service
behaviour, they suggested, was the electoral benefits accruing to elected rep-
resentatives in terms of a personal vote. British MPs and US Congressmen,
operating under a single member plurality (SMP) system, appeared to be
motivated to a high degree in their constituency service activities by strategic-
electoral considerations of vote maximisation (see also Norton and Wood
1993).
Other work conducted in Ireland, whose politicians are elected by STV,
has found high levels of constituency activity by elected representatives, and
a possible link with the candidate-centred electoral system operating there
(Farrell 1985; Wood and Young 1997; though see Gallagher 1987). This is
buttressed by cross-national research (e.g. Bowler and Farrell 1993; also
Farrell and Scully 2003) concerning members of the European Parliament,
where candidate-based ballot structures are seen to favour greater emphasis
on voter contact by politicians. This research shows a relationship between
electoral system design and parliamentary activity.
But some scholars caution that there is a need to take account of coun-
tervailing pressures. For instance, there are the wider cultural factors
principally shared conceptions of the representative role that can affect
how politicians operate (Norris 1997; Searing 1994). Evidence from the
| 104 |
1994 European Elections Study supports this view, leading Katz to con-
clude that the stronger constituency emphasis among British MEPs has
little to do with electoral systems effects, but rather is suggestive of a cul-
tural effect (1997a: 218). I n later work, he warns against predicting that
electoral reform will generate a mechanistic response from elected repre-
sentatives: political culture plays an important part in determining the
political consequences of electoral systems; the same institutions may be
associated with quite different outcomes if the actors pursue a different
mix of objectives (1999: 16). This chapter takes due account of Katzs
warning.
The overarching theme of this chapter, based on our argument in chapter
1, is that electoral politics in Australia faces two competing strains. On the
one hand, there are voter-oriented electoral systems AV and STV both of
which favour voter choice. On the other hand, there are party-controlled
electoral processes, such as how to vote cards, compulsory voting and pref-
erences, ticket voting and so on. The questions this chapter addresses are
twofold. First, how do these diverging pressures affect Australias elected rep-
resentatives? Do they respond more to public pressure than to party pressure,
and how do these pressures affect the way in which they approach their role?
I n theory, the voter orientation of the electoral system should emphasize such
factors as greater attention to constituency affairs and greater efforts to gen-
erate a personal vote during election campaigns. Against this, however, the
party control emphasis that predominates in Australia should manifest itself
in terms of such factors as a greater focus on party work and more emphasis
on party loyalty.
Our second question is derived from the analysis presented in chapter 3
regarding the main features of Australias electoral systems at national level.
We are interested in assessing whether the party control pressure should be
felt more keenly by Senators than by members of the House of
Representatives. Even though Senators are, prima facie, elected under a more
voter-oriented electoral system (in the sense that the voter is able to deter-
mine the fate of several candidates, as against a House of Representatives AV
election in which only one seat is being filled), in practice the party control
elements are far stronger in the Senates STV system, as shown for example
by the use of ticket voting, and by the role of the party selectorates in ranking
their candidate lists. Furthermore, the electoral constituency that senators
operate in an entire state or territory also favours a greater emphasis on
the party rather than on the individual voter.
| 105 |
This chapter examines these competing pressures on Australian elected
representatives, relying on the Australian Candidate Studies that have been
conducted at each federal election since 1987, with the single exception of
1998. The Australian Candidate Study is conducted in parallel with the
Australian Election Study and is designed to collect information about can-
didates activities during the course of the election campaign. We consider
three sets of evidence. We start, in the first section, with an assessment of the
link between parliamentary candidates (for both houses) and their parties,
examining the means by which Australian parties structure the principal
routes of entry into legislative politics. This is followed by an analysis of
survey data measuring the attitudes of Australian politicians about their rep-
resentative role; this analysis again seeks to identify the competing strains
between a voter and a party orientation, and how this plays out across the
memberships of the two houses. Finally, in the third section, we explore the
types of campaign activity engaged in by parliamentary candidates, how this
compares across both houses, and ultimately how it affects the election
prospects of Australian candidates.
Party service among election candidates
While membership of a major political party is an essential criterion for elec-
tion in almost all of the advanced democracies (Webb et al. 2003), in
Australia there are additional requirements: a high level of participation in
the party organisation and, preferably, some form of employment within the
party as well. The strongly disciplined Australian parties ensure that partici-
pation in centrally-organised party activities in the states or in national poli-
tics will bring the greatest rewards in terms of selection for a winnable seat.
Such party service can be defined in many ways, but there is a basic distinc-
tion between party service that is voluntary and unpaid and party service that
involves holding a regular paid job. While the former implies loyalty, it also
suggests a degree of amateurism; if individuals are recruited into politics
through this path, it may involve a conditional commitment, since most will
have had a regular job which could be taken up again if their political careers
prove unsuccessful. By contrast, holding a paid job implies professionalism
and a high level of commitment to making politics the individuals chosen
career. I f the political career fails, there may be no other skill to which the
party professional may turn (McAllister 1997).
Voluntary party service involves strictly party-related activities, such as
| 106 |
holding local or branch office. All of these branch-centred activities demon-
strate a level of party commitment, particularly at the local level. Professional
party service is concerned with employment by the party, either directly in a
national headquarters or in a regional office, or in a ministers or elected rep-
resentatives office. These types of party service are, of course, not mutually
exclusive and may overlap in significant ways; voluntary party service can
(and often does) provide a path to professional service, and vice versa. But we
would expect that, in the strongly disciplined party environment of Australian
politics, professional party service would dominate in legislative recruitment.
The Australian Candidate Studies, starting in 1987, have asked election
candidates a range of questions about their level of party political involve-
ment at different stages during their lives. Table 5.1 shows the proportions
who had held elected party positions and party employment, and their length
of party membership and level of party activity. Most of the candidates
reported that they had held some form of elected party office during their
political careers.
2
The most common type of party office-holding was being
a conference delegate or a local branch official; holding some federal
(national) party position is rarer, but still involves about one in seven candi-
dates. There have been relatively few changes in these patterns of office-
| 107 |
TALE 5.1 PARTY 3ERvlCE AV0N0 ELECTl0N CAN0l0ATE3, 198Z-2001
1987 1990 1993 199 2001 2004
le|d e|ec|ed pos|||or (percer|)
Corlererce de|ega|e ra 19 5 53 51 11
Loca| par|y ollc|a| ra Z2 Z 1 0
3|a|e par|y ollc|a| ra 3 31 28 31 30
Federa| par|y ollc|a| ra 15 13 11 11 15
Par|y ac||v||y
Lerg|| ol reroers||p (years) 11.2 10.Z 12.1 11.8 11.2 11.5
Par|y ac||v||y (|ours per ror||) 9.9 ra 10.1 1.8 1Z.9 22.1
Par|y erp|oyrer| (percer|)
Erp|oyed oy VP, r|r|s|er ra 12 11 1 1 20
Erp|oyed oy par|y ra 5 Z 1
(N) (11) (3Z5) (382) (313) (29Z) (20)
N0TE Par|y ac||v||y relers |o ac||v||y pr|or |o e|ec||or |l ar |rcuroer|. Es||ra|es are lor Laoor, L|oera|, Na||ora| ard Aus|ra||ar
0erocra| card|da|es or|y.
80UR6E 198Z-2001 Aus|ra||ar Card|da|e 3|ud|es.
holding over the period of the surveys. Placing the evidence of party service
found in Australia in a comparative context suggests that the trends are
similar to those found elsewhere. I nternational surveys indicate that candi-
dates and campaign activists have a high level of party involvement. Miller
and Jennings (1986: 77) show that about seven out of ten campaign activists
in the US were party office-holders during the 1980s and in Britain
Lovenduski and Norris (1995: 89) report even higher levels of party office-
holding among British major party candidates in the 1992 general election.
Australian candidates were asked how long they spent on party affairs; in
1987 this was just under 10 hours per month, climbing steadily to reach just
over 22 hours in 2004 reflecting a significant commitment to the party
organisation on the part of the candidate. I n contrast to voluntary party
service, professional party service is much rarer among election candidates; in
1990, 12 percent of candidates said that they had been employed in an MPs
or ministers office, a proportion that increased to one in five in 2004. Party
employment is, of course, an important career path, since it brings the indi-
vidual into contact with many of the major figures in the party leadership and
within the party organisation.
These figures suggest, then, that levels of voluntary party activism have
increased during the 1990s, measured by party activity, while office-holding
has remained constant. Since it is unlikely that party activism has become
more onerous to account for these major increases, it suggests that the
responsibility for running local parties is falling on a smaller number of
people, who must commit more time as party memberships decline.
Although consistent and reliable estimates of party memberships are difficult
to find, the available evidence indicates that the Liberal Party has registered
a consistent decline in membership, from 127 000 in 1967 to 64 000 in 1996
the most reliable estimate available (McAllister 2003: Figure 13.1; Ward
1997) while Labors membership, though higher, is also at an historic low.
3
The burden of maintaining the party organisation is therefore falling on a
smaller number of activists.
How does party service vary between candidates for the House of
Representatives and candidates for the Senate? Looking at the trends for
2004, Table 5.2 suggests that there are significant differences, notably in party
activity, but also in voluntary and professional party service. I n terms of
holding elected party positions, the differences are most marked at state and
federal levels, arguably the most visible types of voluntary party service. I n
this case, Senate candidates are more likely to have held such positions.
| 108 |
Employment by the party is also more likely to be reported by Senate candi-
dates. These results are consistent with other findings, which show that can-
didates with professional party backgrounds are more likely to be selected for
the Senate, where the central party organization, rather than the local
branches, has a much greater say in the selection process (McAllister 1997).
Nor are these patterns limited to the 2004 election. An examination of
earlier Australian Candidate Study surveys shows that Senate candidates
consistently have stronger party backgrounds than their House of
Representatives counterparts. The differences are most consistent, and most
pronounced, in the elected party positions that Senate candidates report
holding at the state and federal level, and also in their employment by the
party, by an MP or by a minister. For example, the earlier results found that
in 1993, 13 percent of all candidates had held a federal party position; disag-
gregated by house, no less than 32 percent of Senate candidates reported this
experience, compared to just ten percent of lower house candidates. I n 1990
the differences were even larger 10 percent in the House compared to 41
percent in the Senate. There is, then, a consistent pattern over an extended
period of time of those with stronger records of party activism and employ-
ment gaining nomination for the Senate.
| 109 |
TALE 5.2 PARTY 3ERvlCE 8Y l0u3E, 2001
house of Reps 8enate 0|fferenoe
le|d e|ec|ed pos|||or (percer|)
Corlererce de|ega|e 11 2 |=2.39, p=.02
Loca| par|y ollc|a| 59 5 |=0.5, p=.52
3|a|e par|y ollc|a| 21 5 |=5.21, p=.00
Federa| par|y ollc|a| 12 35 |=3.Z9 p=.00
Par|y ac||v||y
Lerg|| ol reroers||p (years) 11.Z 1Z.8 |=3.00, p=.00
Par|y ac||v||y (|ours per ror||) 12.1 19.Z |=1.80, p=.0Z
Par|y erp|oyrer| (percer|)
Erp|oyed oy VP, r|r|s|er 18 30 |=1.59, p=.11
Erp|oyed oy par|y 1 8 |=1.12, p=.2
(N) (223) (3Z)
N0TE Par|y ac||v||y relers |o ac||v||y pr|or |o e|ec||or |l ar |rcuroer|. Es||ra|es are lor rajor par|y card|da|es or|y.
80UR6E 2001 Aus|ra||ar Card|da|e 3|udy.
The trends represent clear evidence of the extent to which party service,
through knowledge and familiarity with the party organisation (both local
and central), serves to enhance a candidates election prospects. Service to the
local party will bring the aspiring candidate into contact with the local
power-brokers and party officials, provide an entre to the formal committees
and informal groups that take important local decisions, and convey essential
political intelligence about the constituency as a whole. Service in a profes-
sional capacity will also provide knowledge about formal party decision-
making, but in addition it will enable access to a range of informal networks
and contacts, which can be invaluable when seeking a nomination. Moreover,
from the perspective of this book, the more interesting finding is that the
party-oriented pressures that predominate in the Senate have a significant
impact on the types of individuals who stand for election to that house: those
individuals who have a strong party background are especially favoured.
Candidates views of their representative roles
Another piece of evidence that helps to inform the tension between party and
voter among election candidates is the candidates views of their role as
elected representatives. Notions of political representation in Australia have
followed Westminster conventions, with the British model forming the basis
for the 1901 constitution that established the Federation (Butler 1973; Crisp
1983). However, a distinctive Australian approach to representation has been
moulded by two factors. First, although the constitution adhered to the
Westminster model, incorporating parliamentary government and adversary
politics, a federal arrangement was introduced, based on the American and
Canadian experiences, to provide a mechanism to permit the pre-1901
colonies to retain some degree of political independence. Consequently, the
major political parties have had to operate at two separate levels of govern-
ment, state and federal, and these levels have different electoral systems and
parliamentary procedures.
4
Second, although responsible party government is
at the heart of the Westminster system, the level of discipline that the
Australian parties enforce on their members is among the strictest found in
the Anglo-American democracies. As Jaensch (1983: 213) concludes, the
overriding theme in Australian conceptions of political representation is
government of the people, by the party, for the party.
Prior to 1987, there were few studies of Australian elected representatives.
The only substantial study was Emys (1974: 45599) survey of 75 national
| 110 |
elected representatives in the early 1970s. Emy identified three types of rep-
resentatives: trustees, delegates and what he termed politicos, who sought to
reconcile national and local interest (p. 481). The survey concluded that
politicos were the largest group within the Labor Party, while trustees formed
the majority among Liberals. There was little, however, to suggest that party
political considerations dominated political representation. More recent
studies of national election candidates have found that traditional views of
political representation exist, based on the trustee, delegate and party distinc-
tions, although these views have had little impact on the outlooks or the votes
that candidates attract once party political factors are taken into account
(Studlar and McAllister 1994).
The 1993 Australian Candidate Study included a question asking parlia-
mentary candidates what they considered as important influences on the leg-
islative voting of parliamentarians, and the responses have been factor
analysed (Table 5.3).
5
The first factor is concerned with the influence of the
mass media, parliamentary debate and interest groups on MPs legislative
voting, although few consider these influences to be very important. I n effect,
these items measure national elite opinion, and while not exactly comparable
| 111 |
TALE 5.3 vlEw3 0F P0LlTlCAL REPRE3ENTATl0N AV0N0 ELECTE0 REPRE3ENTATlvE3
HPs shou|d ||sten to...
Peroent say
'very |mportant'
Faotor |oad|ngs
| || |||
Free manoare
... ||e rass red|a 1 .Z .31 -.08
... par||arer|ary deoa|e .Z2 -.33 .10
... |r|eres| groups Z .Z0 .05 .18
Respons|o|e parry
... ||e|r po||||ca| par|y 51 -.09 .81 -.09
... ||e|r par|y |eader 20 .20 Z.2 .25
lmperar|ve manoare
... ||e|r e|ec|ora|e 52 .01 .02 .88
... puo||c op|r|or 18 .3Z .01 .2
... ||e|r core vo|ers 2Z -.0Z .15 .5Z
E|gerva|ue 2.2 1. 1.2
Percer| var|arce exp|a|red 28 20 15
N0TE Fac|or |oad|rgs are var|rax ro|a|ed lac|or |oad|rgs lror a pr|rc|pa| corporer|s lac|or ara|ys|s W||| ur|||es |r ||e
ra|r d|agora|. N=112 pas| ard currer| reroers ol ||e Aus|ra||ar par||arer|.
with the traditional Burkean notion of the free mandate or trusteeship model
of representation, they do reflect a modernised notion of Burkes original con-
ception. The second factor is the familiar party model, reflected here in the
representatives political party (mentioned by 51 percent as very important)
and party leader (mentioned by 20 percent). The third factor is the impera-
tive mandate or delegate model; here, elected representatives are particularly
concerned with the influence of their electorate (mentioned as very important
by 52 percent), and with public opinion and core voters. Again, while not pre-
cisely paralleling the imperative mandate model in the strict sense, the elected
representatives are responding to the dictates of their voters.
The views about political representation of Australian parliamentary
candidates therefore conform closely, but not perfectly, to the three tradi-
tions that dominate the literature. Apart from party influences, legislators
distinguish between national elite opinion influences and influences that
emanate from the voting public, both in a generalised sense (public
opinion) and in particularised sense (their own electorate). This suggests
that Australia does not follow Burkes eighteenth-century British distinc-
tion between delegate and trustee exactly, but the resemblance is sufficiently
close for us to continue to use the terms free mandate and imperative
mandate to describe these two orientations.
| 112 |
TALE 5.4 PARTY, lNCuV8ENCY AN0 l0u3E vARlATl0N3 lN vlEw3 0F P0LlTlCAL REPRE3ENTATl0N
[Heans, zero to 10 soa|es}
Free mandate |mperat|ve
mandate
Respons|b|e
party
[N}
All 1.5 .9 .9 (115)
Parry
Laoor 1.0 .Z Z. (119)
L|oera|-Na||ora| 1. Z.0 .9 (118)
0erocra| 1.9 Z.0 .3 (121)
0reer 1.5 Z.1 Z.1 (2Z)
lnoumoenoy
lrcuroer| 1.1 .3 Z.5 (330)
Nor-|rcuroer| 1. Z.1 .Z (85)
louse
louse ol Represer|a||ves 1.5 Z.0 .8 (312)
3era|e 1.5 .Z Z.1 (Z3)
As we would expect, the most important focus is party-oriented, resulting
in a mean of 6.9 on a zero to ten scale (first line, Table 5.4), followed by the
imperative mandate (6.9) and the free mandate (4.5). I ndeed, given the extent
to which legislative behaviour is party-determined within the Australian par-
liament, it is perhaps surprising that the imperative mandate and free
mandate views of representation attract such a respectable level of support
from legislators.
The remainder of Table 5.4, using a simple comparison of means, shows the
extent to which there are variations in these approaches to political representa-
tion, based on party membership, incumbency and house. The results show that
differences in representation between the major parties are important (cf.
Jaensch 1983: 21315). In line with their individualist, conservative ideology,
Liberal-National legislators are less likely than their Labor counterparts to
emphasise party factors, and are more likely to emphasise the free mandate. The
least party-oriented are, of course, the Greens, in line with their anti-centrist,
populist philosophy. Incumbents and non-incumbents differ significantly on
the imperative mandate and the responsible party models, with incumbents
favouring the party view, challengers the free conscience. Finally, Senators are
more likely to endorse the responsible party model of representation than are
members of the House of Representatives, contrary to the claim that the Senate
is a house of review, above the partisan conflict of the lower house.
Campaign activity by
Australian parliamentary candidates
According to the electoral systems literature, as we reviewed above, preferen-
tial electoral systems provide voters with greater choice between the com-
peting parliamentary candidates, and we should expect this to be reflected in
the types of activities that candidates engage in during election campaigns in
order to increase their personal vote. Studies in recent years, across several
countries, have indicated that under such electoral systems conscientious con-
stituency work can significantly improve the re-election prospects of an
incumbent member (Cain et al. 1987; Norton and Wood 1990; Lancaster and
Patterson 1990; Ferejohn and Gaines 1991). I n Australia, the size of the per-
sonal vote in House of Representatives elections has been estimated to be in
the order of three percent of the first preference vote, although it reaches this
level only in exceptional cases (Bean 1990; also Jackman 2004).
| 113 |
One factor mitigating the potential for a personal vote is, of course, the
size of the country. With just 20 million people in an island continent that
exceeds the size of Europe, elected representatives often have to travel long
distances between their constituencies and the federal capital in Canberra,
thereby reducing the time available for constituent-focused activities. Even
within their constituencies, the distances are often considerable.
6
Within
these vast constituencies voters are often very remote, physically and politi-
cally, from their chosen representatives, just as their representatives are often
very remote from their party colleagues in Canberra. This inevitably has had
a major impact on the relationship between voters and their elected represen-
tatives, and on the ways in which elected representatives can bring themselves
to the attention of their voters.
By any standards, the proportions of House of Representatives candidates
residing in their constituencies is high, and has remained consistent at about
three quarters of all lower house candidates (Table 5.5). I ndeed, the trend
shows a modest increase, if anything: 88 percent said that they lived in their
constituency in 2004, compared to 76 percent in 1987. The period of resi-
dence in the constituency also shows a modest increase over the period of the
surveys, averaging about 16 years in the late 1990s and increasing to 21 years
in 2004. To the extent that there has been any change in community involve-
ments, it has been in the number of community organisation memberships
that candidates report in the surveys. I n 1987, candidates averaged 2.5 mem-
berships out of a total of eight potential organisations; this has declined con-
sistently since then to 1.5 in the most recent survey in 2004. This decline in
social capital is in line with the general decline in social group involvement,
| 114 |
TALE 5.5 C0VVuNlTY lNv0LvEVENT, 198Z-2001
1987 1990 1993 199 2001 2004
L|ve |r cors|||uercy (percer|) Z ZZ 81 Z8 8 88
Lerg|| ol res|derce |r cors|||uercy (rear years) 1Z 1Z 11 1 1 21
Corrur||y orgar|sa||or
reroers||ps (rear ruroer)
2.5 2.3 1.5 1.1 1.1 1.5
(N) (128) (322) (331) (39) (315) (20)
N0TE Es||ra|ed lor louse ol Represer|a||ves card|da|es or|y. Corrur||y |rvo|verer| |s corrur||y serv|ce orgar|sa||or,
re||g|ous orgar|sa||or, Worer's orgar|sa||or, erv|rorrer|a| group, spor|s c|uo, |osp||a| ooard, sc|oo| ooard/PTA, e||r|c
group orgar|sa||or.
80UR6E 198Z-2001 Aus|ra||ar Card|da|e 3|ud|es.
which has been observed by Putnam and others across a range of advanced
democracies (Putnam 2000).
To what extent are election candidates orientations towards their party
and to their voters reflected in how they conduct their election campaign?
Studies of incumbent legislators as campaigners abound (Fenno 1978, 1989;
Parker 1986), but they often suffer from a lack of specification of the com-
petitive conditions facing incumbents in constituency campaigns. One excep-
tion is Fowler and McClures case study, which focuses on an open seat
contest for the USA House of Representatives rather than on the succeeding
contest between an incumbent and a challenger, which the incumbent lost
(Fowler and McClure 1989). The overall conclusion from this work is that
the more individually-oriented campaigns in the USA involve self-motivated
candidates who harbour considerable personal and political ambition, and
voters considerations weigh more heavily than those relating to candidates.
By contrast, in the UK, a study of the 1966 general election (Kavanagh 1970)
TALE 5. TYPE3 0F CAVPAl0N ACTlvlTY, 2001
6ampa|gn aot|v|t|es hours per week Faotor |oad|ngs
| || |||
Par|y
0|scuss|ors W||| |oca| par|y 1.9 .81 -.01 .0Z
0|scuss|ors W||| ra||ora| par|y 2.2 .Z .21 -.11
3pea||rg or |e|ep|ore 8.2 .1 .01 .11
Furd ra|s|rg 2.2 .5Z .3Z .10
(To|a| par|y) (1Z.1)
Ved|a
Rad|o, Tv |r|erv|eWs 2.1 -.09 .91 -.0
Na||ora| reWspaper |r|erv|eWs 0.5 .19 .ZZ .03
3uouroar, |oca| reWspaper |r|erv|eWs 1.Z .03 .2 .15
(To|a| red|a) (1.3)
vo|ers
0oor|roc||rg, carvass|rg 9.5 -.0 .03 .Z2
0rgar|s|rg d|rec| ra|||rg 2.9 .23 .01 .
(To|a| vo|ers) (12.1)
E|gerva|ue 3.0 1.1 1.2
Per cer| var|arce exp|a|red 33 1 13
N0TE Fac|or |oad|rgs are var|rax ro|a|ed lac|or |oad|rgs lror a pr|rc|pa| corporer|s lac|or ara|ys|s W||| ur|||es |r ||e ra|r
d|agora|. N=1ZZ card|da|es.
| 115 |
found that candidates were much more inclined to see their fortunes resting
in national partisan and economic forces rather than in what they could do
individually in their campaigns.
The 2001 Australian Candidate Study asked the candidates how much time
they spent during the election campaign on nine activities, ranging from dis-
cussions with their local party officials to door knocking and canvassing. Based
on a factor analysis of the responses, Table 5.6 shows that these activities can be
categorised into three types of activity: those related to the party, those focusing
on media interviews and those concerned with contact with voters. The most
time-consuming of the three activities was reported to be party-related, taking
up, on average, 17 hours per week for each candidate. This was followed by tra-
ditional campaign activities focused on voters, which accounted for 12 hours
per week, and media interviews, which accounted for 4 hours. In total, the
average candidate spent just over 34 hours per week engaged in election cam-
paign activities of one sort or another. However, the single most important
activity was clearly door knocking and canvassing, which accounted for 9.5
hours, followed by speaking on the telephone, at 8.2 hours.
To what extent do these patterns of activity vary by house, and also by
incumbency and party affiliation? Table 5.7 suggests that there are substan-
tial differences between the groups standing for election in 2001. Party-
related activities were most important for South Australian candidates, and
least for those standing in the ACT, while media activities did not vary sig-
nificantly by state. However, there are very substantial variations based on
voter-related activities. These were highest among candidates in Queensland
and Western Australia, and lowest in the ACT; this probably reflects the dif-
fering sizes of the areas involved and the resulting time required to canvass
voters, at least personally. As we would expect, Senate candidates reported
allocating significantly more time to party than to voter-related activities
than House of Representatives candidates. Again as we would expect, incum-
bents were more active in all of the areas of campaign activity. I n general,
Liberal-National candidates were more active on all forms of campaign
activity when compared to the other party candidates. Among the minor
party candidates, those standing for One Nation were notably energetic on
media activities, and on activities involving voters.
To what extent do the party and voter foci of the election candidates affect
the votes that they receive? We might expect that a party orientation would
affect the vote by ensuring that candidates who spent more time on their
party-related activities would be more likely to gain selection for a safe seat,
| 116 |
and to have more ready access to party resources during the election cam-
paign. By virtue of their party contacts, they might be able to attract the party
leader to their constituency, as well as prominent cabinet and shadow cabinet
ministers. Equally, however, a voter orientation could lead to a personal fol-
lowing, and to a better understanding of the needs of the electorate and to the
groups within it who would be most important to convert. I n other words,
both strains towards the party, or towards the voters may have positive
effects in attracting the vote.
To test these two hypotheses, we estimate an ordinary least squares regres-
sion equation, predicting the first preference vote in the House of
Representatives; because of the different electoral system (and also because of
| 117 |
TALE 5.7 vARlATl0N3 lN LEvEL3 0F CAVPAl0N ACTlvlTY, 2001
Hean hours per week
[N}
Party Hed|a Voters
All 1Z.1 1.3 12.1 (19)
$rare
NeW 3ou|| wa|es 11.8 3.Z 10.Z (111)
v|c|or|a 18.3 1.2 12.3 (101)
0ueers|ard 18. 5.0 15.3 (89)
3ou|| Aus|ra||a 23. 5.2 11.Z (1Z)
wes|err Aus|ra||a 1.1 5.0 11.1 (19)
Tasrar|a 1.Z 3.5 11.0 (20)
ACT 13.Z 3.Z 5. (9)
Nor||err Terr||ory 1Z.1 1.5 13.0 (10)
louse
louse ol Represer|a||ves 1.5 1.1 12.9 (10)
3era|e 23. 1.1 8.Z (3)
lnoumoenr
lrcuroer| 22.5 5.8 15. (102)
C|a||erger 1.1 3.9 11.1 (3Z5)
Parry
L|oera|-Na||ora| 20.3 5.2 20.1 (85)
Laoor 19.5 3.9 18.3 (88)
0erocra| 13.5 1.5 .2 (121)
0reer 1Z.9 3.3 .9 (98)
0re Na||or 1Z.5 5.1 13.Z (Z1)
the small number of Senate respondents, a function of the half-Senate elec-
tion in 2001) we have to exclude Senate candidates. The independent vari-
ables represent party service and the voter orientation, using the measures
discussed in the previous sections. The campaign activities engaged in by the
candidates are also included. Obviously controls are important, and we take
into account incumbency, which will have a significant effect on the vote, and
also party, divided between candidates standing for Labor or a minor party.
7
Liberal-National candidates represent the excluded category.
The results in Table 5.8 suggest that neither a voter orientation nor cam-
paign activities have any statistically significant effect on the vote, once
incumbency and party are taken into account. By contrast, there is a signifi-
cant effect for having been a party employee, which increases the proportion
of first preference vote by just over four percent, net of a wide variety of other
things. There is no comparable effect for holding elected party positions. As
| 118 |
TALE 5.8 PARTY AN0 v0TER 0RlENTATl0N3, AN0 TlE v0TE (l0u3E 0F REPRE3ENTATlvE3
CAN0l0ATE3)
Heans 0L8 est|mates
b beta
Par|y serv|ce
E|ec|ed par|y pos|||ors (r, 0-5) 1.82 -.1 -.02
Par|y erp|oyee (0,1) .13 1.28 .08
vo|er or|er|a||or
Corrur||y reroers||ps (r, 0-Z) 1.13 -.38 -.03
L|ve |r cors|||uercy 0.Z3 -.Z1 -.02
Carpa|gr ac||v|||es (|ours per Wee|)
Par|y 1.18 .00 .00
Ved|a 1.38 .05 .03
vo|er 12.9 -.03 -.02
lrcuroer| 8.1 .19
Par|y (Coa||||or)
Laoor .18 -.88 -.02
V|ror par|y .5 -30.1 -.83
Cors|ar| 35.55
Adj R-squared .81
p<.01, p<.05.
N0TE Par||a| ard s|ardard|sed 0L3 regress|or coellc|er|s pred|c||rg percer| lrs| prelererce vo|e. N=10 louse ol
Represer|a||ves card|da|es.
we would expect, the main effects are party-related, with minor parties
getting around 30 percent fewer votes than their Coalition counterparts, and
incumbents getting just over eight percent more votes than challengers, net
of other things. Clearly, then, there is an important party-related effect on the
vote, which stems from party employment. Since we are taking into account
party-related campaign activities, it would appear that this effect comes about
less because such candidates can access more resources than those who have
not been party employees, but rather through the ability to gain selection for
a winnable seat. I n other words, party employees have more choice in the
types of seats that they gain selection for.
Conclusion
We started this chapter by setting out the competing pressures on Australian
elected representatives. On the one hand, the voter-oriented pressures of the
preferential electoral system should promote a greater emphasis on voters; on
the other hand, the party-dominated features of the electoral process (for
example regarding candidate selection, the distribution of how to vote cards
in AV elections, ticket voting in STV elections) should promote a greater
emphasis on party-related activities. The evidence clearly leads to the conclu-
sion that among Australias elected representatives (or, at least, the parliamen-
tary candidates), the party emphasis predominates. This was shown in three
respects:
the routes of entry to a parliamentary career in Australia emphasise party
activity and/or party employment
survey evidence of how Australian parliamentary candidates envisage
their representative role privileges a party orientation
in their campaign activities, Australian parliamentary candidates spend
more time on party-related activities than on actually chasing votes and
with good reason, as this is shown to correlate closely with electoral
success.
As we might expect, the dominance of party over voter is greater among
Senators than among members of the House of Representatives. This is con-
sistent with the findings in chapter 3 that the Senates system of election has
evolved (particularly since 1983) away from its original voter-oriented roots.
The electoral system has therefore had a distinct and marked effect on the
| 119 |
way in which elected representatives see their role, and in how they allocate
their time to various activities.
I n chapter 1 we referred to the competing pressures on Australian
politicians and their voters in two respects: first, an electoral system that pro-
motes candidate orientation tempered by a strong party orientation; second,
a voter choice emphasis, again promoted by the electoral system, versus wider
elements of voter compulsion. This chapter has dealt with the first of these
pressures, showing the party orientation to be the more dominant. The next
chapter deals with the second pressure.
| 120 |
s i x
Compulsory voting, party
control and the voter
How do Australian voters engage with the preferential electoral systems?
According to their proponents, these systems are, first and foremost, voter-
oriented. I f these systems have a place, it is in their embracing of the notion
of voter choice, in the sense that they provide voters with the maximum pos-
sible opportunity to decide between all the candidates on the ballot paper
(Lakeman 1974; Wright 1980). I t is at this point that ballot structure comes
to the fore: this is where, supposedly, we see the full merit of preferential
voting as an ordinal vote process, as opposed to the either/or categorical
choice offered by SMP or the (closed) list systems (Bowler and Grofman
2000b; Farrell and Gallagher 1998). Critics of preferential electoral systems
have a different perspective, suggesting that the complexity of the ballot (and
counting) process runs the risk of confusing voters (e.g. Jenkins 1998). This
question of voter empowerment versus confusion is also at the heart of the
debate over strategic voting, with some suggesting that the computational
demands made by preferential systems on voters must affect the voters ability
to make strategic use of the ballot (Bowler 1996; Cox 1997).
An added ingredient that is prominent in the Australian context is the
way in which the voter choice element of preferential systems is strongly con-
ditioned by two related factors: first, a compulsory element that runs the
whole gamut of the electoral process; second, legal and strategic manoeuvres
by the political parties to direct how the voters actually use the electoral
system. I f the voter choice dimension of preferential systems invites the risk
| 121 |
of voter confusion, these Australian elements of compulsory voting and party
controls raise the potential of voter alienation.
This chapter focuses on these two features of potential voter reaction,
seeking answers to two questions. First, are Australian voters confused by the
complexities of these electoral systems? Second, are they alienated by their
strictures? We start with some contextual discussion. The details of how the
preferential systems operate have already been dealt with (especially in
chapter 3), and this has included coverage of the party control elements
(notably how to vote cards in AV elections and ticket voting in STV elec-
tions) and the issue of the compulsory expression of preferences. I n the first
section of this chapter we provide additional background on the remaining
elements of compulsion in the Australian electoral process, tracing the origins
and evolution of Australias rules on compulsory registration and voting. The
second section reviews the debate over strategic voting in preferential systems
and its applicability in the Australian context of party control. The picture of
Australian electoral practice that is drawn in these two sections is one in
which many of the elements of voter choice in preferential voting have been
removed, producing the potential double disadvantage of an electoral process
that is at the same time both confusing and alienating.
The remaining two sections examine how Australian voters engage with
the electoral process under these conditions of complexity, compulsion and
party control. The third section assesses the issue of voter confusion through
an examination of voter use of the system, both at national and state level,
revealing significant pockets of voters who are willing and able to vote in a
manner independent of party influence. The chapter concludes with an
examination of survey evidence on voter attitudes to the electoral process.
Compulsory voting in Australia
Chapter 3 provided evidence of the Australian proclivity to require voters to
express a minimum number of preferences (for the most part making them
complete all, or at least most, preferences). This feature alone suggests a
tension in the electoral system between an apparent wish to maximise voter
freedom to choose and a wish to maintain party control over preferences. To
complete this picture, we also need to account for other elements of voter
compulsion that govern Australias electoral institutions: these are compul-
sory registration and compulsory voting. The implications of these will be
assessed in the final section of this chapter.
| 122 |
The decision to make registration (referred to in Australia as enrolment )
compulsory was originally taken on the recommendation of the then Chief
Electoral Officer, R.C. Oldham, who wanted to rationalise the electoral roll.
1
When it was introduced in 1911, the system of compulsory registration had
the intended effect of greatly increasing the size of the electorate: in April
1910 registration was 2 258 482; by the time of the May 1913 election, when
it was compulsory, registration had risen to 2 760 216. Prior to that change,
the compiling of an up-to-date electoral roll was thought by Oldham to be
too cumbersome and expensive as it occurred infrequently and did not
operate effectively. Moreover, there were criticisms of the system of postal
voting, and it was believed that a more secure system of absentee voting
would require complete enumeration of the eligible electorate. This change to
compulsory registration, which established the principle of compulsion and
opened the door to compulsory voting, was thus driven by the bureaucrats,
not by politicians, although the latter were happy to defend it as a necessary
reform.
2
I t was a means of improving the accuracy of the electoral rolls and
was proposed as a simple device for the benefit of electoral officials. The
states quickly followed suit, beginning with Queensland in 1914.
Registered voters as a proportion of the voting age population have varied
in the post war years (Figure 6.1).
3
I n the first two post war elections, in 1946
and 1949, it stood at just over 90 percent, but dropped consistently in each of
the next three elections, reaching a post war low of 78.1 percent in 1955.
Thereafter, the figure has varied from 81 to 85 percent, though there is a clear
downward pattern; over the 1958 to 2004 period, the enrolled proportion of
the age-eligible electorate has been declining by just under 0.1 percent per
year, reaching 81.0 percent in 2004.
4
The compulsory registration provisions are at least as well enforced as in
other English speaking countries, although there appears to be less compli-
ance in Australia. This is due to two factors. First, there are large numbers of
immigrant non-citizens in Australia, particularly from non-English speaking
countries, who cannot enrol to vote until they become citizens. This was
compounded by the change in 1984, whereby United Kingdom and New
Zealand permanent residents were no longer eligible to vote. Second, there is
a relatively high level of non-compliance among the young. Research con-
ducted for the Australian Electoral Commission has consistently found that
younger people are less likely to register (McAllister 1985; Elliott and
Shanahan 1989). Recent surveys conducted for the AEC
5
have found that
just 78 percent of the eligible population aged 18 to 24 are enrolled, compared
| 123 |
| Compul sory voti ng, party control and the voter |
to 93 percent for the eligible population as a whole.
6
The trends suggest that
non-compliance among the young has been increasing in recent years.
When it was first introduced, compulsory voting was argued to be the
natural corollary of compulsory registration. Compulsory voting (or, more
accurately, turnout) is, however, a different matter: several English speaking
countries have compulsory registration, but have stopped short of introducing
compulsory voting. Currently, 43 countries operate some form of compulsory
voting,
7
although only eight Australia, Belgium, Cyprus, Fiji, Luxembourg,
Nauru, Singapore and Uruguay operate strict enforcement of voting,
involving a fine, imprisonment, or both.
8
I n Australia, compulsory voting was first introduced in Queensland by the
Denham Liberal government in time for the 1915 general election. The
measure appears to have been the last desperate attempt of an unpopular gov-
ernment to save itself from defeat. I n the event, the attempt was unsuccessful,
| 124 |
30
1
9
4
6
1
9
4
9
1
9
5
1
1
9
5
4
1
9
5
5
1
9
5
8
1
9
6
1
1
9
6
3
1
9
6
6
1
9
6
9
1
9
7
2
1
9
7
4
1
9
7
5
1
9
7
7
1
9
8
0
1
9
8
3
1
9
8
4
1
9
8
7
1
9
9
0
1
9
9
3
1
9
9
6
1
9
9
8
2
0
0
1
2
0
0
4
40
50
60
70
80
90
100
Vote/registration
Vote/VAP
FIGURE 6.1 PATTERNS OF REGISTRATION, 19462004
SOURCE International IDEA; AEC.
as Labor won a sweeping victory at the election. Over the next three decades
the other states and the Commonwealth followed suit in introducing com-
pulsory voting. The system was introduced for national elections in 1924, fol-
lowing the report of the 1915 Royal Commission, which viewed compulsory
voting as a logical extension of compulsory registration. The delay in its intro-
duction was caused by divisions within the Labor government, which could
agree only on its application to referendums (Goot 1985: 192). I n the states,
Victoria introduced compulsory voting in 1926, New South Wales and
Tasmania in 1928, Western Australia in 1936 and South Australia in 1941.
9
I n all cases, the move to introduce compulsory voting was bipartisan, and
was seen as a logical extension to compulsory training, compulsory educa-
tion, compulsory early closing, and compulsory enrolment (Commonwealth
Parliamentary Debates1915, 6056) in other words, to an underlying culture
that values regulation and compliance. The new system suited the conven-
ience of all parties and was opposed by none, since it would remove the need
for parties to mobilise voters to attend the polling places. I n the case of the
Commonwealth, the entire 1924 debate in both houses took less than three
hours one hour and 26 minutes in the Senate, and 52 minutes in the House
of Representatives and no government minister spoke on the matter. This
is a remarkably short period of discussion for such a major electoral change,
and illustrates the bipartisan support that it attracted.
The effect of the introduction of compulsion on turnout was immediate
and dramatic (Table 6.1). I n the six states and nationally, turnout increased
by an average of 23.2 percent, ranging from 12.6 percent in Queensland
| 125 |
TALE .1 NATl0NAL AN0 3TATE TuRN0uT, 8EF0RE AN0 AFTER TlE lNTR00uCTl0N 0F C0VPuL30RY
v0TlN0
Year |ntroduoed Jur|sd|ot|on E|eot|on pr|or to
|ntroduot|on
E|eot|on after
|ntroduot|on
|norease
[peroent}
1911 0ueers|ard Z5.5 88.1 12.
1921 CorrorWea||| 58.0 91.3 33.1
192 v|c|or|a 59.2 91.8 32.
1928 NeW 3ou|| wa|es 82.5 91.9 12.1
1928 Tasrar|a 81.9 95.0 13.1
193 wes|err Aus|ra||a Z0.1 91. 20.5
1911 3ou|| Aus|ra||a 50.Z 88.5 3Z.8
80UR6E VcA|||s|er, Vac|erras ard 8o|d|s|or (199Z).
(where turnout was relatively high anyway under voluntary voting) to 37.8
percent in South Australia. The average level of turnout under the system of
compulsory voting was 91.6 percent. I n effect, the increase in turnout was
significant, which was the goal of the politicians who introduced it. There was
also a flow-on effect to turnout in state elections, even where voting remained
voluntary: in New South Wales, for example, state turnout rose from 69.1
percent in May 1925 to 82.5 percent in October 1927.
10
I n general, ordinary
citizens accepted the new system with little complaint and there were com-
paratively few instances where voters were fined for not attending the polls
without sufficient reason; as a consequence, the system has been easy to
administer ever since it was introduced.
Clearly, compulsory voting has had the desired effect of keeping levels of
electoral turnout high (even if this is somewhat qualified by the relatively high
levels of non-compliance with the rules on compulsory registration). But,
coupled with the demands put on voters once they have turned out to vote
the rules regarding compulsory preferential voting it brings with it the risk of
voter alienation. We examine this issue in the last section of this chapter (pp.
138145). Before turning to this discussion, we need to explore one more
dimension of Australian electoral practice that may contribute to possible voter
alienation, and this is the role of the parties in seeking to tightly control the
preference-voting proclivities of the voters. We start, in the next section, with a
review of the potential for strategic voting in preferential systems.
Preferential voting and the strategic voter
Chapter 1 reviewed the social choice debates over the properties of electoral
systems and strategic voting. I n chapter 4, we assessed the properties of the
preferential systems, examining how particular features of these systems could
influence the fate of candidates. I n this section we address the theme of
strategic, or insincere, voting, in which voters are said to attach more impor-
tance to preferred outcomes rather than preferred candidates (or parties).
According to the Gibbard-Satterthwaite theorem, such strategic voting is
said to occur under all non-dictatorial electoral systems (Cox 1997;
Ordeshook 1986).
I n the academic literature, strategic voting has been explored in some
detail in the case of such electoral systems as single member plurality (SMP)
and the single non-transferable vote (SNTV) (for a recent overview, see Cox
1997). The general consensus is that the scope for strategic voting declines as
| 126 |
an electoral system becomes more complex (Bartholdi and Orlin 1991; Cox
1997; Saari 2001), the reason being that the computational and information
requirements in more complex systems make it all but impossible for voters
to act in a strategic manner. But, as we argued in chapter 1, a distinction
should be drawn between outcome and intent, for even if the permutations in
complex electoral systems rule out any realistic prediction of the outcome of
a strategic vote, this does not prevent voters from at least trying. For that
matter, there is scope for the parties to attempt to coordinate strategic voting
in their campaigns.
Preferential voting systems and their operation in Australia provide some
correctives to the assumptions underlying the strategic voter debates, first,
with regard to the notion of strategic voting as necessarily insincere voting,
and second, relating to the role of party elites in steering the vote process. We
deal with each of these in turn.
According to the academic debates, strategic voters are seen as insincere
in their voting practice, giving greater priority to outcomes than to candi-
dates, and therefore not voting for their most-preferred candidates. I t is easy
enough to see how this operates under categorical electoral systems where the
voters have simply one choice available to them. But under ordinal systems
such as AV and STV the range of choices available to voters is greater
(Bowler and Grofman 2000; Farrell and Gallagher 1998). I t may still be the
case that a voter places greater priority on outcome than on candidates, in the
sense of trying to prevent the election of her least preferred candidate (be it a
Pauline Hanson or whomever), but this does not necessarily mean that the
voter must vote non-sincerely, in the sense of not giving a preference to her
most preferred candidate. The point is that, if the voter wants to prevent the
election of candidate A, she must firstly not give A any preferences, and sec-
ondly deploy her preference profile in the most effective way to ensure
maximum damage to A. But this does not necessarily preclude her from
expressing a high preference for her most preferred candidate, B.
11
When assessing strategic voting under preferential systems, therefore, the
issue is less one of a voter trying to influence the election outcome by voting
non-sincerely, and more one of trying to do so by deploying her vote in as
efficient a manner as possible. We have already seen how this might occur
under AV, and Hansons fate in 1998 provided a perfect illustration of this (as
discussed in chapter 4). Under STV, inevitably, the range of options available
to the voter becomes even greater. Voters can deploy their preferences effi-
ciently so as to maximise the number of seats won by their preferred party, to
| 127 |
minimise the number of seats won by their least preferred party and to influ-
ence the post-electoral government process.
This can involve four main tactics. First, voters can express top prefer-
ences for all of their preferred partys candidates that is, they can vote loyally
or solidly for all the candidates before expressing preferences for candidates
of any other parties (Gallagher 1978; Sinnott 1995). The object here is to
minimise the leakage of preferences to candidates from other parties: if a
party is fielding, say, three candidates in the constituency, its supporters
should express their first three preferences only for those candidates. Second,
the corollary of this is that the partys supporters should give no (or last) pref-
erences to the candidates of whichever party is seen as the greatest threat to
the electoral success of their preferred party. Third, voters can seek to influ-
ence the government formation process by the direction of their inter-party
preferences, expressing support for candidates of the party most likely to form
a coalition with their preferred party (Laver 2000). Often the parties will seek
to direct this process, establishing transfer deals and encouraging their sup-
porters to give preferences to their intended coalition partner.
Fourth, voters can follow party campaign management tactics to ensure an
even spread of preferences among the preferred partys candidates. I t is gen-
erally argued in the literature that parties aiming to maximise the number of
seats they win should try to spread the preferences as evenly as possible
among their candidates, to give all their candidates a good chance of
remaining in the count long enough to pick up preferences from other
parties. An inappropriate distribution of the preference vote between candi-
dates of the same party can lead to that party losing an otherwise winnable
seat. As Cox puts it, the parties should try to redirect votes from vote-rich to
vote-poorer candidates (1997: 144). According to Gallagher: Under the
assumption that no votes transfer across party lines, a partys ideal strategy
under STV is to ensure that all of its candidates have exactly the same
number of votes at every stage of the count, and that when any one is elimi-
nated, his or her votes transfer evenly among the partys other candidates
(1992: 482; for alternative argument, see Cohan 1979).
I f the nature of STV opens up the potential for sincere or efficient
strategic voting and, certainly, there is evidence from Irish studies of this
happening in practice (Bowler and Farrell 1991, 1991a; Gallagher 1999; Jesse
2000; Laver 2000) the particular form this system takes in Australia (as out-
lined in detail in chapter 3) demonstrates the scope available to political
parties to control and direct the voting process to their best advantage.
| 128 |
There is general recognition of the fact that, axiomatically, voters are not
voting in a political vacuum. They (or, at least, the more attentive among
them) have been informed about poll trends and likely outcomes by the
media and the political parties.
12
There is plenty of evidence that the media
and parties can give voters a cue on how best to deploy their vote tactically. A
good example of this was the anti-Conservative tactical voting that occurred
in the 1997 British general election, which saw the unseating of a number of
Conservative incumbents (Curtice and Park 1999; Evans et al. 1998). I n
similar fashion, the Australian parties deploy how to vote cards in AV elec-
tions to the House of Representatives, guiding their supporters on how best
to deploy their preferences to ensure the election of the partys candidate and
maximum damage to the party seen as the greatest threat.
Under STV, Australia has taken the process of directing strategic voting
to an even higher level of sophistication. With the introduction of ticket
voting in 1983, the parties, in essence, took control over the strategic use of
preferences, thereby ensuring that, in the case of those voters voting above
the line, the use of the preferences would be under party control. Australian
voters are in a double bind: they are forced to turn up at the polling station
and, in order for their vote to count as valid, they must complete all prefer-
ences (or some minimum number) . The rational response is to take short-
cuts, opting to vote above the line (ticket voting). After every election, the
Australian Electoral Commission publishes details on the proportions of
voters opting to vote above the line. As Figure 6.2 shows, over time the use
of this option has become virtually universal: very large and steadily growing
proportions of Australians are ticket voting, up from 85.7 percent in 1984 to
95.2 percent in 2001and 95.8 percent in 2004.
I n terms of the four tactics for strategic voting under STV, which we
reviewed above, we can see just how much control the parties now hold over
the flow of preferences of all ticket voters. I n the case of the first three tactics,
this clearly works to the best advantage of the party in question. First, the
parties achieve high degrees of preference solidarity; there is no leakage of
high preferences to other parties.
13
Second, the most significant competitor
party always receives the last set of preferences. Third, the flow of inter-party
preferences is exactly as the party wants it, because the inter-party ticket deals
are determined by the party elites.
But, for all the control over strategic voting by Australian parties,
paradoxically, in the case of the fourth tactic over the spread of preferences
among a partys candidates the parties appear to operate in a strategically
| 129 |
sub-optimal manner. As we discussed above, academic specialists are gener-
ally agreed that the ideal strategy here is to spread the preferences as evenly
as possible among the partys candidates so as to maximise the number of
seats the party might win. Over the years, Irish parties have adopted prefer-
ence spreading as the principal form of vote management (Marsh 2000). As
the parties have professionalised and centralised their operations, new proce-
dures have been adopted to maximise the efficiency of the preference
spreading strategy. Party statutes have been amended and greater use is made
of existing rules as the central leadership has increased its role in candidate
nomination (Farrell 1994; Gallagher 1988; Mair 1987). The parties issue can-
didate cards (equivalent to how to vote cards) that vary the rank ordering of
their candidates in different parts of the constituency. Opinion polls are used
to calm the nerves of those candidates being asked to sacrifice first prefer-
ences for running mates (Farrell 1993).
I n Australia, the strategy tends to be very different. Australian parties
follow a cascade strategy, concentrating first preference votes on particular
candidates, and letting any surpluses transfer down the line to the other party
| 130 |
0
10
20
30
40
50
60
70
80
90
100
1984 1987 1990 1993 1996 1998 2001 2004
Aust ACT Tasmania
SOURCE Australian Electoral Commission
FIGURE 6.2 USE OF SENATE TICKET VOTES, 19842004
candidates. This is the established practice in all Australian upper house STV
elections, and its origins are most likely in the use of compulsory expression
of preferences combined with a strategy of encouraging donkey voting by
selecting candidates whose names start with letters high in the alphabet (in
early days candidates were ranked alphabetically on the ballot paper). With
the onset of preferential electoral systems, Australian parties distributed how
to vote cards with the intention of directing the flow of preferences down the
line. While there are instances of where certain parties attempted to spread
preferences in an Irish vote-management manner, on the whole such experi-
ments did not prove successful and were not continued.
14
Much like program trading on stock exchanges, under the Australian
procedure thousands of preferences cascade down from one candidate to the
next. Prior to 1983, this practice was promoted by a combination of three
factors: a requirement that voters complete all preferences, large ballot papers,
and the use of how to vote cards to help the bewildered voters cope. The
rational strategy for an average voter, being forced to the polling station as a
result of compulsory voting laws, was to follow the instructions on the how
to vote cards, cascading their preferences down the lines of candidates. As we
saw in chapter 2, this system became institutionalised with the move to ticket
voting for Senate elections in 1983, a move followed by all the upper house
STV systems by the end of the decade (Table 3.4). Of the Australian STV
cases, only Tasmania and the ACT have resisted the move towards a cascade
strategy. I n the case of Tasmania, there was always a resistance to the use of
how to vote cards (Wright 1980: 135-6). But once the state adopted the
Robson procedure of candidate name rotation on the ballot paper a proce-
dure also followed by the ACT when it adopted STV this foreclosed any
possibility of cascading preferences. This is because, even if voters chose to
donkey-vote down the line of candidates, the preferences would still be more
evenly distributed because the order of candidates on the ballot paper had
been altered.
I f preference spreading could result in more candidates being elected, why
are Australian parties not using this strategy for Senate (and other upper
house) elections, instead of cascading? Why pile up preferences, if spreading
them can produce more gains? There are two possible explanations. First, it
could be argued that cascading is really only appropriate in a system charac-
terised by a prominence of friends-and-neighbours voting, in which there is
a high degree of localism in voting behaviour. Such is evidently the case in
Ireland, which is renowned for its bailiwick politics (Farrell 1985), facilitating
| 131 |
the carving up of constituencies between the various candidates of the party,
with the voters varying their candidate rank-orders accordingly. But does this
also apply in Australia? Research by Johnston suggests that there is localism
in Australian voting practice. His examination of the 1974 Senate election in
Victoria found very strong evidence for the existence of friends-and-neigh-
bours effects (1978: 152; also OConnell 1983).
A second explanation is based on the fact that preference spreading is a
risky and complex strategy. I ts success depends upon the size of the party vote
in the constituency, on the degree of solidity in vote transfers, and on the
actions of the other parties in the race. I t is an experiment that can go badly
wrong, and for that reason, it is understandable that Australian parties might
be reluctant to adopt it. This problem need not apply in the Australian case,
however, because of the operation of ticket voting. As we have seen,
Australian parties are virtually guaranteed maximum vote solidity in the
transfer of preferences between their candidates. I ndeed, because of the ticket
voting system, Australian parties have an ideal situation should they want to
adopt preference-spreading strategies. They would only have to change the
electoral law to allow the rotation of candidate names (below the line) on the
ballot paper while maintaining the ticket vote. This would ensure an
extremely efficient preference spreading vote management strategy, guaran-
teeing the spread of preferences across all the candidates.
On the face of it, then, it does seem curious that Australian parties choose
not to use a preference spreading vote-management strategy, especially given
the degree to which ticket voting gives them almost total control over the
process. But as we saw in the last chapter, there is good reason for this,
relating to the theme of control although in this case it is control over the
partys candidates rather than over voters. At the heart of this is the issue of
candidate selection. I n an Irish STV election, where candidates are rank-
ordered alphabetically on the ballot paper, the onus is on each candidate to
chase as many votes (and preferences) as possible if they want to be elected.
However, in Australian Senate elections, using STV, the focus of the candi-
dates attention is more on their selectorates (the party group that selects the
candidate) than their electorates. As Jaensch puts it: To win [the] position of
a guaranteed return to the Senate is the aim of all aspirants; and to achieve
this, good relations with the party are more important than service to the
electorate (1986: 59). While power over candidate selection is always an
important means of control of candidates (Schattschneider 1942), the control
is all the greater under list and upper-house STV electoral systems in which
| 132 |
the rank-ordering of candidates is fixed, and the voters are voting for party
rather than candidate (Bowler et al. 1996). I f rotation of candidate names
were to be implemented on Senate ballot papers, this would remove a strong
element of party control over individual candidates (Sharman et al. 2001).
There is a final point to note about the Australian cascade strategy. While
theoretically it may be sub-optimal (Cox 1997; Gallagher 1992), in the
Australian context, all parties operate the same cascade strategy essentially
whether they like to or not due to the ticket vote system. I t is hard to find
tangible evidence of this affecting the electoral fate of individual candidates,
which suggests a cancelling-out effect.
15
Ticket voting, therefore, facilitates a high degree of control by the party
elites both over individual candidates and over the voters, ensuring that the
party elites can direct the flow of candidate preferences, and adding consid-
erable burdens to those voters wishing to deploy their preferences strategi-
cally for particular candidates. Clearly, the rational strategy for the average
voter, under these circumstances, is to simply vote the ticket, by voting above
the line. I n the next section we assess just how strong an effect this has on
the voting act in Australian politics.
Ticket voting and the preferential voter
For all the parties efforts, there is still scope for the independent-minded
voter to ignore party instructions. There are still instances of voters being pre-
pared to ignore the parties how to vote cards in AV elections, and voting
below the line in STV elections, showing clear signs of preferring to vote in
an independent manner. This is evident, for instance, in those parts of
Australia where ticket voting has not tended to be a dominant feature of sub-
national elections. I n chapter 3, we saw how neither Tasmania nor the ACT
operates ticket voting in their elections. This seems to have a spill-over effect
on the proportion of ticket voting in their Senate elections also. As Figure 6.2
shows, the proportion of ticket voting in Senate elections in the ACT and
Tasmania is on the increase, but in the most recent 2004 elections it was still
some 16 points below the Australian average.
Table 6.2 provides some further evidence of just how much Tasmanian
and (especially) ACT voters can buck the trends. I n recent legislative
assembly elections (which, remember, operate without ticket voting systems),
there were significant proportions of voters who transferred preferences in a
| 133 |
| 134 |
TALE .2 CR033-PARTY PREFERENTlAL v0TlN0 lN TlE ACT lN 2001 AN0 TA3VANlA lN 2002
A6T, 2001 Tasman|a, 2002
[N} 7 [N} 7
L|oera|s (35ZZ2) 1.1 (123) 9.2
Laoor (3912) 1.0 (1329) 10.0
0reers (09) 23.2 (2851) 12.2
Aus|ra||ar 0erocra|s (12982) 2Z.1 (1581) 25.8
N0TE8 Propor||ors ol prelererces ||a| |rarslerred |o card|da|es lror o||er par||es. 0r|y |rc|ud|rg cases W|ere ||e par|y |r
ques||or s|||| |ad card|da|es |r ||e race (|.e. |r|erva| |rarslers). T|e Ns are ||e |o|a| ruroer ol |rarslers |rvo|ved lor ||a| par|y.
80UR6E8 Au||ors' ca|cu|a||ors lror e|ec|ora| re|urrs or ||e ACT ard Tasrar|a e|ec|ora| corr|ss|or Weos||es.
TALE .3 '8EL0w TlE LlNE' v0TlN0 lN TlE NEw 30uTl wALE3 LE0l3LATlvE C0uNClL ELECTl0N 0F 1999
A. Patterns of 'be|ow the ||ne' vot|ng
Va||d preferenoes No. [7}
Less ||ar 15 3892 3.1
15 8880 51.8
1-100 503Z0 10.1
101-200 1203 0.9
201-23 00 0.1
21 19 0.5
To|a| 125591 100.0
. Party breakdowns
Voters for
a
Party's tota|
vote [7}
No. of
oand|dates
'e|ow the
||ne' vote [7}
8o||dar|ty
vote
b
[7}
Exo|us|ve
vote
o
[7}
Laoor 3Z.3 15 1.1 58.9 11.3
L|oera|/Na||ora| 2Z.1 15 1.Z 0.1 1.2
0re Na||or .3 5 5.0 Z2.9 0.1
Aus|ra||ar 0erocra|s 1.0 9 11.58 9.0 0.1
C|r|s||ar 0erocra|s 3.2 5 10.38 Z9. 0.3
0reers 2.9 9 12.95 .2 0.1
3|oo|ers 1.Z 11 5.8 Z8.9 0.
Progress|ve Laoor 1. 2 1.55 58.1 1.0
Var|juara 3ro|ers 1.2 2 8.3 8.2 0.
Relorr Lega| 3ys|er 1.0 2 .8Z 39.9 1.1
a: 0r|y |rc|ud|rg par||es W||| a| |eas| 1 percer| ol ||e vo|e.
o: Propor||or ol vo|ers W|o corp|e|ed a lu|| se| ol prelererces lor a|| ||e par|y's card|da|es.
c: Propor||or ol vo|ers W|o vo|ed or|y lor ||a| par|y's card|da|es.
80UR6E 0reer (2003)
non-solid pattern, with a partys transfers bleeding across party lines even
though the party in question still had candidates in the race.
16
Consistent
with findings in British (Dunleavy et al. 1993) and Irish research (Sinnott
1995), the tendency to transfer across party lines is greatest among supporters
of the smaller parties.
There is also recent research from New South Wales that reveals how,
even under a system of ticket voting, there are some stalwarts who are willing
to take the time to use the option of voting below the line (Table 6.3). The
1999 Legislative Council election was the infamous tablecloth election (see
p. 66), but even though less than four percent of the total number of people
who voted did so below the line, what is extraordinary is just how many
people were prepared to take this extra effort in what, as we have seen, was
a record-breaking election in terms of the total number of candidates being
fielded 264 in total. As Table 6.3 shows, a large proportion of these below-
the-line voters were prepared to express a greater number of preferences than
the legally required minimum of 15. There were even some voters (649 in
| 135 |
0
10
20
30
40
50
60
70
80
90
100
1996 1998 2001 2004
Uses 'how to vote card'
for Representatives
Decides own
preferences in
Representatives
Senate ticket vote
Decides own
preferences in Senate
NOTE The questions were: In voting for the House of representative, did you follow a party How tow to vote card
or did you decide on your own preferences?
And in voting for the Senate, did you vote by placing a 1 in a party box above the line or did you decide you
own preferences below the line?
SOURCE 19962004 Australian Election Studies
FIGURE 6.3 VOTERS FOLLOWING PARTY GUIDANCE IN AUSTRALIA, 19962004
all) who expressed a preference for all 264 candidates, although one of these
was a donkey vote, starting at the top left with Peter Breen, then dutifully
filling in all the remaining squares across and down the ballot paper (Green
2003: 27).
The party breakdowns are also interesting, and consistent with expecta-
tions. The supporters of the larger parties are more inclined to vote above the
line; larger party voters tend to express preferences only for that partys can-
didates (to be more exclusive), although in part this reflects the fact that the
larger parties ran a full list of 15 candidates, so voters were not required to
vote for candidates of other parties in order to meet the minimum require-
ment for a valid vote.
There is survey evidence that provides some clues about the numbers and
types of voters who are prepared to vote independently of party direction at
national level. As Figure 6.3 shows, a large proportion of voters in elections
to the House of Representatives claim to have decided their own preferences,
ignoring the how to vote cards. The proportions across the four elections
from 1996 to 2004 are similar, averaging just under 48 percent, and this sug-
gests that these estimates are accurate. The trend suggests that the numbers
of voters deciding their own preferences is slowly increasing, albeit with a
slight decline in the 2004 election. There is a similar consistency in the small
proportions of voters who claimed to have voted below the line in Senate
elections, averaging just over 13 percent across the four elections (note,
though, how these are double the proportions who actually did so, as shown
in Figure 6.2).
What sort of voters choose to decide their own preferences, rather than
follow a party cue? Table 6.4 assesses the socio-economic and attitudinal
characteristics of these more independently minded preferential voters in the
two houses (in this case, dealing with the trends in 2004). While we have no
direct evidence that such voters are any more or less strategic in their voting
patterns than those voting above the line, the very fact that they are choosing
to make up their own minds on how to rank the candidates is certainly
indicative of a desire not to be overly influenced by the parties in expressing
vote preferences, and therefore there is good reason to assume that a fair pro-
portion of these voters are being strategic in their vote decisions.
The patterns are broadly similar between the House of Representatives and
the Senate. Younger voters are more inclined to decide on their own prefer-
ences in the House of Representatives elections, though not in the Senate race,
and across both houses, male voters are more independently minded than
| 136 |
females, net of other things. The only other socio-economic characteristic that
is important is education, and those with a university degree are significantly
more likely to make up their own minds in House of Representatives elections
compared to those without tertiary qualifications voting for the Senate. The
attitudinal characteristics are much more revealing, showing that preferential
voters are more inclined to be late deciders, and to have a consistently jaun-
diced view on the role of parties.
17
There is a consistent tendency for prefer-
ential voters to eschew a partisan identity.
| 137 |
TALE .4 ClARACTERl3TlC3 0F TlE 'PREFERENTlAL' v0TER, 2001
house of Reps 8enate
Est [8E} Est [8E}
$oo|a| $rruorure
Age -.02 (.00) .00 (.00)
0erder (ra|e) .21 (.11) .33 (.1Z)
8|r||p|ace (Aus|ra||a)
8r|||s| ls|es .23 (.1) -.21 (.33)
0||er cour|ry .18 (.1) -.08 (.2)
Educa||or (ro qua||lca||ors)
Ter||ary .19 (.11) .13 (.21)
Pos|secordary -.25 (.13) -.23 (.20)
$rare (h$wj
v|c|or|a -.2Z (.15) .05 (.25)
0ueers|ard .53 (.15) .10 (.25)
3ou|| Aus|ra||a .2Z (.21) .52 (.32)
wes|err Aus|ra||a .08 (.20) .58 (.30)
Tasrar|a 1.19 (.33) 2.18 (.31)
ACT, Nor||err Terr||ory .0 (.33) 1.3Z (.3Z)
/rr|ruoes
Nor par||sar .Z2 (.1) .1 (.21)
La|e vo|e dec|der .31 (.11) -.03 (.12)
8e||eves par||es ro| reeded .1Z (.05) .31 (.0Z)
Cors|ar| -.10 -3.13
Nage||er|e R-squared .13 .13
(N) (1,Z2) (1,Z22)
, p<.05, , p<.01, oo|| |Wo-|a||ed.
N0TE Log|s||c regress|or es||ra|es s|oW|rg parare|er es||ra|es ard s|ardard errors pred|c||rg W|e||er ||e resporder|
dec|ded ||e|r oWr prelererces (scored 1) or used a par|y oa||o| (scored zero) |r ||e louse ol Represer|a||ves ard ||e 3era|e.
80UR6E 2001 Aus|ra||ar E|ec||or 3|udy.
There are also major differences based on state or territory of residence.
Judged against New South Wales voters (the excluded category), voters in
Tasmania, the Northern Territory and (predominantly) the ACT
18
are much
more likely to want to make up their own minds on how they order the can-
didates. I ndeed, the strength and consistency of the results for Tasmania are
particularly notable. Since the regression models take into account a wide
range of potentially confounding factors, it is clear that the effect of
Tasmanias variant of STV at state level is to imbue a greater sense of empow-
erment among its voters in federal elections. The ACTs experience of a
similar variant of STV is, of course, more limited but there too the territorys
electoral system appears to influence federal voting.
Outside of the ACT and Tasmania, then, the preferential voter is a rather
exotic species in Australia. Between the high levels of party control over the
preference-ordering of candidates and the strong element of voter compul-
sion with which the Australian electoral process is imbued, there are plenty
of grounds for suggesting that if the voters are not confused by the complexity
of the vote process, they may well be alienated by the strong arm of control
they face in trying to operate under it.
Compulsion, complexity and voter engagement
The attitudes of Australian voters to the electoral system and to voting are
likely to be conditioned by a number of features: the extent of party control
over the ballot process, the complexity of the ballot (especially before the
introduction of ticket voting in 1983) and the fact that voting is compulsory.
Between them, they provide good reason for expecting to find evidence of low
degrees of voter engagement with the political process. I n fact, as we shall see,
the evidence points to the opposite conclusion.
Today Australia is one of only a handful of developed democracies to
operate a scheme of compulsory voting,
19
and certainly one of very few to
rigidly enforce it. By law, responsibility is vested in the Australian Electoral
Commission (AEC) to police the process, chasing up non-voters, imposing
fines, and if necessary bringing them to court (see McAllister and Mackerras
1999). I n the light of such rigid enforcement of compulsory voting, we might
expect to find a negative reaction to it among the voting population, and
moves being made to abolish it.
But despite its importance to the electoral system, there have been com-
| 138 |
paratively few debates about compulsory voting during the eight decades it
has been in operation. Proponents of the system have emphasised moral
arguments such as civic duty and responsibility, as well as more practical con-
siderations such as efficiency and inclusiveness (for a review, see Hughes
1968). The arguments used against compulsory voting during the 1920s and
1930s mostly asserted that compulsion was wrong in principle and that it is
not the democratic norm. From the 1980s onwards, some Liberal politicians
have advocated abandonment of the system, and a significant number of
Liberal policymaking bodies have endorsed a return to voluntary voting,
notably the partys highest forum, the Federal Council, which passed motions
in 1988 and 1993 calling for a repeal of the current system ( Jackman 1997).
More recently, there has been a minority report (consisting mainly of
Liberal and National members) of a parliamentary committee charged with
examining all aspects of the conduct of the March 1993 elections. The
minority report argued that compulsion encouraged electoral fraud: because
the law forces people to vote, minimal barriers to registration and voting exist,
so that in practice compulsory voting underpins a system which has very few
checks in place to prevent and detect fraudulent enrolment and voting
( JSCEM 1994: 157). The report went on to suggest that the way in which
the system is applied was undermining popular trust in it, as well as being
used as part of a push for a de facto first-past-the-post [SMP] voting system.
| 139 |
0
10
20
30
40
50
60
70
80
1996 1998 2001 2004
Definitely voted
Probably voted
Might, might not
Probably not voted
Definitely not voted
NOTE The question was: Would you have voted in the election if voting had not been compulsory?
SOURCE 19962004 Australian Election Studies
FIGURE 6.4 LIKELIHOOD OF VOTING UNDER A VOLUNTARY SYSTEM, 19962004
The subsequent parliamentary committee report into the 1996 election also
recommended the abolition of compulsory voting, though not of compulsory
registration. As the report put it, if Australia is to consider itself a mature
democracy, compulsory voting should be abolished ( JSCEM, 1997: 26). The
government, however, rejected the recommendation.
20
Given the operation of compulsory voting, and the degree of controversy
in the political debate over it, it is to be expected that there will be implica-
tions for electoral turnout. I n the four Australian Election Study surveys con-
ducted between 1996 and 2004, respondents were asked if they would have
still have voted if voting had not been compulsory.
21
The results are consis-
tent across all four surveys (Figure 6.4). About two out of every three voters
said that they would definitely have voted, with between 17 and 19 percent
saying that they probably would have voted. I n total, then, the surveys
suggest that a shift to a voluntary system of voting would probably result in
a turnout level of just over 80 percent. This is a level that is currently found
only among the top quintile of established democracies (I nternational I DEA
2002). A further 6 to 8 percent said that they were unsure whether they
would have voted, while around one in 10 said that they either probably or
definitely would not have voted under a voluntary system.
To what extent is there a partisan bias to who would or would not vote if
voluntary voting were to be introduced? Figure 6.5 addresses this question by
showing party vote for the House of Representatives by the probability of
| 140 |
0
10
20
30
40
50
60
70
Definitely voted Probably voted Might, might not Probably not voted Definitely not voted
Coalition
Labor
Other
NOTE Vote is first preference vote in the House of Representatives
SOURCE 2004 Australian Election Study
FIGURE 6.5 PARTY VOTE BY PROBABILITY OF VOTING, 2004
voting. The results confirm the view generally held in the literature that the
less likely the voter is to say that she would have voted, then the less likely she
would be to support the Liberal-National coalition, and the more likely to
support either Labor or the minor parties (cf. Crewe 1981; McAllister 1986;
McAllister and Mughan 1986). For example, among those who said that they
definitely would have voted, 51 percent reported voting Liberal-National in
the election; by contrast, Liberal-National support drops to just 37 percent
among those who said that they probably would not have voted if voting had
been voluntary. I ndeed, among the small group who said that they would def-
initely not have voted, Labor support stands at 58 percent.
22
These are substantial electoral effects, but what net impact would they
have had on the outcome of the election, given the relative sizes of the various
turnout groups? Table 6.5 suggests that the effect would have been small, and
not sufficient to change the outcome of the election. The first row of Table
6.5 for each election shows the first preference vote in the House of
Representatives among those who said that they definitely would have voted.
The remaining rows re-estimate the vote adding in the other four turnout
groups. I n the 2001 election, the estimates suggest that Labor gained almost
2 percent of the first preference vote through compulsory voting, while the
minor parties gained 1.6 percent. However, the Liberal-National vote was
particularly sensitive to turnout in that election. I f only those who were com-
mitted to voting had actually done so, the Liberal-National coalition would
have attracted 50 percent of the first preference vote; compulsory voting
meant that they actually attracted just over 46 percent, a net loss of nearly 4
percent. The estimates for the 2004 election in the second part of Table 6.5
suggest more modest effects, though the broad pattern is the same. Labor
gained about half of one percent of the first preference vote through compul-
sory voting, while the Coalition lost about the same, and the vote for the
minor parties was unchanged.
While these effects are not large enough to have changed the outcome of
either the 2001 or 2004 elections, they would have the potential to determine
the result of a closely fought contest. Moreover, the patterns are consistent,
suggesting a Coalition advantage through voluntary voting, and a Labour
advantage through compulsory voting. Other analyses for earlier elections
also indicate a more pronounced partisan effect; in the 1996 election, for
example, the Coalition loss through compulsory voting was about 5 percent
(McAllister and Mackerras 1999).
For all the differences of opinion among the party elites about the future
| 141 |
of compulsory voting, the survey data reveals strong support for its continu-
ance among voters. Going back to the earliest opinion poll that asked voters
views on the topic, conducted in 1943, the proportion of voters who have
supported compulsory voting has never been less than six out of every ten
(Figure 6.6). Support for the system increased gradually during the 1940s,
1950s and 1960s, peaking at 76 percent in a survey conducted in 1969. This
gradual increase is most likely a reflection of the large number of voters who
have grown up under the system, together with the absence of any political
debate concerning its advantages or disadvantages. Support declined slightly
in the 1970s and early 1980s (perhaps because of the large influx of migrants
from countries lacking democratic traditions), but in recent years support has
strengthened. I ndeed, between 1987 and 2004, support increased by 10 per-
centage points. Surveys that permit the respondents to register the strength
of their opinion for or against the system indicate that voters who favour
compulsory voting have stronger views than those who oppose it (McAllister
1992: 2829). Aitkins (1982: 31) point that compulsory voting is in no sense
seen as an imposition on the electorate and resented by it still applies three
decades later.
| 142 |
TALE .5 PARTY v0TE AT 0lFFERENT TuRN0uT LEvEL3, 2001 AN0 2001
Labor L|b-Nat 0ther Tota| [N}
(2001)
0elr||e|y Wou|d |ave vo|ed 35.1 50.0 11.9 100.0 (1 185)
Add|rg:
Prooao|y Wou|d |ave vo|ed 35.3 19.0 15.Z 100.0 (1 508)
V|g||, r|g|| ro| 3.1 1Z.9 1.0 100.0 (1 50)
Prooao|y ro| 3.8 1Z.1 1.1 100.0 (1 ZZ0)
0elr||e|y ro| 3Z.0 1.1 1.5 100.0 (1 Z8Z)
(2001)
0elr||e|y Wou|d |ave vo|ed 3.2 51.0 12.8 100.0 (1 132)
Add|rg:
Prooao|y Wou|d |ave vo|ed 35.1 52.2 12.1 100.0 (1 100)
V|g||, r|g|| ro| 35.9 51. 12.5 100.0 (1 195)
Prooao|y ro| 3.3 50.9 12.8 100.0 (1 5Z5)
0elr||e|y ro| 3.Z 50. 12.Z 100.0 (1 0)
N0TE Es||ra|es are lrs| prelererce vo|e lor ||e louse ol Represer|a||ves.
80UR6E 2001 Aus|ra||ar E|ec||or 3|udy.
One of the characteristics of electoral systems with compulsory voting is
a high level of invalid (or informal) ballots, and, indeed, Australia has one of
the highest levels of invalid votes among the established liberal democracies.
These votes are set out for the two houses of the national parliament in
Figure 6.7. The Senate results are the more striking: from 1901 to 1917
Senate invalid voting was low, averaging just 5.6 percent. However, that was
during the period of voluntary voting with a simple block plurality system
(see Table 3.2, p. 50) and the introduction of preferential block voting in the
early 1920s soon to be followed by compulsory voting raised invalid
voting to almost one in every ten votes cast. With the introduction of ticket
voting in 1983 (see chapter 2), the situation was transformed. I nvalid voting
dropped dramatically, never rising above 5 percent in all elections held since
then. This suggests that the cause of many of these invalid votes has less to
do with compulsory voting as such, and rather more to do with the complex-
ities of the ballot process prior to 1983.
| 143 |
0
10
20
30
40
50
60
70
80
1943 1955 1964 1965 1967 1969 1974 1979 1980 1987 1993 1996 1998 2001 2004
Compulsory
Voluntary
NOTE Question wordings differ slightly prior to 1987. The AES question is Do you think that voting at Federal
elections should be compulsory, or do you think that people should only have to vote if they want to?
SOURCE Aitkin (1977:31);1979 Australian National Political Attitudes Survey; 1987-2004 Australian Election Study
surveys.
FIGURE 6.6 ATTITUDES TO COMPULSORY VOTING, 19432004
I n both the 1984 and 1987 elections, for the first time in Australian elec-
toral history, invalid voting in Senate elections was less than that for the
House of Representatives, and the same occurred at the 1998 and 2001 elec-
tions. I n subsequent elections the system has stabilised, with an average infor-
mality level of around 3 percent for both houses. This level is still relatively
high by international standards, but low in comparison to earlier Australian
elections. The two most recent elections show a modest increase in invalid
voting; in the 2001 election it stood at 4.8 percent in the House of
Representatives and 4.1 percent in the Senate, and in 2004 the figures were
5.2 percent and 3.8 percent respectively.
I n addition to compulsory voting and ballot complexity, invalid voting in
Australia is also linked to other factors: the regularity of elections in which
Australia competes with the USA in terms of how often voters are asked to
attend the polls and the large variations in what is asked of voters across the
various elections.
23
National elections are held every three years, but usually
more frequently, as are all state elections except New South Wales, where they
must be held every four years (a change introduced in 1981). I n practice, this
means that voters sometimes have to cast a ballot as often as once every 18
months. Not only are elections regular, they also tend to vary in terms of what
| 144 |
0
2
4
6
8
10
12
1974 1975 1977 1980 1983 1984 1987 1990 1993 1996 1998 2001 2004
Representatives
informal
Senate informal
SOURCE McAllister, Mackerras and Boldiston (1997) ; AEC.
FIGURE 6.7 INVALID VOTING BY HOUSE, 1974-2004
is expected of the voters. For instance, there is an increasing tendency to hold
House of Representatives elections in conjunction with Senate elections and
constitutional referendums, all of which have different rules for completing
the ballot. For referendums, the elector writes yes or no in the space pro-
vided, while the Senate ballot paper requires either one number above the line
or a great many more numbers below the line. Finally, as we discussed in
chapter 3, the rules differ for casting a valid vote in state and national elec-
tions. Some Australians live in states with optional preferences (such as New
South Wales or Queensland), while for national elections, preferences are
compulsory. When all these factors are combined, it is not surprising, there-
fore, that the net result is relatively high levels of invalid ballots. But it is
worthwhile reiterating that compulsory voting is only partially to blame for
invalid voting in Australia.
To summarise, the evidence points to high levels of voter support for the
electoral process in Australia, as evidenced by the approval ratings for com-
parative voting in survey data and also by the reduced levels of invalid voting
since the mid-1980s, when the ballot process was simplified. But this latter
set of evidence on invalid voting does lend support to the argument that these
preferential systems are complex and may confuse some voters.
Conclusion
Proponents of preferential systems particularly STV, but to an extent also
AV extol their virtue in providing voters with more choice and limiting the
degree of party control over the electoral process (e.g. Lakeman 1974).
Furthermore, these systems enable voters, if they wish, to vote both strategi-
cally and sincerely. Having said that, the voting act inevitably is more
complex, raising questions over whether voters are capable of making full and
accurate use of the system.
Australia offers a particular twist to all of this. The combination of com-
pulsory turnout, compulsory expression of preferences and complex ballot
papers add greatly to the burdens placed on voters. I n the case of AV elec-
tions, the rational voter follows the advice given in the how to vote cards.
This was also the scenario for STV elections prior to 1983. The introduc-
tion of ticket voting in that year brought party control over the act of voting
to an entirely new level, and, indeed, to a level not seen in any other elec-
toral system. Not only does an Australian party control the voter prefer-
ences for its candidates, it can also determine the direction of inter-party
| 145 |
preferences and thereby guarantee that its supporters will vote in a manner
consistent with intended post-electoral government formation goals. No
other electoral system provides parties with that degree of control over
events in the ballot box.
While recent survey evidence indicates that there is a group of voters who
are less inclined to follow the party line (ignoring the how to vote cards, or
voting below the line), this group is small and declining in number. The fact
that the vast bulk of Australian voters tend to let the parties determine the
flow of their preferences, combined with the other elements of control over
the act of voting (compulsory registration, compulsory turnout, complex
ballot process), would seem to provide many of the ingredients for voter
alienation, as might be manifested through high levels of invalid voting and
survey evidence of voter disengagement. However, as we have seen, there is
little evidence to support the latter, and while there is invalid voting, the indi-
cations are that this has more to do with confusion on the part of some voters
than with widespread alienation.
The juxtaposition of preferential voting with compulsory registration,
voting and expression of preferences seems unusual. The former suggests an
emphasis on voter choice, the latter on party control. This in-built tension
presents a challenge to the electors deciding on whom to vote for, and how
best to deploy their vote; it also influences the strategic environment under
which the politicians and parties must operate, as we saw in chapter 5.
To date, Australian electoral designers appear to have achieved the
admirable feat of maintaining a tight rein over the electoral process while not
alienating the voting population. But this has not been without some cost,
both to the ideals of preferential electoral systems as voter-oriented
allowing greater voter choice than other electoral systems and, in turn, to
the freedom of manoeuvre enjoyed by party politicians.
| 146 |
s e v e n
voter satisfaction
Our analysis of voter attitudes to Australias preferential electoral systems in
the previous chapter was constrained by the interference from the other fea-
tures of Australias electoral institutions notably the elements of voter com-
pulsion and the strong hand of the party over all matters electoral. So even
though the evidence, registered across successive opinion polls, pointed to
high levels of voter support for Australias electoral institutions, we are
unable to use this as categorical proof of the (often polemical) arguments
made regarding the popularity of the various electoral systems that promote
voter choice. The object of this chapter is to look for evidence in support of
such contentions, based on an analysis of comparative trends. I n order to do
so, we link the analysis to international debates over voter satisfaction with
democracy.
Since the onset of the current wave of democratisation, there has been a
growing interest in researching and identifying the institutional factors
underlying citizen support for democracy. This has also, in part, reflected a
renewed scholarly interest in seeking answers to the questions of whether and
how institutions matter (Weaver and Rockman 1993) in this instance, with
regard to the theme of democratic stability. Of all the political institutions
that matter, few would deny that electoral systems are among the most sig-
nificant. They are the central institutional design issue for a new polity to
resolve (e.g. Sartori 1997; Taagepera 2002); and they are also among the most
malleable of the political institutions, a point first stressed by Sartori, who
| 147 |
wrote of the electoral system as the most specific manipulative instrument of
politics (1968: 273). Certainly, when compared with the other fundamental
institutional decisions of a polity (such as deciding between presidentialism
and parliamentarism), electoral systems are generally far easier to change
because they tend not to be constitutionally embedded and are therefore more
open to the whims of politicians and their parties (Bowler, Carter and Farrell
2003).
Electoral system design may be perceived as important by academic
scholars and electoral engineers (Lijphart 1994; Reynolds 2002; Sartori 1997;
Taagepera and Shugart 1989), but what tangible evidence is there of an elec-
toral system actually making a difference to democratic stability?
Developments in large-scale cross-national surveys have facilitated research
into this question, with much of it focused on Arend Lijphart s well-known
framework distinguishing between majoritarian and consensual democracies
(Lijphart 1999). Under this framework, the electoral system is treated as a
core variable and is operationalised in terms of its vote-aggregation proper-
ties as more or less proportional. The basis of Lijphart s argument with regard
to electoral systems is that proportional systems are better than non-propor-
tional systems because they facilitate the representation of all relevant societal
and ethnic groupings.
There is more to this argument than theoretical conjecture. I n the penul-
timate two chapters of his Patterns of Democracy (1999), Lijphart provides
detailed supporting evidence, using aggregate and survey data. This has
prompted others to examine the cross-national evidence, though in some
cases with differing conclusions (Castles 1994; Rose 1992). Two prominent
studies are by Anderson and Guillory (1997) and Norris (1999), both of
whom use cross-national survey data to test the role of proportional represen-
tation electoral systems in promoting citizen satisfaction with democracy.
The Anderson and Guillory study, based on Eurobarometer data, finds clear
evidence of a positive relationship between PR systems and measures of dem-
ocratic satisfaction. By contrast, Norriss findings, using a far wider set of
cases (based on World Values data), are more equivocal: if anything, majori-
tarian institutions tended to produce greater institutional confidence than
consociational arrangements (1999: 233; also Norris 2002, 2004).
Whether and how electoral systems can affect levels of voter satisfaction
with democracy is therefore unresolved; work is required, among other
things, on how best to operationalise the dependent variable (Linde and
Ekman 2003). I n this chapter we argue that the independent variable is also
| 148 |
in need of more attention. The studies to date have treated electoral systems
solely in terms of their vote-aggregation outcomes, with attention focused on
representation in the microcosmic sense (McLean 1991) in terms of who gets
elected and the representation of social and ethnic groups in short, the age-
old issue of proportionality. I t is universally accepted in the electoral systems
literature that the two most important features affecting an electoral systems
aggregate proportionality are district magnitude (the number of politicians
elected in a district) and electoral formula (the counting rule determining
how votes are translated into seats) (e.g. Farrell 2001; Katz 1997a; Lijphart
1994; Rae 1967). But a third feature of electoral systems is ballot structure,
and while its effects on proportionality are negligible (Lijphart 1994), there
is no disputing that it has an effect both on voters, in the sense of determining
the nature and extent of choice available to them on polling day, and on
politicians, who are cognisant of the effect on voters and react accordingly.
The question this chapter seeks to answer is whether ballot structure has
a wider impact on levels of voter satisfaction with democracy and, if so, where
Australia fits into this continuum. We start, in the first section, with a discus-
sion of the ballot structure dimension of electoral systems, how it might be
operationalised, and what effects we might expect it to have on voter attitudes
to politics. Unlike earlier studies that distinguish between proportional and
non-proportional systems, this book has been focusing principally on the dis-
tinction between preferential and non-preferential systems, and this chapter
is consistent with this overriding theme. The second section of this chapter
uses two waves of data from the Comparative Study of Electoral Systems
(CSES) to assess the role of ballot structure, as well as a battery of other elec-
toral system and individual-level variables, in influencing voter satisfaction
with democracy in thirty countries.
1
The third section takes a further look at
the evidence by examining the potential indirect relationships between our
electoral systems variables and measures of satisfaction with democracy.
Ballot structure and democracy
The connection between ballot structure characteristics of electoral systems
and voter attitudes to democracy derives from at least three mechanisms,
one originating directly from the voting act itself, another based more indi-
rectly on the relationship between politicians and their voters, and a third
referring to the ideological tendency within the party political system. I n
the first instance, and as we reviewed in chapter 1, there is the argument,
| 149 |
| Preferenti al systems and voter sati sfacti on |
usually propounded by supporters of preferential systems like STV, that a
principal strength is the maximisation of voter choice (Bowler and
Grofman 2000a; Hallett 1984; Lakeman 1974).
2
I n a context in which
voters are being given more choice in the electoral act, it could be argued
that this should result in a greater sense of efficacy on the part of voters, and
more say in electing their representatives.
Following on from this, a second perspective draws attention to the
linkage between politicians and voters (e.g. Lawson 1980; Mitchell 2000). As
we saw in chapter 5, there is general agreement that ballot structure affects
the representative role of politicians (e.g. Ames 1995; Carey and Shugart
1995; Shugart 2001), and studies based on surveys of politicians have demon-
strated how electoral systems that are characterised by a candidate orientation
in politics and high degrees of preferential voting tend to produce greater
attention to personal vote chasing by politicians and the maintenance of close
links with their electorates (Bowler and Farrell 1993; Farrell and Scully 2003;
though see Katz 1997). There are grounds for expecting a more positive atti-
tude by voters towards their elected representatives in such systems (and, by
extension, towards the wider political system) than might be the case where
politicians are devoting most attention to their internal party selectorates.
Finally, there is the issue of how preferential systems might impact on
party politics more generally. A distinction can be drawn here between cen-
trifugal and centripetal tendencies, in other words of being impelled towards
(or away from) a centrist political position (Cox 1990; Sartori 1976). The key
work here in the context of engineering stability in divided societies is by
Ben Reilly (2001), who builds on earlier work by Timothy Sisk (1995) and
Donald Horowitz (1991). The principal argument is that preferential elec-
toral systems encourage a politics of accommodation between politicians.
Reillys perspective is unlike Lijphart s consensual framework (most recently,
Lijphart 2002), which places stress on the need for strong, cohesive parties
supported by closed list PR, in which the party elites can forge alliances
across party lines and promote a top-down politics of accommodation. Reilly
adopts a more bottom-up approach, in which the electoral incentives pro-
moted by preferential electoral systems encourage politicians to move towards
the centre ground in an effort to sweep up more support in the form of vote
preferences (see also Reynolds 1999, on STV). The implication of this argu-
ment is that such tactics, if successful, should help democratic stability and
therefore should be manifest in higher levels of voter support for democracy.
This last mechanism regarding the possible centripetal tendencies of
| 150 |
preferential systems suggests a significant role for these systems in newer
democracies, in helping them to consolidate.
As we saw in chapter 1, when the focus is on ballot structure, electoral
systems can be differentiated in terms of the natureof the vote choice party-
based versus candidate-based votes and also the extent of the vote choice
in terms of degrees of ordinality, or preferential voting. Systems characterised
by candidate-based voting and high degrees of ordinality the preferential
systems encourage a greater emphasis by candidates on cultivating personal
votes, as opposed to party-centred electoral systems, where the fate of the
candidate is determined largely by the support for their parties.
3
Following
the lead of Matthew Shugart (2001, which in turn builds on his earlier work;
see Carey and Shugart 1995), we develop an index of intra-party efficiency,
which taps the preferential/non-preferential distinction in electoral systems.
4
Like Shugart, we operationalise variations in our sample electoral systems
based on three main characteristics that he terms Ballot, Voteand District, in
which the higher scores across these components indicate a candidate-
centred preferential system and lower scores denote a party-centred categor-
ical system. We have adapted Shugart s original schema for several reasons.
First, our focus is on a wider range of electoral systems than is dealt with in
his paper, and furthermore, unlike him, our focus is less exclusively on testing
the overall efficiency of the mixed-member electoral systems. Second, there
were limitations in our access to full and complete data on all the cases
covered by the CSES data set. Finally, we disagree with some of his categori-
sations, as outlined below.
The Ballot component is designed to measure the degree of party versus
voter control over the ballot placement of candidates, revealing the extent to
which the party leadership (and/or selectorate) can exercise influence over the
partys candidates. The lower the ballot control, the greater the incentive for
candidates to place emphasis on their personal reputation. Our coding
which differs from Shugart s only in the sense that we do not include details
on candidate selection rules
5
is as follows:
1 Voters may not disturb the order of the candidate list.
2 Voters may disturb the order of the candidate list.
3 Ballot access requires first surviving a preliminary round of popular voting
(a runoff election).
4 Ballot access is nearly unrestricted.
| 151 |
Shugart s Votecomponent distinguishes between systems that require voters
to cast a ballot for lists and those where the vote is for individual candidates
(a nominal vote). Our coding of this component also differs from Shugart s,
but in this instance it is because we disagree with some of his underlying
assumptions. I n order to understand the basis for our disagreement, we first
need to provide some background on how Shugart arrived at his codes. The
Shugart codes build on an earlier work (Carey and Shugart 1995), entailing
a merger of what were two distinct components in that paper, Pool
1
and
Votes
1
.
6
The first component, Pool
1
, was designed to measure the extent to which
votes cast for one candidate also contributed to the number of seats a party
wins in the district: the greater the extent of pooling, the lower the incentive
for candidates to cultivate personal votes. On one extreme are electoral
systems that have no pooling, so that a candidates fate is determined solely
by her personal vote chasing activities. Carey and Shugart view Japans (now-
defunct) SNTV system and electoral systems that incorporate candidate pri-
maries as good examples of this. AV and STV are categorised as an
intermediate category, where voters can designate to whom their votes
should be transferred if they are not needed to elect their first choice, or if
their first choice is too unpopular to be elected (Carey and Shugart 1995:
422), whereas the list systems, regardless of their variation in terms of degrees
of ordinality, are subsumed under the one category of pooling across whole
party, and are placed on the opposite extreme to SNTV.
7
Carey and Shugart s placement of AV and especially of STV is problem-
atic, and certainly is at odds with the usual conceptualisation of these systems
as promoting candidate-centred electoral politics (for discussion, see Marsh
2000; Katz 1980). I t may well be true that voters can designate how their
votes will transfer, but what has to be recognised is that the candidates (in
varying degrees of desperation) are seeking to influence the nature of that
designation. The result is a high degree of candidate-centred politics. For that
reason, we find it hard to support the contention that the application of the
Pool
1
component should result in STV being ranked at a midpoint on a scale
of party-centred versus candidate-centred politics.
The Votes
1
component, as outlined by Carey and Shugart (1995), distin-
guishes between single-vote list and nominal systems, with systems of mul-
tiple votes (that is, ordinal systems) comprising an intermediate category.
According to Carey and Shugart, the value to legislative candidates of per-
sonal reputation is highest in nominal systems (1995: 422). They provide very
| 152 |
little indication of how they might distinguish the different multi-candidate
systems, but the following extract is revealing:
[ W]hen multiple votes are cast, personal reputation is not as over-
whelmingly important relative to party reputation as when all candi-
dates are competing simultaneously for the same indivisible support
of each voter. When multiple votes are cast simultaneously, the can-
didates from one party can run as a bloc, rather than running against
each other (p. 422).
Again, this appears to underestimate the degree of intra-party candidate-
centred campaigning that often occurs in preferential elections such as STV
in Ireland. I n some instances, this takes the form of full scale personal battles
between candidates of the same party (e.g. Farrell 1985).
I n his more recent paper, Shugart (2001) merges the Pool
1
and Votes
1
com-
ponents, and his new Votecomponent comprises the following ranking:
1 Vote for list only.
2 Vote list or nominal, but list votes predominate.
3 Vote is nominal only, but vote may pool or transfer to other candidates.
4 Vote is nominal or list, but nominal votes predominate and pool to other
candidates.
5 Vote is nominal only and non-transferable.
Clearly there is no disagreeing with the view that a closed list system lies at
one extreme on this component, with ordered list systems lying in an inter-
mediate position. For the reasons set out above, our biggest area of difference
relates to those systems in which the vote is nominal only and non-transfer-
able the single member plurality systems in which the vote is a simple cat-
egorical choice between candidates.
8
While this is likely to promote greater
degrees of candidate-centred politics than closed list systems, because candi-
dates are striving for personal votes in their district, it is inconceivable that
this would result in higher levels of personal vote chasing than systems where
the vote is nominal only with the vote pooling or transferring to other candi-
dates that is, STV. We also contend that STV systems promote personal
vote chasing to a higher degree than open list systems (where the vote is
nominal or list, but nominal votes predominate and pool to other candidates).
As a consequence, our Votecomponent is coded as follows:
| 153 |
1 Vote for list only.
2 Vote list or nominal, but list votes predominate.
3 Vote is nominal only and non-transferable.
4 Vote is nominal or list, but nominal votes predominate and pool to other
candidates.
5 Vote is nominal only, but vote may pool or transfer to other candidates.
I n the case of the District component, our coding is identical to Shugart s.
This component takes account of an earlier argument developed by Carey
and Shugart (1995) about how the effect of district magnitude (M) can vary
depending on the nature of the ballot structure. I n systems where voters cast
party-based votes, they find that the personal reputation of the candidate
declines in significance as M rises, whereas in systems characterised by can-
didate-based (nominal) votes, as M rises and candidates face more inter-party
and intra-party competitors, the incentives for personal vote chasing
increases.
9
This component is coded as follows:
1 M>1, with Vote=1.
2 M=1.
3 M>1, with Vote>1 and Ballot>1.
| 154 |
TALE 7.1 3C0RlN0 ELECT0RAL 3Y3TEV3 0N TlE lNTRA-PARTY 0lVEN3l0N
|ndex 6omponent soores 0esor|pt|on of system 6ases
10.0 1, 5, 3 3Tv lre|ard
8. 1, 5, 2 Av Aus|ra||a
Z.1 2, 1, 3 0per ||s|, parac|age 0errar|, 3W||zer|ard
5.Z 3, 3, 2 330, |Wo rourds Frarce, u3A
1.Z [VV lx A| VV3 W||| ruroll ru|e lurgary, L|||uar|a
1.3 1, 3, 2 330-p|ura|||y W||| par|y cor|ro| Carada, uK
3. [VV lx 8| VV3 W||| p|ura|||y ru|e 0errary, Japar, Korea, Vex|co, NeW Zea|ard,
Ta|War
2.9 2, 2, 1 0rdered ||s| 8e|g|ur, Czec| Repuo||c, Ne||er|ards, Po|ard,
3|over|a, 3Weder
1.1 1, 1, 1 C|osed ||s| 8u|gar|a, lce|ard, lsrae|, NorWay, Por|uga|,
Rorar|a, 3pa|r
N0TE8 T|e rar||rgs rarge lror ros| card|da|e-cer|red |o ros| par|y-cer|red e|ec|ora| sys|ers. For de|a||s or ||e cod|rg,
see |ex| ard ||e Apperd|x Tao|e A.2.
VV lx A: For VV sys|ers W||| p|ura|||y ru|es |r ||e 330s. Vear ol 5.Z (330-p|ura|||y score) ard 1.1 (c|osed ||s| score)
= 3.. Vear ol 3. ard 5.Z (330-p|ura|||y score) = 1.Z.
VV lx 8: For VV sys|ers W||| ruroll ru|es |r ||e 330s. Vear ol 1.3 (330 |Wo rourds score) ard 1.1 (c|osed ||s| score)
= 2.9. Vear ol 2.9 ard 1.3 (330 |Wo rourds score) = 3..
Using this coding scheme, we can locate our 29 countries
10
on a scale from
predominantly candidate-centred to predominantly party-centred; the results
of this exercise are shown in the first part of Appendix Table A.2. For now,
taking all except the mixed-member (MM) systems into account, this pro-
duces seven different grades. The scores on our index do not carry any
numerical value; the index is simply based on the two logical extremes, taking
account of the seven intervals. The resulting index values are reported in the
first column of Table 7.1.
Following Shugart s lead, a separate set of assumptions apply with regard
to MM systems. These mix nominal and list voting elements: since research
demonstrates that they tend on the whole to place the greater emphasis on
the particularistic features of SMP systems (in terms of the legislative role of
MPs) than on closed list systems (Shugart and Wattenberg 2001), we need to
produce an index rating that places them somewhat closer to SMP systems.
I n order to do this, we follow separate procedures for those MM systems that
use plurality rules for the nominal vote element and those systems that use
runoff rules for the nominal vote element.
Systems that use plurality votes for the nominal vote element in the single
seat districts (or SSDs) include Germany, Japan, Korea, Mexico, New
Zealand, and Taiwan. For these systems, we first take the mean of 4.3 (SSD-
plurality score) and 1.4 (closed list score), which is 2.9. Then, to reflect the
bias in favour of SSDs, we take the mean of 2.9 and 1.4 (SSD-plurality
score), producing a score of 3.6.
Those MM systems that use runoff rules for the nominal vote element are
Hungary and Lithuania. For them, the means are based on the SSD two-
round systems rather than on the SSD-plurality systems, thus producing a
score of 4.7.
The next section tests what role this intra-party variable has in influencing
levels of satisfaction of democracy.
voter satisfaction with democracy
I n the first instance, we can assess where our 29 countries lie on a scale meas-
uring levels of satisfaction with democracy vis--vis their placement on our
intra-party measure. This is set out in Figure 7.1, which uses the intra-party
measure outlined in Table 7.1 and a question about satisfaction with the
| 155 |
| 156 |
10
8
6
4
2
0
0 2 4 6 8 10
Satisfaction with democracy
I
n
t
r
a
-
p
a
r
t
y
C
l
o
s
e
d
O
p
e
n
More Less
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1 Australia
2 Belgium
3 Bulgaria
4 Canada
5 Czech
6 Denmark
7 France
8 Germany
9 Hungary
10 Iceland
11 Ireland
12 Israel
13 Japan
14 Korea
15 Lithuania
16 Mexico
17 Netherlands
18 New Zealand
19 Norway
20 Poland
21 Portugal
22 Romania
23 Slovenia
24 Spain
25 Sweden
26 Switzerland
27 Taiwan
28 UK
29 USA
27
28
29
FIGURE 7.1 BALLOT STRUCTURES AND VOTER SATISFACTION
democratic process, which is scored from a low of zero to a high of 10.
11
The
correlation between the two variables is 0.33, showing that the more candi-
date-centred the electoral system, the higher the level of popular democratic
satisfaction. At one end of the spectrum are ideal country types, including
Ireland and Australia, which have strongly candidate-centred systems and
very high levels of voter satisfaction. At the other end, there is Bulgaria, and
to a lesser extent Slovenia, which exhibit strongly party-centred electoral
systems and the lowest levels of voter satisfaction among the 29 countries in
our sample. I nevitably, there are also countries that do not conform to this
pattern: Norway, the Netherlands, Spain and I celand all have relatively high
levels of voter satisfaction but also maintain electoral systems that are,
according to the coding, party rather than candidate-centred.
These results, of course, are based on the 29 countries each treated as a
unit of observation; and this is reflected in the relatively high correlation (.33)
between the two variables. But voter satisfaction is a consequence of a wide
range of individual circumstances in addition to systematic characteristics,
and all must be taken into account in order to test the robustness of the
finding in Figure 7.1. I n order to control for these potentially confounding
characteristics, we rely on the individual level CSES data, which include
40403 observations across the 29 countries. The results are estimated for
those who reported voting in the previous national election.
12
I n addition to our intra-party variable, the other independent variables
are of two types. First, there are the characteristics of individual voters,
ranging from gender, age and marital status, to education, employment char-
acteristics and income.
13
Following the lead of the studies discussed above
(Anderson and Guillory 1997; Lijphart 1999; Norris 1999), the other char-
acteristics of the electoral system that we take account of in our multivariate
model include the following (see the second half of Appendix Table A.2).
Disproportionality is measured by the Gallagher index:
14
the higher the
figure, the greater the disproportionality produced by the electoral system.
Assembly size is also included, and we should expect greater proportionality
where the size is larger, although this could be countered by higher levels of
disproportionality (Lijphart 1994). Finally, the average voting age popula-
tion per MP measures the proximity of the voter to the elected representa-
tive, and we would hypothesise that smaller ratios of voters to MPs would
increase voter satisfaction (Carey and Shugart 1995) since elected represen-
tatives would be more visible and more accessible to their electors.
15
I n order to estimate the net contribution of the intra-party dimension to
voter satisfaction, we provide four regression models (Table 7.3). The first
model includes just one independent variable the intra-party dimension
and the standardised coefficient therefore represents the correlation between
the two items.
16
The second equation includes the measures of voter charac-
teristics. I n this model, while the intra-party dimension is very slightly
reduced in magnitude, it remains the strongest predictor in the equation;
other variables of note are the importance of family income and unemploy-
ment, with more affluent respondents being more satisfied than their less
affluent counterparts, net of other things. Net of a wide range of individual
characteristics, then, the intra-party measure remains an important predictor
of the degree of voter satisfaction with the democratic process in the partic-
ular country.
The intra-party variable maintains its predictive strength when we intro-
duce the three additional electoral system characteristics into the equation.
Two of the three measures are important assembly size just fails to reach
| 157 |
statistical significance and this is broadly consistent with the studies cited
above (Anderson and Guillory 1997; Lijphart 1999). The fourth equation
shows that much of the importance of the intra-party dimension is attribut-
able to country characteristics, which are measured here by the length of the
democratic experience and GDP. Although the intraparty dimension remains
statistically significant, the coefficient drops from .16 to .04, suggesting that
much of its impact in shaping democratic satisfaction is a consequence of the
accumulated democratic experience and economic development of the
country in question.
17
Electoral systems and political attitudes
So far, we have viewed electoral systems as having a direct effect on satisfac-
tion with democracy, and the results in the previous section have shown that
there is strong empirical support for this proposition, net of a range of other
factors. This supports the view of many electoral system designers that elec-
toral institutions can and do have a significant effect on how voters view the
| 158 |
TALE 7.2 vARlA8LE3, 3C0RlN0 AN0 VEAN3
Var|ab|e 8oor|ng Hean 8t 0ev
0epenoenr var|ao|e
3a||slac||or W||| derocracy 0=very ursa||sled, 3.3=ursa||sled,
.Z=sa||sled, 10=very sa||sled
5.9 2.Z1
vorer onaraorer|sr|os
Age Years 1.1 1.15
0erder 1=ra|e, 0=lera|e .19 .50
Ter||ary educa||or 1=yes, 0=ro .21 .10
urerp|oyed 1=yes, 0=ro .05 .22
Far||y |rcore 0u|r|||es 2.99 1.38
Drner e|eorora| sysrem onaraorer|sr|os
lr|ra par|y |rdex Fror a |oW ol 0 |o a ||g| ol 10 3.83 2.18
0|spropor||ora|||y Fror a |oW ol 1.19 |o a ||g| ol 21.3Z .39 1.2
Assero|y s|ze Nuroer |r |urdreds 2.93 1.
vo||rg age popu|a||or per VP Nuroer |r ||ousards Z.98 80.1
0ounrry onaraorer|sr|os
0erocra||c exper|erce Years s|rce ur|versa| lrarc||se 52.00 33.20
00P per cap||a T|ousard u30 1Z.91 8.15
80UR6E C3E3 coro|red (rodu|e 1 ard rodu|e 2) da|ase| lor ||e 29 cour|r|es ||s|ed |r Apperd|x Tao|e A.2.
democratic system as a whole. A further possibility, which we test in this
section, is that electoral systems also have indirect effects on satisfaction with
democracy, via the more general political views that citizens form about how
the system operates and how effective it is.
18
Electoral systems are here con-
ceptualised by our main intra-party variable, and also by the more conven-
tional features of aggregate disproportionality, assembly size, and the ratio of
MPs to the voting age population (VAP). The analysis in this section pro-
vides a sterner test of our intra-party variable, as well as of the other electoral
system variables, because on this occasion we are looking for evidence of the
actual mechanisms of linkage between the electoral system feature and the
levels of voter satisfaction with democracy.
In the first section of this chapter, we outlined three such mechanisms
relating specifically to our intra-party variable. The first mechanism was a sense
of voter efficacy deriving directly from the voting act itself: voters in preferen-
| 159 |
TALE 7.3 PRE0lCTlN0 v0TER 3ATl3FACTl0N lN F0uR V00EL3
Eq 1 Eq 2 Eq 3 Eq 4
b beta b beta b beta b beta
lntre-perty dimension .1 .12 .15 .12 .1 .13 .01 .03
Age .00 .01 .00 .00 .00 .00
0erder .1Z .03 .1Z .03 .11 .03
Ter||ary educa||or .18 .03 .10 .02 .00 .00
urerp|oyed -. -.05 -.2 -.05 -.1Z -.01
Far||y |rcore .11 .0Z .11 .0Z .11 .0Z
0|spropor||ora|||y -.01 -.08 -.01 -.0
Assero|y s|ze . -.02 -.02 .02 .01
vo||rg age pop per VP .01 .08 .00 .00
0erocra||c exper|erce .01 .15
00P .01 .13
Cors|ar| 5.10 1. 1.83 3.Z5
Adj R-squared .02 .03 .01 .08
s|a||s||ca||y s|gr|lcar| a| p<.01, |Wo-|a||ed.
N0TE8 0rd|rary |eas| squares regress|or es||ra|es s|oW|rg par||a| (o) ard s|ardard|zed (oe|a) coellc|er|s pred|c||rg ||e
prooao||||y ol sa||slac||or W||| derocracy. 3ee Tao|e Z.2 lor de|a||s ol var|ao|es ard scor|rg. N=5 152.
80UR6E C3E3 coro|red (rodu|e 1 ard rodu|e 2) da|ase|, lor ||e 29 cour|r|es ||s|ed |r Apperd|x Tao|e A.2.
tial systems felt greater empowerment over the political system, compared to
voters in other electoral systems. The second was the degree of linkage between
politicians and voters, so that voters in preferential systems had easier access to
their elected representatives. Third was ideological tendency within the party
political system. In large part, the first of these mechanisms formed the basis of
the analysis in the second section of this chapter, which examined the direct
relationship between the nature of the electoral system and the attitudes of
voters towards the political system. The remaining two mechanisms suggest a
more indirect relationship with, in turn, politicians and parties playing key
mediating roles. Given the focus of our other electoral system variables (partic-
ularly disproportionality and assembly size) on the aggregate proportionality of
the election result, we should add a fourth mechanism of linkage, namely the
degree to which the election result is perceived as fair.
Taking our (remaining) three mechanisms together, therefore, we have the
ingredients for a systematic examination of how electoral systems link with
voter attitudes to democracy. I n the first instance, there are voter attitudes to
the role of political parties. We would expect that the more open the system
is to minor parties and independents, the greater the support for parties as an
organising element within the political system. Conversely we might expect
a negative relationship between our intra-party variable and parties, reflecting
a greater emphasis on candidate-centred voting. Second, the extent to which
voters believe that their elected representatives are responsive to and knowl-
edgeable about their needs should be related to smaller assembly sizes, to the
lower ratios of elected representatives to electors and also to the nature of the
ballot structure. Third, we would expect that the perceived fairness of the
electoral system would be a consequence of disproportionality and possibly
also the assembly size. I n this case, we might expect that more proportional
systems will be associated with greater support for the view that the election
was fair. The perceived fairness could also be related to ballot structures, in
the sense that voters in candidate-centred systems might feel a greater sense
of ownership over the electoral process due to their determination of the fate
of individual candidates.
Questions measuring these three dimensions were asked in the first CSES
module and are shown in Table 7.4. Unfortunately, the questions were not
asked in the second CSES module, and for that reason the results presented
in Table 7.4 are based on 26 countries, and exclude Ireland, Bulgaria and
France. I n most countries, political parties are regarded as necessary in order
to make the system work; almost half of the respondents in Table 7.4 support
| 160 |
this proposition, and just one in ten hold the opposing view. Similarly, just over
half see the last election in their country as having been conducted fairly; once
again, just one in ten of the respondents take a negative view. Opinions are
more evenly balanced on whether or not elected representatives know what
ordinary voters think; while one-third take a neutral view, 40 percent believe
that they are not in touch compared to 28 percent who believe they are.
These attitudes are related to the objective aspects of the electoral
system in the models outlined in Table 7.5. The intraparty dimension is sig-
nificant in shaping attitudes in two out of the three cases: voters in more
candidate-centred systems are more likely to believe that elections have
been conducted fairly, and, counter-intuitively, to consider that elected rep-
resentatives are less likely to know what ordinary people think. I n each case
these findings are net of a wide range of other factors, including voter, elec-
toral system and country characteristics. The other three electoral system
characteristics are also consistently important. The greater the dispropor-
tionality of the system, the more negative the voters views of how the
system operates as we would expect. I ndeed, in the case of how fair
| 161 |
TALE 7.4 ATTlTu0E3 T0wAR03 TlE P0LlTlCAL 3Y3TEV
Po||t|oa| part|es neoessary? E|eot|ons oonduoted fa|r|y? HPs know what peop|e th|nk?
1. Necessary 19 1. Fa|r 5 1. KroW 9
2. 2 2. 22 2. 19
3. 1 3. 13 3. 32
1. 5 1. 5 1. 22
5. No| recessary 1 5. urla|r 1 5. 0or'| |roW 18
To|a| 100 100 100
(35 10Z) (32 03) (35 198)
N0TE8 T|e ques||ors Were:
'3ore peop|e say ||a| po||||ca| par||es are recessary |o ra|e our po||||ca| sys|er Wor| |r [cour|ry|. 0||ers |||r| ||a| po||||ca|
par||es are ro| reeded |r [cour|ry|. us|rg ||e sca|e or |||s card, (W|ere 0NE rears ||a| po||||ca| par||es are recessary |o
ra|e our po||||ca| sys|er Wor|, ard FlvE rears ||a| po||||ca| par||es are ro| reeded |r [cour|ry|), W|ere Wou|d you p|ace
yourse|l?'
'lr sore cour|r|es, peop|e oe||eve ||e|r e|ec||ors are corduc|ed la|r|y. lr o||er cour|r|es, peop|e oe||eve ||a| ||e|r e|ec||ors
are corduc|ed urla|r|y. T||r||rg ol ||e |as| e|ec||or |r [cour|ry|, W|ere Wou|d you p|ace || or |||s sca|e ol ore |o lve W|ere
0NE rears ||a| ||e |as| e|ec||or Was corduc|ed la|r|y ard FlvE rears ||a| ||e |as| e|ec||or Was corduc|ed urla|r|y?'
'3ore peop|e say ||a| reroers ol [Corgress/ Par||arer|| |roW W|a| ord|rary peop|e |||r|. 0||ers say ||a| reroers ol
[Corgress/Par||arer|| dor'| |roW ruc| aoou| W|a| ord|rary peop|e |||r|. us|rg ||e sca|e or |||s card, (W|ere 0NE rears
||a| ||e reroers ol [Corgress/Par||arer|| |roW W|a| ord|rary peop|e |||r|, ard FlvE rears ||a| ||e reroers ol [Corgress/
Par||arer|| dor'| |roW ruc| aoou| W|a| ord|rary peop|e |||r|), W|ere Wou|d you p|ace yourse|l?
80UR6E C3E3 rodu|e 1.
elections are perceived to be, the disproportionality of the system is the
important variable in the equation. Larger assemblies also promote more
positive views, as do smaller district sizes again as we would expect.
I t remains to test the hypothesis that the intra-party dimension increases
favourable attitudes towards democracy, net of a wide range of other factors,
including the attitudes that voters hold about the political system itself. This
hypothesis is tested in Table 7.6, which presents the results of a regression
analysis, predicting satisfaction with democracy from a wide range of other
characteristics. The results show that the intra-party dimension does indeed
remain a consistent predictor of satisfaction, net of individual and system-
level characteristics. While the total effect is not large, it remains statistically
significant and is similar in magnitude to disproportonality and assembly
size, or to any of the individual characteristics of voters. The most important
electoral system predictor of satisfaction is district size.
| 162 |
TALE 7.5 PRE0lCTlN0 ATTlTu0E3 T0wAR03 TlE P0LlTlCAL 3Y3TEV
Part|es neoessary E|eot|ons fa|r HPs know
b b beta b beta
lr|ra-par|y d|rers|or .00 .01 .0 -.01 -.01
vo|er c|arac|er|s||cs
Age .00 .05 .01 .09 .00 .00
0erder .02 .03 .03 .05 .01 .02
Ter||ary educa||or .02 .02 .01 .02 .03 .01
urerp|oyed -.02 -.02 -.03 -.02 -.02 -.01
Far||y |rcore .01 .05 .01 .0Z .01 .0
0||er e|ec|ora| sys|er c|arac|er|s||cs
0|spropor||ora|||y -.01 -.10 -.01 -.20 -.01 -.08
Assero|y s|ze .02 .13 .02 .15. .00 .00
vo||rg age pop per VP -.01 -.15 -.01 -.1 -.00 -.03
Cour|ry c|arac|er|s||cs
0erocra||c exper|erce .00 .0Z .00 .1 .00 .02
00P -.00 -.01 .00 .01 .00 .00
Cors|ar| .Z0 .2 .1
Adj R-squared .03 .08 .02
(N) (35 10Z) (32 03) (35 198)
N0TE8 0rd|rary |eas| squares regress|or es||ra|es s|oWr par||a| (o) ard s|ardard|zed (oe|a) coellc|er|s pred|c||rg ||e
prooao||||y ol pos|||ve a||||udes |oWards ||e po||||ca| sys|er. 3ee Tao|e Z.2 lor de|a||s ol var|ao|es ard scor|rg.
Among the other variables, attitudes to the electoral system are consis-
tently important, and the belief that elections are conducted fairly is easily the
most important predictor in the equation. Second in importance is the length
of democratic experience, followed by believing that MPs know what people
think, and taking the view that political parties are necessary in making
democracy work.
Overall, then, the analysis presented in Table 7.6 provides additional
support for the argument that electoral systems do indeed affect levels of
voter satisfaction with democracy, and in this instance we have evidence of
| 163 |
TALE 7. 0EV0CRATlC 3ATl3FACTl0N, ATTlTu0E3 T0wAR03 TlE P0LlTlCAL 3Y3TEV AN0 ELECT0RAL
3Y3TEV ClARACTERl3TlC3
0emoorat|o sat|sfaot|on
b beta
lntre-perty dimension .01 .03
/rr|ruoes ro rne e|eorora| sysrem
Par||es are recessary 1.13 .11
E|ec||ors corduc|ed la|r|y 2.10 .22
VPs |roW W|a| vo|ers |||r| 1.0 .12
Age -.00 -.02
0erder .09 .02
Ter||ary educa||or -.08 -.01
urerp|oyed -.28 -.02
Far||y |rcore .0Z .01
0|spropor||ora|||y .02 .03
Assero|y s|ze -.0Z -.01
vo||rg age pop per VP .01 .09
0erocra||c exper|erce .01 .15
00P -.01 -.02
Cors|ar| 1.95
Adj R-squared .15
(N) (35 105)
N0TE8 0rd|rary |eas| squares regress|or es||ra|es s|oWr par||a| (o) ard s|ardard|zed (oe|a) coellc|er|s pred|c||rg ||e
prooao||||y ol sa||slac||or W||| derocracy. 3ee Tao|e Z.2 lor de|a||s ol var|ao|es ard scor|rg.
the nature of the linkage between these two sets of variables. The results in
Table 7.5 show that the intra-party nature of the electoral system has an
impact on popular perceptions of electoral fairness and, through that, to sat-
isfaction with democracy (as well as having a direct effect in its own right).
While this analysis provides further support for the argument that the macro
measure of systemic proportionality is the most important electoral system
feature, there are also grounds for arguing that the micro aspects of electoral
systems in this case the intra-party dimension affected particularly by the
ballot structure have a role to play, and our analysis in Table 7.6 shows that
this role is far from being insignificant.
Conclusion
With the mushrooming of new democracies over the past few decades, elec-
toral engineering has become big business. On an ever-increasing basis, polit-
ical scientists are invited to comment on the design of new electoral systems;
indeed, this has also included consideration of fundamental reform of elec-
toral systems in established democracies (I taly, Japan, New Zealand, I srael,
the UK, and Canada, for example). While few today would demur from
Richard Katzs contention that the answer to the question which electoral
system is best depends on who you are, where you are, and where you want
to go (1997a: 308), this has not deterred specialists from proffering advice on
where to place greatest emphasis in electoral system design (for a recent illus-
tration, see Taagepera 2002). Among the items that seem to feature in these
deliberations, two are especially prominent: the first is the level of overall pro-
portionality of the system; the second is the nature and degree of linkage
between politicians and voters. Certainly this is seen, by at least some
scholars, as a major factor behind the recent fashion for mixed systems, whose
principal virtue is apparently that they represent the best of both worlds in
terms of these two characteristics (Shugart and Wattenberg 2001).
This chapter started from a similar interest in promoting the joint virtues
of proportionality and good linkage, although in this case (and in contrast to
most scholars, who are inspired by Lijphart) we have placed even greater
emphasis on the latter than the former. Our review of the evidence using
CSES data provides firm support for the view that preferential voting can
make a difference in this instance to levels of voter satisfaction with democ-
racy. Using an intra-party measure that identifies the main characteristics of
preferential system, our comparative analysis has shown that such systems
| 164 |
promote a greater sense of fairness about election outcomes among citizens,
which in turn is a major component of the publics satisfaction with the dem-
ocratic system. Electoral engineers being asked to comment on electoral
system design for the next new democracy (or, for that matter, an established
democracy engaging in a major overhaul of its electoral institutions) might
want to take note.
What do these comparative results imply about the design and operation
of the Australian electoral system? Most importantly, Australia registers one
of the highest levels of satisfaction with democracy in the world. While many
factors feed into that popular evaluation, the electoral system is one major
component of it. The preferential system, the absence of corruption and, not
least, widespread and sustained popular support for the system of compulsory
voting are obviously important elements of this overall assessment. At one
level, this high level of satisfaction with such a complex electoral system,
which displays strong elements of compulsion and party control, may seem
strange. But the system has evolved within a political culture that emphasises
innovation as well as compliance and regulation. I t is the interaction between
all of these elements that generates the levels of population satisfaction that
we see in the results presented in this chapter.
| 165 |
e i g h t
Preferential voting and
electoral system design
This book has been concerned with examining two prominent examples of
preferential electoral systems, AV and STV, with a particular focus on
Australia. Additionally, it has placed the Australian experience in a compar-
ative perspective. We have considered the origins and evolution of the
Australian systems, how they vary at both national and state or territory level,
and their electoral and behavioural consequences. Our examination of these
two systems and, more particularly, our consideration of how they are
embedded in the Australian context, have facilitated a close and detailed
scrutiny of a range of their features. Consistent with the over-arching theme
of the book, we can review our principal findings in terms of both the com-
parative debates over electoral systems and the specific themes relating to
preferential systems. I n the former case, our focus on AV and STV provides
insights into our comparative understandings of electoral systems. I n the
latter, the Australian focus provides important lessons for our understanding
of the place of preferential systems.
I n this concluding chapter, we draw together the main findings of our
study and conclude with a discussion about the current state of Australian
electoral system design and the potential for future reform. Our review of our
main findings covers three themes. We start, in the first section, with the pol-
itics of electoral system design. The second section revisits the issue of varia-
tions within particular electoral system families. I n the third section, we
review the principal electoral and behavioural consequences of the AV and
| 166 |
STV systems. Next, in the fourth section, we address the question of electoral
reform, or more accurately its absence from Australian political debate over
the past twenty years. The chapter concludes with consideration of some
potential areas for future electoral reform.
The politics of electoral system design
This book provides interesting insights into a (recently) developing research
agenda on the origins, design and evolution of electoral systems. The
Australian experience supports the contention that institutional design is
contingent on a range of factors peculiar to the country in question, in this
case a combination of colonial heritage, pressures from state-level experi-
ments and the prominent role of key actors. But in seeking to explain why
Australia has evolved as one of only a tiny handful of countries actually using
AV or STV for national (or even state or territory) level elections, we need to
delve more deeply, because the very rarity of such systems poses the question:
why Australia? I n many of the recent debates over electoral reform in devel-
oped democracies (for example, New Zealand or the UK), STV is a low pri-
ority on the list of systems that are preferred by the political elite.
1
The main
reason seems to be a fear on the part of the political elite that they will lose
control; for them, a system like STV places too many decision in the hands
of voters. As for AV, apart from some small scale and generally unsuccessful
experimentation in the Oceania region, it simply does not feature as a serious
option.
Given this background, it does seem curious that Australia adopted AV
for House elections in 1918 and STV for Senate elections in 1948. While
short-term political expediency has played a role, there is good reason for
supporting the revisionist view that this was essentially a final chapter in the
pursuit of a long-term goal for a particular form of institutional design
(although 1983 shows it was not really the final chapter). As we suggest in
chapter 2, there is a sense in which elite opinion froze around the choice
between two main types of electoral systems (plurality versus preferential)
that predominated in debates at the turn of the twentieth century. This was
very much a debate centred on elite circles in the UK, where there was little
support for list PR systems. I ndeed, in the same way that mixed member elec-
toral systems are fashionable in the new democracies at the turn of this
century, in the debates at the turn of the last century in first-wave democ-
racies with an Anglo-Saxon heritage preferential systems were derigueur.
| 167 |
| Preferenti al voti ng and el ectoral system desi gn |
This interpretation that Australian elite opinion settled on preferential
systems as the only viable alternative to plurality systems is supported by
our examination in chapter 3 of other electoral systems that have been the
subject of experiments at state and territory level. I t is clear throughout that
there is a penchant for preferentialism, and this reveals itself in its starkest
form in the excessively complex variant of list PR used briefly in the ACT, a
variant that consisted of a blending of list and preferential electoral systems.
Although the system was quickly discarded after its complexity became clear,
it provides ample evidence of the extent to which preferentialism dominates
elite thinking about electoral system design.
A further feature of electoral system design that has been prominent in
the Australian case is the role of electoral systems experts. This suggests some
revision of the view, expressed classically by K. C. Wheare in the 1960s, that
politicians manifest an inability to understand or listen to the explanations of
electoral experts (cited in Goot 1985: 180). The Australian experience indi-
cates that the politicians do listen although whether they fully understand
is a moot point. I n the parliamentary debates of 1902, Edward Nanson and
Catherine Helen Spence (and, to a degree, the Ashworth brothers and, indi-
rectly, I nglis Clark) were very significant players. Nanson designed the prin-
cipal features of the systems being proposed; Spence sought to alter some of
those same features. Similarly in 1983, the introduction of ticket voting and
the inclusive Gregory method for Senate elections were influenced by experts
advising the Joint Select Committee on Electoral Reform ( JSCER, the pre-
cursor to todays JSCEM).
I n all these changes, it is interesting to speculate about just how much the
political masters at the time fully understood the implications of the specific
types of electoral system being proposed. As we saw in chapter 2, the form of
STV that Nanson promoted in 1902 was unique in a number of respects, and
while this might have allayed some social choice concerns about this system,
if it had been introduced it is likely that its peculiar features (for example
regarding bracketed preferences) would have surprised many politicians on its
first outing. I n 1983, the full implications of the inclusive Gregory method
seems to have passed many legislators eyes unnoticed. These two examples
suggest that, in studies of electoral system design, greater attention should be
paid than previously to the role of electoral system specialists.
Another facet of Australian electoral system design that has implications
for the comparative study of electoral systems is the penchant for change.
2
The long-held assumption in the electoral systems literature has been that
| 168 |
electoral systems are stable political institutions (e.g. Taagepera and Shugart
1989). Events in recent years (notably in New Zealand, I taly, I srael, Japan,
Venezuela, the British regions and the Canadian provinces) has caused some
revision of this perspective, in the sense that electoral reform is seen now as
more likely than it used to be. Our review of Australian trends (over the
course of the previous century) suggests that even more revision of this per-
spective is necessary. Contrary to received wisdom, Australia has provided a
prominent example of how electoral systems do change, and in this instance
were subject to regular and major change. Furthermore, the Australian case
points to the importance of taking account of temporal aspects in the study
of electoral systems (as stressed, for instance, by Blau 2003).
The contrary view is, of course, that for the most part the changes to the
electoral system were very small (although certainly not so in 1918 and 1948).
But even if this is so, the Australian case does suggest a need to take greater
account of the true extent of minor changes. Taken together over time, they
can amount to a significant trend (for example, rules regarding the minimum
number of preferences, the move towards grouping of candidates on the
ballot paper and, ultimately, the adoption of ticket voting). And, on occasions
(following Coram 1996), even seemingly minor changes can produce quite
significant results for example, the adaptation to the Gregory method for
transferring vote surpluses in 1983, which, as we saw in chapter 4, can mate-
rially affect the electoral fate of candidates.
Variations in electoral system families
I n chapter 3 we reviewed the main variations in the different forms of AV and
STV systems, building on the work of Ben Reilly (2001) on AV, and also on
the findings of the edited collection on STV produced by Shaun Bowler and
Bernard Grofman (2000). Both electoral systems are prone to variation, par-
ticularly STV, and the latter finding lends strong support to Tidemans (1995)
contention that it is more accurately described as an evolving family of vote-
counting rules rather than a single rule. Given the extent of variation that is
possible in these systems, questions can be asked about some of the compar-
ative treatments of them in the literature, in which conclusions are often
drawn on the basis of a single countrys experience Australian national AV
elections and Irish STV elections. I f, as we have shown, the systems can vary,
with implications for their wider effects on the polity, then the standard
descriptions of these systems may need revision. And certainly, in the light of
| 169 |
this, Bernard Grofmans proposal that electoral systems be treated as nation-
ally embedded institutions finds strong support in this study.
Contrary to the general picture that preferential systems present voters
with greater choice than other electoral systems, the Australian case provides
an important example of how party politicians can develop methods for con-
trolling this. More than anything else, it shows how effective political parties
can be in adapting not only themselves, but also the political institutions they
work with, to ensure their survival. I n the party politics literature there is a
prominent debate about the ongoing adaptation of parties to the more com-
petitive electoral market. According to Katz and Mair (1995), this manifests
itself in the development of party cartels, in which the parties cushion them-
selves against electoral change by seeking additional state resources and also
by introducing institutional restrictions to limit the rise of competitors. I n
similar fashion, the efforts of Australian politicians in electoral system design
have, over time, tended to favour party political control over the features of
voter choice that are supposed to characterise these preferential electoral
systems.
| 170 |
Candidate-oriented
V
o
t
e
r

c
o
m
p
u
l
s
i
o
n
Party-oriented
V
o
t
e
r

c
h
o
i
c
e
STV
AV
?
?
FIGURE 8.1 TEMPERING THE VOTER CHOICE CANDIDATE-ORIENTATION OF PREFERENTIAL
SYSTEMS: ELECTORAL SYSTEM DESIGN IN AUSTRALIA
The consequences for the electoral systems have been very significant, as
Figure 8.1 suggests. The requirement that voters complete all preferences in
Australian AV elections, and the how to vote cards by which the politicians
manipulate those preferences, have resulted in a shift in the AV system away
from its voter choice roots towards a scenario of voter compulsion one that
is buttressed by laws requiring compulsory registration and attendance at the
polls. The use of how to vote cards which, as we saw, are used by about half
of all voters to guide their preferences also suggests a shift in emphasis
within the electoral system away from a candidate orientation towards a more
party-oriented system.
Matters are compounded in the case of Australian STV elections by the
combination of a more complex electoral system and (largely in response to
this) the adoption of ticket voting in 1983. As a consequence, STV
Australian-style (at least outside Tasmania and the ACT) is characterised not
by real voter choice but by high degrees of voter compulsion and strong party
control over the voting process. The adaptation here is so profound that it
calls into serious question the degree to which Australian Senate-style STV
can really be classified as STV as generally understood in the electoral
systems literature.
Electoral and behavioural consequences
Our examination of electoral system consequences (chapters 46) produced
three sets of implications for the study of electoral systems. These relate to the
macro-systemic focus on election outcomes, the recently more prominent
interest in micro-behavioural consequences of electoral systems and to efforts
at categorising electoral systems.
Traditionally, much of the interest of electoral systems research has been
on election outcomes. The general presumption is that AV, as a non-propor-
tional system, can be slotted in with the other non-proportional systems,
while STV, as a system whose complexities necessarily require a small district
magnitude, is probably best located on the lower end of the proportionality
scale, and may even be classified as quasi-proportional. This study provides
correctives to such presumptions. First, while AV may produce dispropor-
tional electoral results, this is not the same as saying that the results produced
by AV are equivalent to what would be produced by other non-proportional
systems. I ts preferential nature has consequences, and there is evidence that
this is of increasing significance in recent elections. As for STV, contrary to
| 171 |
expectations, there are instances of STV elections with large district magni-
tudes (large constituencies, or electorates), and by extension this means that
STV is fully capable of producing low levels of disproportionality comparable
with that found in list PR elections.
As to the electoral fate of individual candidates, in chapter 4 we produced
evidence in support of the contention of social choice critics that AV and
STV are capable of producing quasi-chaotic results. We saw this in two
respects. First, we saw how an alteration in the rules regarding the numbers
of preferences available to voters in AV elections may produce non-monot-
onic results; in the example given in chapter 4, this related to whether Pauline
Hanson might have won the seat of Blair in 1998 if there had been a reduc-
tion in the number of preferences available to voters. Second, we saw how an
apparently tiny change in the procedures for transferring vote surpluses in
STV election counts (the various Gregories examined in chapter 4) can
affect a candidates electoral chances.
This study demonstrates how electoral systems can have profound effects,
not only on who gets elected, but also with regard to the nature of represen-
tative politics. The electoral system influences the types of individuals who
are recruited to stand for election, the individuals who gain election, the
means by which they seek election and how they envisage their representative
role. While much of this can be further mediated by other features of a
countrys electoral institutions (relating, for instance, to built-in compulsory
features and the role of parties), the fact remains that having a preferential
electoral system has implications for the nature and degree of voter emphasis
by parliamentary candidates and elected parliamentarians.
Similarly, there are behavioural consequences for the voters, not only with
regard to voter confusion (and/or exhaustion) but also with regard to the
potential for strategic voting. Our first finding here is that complex electoral
systems do not necessarily confuse voters a finding that was also demon-
strated in Ireland in 1918 and again in Northern Ireland in 1973, when STV
was used for the first time. I nvalid voting in Australia at least in recent elec-
tions is lower than might be expected, particularly given the existence of
compulsory voting, the large number of non-English speaking immigrants,
the frequency of elections and the complex and different systems that exist at
state and federal level. Moreover, there is little sign of voter alienation with
the system. Secondly, preferential systems like AV and STV suggest that
standard conceptualisations of strategic voting as insincere voting may
require revision, and we argue that more account should be taken of
| 172 |
preferential electoral systems by scholars working in this area. Furthermore,
and more specifically in the Australian context, while the Gibbard-
Satterthwaite theorem (that all non-dictatorial electoral systems are prone to
strategic voting) may well be correct, account needs to be taken of the ability
of politicians to influence the process.
This discussion serves to buttress the remarks of other scholars in this area
about the need for electoral system researchers to once and for all break free
from the fixation with analysing electoral systems principally in terms of their
macro-systemic consequences, which produces a focus on such features of
electoral systems as their district magnitudes or electoral formulas. Like
Bowler and Grofman (2000b), we suggest that this tends to produce a dis-
torted hierarchy of electoral systems in which the proportional family seen
primarily as including list and certain mixed-member systems, and, by some
scholars, possibly also STV stand out as an ideal. I n this book we have
placed greatest emphasis on ballot structure as a factor in electoral system
design. On that basis (and, as stressed, particularly in the comparative
overview provided in chapter 7), we find that this can result in quite a dif-
ferent hierarchy of electoral systems, in which AV and STV can be distin-
guished from most of the other major electoral systems currently in use for
national-level elections.
Preferential voting has an important role in our understanding of the
implications of electoral system design. The emphasis such systems place on
voter choice directs the interests of scholars away from their longstanding fix-
ation on the macro-systemic consequences of electoral system design and
towards the micro-behavioural effects of that design. Similarly, these systems
provide fresh insights into the effects of ballot structure design, a feature of
electoral systems that, on the whole, tends to be downplayed in comparative
study.
I f the study of preferential electoral systems adds significant new findings
to the literature on electoral system design and electoral system consequences,
it is also evident that a close examination of such systems as embedded insti-
tutions adds greatly to our appreciation of exactly how these systems work,
how they vary and how their consequences can be tempered by other features
of institutional design. Australia has, not for the first time, offered important
new insights into our understanding of institutional design, and as a conse-
quence, its status as a democratic laboratory is assured. I t is, however, a status
that in recent years is starting to come into question as we discuss next.
| 173 |
From innovation to stagnation:
electoral system design
I nstitutional evolution has been one of the prominent themes of this book.
As we have seen, over the course of the last century Australia had a distin-
guished record of innovation and experimentation in electoral system design,
blazing a trail of new electoral systems and establishing international best
practice in electoral procedures. And while there is some evidence at sub-
national level of a continuing tendency to tweak and change electoral systems
as deemed necessary (for instance, the adaptation of ticket-voting procedures
in New South Wales in 2002 or the ongoing discussions over surplus transfer
procedures in Western Australia), the same cannot be said about trends at
national level. I ndeed, and somewhat ironically, the lack of discussion today
over Australian electoral system reform is the reverse of international trends,
where electoral reform is increasingly on the agenda.
Twenty years ago, the eminent electoral systems scholar, Dieter Nohlen,
famously observed that fundamental electoral system reform was rare, occur-
ring only in extraordinary historical situations (1984: 218). Since then, the
growing list of established democracies that have undergone large scale elec-
toral reform, or are discussing it (Colomer 2004; Gallagher and Mitchell
2005), necessitates some revision of Nohlens dictum. The interesting cases
are now those that appear to have avoided any serious debates over electoral
reform such as Australia. Where once Australia used to be the electoral
system innovator, the electoral reformer par excellence, today this accolade is
increasingly difficult to maintain. Australia now stands as a prominent and
increasingly rare example of a country that is not giving serious consideration
to electoral system reform. The signals from the party political elite appear to
be that there is no need for further change, that the current state of the elec-
toral system is fine. I f this is what the political elite truly believe, then they
are being overly sanguine. Resting on your laurels is never a good thing;
indeed, as Mercurio and Williams comment, in the light of international
trends in electoral system reform, it is something of a conceit for Australia to
continue to claim to be a leading innovator in electoral law and practice
(2004: 21; also Orr et al. 2003).
I t is not as if there have not been calls for electoral reform, whether from
such obvious quarters as the Proportional Representation Society of
Australia, the smaller political parties, or prominent academic research for
example, the regular reports and bulletins produced by the Democratic Audit
| 174 |
of Australia (http://democratic.audit.anu.edu.au; see also Orr et al. 2003).
Many of the themes raised by these calls are beyond the remit of this book;
such themes include campaign finance reform, the regulation of political
advertising and the malapportionment resulting either from rural weighting
in certain parts of Australia or from rules regarding state and territory repre-
sentation in the federal houses. We do not deny that these are important
matters, but they are outside the scope of this book, whose focus has been
specifically on Australias voting rules.
There are some themes relating to the electoral system that are, we would
argue, not specific to Australia, and therefore also not of direct relevance to this
study. A good example of this relates to the rights of voters with special needs,
such as print-handicapped voters, who at present are denied a secret ballot.
Clearly it would be relatively simple to adapt existing procedures to resolve this
problem: for example, ballot papers could include braille text or colour photo-
graphs, or computerised voting with voice prompts could be introduced. We
would endorse arguments favouring such moves (e.g. Mercurio and Williams
2004). In passing, it is worth speculating about how such an issue would have
been dealt with by the political elite at the turn of the last century, had it been
a concern then; yet again, the absence of serious moves to resolve something
that is so simple to resolve underscores the degree of complacency about elec-
toral matters among the current members of the political establishment.
This complacency is ill judged, because Australias electoral systems, at
least at national level, are long overdue for reform both to address the needs
of voters and to resolve blatant inequities. I n the first instance, too much is
being demanded of Australias voters, and even if they do not seem to mind
(as we saw in chapter 6), there is nevertheless scope for resentment and alien-
ation to fester. Australias voters are required by law to register to vote and to
turn out on polling day. I f they wish to make a valid vote, they must complete
all the preferences on the ballot paper or take the shortcut of ticket voting in
the Senate election (see the first entry on Table 8.1). Even if technically they
are permitted to leave the final preference unmarked so that they do not feel
forced to vote for a candidate they find reprehensible, the reality is that this
is counted as if they had completed their final preference. I n no other elec-
toral system are voters required to express a preference for candidates they do
not support. The problem is compounded by the operation of Senate ticket
voting, where by default the vast bulk of Australias voters sign away their
right to a secret ballot to the party leaders who carve up ticket deals between
themselves. Once the voter has expressed a ticket vote for a party, all her
| 175 |
remaining preferences are decided for her by the party; her only alternative is
to vote below the line, but even here (as for AV elections) there is no escape
from the requirement to vote for all the candidates whether she supports
certain candidates or not.
An additional problem for Australias voters is the potential confusion
resulting from significant variations in voting rules, both at national level and
also between the national and state/territory systems. At the national level,
the voters are required to vote in two very different ways for the two houses:
they must rank-order all candidates on the AV ballot paper for the House of
Representatives, whereas assuming they opt to vote above the line, as well
over nine in ten voters do (see Figure 6.2) on the STV ballot for Senate
elections they merely express a single preference for one party. No wonder
more than half a million voters had their votes declared invalid in the 2004
Commonwealth election. A significantly a greater proportion (5.2 percent)
spoiled their vote in the easier AV election for the lower house than did for
the Senate race (3.8 percent) this trend reflecting the fact that most Senate
voters vote above the line.
| 176 |
TALE 8.1 TlE ELECT0RAL 3Y3TEV TERRAlN FAClN0 Au3TRALlA'3 v0TER3
Jur|sd|ot|on Lower house e|eotora| system Upper house e|eotora| system
CorrorWea||| /v u|rn oompu|sory express|on ol a||
prelerenoes
$Tv u|rn oompu|sory express|on ol a||
prelerenoes or aoove rne ||ne r|oker vor|ng
NeW 3ou||
wa|es
Av W||| op||ora| prelerer||a| vo||rg 3Tv W||| corpu|sory express|or ol 15
prelererces or aoove ||e ||re op||ora| rar|-
order|rg ol par||es
v|c|or|a /v u|rn oompu|sory express|on ol a||
prelerenoes
3Tv W||| corpu|sory express|or ol 5
prelererces or aoove ||e ||re ||c|e| vo||rg
0ueers|ard Av W||| op||ora| prelerer||a| vo||rg r.a.
3ou|| Aus|ra||a Av W||| corpu|sory express|or ol a||
prelererces
Tasrar|a 3Tv W||| corpu|sory express|or ol 5
prelererces (|rc|ud|rg Roosor Ro|a||or)
Av W||| corpu|sory express|or ol 3
wes|err
Aus|ra||a
/v u|rn oompu|sory express|on ol a||
prelerenoes
Nor||err
Terr||ory
/v u|rn oompu|sory express|on ol a||
prelerenoes
r.a.
ACT 3Tv W||| corpu|sory express|or ol 5 or Z
r.a.
Jur|sd|c||ors s|ar|rg ||e sare e|ec|ora| sys|ers as ||e CorrorWea||| sys|ers are |rd|ca|ed oy ||a||cs
80UR6E8 Tao|es 3.1, 3.3, 3.1
The potential for voter confusion is increased by the range of variations in
AV and STV in use across Australias various jurisdictions, as was outlined in
some detail in chapter 3. Table 8.1 summarises the electoral terrain facing
Australias voters: except in South Australia, Western Australia and the uni-
cameral Northern Territory, electoral systems in the states and territories
impose requirements that are different from those under the national system.
There are three principal variations: STV rather than AV for lower house
elections in Tasmania and the ACT (and AV rather than STV for the
Tasmanian upper house election); different rules about the compulsory
expression of preferences across the remaining jurisdictions; and Robson
rotation of candidate name order on the ballot paper in Tasmania and the
ACT.
Quite apart from the potential alienation and/or confusion of Australian
voters, there are other features of the current Australian electoral systems,
particularly the STV system for the Australian Senate, that raise concerns
over political equity. The most obvious of these is the operation of ticket
voting. As the New South Wales tablecloth election of 1999 demonstrated
(see p. 66), this effectively changes the STV system into an extreme version
of closed list PR (extreme because no other list system provides party leaders
with so much control over voter preferences). I n addition, it opens up the
possibility for game playing by the parties, funnelling preferences through
paper parties hardly matching the ideals of a properly constituted electoral
process. For that matter, the practice of ticket deals can result in certain
parties being effectively frozen out by the other parties, as has happened to
One Nation; and while, on occasions, that might be for laudable enough
reasons, it nevertheless is worrying that an electoral system can facilitate such
a practice.
There is one other feature of the ticket voting system that is a cause for
concern: this relates to the treatment of independent candidates who (unless
they happen to be incumbents) are listed only below the line, where they are
grouped together at the far right of the ballot paper. This provides no oppor-
tunity for the voter to express a first preference for them, short of opting for
the more arduous task of voting below the line.
I n existing Australian electoral practice, ticket voting is clearly the most
significant feature in need of fixing. But there are other features that could be
addressed, among them the order of candidate names (and, indeed, party
names) on ballot papers and certain details in the counting procedures in
Senate-style STV. These are two useful examples of problems that are
| 177 |
relatively easy to fix. The first of these relates to the practice of fixing the
order in which candidates and parties appear on ballot papers, which privi-
leges those listed high (or to the left) on ballot papers due to alphabetical or
donkey voting. I n Tasmania and the ACT the adoption of Robson rotation
(see p. 62) has prevented this problem (with the added attraction that the
parties cannot, even if they were allowed, foist how to vote cards on hapless
voters in these jurisdictions). The second problem is the use of the inclusive
Gregory method (see pp. 6264) to determine the transfer of surplus votes in
elections under the Senate-style STV system; this method can result in sig-
nificant electoral anomalies. A switch to the unmodified Gregory method, as
used in Tasmania, or to weighted inclusive Gregory (as, at the time of writing,
is being considered for Western Australia) would solve this problem.
Where to now?
The current electoral systems have served Australian voters and policy-
makers well, but they are now in need of modification. Twenty years after the
last major electoral reform, over a hundred years after the establishment of
the Commonwealth, in a world in which institutional reform and questions
of good governance are much higher on the political agenda than ever before,
it is hard to imagine a more suitable time to be raising the question of change.
But change in which sense, and in which specific areas? Our start and finish
point is that Australia has designed for itself two systems that have consider-
able strengths, and that bear favourable comparison with any other systems
that might be offered as alternatives: in other words, we do not favour funda-
mental electoral reform to entirely new electoral systems. Nevertheless, some
change is required, and we identify three areas, the last of which has signifi-
cant implications. All three changes have in common one fundamental ideal,
namely to achieve consistency in electoral system design across all Australian
jurisdictions.
The first two proposals can be simply stated without any further need for
explanation or elaboration. First, Robson rotation of candidate (and party)
names should be introduced for all elections at all levels; second, the inclusive
Gregory method of transferring surplus votes in Senate STV counts should
be replaced either by Gregory or by weighted inclusive Gregory. Our third
proposal is the key reform that would do so much to return Australian elec-
tions to the original ideals of preferential electoral systems, inspired by voter
| 178 |
choice. I n principle, it requires only a very small change to existing practice.
The proposed change is that the compulsory expression of preferences should
be replaced by optional preferential voting, as used for AV elections in New
South Wales and Queensland and for STV elections in Ireland. I f this move
is considered too bold, a possible compromise would be to require some
minimum number of preferences of voters, such as for Tasmanian and ACT
STV and Tasmanian AV, where voters must express as many preferences as
there are seats to be filled. Without the compulsory expression of preferences
there would be no need for above-the-line ticket voting in STV elections, nor
would voters have to rely so heavily on guidance from the parties how to vote
cards in AV elections. The degree of party control over voters preferences
would be considerably reduced, particularly for the Senate system, enabling
that system to once again be treated as a proper member of the STV family.
I nevitably, full or partial optional expression of preferences is not without
potential risks, particularly in the case of the AV system, where Just Vote 1
strategies (as practiced, for example, by the ALP in Queensland in 1998 and
2001) could so weaken the value of preferences as to result by default in single
member plurality elections. But this need not be the case, and other jurisdic-
tions that allow optional expression of preferences (for example Ireland) have
not witnessed such trends.
Twenty years ago Australia stood apart from other nations as the country
that showed greatest willingness to experiment and innovate in electoral
system design. At a time when the tendency was to leave well alone,
Australian policy makers allowed the countrys electoral systems to evolve to
meet changing needs but no more. Today Australia stands apart once again,
but now as one of a dwindling number of countries manifesting little willing-
ness to embrace further electoral reform. I t would be a pity if such a situation
was allowed to continue.
| 179 |
Appendix
| 180 |
TALE A.1 P03T-wAR 3Tv 3TATl3TlC3: Au3TRALlA (NATl0NAL, 3TATE AN0 TERRlT0RY), lRELAN0 AN0
VALTA
8ystem E|eot|on H A8 T
eff
S| N
v
N
s
ACT 1995 5.Z 1Z 11.2 5.2 3.5 3.18
ACT 1998 5.Z 1Z 11.2 8.Z3 1.13 3.18
ACT 2001 5.Z 1Z 11.2 8.51 3.15 2.51
ACT 2001 5.Z 1Z 11.2 Z.1Z 2.85 2.21
wA 1989 5.Z 31 11.2 11.59 2.9 2.33
wA 1993 5.Z 31 11.2 Z.5 2.91 2.52
wA 199 5.Z 31 11.2 1.21 3 2.59
wA 2001 5.Z 31 11.2 5.8 3. 3.32
wA 2005 5.Z 31 11.2 .3Z 3 2.38
3A 1982 11 11 .2 1.53 2.19 2.3Z
3A 1985 11 11 .2 .02 2.5Z 2.3Z
3A 1989 11 11 .2 5.91 2.91 2.3Z
3A 1993 11 11 .2 Z.0 2.85 2.28
3A 199Z 11 11 .2 .9 3.Z 3.2Z
3A 2002 11 11 .2 .3 3. 2.81
N3w 19Z8 15 15 1. 5.11 2.3 1.92
N3w 1981 15 15 1. 2.82 2.55 2.1Z
N3w 1981 15 15 1. 3. 2.1 2.2Z
N3w 1988 15 15 1. 1.15 2.Z9 2.59
N3w 1991 15 15 1. 1.0 2.81 2.59
N3w 1995 21 21 3.1 1.59 3. 3.32
N3w 1999 21 21 3.1 5.9 1.19 1.12
N3w 2003 21 31 3.1 1.18 3.22 2.85
Tas 191 30 10.Z 1.9 2.2 2.25
Tas 1918 30 10.Z 2.12 2.5 2.11
Tas 1950 30 10.Z 1. 2.1 2.11
Tas 1955 30 10.Z 3.Z8 2.0Z 2
Tas 195 30 10.Z 5.15 2.25 2
Tas 1959 Z 35 9.3 .12 2.Z1 2.25
Tas 191 Z 35 9.3 .91 2.11 1.99
Tas 199 Z 35 9.3 3.Z 2.3 2.12
Tas 19Z2 Z 35 9.3 1.Z 2.22 1.92
Tas 19Z Z 35 9.3 3.3 2.11 2
Tas 19Z9 Z 35 9.3 3.05 2.11 1.9
Tas 1982 Z 35 9.3 5.31 2. 2.2
| 181 |
| Appendi x |
eff
S| N
v
N
s
Tas 198 Z 35 9.3 3.93 2.38 2.2
Tas 1989 Z 35 9.3 2.Z 2.88 2.51
Tas 1992 Z 35 9.3 2.Z3 2.51 2.12
Tas 199 Z 35 9.3 3.Z8 2.89 2.2
Tas 1998 5 25 12.5 9.8 2.Z9 2.1
Tas 2002 5 25 12.5 3.35 2.5 2.39
Aus 1919 Z 12 9.3 3.12 2.19 1.98
Aus 1953 5.33 32 11.8 3.29 2.2 1.99
Aus 1955 5 30 12.5 .52 2.1 2.0Z
Aus 1958 5.33 32 11.8 .18 2.53 2.12
Aus 191 5.1Z 31 12.1 9.Z5 2.59 2.13
Aus 191 5 30 12.5 2.0 2.11 2.2Z
Aus 19Z 5 30 12.5 3.8 2.53 2.11
Aus 19Z0 5.33 32 11.8 3.1 2.9Z 2.Z1
Aus 19ZZ 1.25 31 11.2 Z.3 2.81 2.21
Aus 1980 1.25 31 11.2 1.51 2.5 2.52
Aus 1981 5.Z5 1 11.1 5.35 2.89 2.5
Aus 1990 5 10 12.5 1.39 2.95 2.1
Aus 1993 5 10 12.5 3.33 2.5 2.15
Aus 199 5 10 12.5 1.51 2.98 2.5Z
Aus 1998 5 10 12.5 .9 3.39 2.81
Aus 2001 5 10 12.5 .33 3.33 2.ZZ
Aus 2001 5 10 12.5 Z.2 1.39 2.28
Aus00 1951 10 0 .8 3.03 2.18 1.99
Aus00 19Z1 10 0 .8 3.Z2 2.39 2.11
Aus00 19Z5 8 1 8.3 3.08 2.3 2.1
Aus00 1983 8 1 8.3 3.3Z 2. 2.1
Aus00 198Z 9.5 Z Z.1 2. 2.Z3 2.59
lRL 1918 3.8 11 1 5.52 1.11 3.
lRL 1951 3.8 11 1 2.12 3.38 3.2Z
lRL 1951 3.8 11 1 1.9Z 3.25 3.01
lRL 195Z 3.8 11 1 1.31 3.1 2.Z2
lRL 191 3.Z9 113 15. 1.38 3.23 2.Z8
lRL 195 3.Z9 113 15. 2.11 2.Z2 2.1
lRL 199 3.13 113 1.9 5.3 2.83 2.1
lRL 19Z3 3.13 113 1.9 2.2Z 2.81 2.59
lRL 19ZZ 3.52 11Z 1.5 1.8Z 2.Z5 2.3
| 182 |
eff
S| N
v
N
s
lRL 1981 1.02 15 11.9 2. 2.8Z 2.2
lRL 1982 1.02 15 11.9 1.3 2.9 2.53
lRL 1982 1.02 15 11.9 2.Z2 2.Z2 2.52
lRL 198Z 1.02 15 11.9 5.12 3.1Z 2.89
lRL 1989 1.02 15 11.9 3.5 3.38 2.9Z
lRL 1992 1.02 15 11.9 3.85 3.89 3.1
lRL 199Z 1.05 1 11.8 .93 3.95 2.99
lRL 2002 1.05 1 11.8 .1 1.11 3.1Z
Va| 191Z 5 10 12.5 3.0 2.12 2.15
Va| 1950 5 10 12.5 3.9Z 1.22 3.9Z
Va| 1951 5 10 12.5 2.29 3.39 3.29
Va| 1953 5 10 12.5 .8Z 2.Z8 2.31
Va| 1955 5 10 12.5 2.Z1 2.0 1.9
Va| 192 5 50 12.5 .29 3.22 2.Z1
Va| 19 5 50 12.5 Z.29 2.39 1.9Z
Va| 19Z1 5.5 55 11.5 1.0 2.01 2
Va| 19Z 5 5 12.5 0.ZZ 2 2
Va| 1981 5 5 12.5 3.21 2 2
N0TE8 we lo||oW L|jp|ar|'s corver||or ||a| a|| |r|egers |rd|ca|e exac| ard urc|arg|rg ruroers, a|| o||er ruroers |rd|ca|e
averages: oecause T
ell
|s or|y ar approx|ra|e va|ue, || |s recorded |o jus| ore dec|ra| p|ace.
V = 0|s|r|c| ragr||ude.
A3 = Assero|y s|ze, |.e. ||e ruroer ol |eg|s|a|ors oe|rg e|ec|ed |r a par||cu|ar e|ec||or (|r ||e case ol s|aggered e|ec||ors,
suc| as N3w or 3A, |||s reasure does ro| |a|e accour| ol ||e ac|ua| assero|y s|ze).
T
ell
= Ellec||ve ||res|o|d.
0l = 0a||ag|er |rdex ol d|spropor||ora|||y.
N
v
= Laa|so ard Taagepera |rdex ol ||e ellec||ve ruroer ol e|ec|ora| par||es.
N
s
= Laa|so ard Taagepera |rdex ol ||e ellec||ve ruroer ol par||arer|ary par||es.
80UR6E8 Re|evar| e|ec|ora| corr|ss|or Weos||es.
| 183 |
| Appendi x |
TALE A.2 ELECT0RAL 3Y3TEV3 0ATA F0R TlE C3E3 C0uNTRlE3
|ntra-party d|mens|on ood|ng 0ther e|eotora| system var|ab|es
a||ot Vote 0|str|ot |ndex 0|spropor-
t|ona||ty
a
Assemb|y
s|ze
VAP per HP
Aus|ra||a (199) 1 5 2 8. 10.2 118 91500
8e|g|ur (1999) 2 2 1 2.9 3.12 150 53300
8u|gar|a 1 1 1 1.1 Z.09 210 232
Carada (199Z) 1 3 2 1.3 15.ZZ 301 Z8300
Czec| Repuo||c (199) 2 2 1 2.9 1.11 200 39300
0errar| (1998) 2 1 3 Z.1 1.85 1Z9 23000
Frarce 3 3 2 5.Z 21.3Z 5ZZ ZZ10
0errary (1998) - - - 3. 3.38 5 100000
lurgary (1998) - - - 1.Z Z.5 38 20000
lce|ard (1999) 1 1 1 1.1 2.13 3 3120
lre|ard 1 5 3 10.0 5.39 1 1155
lsrae| (199) 1 1 1 1.1 2.50 120 30Z00
Japar (199) - - - 3. 10.Z8 500 193100
Korea (2000) - - - 3. 8.81 299 111900
L|||uar|a (199Z) - - - 1.Z Z.18 111 2100
Vex|co (2000) - - - 3. Z.08 500 110800
Ne||er|ards (1998) 2 2 1 2.9 1.58 150 80000
NeW Zea|ard (199) - - - 3. 1.3 120 21100
NorWay (199Z) 1 1 1 1.1 3.92 15 20000
Po|ard (199Z) 2 2 1 2.9 .3 10 0Z00
Por|uga| (2002) 1 1 1 1.1 5.0Z 230 3819
Rorar|a (199) 1 1 1 1.1 8.50 313 18800
3|over|a (199) 2 2 1 2.9 1.19 90 1Z000
3pa|r (2000) 1 1 1 1.1 .3 350 8800
3Weder (1998) 2 2 1 2.9 2.13 319 19800
3W||zer|ard (1999) 2 1 3 Z.1 3.2 200 28Z00
Ta|War (199) - - - 3. 3.11 331 12900
uK (199Z) 1 3 2 1.3 15.19 59 8100
u3A (199) 3 3 2 5.Z 1.1Z 135 13Z00
a. 0a||ag|er |rdex. For ros| 0EC0 cases, We prov|de decerr|a| averages lor ||e 1990s, |r a|| o||er cases ||e ros| recer|
e|ec||or. (For ||e assero|y s|ze ard vo||rg age popu|a||or da|a, We repor| ||e |rerds lor ||e ros| recer| e|ec||or.)
80UR6E8 8oW|er e| a|. 2003, Farre|| 2001, Norr|s 2002, Rose 2000, |||p://WWW.essex.ac.u|/e|ec||ors/, WWW.|dea.|r|,
WWW.e|ec||orWor|d.org, |||p://psep|os.adar-carr.re|.
Notes
Chapter 1
1 Known as preferential voting in Australia, in the USA it is sometimes referred to as
instant run-off voting.
2 Also known as the Hare or Hare-Clark system, in the USA it is sometimes referred to
as choice voting.
3 The most commonly used measure of disproportionality is known as the Gallagher
index, and this is derived as follows: GI =
1/2 (v
i
s
i
)
2
(where v = votes and s =
seats). The standard measure of party numbers is known as the Laakso and Taagepera
effective number of parties index, and this is derived as follows: N
s
= 1/s
i
2
.
4 The catalogue of shortcomings of AV and STV is large, including the following: an
inability to ensure the election of a Condorcet winner (i.e. the election of a candidate
who beats all other candidates in a series of pairwise competitions), non-monotonic
tendencies (in which, paradoxically, a candidates chances of election can be harmed if
her vote increases), and the inconsistent treatment of preferences (with some voters
preferences being given more weight than others). Attention is also given to the prob-
lems of voters acting in an insincere manner, attaching more importance to preferred
outcomes rather than preferred candidates, a point encompassed by the Gibbard-
Satterthwaite theorem that strategic voting occurs under all (non-dictatorial) electoral
systems (see Cox 1997; Ordeshook 1986). The Oxford logician Michael Dummett is
probably the most trenchant social choice critic of AV and STV, arguing that they are
quasi-chaotic, exceptionally erratic in operation, producing results that are virtu-
ally random (Dummett 1997: 142, 151). He goes so far as to dub STV the second
worst electoral system ever devised (Dummett 1992: 111).
5 As we shall see in chapter 2, this may be one reason why Edward Nanson seemed pre-
pared to drop his theoretical doubts over preferential systems when advising on the
1902 Commonwealth Electoral Bill.
6 The coalition between the Liberal and National parties has remained in existence since
the 1920s, except for short periods in 1973-74 and 1987. Despite the permanent
nature of the arrangement, there has been little pressure for a merger; two conserva-
tive parties, one catering for urban dwellers, the other appealing to farmers living in
the bush, has suited those on the anti-Labor side of politics.
| 184 |
7 Namely, 1943 (16.3 percent), 1958 (10.6 percent), 1977 (12.2 percent), 1990 (17.2
percent), 1993 (10.8 percent), 1996 (14.3 percent) and 2004 (16 percent).
8 I n the 1990 federal election the Australian Democrats won 11.3 percent of the first
preference vote.
9 State and territory electoral system developments are also discussed in chapter 2, and
feature prominently in chapter 3.
Chapter 2
1 I nevitably, this chapter focuses on national-level trends, but where relevant we also
refer to trends at state/colony level (see also chapter 3).
2 The block vote system is a plurality system in multi-member constituencies.
3 Ultimately, STV was introduced for electing MPs from four university seats.
4 Mention should also be made of the contribution by the Ashworth brothers, particu-
larly by the publication of their book, Proportional Representation Applied to Party
Government in 1900. This was an influential study, although, their proposed hybrid
system (somewhat akin to a list system) was not taken very seriously. I n essence, their
principal role was to provide useful ammunition for critics of the electoral systems pro-
posed in the 1902 Bill.
5 Droop proposed a similar amendment in a paper published several years later
(McLean and Urken 1995: 47).
6 The Droop quota is calculated as follows: ([v/s+1]+1), where v = votes and s = seats.
7 The Hare quota is calculated as follows: (v/s).
8 J. B. Gregory (Melbourne) in the 1880s devised a scheme for transferring surplus votes
involving the transfer of all the ballots at a fraction of their original value. This scheme
removes the element of chance caused by transferring only surplus ballots at their full
value (see also chapter 3 below, p. 63).
9 Matthew Clarke, Federal Convention Debates, Second Session, Sydney, September 13
1897, p. 368. A few days after this comment, on September 21, the Tasmanian dele-
gation proposed an amendment to allow each state to determine its own method for
electing Senators, but the move attracted no support.
10 See, for instance, the comments of Barton and Deakin, both on April 15 1897, Federal
Convention Debates, First Session, Adelaide, p. 673. Also, see Deakin on March 7,
1898, Federal Convention Debates, Third Session, Melbourne, pp. 1925ff; on this occa-
sion, Barton states (p. 1925) that this is a matter for the future.
11 I n other words, depending on whether the election involved a half or a full dissolution,
district magnitude (i.e. the number of seats per constituency) would be either three or
six. Note that throughout this volume we use the term constituency to refer to the
electoral district in which a candidate is elected. I n Australia it is common practice to
refer to this as the electorate.
12 I n the early years of the Commonwealth, it was common for Bills to be initiated in the
Senate. The Senate encouraged this as a means of establishing the principle that min-
isters could and should speak in both Houses. While this may have been a break from
British parliamentary practice (Crisp 1949: 186), it was a familiar practice in
colonial/state upper houses.
13 Reid and Forrest end their treatment of this debate by stating, incorrectly, that
OConnors amended clause was accepted without a division (1989: 104). Reid and
Forrest are also incorrect in summarising the 1902 Act as allowing, for the Senate,
optional voting, up to six with crosses (1989: Table 3.1, p. 100). I n fact, the block vote
system eventually passed required voters to express a vote for as many candidates (no
more, no less) as there were seats to fill.
14 Strictly speaking, what OConnor was proposing was to offer the voter the option of
either voting as if the system was a block vote system casting as many votes as there
are seats to fill or else voting as if the system were a limited vote system casting
| 185 |
| Notes to pages 14 33 |
fewer votes than the number of seats to be filled. I t is likely that smaller parties would
have used this opportunity to encourage their supporters to treat this as a limited vote
system and to vote in the appropriate strategic manner to the benefit of the relevant
party.
15 The prime example is the single non-transferable vote system, used in Japan from
194894, a limited vote system in which the voter has just one vote and which has
been demonstrated to produce results which compare favourably (in terms of propor-
tionality) with some PR systems (Cox 1991).
16 At this point, Senator De Largie introduced another amendment (CPD 11090) that
would have turned this into the cumulative vote system, another semi-proportional
system with similar properties to the limited vote (Farrell 2001). I n this case the voter
has the right to cumulate votes on one or more candidates. The proposal attracted
very little support and was defeated on a vote.
17 See the comments of Cook: There has never been a greater outrage on responsible
government than the backing and filling of Ministers over this important Bill (CPD
16467) and Reid, who referred to this episode as a travesty of responsible govern-
ment (CPD 16466). They had a point, as shown by this summary of government posi-
tions on the clause: (1) first Senate debate, having lost STV, the government supports
limited vote (LV); (2) first House debate, the government opposes LV; (3) second
Senate debate, the government supports LV; (4) second House debate, the government
supports LV; (5) third Senate debate, the government opposes LV; (6) third House
debate, the government opposes LV.
18 Though Reid and Forrest (1989: 96) rightly single out Senator OConnor for the con-
summate skill he showed in steering the passage of the Bill through the Senate.
19 I t is clear in many of these debates that Nanson was still playing an active role behind
the scenes, as well as in public debates (most notably in a series of articles in the
Melbourne Age). I n introducing the 1918 bill, Patrick Glynn referred to meetings
with, and correspondence from, Edward Nanson.
20 The latter point was put most colourfully by Page, who exclaimed: Has Rip Van Winkle
been asleep on the Grampians? Let him wake up. Does he know there is a drought in
Queensland, and that bushfires are burning his sheep? (CPD October 31 1918, 7325).
21 At one point, Glynn suggested that the list system might be worth considering for the
Senate. (CPD October 4 1918, 6680).
22 Respectively, McDougall in CPD October 15 1919, 13322; Tudor in CPD October 21
1919, 13628; Hill in CPD October 21 1919, 13640. Edward Nanson was also an out-
spoken critic of this system. Even Patrick Glynn, on introducing the bill before the
House on October 21 (CPD, 13622), revealed his lack of enthusiasm for this proposal.
23 The formula for determining the minimum number of preferences was twice the
number of seats in the constituency plus one.
24 This occasion prompted two more attempts to propose an amendment bringing in
STV (CPD August 9 1922, 1203-22; CPD September 19 1922, 2377-93).
25 I n other words, the minimum threshold for being elected was [v/1+1] +1; very dif-
ferent from the Droop quota ([v/s+1] +1).
26 I n STV, when preferences are being transferred from winning candidates, one of two
systems apply: (1) only those votes surplus to the quota are transferred, or (2) all votes
are transferred, but based on a fraction of their true value. See chapter 3 below.
27 Most notable, perhaps, was the majority report of the 1919 Royal Commission on the
Constitution, which called for STV.
28 Under s.24 of the Constitution (the nexus clause), it is a requirement that the mem-
bership of the House should be approximately twice that of the Senate.
29 For coverage of the other changes introduced by electoral reform measures in the
1980s, see Uhr (2000: 24ff.).
30 As we shall see in the next chapter, however, PR-list systems have not been completely
ignored by Australian politicians.
| 186 |
Chapter 3
1 For that matter, the table also does not show the range of variation across Australia at
local and municipal level. The contemporary mix can be summarized as follows: New
South Wales uses STV; Northern Territory uses a mix of AV, SMP and consensus
voting; Queensland uses a mix of AV and block vote; South Australia uses a mix of AV
and STV; Tasmania uses STV; Victoria uses a mix of AV and STV; and Western
Australia uses SMP. The ACT does not have local government.
2 As we saw in chapter 2, STV was used in only two districts, and it was dropped the
following year.
3 These distinctions are clarified in the second section of this chapter (see page 52).
4 Queensland reverted to SMP from 1944-1960.
5 This discussion has benefited greatly from the recent work on this subject by Ben Reilly.
6 I n all these cases, there is little to distinguish them in terms of district size (today all
have single member constituencies), ballot paper design (a single vertical list deter-
mined by the parties; in the Northern Territory the ballot papers include candidates
photographs), or casual vacancies (use of by-elections). As we shall see in the next
section, these form additional important points of distinction in the STV systems.
7 I n Tasmania voters are required to express as many preferences as there are seats to fill.
As for its upper house elections (see below), South Australia also uses a ticket vote
mechanism for lower house elections to reduce the proportion of invalid votes. For
instance, in 2003, this reduced the proportion of invalid votes from seven percent to
three percent (information supplied by Jenni Newton-Farrelly). Western Australia
experimented with a similar procedure in the 1980s, but abandoned it for fear of voter
confusion (information supplied by Antony Green).
8 The committee held to precisely the same position in its report relating to the 2001
election ( Joint Standing Committee on Electoral Matters 2003).
9 We also include details of Victorias new STV system for upper house elections, which
will be used for the first time in 2006.
10 I ncluding Ireland and Malta increases our N of cases for the statistical analysis of
electoral trends in chapter 4 below. Throughout this section we are dealing with the
situation in Malta up to and including 1981. I n 1987, a contingent upper tier was
introduced, designed to benefit the winning party, and potentially so distorting the
election result as to question the degree to which the system could still be classified as
part of the STV family (in general, see Hirczy de Mino and Lane 2000).
11 Ballot paper design can also have important consequences for the voters. A question
can be raised over which is fairer: a ballot paper that sorts by candidates (Ireland),
where the voters has to work harder to find the relevant party labels, or a ballot paper
that sorts by parties (all other systems), where the voter has to work harder to find the
relevant candidates. For further discussion, see Darcy and Marsh (1994).
12 It is worth noting in this context that not only do Tasmania and the ACT not operate
ticket voting for their state/territory elections, they also ban the use of how-to-vote cards.
13 Given that the count in New South Wales is now carried out by computer, it has been
suggested that, in time, perhaps the state might finally move over to a Gregory method
of vote transfer (Green 2003: 18).
14 The Proportional Representation Society of Australia produced reports on NSW elec-
tions in the early 1980s that demonstrated the significance of this random element in
cases of recounts when new random samples are drawn (personal communication from
Bogey Musidlak).
15 The transfer value is calculated as follows: TV =s/a, where sis the candidates surplus
vote, and ais the total number of ballot papers received by the candidate (in the case
of surplus transfers arising from the first count), or the last parcel of ballot papers
received (in the case of surplus transfers in later counts).
16 We are grateful to Paul Wilder for drawing our attention to this issue, and to Bogey
| 187 |
Musidlak for guiding us through its intricacies. I n general there has been a tendency
to overlook this modification of the Gregory method (Fischer 1988: 142; Gallagher
and Unwin 1986: 247; McLean 1996: 377; Reilly and Maley 2000: 58).
17 TV =s/n, where nis the total number of ballot papers in the candidates pile. We refer
to this as inclusive because of the way it takes accounts of all ballot papers.
18 Some of the following discussion is based on personal communications from Alastair
Fischer, Michael Maley and Bogey Musidlak.
19 See the comments of the chair of the Joint Select Committee on Electoral Reform
( JSCER; the precursor to todays Joint Select Committee on Electoral Matters), on
August 8, 1985, p. 1057. JSCER Report No. 2, Volume 7, Official Hansard Record,
December 1986.
20 See for instance the submission by the Proportional Representation Society of
Australia to the JSCER in 1983, which, while acknowledging the possibility that a
modified form of the Gregory method to take account of all ballot papers might be
adopted, warned that this would impose a very considerable additional burden on the
Electoral Office ( JSCER Official Hansard Report, 21-22 June 1983, p. 739). At no
other point, either in its written submission or in its public evidence, does the PR
Society make any further reference to this possible modification. Alastair Fischer has
also informed us (private communication) that he was unaware that such a modifica-
tion was being considered.
21 The closest parliamentarians came to addressing this issue was when Senator Missen
read out a letter he had received from Jack Wright (CPD, 30 November 1983: 3055)
which warned of several problems in the proposed reforms of the Senate system, one
of which was the modifications to the Gregory method. Hardly a model of clarity,
most of Wright s warnings appear to have fallen on deaf ears.
22 The merits of the inclusive-Gregory method was given far greater prominence in the
subsequent report of the JSCER in 1986, in which, in particular, the Proportional
Representation Society of Australia fought a rearguard action to have it replaced by a
weighted inclusive version (discussed in chapter 4 below). While this proposal received
a sympathetic hearing from some committee members, ultimately it was deemed too
cumbersome to implement. I n addition, the Electoral Commission staff argued in
favour of retaining the inclusive-Gregory method.
23 As we saw in the previous chapter, since 1983 it has been accepted that a ballot paper
that has 90 percent of the preferences completed (with no more than three omissions
or duplication in sequential voting) is counted as valid; however, this fact is not adver-
tised and therefore not generally known.
24 As Sharman (1982: 103) argues, in most cases, the adoption of Senate-style STV has been
by the ALP, which favours those electoral systems prompting strong party discipline.
25 We are grateful to Gareth Griffith for providing us with press clippings of the debates
that followed in the wake of this election.
26 Known technically as the effective threshold, the electoral threshold is an estimate of
the minimum proportion of votes that parties need before they start to win seats. I n
New South Wales, the effective threshold was 3.4 percent. For more details about
effective thresholds generally, see chapter 4, pp. 8487.
27 However, this did not stop the number of candidates rising from 264 in 1999 to 284 in 2003.
28 Technically, there is a ticket-deal between parties, because the parties are obliged to
nominate a second party to receive preferences (i.e. those not otherwise allocated by
those voters who opt to express more than one preference above the line). This is in
order to ensure that a vote is not declared invalid if, for example, a voter has expressed
just one preference above the line and the party in question loses its 15th candidate
due to an unexpected death which did not apply in 2003.
29 Clearly this new procedure based on a proposal from the Green Party (Rhiannon
1999) was facilitated by the fact that in New South Wales council elections a voter
is only required to complete 15 preferences (Table 3.4), which, it so happens, is the
minimum number of candidates a party needs in order to have a space above the line.
| 188 |
I n other words, by expressing one preference above the line, the vote is valid because
this represents 15 preferences below the line. For details on the electoral reform, see
New South Wales Parliamentary Debates November 18 1999, 338183. For an
analysis of how such a reform might operate in South Australia, see Newton-Farrelly
(2000: 2123). I n its 2003 report, the Joint Select Committee on Electoral Matters
referred to the possibility that such a change might be introduced in the future for
elections at national level, and it proposed to return to this question on a future occa-
sion ( JSCEM 2003: para. 7.40).
30 The Constitution originally provided for Senate casual vacancies to be filled by by-
election (s.15 un-amended); an additional member would be elected at the next half-
Senate election. I f there was a House of Representatives election before the next
Senate election, a single senator would be elected to that state. Until the by-election,
the state parliament (or the state government if the parliament was not in session)
would appoint a replacement.
31 Bizarrely, in the original version of STV that New South Wales used to elect its
Legislative Assembly between 1918 and 1926, no provision was made for dealing with
casual vacancies (Wright 1980: 121). This was subsequently corrected (Hogan 2001a).
32 By-elections were more common in STV systems in the past. For instance, Tasmania used
by-elections until 1917; the Australian Senate used them until as recently as the 1970s.
33 Ireland also represents one of the few cases of a PR system that incorporates by-elections.
As a result of its electoral reform in the early 1990s, New Zealand is another example.
34 Of course, there is nothing to stop a state breaking with this lower house/upper house
distinction in the future.
35 I n largest remainder systems, the Droop quota is generally referred to as Hagenbach-
Bischoff . I t is calculated as follows: v/(s+1), where v = votes and s= seats. This pro-
duces slightly less proportional results than the Hare quota (v/s), which is most
commonly used in largest remainder systems.
36 [100/11+1]/2 = 4.2.
37 Bennett (1992: 170) observes that the new electoral system finally adopted in South
Australia in 1973 represented the first example of ticket voting in Australia, in which
voters voted for groups of candidates rather than for individual candidates (see the
sample ballot paper in Bennett 1992: 171). As we saw in chapter 2, this South
Australian experiment was to influence the decision to adopt ticket voting at national
level a decade later.
38 I n its original conception, the intention had been to have a closed list system. See
JSCEM Report No. 5, I nquiry into the ACT Election and Electoral System, 1989: 50.
39 What follows is a simplified summary from the national government s JSCEM
I nquiry into the operation of the electoral system ( JSCEM 1989: ch. 4).
40 100/17+1 = 5.56.
41 This is the most common form of divisor used in the highest average list system,
whereby the parties votes are divided by a series of divisors (1, 2, 3, etc.) and a seat is
awarded to the party with the highest average vote after each division (for more
details, see Farrell 2001: 208-09).
42 Disproportionality is a measure of the average disparity between the proportions of
votes won by parties in an election and the proportions of seats they are awarded.
Various indices of disproportionality are available, of which the most widely used is the
Gallagher index (see chapter 1, note 3). The higher the score, the greater the dispro-
portionality.
Chapter 4
1 I n this chapter we are following the same procedure adopted by Arend Lijphart (1994)
in his influential cross-national study of electoral systems hence our reference to
| 189 |
Lijphart s rule. This is the only aspect in which AV has varied over time; in all its
other features (district magnitude, electoral formula, ballot structure), AV has
remained unchanged across the post-war period.
2 Throughout this book, we use the Gallagher I ndex (GI ) of disproportionality (see
chapter 1, note 3).
3 The two-party preferred vote is calculated by allocating all remaining preferences in a
constituency count to either the ALP or Coalition candidates, thereby providing a
final record of the share of the vote to either of the two main political groupings.
Given that we are now dealing with just two parties, by simply subtracting the differ-
ence between vote and seat shares for one of the parties (in this case ALP), we have
the equivalent of a disproportionality score.
4 I n section 4.3 we will see cases of how, under certain circumstances, preferences can
matter for the determination of the final result.
5 As Gallagher (1992: 480) observes, this assumption may have some credence in Malta
(as we have seen, also in the case of Australian Senate elections), but it should be made
clear that this does not give a full picture of the way [STV] can operate in practice.
6 After some debate over the correct derivation of T
eff
, the general consensus has settled
on the following: 75%/(M + 1) (see Lijphart 1997; Penads 1997; Taagepera 1998).
7 I t is a mirror image of M, which explains why the T
eff
trends in Table 4.3 are the
inverse of the M trends.
8 There was an additional change at the start of the period which was due to the adjust-
ment to STV in 1949. Coinciding with the switchover, there was an enlargement of
the Senate requiring a one-off election of seven senators per state in 1949 instead of
the usual five. This accounts for Australia1.
9 I n 1999 the New South Wales government also attempted to reduce average M in
the light of the tablecloth election that year but was prevented from doing so due
to strong opposition from other parties, as well as splits in its own ranks.
10 We have not included I srael and the Netherlands in the Lijphart averages because they
have unusually large Ms owing to the fact that in both cases the entire nation is one
constituency. We have also excluded Lijphart s European Parliament cases. Note that
in his 1994 book, Lijphart used a slightly different formula for calculating T
eff
:
[50%/(M + 1)] + [50%/2M]. He subsequently converted to the formula we are
employing here (Lijphart 1997). For this reason, we do not compare the T
eff
scores.
11 Disproportionality is measured by the Gallagher index (GI ). The effective number of
parties is derived as follows: N
v
= 1/v
i
2
; N
s
= 1/s
i
2
(Laakso and Taagepera 1979).
12 The effective number of parties (N
v
based on votes; N
s
based on seats) is a summary
measure of the number of parties in a political system which takes account of their rel-
ative sizes. The higher the index the greater the number of parties (see Laakso and
Taagepera 1979).
13 For details on the nine variants of STV, see Table 3.4. I n this instance, we are treating
each system in which STV operates as a separate case; we are also treating half-Senate
(Australia) and full-Senate (AustraliaDD) elections as separate variants. To maximise
the number of cases in our regression analysis we have treated each election in each of
the STV variants as one unit of measure.
14 The low disproportionality score for Malta, in part, probably reflects our small number
of cases and some high inter-correlations between the variables; but it may also be due
to the relatively small number of parties fielding candidates in the election.
15 For more sustained analysis of how the number of preferences available to voters in AV
systems can have a bearing on elections results, see Cox (2003).
16 Authors calculations based on raw data in Green (2003a).
17 Another variant that should also be mentioned in passing is the Warren method, but
to date this has not been adopted anywhere (Tideman and Richardson 2000).
18 As Narelle Miragliotta points out, however, because the Meek method also recalcu-
lates the quota throughout the count process and transfers ballot papers to candidates
already elected, it is possible that it might occasionally produce slightly different elec-
| 190 |
tion results to the weighted inclusive Gregory system (Miragliotta 2002, 33).
19 This example is based on one produced by the Proportional Representation Society of
Australia.
20 According to Bogey Musidlak (personal communication), this was found in an
analysis of periodic national elections for the Aboriginal and Torres Strait I slander
Commission (ATSI C) which also uses the inclusive Gregory method.
21 The matter was debated by the Standing Committee on Legislation, resulting in a
minority recommendation to replace the inclusive Gregory method with the Gregory
method (Report of the Standing Committee on Legislation, No. 8, November 2001).
This recommendation was not adopted. At the time of writing (March 2005), Western
Australia is considering proposals to switch to a weighted inclusive Gregory method
for future elections, this move in part prompted by advice from the current authors.
22 According to information supplied by the Western Australian Electoral Commission
during a debate in the Legislative Councils Estimates Committee, 4 252 of Nevills
4 397 ballot papers consisted of ticket votes, and of the remaining 145 ballot papers,
57 placed Smith above Chapple, and 88 placed Chapple above Smith (WALC,
Estimates Committee Supplementary I nformation, E855, 15-18 October, 2001).
23 WALC, 20 February 2002, 7657-58; 21 February 2002, 777375.
Chapter 5
1 To put this slightly more technically, our emphasis in this chapter is on the micro-
behavioural outcomes of electoral systems rather than the macro-systemic features
explored in the previous chapter.
2 Joan Rydon (1986: 115) suggests a similar level of party involvement among
Australian national MPs
3 On the eve of the Second World War, the ALP had 370 000 members, nearly seven
times its current membership (Crisp 1983: 20).
4 The result is comparatively little overlap within the major parties between their state
and national components in terms of either personnel or practice. For example, com-
paratively few state elected representatives use their state experience as a stepping
stone to federal politics, or vice versa. I n the 1990 parliament, for example, only seven
members of the cabinet or shadow cabinet had previously been elected to state parlia-
ments (McAllister 1992: 228).
5 The question (which was not asked in the subsequent Australian Candidate studies)
referred to legislative voting so that it would be applicable both to House of
Representatives candidates (who have constituency responsibilities) and Senate candi-
dates (who do not). A question asking solely about constituency versus national or leg-
islative responsibilities would have applied only to the former group. Hereafter we
refer to members of both houses as Members of Parliament (MPs) or as legislators.
6 For example, the Western Australian constituency of Kalgoorlie is the largest single-seat
parliamentary constituency in the world; in total size it exceeds the landmass of France.
7 The minor parties are Democrat, Green and One Nation. I ncluding the parties sepa-
rately produces the same results, so for the sake of parsimony, a single variable is used.
Chapter 6
1 As Oldham (1911: 162) put it: a thoroughly efficient roll can only be continually pre-
served under a system of compulsory registration, or by the employment of an army of
electoral inspectors to continuously watch the movements of people throughout the
Commonwealth.
2 Senator Pearce, who introduced the item in Parliament, argued that compulsory
registration was a duty, not a privilege, and saw it as analogous to other compulsory
| 191 |
measures, such as schooling and medical vaccination.
3 Estimates are difficult to make prior to 1946.
4 The 2001 and 2004 figures for voting age population are estimated from ABS data.
5 Monthly surveys conducted for the AEC since November 1997 have shown an
average of 78.2 percent of 1824 year olds are enrolled. Our thanks to Brian Hallett
at the AEC for providing this information.
6 These figures explain how the I nternational I nstitute for Democracy and Electoral
Assistance (1997: 14) can show Australia as having a turnout of only 82.7 percent
during the 1990s, ranking it 20th in the world in terms of turnout. Estimating turnout
based on registration rather than eligible population puts Australia in second place,
having a turnout at national elections during the 1990s of 96.2 percent, just behind
Malta with 96.7 percent.
7 See http://www.int-idea.se/voter_turnout/compulsory_voting.htm.
8 A non-voter who fails to provide an acceptable explanation is fined; if the fine remains
unpaid, he or she may incur a prison sentence. Such cases are usually classified as
failure to pay the fine rather than failure to vote, hence estimates of the number of
individuals who fall into this category are difficult to make. Technically speaking, of
course, an Australian voter is not actually forced to vote. Under the existing legislation,
she is required to turn up to the polling station, take a ballot paper to a polling booth,
and fold it. She can then opt to return the ballot paper to the presiding officer
(Commonwealth Electoral Act, 1918, ss. 23133).
9 Note that South Australia did not introduce compulsory voting for Council elections
until the 1970s. I n general there was no agreement on the introduction of compulsion
for local government elections and it has applied in only a few states, some of which
have veered between voluntary and compulsory systems in the post war years.
10 I n October 1930, the first NSW state election held under compulsory voting, turnout
rose further, to 94.9 percent.
11 Needless to say, there are scenarios under which it might be advisable not to give a first
preference to B, such as in the event that B might have so many first preferences that
her ballot papers are transferred late in the count and therefore have no influence on
the final outcome. While accepting that the form of strategic voting is very different
under preferential electoral systems, Dummett (1997: 94), nevertheless, forcibly argues
that it is no less invidious.
12 Mock-ballot surveys provide one instance in which voters (as survey respondents) are
asked to make strategic decisions without any influence from party strategists (see
Bowler and Farrell 1996).
13 There is also evidence that party solidarity is high in Malta, which does not have a
system of ticket-voting; though given that Malta has a simple two-party system it is
understandable why a large proportion of voters should choose to stay within one party
when transferring preferences.
14 For instance, in the 1920 New South Wales Legislative Assembly election (which then
used STV), the practice was for a party to use the how to vote cards solely for indi-
cating the preferences for candidates of other parties, leaving the first five preferences
(against that partys candidates) blank so as not to give precedence to any of its candi-
dates (Hogan 2001: 21314). I n 1922, the Labor party explicitly sought to spread the
preferences; but the instructions on the how to vote cards were hardly a model of
clarity: Electorate divided into 2 parts. Mr OHALLORAN will receive No. 1 Vote
in the Randwick side of Oxford Street and Bronte tramline; Mr DWYER will receive
No. 2 Vote in this section. Mr DWYER will receive No. 1 Vote on the Woollahra side
of Oxford Street and Bronte tramline; Mr OHALLORAN will receive 2nd
Preference Vote in this area (quoted in Hogan 2001a: 246). I n the 1925 election,
Labor abandoned efforts to spread the preferences and moved over to the cascade
system (Hogan 2001a: 264).
15 An examination of the electoral returns in four sets of national elections (19932001) revealed
no instances of where the cascade strategy harmed the electoral chances of a candidate.
| 192 |
16 We recognise that a proportion of these transfers consist of ballot papers received in
earlier counts from other candidates, and therefore also from other parties, which means
that our figures are likely to over-estimate the true extent of cross-party leakage.
17 We tested for partisan differences in whether or not the respondents followed party
cues, but these were found to be non-significant so they are not examined in the results
presented in Table 6.4.
18 There are 42 respondents in the ACT, but just 11 in the Northern Territory.
19 Comparative research has shown that compulsory voting is far more common in newer
democracies (I nternational I DEA 1997).
20 The JSCEM report on the 1998 election took no position on compulsory voting:
Although there are strong views regarding compulsory voting, this Committee has no
plans to pursue the issue of voluntary voting ( JSCEM 2000). I n its most recent report,
the JSCEM held to the same position ( JSCEM 2003).
21 Survey estimates of expected turnout will, as Jackman (1997) notes, underestimate the
decline in turnout due to voluntary voting. This is caused by the tendency of opinion
surveys to under sample disinterested voters who would be less likely to vote under a
voluntary system. I t is difficult to estimate the likely error caused by survey non-
response in predicting turnout, but it is likely to be small. I n any event, our interest
here is in the broad patterns rather than in precise estimates.
22 These effects for vote are similar across the four surveys for which we have data.
23 Also of relevance in this context is the fact that Australia has a high proportion of non-
English speaking immigrants (McAllister and Makkai 1993), increasing the potential
for voter confusion.
Chapter 7
1 The CSES dataset (both modules 1 and 2) are available for download at
<http://www.cses.org/>. The website also contains details of the participating coun-
tries, technical details of the surveys and an extensive bibliography of works using the
CSES datasets.
2 Critics of this feature argue that it brings with it voter confusion and the risks of voting
paradoxes; see e.g. Dummett (1997).
3 Unlike the earlier efforts to examine the relationship between ballot structure charac-
teristics and levels of democratic satisfaction (Norris 2004), the merit of the approach
to be outlined here is that we provide a more nuanced delineation of electoral systems
on this dimension, which, for instance, allows us to more accurately locate the mixed
electoral systems vis--vis the other more established systems.
4 We lack sufficient detail on all the CSES country cases to replicate Shugart s other
index of inter-party efficiency.
5 Since these selection rules are not readily available for all the countries in the CSES
sample, our coding is similar to that of Carey and Shugart (1995: 420-21).
6 We add the subscripts to distinguish between the Carey and Shugart (1995) codes and
the Shugart (2001) codes.
7 I n a footnote, Carey and Shugart give due recognition to the fact that list systems can
vary in terms of ordinality, but in this paper they do not attempt to capture these sorts
of variation (1995: 435).
8 We have discussed our coding privately with Matthew Shugart, and there is general
agreement that, were SNTV to still be a relevant system requiring coding in this
analysis, then this Votecomponent would need further refinement to take account of the
distinction between SNTV and SMP (and other nominal categorical) systems. Under
Shugart s schema, they are coded together on this component, but whereas it would be
correct to locate SNTV further on the extreme on this dimension than STV, this would
not be the same for SMP. Given that we are not dealing with SNTV systems in the
analysis that follows, we shall simply leave them out of consideration here.
| 193 |
9 It is interesting to note how for Shugart, intra-party competition does feature in the case of
this component even if its significance tends to be downplayed for the other components.
10 The 29 countries are listed in Figure 7.1 and the Appendix Table A.2. Of the 29 countries,
26 were derived from module 1 of the CSES dataset <http://wwwcses.org>, and three
(Bulgaria, France and Ireland) from module 2. Note that we are only including democratic
countries in our CSES sample (based on the usual Freedom House scores). A merged
dataset was created from the common items of interest in the two datasets. The fieldwork
dates for the surveys are shown in Appendix Table A.2. Belgium was separated in module
1 between Flanders and Walloon; in our analysis the two samples are combined.
11 The satisfaction question was: On the whole, are you very satisfied, fairly satisfied, not
very satisfied, or not at all satisfied with the way democracy works in [country]? Very
satisfied was scored 10, fairly satisfied 6.7, not very satisfied 3.3, and not at all satis-
fied 0. This coding was used to assist with interpretation.
12 One country that we would have liked to include in the analyses Thailand did not
include a question on turnout at the previous election and therefore had to be
excluded. A second issue with the turnout measure in the CSES data is the over-esti-
mation of turnout. Few of the CSES countries provide validation of voters, thereby
over-estimating the actual number of voters. I n reducing the error associated with this
measure we have followed the approach of Karp and Banducci (2004: 11) by assuming
that non-responses to the turnout question are non-voters, and weighting the coun-
tries to compensate for unequal sample sizes.
13 A range of additional socio-economic variables were examined in the preliminary
analyses, but eventually excluded because they added little or nothing to the explana-
tory power of the model.
14 See chapter 1, note 3.
15 An additional control variable would be number of elapsed years since the last elec-
tion, on the grounds that respondents who had voted most recently would be more sat-
isfied than voters with more distant memories of the event. However, since the CSES
data are based on national election surveys, in almost all cases the elapsed time since
the last election was zero.
16 The correlation of -.16 is smaller than the correlation of -.33 noted earlier because the
units of analysis in the first calculation are 40 403 individual respondents and in the
second equation, 29 countries.
17 Another possibility is that there are country-specific variations, above and beyond the
variables controlled for in the fourth model, which account for the differences. Several
multilevel models were estimated to test the random and fixed country effects.
However, the 29 countries were too closely correlated with the electoral system and
country specific variables in order to follow this line of enquiry further.
18 This is consistent with previous studies that have considered indirect causal effects
between electoral systems and measures of democratic satisfaction. For instance,
Anderson and Guillory (1997) and Norris (1999) assess the mediating role of being
losers or winners in a polity as the basis for their examination of the relationship
between electoral institutions and levels of democratic satisfaction.
Chapter 8
1 The recent referendum for the possible adoption of STV in the Canadian province of
British Columbia supports this contention, for this STV option resulted from the delib-
erations of an assembly comprising a representative sample of the provinces citizens.
2 However, see our comments on p. 174 about the current Australian tendency not to
embrace change.
| 194 |
Bibliography
Aitkin, Don (1982) Stability and Changein Australian Politics, 2nd ed. Canberra: Australian
National University Press.
Aitkin, Don & Francis Castles (1989) Democracy Untrammelled, in K. Hancock (ed.)
Australian Society. Melbourne: Cambridge University Press.
Ames, Barry (1995) Electoral Strategy under Open-List Proportional Representation,
American Journal of Political Science. 39: 40633.
Anderson, Christopher & Christine Guillory (1997) Political I nstitutions and Satisfaction
with Democracy: A Cross-National Analysis of Consensus and Majoritarian Systems,
American Political ScienceReview. 91: 6681.
Arrow, Kenneth (1951) Social Choiceand Individual Values. New Haven: Yale University
Press.
Ashworth, T. R. & H. P. C. Ashworth (1900) Proportional Representation Applied to Party
Government: A New Electoral System. Melbourne: George Robertson and Co.
Bartholdi, J. & J. Orlin (1991) Single Transferable Vote Resists Strategic Voting, Social
Choiceand Welfare. 8: 34154.
Bawn, Kathleen (1993) The Logic of I nstitutional Preferences: Electoral Law as a Social
Choice Outcome, American Journal of Political Science. 37: 96589.
Bean, Clive (1990) The Personal Vote in Australian Federal Elections, Political Studies38:
253268.
Beitz, Charles (1989) Political Equality: An Essay in DemocraticTheory. Princeton, NJ:
Princeton University Press.
Bennett, Scott (1992) Affairs of State: Politics in theAustralian States and Territories. Sydney:
Allen & Unwin.
(1995) These New Fangled I deas: Hare-Clark 18961901, in Marcus Haward &
James Warden (eds), An Australian Democrat: The Life, Work, and Consequences of
Andrew Inglis Clark. Hobart: Centre for Tasmanian Studies.
Black, David (1989) LegislativeCouncil of Western Australia: Membership Register, Electoral
Law and Statistics, 18901989. Perth: Western Australia Parliamentary History
Project.
Black, Duncan (1996) The Principles of Parliamentary Representation, in I ain McLean,
Alistair McMillan & Burt Monroe, A Mathematical Approach to Proportional
Representation: Duncan Black on Lewis Carroll. Boston: Kluwer Academic Publishers.
| 195 |
Blais, Andr & R. K. Carty (1991) The Psychological I mpact of Electoral Laws: Measuring
Duvergers Elusive Factor, British Journal of Political Science. 21: 7993.
Blais, Andr & Louis Massicotte (1996) Electoral Rules, in Richard Niemi, Lawrence
LeDuc & Pippa Norris (eds), Comparing Democracies. London: Sage.
Blau, Adrian (2003) The Comparative Study of Electoral Systems: Time, Space, and the
Translation of Votes to Seats(mimeo).
Blondel, Jean (1969) An Introduction to ComparativeGovernment. London: Weidenfeld and
Nicolson.
Bogdanor, Vernon (1981) ThePeopleand theParty System: TheReferendumand Electoral
Reformin British Politics. Cambridge: Cambridge University Press.
(1984) What is Proportional Representation? A Guideto theIssues. Oxford: Martin
Robertson.
Bowler, Shaun (1996) Reasoning Voters, Voter Behaviour and I nstitutions: The Decision
Dependence of Voting Behaviour, in David Farrell, David Broughton, David Denver
& Justin Fisher (eds), British Elections and Parties Yearbook, 1996. London: Frank Cass.
Bowler, Shaun, Elisabeth Carter, & David Farrell (2004) Changing Party Access to
Elections, in Bruce Cain, Russell Dalton, & Susan Scarrow (eds), Democracy
Transformed? Expanding Political Opportunities in Advanced Industrial Democracies.
Oxford: Oxford University Press.
Bowler, Shaun, Todd Donovan & David Farrell (2000) Party Strategy and Voter
Organization under Cumulative Voting in Victorian England, Political Studies. XLVI I :
90617.
Bowler, Shaun, Todd Donovan, David Brockington & Sheila Mello (2002) Electoral Reform
and Minority Representation: Local Experiments with AlternativeElections. Columbus,
Ohio: Ohio State University Press.
Bowler, Shaun & David Farrell (1991) Party Loyalties in Complex Setting: STV and Party
I dentification, Political Studies. 39: 35062.
(1991a) Voter Behavior under STV-PR: Solving the Puzzle of the Irish Party System,
Political Behavior. 13: 30320.
(1993) Legislator Shirking and Voter Monitoring: I mpacts of European Parliament
Electoral Systems Upon Legislator-Voter Relationships, Journal of Common Market
Studies. 31: 4569.
(1996) Voter Strategies under Preferential Electoral Systems: A Single Transferable
Vote Mock Ballot Survey of London Voters, in Colin Rallings, David Farrell, David
Broughton & David Denver (eds), British Elections and Parties Yearbook 1995. London:
Frank Cass.
Bowler, Shaun, David Farrell & I an McAllister (1996) Constituency Campaigning in
Parliamentary Systems with Preferential Voting: I s There a Paradox?, Electoral Studies.
15: 46176.
Bowler, Shaun, David Farrell & Robin Pettitt (2005) Expert Opinion on Electoral
Systems: So Which Electoral System is Best?, Journal of Elections, PublicOpinion and
Parties. 15: 320.
Bowler, Shaun & Bernard Grofman (eds) (2000) Elections in Australia, Ireland, and Malta
under theSingleTransferableVote: Reflections on an Embedded Institution. Ann Arbor:
University of Michigan Press.
(2000a) I ntroduction: STV as an Embedded I nstitution, in Shaun Bowler & Bernard
Grofman (eds) Elections in Australia, Ireland, and Malta under theSingleTransferable
Vote: Reflections on an Embedded Institution. Ann Arbor: University of Michigan Press.
(2000b) Conclusion: STVs Place in the Family of Electoral Systems, in Shaun
Bowler & Bernard Grofman (eds) Elections in Australia, Ireland, and Malta under the
SingleTransferableVote: Reflections on an Embedded Institution. Ann Arbor: University
of Michigan Press.
Brams, Steven & Peter Fishburn (1983) Approval Voting. Boston: Birkhauser.
(1984) Some Logical Defects of the Single Transferable Vote, in Arend Lijphart & Bernard
Grofman (eds) Choosing an Electoral System:Issuesand Alternatives. Westport: Praeger.
| 196 |
Butler, David (1963) TheElectoral Systemin Britain Since19182nd ed. Oxford: Clarendon
Press.
(1973) TheCanberra Model: Essays on Australian Government. Melbourne: Cheshire.
Cain, Bruce, John Ferejohn & Morris Fiorina (1987) ThePersonal Vote:Constituency Service
and Electoral Independence. Cambridge, MS: Harvard University Press.
Carey, John & Matthew Shugart (1995) I ncentives to Cultivate a Personal Vote: A Rank
Ordering of Electoral Formulas, Electoral Studies. 14: 41739.
Carstairs, Andrew McLaren (1980) A Short History of Electoral Systems in Western Europe.
London: George Allen & Unwin.
Castles, Francis G. (1994) The Policy Consequences of Proportional Representation: A
Sceptical Commentary, Political Science. 46: 16171.
Chaples, Ernie (1985) Wran in Government, 197681, in Ernie Chaples, Helen Nelson &
Ken Turner (eds), TheWran Model: Electoral Politics in New South Wales 1981 and 1984.
Melbourne: Oxford University Press.
Chubb, Basil (1963) Going About Persecuting Public Servants: The Role of the Irish
Parliamentary Representative, Political Studies. 11.
Coakley, John & Gerald ONeill (1984) Chance in Preferential Voting Systems: An
Unacceptable Element in Irish Electoral Law?, Economicand Social Review. 16: 118.
Cohan, A. S. (1979) Rejoinder, British Journal of Political Science9: 36970.
Collins, Hugh (1985) Political I deology in Australia: the Distinctiveness of a Benthamite
Society, Daedalus114: 14769.
Colomer, Josep (ed.) (2004) Handbook of Electoral SystemChoice. Houndmills, Basingstoke:
Palgrave Macmillan.
Colston, Malcolm (1975) TheOdd OneOut: An Account of the1974 SenateElections in
Queensland. Colonial Press.
Coram, Bruce Talbot (1996) Second Best Theories and the I mplications for I nstitutional
Design, in Robert Goodin (ed.), The Theory of Institutional Design. Cambridge:
Cambridge University Press.
Cox, Gary (1990) Centripetal and Centrifugal I ncentives in Electoral Systems, American
Journal of Political Science. 34: 90335.
(1991) SNTV and dHondt are Equivalent , Electoral Studies. 10: 11832.
(1997) Making Votes Count: Strategic Coordination in theWorlds Electoral Systems.
Cambridge: Cambridge University Press.
(2003) Instant Runoff Voting with Restricted Voting(mimeo).
Crewe, I vor (1981) Electoral Participation, in David Butler, Howard R. Penniman &
Austin Ranney, eds, Democracy at thePolls. Washington American Enterprise I nstitute.
Crisp, L. F. (1949) The Parliamentary Government of the Commonwealth of Australia.
London: Longmans, Green & Co.
(1978) TheAustralian Federal Labour Party, 19011951. Sydney: Hale & Iremonger.
(1983) Australian National Government 5th ed. Melbourne: Longman Cheshire.
Curtice, John & Alison Park (1999) Region: New Labour, New Geography?, in Geoffrey
Evans & Pippa Norris (eds), Critical Elections: British Parties and Voters in Long-Term
Perspective. London: Sage.
Dalton, Russell J. & Martin P. Wattenberg (eds) (2001) Parties Without Partisans. Oxford:
Oxford University Press.
Darcy, Robert & I an McAllister (1990) Ballot Position Effects, Electoral Studies. 9: 517.
& Malcolm Mackerras (1993) Rotation of Ballots: Minimizing the Number of
Rotations, Electoral Studies. 12: 7782.
& Michael Marsh (1994) Decision Heuristics: Ticket-Splitting and the Irish Voter,
Electoral Studies. 13: 3849.
Doron, Gideon & Richard Kronick (1977) Single Transferable Vote: An Example of a
Perverse Social Choice Function, American Journal of Political Science. 21: 30311.
Dummett, Michael (1984) Voting Procedures. Oxford: Clarendon.
(1992) Towards a More Representative Voting System: The Plant Report, New Left
Review. 194: 98113.
| 197 |
| Bi bl i ography |
(1997) Principles of Electoral Reform. Oxford: Oxford University Press.
Dunleavy, Patrick, Helen Margetts & Stuart Weir (1992) Replaying the 1992 General
Election: How Britain Votes under Alternative Electoral Systems, LSE PublicPolicy
Paper No. 3. London: London School of Economics.
(1993) The 1992 Election and the Legitimacy of British Democracy, in David
Denver, Pippa Norris, David Broughton & Colin Rallings (eds), British Elections and
Parties Yearbook 1993. Hemel Hempstead: Harvester Wheatsheaf.
Duverger, Maurice (1954) Political Parties: Their Organization and Activity in theModern
State. London: Methuen.
Elliott & Shanahan Research (1989) Identifying Barriers and Motivations for Enrolment for
Young People. Sydney: Elliott & Shanahan Research Report.
Engstrom, Richard L (1998) Minority Electoral Opportunities and Alternative Election
Systems in the United States, in Mark E. Rush (ed.), Voting Rights and Redistricting in
theUnited States. Westport: Greenwood Press.
Evans, Geoffrey, John Curtice & Pippa Norris (1998) New Labour, New Tactical Voting,
British Elections and Parties Review. 8: 558572.
Farrell, Brian (1985) Ireland: From Friends and Neighbours to Clients and Partisans, in
Vernon Bogdanor (ed.), Representatives of thePeople? Aldershot: Gower.
Farrell, David (1993) Campaign Strategies, in Michael Gallagher & Michael Laver (eds),
How Ireland Voted 1992. Dublin: Folens/PSAI Press.
(1994) Ireland: Centralization, Professionalization and Campaign Pressures, in
Richard S. Katz & Peter Mair (eds), How Parties Organize: Adaptation and Changein
Party Organizations in Western Democracies. London: Sage.
(2001) Electoral Systems: A Comparative Introduction. London and New York:
Palgrave.
(2001a) The United Kingdom Comes of Age: The British Electoral Reform
Revolution of the 1990s, in Matthew Shugart & Martin Wattenberg (eds), Mixed-
Member Electoral Systems:TheBest of Both Worlds? Oxford: Oxford University Press.
& Michael Gallagher (1998) Submission to theIndependent Commission on theVoting
System. London: McDougall Trust.
Malcolm Mackerras & I an McAllister (1996) Designing Electoral I nstitutions: STV
Systems and their Consequences, Political Studies. 44: 2443.
& Roger Scully (2003) Electoral Reform and the British MEP, Journal of Legislative
Studies. 9: 1437.
Fenno, R. (1978) HomeStyle: HouseMembers in their Districts. Boston, MA: Little Brown.
(1989) TheMaking of a Senator: Dan Quayle. Washington: CQ Press.
Ferejohn, John & Brian Gaines (1991) The Personal Vote in Canada, in Herman Bakvis
(ed.), Representation, Integration and Political Parties in Canada. Toronto: Dundurn
Press.
Fischer, Alastair (1978) The Random Sampling on Votes in the Australian Senate,
Economics Department Paper No. 7. University of Adelaide.
(1981) Aspects of the Voting System for the Senate, Politics. 16: 5762.
(1988) Electoral Distortion under STV Random Sampling Procedures: A Comment,
British Journal of Political Science. 18: 14243.
Fowler, L.L. & McClure, R.D. (1989) Political Ambition. New Haven: Yale University
Press.
Gallagher, Michael (1975) Disproportionality in a Proportional Representation System:
The Irish Experience, Political Studies23: 50113.
(1978) Party Solidarity, Exclusivity and I nter-Party Relationships in I reland
19221977: The Evidence of Transfers, Economicand Social Review. 10: 122.
(1987) Does Ireland Need a New Electoral System? Irish Political Studies2: 2748.
(1988) Ireland: The I ncreasing Role of the Centre, in Michael Gallagher & Michael
Marsh (eds), CandidateSelection in ComparativePerspective:TheSecret Garden of Politics.
London: Sage.
(1992) Comparing Proportional Representation Electoral Systems: Quotas,
| 198 |
Thresholds, Paradoxes and Majorities, British Journal of Political Science22.
(1999) The Results Analysed, in Michael Marsh & Paul Mitchell (eds), How Ireland
Voted 1997. Boulder: Westview Press.
Gallagher, Michael, Michael Laver & Peter Mair (2000) RepresentativeGovernment in
Modern Europe3rd ed. New York: McGraw Hill.
Michael & Paul Mitchell (eds) (2005) ThePolitics of Electoral Systems. Oxford: Oxford
University Press.
Michael & A. R. Unwin (1986) Electoral Distortion under STV Random Sampling
Procedures, British Journal of Political Science. 16: 24368.
Galligan, Brian & James Warden (1986) The Design of the Senate, in Gregory Craven
(ed.), TheConvention Debates, 18911898: Commentaries, Indices and Guide. Sydney:
Legal Books.
Goodin, Robert (ed.) (1996) TheTheory of Institutional Design. Cambridge: Cambridge
University Press.
Goot, Murray (1985) Electoral Systems, in Don Aitkin (ed.), Surveys of Australian Political
Science. Sydney: George Allen & Unwin.
Graham, B. D. (1968) The Choice of Voting Methods in Federal Politics, 19021918, in
Colin Hughes (ed.), Readings in Australian Government. St. Lucia: University of
Queensland Press.
Green, Antony (1999) Two Votes Can Get you on your Way to the Council, Sydney
Morning Herald. March 17: 9.
(2000) New South Wales Legislative Council Elections 1999, NSW Parliamentary
Library Research Service: Background Paper. No. 2.
(2003) Prospects for the 2003 Legislative Council Election, NSW Parliamentary
Library Research Service: Background Paper. No. 3/03.
(2003a) 2003 New South Wales Election: Preliminary Analysis, NSW Parliamentary
Library Research Service: Background Paper. No. 4/03.
Griffith, Gareth & Sharath Srinivasan (2001) State Upper Houses in Australia, NSW
Parliamentary Library Research Service: Background Paper. No. 1.
Grofman, Bernard (1999) Preface, in Bernard Grofman, Sung-Chull Lee, Edwin
Winckler, & Brian Woodall (eds), Elections in Japan, Korea, and Taiwan under theSingle
Non-TransferableVote: TheComparativeStudy of an Embedded Institution. Ann Arbor:
Sung-Chull Lee, Edwin Winckler, & Brian Woodall (eds) (1999) Elections in Japan,
Korea, and Taiwan under theSingleNon-TransferableVote: TheComparativeStudy of an
Embedded Institution. Ann Arbor: University of Michigan Press.
(1999a) I ntroduction, in Bernard Grofman, Sung-Chull Lee, Edwin Winckler, &
Brian Woodall (eds), Elections in Japan, Korea, and Taiwan under the Single Non-
Transferable Vote: The Comparative Study of an Embedded Institution. Ann Arbor:
Hallett, George (1984) Proportional Representation with the Single Transferable Vote: A
Basic Requirement for Legislative Elections, in Arend Lijphart & Bernard Grofman
(eds). Choosing an Electoral System: Issues and Alternatives. Westport: Praeger.
Hancock, W. Keith (1930) Australia. London: Ernst Benn.
(1947) Politics in Pitcairn: And Other Essays. London: Macmillan.
Hart, Jenifer (1992) Proportional Representation: Critics of the British Electoral System,
18201945. Oxford: Clarendon.
Haward, Marcus & James Warden (eds) (1995) An Australian Democrat:TheLife, Work, and
Consequences of Andrew Inglis Clark. Hobart: Centre for Tasmanian Studies.
Hill, I . (1994) Meek or Warren Counting, Voting Matters. 1: 1314.
Hirczy de Mino, Wolfgang & John Lane (2000) Malta: STV in a Two-Party System, in
Shaun Bowler & Bernard Grofman (eds), Elections in Australia, Ireland, and Malta
under theSingleTransferableVote: Reflections on an Embedded Institution. Ann Arbor:
Hogan, Michael (1995) The New South Wales State Election 1922, TheNew South Wales
| 199 |
| Bi bl i ography |
StateElection Series. New South Wales Parliamentary Library and Department of
Government, University of Sydney.
(2001) 1920, in Michael Hogan & David Clune (eds), ThePeoples Choice: Electoral
Politics in 20th Century New South Wales. Volume1: 19011927. Sydney: Parliament of
New South Wales and University of Sydney.
(2001a) 1922, in Michael Hogan & David Clune (eds), ThePeoples Choice: Electoral
(2001b) 1913, in Michael Hogan & David Clune (eds), ThePeoples Choice: Electoral
Horowitz, Donald (1991) A Democratic South Africa? Constitutional Engineering in a
Divided Society. Berkeley, CA: University of California Press.
(2002) Constitutional Design: Proposals versus Processes, in Andrew Reynolds (ed.),
TheArchitectureof Democracy. Oxford: Oxford University Press.
Hughes, Colin (1968) Compulsory Voting, in Colin A. Hughes (ed.), Readings in
Australian Government. St. Lucia: University of Queensland Press.
(1980) TheGovernment of Queensland. St. Lucia: University of Queensland Press.
(2000) STV in Australia, in Shaun Bowler & Bernard Grofman (eds), Elections in
Australia, Ireland and Malta under theSingleTransferableVote. Ann Arbor: University of
Michigan Press.
Humphries, David (1999) Stronger Vote, Weaker House: Libs Offer Deal, Sydney Morning
Herald. September 3: 5.
Humphreys, John (1911) Proportional Representation. London: Methuen.
I nternational I DEA (1997) Voter Turnout from1945 to 1997: A Global Report. Stockholm:
I nternational I DEA.
Jackman, Simon (1997) Non-Compulsory Voting in Australia? What the Surveys Can
(and Cant) Tell Us. Paper presented to the Annual Meetings of the American Political
Science Association, Washington DC, August.
Jackson, Sally & Trudy Harris (1999) Multi-party Politics Gets Microscope, The
Australian. March 20: 4.
Jaensch, Dean (1981) Electoral Reform, in Andrew Parkin & Allan Patience (eds), The
Dunstan Decade: Social Democracy at theStateLevel. Melbourne: Longman Cheshire.
(1983) TheAustralian Party System. Sydney: Allen & Unwin.
(1986) Getting Our Houses in Order: Australias Parliament, How it Works and theNeed
for Reform. Ringwood, Victoria: Penguin.
(1986a) Parliament and Government, in Dean Jaensch (ed.), TheFlinders History of
South Australia. Netley: Wakefield Press.
(1994) Power Politics: Australias Party System3rd ed. St. Leonards, NSW: Allen &
Unwin.
Jenkins, Roy (1998) TheReport of theCommission on theVoting System. Volume 1. London:
HMSO <www.official-documents.co.uk/document/cm40/4090/4090.htm>.
Jesse, Neal (1999) Candidate Success in Multi-Member Districts: An I nvestigation of
Duverger and Cox, Electoral Studies. 18: 32340.
(2000) A Sophisticated Voter Model of Preferential Electoral Systems, in Shaun
Bowler & Bernard Grofman (eds), Elections in Australia, Ireland and Malta under the
SingleTransferableVote: Reflections on an Embedded Institution. Ann Arbor: University
of Michigan Press.
Johnston, R. J. (1978) Friends-and-Neighbours Voting in Victoria: A Note, Politics13.
Karp, Jeffrey & Susan A, Banducci (2004) Political Efficacy and Participation in Twenty-
FiveDemocracies. How Electoral Systems ShapePolitical Behavior. Unpublished paper.
Katz, Richard (1980) A Theory of Parties and Electoral Systems. Baltimore, MD: Johns
Hopkins University Press.
(1984) The Single Transferable Vote and Proportional Representation, in Arend
Lijphart & Bernard Grofman (eds), Choosing an Electoral System:Issues and Alternatives.
| 200 |
New York: Praeger.
(1997) Democracy and Elections. Oxford: Oxford University Press.
(1997a) Representational Roles, European Journal of Political Research. 32: 21126.
(1999) Electoral Reform and its Discontents, British Elections and Parties Review9.
Katz, Richard & Peter Mair (1995) Changing Models of Party Organization and Party
Democracy: The Emergence of the Cartel Party, Party Politics. 1: 528.
Kavanagh, D. (1970) Constituency Electioneering in Britain. London: Longman.
Kelley, J. & I an McAllister (1984) Ballot Paper Cues and the Vote in Australia and Great
Britain: Alphabetic Voting, Sex, and Title, PublicOpinion Quarterly. 48: 45266.
Kuklinksi, James H. (1979) Representative-Constituency Linkages: a Review Article,
LegislativeStudies Quarterly. 4: 121140.
Laakso, M. & R. Taagepera (1979) Effective Number of Parties: A Measure With
Application to West Europe, ComparativePolitical Studies. 12: 327.
Lakeman, Enid (1974) How Democracies Vote: A Study of Majority and Proportional Electoral
Systems 4th ed. London: Faber & Faber.
(1982) Power to Elect:TheCasefor Proportional Representation. London: Heinemann.
Lancaster, Thomas D. & W. David Patterson (1990) Comparative Pork Barrel Politics:
Perceptions from the West German Bundestag. Comparative Political Studies 22:
458477.
Laver, Michael (2000) STV and the Politics of Coalition, in Shaun Bowler & Bernard
Grofman (eds), Elections in Australia, Ireland and Malta under theSingleTransferable
Vote: Reflections on an Embedded Institution. Ann Arbor: University of Michigan Press.
Lawson, Kay (ed.) (1980) Political Parties and Linkage: A ComparativePerspective. New
Haven: Yale University Press.
LeDuc, Lawrence, Richard Niemi & Pippa Norris (eds) (1996) Comparing Democracies.
Thousand Oaks: Sage.
Levin, Jonathan & Barry Nalebuff (1995) An I ntroduction to Vote-Counting Schemes,
Journal of EconomicPerspectives. 9: 326.
Lijphart, Arend (1986) Degrees of Proportionality of Proportional Representation
Formulas, in Bernard Grofman & Arend Lijphart (eds), Electoral Laws and Their
Political Consequences. New York: Agathon Press.
(1987) The Demise of the Last Westminster System? Comments on the Report of
New Zealands Royal Commission on the Electoral System, Electoral Studies. 6:
97103.
(1994) Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies,
19451990. Oxford: Oxford University Press.
(1997) The Difficult Science of Electoral Systems: A Commentary on the Critique
by Alberto Penads, Electoral Studies. 16: 7379.
(1999) Patterns of Democracy: Government Forms and Performance in Thirty-Six
Countries. New Haven: Yale University Press.
(2002) The Wave of Power-Sharing Democracy, in Andrew Reynolds (ed.), The
Architectureof Democracy: Constitutional Design, Conflict Management, and Democracy.
Oxford: Oxford University Press.
& Bernard Grofman (1984) Choosing an Electoral System, in Arend Lijphart &
Bernard Grofman (eds), Choosing an Electoral System:Issues and Alternatives. New York:
Praeger.
Rafael Lopez Pintor, Yasunori Sone (1986) The Limited Vote and the Single Non-
Transferable Vote: Lessons from the Japanese and Spanish Examples, in Bernard
Grofman & Arend Lijphart (eds), Electoral Laws and their Political Consequences. New
York: Agathon Press.
Linde, J. & J. Ekman (2003) Satisfaction with Democracy: A Note on a Frequently Used
I ndicator in Comparative Politics, European Journal of Political Research. 42: 391408.
Lipset, S. M. & S. Rokkan (1967) Cleavage Structures, Party Systems and Voter
Alignments: an I ntroduction, in S. M. Lipset & S. Rokkan (eds), Party Systems and
Voter Alignments. New York: Free Press.
| 201 |
| Bi bl i ography |
Loveday, P. (1977) The Federal Parties, in P. Loveday, A. W. Martin & R. S. Parker (eds).
TheEmergenceof theAustralian Party System. Sydney: Hale & Iremonger.
Mackerras, Malcom (1970) Preference Voting and the Donkey Vote, Politics. 5: 6976.
(1995) The Operation and Significance of the Hare-Clark System, in Marcus
Haward & James Warden (eds), An Australian Democrat: The Life, Work, and
Consequences of Andrew Inglis Clark. Hobart: Centre for Tasmanian Studies.
Magarey, Susan (1985) Unbridling theTongues of Women: A Biography of CatherineHelen
Spence. Sydney: Hale & Iremonger.
Mair, Peter (1987) TheChanging Irish Party System. London: Pinter.
& Michael Laver (1975) Proportionality, PR and STV in Ireland, Political Studies23:
491500.
Marsh, Michael (2000) Candidate Centered but Party Wrapped: Campaigning in Ireland
under STV, in Shaun Bowler & Bernard Grofman (eds), Elections in Australia, Ireland
and Malta under theSingleTransferableVote: Reflections on an Embedded Institution. Ann
Arbor: University of Michigan Press.
Massicotte, Louis & Andr Blais (1999) Mixed Electoral Systems: A Conceptual and
Empirical Survey, Electoral Studies. 18: 34166.
McAllister, I an (1985) Enrolment/Population Ratios. Canberra: Australian Electoral Office
Research Report No. 9.
(1986) Compulsory Voting, Turnout and Party Advantage in Australia. Politics21:
8993.
(1992) Political Behaviour: Citizens, Parties and Elites in Australia. Melbourne:
Longman Cheshire.
(1997) Australia, in Pippa Norris (ed.), Pathways to Power: LegislativeRecruitment in
Advanced Democracies. Cambridge: Cambridge University Press.
(2002) Political Parties in Australia: Party Stability in a Utilitarian Society, in Paul
Webb, David Farrell & I an Holliday (eds), Political Parties in Advanced Industrial
Democracies. Oxford: Oxford University Press.
& Malcolm Mackerras (1999) Compulsory Voting, Party Stability and Electoral
Advantage in Australia. Electoral Studies18: 21733.
Malcolm Mackerras & Carolyn Boldiston (1997) Australian Political Facts2nd ed.
Melbourne: Macmillan.
& Toni Makkai (1993) I nstitutions, Society or Protest? Explaining I nvalid Votes in
Australian Elections. Electoral Studies 12: 2340.
& Anthony Mughan (1986) Differential Turnout and Party Advantage in British
General Elections, 196483. Electoral Studies5: 14352.
McLean, I ain (1991) Forms of Representation and Systems of Voting, in David Held (ed.),
Political Theory Today. Cambridge: Polity Press.
(1996) E. J. Nanson. Social Choice and Electoral Reform, Australian Journal of
Political Science. 31: 36985.
McLean, I ain & F. Hewitt (eds) (1994) Condorcet: Foundations of Social Choiceand Political
Theory. Cheltenham: Edward Elgar.
McLean, I ain & Arnold Urken (1995) General I ntroduction, in I ain McLean and Arnold
Urken (eds), Classics of Social Choice. Ann Arbor: University of Michigan Press.
McLean, I ain, Alistair McMillan & Burt Monroe (eds) (1996) A Mathematical Approach to
Proportional Representation: Duncan Black on Lewis Carroll. Boston: Kluwer Academic
Publishers.
Meek, B. (1994) A New Approach to the Single Transferable Vote, Voting Matters. 1: 111.
Mercurio, Bryan & George Williams (2004) Australian Electoral Law: Free and Fair?,
Federal Law Review. 32: 131.
Miller, W. E. & M. K. Jennings (1986) Parties in Transition: A Longitudinal Study of Party
Elites and Party Supporters. New York: Russell Sage.
Miragliotta, N. (2002) Determining theResult: Transferring Surplus Votes in theWestern
Australian LegislativeCouncil. Perth: Western Australian Electoral Commission.
Mitchell, Paul (2000) Voters and their Representatives: Electoral I nstitutions and
| 202 |
Delegation in Parliamentary Democracies, European Journal of Political Research. 37:
33551.
Morris, Rachel (1999) How the Micro Parties Plot to Get Elected, TheDaily Telegraph.
March 25: 18.
Mozaffar, Shaheen (1997) Democratization, I nstitutional Choice and the Political
Management of Ethnic Conflict: Africa in Comparative Perspective, paper presented
at American Political Science Association annual meeting, Washington, DC.
Nanson, E. J. (1995 [1882]) Methods of Election, in I ain McLean & Arnold Urken (eds),
Classics of Social Choice. Ann Arbor: University of Michigan Press.
Newman, Terry (1992) Hare-Clark in Tasmania: Representation of All Opinions. Hobart:
Joint Library Committee of the Parliament of Tasmania.
Newton-Farrelly, Jenni (2000) The South Australian Legislative Council: Possible
Changes to its Electoral System, Parliamentary Library of South Australia, Information
Paper No. 17.
Norris, Pippa (1985) Womens Legislative Participation in Western Europe, West European
Politics. 8: 90101.
(1997) The Puzzle of Constituency Service, Journal of LegislativeStudies. 3: 2949.
(1999) I nstitutional Explanations for Political Support, in Pippa Norris (ed.), Critical
Citizens: Global Support for DemocraticGovernance. Oxford: Oxford University Press.
(2002) Ballots not Bullets: Testing Consociational Theories of Ethnic Conflict,
Electoral Systems, and Democratization, in Andrew Reynolds (ed.), TheArchitectureof
Democracy: Constitutional Design, Conflict Management, and Democracy. Oxford: Oxford
University Press.
(2004) Electoral Engineering: Voting Rules and Political Behavior. Cambridge:
Norton, Philip & David M. Wood (1990) Constituency Service by Members of
Parliament: Does I t Contribute to a Personal Vote? Parliamentary Affairs43: 196208.
(1993) Back fromWestminster: British Members of Parliament and their Constituents.
Lexington: University Press of Kentucky.
Nurmi, Hannu (1996/97) I t s not Just the Lack of Monotonicity, Representation. 34: 4852.
OConnell, Declan (1983) Proportional Representation and I ntra-Party Competition in
Tasmania and the Republic of Ireland, Journal of Commonwealth and Comparative
Politics. 21: 4570.
Oldham, R. C. (1911) Compulsory Enrolment: Paper Prepared by the Chief Electoral
Officer, Commonwealth of Australia Parliamentary Papers. Vol. 2, No. 27: 16162.
Ordeshook, Peter (1986) Game Theory and Political Theory. Cambridge: Cambridge
University Press.
Orr, Graeme, Bryan Mercurio & George Williams (eds) (2003) Realising Democracy:
Electoral Law in Australia. Sydney: Federation Press.
Overacker, L. (1952) TheAustralian Party System. New Haven, CT: Yale University Press.
Parker, G.R. (1986) Homeward Bound: Explaining Changes in Congressional Behavior.
Pittsburgh: University of Pittsburgh Press.
Penads, Alberto (1997) A Critique of Lijphart s Electoral Systems and Party Systems,
Phillips, Anne (1995) ThePolitics of Presence:ThePolitical Representation of Gender, Ethnicity
and Race. Oxford: Clarendon Press.
Pitkin, Hannah (1968) The Concept of Representation. Berkeley, CA: University of
California Press.
Putnam, Robert D. (2000) Bowling Alone:TheCollapseand Revival of American Community.
New York: Simon & Schuster.
Rae, Douglas (1967) The Political Consequences of Electoral Laws. New Haven: Yale
University Press.
Reeves, William Pemper (1902) StateExperiments in Australia and New Zealand. London:
Alexander Moring Ltd.
Reid, G. S. & Martyn Forrest (1989) Australias Commonwealth Parliament, 19011988:Ten
| 203 |
| Bi bl i ography |
Perspectives. Carlton: Melbourne University Press.
Reilly, Ben (1997) The Plant Report and the Supplementary Vote: Not So Unique After
All, Representation. 34: 95102.
(1997a) Preferential Voting and Political Engineering: A Comparative Study, Journal
and Commonwealth and ComparativePolitics. 35: 119.
(2001) Democracy in Divided Societies: Electoral Engineering for Conflict Management.
& Michael Maley (2000) The Single Transferable Vote and the Alternative Vote
Compared, in Shaun Bowler & Bernard Grofman (eds), Elections in Australia, Ireland
and Malta under theSingleTransferableVote: Reflections on an Embedded Institution. Ann
Arbor: University of Michigan Press.
& Andrew Reynolds (1999) Electoral Systems and Conflict in Divided Societies.
Washington, DC: National Academy Press.
Reynolds, Andrew (1999) Electoral Systems and Democratization in Southern Africa. Oxford:
Oxford University Press.
(ed.) (2002) TheArchitectureof Democracy. Oxford: Oxford University Press.
Reynolds, Andrew & Ben Reilly (1997) (eds) TheInternational IDEA Handbook of Electoral
System Design. Stockholm: I nternational I nstitute for Democracy and Electoral
Assistance.
Rhiannon, Lee (1999) Minor Parties Stand to Lose Out in Major Way, Sydney Morning
Herald. June 4: 15.
Rose, Richard (1992) What aretheEconomicConsequences of PR?London: Electoral Reform
Society.
(ed.) (2000) International Encyclopedia of Elections. London & New York: Macmillan
Palgrave.
Rothstein, Bo (1998) Political I nstitutions: An Overview, in Robert Goodin & Hans-
Dieter Klingeman (eds), A New Handbook of Political Science. Oxford: Oxford
University Press.
Rule, Wilma (1987) Electoral Systems, Contextual Factors, and Womens Opportunity for
Election in Twenty-Three Democracies, Western Political Quarterly. 40: 47798.
Ruostetsaari, I lkka & Mikko Mattila (2002) Candidate-Centred Campaigns and their
Effects in an Open List System: the Case of Finland, in David Farrell & Rdiger
Schmitt-Beck (eds), Do Political Campaigns Matter? Campaigns Effects in Elections and
Referendums. London & New York: Routledge.
Rydon, Joan (1968) Electoral Methods and the Australian Party System 19101951, in
Colin Hughes (ed.), Readings in Australian Government. St. Lucia: University of
Queensland Press.
(1986) A Federal Legislature: TheAustralian Commonwealth Parliament 19011980.
Saari, Donald (2001) ChaoticElections! A Mathematician Looks at Voting. Washington, DC:
American Mathematical Society.
(2001a) Decisions and Elections: Explaining theUnexpected. Cambridge: Cambridge
University Press.
Sacks, Paul M. (1977) TheDonegal Mafia: MachinePolitics in Ireland. New Haven, Conn:
Yale University Press
Sartori, Giovanni (1968) Political Development and Political Engineering, PublicPolicy.
17: 26198.
(1976) Parties and Party Systems: A Framework for Analysis. Cambridge: Cambridge
University Press.
(1997) ComparativeConstitutional Engineering: An Inquiry into Structures, Incentives
and Outcomes 2nd ed. London: Macmillan.
Sawer, Marian. (ed.) (2001) Elections Full, Freeand Fair. Sydney, NSW: The Federation
Press.
Schattschneider, E. E. (1942) Party Government. New York: Farrar & Rinehart.
Searing, D. (1994) Westminsters World. Cambridge, MA: Harvard University Press.
| 204 |
Sen, Amartya. (1995) How to Judge Voting Schemes, Journal of EconomicPerspectives. 9:
9198.
Sharman, Campbell (1982) Diversity, Constitutionalism and Proportional Representation,
in Michael James (ed.), The Constitutional Challenge: Essays on the Australian
Constitution, Constitutionalism and Parliamentary Practice. Canberra: Centre for
I ndependent Studies.
A. M. Sayers & N. Miragliotta (2002) Trading Party Preferences: The Australian
Experience of Preferential Voting, Electoral Studies 21:54360.
Shugart, Matthew (2001) Electoral Efficiency and the Move to Mixed-Member Systems,
& Martin Wattenberg (2001) (eds) Mixed-Member Electoral Systems:TheBest of Both
Worlds? Oxford: Oxford University Press.
Sinnott, Richard (1995) Irish Voters Decide: Voting Behaviour in Elections and Referendums
Since1918. Manchester: Manchester University Press.
Sisk, Timothy (1995) Democratization in South Africa: The Elusive Social Contract.
Princeton: Princeton University Press.
Spence, Catherine (1910) Catherine Helen Spence: An Autobiography. Adelaide: W. K.
Thomas & Co.
(1987 [1896]) What is Effective Voting and How is it to be Secured?, in Helen
Thomson (ed.), CatherineHelen Spence. St. Lucia: University of Queensland Press.
Stone, Bruce (2002) Bicameralism and Democracy: The Transformation of State Upper
Houses, Australian Journal of Political Science. 37: 26781.
Taagepera, Rein (1998) Effective Magnitude and Effective Threshold, Electoral Studies. 17:
393404.
(1998a) How Electoral Systems Matter for Democratization, Democratization. 5:
6891.
(2002) Designing Electoral Rules and Waiting for an Electoral System to Evolve, in
Andrew Reynolds (ed.), TheArchitectureof Democracy: Constitutional Design, Conflict
Management, and Democracy. Oxford: Oxford University Press.
& Matthew Shugart (1989) Seats and Votes: TheEffects and Determinants of Electoral
Systems. New Haven: Yale University Press.
Tideman, Nicolaus (1995) The Single Transferable Vote, Journal of EconomicPerspectives. 9:
2738.
& D. Richardson (2000) A Comparison of I mproved STV Methods, in S. Bowler &
B. Grofman (eds), Elections in Australia, Ireland, and Malta under theSingleTransferable
Vote. Ann Arbor: University of Michigan Press.
Townsley, W. A. (1956) The Electoral System and the Constituencies, in F. C. Green (ed.),
Tasmania: A Century of Responsible Government, 18561956. Hobart: Government
printer.
Turner, Ken (1985) The New Rules of the Game, in Ernie Chaples, Helen Nelson & Ken
Turner (eds), TheWran Model: Electoral Politics in New South Wales 1981 and 1984.
Uhr, John (1998) DeliberativeDemocracy in Australia: TheChanging Placeof Parliament.
(1999) Why We Chose Proportional Representation, in Marian Sawer & Sarah
Miskin (eds), Representation and Institutional Chance: 50 Years of Proportional
Representation in theSenate. Canberra: Department of the Senate.
(2000) Rules for Representation: Parliament and the Design of the Australian
Electoral System, Commonwealth of Australia, Department of the Parliamentary
Library, Research Paper No. 29.
(2002) Catherine Helen Spence and Australian Electoral Reform, Australian Political
Studies Association Jubilee Conference, ANU, Canberra, August.
Wanna, John (2003) Queensland, in Jeremy Moon & Campbell Sharman (eds), Australian
Politics and Government: TheCommonwealth, theStates and theTerritories. Cambridge:
| 205 |
| Bi bl i ography |
Ward, I an (1997) Party Organization and Membership Participation, in John Warhurst
(ed.), Keeping theBastards Honest:TheAustralian DemocratsFirst Twenty Years. Sydney:
Allen & Unwin.
Weaver, R. Kent & Bert A. Rockman (eds) (1993) Do Institutions Matter? Government
Capabilities in theUnited States and Abroad. Washington: Brookings.
Webb, Paul, David Farrell & I an Holliday (eds) (2003) Political Parties in Advanced
Industrial Democracies. Oxford: Oxford University Press.
Wood, David, & G. Young (1997) Comparing Constituency Activity by Junior Legislators
in Great Britain and Ireland, LegislativeStudies Quarterly. 22: 21732.
Wright, J. F. H. (1980) Mirror of the Nations Mind: Australias Electoral Experiments.
Sydney: Hale & Iremonger.
(1983/84) Changes in Australias Federal Electoral Laws in 1983, Representation. 24:
14.
| 206 |
Adams, John 102
Aitkin, Don 1, 1415, 142
alphabetic voting, seedonkey voting
alternative vote 2, seealsopreferential
voting
ballot papers 34, 131ff
ballot structure 2, 1112, 149ff
House of Representatives use of 36ff
origins of 17, 21ff, 52ff
politicians and 103ff
Senate use of 4041
STV and 79ff
variants of 53ff
Ames, Barry 103
Amnesty I nternational 57
Anderson, Christopher 148, 157
Andrae, Carl 24, 28, 58
Arrow, Kenneth 10
Australian ballot, seesecret ballot
Australian Candidate Study 19, 106ff
Australian Capital Territory 86, 131, 138,
171
list system in 70
PR systems in 18, 48, 60ff, 69, 72ff
Australian Democrats 76
Australian Election Study viii, 1516, 106
Australian Electoral Commission vii, 56,
64, 101, 12324, 129, 138
Australian Labor Party 13ff, 30, 66, 108ff,
179
optional preferential voting and 5556
preferential block voting and 4142
Bartholdi, J. 12, 127
Barton government 3031, 38
Barton, Sir Edmund 23, 30
Bawn, Kathleen 6
Bean, Clive 113
Belgium 11, 23, 124
Bennett, Scott vii, viii, 7, 27
Bentham, Jeremy 13
Black, David vii
Blair, electorate of 9293, 100, 172
Blais, Andre 79
Blau, Adrian 169
block vote 33, 143
Blondel, Jean 83
Bogdanor, Vernon 24, 33
Boman, Vena vii, viii
Bonner, Neville 64, 98
Borda Count 10
Bowler, Shaun viii, 2, 1011, 13, 24, 57,
81, 104, 121, 127, 133, 148, 150, 173
Brams, Steven 9, 10
Breen, Peter 136
Britain 110, 135, seealsoUnited Kingdom
electoral reform debates 9, 24ff, 33,
169
STV and 58ff
Burke, Edmund 111ff
Butler, David 24, 82, 110
by-elections 38, 39, 52, 68
Cain, Bruce 104, 111
Canada 52, 164
| 207 |
Index
Canberra 114
Carey, John 103
Carr Labor government 66
Carroll, Lewis 9
Carstairs, Andrew McLaren 23, 25
Carter, Elisabeth 148
Castles, Francis 1, 148
casual vacancies 6667
Chaples, Ernest 73
Chapple, Robin 98
Chifley Labor government 6, 21, 42
Christian, Andy vii
Chubb, Basil 58
Clark, I nglis 5, 6, 25ff, 42, 58ff, 168ff
Clarke, Virginia 92
Coakley, John 63
Cohan, A.S. 128
Collins, Hugh 14
Colomer, Josep 174
Colson, Malcolm 64, 94, 98
Commonwealth Electoral Act 1902 7, 16,
23, 29ff, 4344
Commonwealth Electoral Act 1983 56
Commonwealth Electoral Act 1998 56
Commonwealth Parliamentary Library vii
comparative study of electoral systems viii,
20, 149ff
compulsory preferences 19, 42, 54, 65,
9394, 122, 179
compulsory registration 123ff
compulsory voting 12, 14, 16, 19, 121ff
Condorcet 101
contingent vote 7, 25, 32, 49, 54, 91
Coram, Bruce 18, 69, 169
Corangamite, electorate of 38
Cox, Gary 12, 103, 126, 133, 150
Crewe, I vor 141
Crisp, L.F. 29, 35, 42, 110
Cyprus 124
Dalton, Russell 14, 16
Darcy, Robert 6062
DeGaris, Ren 71
Democratic Audit 17475
Democratic Labor Party 82
Denham Liberal government 124
Denmark 24, 28
dHondt systems 72, 87
Dickson, Guy vii
disproportionality, seeproportionality
district magnitude 33, 58, 59, 65, 78, 83ff,
103, 152ff
Dodgson, C.L. 9
donkey voting 6061, 131
Donovan, Todd 24
Doro, Gideon 9
Droop quota 26, 28, 31, 36, 42, 58, 70, 72,
74, 84
Dummett, Michael 9, 12, 41, 95, 100
Dunleavy, Patrick 13, 135
Dunstan, Don 6970, 73
Duverger, Maurice 103
elected representatives 104ff, 156ff
election campaigns 113ff
elections
1901 27, 3031
1903 1, 35, 37
1910 30, 37, 52, 69
1913 37, 123
1914 3638
1917 37, 143
1919 4041
1946 123
1949 81, 123
1955 81, 123
1958 81
1963 81
1966 81
1974 64, 132
1975 71, 76, 81
1977 81
1980 81
1984 62, 80, 144
1987 144
1993 56, 139
1998 5657, 100, 106, 144, 179
2001 81, 144, 179
2004 136ff, 176
electoral systems
design of 56, 8ff, 47ff, 166ff
functions of 6ff
in states 4750
voter choice in 10ff
voter satisfaction with 147ff
Emy, Hugh 11011
Engstrom, Richard 34
Estonia 57
European Elections Study 104
Family First 66
Farrell, David M. 3, 7, 8, 1011, 13, 24,
33, 52, 54, 70, 79, 80, 102ff, 121, 127,
148, 149, 153
Federal Convention 189798 38
Fenno, Richard 104, 115
Ferejohn, John 113
Ferranda, Rosa vii
| 208 |
Fiji 124
Finland 23
Fischer, Alistair vii, 63, 64, 94
Fischer, John 9899
Fishburn, Peter 9
Flinders Deal 6, 21, 3839
Ford, Jon 99
Forrest, Martyn 21, 26, 2829, 33, 35
Fowler, Linda 115
France 69
franchise 29, 52, 70, 122ff, seealsowomen
Free Trade Party 33
Gaines, Brian 113
Gallagher, Michael vii, 8, 11, 63, 80, 104,
121, 12728, 133, 174
Galligan, Brian 29
Gardner, Senator 39
Germany 6, 70, 155
Glynn, Patrick 3839
Goot, Murray vii, 4, 5, 7, 45, 69, 82, 125,
167
Graham, B.D. 5, 6, 21, 23, 31, 33, 3536,
45
Green, Antony viii, 66, 136
Gregory method 2728, 42, 44, 63ff, 94ff,
168, 172, 177
Gregory, J.B. 63
Griffith, Gareth viii, 66
Grofman, Bernard 2, 10, 13, 18, 57, 121,
127, 150, 170, 173
Groom, Littleton 34
Guillory, Christine 148, 156
Hall, Steele 70
Hallett, Brian vii
Hallett, George 10, 11
Hancock, W. Keith 14
Hanson, Pauline 56, 9294, 100, 127
Harbord, Justin vii
Hare, Thomas 24, 26ff, 36, 5859
Hare-Clark system
origins of 27, 5859
use of 6869
Harris, Trudy 66
Hart, Jenifer 9, 24, 58
Hartz, Louis 13
Haward, Marcus 26
Hawke government 43
High Court of Australia 56
Hill, I . 95
Hirczy-de-Mino, Wolfgang vii
Hobart 2728
Hogan, Michael 69
Horowitz, Donald 150
House of Representatives 21, 28ff
elections in 17
origins of electoral system 28ff
members of 106ff
how-to-vote cards 19, 122, 131ff
Hughes Labor government 21
Hughes, Colin vii, 139
Humphrys, John 11, 94
I celand 156
immigrants 61, 123, 172
I ndia 24
informal votes 4344, 143ff, 172, 176
I nternational I DEA 140
invalid votes, seeinformal votes
Ireland 2, 8, 1819, 24, 48, 52, 57ff,
6768, 79, 87ff, 9596, 100, 104, 13032,
172, 179
I srael 164, 169
I taly 6, 164
Jackman, Simon 113, 139
Jackson, Sally 66
Jaensch, Dean 14, 30, 62, 69, 70, 110,
113, 132
Japan 155, 164, 169
Jenkins, Roy 12, 121
Jennings, M. Kent 108
Jesse, Neal 13, 128
Jessop, T. vii
Johnston, R.J. 132
Joint Select Committee on Electoral
Reform 64, 168
Joint Standing Committee on Electoral
Matters 4344, 5657, 97, 139, 168
Katz, Richard 9, 58, 78, 79, 83, 84,
103104, 14950, 152, 164, 170
Kavanagh, Dennis 115
Korea 155
Kuklinski, James H. 104
Labor Party, seeAustralian Labor Party
Lakeman, Enid 11, 12, 33, 63, 103,121,
145
Lancaster, Thomas D. 113
Lane, John vii
Langer, Albert 5657
Launceston 2728
Laver, Michael 13, 128
Levin, Jonathan 10, 12, 79, 91
Liberal and Country League 69
Liberal Party 13ff, 30, 55, 56, 66, 9294,
| 209 |
| I ndex |
108ff, 139ff
Lijphart, Arend 2, 8, 24, 34, 57, 79, 80,
83ff, 102103, 14849, 156
Lipset, Seymour Martin 22
list systems 8, 45, 58, 6970, 73, 76, 84,
121, 152
Lithuania 155
London 91
Loveday, Peter 30, 3637
Lovenduski, Joni 108
Luxembourg 124
Lyne, William 34
Mackerras, Malcolm vii, 28, 6062, 138,
141
Magarey, Susan 26
Mair, Peter 84, 130, 170
Makkai, Toni 43
Maley, Michael vii, 55, 9495
Malta 2, 8, 1819, 24, 48, 57ff, 67, 79,
86ff, 100
Marsh, Michael 130, 152
Massachusetts 57
Massicotte, Louis 79
McAllister, I an 13, 14, 43, 55, 60, 106,
108, 111, 123, 138, 14142
McClure, R.D. 115
McLean, I ain 9, 17, 26, 28, 30, 32, 102,
149
Meek, B. 95, 101
Menzies, Sir Robert 21
Mercurio, Bryan 175
Mexico 155
Mill, John Stuart 9, 24, 26
Miller, Warren E. 108
minor parties 1314, 71ff, 118ff, 141
Miragliotta, Narelle vii
Miskin, Sarah vii
Mitchell, Paul 174
mixed member systems 58
modified dHondt, seedHondt systems
Moon, Jeremy vii
Morris, Rachel 66
Mughan, Anthony 141
Musidlak, Bogey vii
Nalebuff, Barry 10, 12, 79, 91
Nanson, Edward 5, 6, 7, 9, 17, 25ff, 36,
42, 4546, 58ff, 168ff
National Party 13, 38, 94, 113ff
Nationalist government 37
Nauru 124
Netherlands 156
Nevill, Mark 98
New South Wales
1999 election 6667, 177
electoral experiments in 5455, 6970
electoral systems in 23ff, 48ff, 62ff,
86ff, 174
parliamentary terms 15
ticket voting in 7, 6566, 13536
New Zealand 1, 6, 69, 9596, 123, 155,
164, 16768
Newman, Gerry vii, 27
Newman, Terry vii, viii
Newton-Farrelly, Jenni viii
Nohlen, Dieter 174
Norris, Pippa 103, 108, 148, 156
Northern Ireland 57, 172
Northern Territory 48, 52, 86, 138, 177
Norton, Phillip 104, 113
Norway 156
Nurmi, Hannu 9
OConnell, Declan 132
OConnor, Richard 3233
ONeill, Gerald 63
Oldham, R.C. 122
One Nation Party, seePauline Hansons
One Nation Party
optional preferences 3031, 38, 54ff,
5657, 72, 145
Ordeshook, Peter 126
Orlin, J. 12, 127
Orr, Graeme 4, 174
Outdoor Recreation Party 66
Overacker, Louise 1
Parker, G.R. 115
party identification 13ff
party leaders 111
Patterson, W. David 113
Pauline Hansons One Nation Party 56,
98, 116, 177
Phillips, Anne 103
Pitkin, Hannah 102103
Playford, Thomas 70
political culture 1314
political parties 104ff
compulsory voting and 122ff
and disproportionality 84, 89ff
and electoral rules 30ff
strength of 106108
postal voting 38, 123
preferential voting, seealsoalternative vote
block system 4041
institutional design 45, 173ff
order of candidates 12
| 210 |
social choice and 910
strategic voting in 1213, 122ff
voters use of 121ff, 133ff, 149ff
Proportional Representation of the UK 9,
24ff
Proportional Representation Society of
Australia vii, 44, 95, 101, 17475
proportionality 8ff, 18, 19, 28, 40, 59, 66,
74ff, 102ff, 149, 157, 159.
Protectionist Party 17, 30
Pulsford, Senator 36
Putnam, Robert 115
Queensland 4, 17
alternative vote in 2425, 3031
contingent vote in 7, 33
electoral experiments in 5, 5455, 76
electoral systems in 23ff, 48ff
Rae, Douglas 78, 81, 84
Reeves, William Pemper 25
referendums 6, 49, 73, 125, 145
Reid, Gordon S. 21, 26, 2829, 33, 35
Reilly, Ben 2, 7, 52, 54, 55, 82, 91, 9394,
150
representation 41, 58, 86, 103ff
Residents Rally 75
Reynolds, Andrew 5253, 148, 150
Robinson, Kirsten vii
Robson rotation 6263, 131, 17778
Rockman, Bert A. 147
Rokkan, Stein 22
Rose, Richard 148
Rule, Wilma 103
Rydon, Joan vii, 7, 21, 82
Saari, Donald 10, 91, 127
Sacks, Paul M. 58
Sartori, Giovanni 147, 150
Sawer, Marian vii, 1, 4
Schattschneider, E.E. 132
Schumpeter, Joseph 12
Scully, Roger 11, 104, 150
Searing, Donald 104
secret ballot 1
Sen, Amartya 10, 79, 91
Senate
1948 reforms 4243
1983 reforms 4446, 6364, 119, 131,
169
design of electoral system 23ff, 33ff,
40ff
elections in 17
electoral systems in 23ff
Gregory method in 100
members of 106ff
ticket voting in 129ff, 174
use of STV in 21ff, 43ff
Sharman, Campbell viii, 81, 82
Shugart, Matt 6, 8, 57, 58, 150ff
Simms, Marian vii
Singapore 124
single member plurality 10, 32, 49, 104,
121
differences with AV 7, 79ff
House of Representatives and 35, 139
origins of 22, 2728
states and 5556
single non-transferable vote 12627
single transferable vote 2, 11
alternative vote and 4142
ballot papers 23, 4
ballot structure 23, 150ff
debates in Britain over 2425
elected representatives and 104ff
proportionality of 1819
use in Senate 21ff, 132ff
use in states 49ff, 75ff
variations in 57ff, 74ff
Sinnott, Richard 128
Sisk, Timothy 150
Smith, Greg 98
social choice theory 9, 12, 17, 41, 78, 100
South Australia
electoral systems in 23ff, 44, 48ff,
7172, 7576
parliamentary terms 14
PR systems in 18 33
secret ballot 1
votes for women 1, 29
Spain 11, 156
Spence, Catherine Helen 5, 6, 17, 26ff,
36, 38, 42, 58ff, 168ff
spoiled votes, seeinformal votes
Sri Lanka 52
Srinivasan, Sharath 66
Stone, Bruce vii
strategic voting 1213, 122ff, 173
Studlar, Donley 111
Styles, Senator 36
Swan, electorate of 38
Sweden 23
Taagepera, Rein 8, 58, 79, 83, 147
Taiwan 155
Tasmania 17, 47, 86, 131, 138, 171
| 211 |
| I ndex |
electoral experiments in 5, 4243
Hare-Clark system and 27ff, 6869, 73
Legislative Council in 27
Thompson, Cameron 92
ticket voting 61, 6566, 7273, 122, 129ff,
171, 17475
Tideman, Nicolaus 9, 18
Townsley, W.A. 2728
Turner, Ken 73
turnout 123ff, seealsocompulsory voting
Uhr, John vii, 4, 5, 6, 21, 28, 29, 33, 36,
4243, 45
United Kingdom 115, 123, 164, 167, see
alsoBritain
electoral choice in 11, 104
United States 52, 58, 104, 108, 115, 144
Unwin, A. R. 63
Uruguay 124
values and beliefs 1314
Venezuela 169
Victoria 47
electoral systems in 23ff, 48ff
preferences 38
secret ballot 1
Victorian Farmers Union 38
vote splitting 55
voting age population 156ff
Wanna, John viii
Warden, James 26, 29
Ware, W.R. 32
Wattenberg, Martin 6, 14, 16, 57, 164
Weaver, R.Kent 147
Webb, Paul 106
Western Australia
block vote in 33
electoral systems in 23ff, 48ff, 174
preferences in 38
Western Australian Electoral Commission
vii
Westminster model 110
Wheare, K.C. 168
Whitlam government 43
Wilder, Paul vii
Williams, George 175
Wills, Roger vii
women
extension of vote to 1, 29
Wood, David 104, 113
Wran, Neville 73
Wright, Jack 4, 52, 64, 121, 131
Young, G. 104
| 212 |
Also published by UNSW Press
TERMS OF TRUST
Arguments over ethics in Australian government
John Uhr
Public trust in Australias governments and our political leadership is at an
all-time low, according to many opinion polls. Shaken by years of broken prom-
ises, political spin, shady deals, obfuscation, and even brazen lies, many con-
sider the condition terminal. According to John Uhr the situation is not
irretrievable, though serious, and immediate consideration needs to be given
to finding means by which our political leaders can re-earn the confidence of
the community.
In Terms of Trust John Uhr poses, and answers, fundamental questions about
our governments leadership:
How can we make political leadership compatible with ethical leadership?
How relevant is personal character to public life?
Why do we need to widen the scope of leadership to include all public
officials and not just those at the top?
How can citizens become more informed about the extent of government
trustworthiness?
With such questions in mind, Australians can now get down to business and
spell out the terms of trust compatible with the exceptional high hopes but
also the everyday low practices of our democracy.
Terms of Trust is the first book-length analysis of the role of ethics in Australian
government. It scrutinises what actually happens in practice against the demo-
cratic theory, and identifies the strengths and weaknesses of public-sector
ethics.
ISBN 0 86840 639 2
| 213 |
WESTMINSTER LEGACIES
Democracy and Responsible Government in Asia and the Pacific
Edited by Haig Patapan, John Wanna and Patrick Weller
Westminster Legacies examines the ways in which the Westminster system
has been influential in shaping responsible government and democracy across
Asia, Australasia and the Pacific. It devotes chapters to each of the following
countries: India, Pakistan, Nepal, Singapore, Malaysia, Australia, New
Zealand, Papua New Guinea, Fiji and the smaller Pacific island nations.
Westminster Legacies explores the way Westminster understandings of the
executive, bureaucracy, parliament and responsible government have been
influential in these countries home to such diverse histories, cultures and
traditions. It examines the ways the Westminster system has been adapted in
the light of local practices and traditions, and considers how Westminster
remains important for understanding political institutions and practices in
these countries.
It also looks at the conditions under which Westminster legacies have taken
root and endured, and those conditions that have eroded or significantly
changed its influence. Some of the countries Westminster Legacies surveys
have teetered on the edge of becoming failed states (especially in terms of
legitimate democracies), while others remain robust adversarial democracies.
ISBN 0 86840 848 4
| 214 |
YES, PREMIER
Labor Leadership in Australias States and Territories
Edited by John Wanna and Paul Williams
Since 2001 and for the first time since Federation all state and territory
governments in Australia have been held by the Australian Labor Party.
Yes, Premier features a chapter on each state or territory, and in each case
focuses on the leader who has led their party to power and so created this
unprecedented and historical situation. It examines each Labor premier and
territory chief ministers individual rise to power and their political and per-
sonal style.
The chapters are written by experts in the politics and political culture of
their home state. Part personal and part political biographies, the chapters also
assess the instruments of leadership any modern state or territory leader must
master in order to remain in the top job.
Students and media commentators will find this book invaluable.
Diane Carlyle and Nick Walker, Australian
Published at a time of the historic simultaneous dominance of all state and
territory governments by the Labor Party, this is a useful and readable addi-
tion to the literature on Australian state politics.
Paul Rodan, Australian Policy Online
Yes, Premier provides a good examination of the political leadership, charac-
teristics and context in which the leaders operate. Each of the contributors
looks at them as policy brokers, decision makers, crisis managers and election
campaigners.
Alan Fitzgerald, Independent Australian
ISBN 0 86840 840 9
| 215 |

U NS W P R E S S
UNSW
PRESS
The Australian Electoral System has
many strengths, chief among them
being the authors ability to combine
detailed analysis of Australias varied
electoral experience with the broader
issues raised by the international
literature on elections and electoral
behaviour. This publication makes a
major contribution to the study of
electoral politics and to Australian
political science.
G. CAMPBELL SHARMAN, DEPARTMENT OF
POLI TI CAL SCI ENCE, UNI VERSI TY OF BRI TI SH
COLUMBI A, VANCOUVER
The Australian Electoral System is an
outstanding exploration of the range
of electoral systems employed in
Australia. It benef its greatly from being
informed by the broader comparative
electoral systems literature and it is a
book that all those interested in electoral
systems generally will want to read.
MI CHAEL GALLAGHER, DEPARTMENT OF
POLI TI CAL SCI ENCE, TRI NI TY COLLEGE, DUBLI N
A key aim of this clearly written
and polemic-free book is to place
the origins, evolution and operation
of preferential election methods in
Australia in comparative context
This volume should be of great
interest to all students of elections and
representation, and to students of
comparative politics more generally
It will have special appeal to those
interested in electoral reform in
Australia.
BERNARD GROFMAN, PROFESSOR OF POLI TI CAL
SCI ENCE, UNI VERSI TY OF CALI FORNI A, I RVI NE
The Austral i an
El ectoral
System
or i gi ns,
var i at i ons
and
consequences
DAV I D M. F A R R E L L I A N Mc A L L I S T E R
FARRELL
McALLISTER
T
h
e

A
u
s
t
r
a
l
i
a
n

E
l
e
c
t
o
r
a
l

S
y
s
t
e
m
AustElecSystrover copy.indd 1 11/10/05 11:11:20 AM

Farrell

Transféré par

Informations du document

Description originale:

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Farrell

Transféré par

Droits d'auteur :

Formats disponibles

U NS W P R E S S

Vous aimerez peut-être aussi