SSE
i; .
. Lime

 

 

2..
.. ya} . .
.\\L\:5:Lu.m

.. .5.

 

5.
x y: .
‘ hﬁﬁmu. ma
n.n”..ur....i.3en ,
JV 09$! 1 ,
x. . I;

.rdﬂm
, “may 95.. ...L:u.m . .

”(ﬁg .
v ‘ . .
”sh sum: 4. . ,

 
 

 

 

ft
5!.»

*4
$.51

 

.1 we? r S.
x .. . L .
._...Xr.\...,..m..sv.uv.n
59.... .u .35.
$3.11....- 1 u
(1.: till... :3 7
‘4: ‘5q\ .2:
gig-7.x!
ti ‘A
.1,

:31? his?
5.251355.
(‘2; :I‘n

 

23.1.:
A): ‘s‘ .l

 

n .o‘n-‘Hi .h.
Efrﬁé

  

l‘l‘

 

ﬁrst}? _LIBRARY
200? Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

COPING WITH POLICY-MAKING COMPLEXITY:
ELECTORAL INSTITUTIONS, DIVERSITY, AND POLICY
PROBLEM-SOLVING

presented by

Salomon E. Orellana

has been accepted towards fulﬁllment
of the requirements for the

Doctoral degree in Political Science

 

 

(W M
Major Professor’s Signature

// //=a / 06008

Date

 

 

MSU is an Afﬁrmative Action/Equal Opportunity Employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5’08 K:IProj/Acc8Pres/ClRC/Date0ue.indd

 

COPING WITH POLICY-MAKING COMPLEXITY: ELECTORAL INSTITUTIONS,
DIVERSITY, AND POLICY PROBLEM-SOLVING

By

Salomon E. Orellana

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Political Science

2008

ABSTRACT

COPING WITH POLICY-MAKING COMPLEXITY: ELECTORAL INSTITUTIONS,
DIVERSITY, AND POLICY PROBLEM-SOLVING

By

Salomon E. Orellana

Political thinkers have long sought to identify institutions that help produce good
governance and this concern extends to recent research showing that electoral institutions
inﬂuence governmental performance on a wide range of policy outcomes. In particular,
countries with institutions such as proportional representation have been found to
outperform countries with maj oritarian institutions on a wide range of environmental,
social health, and governmental performance indicators. This dissertation contributes to
that line of thought by elaborating a framework that explains these results. I apply to the
political arena insights from an interdisciplinary body of literature that identiﬁes diverse
perspectives as a factor enhancing the problem-solving capacity of small groups. I argue
that electoral institutions inﬂuence the degree to which diverse perspectives are included
in the policy-making process of a country and that this inﬂuence produces consequences

important to a country’s ability to deal with a wide range of policy problems.

Copyright by
SALOMON E. ORELLANA
2008

This dissertation is dedicated to mi familia.

ACKNOWLEDGEMENTS

I would ﬁrst like to thank my family. This project could not have been completed
without their patience and support. I would also like to thank many people in my
department: my adviser, Carlos Pereira, who was extremely gracious with his time; Brian
Silver and Eric Chang, who provided detailed feedback on this project; the various
faculty members whose classes I attended and who thereby provided feedback on various
parts of this project; and the various graduate students who provided feedback, either in
class or at various presentations. I would also like to thank Stan Kaplowitz who served
as the Dean’s representative on my committee and offered several useful comments, and I
thank Burt Monroe, who worked extensively with me on Chapter 3 of this project.
Finally, I thank the political science department and Michigan State University for

providing funding and other forms of support as I completed this dissertation.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................ VII
LIST OF FIGURES ............................................................................. IX
CHAPTER 1 - INTRODUCTION ............................................................. 1
CHAPTER 2 - ELECTORAL INSTITUTIONS, DIVERSITY, AND POLICY
PROBLEM SOLVING ......................................................................... IO
Mechanisms of Diversity .............................................................. 11
Information and the Public ............................................................. l3
Electoral Institutions Inﬂuence Elite Diversity ..................................... 14
Electoral Institutions, Dissent, and Reactionary Policies .......................... 18

CHAPTER 3 - ELECTORAL INSTITUTIONS AND INFORMATION

DIVERSITY ...................................................................................... 21
Electoral Institutions and Political Information ..................................... 22
Party Systems and the Media .......................................................... 25
Empirical Investigation ................................................................. 27
The Distribution of Media Attention to Political Parties .......................... 28
The Distribution of Media Attention to Political Issues ........................... 31
The Distribution of Media Attention to Issue Positions ........................... 37
The Groupthink Dynamic ............................................................. 46
Discussion ............................................................................... 49
CHAPTER 4 - ELECTORAL INSTITUTIONS AND PUBLIC OPINION ............ 51
Electoral Institutions, Information Diversity, and Cosmopolitanism ............ 53
Electoral Institutions, Information Diversity, and Reactionary Preferences. . .58
Aggregate Level Evidence ............................................................. 59
Multilevel Analysis ..................................................................... 68
Conclusion ............................................................................... 77
CHAPTER 5 - ELECTORAL INSTITUTIONS AND POLICY OUTCOMES ........ 81
Electoral Institutions, Policy Innovation, and Reactionary Policies. . . . . . . . .....82
Measurement ............................................................................ 86
Results .................................................................................... 88
Discussion and Conclusion ............................................................ 92
CHAPTER 6 — CONCLUSION ................................................................ 95
Broader Implications .................................................................... 97
Loose Ends .............................................................................. 99
Causality ................................................................................ 101

vi

REFERENCES .................................................................................. 105

vii

LIST OF TABLES

Table 3.1 Votes, Seats, and Media Attention .................................................. 30
Table 3.2 Issue-Speciﬁc Mentions in Media and the House of Commons ................. 36
Table 3.3 Semi-sentences on law and order appearing in New Zealand Herald .......... 48
Table 4.1 Dependent Variables .................................................................. 60
Table 4.2 Electoral Systems and Public Preferences ......................................... 64
Table 4-3 Electoral Systems and Public Preferences .......................................... 65
Table 4.4 Multilevel Model of Cosmopolitanism ............................................. 76
Table 5.1 OLS Analysis of Environmental Performance and Event History

Analysis of Civil Union Legislation ............................................................. 90

Table 5.2 OLS Analysis of Incarceration Rates and Ordered Probit Analysis of

Capital Punishment ................................................................................ 92
Table 6.1 Women's Suffrage ................................................................... 104

viii

LIST OF FIGURES

Figure 3.1 Positions and Media Mentions, Taxation ...................................... 39—40
Figure 3.2 Positions and Media Mentions, Education ..................................... 42-43
Figure 3.3 Positions and Media Mentions, Law & Order ................................. 44-45
Figure 6.1 Relationship Between Insitutions, Public Opinion, and Policy ............... 101

Chapter 1 Introduction

Political thinkers have long sought to identify institutions that help produce good
governance. F amously, Plato promoted a government run by ascetic philosophers;
Aristotle favored a hybrid of oligarchy and democracy; Hobbes recommended that power
should be centralized (particularly in the hands of a benevolent monarch); and
Montesquieu supported the decentralization of power with checks and balances that
would stiﬂe ambition. Many more examples exist. This dissertation project is inﬂuenced
by this framework, by this search for political institutions that contribute to good
governance.

A useful ﬁrst step is to identify the problem that institutions are meant to address
and here I deﬁne the problem in terms of its process. I ask in this project what a society
can do to improve its decision-making capacity, and so I seek to identify the institutions
that enhance a state’s ability to deal with the related problems of human cognitive
limitation, complexity, uncertainty, and the difficulty of prediction. To a great extent,
this project is inﬂuenced by Karl Popper’s effort to identify political institutions that
promote good governance. In particular, Popper focused on the performance of
democracy relative to authoritarianism. Popper described the cognitive problem faced by
humans, especially as decision-makers, and how democracy improves the acquisition of

knowledge:

The position here adopted is very different from the popular . . . view of reason as
a kind of ‘faculty’, which may be possessed and developed by different men in
vastly different degrees. . . . According to our view, however, we not only owe
our reason to others, but we can never excel others in our reasonableness in a way
that would establish a claim to authority; authoritarianism and rationalism in our
sense cannot be reconciled, since argument, which includes criticism, and the art
of listening to criticism, is the basis of reasonableness. . . . Reason, like science,
grows by way of mutual criticism; the only possible way of ‘planning’ its growth
is to develop those institutions that safeguard the freedom of criticism, that is to

say, the freedom of thought. (Popper 1971, 226-227)

Thus, Popper was advocating for institutions that help society deal with the
complexity and uncertainty associated with the world and the limitations associated with
the human mind. Institutions should enhance openness and the tolerance of criticism
because these mechanisms permit humans to reach higher levels of objectivity and
provide the best opportunity for approximating truth (1971, 225). Openness ensures that
others can examine an individual’s claims about the way the world works and this
mechanism creates incentives for individuals to pursue objectivity. Popper also believed
the openness is a part of a piecemeal learning process; individual claims make piecemeal
contributions to the community’s understanding of the world, because the community can
reject a claim, accept it, or develop it. Popper is well known for applying this logic to the
philosophy of science, but he also extended this idea to the political realm, claiming that

democracies have an advantage over authoritarian regimes, with respect to the

2

development of knowledge, because they institutionalize openness and the tolerance of
criticism.

In recent political science work, we can see a related preoccupation with
governance. Like Popper, some of this work focuses on the relationship between
democracy and governance. For example, Przeworski et al (2000) showed that electoral
democracies achieve higher levels of well-being for their citizens than do authoritarian
regimes. Moreover, various branches of institutional research have focused on particular
institutions and particular mechanisms by which these institutions affect governmental
performance. An important example of this approach focuses on the effects of
presidentialism on the democratic performance. Much of this research is concerned with
how presidentialism and parliamentarism affect democratic stability, but this research
also touches on how these regimes affect democratic quality (e. g., Linz and Valenzuela
1994; Stepan and Skach 1993; Shugart and Carey 1992). Stepan and Skach, for example,
used Vanhanen’s Index of Democratization (ID) and his Index of Power Resources (IPR)
to categorize countries as democratic overachievers or as democratic underachievers, and
they found that presidential systems had a democratic underachiever rate 3.4 times
greater than parliamentary systems, while parliamentary systems were 1.8 times more
likely than presidential systems to be democratic overachievers.l

A more direct connection between institutions and the performance of

governments comes from research on electoral systems. Decades of research shows that

 

' The ID is based on total percentage of vote received by all parties except largest, and the
percentage of population that votes. The [PR is based on degree of decentralization of
nonagricultural resources, percentages of total agricultural land owned as family farms, and the
percentages of the population in universities, in cities, that is literate, and that is not employed in
agriculture.

3

electoral systems inﬂuence many areas of politics. The nucleus of this research has
established that electoral institutions inﬂuence the character and nature of party systems.
They inﬂuence the number and type of parties that compete and succeed in a democracy
(Cox 1997; Duverger 1954; Lijphart 1994; Rae 1971; Taagepera and Shugart 1989); and
they are argued to have indirect effects on virtually every other important characteristic
of a democracy, including: minority representation (Bowler et al. 2003), the success of
interest groups and social movements (Heidenheimer 1973; Kitschelt 1986),
responsiveness to public opinion (Powell 2000), legitimacy (Banducci et al. 1999;
Banducci and Karp 1999), conﬂict mitigation (Lijphart 1969), and governability (Pereira
and Orellana 2008).

We are also learning that electoral institutions inﬂuence a broad array of policy
outcomes. Electoral institutions, for example, have been credited with inﬂuencing wealth
redistribution (Huber and Stephens 2001; Alesina and Glaeser 2004), environmental
performance (F redriksson and Millimet 2004), economic policy (Persson and Tabellini
2003) and corruption (Chang and Golden 2006). Indeed, based on the kinds of ﬁndings
produced by this research it has been suggested that countries with institutions such as
proportional representation tend to outperform countries with majoritarian institutions on
a wide range of indicators of performance. Lijphart (1999), for example, suggested that
little separates consensual and majoritarian systems in terms of economic performance,
but consensual systems signiﬁcantly outperform majoritarian systems in terms of
“kinder-gentler” political outcomes. These outcomes include democratic quality
(measured in several ways), social welfare, protection of the environment, criminal

justice. and foreign aid. Gerring et al. (2005) argued, further, that “centripetal”

4

institutions (which include parliamentarism and proportional representation) are
associated with higher achievement in several areas of governance, including political,
economic, and human development.

A key mechanism identiﬁed by these studies is the degree of “collegiality” found
in a legislature—consensual consensual systems, in general, are assumed to increase
collegiality. Little has actually been said, however, about what this collegiality is and
how it is linked to the quality of policies produced by governments. The goals of this
dissertation are to identify concrete mechanisms by which electoral institutions affect
government performance, and in doing so, to provide an overarching framework that
explains a broad range of the many relationships that have already been identiﬁed
between electoral institutions and policy outcomes.

This endeavor requires a move away from (or the re-deﬁnition of) the concept of
collegiality, toward a focus on the concept of diversity. This is a concept consonant with.
Karl Popper’s defense of democracy and openness. That is, we can View diversity as a
form of openness, and as such, it should lead to a greater ﬂow of information. A crude
but useful way to think about the independent effects of diversity is to imagine that a
country may employ democratic electoral processes in its decision-making, but if it is
perfectly lacking in diverse agents (e. g., if everyone in this society thinks the same way),
then the level of criticism and information it generates may be as limited as what is found
under more authoritarian regimes. On the other hand, a democracy with a high degree of
agent and discourse diversity can be seen, in this instance, as approximating a

maximization of openness.

Indeed, Przeworski and his colleagues were aware of the interplay between
institutions and information. They speculated on the possibility that democracies
enhance the free-ﬂow of information and hence can improve the quality of economic
decisions (Przeworski 2000, 144), but they did not explore the speciﬁc mechanisms that
might be involved in this process. Such information mechanisms remain largely
unexplored, perhaps because of the unruly nature of the data required. Nevertheless,
exploring the impact of information on the quality of policy-making may be facilitated by
shifting the focus from the differences in information ﬂows between democracies and
non-democracies to the differences in information ﬂows that exist between democracies,
and more speciﬁcally. with the differences in “openness” (i.e., diversity) effectuated by
electoral institutions.

The idea that diversity can affect problem solving endeavors has received ample
attention from several ﬁelds (including economics, psychology, organization theory, and
computer science). Varied perspectives, knowledge, skills, and approaches have been
found to enhance a group’s capacity for problem solving, by several mechanisms. Most
of this research is focused on small groups, and accordingly, this project is in large part
an attempt to extend the insights provided by this small-group literature to the larger
areas of politics.

The key questions this study seeks to address, then, are the following. First, do
electoral institutions signiﬁcantly affect the diversity of perspectives included in the
policy—making process? And, second, does this diversity affect a country’s “problem-
solving” capacity in noticeable ways? In this dissertation it is argued that the answer to

these questions is yes. Electoral institutions do signiﬁcantly shape the diversity of

6

perspectives included in the policy-making process and this diversity produces two
concrete mechanisms that affect a society’s capacity for policy-related problem-solving.
In short, through their inﬂuence on the diversity of the legislature and the
executive, electoral institutions ﬁrst inﬂuence the diversity of the political discourse
produced by elites and that captured by the media. This mechanism inﬂuences how
quickly political issues are debated and addressed. Via a second mechanism, majoritarian
systems tend to produce a style of political competition that privileges short-term
solutions to problems. This mechanism produces more aggressive, short-term policy
responses to problems. Both of these mechanisms produce important consequences for

public opinion and policy outcomes.

Plan of Book
In Chapter 2, I elaborate a theory explaining how electoral institutions affect the diversity

of the political information environment and consequently inﬂuence public learning and
policymaking. I use small-group and public opinion literature to derive three
mechanisms by which ideological diversity can inﬂuence policy innovation and long-
terrn planning.

In Chapter 3, I provide empirical evidence showing how electoral institutions
inﬂuence political information and media coverage. I use data from legislative speeches,
the Comparative Manifestos Project (Klingemann et al. 2006), and perform a content
analysis of New Zealand’s largest newspaper, the New Zealand Herald, to provide
empirical evidence showing that electoral institutions do inﬂuence political debate and
the information environment. New Zealand’s experience with major electoral reform (a
move from a majoritarian to a proportional representation system) increased the diversity

7

of perspectives expressed in the political system and that covered by the media.
Signiﬁcantly more coverage was given to minor parties and many of the “marginalized”
issues they promoted. The diversity of solutions offered for several policy problems also
increased. I also show that that on issues related to personal security (such as law and
order) New Zealand’s political discourse under the Westminster electoral system suffers
from groupthink and can be described as more aggressive than the political discourse
emerging after New Zealand’s adoption of a more proportional electoral system.

In Chapter 4, I investigate how the information effects produced by electoral
institutions inﬂuence public learning. If electoral institutions affect the kind of
information that citizens can access, then we should expect them to inﬂuence citizen
preferences on various issues. I found this is the case especially for what I call
“sensitive” issues. In particular, minor parties in more permissive electoral institutions
(e. g., proportional representation) have more space to introduce controversial or sensitive
issues into the political discourse. This means that issues such as same-sex rights and
environmental protection will be debated earlier in countries with electoral institutions
that permit a more diverse political discourse. Consequently, citizens in countries with
parliamentary, proportional electoral systems tend to show more tolerance toward
sensitive issues such as homosexuality, abortion, and euthanasia, and they tend to be
more supportive of protecting the environment. Furthermore, because majoritarian
systems limit dissent, relative to more permissive systems, the public is more likely to
favor aggressive, short-term responses to policy problems related to personal security

(such as crime).

In Chapter 5, I investigate how the information mechanisms inﬂuence policy
problem-solving. First I consider how electoral institutions can inﬂuence policy
innovation, generally through their inﬂuence on how early certain issues can be debated.
If electoral institutions affect a society’s ﬂow of information, and thereby the kind of
information that elites have access to, and if they affect the preferences of citizens, then
we should expect them to inﬂuence how quickly governments take action on certain
issues. I show that countries with electoral institutions that permit the earlier discussion
of sensitive issues take earlier action on those issues. For example, countries with
proportional representation systems tend to act earlier on issues such as environmental
protection and the adoption of civil union legislation.

A second policy effect I consider in this chapter is how limitations on dissent
encourage the advancement of short-term solutions, especially on issues related to
personal security. I show that the mechanism of groupthink helps explain why countries
with majoritarian institutions are more likely to use more aggressive policies to deal with
problems in the areas of law and order and foreign affairs.

In Chapter 6, I conclude by arguing that my ﬁndings help explain the
relationships uncovered by past research and they generate important implications. For
example, I ask if the evidence presented here suggests that some countries have
competitive advantages over others as a result of the mechanisms described in this study.
I also discuss a variety of points that should be addressed by future work, including how
diversity may not always be a positive inﬂuence and how different kinds of diversity can

produce differing results.

Chapter 2 Electoral Institutions, Diversity, and Policy Problem
Solving

When attempting to solve problems, humans face a number of limitations. We suffer
from numerous biases and cognitive constraints (Nisbett and Ross 1980; Holland et al.
1989); moreover, we often face difﬁcult decision-making situations, i.e., multi-
dimensional and highly nonlinear problems (Senge 2006). In this study, I claim that
electoral institutions affect a society’s capacity for problem—solving by affecting the
diversity of perspectives that are included at various stages of the policy-making process.
Many of the insights for this claim come from interdisciplinary research (from ﬁelds
including economics, psychology, organization theory, and computer science) suggesting
that diversity provides problem-solving beneﬁts to small groups. In these ﬁelds,
heterogeneity in perspectives, knowledge, skills, and approaches has been found to
enhance a group’s capacity for problem solving, through a variety of mechanisms.

This study represents an effort to extend the insight of such ﬁndings to the
political arena, by exploring whether countries with electoral institutions that more
readily permit the expression of diverse points of view (e. g., systems with proportional
representation) possess a problem-solving advantage over countries with electoral
systems that limit the expression of diversity (i.e., majoritarian systems). Because of
time constraints 1 narrow my focus to only a few mechanisms by which diversity

generates positive consequences for state policy-making.2 I focus on two mechanisms

 

2 Indeed, there are other mechanisms of diversity (that produce both positive and negative
consequences) which will have to be explored in future efforts. I discuss some of these in greater
detail in the concluding chapter.

10

discussed extensively in the small-group literatures, and on a third mechanism that can be
deduced from public opinion research. These mechanisms can be considered to overlap,
but it should also become clear that each of them helps to directly explain particular

features of political debate and policy making.

Mechanisms of Diversity
The ﬁrst mechanism suggests that the inclusion of broader range of perspectives in the

decision making process takes advantage of human variation in interpretation and
heuristics. In this view, diversity increases the probability that a person with the right
perspective to solve a particular problem will become available in a given decision-
making situation (Page 2007). This mechanism implies that breakthroughs and
innovations are the result of an individual in a group applying a novel perspective to a
problem. The power of this mechanism is exempliﬁed by studies such as Hong and Page
(2004) and Page (2007), which used simulations to demonstrate various conditions under
which a functionally diverse group with less ability will outperform a homogeneous
group with higher ability.

A second mechanism relevant to problem-solving endeavors is derived from the
expectation that diversity inﬂuences the nature of dissent in a group or society. It is
assumed that compliance and conformity are more likely to occur in homogeneous rather
than heterogeneous groups, or conversely, that heterogeneous groups should experience
higher levels of dissent from minorities (De Dreu and West 2001; Hoffman and Maier
1961). Although cohesiveness and conformity in small groups can facilitate coordination
and task performance, much research in psychology focuses on their several downsides.
Janis (1972), for example, observed that conformity pressures and extreme concurrence

11

seeking may lead to defective decision making with sometimes disastrous consequences.
Strains to uniformity, or “Groupthink”, were found to cause the adoption of a preferred
solution without adequate consideration of information or alternatives. Similarly,
Hackman and Morris (1975) argued that one reason why groups sometimes fail to
outperform individuals is their premature movement to consensus—with dissenting
opinions being suppressed or dismissed.

Much research focuses on mechanisms that can mitigate groupthink, and here I
focus on a mechanism that is particularly relevant to the political arena: the mechanism of
dissent. Dissent can mitigate groupthink in several ways (Nemeth and Nemeth-Brown
2003). As famously documented by researchers such as Solomon Asch (1955, 1956)
dissent can diminish the negative inﬂuences of social pressures. For example, large
proportions of participants in Asch’s studies chose incorrect answers to easy questions
when the rest of the group unanimously chose the wrong answer. The presence of even
one dissenter reduced conformity to less than 10%. Dissent also stimulates thinking
about issues or problems from multiple perspectives (Nemeth 1986). Minority
viewpoints, even when wrong, can stimulate a search for information by creating a
tension that the majority frequently wants to resolve, generally by taking the minority
members’ perspectives into consideration. This act tends to stimulate reappraisal of the
situation and consideration of more features of the situation.

Small-group studies tend to ﬁnd that these ﬁrst two mechanisms increase creative
thought and solutions. Group homogeneity stimulates more conventional thought while
group heterogeneity and exposure to dissent stimulates more novel thinking. In

particular, people exposed to dissent develop ideas that are statistically infrequent. I

12

argue that these ﬁrst two mechanisms are largely translatable to the broader political
arena. They should operate in the decision-making processes undertaken by government

elites.

Information and the Public
1 also identify a third, citizen-level mechanism linked to diversity. We should expect that

the information produced by political systems inﬂuences the political sophistication of
the public. In general, political systems with diverse sources of information should
increase the amount of information available to citizens and this should be beneﬁcial to
voters, especially in their comprehension of issues and their development of preferences.
I focus on public Opinion in this study because much research suggests that public
opinion produces important effects on policy-making (see, for example, seminal studies
by Page and Shapiro (1983) and Monroe (1983)).

Experimental studies provide solid evidence for the claim that access to
information beneﬁts individuals. One particularly interesting and relevant study comes
from Druckman (2004) who found that framing effects (the degree to which an
individual’s preferences can be altered simply by asking a question in a particular
manner) can be mitigated by contextual factors such as elite competition and
heterogeneous group deliberation. That is, individuals will tend to be less affected by
framing if they are exposed to “counter-framing” by opposing elites, or if they engage in
interpersonal debates that introduce alternative perspectives. At the mass-level, Gordon
and Segura (1997) argued and demonstrated that because electoral institutions shape the

availability, clarity, and usefulness of political information, citizens in multiparty

13

countries tend to outperform citizens in countries with fewer parties on indicators of
political knowledge.

The key questions I seek to answer in this study, then, are the following. First, do
electoral institutions signiﬁcantly affect the diversity of perspectives included in the
policy-making process? And, second, does this diversity produce some of the

3 66

mechanisms found to occur in small groups in ways that affect a country 5 problem-
solving” capacity? In this dissertation it is argued that the answer to these questions is
yes. Through their inﬂuence on the diversity of the legislature and the executive,
electoral institutions ﬁrst inﬂuence the diversity of the political discourse produced by
elites and that captured by the media. This mechanism inﬂuences how quickly political
issues are debated and addressed. Via a second mechanism, the limitation of dissent,
majoritarian systems tend to produce groupthink, and this produces a style of political
competition that privileges short-temt/suwival-oriented solutions to problems. This
mechanism produces more aggressive, short-term policy responses to problems. Both of

these mechanisms produce important consequences for public opinion and policy

OUICOITICS.

Electoral Institutions Inﬂuence Elite Diversity
The question of whether electoral institutions inﬂuence the diversity of perspectives

included in the policy-making process is answered by perhaps one of the most extensive
bodies of literature. Decades of research conﬁrm that electoral institutions impact the
amount of diversity that is expressed at various stages of the policy-making process,
including 1) the campaign process, 2) the legislature, 3) the executive, and ultimately 4)
within the actual society.

14

Electoral institutions are expected to inﬂuence the information produced by the
political system, ﬁrst, by inﬂuencing the heterogeneity of party systems. This is, of
course, the Duvergerian theme at the core of electoral studies. Countries with institutions
such as proportional representation (PR) tend to have more parties contesting elections
(Katz 1997); they also tend have more parties winning seats in the legislature (Duverger
1954; Cox 1997; Lijphart 1994; Rae 1971); and more parties in the legislature also
ensures that more issue dimensions will be contested (Lijphart 1999). As noted by
Powell (2000), PR is also associated with committee rules facilitating opposition
inﬂuence in the legislature, and executives in PR systems tend to include more parties,
through coalition govemments (Blais and Carty 1987). Consequently, we should expect
proportional systems to increase the diversity of issues discussed in a political system and
also the various solutions offered for given policy problems.

A second set of expectations on the way in which electoral systems can inﬂuence
information processes follow early assertions by both Duverger (1954) and Downs
(1957) and have been elaborated from a variety of theoretical perspectives (e.g., Cox
1990; lnglehart 1990: p. 270-2). It is generally expected that majoritarian/pluralitarian
electoral systems produce greater centripetal incentives that encourage politicians and
parties to gravitate toward the median voter’s preferences and this tendency should limit
the range of political discourse.

According to Downs, majoritarian institutions should also encourage parties to
adopt “catch-all” strategies, where they try to appeal to the greatest number of voters by
employing tactics such as ambiguity and emphasizing popular issues and positions while

avoiding those that are controversial. Thus, two-party systems are expected to provide

15

less policy information and encourage voters to base evaluations of parties on non-
ideological/non-policy factors, such as the horse race — who is winning or moving in the
latest poll —— and the personal characteristics of the candidates.

These incentives contrast sharply with the promulgation incentives produced by
multiparty systems. Parties operating under proportional representation do not have to
seek voters by gravitating to the median voter’s position, but instead can win seats by
capturing enough support to overcome a relatively small electoral threshold. In this case,
parties have a greater incentive to distinguish themselves from one another and they will
therefore promulgate more distinct policy programs. Again, we should not expect that
parties in majoritarian systems will always avoid discussing sensitive issues, but we
should expect that parties in proportional systems (especially minor parties) will be
relatively freer to explore sensitive issues.

There are claims that two party systems are not particularly convergent (Poole and
Rosenthal 1984; Grofman 2004; Rabinowitz and Macdonald 1989) and multiparty
systems may not be particularly divergent (Lin et al. 1999). In general, however, we
should expect that two party systems are more convergent than multiparty systems. In a
study that compared levels of convergence and divergence across electoral systems rather
than in a single system—as virtually all other studies have done—Dow (2001) found that
parties and candidates in the majoritarian/pluralitarian systems are located signiﬁcantly
closer to the center than parties and candidates in proportional systems. Thus, while
parties in majoritarian systems may not gravitate exactly to the center, Dow’s study
suggests that the breadth of political discourse will, nevertheless, tend to be wider in

proportional systems than in pluralitarian systems.

16

The diversity produced by electoral institutions also reshapes the diversity found
in the society. Downs suggested that “in a plurality structure, since a two-party system is
encouraged and the two parties usually converge, voters’ tastes may become relatively
homogeneous in the long run; whereas the opposite effect may occur in a proportional
representation structure” (1957, 125). Evidence for this effect comes from Norris’ (2004,
1 19-121) ﬁnding that cleavage politics are stronger under PR than under majoritarian
systems. That is, majoritarian systems produce more centrist voting patterns than
proportional systems.

Thus, based upon the above dynamics, it is expected that electoral institutions will
particularly affect public preferences on “sensitive” issues and “sensitive” policy
positions. Formally, my conceptualization of sensitivity involves two concerns. First,
sensitive issues and sensitive policy positions are those which politicians ﬁnd more
difﬁcult to discuss. Many of these kinds of issues and positions can only be introduced
by minor parties. Second, sensitivity is a time-dependent concept. An issue that is
sensitive at a particular point in time is one that could be ordinary in another. For
example, an issue like racial equality was once a sensitive issue in most countries and
today it is largely taken for granted that discrimination based on race should be illegal.
More recently, an issue like homosexuality appears to be undergoing the transition from
sensitive issue to ordinary.

It should be emphasized that I am not suggesting that this evolution of issues does
not occur in majoritarian systems. I am suggesting, rather, that the evolution is more
rapid under electoral institutions that facilitate the expression of diverse ideological

perspectives. All democratic systems permit the ﬂow of information, but electoral

17

institutions such as proportional representation should facilitate the discussion of
sensitive issues and the expression of sensitive policy positions on various issues, enough
to affect public consideration of such issues and positions, and enough that electoral
institutions can help explain the variation in attitudes towards sensitive issues that exists

across countries.

Electoral Institutions, Dissent, and Reactionary Policies
Not only should we ﬁnd a greater degree of diversity under PR systems, but I argue

further that majoritarian/pluralitarian systems are more limiting of dissent and are
therefore more likely to suffer from groupthink. For many issues, political competition
encourages the largest parties to appeal to survival-oriented motives of voters and this
approach often engenders a reactionary style of policy making, especially for issues
linked to personal security. Dissent can mitigate this dynamic, but in majoritarian
systems dissent is limited and marginalized.

If we assume that survival is at the core of human motivation (Maslow 1943),
then it is reasonable to expect that parties and candidates in all political systems have an
incentive to compete for support by exploiting survival-oriented motivations.
Majoritarian systems, however, tend to exacerbate this phenomenon because they limit
dissent. Winning a plurality of votes requires that parties in majoritarian systems avoid
sensitive topics and solutions, and also that they emphasize “safe” (valence) issues and
solutions. Perhaps the easiest place to look for safe issues and solutions is along a
survival-security dimension. In majoritarian systems, moreover, the kind of political
debate that tends to emerge from this emphasis on survival often takes on a “one-
upsmanship” dynamic, especially in issue areas such as law and order and economic

18

security. For example, candidates may seek to emphasize their “tough-on-crime”
credentials and suggest more and more punitive measures for dealing with crime, they
may attack one another for not doing enough to keep gas prices down, or they may attack
each other for suggesting that taxes may need to be raised.

Maslow (1943) suggested that abundance and security will decrease the
monopolization of consciousness by survival needs, and here it is suggested that diversity
and dissent also produce mitigating effects. Under most decision-making situations, it is
perhaps impossible to achieve complete consensus and a contrarian approach is always
likely to be present and supported by at least a small proportion of a group or population.
Whether or not this contrarian position is “correct,” it can serve a useful purpose by
introducing criticisms and alternative solutions that deepen deliberation. In a multi-party
setting, these contrarian positions are more likely to be represented by a legislative party,
they are more likely to receive considerable public attention, and they are ultimately more
likely to become part of the decision-making calculus.

In two-party systems, on the other hand, contrarian positions are more frequently
ignored, especially if a super-majority of the public prefers an emotional response. Both
major parties will likely adopt the response preferred by a super-majority of the citizenry,
and the contrarian response is likely to be marginalized. For example, if a super-majority
of the public supports the use of capital punishment, low gasoline taxes, or opposes same-
sex marriage, then major parties in maj oritarian systems will likely adopt those positions.
To a great extent, thus, parties in majoritarian systems have a greater incentive to follow
the preferences of a majority rather than lead with alternative ideas. They will alter their

position on an issue only when public (or a powerful lobby’s) support for that position

19

has risen above a certain level (perhaps, somewhere around a near majority). It can be
expected that the public responds to this dynamic by accepting, in the main, that survival-
oriented approaches are the most legitimate and “correct,” while also viewing contrarian
ideas as less legitimate and eccentric.

As a rough example, the American public may believe that gas should be
inexpensive because both major parties say it should be, and yet the fact that taxes are
high in most other advanced industrial democracies suggests that the idea of higher gas
taxes is at least worthy of debate. In the US. setting, any party suggesting that gasoline
taxes should be raised (to help prevent global warming, to reduce dependency on foreign
oil suppliers, or for any number of other reasons) hurts its chances of winning an election.
In a multiparty setting, a small party supporting gasoline taxes can win representation
with a small proportion of support and therefore can take the risky position that voters

may not want to hear.

20

Chapter 3 Electoral Institutions and Information Diversity

As argued in the previous chapter, because electoral institutions affect political
competition they also inﬂuence the nature of political communication. Here I
concentrate on how electoral institutions inﬂuence information diversity in the political
discourse and in the media. The idea that electoral institutions inﬂuence political
information has already been suggested in the literature. Downs (1957) hypothesized
about this relationship, speciﬁcally noting that two-party systems would tend to
encourage message convergence between competitors, whereas message divergence
would tend to occur in a proportional representation structure. Since Downs, several
studies (including for example Gordon and Segura (1997) and Westholm and Niemi
(1992)) have assumed that electoral institutions inﬂuence the political information
environment, and consequently various elements of public opinion and behavior.
Despite this attention, the relationship between electoral institutions and political
information has not been formally examined, and this result is likely an artifact of
empirical circumstances. For the most part, electoral systems vary and provide empirical
leverage only cross- or sub-nationally, while substantial variations across time (major
electoral reforms) are relatively rare. F urtherrnore, while media systems certainly vary
cross-nationally, linguistic barriers and confounding cultural variation have sharply
limited cross-national comparative research on political media (Curran and Park 2000;
Hallin and Mancini 2004). To overcome these obstacles, I provide evidence for these
claims through an analysis of New Zealand, which underwent an electoral reform from a

plurality system to proportional representation in the 19903 and experienced all of the

21

expected party system effects. This case provides a unique quasi-experimental
opportunity to gain empirical leverage on the expected effects of electoral systems on
political discourse and media coverage of politics.

The empirical evidence presented in this chapter shows that electoral systems do
affect the nature of political information produced by the political system and the
information reported in the media. These effects include relatively intuitive relationships
(e. g., if an electoral system encourages a more fragmented and dispersed party system it
will also encourage a more diverse political discourse and news agenda), but it also
includes various effects that are consonant with the spatial dynamics described by
Downs. Finally, the lower levels of diversity in majoritarian systems (in relation to
proportional systems) are associated with lower levels of dissent, and this dynamic
encourages the formulation of more aggressive discourse, especially on issues related to

personal security.

Electoral Institutions and Political Information
As discussed in chapter 2, the theoretical preconditions for my hypotheses relating the

electoral system and political information are the relationships between the electoral
institution and the party system. To brieﬂy review these preconditions, electoral
institutions are expected to inﬂuence the information produced by the political system,
ﬁrst, by inﬂuencing the heterogeneity of party systems, and thereby we should expect
proportional systems to increase the diversity of issues discussed in a political system and
also the various solutions offered for given policy problems.

Second, electoral systems can inﬂuence information processes through spatial
incentives that encourage politicians and parties to gravitate toward or away from the

22

median voter’s preferences and this tendency should limit the range of political discourse.
Moreover, these parties will attempt to resemble one another and will adopt a catch-all
strategy, where they try to appeal to the greatest number of voters by employing tactics
such as ambiguity and emphasizing popular issues and positions while avoiding those
that are controversial. Thus, in comparison to multi-party systems, two-party systems are
expected to provide less policy information and encourage voters to base evaluations of
parties on non-ideological/non-policy factors, such as the horse race and the personal
characteristics of the candidates.

The ﬁnal precondition is derived from the expectation that diversity can mitigate
the problem of groupthink, especially by facilitating minority dissent. Minority
viewpoints are important because they stimulate divergent attention and thought, and thus
deepen deliberation and encourage more complete and innovative solutions to problems.
1 frame this last expectation in reciprocal terms. That is, we should ﬁnd a greater degree
of dissent under PR systems, in relation to Majoritarian systems, but because majoritarian
systems reduce dissent I expect them to encourage a form of groupthink.

As we will see, groupthink in this case is characterized by a style of political
competition that privileges reactionary solutions to problems. That is, majoritarian
systems encourage parties to seek votes by emphasizing survival-oriented issues and
solutions, and they marginalize contrarian/dissenting perspectives. This dynamic is
particularly visible for issues related to personal security, such as law and order and
national security. For example, candidates may seek to emphasize their “tough-on-
crime” credentials and suggest more and more punitive measures for dealing with crime,

or they may attack one another for not doing enough to keep gas prices down.

23

In light of the discussion below of the New Zealand case, 1 frame these
expectations in terms of a reform from a standard plurality system to a standard
proportional representation system. Broadly, a move from a plurality electoral system to
a proportional representation system encourages a more fragmented party system with
parties as a group advocating a more diverse and dispersed set of policy positions. After
long serving as a better example of the two—party / plurality / majoritarian Westminster
model than even the original (Lijphart 1984), New Zealand underwent a reform process
in the mid-19905 leading ultimately to use of a new mixed-member proportional system
based roughly on the German model. In New Zealand’s post-reform system, parties can
win seats with 5% of the national vote (or less, if they can win a single-member district).
To reach this threshold and its electoral rewards, small parties can seek out fairly small
niches of ideology or policy space, picking at the major parties from all sides. Moreover,
the major parties may ﬁnd it relatively advantageous to ﬁght off attacks from their
ideological ﬂanks, rather than compete with each other for the center. In general then, we
should expect parties as a whole to become more programmatic in post-reform New
Zealand.

While New Zealand’s reform has had a number of consequences, some
anticipated and some not (Barker et al. 2001; Boston et al. 2000; Vowles et al. 1998;
Vowles et al. 2002), there is little debate as to whether the basic expected consequences —
more political parties, more programmatic parties, etc. — have been observed. Below, I
revisit some of these consequences and compare them with consequences for political

news coverage.

24

Party Systems and the Media
With respect to the media system, I argue that the media in New Zealand abide by the

principal of political parallelism, according to which the media tend to reﬂect or parallel
the party system. Simply put, if an electoral reform induces changes in the party system,
we should expect to see parallel changes in the political news. Parallelism demands that
the media demonstrate responsiveness to changes in the party system. Therefore, as a
party grows or shrinks in popular support, coverage of that party and its policies should
also grow or shrink.

The notion of parallelism comes from Hallin and Mancini (2004). Their study
represents a substantial advance in the comparative study of politics and the media, and
appears to be the only prior study that explicitly notes the relationships between political
institutions and media systems. Focusing on North American and West European cases,
they group countries into three broad cultural / institutional “systems” — Liberal,
Democratic Corporatist, and Polarized Pluralist (these have geographic identities as
well: respectively, North Atlantic, North/Central European, and Mediterranean) — which
they note have partial parallels in Lijphart’s (1984, 1999) notions of majoritarian and
consensual government. Their Liberal systems — the United States, Canada, Ireland, and
the United Kingdom — in addition to all being Anglophone, also are broadly majoritarian
systems in Lijphart’s formulation, especially in their electoral and party systems.3

Countries in the other two Hallin/Mancini groupings are all broadly consensual nations in

 

3 Though Ireland, with single transferable vote and a multiparty system is at least a partial
exception.

25

continental Europe, all using some form of proportional representation and supporting
multiple parties.

Prior to its electoral reform and subsequent party system changes, New Zealand
met all of Hallin and Mancini’s criteria for a Liberal media system. Hallin and Mancini
note that the Liberal systems all have media systems in which the parallelism is internal,
meaning that each media outlet is pluralistic in its coverage and each provides a full
reﬂection of the political environment.4 In the continental systems, parallelism is largely
external, meaning that outlets are tied to particular partisan viewpoints and pluralism is
achieved at a systemic level. My claim is conditioned on the presence of internal
pluralism as, certainly, are my empirical tests using a single media outlet in New
Zealand.

This brings us to formal expectations about how electoral reform affects political
information. Combining the theoretical precondition of media parallelism with
theoretical preconditions on the party system effects of electoral reform produces six
general hypotheses. Framing these hypotheses to parallel the reform in New Zealand

from plurality to proportional representation, they are as follows:

H1: Relative shares of media attention allocated to political parties will be
responsive to parties’ expected seat share, under both plurality rule and

proportional representation.

 

4 Here the United Kingdom is a partial exception; I also return to this below

26

H2: The effective number of parties, as represented by media attention, will be

greater under proportional representation.

H3: The media attention given to speciﬁc issues (relative to horse race and other

non-issue coverage) will be greater under proportional representation.

H4: The relative media attention given to “new politics” issues and policy

positions will be greater under proportional representation.

H5: The dispersion of policy positions discussed (or solutions offered) on any
given issue, weighted by media attention to those positions, will be greater under

proportional representation.

H6: There will be more reactionary discourse especially on issues related to

personal security.

Empirical Investigation
As noted, I use the quasi-experiment of the New Zealand electoral reform as the setting

for the empirical investigation. While not a case in the Hallin and Mancini study, pre-

reforrn New Zealand is a clear example of their Liberal model. I therefore expect the

hypotheses to be reﬂected in the pre- and post-reform content of any given newspaper,

ceteris paribus. For this study, I examined the election coverage in the New Zealand

Herald, the largest circulation newspaper in New Zealand. The Herald has the additional

feature of not having been a participant in “public journalism” reforms undertaken by

27

other papers at the same time, allowing for a clear separation between my explanation
and that of McGregor et al. ZOOO—which I will describe further below.

To get clear comparisons of pre- and post-reform media coverage, I focus
primarily on the elections of 1990 and 1999. During the election of 1993, there was
already some anticipation of possible reform (including the referendum that would
introduce a new system in the next election), so this election may not give a clear
indication of pre-reform behavior. Conversely, during the election of 1996 there was no
experience with. the new institutions, so it may be unreasonable to expect behavior to
have yet approached equilibrium. Primary data, then, are based on content taken from
the New Zealand Herald’s special campaign coverage for the three-week period prior to
the 1990 and 1999 elections. For almost all the days analyzed in this period there were
two pages of special campaign coverage that appeared each day, and this format
remained conveniently consistent across the years examined.5 1 provide more details of

speciﬁc measurements in the hypothesis testing discussions below.

The Distribution of Media Attention to Political Parties
The ﬁrst set of hypotheses make predictions about how media attention will be

distributed to political parties relative to their expected electoral and parliamentary
strength. I measure media attention as a simple count of articles or editorials in the

Herald election coverage sample that mention each political party. Election results, votes

 

5 It is worth noting that the only records of the Herald available prior to 1997 were microﬁche
versions stored at the Library of Congress, which had to be copied and manually coded. As a
result, each data point was relatively expensive.

28

or seats as appropriate, serve as the relevant comparison standard.6 These data are shown
in Table 3.1. First note that the electoral result data comport completely with standard
theoretical preconditions. There are clearly fewer parties winning representation under
plurality, in 1990, and more winning representation under proportional representation, in
1999. To put this more formally, I look at the standard measure of effective number of
parties winning representation in the legislature, which rose from 1.74 in 1990 to 3.46 in
1999.

With respect to the ﬁrst two hypotheses, the evidence supports the notion that
media attention is positively related to expected seat share and that minor parties receive
more coverage under proportional representation. In Table 3.1, we can see that media
coverage abides by the Duvergerian effect toward more parties in media coverage. The
effective number of parties receiving mentions in the media rose from 2.54 to 5.78.7 In
fact, it appears that the norms of media neutrality cause the Duvergerian difference to be
more dramatic in media coverage than that reﬂected in actual votes. The media depicted

3.24 more parties after reform while the election results reﬂected only 1.72 more parties.

 

6 Media attention is delivered before votes and seats, of course, but the actual results are
completely reasonable estimates of what results were expected weeks before the election.

7 Hayward and Rudd (2000) report a measure of space [sq. cm.] given to parties in the New
Zealand Herald in 1999 over a different sample frame, four weeks rather than three. There are
differences for particular parties, but the implied effective number of parties is roughly identical
to ours, 6.04.

29

Table 3-1 Votes, Seats, and Media Attention

 

 

 

 

1990, 96 1999, 96
Party Votes Seats Mentionsa Votes Seats Mentionsa
National 48.2 70.1 47.2 (134) 30.5 32.5 23.5 (99)
Labour 35.1 28.9 41.2 (117) 38.7 40.9 23.7 (100)
New Labour 5.2 1.0 2.5 (7) - - —
Social Credit 0.9 0.0 1.1 (3) - - -
New Zealand 0.0 0.0 0.7 (2) - - -
Mana Motuhake 0.6 0.0 2.8 (8) - - -
Green 6.8 0.0 1.4 (4) 5.2 5.8 15.9 (67)
Christian Heritage 0.5 0.0 1.1 (3) 2.4 0.0 0.1 (2)
Democratic 1.7 0.0 1.1 (3) - - -
McGillicuddy Serious 0.0 0.0 0.7 (2) 0.2 0.0 0.1 (2)
Communist League 0.0 0.0 0.4 (1) 0.0 0.0 0.0 (0)
Alliance - - - 7.7 8.3 12.3 (52)
ACT - - - 7.0 7.5 11.8 (50)
New Zealand First - - - 4.3 4.2 8.5 (36)
United - - - 0.5 0.8 0.1 (4)
Future NZ - - — 1.1 0.0 0.1 (2)
Legalise Cannabis — - - 1.1 0.0 0.1 (2)
Piri Wiri Tua - - — 0.0 0.0 0.1 (3)
Mana Maori - - — 0.3 0.0 0.1 (2)
Maori Pacific - - - 0.2 0.0 0.0 ( 1)
Eff. if of Parties, N g 2.75 1.74 2.54 3.89 3.45 5.78

 

 

 

 

a .
Raw counts in parentheses.

Under plurality rule, the media depicts a nearly perfect two-party-and-change
system, with equalizing coverage for the two large parties and trivial coverage for the

remainder. Under proportional representation, the large parties lose only 15% of their

30

collective vote share, but 50% of their media attention, leaving a media picture of a party
system more highly fragmented than it really is.

There are further subtleties here that reinforce the notion that it is in fact the
electoral system, and not other coincidental changes, impacting media behavior. Note in
particular the coverage of the Green Party detailed in Table 3.1. In 1990, they received
nearly 7% of the vote, but anticipating correctly that this would translate under plurality
into no seats and no post-election relevance, the Herald accorded them only 1% of its
coverage. In 1999, the Greens’ popular support dropped to 5%, but anticipating that this
would now translate into seats and even possible membership in a governing coalition,
the Herald accorded them 16% of its coverage. A similar effect is seen with New Labour
and its partial successor, the Alliance. Under plurality, 5% of the vote translates into half
as much (2.5%) coverage, but under proportional representation, 8% of the vote translates
into increased (12%) coverage. The media attends to parties not in response to their
expected popular support (vote share) but to their expected relevance (seat share). In

short, when the electoral system began taking small parties seriously, so did the Herald.

The Distribution of Media Attention to Political Issues
I now turn the focus from the media attention that parties receive to the media attention

that topics receive. 1 am concerned both with (a) how much attention the media gives to
issues rather than non-issue campaign coverage, and (b) how much attention the media
gives to “new” rather than traditional issues. My expectations are that parties will
increase their issue-speciﬁc campaigning and their relative focus on new issues, and these

increases will be reﬂected in media attention.

31

The expected effect is then that the media will make less reference to valence and
purely political issues and more references to speciﬁc policy positions. Again, for my
purposes, only the comparative statics — more issue competition on more issues — need to
hold for the story to have bite. Unfortunately, there are no current means (using the
historical data available) to gain separate leverage on the theoretical precondition (that
the parties collectively focus more on issues) and Hypothesis 3 (the media focus more on
issues). Nonissue campaigning occurs largely outside the realm of historically available
sources like manifestos, which are consistently issue-dominated documents, and analysis
of media coverage alone would conﬂate the two. Thus, with the current data, I can only
assume the theoretical precondition and look for the net effect of increased media issue
coverage (Hypothesis 3).

I did not expend resources collecting horse race and valence mentions, because
Hypothesis 3 is strongly supported in the work of others. In their analysis of New
Zealand’s newspaper coverage, McGregor et a1. (2000) look at this issue speciﬁcally.
They reported a signiﬁcant drop in “horse race”, personality, and other traditional
campaign coverage, relative to issue-reporting, from 1993 to 1996. They found that
horse race coverage dropped from 16.5% of coverage in 1993 to 15.6% in 1996. More
dramatically, coverage that focused on candidates (as opposed to policies or issues) fell
from 33.4% to 21.3%. (McGregor, et al. 2000: 143, Table 3). I have noted that the party
system - and in turn, news coverage — had already begun to anticipate the new multiparty
reality by 1993, so this effect might appear stronger with data from 1990 or earlier. It

should be noted also that McGregor, et al. attributed this effect not to electoral reform but

32

to media reform, speciﬁcally a turn to “public journalism”—an issue further addressed
below.

Hayward and Rudd (2000) similarly reported an increase in “interpretive”
coverage, relative to campaign coverage, from 1987 to 1999. Their measure is space (sq
cm) and they examine the ﬁve largest New Zealand newspapers. They found 9% of
space devoted to interpretive news in an earlier study of the 1987 election, but 34% in the
1999 election. This varied considerably across papers, with the New Zealand Herald
having the highest ratio of interpretive news to all others. Like McGregor, et al.,
Hayward and Rudd were also not looking at this as an electoral system effect, but rather
as evidence against the hypothesis that election news has followed a secular trend of
becoming more “trivialized”.

So, while collected for different purposes and in support of different theories, the
data gathered by McGregor et al. and Hayward and Rudd supports Hypothesis 3. On this
particular point, the expectations are observationally equivalent. Note, however, that
Hayward and Rudd’s is a null theory that does not produce any hypotheses in the other
areas I examine. To the extent that McGregor et al.’s expectations overlap with my
own—and one could argue that public journalism might produce some of the same
effects independent of electoral reform—they should do so only in papers that undertook
public journalism reforms. The Herald did not do so, and therefore, the effects that
appear in either my data or Hayward and Rudd’s would not do so if the explanation were
public journalism rather than electoral reform.

More interesting, perhaps, is the expectation that competition will become

multidimcnsional. In a majoritarian system like pre-refonn New Zealand, I expect

33

competition between the two major parties to be centripetal and to be primarin on the
traditional economic / materialist left-right dimension. Post-reform, I expect minor and
new parties to attempt to carve out niches by seeking out opportunities to meet latent
demand in other dimensions. Through much of the industrialized world, the prominent
emerging second dimension in electorate preference has been “postrnaterialism” or
“postmodernism” (lnglehart 1990, 1997, 2000), deﬁned by attitudes towards such general
areas as the environment, minority rights, women’s rights, and cultural permissiveness.
The New Zealand Election Study found postrnaterialism to be emerging as a second
dimension in voter attitudes prior to reform (Vowles and Aimer 1993). So, as more
relevant parties emerge to press the saliency of such new issues, we should expect
parallelism to create greater media discussion of news issues (Hypothesis 4).

Economic issues fall most clearly in the “old” materialist dimension. Issues
related to the environment, race relations, gender, nuclear weapons, and “enlightened”
approaches to law and order are among those that are categorized most clearly in the
“new” postmaterialist dimensions. Note that this does not just mean typical postmaterial
positions, but also positions in direct opposition, such as tougher immigration practices or
returns to family values. We should expect to see an increase in the attention given to
these sorts of issues, in relative terms. Issues associated with the welfare state conﬂate
both materialist and postmaterialist dimensions (see, e. g., Vowles, et al. 1993: 65), and
consequently, I have no ﬁrm expectation about this issue area.

To develop these measures, I count articles or editorials in the sample that
mention 22 particular issue categories. These are reported in Table 3.2. Like the party

measures, 1 count only one mention of each issue per article. Since exogenous factors are

34

likely to affect the relative attention given to any particular issue in any given campaign,
and this effect is likely to be more substantial than any anticipation or learning effects
outside of the 1996 election, data were collected from for samples from four election
years. The two elections prior to 1996 (1990 and 1993) constitute the pluralitarian
sample and the two elections subsequent to 1996 (1999 and 2002) constitute the
proportional representation sample.

I also rely on a second source of evidence: the attention accorded these issues by
parties, weighted by party size and not ﬁltered through the media. I measure this
attention through a count of mentions of relevant word stems in the Parliamentary
Debates (Hansard) of the New Zealand House of Representatives, the official record of
parliament for two periods pre- and post-reform, 1988-1992 and 1997-2003.8 These data
are also reported in Table 3.2.

I operationalize attention to new issues as the proportion of mentions given to
environmental issues, law and order, immigration, minority relations. and
children/families. It is important to reiterate that many of these new issue mentions are
for policies that could perhaps be described as opposed to postmodern values: tougher
immigration policy, crackdowns on abuses of the Treaty of Waitangi, advocacy of family

values, and mandatory sentencing laws.

 

8 These are as archived by the Virtual Democratic Infrastructure Group, www.vdig.net, which has
made Hansard available, with a search engine, for Nov. 1987 through the present. The counts are
of “hits” in the database, with a hit being a speech or debate contribution that contains the word
stem at least once. Some debate contributions will be counted more than once if contributing to
more than one topic; this is also true of articles in the Herald data.

35

Table 3-2 Issue-Specific Mentions in Media and the House of Commons

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

New Zealand Herald 1990 1993 1999 2002 Hansard 1988-1992 1997-2003
Taxation 4 15 25 l 1 "tax..." 3836 4747
Employment 18 24 8 "job..." 5201 7340
Trade 8 4 5 "trade..." 3 817 4708
Privatization / Nationalization 5 14 10 15 "privatis..." 516 827
Regulation 0 1 9 4 "regulat..." 1584 3024
Inﬂation 1 1 O 1 "inﬂation" 1679 674
Wage Rate 8 5 2 1 "wage..." 1555 1821
Other Industrial Relations 9 6 13 5 "union..." 1931 2138
Agriculture / Rural Issues 2 1 4 0 "rural..." 1712 1367

Total Economy: 45% 47% 33% 18% 47% 43%
(4%) (4%) (3%) (2%) (0.2%) (0.2%)
Superannuation l 8 6 9 "superannuat..." 1721 1655
Education 12 16 19 28 "educat..." 43 16 5409
Health 6 14 13 14 "heath..." 4063 6356
Other Welfare 5 7 9 6 "welfare..." 3327 1970
Total Welfare State: 17% 39% 19% 21% 28% 25%
(3%) (4%) (2%) (2%) (0.2%) (0.2%)
Foreign Affairs 8 2 4 11 "defenc..." 1352 2173
Total Foreign Policy: 6% 1% 2% 4% 3% 3%
(2%) (1%) (1%) (1%) (0.1%) (0.1%)
Drugs / Cannabis l 0 14 11 "drug..." 491 902
Other Law and Order 15 1 l 19 41 "crim..." 1113 2270
Total Law & Order: 11% 7% 13% 19% 3% 5%
(2%) (2%) (2%) (2%) (0.1%) (0.1%)
Children / Families 3 6 1 1 12 "family/ies" 2771 5291
Maori / Treaty of Waitangi 12 10 27 17 "Waitangi" 794 1502
Immigration 1 2 3 18 "immigra..." 501 1477
Total (Other) Social: 11% 12% 16% 17% 9% 13%
(3%) (3%) (2%) (2%) (0.1%) (0.1%)
Roads / Public Transport 0 0 11 9 "transport..." 1878 2499
Other Environment 14 3 31 52 "environment..." 2230 3972
Total Environment: 10% 2% 17% 22% 8% 10%
(3%) (1%) (2%) (2%) (0.1%) (0.1%)
Total: 145 149 251 278 46388 62122
Proportion "Wis-“rues, V.-S= 27% (3%) 52% (2%) 20% (0.2%) 28% (0.2%)

 

 

 

Standard errors of proportions in parentheses.

36

Table 3.2 reports the full data set. Given these assumptions, the theoretical
precondition is met with parliamentary attention to new issues rising from 20% to 28%.
Hypothesis 4 is strongly supported with media attention to new issues rising from 27% to
52%. As was the case with party mentions, media attention exaggerates the information
produced by the political system; the increase in media attention of 25% is much larger
than the parliamentary increase of 8%.

The Distribution of Media Attention to Issue Positions

Perhaps even more interesting than the distribution of attention to parties and issues, is
the distribution of attention to ideas. This is where we get some concrete sense of
institutions shaping the richness and diversity of the political information system; and as
it turns out, the impact of electoral reform is dramatic. As noted earlier, under the
Downsian incentives of pre-reform New Zealand the two major parties are expected to be
roughly convergent, espousing similar positions on any given policy issue. In the post-
reforrn world, small parties can ﬁnd electoral success by carving out small portions of
space not just in secondary dimensions, but also on the extreme ﬂanks of the major
parties. Not only should we see a greater variety of positions expressed directly by these
newly successful parties. but we should expect them to exert some centrifugal pull on the
major parties as well (Cox 1990). For any given topic area. therefore, we should expect
relatively greater attention given to more extreme issue positions (Hypothesis 5).

I develop variance measures using more nuanced versions of the issue mention
data from 1990 and 1999. These consist of the text surrounding each issue mention

(roughly the same as a “quasi-sentence”, the coding unit in the Comparative Manifestos

37

Project (Budge et al. 2001)).9 As a comparison, I also have sentences describing the
positions of the parties on each issue, as described in its manifesto or similar declaration.
The hypotheses should hold for any given issue. Here, results are reported for three of
the largest issue areas: taxation, pre-tertiary education, and law and order.

Within any single issue category, the quasi-sentences are coded as positional data.
Bundling the Herald and manifesto statements together, with party labels removed, the
positions are placed on a seven-point scale from —3 to 3. I use —3 to represent the most
“left” position expressed, +3 the most “right” position, and 0 for an endorsement of the
status quo or for statements that are vague in left-right terms.

Figure 3.1 shows the results for taxation. Each graph depicts two distributions.
The distribution made up of the darker bars on the left of each pair depicts the
distribution of policy positions in the party programs — the height of a bar indicates the
combined proportion of the vote at that particular election for the parties (party list votes
in 1999) coded at that policy position. The lighter bar on the right of each pair depicts
the proportion of issue mentions in the media sample that are coded at that location. For
several positions, the ﬁgure gives examples of manifesto or media statements that have
been coded at that location. On the left I code statements favoring more progressive tax

rates — higher for hi gh-income and/or lower for low-income - or shifts in the mode of

 

9 I do not use the CMP coding method directly because it does not translate well to the media
data, which can include negative statements about others’ positions, defenses against purported
misstatements, and so on. I note that Gibbons (2003) analysis of the New Zealand manifestos
using the CMP method gives unambiguous evidence of the theoretical precondition: dispersion of
the parties in multiple dimensions.

38

Figure 3.1 Positions and Media Mentions:

Taxation

 

 

Proportion

0.5

0.4

0.3

0.2

0.1

0.0

l

(a) Taxation, 1990

 

 

 

l Vote-weighed Marifesto Positions, “=00, 02:1.42

 

 

‘3 Mertions in NewZea/and Herald, 750.0, 32:158. n=4

 

 

 

 

Left-Hg“ Position

 

39

 

 

 

Figure 3.1 (Continued)

 

Proportion
0.3 0.4
r 1

0.2

0.1

 

0.0

 

(b) Taxation, 1999

 

I Vote—weighed Marifesto Positions, n=o.1, 02=2.60
E3 Mentions in NewZea/and Herald, 2:02, s2=2.72, n=25

 

 

 

 

 

   

Left-Rigtt Position

 

40

 

taxation away from GST (sales tax) or toward taxes on capital gains, ﬁnancial
transactions, transfer (estate) tax, corporations, pollution, and that sort of notion.

With regard to the former, for example, in 1999 each party made an explicit pledge about
the level at which the top tax rate (then at 33%) should be set, with proposals ranging
from 20% (ACT) to 47% (the Alliance). Thus, the data on taxation are supportive of the
theoretical precondition, with an 85% increase (2.60/1 .42=1 .85) in the dispersion of
manifesto positions from 1990 to 1999. Hypothesis 5 is also supported with an increase
in dispersion of media mentions of 72%.

Figure 3.2 shows more compelling results for pre-tertiary education.l0 The
primary left-right distinction is a public-private debate, with full state funding of a
completely free education on the left, and a largely private system supported by a
voucher system on the right. Other important issues were “bulk funding”, teacher pay,
and national testing. Again we can see a centered unimodal distribution in 1990, with
manifesto pledges on valence issues like truancy and bland slogans such as “education is
working”. It is worth noting that left-leaning positions of the Greens, the New Labour
Party, and Social Credit (aka the Democrats) were completely ignored in the 1990 media
sample. The theoretical precondition again holds, with an 89% increase in dispersion of
manifesto positions. As predicted by Hypothesis 5, there is a dramatic increase in the
dispersion of media mentions of 237%, and again, the media increase is 78% larger than

that in the manifestos.

 

'0 This is a conservative narrowing of the education topic. Higher education issues of student
loans and university funding were widely discussed, with highly variable opinions, in 1999, but
mostly unmentioned in 1990. If this is included, the support for the hypotheses is even stronger.

41

Figure 3.2 Positions and Media Mentions:

Education

 

Proportion

 

0.2

0.1

0.0

(a) Education (pre-tertiary), 1990

L

 

 

I Vote-weighed Marifesto Positiors, n=o.o, 62:1.76
El Mertions in NewZeaIand Herald, 7:00, s2=1.24, n=7

 

 

 

 

 

 

 

 

 

 

-2 -1 0 1

Leﬂ—Rigtt Position

 

42

 

 

Figure 3.2 (Continued)

 

 

Proportion

0.5

0.4

0.3

0.2

0.1

0.0

(b) Education (pm-tertiary), 1999

 

 

I Vote—weighted Marifesto Positions, n=o.1, o2=3.43
D Mentions in NewZea/and Herald, i=o.o, 52:418. n=10

 

 

 

 

 

 

 

 

 

 

 

 

.....__.
-2 -1 0 1 2 3
Left-Rigtt Position

 

43

 

Figure 3.3 Positions and Media Mentions: Law &

Order

 

Proportion

 

0.5

0.4

0.3

0.2

l

0.1

0.0

l

 

(a) Law & Order (including cannabis), 1990

1

 

 

I Vote-weighted Manifesto Positions, u=0.0, oz=1.42
D Mentions in NewZea/and Hera/d, 7:01, 32:0.80, n=16

 

 

 

 

 

 

 

 

 

 

 

Left-Right Position

 

44

 

Figure 3.3 (Continued)

 

Proportion

 

0.5

0.3 0.4

0.2

0.1

0.0

l

(b) Law 8. Order (including cannabis), 1999

 

 

I Vote-weigtted Manifesto Positions, ”=01, 02:3.57
D Mertions in NewZea/and Herald, §-—0.1, 52:5.50, n=33

 

 

 

 

.3

 

 

 

 

 

 

-2 -1 0 1 2
Left-Rigtt Position

 

45

 

Figure 3.3 shows the most dramatic results, for law and order issues including
cannabis legalization. Again, we see largely unimodal distributions of manifesto
positions and media mentions in 1990, ignoring the far left and centered slightly to the
right of status quo. In 1999, the Greens, Alliance, and Legalize Cannabis parties struck
positions on the far left (e. g., drug legalization, rehabilitation centers instead of prisons)
while the other major parties all took positions to the right of the status quo, ranging from
mildly so (e.g., tougher bail requirements) to sharply so (e.g., doubling prison sentences,
hard labor). The evidence for the hypothesis is dramatic in this case, with a clear increase
in attention to extreme positions on law and order. This is largely driven by the
controversy over drug legalization, especially as sparked by revelations of a criminal
background of one Green candidate (that would have likely been ignored under the
previous system), but we also see “over-attention” to policies on the far right. With this
issue, the increase in dispersion is extreme and the effect is unambiguous. There is
strong evidence for the theoretical precondition (manifesto dispersion increased 157%)
and for Hypothesis 5 (media mention dispersion increased 589%). Again, the increase in

media dispersion is 167% greater than the increase in manifesto dispersion.

The Groupthink Dynamic
The evidence from the previous section also hints at hypothesis 6 — that there will be a

greater tendency for groupthink in majoritarian systems. For the issue areas discussed
here, it appears that majoritarian systems are especially likely to attenuate perspectives
from the left and there is a strain toward a center-right consensus. It is reasonable to
expect that this pattern is especially relevant for issues associated with personal security
because these kinds of issues are most likely affected by human instinct to focus on short-

46

term responses rather than reﬂecting on root causes. The introduction of unpopular
alternatives, that may address the root causes of problems, is not likely to occur if most
voters support the short-term solution. Indeed, in majoritarian systems, if a party or
candidate proposes long-term solutions for these kinds of issues, the other party may
suggest that it or he is weak.

We can gain a little more insight on this dynamic by taking a closer look at the
discourse on law and order. Table 3.3 displays the complete list of crime-related
statements (in a “quasi-sentence” format) that appeared in the New Zealand Herald in the
three weeks before the 1990 and 1999 elections.“ It is possible to note that when New
Zealand operated under a maj oritarian system, in 1990, the campaign discourse that
unfolded emphasized punishment and enforcement as solutions to crime. The statements
from the conservative National party were particularly aggressive, while Labour’s
statements were mainly defensive and at times also aggressive. In general, there were no
statements about reducing crime through preventative measures.

When New Zealand switched to MMP, a wider variety of positions emerged. In
1999 (the second election under MMP) there were still calls for punishment and
enforcement but there was also a promulgation of alternative solutions aimed at
addressing root causes. For example, the Green Party and the Alliance favored the ideas
of “restorative justice” and spending money on mental health, and these kinds of ideas

therefore received media attention.

 

” Again. this involved a content analysis of usually two pages of special campaign coverage.

47

Table 3-3 Semi-sentences on Law & Order in New Zealand Herald

 

Law and Order in 1990 Under Majoritarian System

Law and Order in 1999 Under Mixed Member
Proportional Representation

 

I think big things are help for more police in trying
to control crime (Cit)

impressed with L's ﬁghting crime w/ more family
support not prisons (Cit)

 

gave party's plans for crime (N)

dismisses N's move to give police powers to take
DNA from burglary suspects (Cit)

 

N's advertisement deceptive, claiming crime had
increased during L's term in gov't (L)

he is supporter of Act's tougher longer prison
sentencing, Nat's DNA test plan (Cit)

 

latest official figures showed decline in crime;
problem was inherited got out of control under N

(L)

nervous about Act, it is now edging harder on
crime (N)

 

false advertising about crime rates

NZ voters impressed with tough law-and-order
policy (Act)

 

N spokeswoman said she would not be opposed to
use of mace to disable attackers

restorative justice: make offenders accept

 

under N people would understand they had more
than even chance of being dealt with if they
offended (N)

responsibility, apologize, pay restitution (GP, All)
wanted to register all guns as well as owners
(All)

 

promised NZlanders would no longer be too
frightened to go out at night (N)

spend extra on mental health, for helping
families of criminally ill (All)

 

question from floor for former PM (L), whether an
extra 900 police would reduce crime

N/Act outbid each other on who can be toughest
on crime, but go quiet on how to stop crime (All)

 

NZ did not have effective deterrents (former L PM)

claims GP is against prisons and for breaking the
law (N)

 

promised legislation to deny bail to violent
criminals charged w/ similar offences (N)

supports prison for violent offenders, is against
breaking laws, unless no other option (GP)

 

N would amend legislation to give police powers to
deal w/ young criminals

continue conditions for security (N)

 

introducing tougher bail laws would be one of top
priorities for N

making its stance on longer prison sentences for
criminals a bottom line in coalition (Act)

 

L attacked N's bail proposal, 99% of people do not
commit crimes of violence while on bail

N has tried to crush GP on candidates’
confrontations w/ law

 

he laid out themes: collar the crims (N)

in his decent society, police have resources to do
their job (L)

 

referendum: favour reform emphasizing needs
of victims, imposing minimum sentences hard
labor (Cit, NZF, N, Act)

 

favour restorative justice aspects of crime
petition, are softer on longer sentences (L, All,
GP)

 

 

 

she laid down same hand: a harder line on
burglaries (L)

 

Abbreviations: Act=Act Party, All=Al|iance, CH=Christian Heritage, FNZ=Future New

Zealand, GP=Green Party, L=Labour Party, ALCP=Legalise Cannabis Party, MP=Maori Paciﬁc,
MMM=Mana Maori Movement, MS=McGillicuddy Serious Party, N=National Party, NZF=New
Zealand First, PWT=Piri Wiri Tua party, UP=United Party, Cit=citizen

48

 

Discussion
This chapter provides conﬁrmatory evidence for what researchers such as Downs

theorized (and others have assumed) about how electoral institutions affect information
processes. Electoral institutions inﬂuence the diversity of messages produced by the
political system, with proportional systems tending to encourage the representation of
more perspectives than majoritarian systems. The political discourse is also inﬂuenced
by the spatial incentives associated with each type of electoral system, with PR systems
encouraging a more programmatic discourse.

Morevoer, I believe this is the ﬁrst study to (a) specify theoretical relationships
between electoral institutions and media coverage of politics, and (b) provide empirical
evidence of their existence. The ﬁndings here suggest that we need to be aware of
institutional context in otherwise institution-free media studies. In the context of a media
system roughly corresponding to the Hallin and Mancini Liberal model, a more
permissive and fragmentation-encouraging electoral system will lead to a more diverse
political news agenda. More parties will get attention, more issues will get attention, and
a more diverse array of policy options will get attention. F urtherrnore, the norm of
political neutrality appears to amplify these effects such that increases in the diversity of
the political system are magniﬁed in their reﬂections in the media.

We already have a good idea that this affect on information can inﬂuence
important political phenomena, such as voter behavior, but the results in this study may
point to many other important and unexplored consequences of these information

processes for public opinion and policy—making. I explore some of these consequences in

49

the following chapters, including how these information processes affect public opinion
and public policy toward new issues and toward issues linked to personal safety.

There are important caveats to consider. First, the research design is quasi-
experimental, and comes with all of the weaknesses inherent in such a design. Other
changes in the New Zealand system at roughly the same time as electoral reform could
contaminate the inferences, and the same can be said of any endogeneity in the electoral
reform — i.e., that increases in party system fragmentation caused electoral reform, and
not the reverse. Some might note the Green Party’s relative electoral success prior to
reform (7% in 1990) as evidence of this, but the dramatic increase in media attention to
the Greens post-reform gives a much more comfortable ﬁt with my narrative. Second,
the evidence might be viewed as conditional in various ways. It is certainly conditional
on the media system meeting the basic parameters of Hallin and Mancini’s Liberal
model. Furthermore, it is possible that there are peculiarities—about New Zealand, about
the New Zealand Herald, or about the sampling frame—that could affect the results
observed.

Ultimately, addressing these concerns will require the accumulation of expensive
data. With respect to the case of New Zealand, this may require the examination of a
larger sample from more newspapers. Another tactic is to conduct more quasi-
experimental analyses, particularly cases with signiﬁcant shifts from pluralitarian to
proportional electoral systems, or vice versa. A ﬁnal approach is to conduct cross-
sectional examination of several cases—some pluralitarian and some proportional.
Unfortunately, pursuing these approaches will require relatively extensive resources and

will likely not become available for some time.

50

Chapter 4 Electoral Institutions and Public Opinion

If electoral institutions affect the kind of information that citizens can access, then we
should expect them to inﬂuence a crucial component of the policy-making process:
public opinion and behavior. As noted earlier, this possibility has already received
attention. Several studies, for example, have found that electoral institutions affect
participation in elections, satisfaction with political institutions, the tendency to cast
tactical votes, and attachments to parties (Karp et al. 2002; Niemi et al. 1992; Norris
2004)

Studies of this kind have argued that institutional context inﬂuences the type of
information produced by political systems and thereby inﬂuences various elements of
public opinion. Westholm and Niemi (1992) noted, for example, that the transfer of
ideological orientations and the relationship between ideological orientation and
partisanship is stronger in multiparty systems than in systems with fewer parties because
parties in the former systems have a strategic incentive to maintain ideological
distinctions and because multiparty systems increase the need for voters to employ a
cognitive spatial ordering of some kind. Similarly, Gordon and Segura (1997) argued
and demonstrated that because electoral institutions shape the availability, clarity, and
usefulness of political information, citizens in multiparty countries tend to outperform
those in countries with fewer parties on indicators of political knowledge.

This chapter extends this line of work by arguing that electoral institutions
actually help shape public preferences toward speciﬁc political issues and may even help

shape cultural change in ways that impact policy problem-solving. If the diversity of

51

multi-party systems does facilitate public consideration of sensitive issues (such as
material sacriﬁces for the sake of protecting the environment or the extension of rights to
same-sex couples), then can we expect that there will be greater public support for these
issues in diverse systems? Conversely, if majoritarian systems encourage a reactionary
style of political debate, can we expect there to be greater public support for short-term
policy responses in such systems?

Data from the 1999 World Values Survey and the International Crime
Victimization Survey are used to show that the answer to both of these questions is yes.
Electoral institutions do help explain the variation in attitudes observed across countries
on several sensitive issues. I show that citizens from countries with proportional electoral
systems (i.e., countries with institutions that increase the ﬂow of more diverse ideological
information) are more likely to support the sensitive issues promoted by parties on the
left of the political spectrum. This ﬁnding matches expectations derived from an
interdisciplinary body of research showing that individuals with access to more
information tend to hold more “cosmopolitan” or “postmodern” social attitudes. I show
also that citizens in countries with majoritarian electoral institutions tend to hold more
reactionary policy preferences, especially on issues related to crime.

It is argued in this chapter that these results can be interpreted as showing a) that
proportional systems facilitate social learning and permit the public to think in a long-
term manner and b) that majoritarian systems encourage survivalistic preferences.
Tolerance toward the sensitive issues discussed here generally reveals that people are
willing to think outside the box and that they are willing to accept short-term costs for the

sake of long-term gains. This is especially true for the environment where the long-term

52

gains are often long term and not immediately apparent. Support for punitive responses
to crime shows that people are thinking about short-term solutions that do not necessarily
address the root of the social problem. I show in chapter 5 how these ﬁndings are

connected to policy-making.

Electoral Institutions, Information Diversity, and Cosmopolitanism
When political elites settle on the electoral institutions their countries should adopt they

probably give little thought to how these institutions may inﬂuence public opinion. It
may seem a leap to expect that electoral institutions are somehow related to citizen
preferences, and yet, researchers have begun to uncover the contours of just such a
relationship. Several studies document how electoral institutions inﬂuence various
elements of public Opinion and behavior via their inﬂuence on the political information
environment (see for example Gordon and Segura (1997) and Westholm and Niemi
(1992)). Here, I show how the information mechanisms associated with electoral
institutions inﬂuence public attitudes in rather ﬁne ways; they actually inﬂuence citizens’
preferences on speciﬁc issues.

The information mechanisms described in chapters 2 and 3 provide clear
expectations for public opinion. Recall that electoral institutions affect the diversity of
party systems, issue dimensions, and policy ideas; they inﬂuence the degree to which
campaigns provide programmatic versus candidate centered information; and they
inﬂuence the degree to which dissent will be expressed in a political system. These
mechanisms suggest that electoral institutions will affect how frequently the public is
exposed to certain messages. Mainly, I expect electoral institutions to affect access to
information on what I refer to here as “sensitive” issues and “sensitive” policy positions.

53

My conceptualization of sensitivity involves two concerns. First, sensitive issues
and sensitive policy positions are those which more centrist politicians will ﬁnd difficult
to discuss. Consequently, many of these kinds of issues and positions are more likely to
be introduced by minor parties. Second, sensitivity is a time-dependent concept. An
issue that is sensitive at a particular point in time is one that could be ordinary in another.
For example, an issue like racial equality was once a sensitive issue in most countries and
today it is largely taken for granted that discrimination based on race should be illegal.
More recently, an issue like homosexuality appears to be undergoing the transition from
sensitive issue to ordinary.

Based upon the New Zealand media studies, I expect therefore that, for the period
analyzed in this chapter (1999), the institutional dynamics discussed above will be
particularly relevant to “new politics” issues.l2 In the New Zealand media studies it is
clear that the issues promoted by the Green Party received greater attention under
proportional representation than under majoritarianism. On the other hand, the inclusion
of more diverse perspectives also led to greater attention for marginal issues and
positions (such as immigration) promoted by small parties on the right of the political
spectrum. How, then, will exposure to these marginal topics affect public opinion?
Should we expect that exposure to greater ideological diversity will lead the public to
become more accepting of sensitive issues and positions from the left and right, or might

we expect that certain sensitive issues are more likely to gain support than others?

 

'2 For a different time period, perhaps circa 1950, the relevant set of issues might have been
linked more closely to economic issues Exploring this possibility appears interesting but less
feasible due to data constraints.

54

Insights into this question can be gleaned from an interdisciplinary body of
research on the effects of information on individual attitudes and cultural change. First,
modernization theorists have argued that socioeconomic development is associated with
technological innovations, increases in labor productivity, occupational specialization,
rising levels of education, rising levels of income, and the diversiﬁcation of human
interactions, and these forces consequently produce coherent changes in culture and
political life (lnglehart and Welzel 2005; Rokeach 1960; Inkeles 1983; Inkeles and Smith

1974; Lerner 195 8). In general, these forces encourage the development of more

’9 99 ‘6

“cosmopolitan, “secular, tolerant,” and “self-expressive” social attitudes. With
respect to particular attitudes and behaviors, lnglehart and Welzel (2005) note that in
recent decades wealthy societies have experienced cultural changes that include the
supplanting of institutionalized religion with individualized spiritual concerns, greater
support for protecting the environment, and greater tolerance toward issues such as
homosexuality, gender equality, and divorce.

Modernization theorists such as lnglehart have emphasized the importance of
existential security in producing these cultural changes, but much of modernization
research concentrates on the importance of cognitive mobilization. Modernization
theorists such as Lerner (195 8), Inkeles and Smith (1974), and Inkeles (1983) emphasized
the importance of knowledge and education in the modernization process, noting that
individuals with the highest levels of education tended to hold more secular and modern

worldviews. Cognitive mobilization is also enhanced by the emergence of service and

knowledge sectors that permit citizens to exercise individual judgment and creativity, and

55

the evolution of mass media and information technology that give people access to
knowledge and increase information autonomy.

We can draw similar insights from social psychological studies on the differences
between urban and rural individuals. These studies rely on a classical theory from Wirth
(1942) who suggested that the differences observed across urban and nonurban cultures
could be largely explained by three factors of residential space: size, density, and
heterogeneity. He expected that, in general, individuals residing in heterogeneous, more
densely populated areas will tend to be tolerant, “psychologically sophisticated,”
impersonal, and generally cosmopolitan. On the other hand, individuals residing in areas
that are smaller, more sparsely populated, and with more homogeneous groups will tend
to be more conservative, judgmental, parochial. and to maintain more personal
relationships. Empirical studies in this ﬁeld have consistently supported this theory.
Recently, for example, Carter and Borch (2005) found that people living in more urban
areas are likely to espouse more liberal attitudes toward the changing roles of women in
varying spheres of social life.

Finally, a prediction of “cosmopolitanism” can be inferred from a smaller
literature in political science on information effects and public preferences (see Althaus
1998; Bartels 1996; Delli Carpini and Keeter 1996; Gilens 2001). These studies use
statistical models to impute “fully informed” preferences to less informed respondents.
Each respondent’s political preferences are adjusted to match the predicted preferences of
a respondent who shares the same set of demographic characteristics but who also
possesses the highest level of political information. Thus, the imputed, fully-informed

preferences can be compared with the sample’s observed preferences. Althaus (1998)

56

used this approach to ascertain fully informed preferences across a wide range of issues.
In general, he found that these preferences were more liberal on a number of social
issues. Fully informed preferences included: more dovish attitudes on foreign policy, and
more support for abortion and gay rights. Attitudes toward government spending were
more ambivalent: fully informed preferences included opposition to the idea of “big
government,” but also a greater willingness to pay more taxes to pay down the deﬁcit and
to fund social services.

Based on these literatures, therefore, I expected that because electoral institutions
help shape the political information environment they should produce attitudinal and
cultural consequences similar to those we observe with other information mechanisms
(e. g., education). Electoral institutions that facilitate the expression of a greater number
of perspectives should help produce “more cosmopolitan” dispositions that lead to greater
support for sensitive issues from the “left” of the political spectrum and not necessarily
for those of “right.” I hypothesize, then, that citizens in countries with electoral
institutions that permit a more diverse political discourse (e.g., proportional
representation) will tend to have more positive attitudes toward sensitive issues and
policy positions promoted by the left (such as the environment and homosexuality) than
citizens in countries with electoral institutions that limit the diversity of the political
discourse.

It should be noted that I am not suggesting that this evolution of issues does not
occur in majoritarian systems. I am suggesting, rather, that the evolution is more rapid
under electoral institutions that facilitate the expression of diverse ideological

perspectives. All democratic systems permit the ﬂow of information, but electoral

57

institutions such as proportional representation should facilitate the discussion of
sensitive issues and the expression of sensitive policy positions on various issues, enough
to affect public consideration of such issues and positions, and enough that electoral
institutions can help explain the variation in attitudes towards sensitive issues that exists

across countries.

Electoral Institutions, Information Diversity, and Reactionary Preferences
As shown in the New Zealand study, electoral institutions inﬂuence not only the

perspectives to which citizens are exposed, but also majoritarian systems exaggerate the
prestige of reactionary issues and policy positions. This dynamic should be especially
relevant for issues tied to personal security, such as law and order. The question in this
case, then, is how the exaggeration of reactionary approaches will inﬂuence public
attitudes.

In this situation, 1 base expectations on a study by Druckman (2004), which
essentially shows how asymmetrical information can inﬂuence individuals. Druckman
showed that when they are only exposed to one frame individuals are susceptible to
framing effects. That is, their preferences can be altered based simply on how questions
are asked. When exposed to multiple frames at the same time (if they are exposed to
“counter-framing” by opposing elites, or if they engage in interpersonal debates that
introduce alternative perspectives), the framing effects are drastically mitigated or
disappear.

Thus, I expect that if individuals are being almost exclusively exposed to
reactionary positions on issues (as was the case in New Zealand’s political discourse on
law and order), then these individuals will be more likely to support these positions. I

58

hypothesize, then, that citizens in countries with maj oritarian systems will be more likely
to support reactionary policy positions. On an issue like law and order, this will mean
that citizens in majoritarian countries are more supportive of punitive approaches to

dealing with crime.

Aggregate Level Evidence

Measurement
To explore the cosmopolitanism hypothesis, I use data from the World Values Survey

(European Values Study Group and World Values Survey Association 2004) for the year
1999. Table 4.1 lists 16 issues appearing in the World Values Survey. These include
several “new politics” issues, such as protecting the environment and the justiﬁability of
homosexuality. For this time frame, I expect to ﬁnd differences in attitudes across
countries on several “new politics” issues. These are issues that have more recently
become salient—relative to more traditional economic issues which have been salient for
a much longer period of time. The list also includes a question addressing the tolerance
of immigration, an issue that should be promoted mainly by parties on the right of the
political spectrum. Analyzing this issue should provide insight into how messages from
the right affect public preferences. The list also includes four questions related to
economic issues. Examining these economic issues allows for comparison with more
traditional issues. Table 4.1 also shows the responses to the survey questions in the
binary form in which I recoded them, at the individual level. A value of 1 signiﬁed a
positive response and a value of 0 signiﬁed a negative response. The aggregate score for
each country. then. is the proportion of citizens that revealed a positive response to each

question.

59

Table 4-1 Dependent Variables

 

Disagree that protecting environment shouldn't cost me, so 1 = support for protection

 

Justiﬁability of homosexuality, 1 = support for

 

Justiﬁability of abortion, 1 = support for

 

Justiﬁability of divorce, 1 = support for

 

Justiﬁability of euthanasia, l = support for

 

Justiﬁability of taking marijuana, 1 = support for

 

Justiﬁability of experiments with human embryos, 1 = support for

 

Agree that marriage is outdated, 1 = yes

 

Disagree that men make better leaders, 1 = yes

 

Do not attend religious service once month or more, 1 = yes

 

Tolerant of immigration, 1 = yes

 

Should country provide more aid to poorer countries, keeping it the same or giving
more = 1

 

l = supmrt for incomes being more equal

 

1 = support for more government ownership of business and industry

 

1 = supmrt for government taking more responsibility

 

 

l = agree that competition is harmful

 

To explore the reactionary hypothesis, I use data from the International Crime

Victimization Survey (ICVS International Working Group et al. 2003) which asked

 

respondents what punishment they considered most appropriate for a recidivist burglar--a

man, aged 21, who is found guilty of burglary for the second time. Respondents could
choose between the following alternatives: community service, imprisonment, a ﬁne,
suspended sentence, other sentence, and don’t know. For the aggregate-level analysis, I
concentrate on the percentage of people in each country that support imprisonment. I

interpret this choice as reﬂecting a public preference for short-term solutions to the

problem of crime, rather than considering solutions that might address the root causes of

crime.

60

The independent variables used in the aggregate level analysis are the following:

Effective Threshold. Rather than relying on categorical measures that distinguish
between majoritarian, mixed-member, and proportional systems, this study employs
Lijphart’s (1994) effective threshold measure,13 which provides a continuous measure of
the type of electoral system employed by a country. The effective threshold can be
expected to affect diversity via its impact on the number of parties that win representation
and via the incentives it generates for political elites. The effective threshold is a more
reﬁned indicator of the barriers parties must overcome to win seats, taking into account
the average district magnitudel4 (the number of seats up for grabs in a district) and any
national legal threshold present in a country. As such, it is also a good indicator of the
incentives parties will face when they compete for support and thus it captures the degree
to which parties may need to moderate their policy agendas and messages in order to
receive enough votes to win legislative seats. More moderation and ambiguity will be
required in more maj oritarian systems (such as in the U nited States, which are coded as
requiring 35 percent of the vote in order for a party to win a seat in a single-member
district), while parties (and especially small parties) will be quite free to explore a wider
range of policy issues and positions in systems such as the Netherlands, where the

effective threshold is .67 percent. Data from the World Bank Database of Political

 

° 5 9/ . . . .
'3 Effective Threshold = 50 /° + i O 0 where M = the average district magnitude. If a national

(M +1) 2M
legal threshold represents a higher barrier to party entry, then that value is used instead.

 

'4 According to Cox ( 1990), district magnitude is the factor that may have the largest inﬂuence on
spatial incentives.

61

Institutions (Beck et al. 2000) are used to calculate the average effective threshold for the
lower chamber in each country over the years 1975 to 1999.15

Demographic Heterogeneity. Demographic heterogeneity can be considered
another source of diversity. In the small-groups literature its effects are found to be
ambiguous. For example, Jehn et al (1999) ﬁnd that information diversity tends to
improve group performance while value and demographic diversity have complicated
effects, and Polzer et al (2002) ﬁnd that when groups have high interpersonal congruence
(i.e. the congruence between how one views himself and how he is viewed by his group)
demographic diversity can enhance creative task performance. In macro-social contexts,
social heterogeneity has been found to negatively affect a public’s support of social
programs (Gilens 2000) and consequently wealth redistribution (Alesina and Glaeser
2004). To capture the effect of this factor, data are taken from Yeoh Kok Kheng’s (2001)
measure of ethnic fractionalization, which accounts for religious, racial, and linguistic
cleavages.

Years Free. The effects of electoral institutions require time to take hold. It is
possible, for example, that a country may have several political parties present in the

political system, but if these parties lack the democratic means to express their

 

'5 For this study, one adjustment is made to this measure. Rather than coding runoff systems as
requiring 35 percent of the vote (as done by Lijphart), they are coded here as requiring 20 percent
of the vote. This is done because runoff systems tend to increase the number of candidates
competing for ofﬁce (Shugart and Carey 1992), and should therefore encourage multi-partism
and a broader political discourse (especially during campaigns) at least more so than majoritarian
systems that rely solely on plurality rules. Runoff systems can also affect public expectations
regarding elite behavior. Westholm and Niemi argue that “systems with proportional
representation. . .or runoff elections. . .will force voters to think in terms of coalitions and second
choices” (1992, 31).

62

perspectives, then their impact will be signiﬁcantly limited. For this reason, the number
of years that a country is regarded as free (by Freedom House) is controlled for. For most
models the values are taken for the period from 1972 to 1999. For the civil union
legislation model, the values are taken for the period from 1972 to the year when the
country adopted the civil union legislation.

Economic Development. Modernization studies suggest that citizens with greater
material wealth begin thinking less about “lower order” issues such as economic security
and more about “higher order” issues such as quality of life and concern for the
environment (lnglehart and Welzel 2005). For this reason, economic development is
controlled for by including a measure of GDP per capita taken from the CIA World
Factbook for the year 1999.

Religious Background. Historically Roman Catholic societies have tended to
resist secularization even in the face of communist rule and socioeconomic development
(lnglehart and Welzel 2005, 75). This has direct consequences for issues such as the
adoption of civil union legislation. Protestant and Muslim backgrounds are also

controlled for.

Results
To analyze how political information systems inﬂuence public preferences across

countries, I regressed the proportion of positive responses for each issue, in each country,
on the log of the effective threshold—the log should reﬂect the expectation that there is a
diminishing rate of return from moving to more and more proportional electoral systems.

Tables 4.2 and 4.3 display the results of the regressions for a sample of countries

63

considered democratically free by Freedom House for at least 10 years. I only report the

regressions for which the effective threshold produced statistically signiﬁcant effects.

Table 4-2 Electoral Systems and Public Preferences

 

 

Environment Homosexuality Abortion Divorce Euthanasia
Effective threshold -0.062** -0.080*** ~0.046** -0.059*** -0.049**
(logged) (0.037) (0.026) (0.023) (0.021) (0.026)
Ethnic fractionalization -0.167 0.254“ 0.045 0237*” 0.226"
(0.198) (0.115) (0.101) (0.091) (0.108)
Years free -0.001 0.011M 0.006 0.011" 0.003
(0.008) (0.006) (0.005) (0.004) (0.005)
GDP per ca pita 0.010 0.007 0.004 -0.002 0.008”
(0.007) (0.004) (0.004) (0.004) (0.004)
Catholic -0.108 -0.135 -0.203**"‘ -0.269*** -O.289***
(0.105) (0.088) (0.078) (0.070) (0.082)
Protestant 0.154 0.082 0.086 -0.038 -0.016
(0.117) (0.093) (0.082) (0.074) (0.087)
Muslim 4.027" -1.633 -0.360 -2.771*** -1.584
(2.233) (1.202) (1.059) (0.955) (1.320)
Constant 0378*“ 0.175 0269*“ 0.495*** 0390*"
(0.142) (0.114) (0.100) (0.090) (0.107)
R2 0.619 0.703 0.698 0.641 0.684
N (countries) 26 35 35 35 34

 

* p s .1, ** p s .05, *** p s .01 (one-sided). Standard errors in parentheses.

We can see in the table that the effective threshold signiﬁcantly predicts attitudes

toward several “new politics” issues, including making material sacriﬁces to protect the

environment, and tolerance of: homosexuality, abortion, divorce, euthanasia,

experimentation on human embryos, attendance of religious services, and agreement with

the idea that marriage is outdated. As the effective threshold increases, as we move

toward greater majoritarianism, attitudes on all these issues become more negative.

Conversely, I interpret this pattern as suggesting that as we move towards more

proportional systems citizens become more cosmopolitan.

64

Table 4-3 Electoral Systems and Public Preferences

 

 

Embryonic Religious Marriage Immigration Prison for
Services Outdated Recidivist
Effective threshold -0.039** -0.063** -0.024* 0.003 7.831***
(logged) (0.017) (0.036) (0.017) (0.038) (2.999)
Ethnic fractionalization 0.359M 0.166 0188*" 0.183 1.625
(0.152) (0.148) (0.069) (0.168) (16.360)
Years free 0.036” 0.002 0.000 0.008 -1.110
(0.015) (0.007) (0.003) (0.008) (1.036)
GDP per ca pita -0.011** 0.003 0.003 -0.003 0.045
(0.005) (0.006) (0.003) (0.006) (0.820)
Catholic -0.009 -0.474*** 0.000 -0.040 -43.884***
(0.039) (0.112) (0.052) (0.119) (13.104)
Protestant 0.036 0.026 -0.055 -0.053 -33.322***
(0.040) (0.120) (0.056) (0.128) (11.573)
Muslim -3.784* -2.637* -0.406 -3.644** -68.179
(2.250) (1.812) (0.841) (1.983) (411.079)
Constant —0.632** 0856*" 0.164" 0.376“ 62.489***
(0.303) (0.147) (0.068) (0.156) (23.240)
R2 0.741 0.656 0.364 0.172 0.649
N (countries) 12 34 34 33 21

 

* p s .1, *" p s .05, "* p s .01 (one-sided). Standard errors in parentheses.

Substantively, a change in the effective threshold from its lowest value to its

highest value (which is equivalent to switching from a nearly pure proportional system

like the Netherland’s to a pluralitarian system like Great Britain's) is associated with a

6.2% drop in support for the environment, an 8% drop in support for homosexuality, a

4.6% drop in support for abortion, a 6% drop in support for divorce, a 5% drop in support

for euthanasia, a 4% drop in support for experimenting with human embryos, a 6.3%

drop in respondents who say they attend religious services once a month or more, and a

2.4% drop in agreement with the idea that marriage is outdated.

65

The effective threshold was not statistically related to any other of the 16 World
Value Survey issues considered here. For several of these issues, this ﬁnding may largely
reﬂect two phenomena. First, the lack of a relationship may suggest that some variables
were too controversial (in 1999), even for countries with low effective thresholds. For
example, it is possible that tolerance for prostitution will not increase until sometime in
the future, or it is possible that support for this issue may never increase, regardless of the
electoral system. Second, the lack of a relationship may also result from a process of
cross-national attitude convergence. A convergence of attitudes is more likely to occur
for issues that have been salient for a longer period of time (such as gender equality and
economic issues). Thus, the idea of gender equality has become prevalent in all advanced
democracies to a degree that the effective threshold will no longer be useful in predicting
attitudes toward this issue. Longitudinal analyses may help reveal if both of these
phenomena are indeed occurring, however, the required data is not readily available and
therefore such an approach should be addressed by future work.

In Table 4.3 we can also see that there is no statistical relationship between the
effective threshold and immigration, suggesting that though proportional representation
likely increases exposure to more intense anti-immigrant messages citizens in countries
with proportional representation are not more likely to hold anti-immigrant views than
citizens in majoritarian systems. In the sample of countries used here, Sweden has the
highest level of tolerance for immigration (70%), while Denmark has the lowest (31%).
Support in the United Kingdom is low (at 39%), while it is moderately high in the United
States (57%). This ﬁnding suggests that other factors are more relevant for explaining

attitudes toward immigration. More importantly, this result suggests that citizens are

66

making a distinction between sensitive ideas from the left of the political spectrum and
sensitive ideas from the right. This is only a preliminary glance at a topic that should be
addressed more extensively in future research, but the results do suggest that in a diverse
information environment, ideas on social issues from the left of the political spectrum
have an advantage over ideas coming from the right.

Table 4.3 also shows that countries with higher effective thresholds (or more
majoritarian electoral systems) tend to have higher levels of support for more punitive
approaches toward addressing crime. Citizens in more majoritarian countries were
signiﬁcantly more likely to recommend prison sentences for a man, aged 21, who is
found guilty of burglary for the second time. A change in the effective threshold from its
lowest value to its highest value (which is equivalent to switching from a nearly purely
proportional system to a pluralitarian system) is associated with a 7.8% increase in
support for incarceration. This result comports with the expectation that because
majoritarian institutions privilege punitive responses to crime and limit dissenting
positions on addressing crime citizens in countries with these systems will be more likely
to support punishment.

No other variable predicts the attitudes considered here as consistently as the
effective threshold, but it is interesting to note that ethnic fractionalization predicts
positive support for several issues, including: homosexuality, divorce, euthanasia,
embryonic experimentation, and the idea that marriage is outdated. This ﬁnding suggests
that ethnic fractionalization acts somewhat like ideological diversity; it also serves to
introduce individuals to diverse perspectives, especially on social issues, and this

introduction to different perspectives encourages cosmopolitanism.

67

As for the control variables, the religious variables generally work as expected,
though their inﬂuence is less consistent than that of the effective threshold. Catholicism
does signiﬁcantly decrease support for abortion, divorce, euthanasia, and it increases the
likelihood that respondents attend religious services. Catholicism also strongly reduces
support for the use of prison sentences to deal with recidivism. This result is likely
consonant with the Church’s tradition of promoting social justice. Respondents with a
Muslim background are more likely to oppose divorce, immigration, and (weakly)
embryonic experimentation, and they are more likely to attend religious services.
Muslims are also more likely to support protecting the environment, and this result may
be tied to religious principles which I have insufﬁcient knowledge to properly expound
on.

It should also be noted that economic development only affects euthanasia and
embryonic experimentation (in an unexpected direction) and democratic freedom only
affects homosexuality, divorce, and embryonic experimentation (all in the expected
direction). These results should be interpreted with caution because countries were in
part selected on these variables; most countries in the sample are at least relatively
wealthy and democratic. Moreover, these two variables, democratic freedom and
economic development, are known to be collinear.

Multilevel Analysis

The clear pattern that emerges from the previous section is that the electoral institutions
which increase the ﬂow of information (e.g., proportional representation) signiﬁcantly
increase support for social issues that have a more recent history of salience and that tend
to come from the left of the ideological spectrum. In this section I seek to take advantage

68

of the multilevel nature of the data to explore the effects of individual-level factors.”
Based on the results of the aggregate-level regressions, which ﬁt well with the insights
from modernization theory and the institutional mechanisms outlined in the previous
section, 1 construct a sensitive-issues-and-policy-positions (SIPP) scale that I then
analyze with a hierarchical linear model (HLM).l7

The SIPP scale is a summated rating scale composed of the issues for which the
effective threshold was a signiﬁcant predictor, including: the environment,
homosexuality, abortion, divorce, euthanasia, experimentation on human embryos,
attendance of religious services, and agreement with the idea that marriage is outdated.18
Not all the questions were asked in all of the countries, but I included all the questions in
the scale because summated rating scales “standardize” the scores: they divide a
respondent’s total score by the number of questions answered.19 Higher values on the
SIPP scale can be interpreted as indicating higher levels of support or tolerance toward
sensitive issues and sensitive policy positions, particularly those promoted by parties

from the left of the political spectrum.

 

'6 An HLM analysis also offers two key beneﬁts to this study. First, because they decompose the
relationship between variables into separate level-1 and level-2 components, HLMs reduce the
loss of information encountered with strictly aggregate-level analyses. Second, HLMs account
for the misestimation of standard errors that could result from the dependence of individual
responses within the same organization (Raudenbush and Bryk 2002).

'7 Unfortunately, because not all of the ﬁfteen questions were asked in all of the countries
analyzed, a factor analysis could not be conducted to determine the presence of attitude
dimensions that underlie the issues examined here. In any case, scales created by using those
techniques would likely miss the “sensitive issues dimension” that I am aiming to capture because
they do not account for changes in salience across time.

'8 The Cronbach’s alpha statistic for this scale is equal to 0.767, which suggests a high level of
reliability. Various versions of this scale (where one or two questions were removed or one or
two questions were added) produced similar alpha and regression results.

'9 I also attained similar results using variations of the scale, where I dropped various issues or

added others.
69

It should be repeated that the HLM analysis of this scale is not speciﬁcally meant
to examine the effects of the effective threshold; instead, it is meant to explore the effects
of individual-level predictors. Given that the scale is composed of issues for which the
effective threshold produced signiﬁcant effects, we should expect that the effective
threshold will have a signiﬁcant effect on the SIPP scale. It should be noted, however,
that the SIPP scale can be interpreted as telling us something about cultural change. The
scale created here is similar to indexes of postmodernism developed in the modernization
literature, including for example, Chamock and Ellis’ (2004) postmodernism index,
which they based on Inglehart’s suggested new, non-class based dimension. They
showed that, for Australia, a postmodernism index created from 23 items (including
issues such as euthanasia, gender equality, and the environment) is independent of the

traditional left-right dimension.20

Explanatory Variables

Controls are introduced for several individual-level demographic variables, available in
the World Values Survey. I explore the effects of the respondent’s gender, education,
age, and political interest, and religious background.2| In accordance with
modernization theory, it is expected that education increases a respondent’s exposure to

information and his/her capacity to process it, and should therefore increase levels of

 

20 Indeed, indexes of postmodernism that include responses to several issue items are likely less
controversial than a measure often used by lnglehart. Inglehart‘s measure of postrnaterialism--
based on either a single or multiple ranking exercise of four national aims (two materialist and
two postmaterialist) and which results in classiﬁcations of survey respondents as “materialist”,
“postmaterialist”, or “mixed”—has been criticized as invalid and as not representing an actual
value dimension (Davis and Davenport 1999; Davis et al. 1999).
2' Other interesting variables, such as the size of the respondent‘s town, were available for some
countries, but unfortunately they were not asked across a sufﬁcient number of countries.

70

cosmopolitanism or liberalism. Modernization theorists have also argued that cultural
change occurs through generational replacement (Abramson and lnglehart 1992) and we
should therefore expect that younger respondents will tend to be more cosmopolitan than
older respondents. I also expect that an interest in politics means that a respondent will
be exposed to higher levels of information and this variable should therefore be positively
associated with cosmopolitanism.
The education variable is a three-category variable. The lower category
(Education 1) includes:
1. Inadequately completed elementary education
2. Completed (compulsory) elementary education
3. (Compulsory) elementary education and basic vocational qualiﬁcation
The medium category (Education 2) includes:
4. Secondary, intermediate vocational qualiﬁcation
5. Secondary, intermediate general qualiﬁcation
6. Full secondary, maturity level certiﬁcate
The upper category (Education 3) includes:
7. Higher education - lower-level tertiary certiﬁcate
8. Higher education - upper-level tertiary certiﬁcate
Age is also a three category variable: 15-29 years (Age 1), 30-49 years (Age 2),
and 50+ years (Age 3). A respondent’s political interest is measured by a summated
rating scale created from responses to the following questions:
(A062) When you get together with your friends. would you say you

discuss political matters frequently, occasionally or never?

71

(E023) How interested would you say you are in politics?
(E150) How often do you follow politics in the news on television or on
the radio or in the daily papers?
For religious background, I now use individual-level data on Roman Catholic,
Protestant, and Muslim backgrounds. Recall that these religious factors are expected to

reduce the tendency toward secularization and cosmopolitanism.

Model Construction
Here, I outline the steps taken in constructing the HLM. The ﬁrst step is to run a one-

way ANOVA model with random effects to determine how much of the variation on the
sensitive issues scale lies within and between countries. The results indicate that 26
percent of the variance is between countries.22 Moreover, the ANOVA model shows that
the variability in country means is signiﬁcantly different from zero (x2=15158.686, p <
.001)

As a second step, I run a model with only the individual-level variables. This step
provides a baseline to compare the effects of individual level variables with country-level

variables. The individual level model is:
Yii = 80 + [31 *(FEMAI.E),J~ + [32*(AGE 2)” + 83*(AGE 3)”- + 04*(EDUCATION
2)ij + 85*(EDUCAT ION 3),J- + B6*(CATHOLIC)U + 137*(PROTESTANT)ij

+ B8*(M.USLIM),J- + 09*(POLITICAL INTEREST)ij + R

 

22 Proportion of variance in cosmopolitanism between countries = between group
variability/(between group variability + within-group variability) = 0.76712/(0.76712 + 2.21203)
= 0.26

72

where Yij is the score on the sensitive issues scale for an individual i in country j and [30
is the individual level intercept. The remaining coefﬁcients match the descriptions in the
previous section.

The third step is to introduce the country-level variables only for the model
intercept [30. This step will examine the effects of the country-level variables on the

sensitive issues scale. The country-level equation for the model intercept is:

B0 = m, + yo. *(EFFECTIVE THRESHOLD J-) + 702*(ETHNIC
FRACTIONALIZATIONJ) + 703*(601) CAPj) + y04*(YEARS FREEj) + 00,-

where 700 is the country-level intercept and yoj, yoz, y03, and y04 are the effects of the

country-level variables.

Finally, HLM permits us to explore whether the electoral institutions signiﬁcantly
predict the within-country slopes. That is, we can explore if countries with high effective
thresholds differ from countries with low effective thresholds in terms of the strength of
the association between an individual-level variable and scores on the sensitive issues
scale. Cross-level interaction terms can introduce model instability mainly because of the
multicollinearity they introduce. Therefore, this part of the analysis is limited to the
variables of highest theoretical interest. I explored the interaction of the effective
threshold on education. The question, then, is whether the effective threshold
signiﬁcantly inﬂuences how strongly education is associated with high scores on the
sensitive issues scale. Is the slope for education signiﬁcantly steeper in countries with

higher or lower effective thresholds? In other words, is the effect of education

73

signiﬁcantly stronger in countries with greater information diversity? The cross-level

interaction equation examined here is the following:

135 = 750 + y5| (EFFECTIVE THRESHOLD) + 0,,-

where y50 is the intercept for the higher education slope and Ysr is the coefﬁcient of the

effective threshold on this slope.

Results
Table 4.4 reports the results of the microlevel model (Model 1), the country-level model

(Model 2), and the full model with interaction effects (Model 3). The microlevel model
includes only the individual-level predictors, controlling for the multilevel nature of the
data structure. Here, all of the individual-level parameter estimates are consistent with
the expectations derived from modernization theory and other literatures.

Respondents with higher levels of political interest and education score
signiﬁcantly higher on the SIPP scale. Increasing political interest from the lowest score
to the highest increases the value of the SIPP scale by .4 points, or by about 6%. The
shift to a secondary education is associated with an increase in the SIPP scale of .32
points, or approximately 5%. Tellingly, the shift to a tertiary education is associated with
a considerably stronger increase (.6 points or 9%) in the SIPP scale.

Older respondents and respondents identifying as Catholic, Protestant, or Muslim
score signiﬁcantly lower on the SIPP scale. The shift to the second age group (30-49
years) is associated with an increase in the SIPP scale of .21 points, or approximately 3%,
while the shift to the third age group (50+ years) is associated with a considerably
stronger increase (.71 points or 10%) in the SIPP scale. Identifying as Catholic decreases

the SIPP score by .66 points or 9%; identifying as protestant decreases the SIPP score by
74

.49 points or 7%; and identifying as Muslim decreases the SIPP score by 1.5 points or
21%.

The variance components at the bottom of Table 4.4 can be used to assess
whether the introduction of the country-level variables in Models 2 and 3 improves our
ability to explain cross-national variations on the SIPP scale. Using the components in
Model 1 as a baseline, we can gauge the explanatory power of the country-level variables
by Observing the change in size of the components as we move from Model 1 to Models
2 and 3.

When the country-level variables are introduced, for Model 2, we can note ﬁrst
that the variance component of this model is considerably smaller than for Model 1.
With the introduction of the country-level variables there is a 32% reduction in the
unexplained cross-national variance over the baseline model. As expected, the effective
threshold is signiﬁcantly and negatively related to the SIPP scale (p < .05).
Substantively, a change in the effective threshold from its lowest value to its highest
value (i.e., moving towards a more majoritarian system) is associated with a decrease of
.28 points on the sensitive issues scale. This represents about 4% of the range on the
sensitive issues scale.

With respect to other country-level variables, only economic development
signiﬁcantly affects scores on the SIPP scale (p < .1). A ten thousand dollar increase in a
country’s GDP per capita increases an individual’s score on the scale by .4 points. This
result matches the ﬁndings in the modernization literature and should be considered

particularly robust given that the inﬂuence is signiﬁcant even though all of the countries

75

Table 4—4 Multilevel Model of Cosmopolitanism

 

 

Model 1 Model 2 Model 3
Microlevel Country-level Full Model with
Only Intercept Effects Interaction Effects
Macro-level

Effective threshold (logged) -0.278** (0.145) -0.222** (0.134)
Ethnic Fractionalization 0.210 (0.530) 0.209 (0.530)
Years free 0.038 (0.030) 0.037 (0.030)
GDP per capita 0.038* (0.027) 0.039* (0.026)

Individual-level
Political Interest 0.099***(0.030) 0098*" (0.030) 0100*“ (0.029)
Female 0.020 (0.022) 0.020 (0.022) 0.018 (0.023)
Age 2 -0.208*** (0.031) -0.208*** (0.031) -0.211*** (0.030)
Age3 -0.714*** (0.067) -0.714*** (0.067) -o.719*** (0.066)
Education 2 0324*“ (0.040) 0.324*** (0.040) 0326*" (0.038)
Education 3 0599*” (0.079) 0599*“ (0.079) 1086*” (0.278)
Effective Threshold -0.202** (0.108)
Catholic -0.662*** (0.070) -0.662*** (0.070) 0661*“ (0.070)
Protestant -0.485*** (0.082) —0.486*** (0.082) -0.484*** (0.082)
Muslim -1.541*** (0.632) -1.s41*** (0.633) -1.S36** (0.624)
Constant 3863*" (0.130) 2916*" (0.512) 2777*" (0.499)

Random Effect
Variance Component
Percent Explained
N (35 countries)
Degrees of freedom

x2

0.767

46885
34
14159.721

0.525
58.5
46885
30
8454.654

0.520
56.8
46885
30
8404.645

* p s .1, ** p s .05, *** p s .01 (one-sided). Robust standard errors in parentheses. HLM 6.04
was used to calculate restricted maximum likelihood estimates.

included in this study are at least moderately developed. As noted earlier in this chapter,

the results for democratic freedom should be interpreted with caution because countries

in the sample were in part selected on this variable and because democratic freedom is

known to be collinear with economic development.

76

The results for ethnic fractionalization suggest that this type of diversity does to
an extent differ from the ideological diversity produced by electoral institutions, at least
with respect to the SIPP scale. Ethnic fractionalization is positively correlated with the
SIPP scale, but the inﬂuence is not statistically signiﬁcant. Future work should give
greater attention to the distinction between ideological and ethnic diversity. It would be
particularly useful to focus on whether ethnic fractionalization produces signiﬁcant
results for other issue areas, especially those associated with ethnic interests and
competition.

The ﬁnal model, with the cross-level interaction of education and the effective
threshold, does not improve the variance explained over Model 2, and in fact, decreases
it. There is also little change in the coefﬁcients for most variables, though the magnitude
of the effective threshold declines from .28 to .22. Of primary interest for this model,
there is a tendency for countries with higher effective thresholds to have smaller within-
country slopes for higher education. This ﬁnding indicates that higher education is more
strongly related to SIPP scores in countries with lower effective thresholds. People with
higher education are even more likely to support sensitive issues and positions when they
reside in countries with institutions that facilitate the expression of diverse ideological
perspectives. This result is theoretically reasonable because it suggests that when two
mechanisms of information are combined, the effects on individual attitudes are

considerably stronger.

Conclusion
The ﬁndings of this chapter are to some extent preliminary because one time point is used

to explore what is likely a dynamic process that ideally should be analyzed with data

77

from many time points and over several decades. The evidence presented here is not rich
enough to detail the dynamic effects of electoral institutions on cultural change, but it is
strong enough to support the conclusion that among the many unintended consequences
of constitutional engineering lie its important effects on how and what citizens think.
This study shows that, via their inﬂuence on the diversity of information to which
citizens have access, electoral institutions inﬂuence citizen preferences even on particular
political issues.

This study also identiﬁes a speciﬁc mechanism by which modernization leads to
cultural change. Speciﬁcally, it suggests that information diversity is a key mechanism
contributing to cultural change. Though research on information processes has focused
on factors such as knowledge, education, and communications systems, diversity has
been largely overlooked in political science. Yet, diversity clearly increases the ﬂow of
information and facilitates the introduction of new information, and the ﬁndings in this
paper suggest that these forces signiﬁcantly inﬂuence public preferences and cultural
change.

Particularly, while information diversity means that the public will be exposed
more frequently to sensitive issues from the right and left of the political spectrum, it
appears that diverse information environments tend to favor the formation of
“cosmopolitan” or “postmodern” outlooks that are more favorable to the acceptance of
issues from the left. It is beyond the scope of this paper to establish possible cognitive
processes that explain these results, but it has been suggested by lnglehart and Welzel
(2005, Ch. 6) that the factors associated with sociO-economic development (such as

existential security and cognitive mobilization) help reveal the most fundamental human

78

motivations, including the desire to exercise autonomous choices. That is, humans have a
fundamental desire to be free, but this desire is suppressed by survival needs and a lack Of
cognitive mobilization.

With respect to the relationship between electoral institutions and other issue
areas, such as government intervention in the economy, I suggest that the lack of a
relationship is surprising. Cross-national studies of the welfare state show that countries
with fewer veto players and with proportional representation tend to have higher levels Of
wealth redistribution (Alesina and Glaeser 2004; Huber and Stephens 2001), and this
suggests that we should ﬁnd attitudes more supportive of government intervention in the
economy in countries with proportional representation. There are three plausible
explanations for the absence of relationship. First, as noted earlier, the survey data were
taken at a speciﬁc point in time, and this limits our ability to see how attitudes toward
government intervention in the economy changed over time. It is possible that economic
attitudes differed some decades ago and have now converged. This explanation would
suggest that we may eventually see a convergence on sensitive issues, as well.

The second explanation is that economic issues may not be dealt with through the
public opinion mechanism. That is, government policy on economic intervention may
not follow public preferences but instead may be shaped by political competition at the
elite level. Therefore, policy outcomes in areas such as economic intervention may not
match public opinion patterns. The third possibility is that the questions in the survey
simply did not capture public attitudes on government intervention. Addressing the ﬁrst

two explanations would likely be aided by time series data, which currently do not exist.

79

The third explanation may be addressed by the use of alternative cross-national survey
sources.

A ﬁnal point to consider is how differences in the rate of cultural change are
consequential to the policies of a nation. The importance Of the dynamics described in
this chapter is magniﬁed by the extent to which public opinion inﬂuences public policy.

In fact, this is the subject matter of chapter 5.

80

Chapter 5 Electoral Institutions and Policy Outcomes

We have seen that electoral institutions can inﬂuence the political information
environment and public opinion. I show in this chapter that the mechanisms and effects
identiﬁed in this project are also relevant for policymaking. Particularly, I argue that
these information mechanisms help explain two features of policy-making: policy
innovation (how quickly govemments take action on certain issues) and the tendency to
use reactionary policy responses.

As I have argued, diversity is a form of openness and, as such, it should affect a
society’s ﬂow of information. This implies that many policy issues will be debated
earlier in environments that include a more diverse range of perspectives, and that
consequently, certain policies will be adopted earlier in those societies. Because electoral
institutions signiﬁcantly inﬂuence the diversity of perspectives that a political system
brings to bear on policy problems. they should inﬂuence various processes that contribute
to policy innovation. They should inﬂuence how quickly a country’s political system
will debate particular problems, how quickly public attitudes will change on given policy
positions, and as a result, how quickly certain policies are adopted (or policy innovation).

This mechanism of diversity should most clearly apply to sensitive or
controversial issues that are more likely to receive attention in diverse systems. As
argued earlier, these sensitive issues largely include “new politics” issues, such as
environmental protection and homosexuality. I show here that countries with electoral
institutions that increase the number of perspectives included in the policy-making

process (i.e., proportional representation systems) tend to act earlier on these kinds of

81

issues. They tend to have higher scores on environmental performance indicators and
tend to be early adopters Of civil-union legislation.

Additionally, as demonstrated in previous chapters, majoritarian institutions not
only limit dissent, but they also encourage reactionary discourse and public preferences.
I show in this chapter that this mechanism helps explain why countries with majoritarian
systems are more likely to use reactionary policies, especially when dealing with issues
linked to personal security, such as crime. Thus, countries with more majoritarian
electoral systems tend to have higher incarceration rates and are more likely to use the

death penalty.

Electoral Institutions, Policy Innovation, and Reactionary Policies
Studies of policy innovation have by now identiﬁed many factors that lead entities (e. g.,

organizations, countries, or American states) to adopt various policies sooner than other
entities (for extensive reviews of policy innovation research see Berry (1994), Dobbin et
al. (2007), Dolowitz and Marsh (1996), and Wejnert (2002)). As noted by Dobbin et al.
(2007), an important category of factors contributing to policy innovation is associated
with information and social learning.23 The general theme in this category is that policy
actors change their beliefs as new evidence emerges. Studies based on this approach all
say something about the information environment and its effects largely on elites.
Information is identiﬁed as originating from a variety Of sources, including elite
networks, and observation or quasi-experimental evidence from an organization’s own

experiences, and more frequently from the experience of other organizations.

 

23 This branch of factors can be further subdivided into three approaches: the political science
approach on social knowledge, Bayesian learning, and channeled learning.
82

Despite this concern for the information environment, policy innovation research
has overlooked the importance of diversity. Yet, the diversity of perspectives expressed
in a political system clearly affects the kind of information that elites and citizens will
encounter and should therefore inﬂuence policy innovation. Indeed, it is a common
theme in the small-group literature that varied perspectives, knowledge, skills, and
approaches tend to increase a group’s level of creativity and innovativeness, under certain
conditions (see Milliken et al. (2003) and Nemeth and Nemeth-Brown (2003)).

The inclusion of more perspectives in the decision making process takes
advantage of human variation in interpretation and heuristics, and consequently increases
the probability that a person with the right perspective to solve a particular problem will
become available (Page 2007). This mechanism implies that breakthroughs and
innovations are the result of an individual seeing a problem in a novel way. Moreover, as
suggested by De Dreu and West (2001, 1192), diversity is linked to minority dissent,
which appears to prevent groupthink, or a premature movement tO consensus. Minority
viewpoints, then, are important because they stimulate divergent attention and thought
(N emeth 1986) and consequently deepen deliberation and lead to the identiﬁcation of
new ideas and solutions.

The diversity framework, then, appears particularly relevant to policy innovation
research. But it is also a framework tailor-made for cross-national studies of innovation.
Indeed, few settings present such a clear variation in the number of perspectives that are
included in decision making processes. I expect that because electoral systems affect
how quickly new issues and positions will be introduced to the arena of political

competition they will consequently affect how quickly certain policies will be adopted.

83

The ﬁndings in the previous two chapters tell us that we should expect electoral
systems to most clearly affect innovation on new politics issues. These are issues that for
the period analyzed here can be considered sensitive or difﬁcult to introduce. In this
chapter, attention is given to the protection of the environment and the extension of rights
to same-sex couples.24 These two issue areas are relatively difﬁcult for policy-makers to
address because they are imbued with controversy and political costs. Protecting the
environment requires asking the public to make material sacriﬁces while extending rights
to same-sex couples involves advancing the interests of a historically unpopular minority
group. Environmental sustainability, moreover, represents the type of problem that
requires long-term, “nonlinear” planning (identifying economic externalities, for
example).

Diverse systems can more readily inform their citizens of the long-term costs of
ignoring economic externalities and they can acquaint the public with the plight of a
historically unpopular minority. Under PR, we are more likely to see the emergence of
what Kitschelt (1988) calls “left-libertarian parties” (a category that includes green
parties) that routinely emphasize “quality Of life” and environmental sustainability over
increasing economic prosperity. These same minor parties are also most likely to take

controversial positions on social issues such as the rights Of same-sex couples. Under

 

3" To explore the relationship between electoral institutions and policy innovation, regressions
were run only for the issues of environmental performance and homosexuality. There is not yet
enough variation across the issue of euthanasia because at this time too few countries have
adopted relevant legislation. It should be noted, however, that the only two countries to formally
legalize euthanasia--The Netherlands and Belgium-~are countries with consensual electoral
institutions. For the issues of abortion and divorce, data from earlier time periods is required for
the independent variables. This data is not readily available, especially the data for electoral
institutions.

84

PR, therefore, we should see that minor parties, in particular, are more capable of taking
up difﬁcult issues and that their presence increases the likelihood that controversial
approaches to problems will become part of the policy-making calculus.

Chapters 3 and 4 also identiﬁed a second information mechanism—the
groupthink dynamic. As discussed, the groupthink dynamic limits dissent and thereby
privileges the promotion Of survivalistic messages, especially for issues linked to
personal security. This mechanism is most likely to emerge under majoritarian systems
because these systems lack the minor parties that can provide the dissenting perspectives
that mitigate groupthink. In fact, a clear example of this dynamic can actually be
observed in the current American presidential elections. On foreign policy, the two
major parties actually engage in a one-upsmanship dynamic, where they Offer tougher
and tougher approaches to dealing with international security threats.

In future work, I will explore how electoral systems inﬂuence foreign policy, but
here I concentrate on the issue of law and order. Again, in the United States, both
presidential candidates have staked tough-on-crime positions, including support for the
death penalty. And no dissenting candidates are receiving any signiﬁcant media
attention. Thus, I expect this groupthink dynamic to increase the tendency for a country
to rely on punitive approaches when dealing with crime and this should help explain why
countries with majoritarian institutions have higher rates of incarceration and why they
are more likely to use the death penalty.

Thus, the hypotheses explored in this chapter are the following:

H1: Electoral institutions that facilitate the expression of diverse

perspectives (i.e., PR) should have an advantage in addressing

85

controversial problems, and in particular they should be innovators on
issues such as environmental protection and the extension of rights to
historically unpopular minorities—in this case, same-sex couples.

H2: Majoritarian electoral institutions (i.e., those that use small district
magnitudes) should privilege survival-oriented solutions to problems and
this should be revealed most clearly by outcomes in the arena of law and
order, where we should see a greater reliance on punishment as a solution

to crime.

Measurement
Data for this study are taken from 21 to 60 countries, the number of observations

depending mainly on the availability of data for each of the dependent variables. The
independent variables are those used in Chapter 3, including: the effective threshold,
ethnic fractionalization, the number of years a country is considered democratic, GDP per
capita, and religious background (Catholic, Protestant, Muslim). Countries include a mix
of advanced democracies (especially for the smaller datasets) and developing
democracies. The sample also includes countries from several regions of the world.
There are four dependent variables analyzed in this chapter. As noted above, I
analyze environmental performance and tolerance of homosexuality because data is
readily available for these issues. Protecting the environment and extending rights to
same-sex couples can be considered difﬁcult to address because they tend to involve high
political costs-——costs which minor parties are more willing to bear and actually proﬁt

from.

86

The index used here to measure environmental performance is the Pilot
Environmental Performance Index (EPI) developed for 2002 by the Global Leaders of
Tomorrow World Economic Forum in collaboration with the Yale Center for
Environmental Law and Policy and the Center for International Earth Science
Information Network of Columbia University (2002). It is designed to measure
environmental results at a national level, by gauging performance on and rates of change
in air quality, water quality, greenhouse gas emissions, and land protection. The
extension of rights to same-sex couples is operationalized as the amount of time it takes a
country to adopt civil union legislation. The period under consideration begins in 1988
(the year before Denmark became the ﬁrst country to adopt civil union legislation) and
ends in 2006. Each country received a score based on the number of years it took to
adopt the legislation.

The impact of the survival-oriented dynamics produced by majoritarian systems is
explored in the arena of law and order. As a result of a greater emphasis on survival,
countries with majoritarian systems should tend to more frequently rely on punishment as
a solution to crime. Consequently, I expect them to have higher incarceration rates and
they should be more likely to rely on capital punishment. The data for incarceration rates
are taken from the International Centre for Prison Studies (2006) and are measured as the
number of people in prison per 100,000 of the national population. For most of the
countries the data are for 2004. The data for capital punishment come from Lijphart
(1999), who used a three-point scale to distinguish between countries that use capital

punishment for ordinary crimes (coded 2), countries that had capital punishment but only

87

for exceptional crimes or that had it but did not use it (coded l), and countries that did not
have it in law at all (coded 0).

To examine the relationships between electoral institutions and policy outcomes,
ordinary least squares is used to analyze the continuous dependent variables of
environmental performance and incarceration rates, and an ordered probit model is used
to analyze the categorical dependent variable of capital punishment. An event history
model is used to analyze the adoption of civil union legislation because the sample of
countries can be considered censored. Through 2006, only 21 countries had adopted civil
union legislation and this ﬁgure is likely to rise with the passage of time. Event history
models permit the modeling of the probability that an event will happen, given that,
through the period observed, the event has not happened (see Box-Steffensmeier and
Jones 2004). A Weibull hazard model is used to analyze a total of 60 countries (for
which data are available) because it is capable of handling the expectation that the hazard
rate changes as time passes. In this case, it is expected that once the ﬁrst nations have

adopted civil union legislation it should become easier for other nations to do so.

Results
Results for environmental performance and civil union legislation are reported in Table

5.1. In the case of environmental performance, the results of the OLS regression suggest
that the effective threshold has a negative, nonlinear (logged), and signiﬁcant impact (at
the .01 level). Holding all other explanatory variables constant, a move from a system of
pure proportional representation to a pure majoritarian system (equivalent to moving
from the Netherlands to Great Britain) would decrease the environmental performance
index by approximately 6%. Ethnic fractionalization, an alternative source of diversity,

88

has a positive inﬂuence on environmental performance, but it is only weakly signiﬁcant
at the .1 level

With respect to other control variables, only Protestant background has a
signiﬁcant effect on environmental performance. An explanation for this outcome that
takes account of the nature of each religious tradition cannot be offered here, but it is
possible that this result is linked to other cultural features associated with Northern
Europe. All other variables were not signiﬁcant and to some degree this is likely because
the effects of variables such as economic development and democratic freedom are
reduced by the use of a sample that over-represents wealthy, democratic countries, and by
collinearity.

Table 5.1 also demonstrates that the effective threshold has a negative and
signiﬁcant effect (at the .01 level) on the adoption of civil union legislation. The more
majoritarian a country the less like it is to adopt this kind of legislation. In particular, a
change from a pure proportional system to a pure majoritarian system is associated with a
64% decrease in the “hazard” or “risk” that a nation will adopt civil union legislation.25

Protestantism weakly predicts the adoption of civil union legislation (at the .1
level), and again, this may be related to cultural features associated with Northern
Europe. All other explanatory variables do not signiﬁcantly predict the adoption of civil
union legislation, even though this model does include more developing countries.

Again, democratic freedom and economic development may suffer from collinearity.

The duration parameter (p) for this model is signiﬁcantly above 1.0. at 2.83, indicating

 

25 The percentage change in the hazard associated with a unit change in a covariate is:
100*(Hazard Ratio — 1')
89

that the risk that a nation will adopt civil union legislation does indeed increase with time.

Once the ﬁrst country adopts civil union legislation it becomes easier for other countries

to do so.

Table 5-1 OLS Analysis of Environmental Performance and Event History Analysis of

Civil Union Legislation

 

Environmental Performance

Civil Union Legislation

 

 

Index (OLS) (Weibull Model)
B
Predictors B (Stndrdzd) [3 Hazard Ratio
Effective threshold (logged) -S.912*** -0.510 -1.025*** .359
(2.137) (0.322) (0.116)
Ethnic Fractionalization 19.753* 0.397 1.070 2.915
(12.097) (1.110) (3.236)
Years free 0.001 0.0003 - -
(through 1999) (0.447)
Years free - - 0.034 1.035
(before civil union) (.050) (0.051)
GDP per capita —0.031 -0.020 .020 1.020
(0 .439) (0.051) (0.052)
Catholic 10.300 0.333 -0.267 0.765
(6.266) (0.963) (0.737)
Protestant 30.624**"' 0.842 1.525“ 4.593
(7.111) (1.017) (4.671)
Muslim -461.114 -0.298 -.506 0.603
(299.339) (3.289) (1.983)
Constant 51.009*** - -8.275***
(9.542) (1.803)
p (duration parameter) NA 2.826
(0.561)
N (No. countries with civil- 21 60(21)
union)
R2 0.717 NA
Log-likelihood NA -30.935

 

* p .<. .1, ** p s .05, *** p s .01 (one—sided). Standard errors in parentheses.

Table 5.2 displays the regression results for the law and order issues. The

effective threshold is found to have a positive and signiﬁcant effect (at the .05 level) on

incarceration rates. That is, as we move towards a higher level of majoriatarianism the

90

incarceration rate increases. Moving from a pure proportional system to a pure
majoritarian system increases the incarceration rate by approximately 49%. This is a
remarkable result, considering that the sample includes Israel, a strong outlier. Israel has
one of the most proportional systems in the sample but it also has a relatively high
incarceration rate—likely related to national security issues. The effect of ethnic
fractionalization is not statistically signiﬁcant for this model, suggesting that social
diversity does not increase the tendency for incarceration. No other explanatory variables
have signiﬁcant effects on incarceration rates. Again, economic development and
democratic experience may not be signiﬁcant because the sample includes mainly
wealthy, democratic countries, and because these two variables are also known to be
collinear.

For the capital punishment model, the ordered probit estimates reveal that the
effective threshold has a positive and signiﬁcant effect (at the .01 level) on the
probability that a nation will employ capital punishment. Moving from the minimum
effective threshold (e. g., in the Netherlands) to its maximum (e.g., in the United States)
increases the predicted probability of using the death penalty by .59, holding all other
variables constant at their means. The effects of ethnic fractionalization are again not
signiﬁcant, suggesting that, while this form of diversity does not reduce the likelihood of
using the death penalty. it also does not increase it. All other control variables are not

signiﬁcant.

91

Table 5-2 OLS Analysis of Incarceration Rates and Ordered Probit Analysis of Capital
Punishment

 

 

 

Incarceration Rates Capital Punishment
(OLS) (Ordered Probit)
Predictors B [3
Effective threshold (logged) 49.087" --
(28.171)
Effective threshold - 0097*"
(0.029)
Ethnic fractionalization 46.505 0562
(116.555) (1.474)
Years free -3.672 0.019
(5.280) (0.116)
GDP per ca pita 3.066 0.005
(3.670) (0.044)
Catholic 32.038 0.144
(29.885) (0.327)
Protestant 24.924 -0.902
(78.875) (1.088)
Muslim 175.391 7.833
(794.265) (8.032)
Constant 37.438 --
(109.771)
N 44 36
R2 0.166 NA
Pseudo- R2 NA 0.325
Log-likelihood NA -22.796

 

* p s .1, ** p s .05, *** p s .01 (one-sided). Standard errors in parentheses.

Discussion and Conclusion

The argument presented in this chapter is that diversity inﬂuences a political system’s
ﬂow of information and therefore its capacity for policy innovation and its capacity to
limit the effects of groupthink and its concomitant reactionary policy making. It was
shown that countries with consensual institutions that encourage early debate on sensitive

issues (i.e., proportional representation) tend to act earlier on those issues. Moreover,

countries with majoritarian institutions that limit dissenting perspectives (i.e.,

92

pluralitarian systems) are more likely to rely on punitive policies to deal with issues like
crime.

It should be noted that compared to the results in chapter 4 the effects of
demographic diversity diverge from ideological diversity. Its effects are not signiﬁcant
for any of the models, in either direction, which indicates that demographic diversity
neither encourages nor hinders policy innovation, and it does not encourage the formation
of punitive policies for issues like crime. This suggests that, with respect to policy
making, demographic diversity does work differently from ideological diversity or
produces much weaker effects. Further research will be required to uncover if there are
conditions under which demographic diversity can encourage innovation or hinder it.
Further research may also identify issues for, and conditions under which, ethnic
diversity may produce conﬂict that damages policy problem solving. We also may wish
to focus on situations in which ideological diversity works differently from what has been
found in this project.

This chapter only scratches the surface of the importance of the diversity
framework for policy problem-solving. For example. it is highly likely that this
framework extends to a much broader range of issues and this may be revealed by greater
attention to the effects of time. My explorations into other issue areas were cursory and
did not account for how policy convergence reduces policy variation over time. Thus, for
example, it is possible that if time could be accounted for, the innovation patterns
observed in this chapter may be observed in issue areas such as welfare state innovation.
This was not possible to analyze with the public opinion data used in this study because

the earliest wave of the World Values Survey was collected in 1981. well after most

93

welfare states took on their characteristic forms. Historical data on electoral institutions
would also need to be gathered.

It was the aim of this chapter to pair public opinion phenomena with policy
actions, and this reduced the number of policy actions that could be examined. That is,
the issues examined here required public opinion and policy outcome data. It may be
possible to extend this analysis to several other issues, if public opinion data (which are
not readily available) need not be gathered. For example, the reactionary dynamic could
have more readily been applied to issues such as military spending. In the future, I will

likely rely more frequently on this approach.

94

Chapter 6 Conclusion

I have demonstrated in this project that, through their effects on the information
environment, electoral institutions inﬂuence important elements of policy making. They
inﬂuence the information available to citizens and elites and thereby affect their
preferences. They speciﬁcally inﬂuence how quickly certain policy issues and solutions
can be debated and thereby inﬂuence how quickly governments will respond to these
issues and solutions. This dynamic helps explain why countries with more proportional
electoral systems tend to act earlier on “new politics” issues such as protecting the
environment and extending rights to same-sex couples. Furthermore, electoral
institutions inﬂuence the prevalence of dissent and thereby inﬂuence the prevalence of
groupthink. Especially for issues connected to personal security, majoritarian systems
limit dissent and therefore favor the use of more aggressive, short-term policies. This
dynamic helps explain why maj oritarian systems tend to have high rates of incarceration,
more frequently rely on capital punishment to deter crime, and have publics that are more
willing to rely on punishment to deal with crime.

The mechanisms described here provide a step toward ﬂeshing out the black-box
that lies between political institutions and the numerous policy outcomes they are thought
to affect. To date, the literature on electoral institutions and policy outcomes has focused
on the behavior of elites and depended on correlations between measures of electoral
systems and policy outcomes. For example, a common mechanism advanced is based on

the assumption that policies can be divided into those that beneﬁt the population broadly

95

and those that beneﬁt narrow, geographic constituencies.26 In this view, electoral
institutions are said to inﬂuence which of these types of policies parties will favor. In
majoritarian systems political parties only need to win half the districts and to win each
district only requires a party to win half the vote. This will encourage parties to pursue
narrow, geographically oriented policies a) because they are trying to win over voters that
are distributed in this manner and b) because they do not need to maximize their share of
the vote (50% plus 1 is good enough). Under proportional representation a party has a
greater incentive to pursue broad-based policies. Parties in pure PR systems will not
have geographically distributed voters and they will seek to maximize their share of the
vote and will therefore favor policies that appeal to a broader portion of the electorate.
Evidence for this mechanism (and for others offered in the literature) tends to
come from correlations between measures of electoral systems and policy outcomes. My
information-based analysis shows how electoral institutions inﬂuence the information
environment. I also develop test expectations about public opinion—«an intermediary
factor in policy making-“from these patterns of information. And ﬁnally, I connect my
ﬁndings on the information environment and public opinion to speciﬁc policy outcomes.
Researchers have already hinted at the importance of diversity, but they have
conceptualized it more as collegiality (Gerring et a1 2005. Lijphart 1999). My project
suggests that we should focus on the number of perspectives that are expressed in the

deliberation process—e.g., in legislative discourse, in campaign discourse, and in the

 

26 Some examples of work employing this theme include: Persson and Tabellini (2003), Milesi-
F erretti et al. (2002), Lizzeri and Persico (2001), Alesina and Glaeser (2004), and Fredriksson and
Millimet (2004).

96

discourse that appears in the media. Steiner et al. (2004) have already developed a
Deliberation Quality Index (DQI) that measures the quality of deliberation in several
countries, and they ﬁnd that consensual (proportional) systems tend to produce higher
quality deliberation than majoritarian systems.27 The problem from my perspective is
that this index is aimed more at measuring collegiality rather than the diversity of

perspectives included deliberation.

Broader Implications
The mechanisms described here may also help explain and predict outcomes across a

broader range of policy areas. The dynamics described here could, for example, extend
to areas such as drug-policy, scientiﬁc research (such as genetic engineering), military
spending, and even international conﬂict. In fact, Leblang and Chan (2003) have already
found that democracies with proportional representation are signiﬁcantly less likely to
engage in international conﬂict than democracies that employ majoritarian systems.
Though the authors found that the relationship is strong, they suggested that more work
needs to be done to identify the mechanisms producing this result.

Many of the broader implications of the ideas outlined in this dissertation have a
normative bent, but ultimately these are questions worth asking and exploring in future
work—especially for those who might be interested in questions of cross-national
competitiveness. We can ask, then, if countries with institutions that permit the

expression of diverse perspectives possess advantages over countries that limit diversity.

 

27 The researchers also explored the impact that quality of deliberation had for policy outcomes.
They found that when quality of deliberation was high, support for legislation tended to be
higher. They also found that high quality deliberation is not associated with policies that produce
greater redistribution.

97

Do countries with institutions such as proportional representation diversity produce
societies that run more efﬁciently? Are these societies more sustainable in the long term?

For example, based on my ﬁndings, 1 suggest that Continental Europe has an
advantage over other regions of the world. These countries tend to employ policies that
focus on long term consequences rather than short term consequences. Thus, for
example, we can see that because Europe tends to spend more on social programs that
beneﬁt people early in life they then use fewer resources on law enforcement.
Conversely, my ﬁndings may help explain why countries like the United States struggle
with these policy areas. The US. is currently struggling with a skyrocketing prison
population, such that, according to a report from the Pew Center on the States (2008), the
US. now has one in every one hundred adults conﬁned in a jail or prison. Furthermore,
as the study suggests, the tough approaches to solving this problem have not produced a
clear impact on recidivism or overall crime. Indeed, many American states are ﬁnding it
difﬁcult to deal with the soaring costs of incarceration.

Another example is Continental Europe’s lead in protecting the environment.
While environmentalism has recently become more popular in the United States, such
that corporations and even universities are touting their green credentials, many countries
in Continental Europe have for decades allowed environmentalist parties to inﬂuence
public policy. The results are reﬂected in the Environmental Performance Index
(discussed in Chapter 5) but we can also see it in responses to global commitments, like
the Kyoto Protocol. While Sweden is ahead of schedule in meeting its obligations under

the treaty (Sweden 'is beating Kyoto emissions targets" 2007), majoritarian countries

98

such as the UK are struggling with the requirements (Black 2005) and the US. has yet to
sign on to the protocol.

It should be noted, in this discussion of implications, that most democracies are
open and allow for the emergence of a great degree of diversity, and we should therefore
expect that they all have a high capacity for coping with the complexities of policy-
making.28 Nevertheless, it is important to think about the long-term consequences of
differences in problem-solving capacities. Even if the effects are incremental, could large
consequences emerge over the course of several decades, such that countries PR systems
improve their position in the world on several fronts?

If diversity is ultimately deemed important to good governance, then the question
is: how do we procure it? Does this mean that countries should move to adopt more
proportional electoral systems, or can they ﬁnd other means of expanding diversity? Can
governments and media in majoritarian systems ﬁnd ways to include the perspectives
offered by minor parties? Great Britain, for example, already gives minor parties the
right to free air time during election campaigns. Is it sufﬁcient and feasible to expand
programs like these such that minor parties could become a regular part of a nation’s

political deliberation, even if they cannot win seats in their respective legislatures?

Loose Ends
The subject matter of this dissertation is substantial and labyrinthine. meaning naturally,

that many concerns have been overlooked. To begin with, a more complete analysis of
this topic will require ﬁeldwork. An extended stay in Northern Europe, for example, may

allow for the identiﬁcation of other important elements tied to the processes I have

 

23 This is probably true even in more authoritarian societies.
99

described. It would also allow for a closer analysis of efﬁciency, or how policies interact
to help societies function more effectively. Do social welfare and education policies
affect the long-term costs of law and order?

Another concern is that, while the small group literature has discovered that
diversity produces a number of beneﬁts, the literature has also discovered that diversity
produces many drawbacks. In fact, there are a wide range of factors—such as fear of
evaluation or breakdowns in communication—that can lead diverse groups to
underperform more homogeneous groups (Milliken et al. 2003). Fortunately, many of
these factors are not present in legislative deliberation-«indeed, for this reason I would
suggest that legislatures actually represent ideal places to study the dynamics of diversity.
For example, we are not likely to see legislators and parties fearing social judgment in
ways that produce breakdowns in legislative communication.

Nonetheless, it is still necessary to consider conditions under which diversity
might produce negative results in legislative deliberation. Future research may ﬁnd, for
example, that ideological and ethnic diversity produce negative results in less developed,
less democratic countries. In this study, ethnic diversity did not produce any
considerably negative results and, in fact, was more likely to produce positive results.
Yet, results from other studies do suggest that ethnic diversity can produce negative
outcomes. Ethnic diversity, for example, has been found to undermine wealth
redistribution (Alesina and Glaeser 2004; Gilens 2000) and is seen by a number of
political scientists as a hindrance to democratization (e.g., Dahl 1971; Lijphart 1977)—

though on this latter view a dissenting position is emerging (see Fish and Brooks 2004).

100

Causality

A particularly large concern is that the relationships described in this study involve some
degree of endogeneity. The hypothesized dynamic between the key variables, depicted in
Figure 6.1, likely involves a tri-angular loop, with a reciprocal relationship between
electoral institutions and public preferences, and a reciprocal relationship between public
preferences and policy outcomes. Indeed, there is a growing body of literature concerned
with the factors involved in the selection of institutions and whether the existence of
these factors suggests that institutional effects suffer from endogeneity (see Shvetsova
2003). Unfortunately, the use of statistical methods that can address endogeneity is
hampered by a lack of appropriate instrumental variables and the relatively small number

of countries for which data are available. Indeed, testing the full model in Figure 6.1 was

not possible with the data available for this study.

 

 

i Policy
(Outcomes i

    
    

i i
: Electoral !

l"— W ‘7

; Public

1 Preferences

 

,s t l
L ysem’r

 

L_____.-- ,_-

l
|

|
l

 

Figure 6.1 Relationship Between Institutions, Public Opinion, and Policy

101

 

I have, however, tested for simultaneity on the relationship between electoral
institutions and public preferences because this is the part of the model most likely to
raise questions about causality. A Hausman test of endogeneity showed that there was no
statistically signiﬁcant simultaneity between attitudes toward each of the issues explored
in this study and the electoral institutions.29 Theoretically, it is a reasonable to expect this
result because, as argued by the modernization literature, issues such as protecting the
environment and extending rights to same-sex couples may represent the emergence of a
new dimension of political competition that has only in the past several decades reached
the forefront of the political agenda. This is long after most of the countries in this study
established their electoral institutions. Moreover, for the issues considered here, the
changes in public preferences and in policy outcomes exhibit a pattern of cross-national
convergence. As argued by lnglehart (1997), support for the protection of the
environment and tolerance for homosexuality has increased over the past several decades
in all modernizing countries. Here, I am suggesting that these changes should unfold
earlier and more extensively in political systems that foster diversity.

Cox (1997, 19-27) and Duverger (1954, 218-220) have furthermore pointed to
evidence suggesting that electoral institutions independently affect societal behavior.

Cox controlled for the social diversity of a society by comparing the number of parties

that won seats in the house and senate of a country when the electoral institutions for

 

29 The instrumental variables used here for the effective threshold are geographical area and an
interaction between geographical area and ethnic fractionalization. Boix (1999) found that these
variables strongly predict the type of electoral system used by a country. The instrumental
variables fo.r public preferences include economic development, number of years a country is
free.

these chambers differed. He found that different electoral systems do produce different
party systems, even in the same society. Duverger noted that when proportional
representation was tried in local settings in the United States—which is known for its
rigid two-party structure—multiparty systems emerged (as in New York between 1936
and 1947). Thus, the effects of electoral institutions on social structure are likely more
consistent and systematic than are the effects of social structure on electoral institutions.

Finally, we can also gain some leverage on causality by examining historical
evidence. The extension of voting rights to women is an issue that became salient for
many democracies at the end of the 19th century. In Table 6.1 it is possible to observe
that the countries that became today’s multi-party systems (and that therefore tend to act
more readily on modern sensitive issues) have not always been at the forefront of policy
innovation. On the contrary, the ﬁrst two countries to grant women the right to vote
(New Zealand and Australia) employed majoritarian electoral systems, and several other
maj oritarian countries (Canada, Ireland, the United Kingdom, and the United States) were
not far behind. This indicates that at some point countries with majoritarian systems held
the “openness” (information ﬂow) advantage, likely because they tended to have been
democratic for a longer period of time.

Circa 1900, three of the four most democratic countries (with a polity score of 10)
were majoritarian countries, while most of the countries that became today’s multi-party
systems tended to have autocratic characteristics (indicated by the negative polity scores).
This suggests that the development of openness appears to proceed in a dialectical
manner. At one point, the development of free and fair elections provided countries with

a signiﬁcant openness advantage. More recently, the openness advantage appears to have

103

shifted to countries that have not only developed free and fair elections but that have also
developed institutions that foster the expression of diverse perspectives. Cultural

approaches do not easily account for this historical pattern.

 

 

Table 6-1 Women's Suffrage

Year Rights Polity Score
Countries Extended Circa 1900
New Zealand 1893 10
Australia * 1902 10
Finland 1906 8 (1917)
Norway * 1913 10
Denmark, Iceland 1915 -3
Canada 1918 9
Ireland * 1918 8 (1921)
United Kingdom * 1918 7
Austria 1918 -4
Georgia " 1918 NA
Germany 1918 1
Estonia 1918 8 (1917)
Kyrgyzstan 1918 NA
Latvia 1918 7 (1920)
Poland 1918 8 (1918)
Russian Federation 1918 ~10
Belarus, 1919 NA
Belgium * 1919 6
Luxembourg 1919 NA
Netherlands 1919 -2
Sweden * 1919 -4
Ukraine 1919 NA
United States 1920 10
Albania 1920 -2
Czech Republic 1920 7 (1918)
Slovakia 1920 7 (1918)

 

Source: Inter-parliamentary Union (1997).

Polity scores are for the year 1900 or the earliest year for which data are
available for a country. The dates for the latter scores are noted in
parentheses.

* Right subject to conditions or restrictions

104

REFERENCES

Abramson, Paul R., and Ronald lnglehart. 1992. "Generational Replacement and Value

Change in Eight West European Societies." British Journal of Political Science
22 (2):]83-228.

Alesina, Alberto, and Edward L. Glaeser. 2004. Fighting poverty in the US and Europe:
a world of difference. Oxford: Oxford University Press.

Althaus, Scott L. 1998. "Information Effects in Collective Preferences." American
Political Science Review 92 (3):545-58.

Asch, Solomon E. 1955. "Opinions and social pressure." Scientiﬁc American 193 (5):31-
5.

 

. 1956. "Studies of independence and conformity: A minority of one against a
unanimous majority." Psychological Monographs 70 (9):1-70.

Banducci, Susan A., Todd Donovan, and Jeffrey A. Karp. 1999. "Proportional
representation and attitudes about politics: results from New Zealand." Electoral
Studies 18 (4):533-55.

Banducci, Susan A., and Jeffrey A. Karp. 1999. "Perceptions of Fairness and Support for
Proportional Representation." Political Behavior 21 (3):217-38.

Barker, Fiona, Jonathan Boston, Stephen Levine, Elizabeth McLeay, and Nigel S.
Roberts. 2001. "An Initial Assessment of the Consequences of MMP in New
Zealand." In Mixed-Member Electoral Systems: The Best of Both Worlds?, ed. M.
S. Shugart and M. P. Wattenberg. Oxford: Oxford University Press.

Bartels, Larry M. 1996. "Uninforrned Votes: Information Effects in Presidential
Elections." American Journal of Political Science 40 (1)2194-230.

Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer, and Patrick Walsh. 2000.
"New Tools and New Tests in Comparative Political Economy: The Database of
Political Institutions." World Bank.

Berry, Frances Stokes. 1994. "Sizing Up State Policy Innovation Research." Policy
Studies Journal 22 (3):442-56.

Black, Richard. 2005. UK could ’miss Kyoto gas target' BBC News; 1 April, 2005;
available from http://news.bbc.co.rrk/2/hi/science/nature.’4399323.stm: Internet;
accessed 15 September 2008.

 

105

Blais, A., and R. K. Carty. 1987. "The impact of electoral formulae on the creation of
majority governments." Electoral Studies 6 (3):209-18.

Boix, Carles. 1999. "Setting the Rules of the Game: The Choice of Electoral Systems in
Advanced Democracies." American Political Science Review 93 (3):609-24.

Boston, Jonathan. Stephen Church, Stephen Levine, Elizabeth McLeay, and Nigel S.
Roberts. 2000. Left Turn: The New Zealand General Election of 1 999.
Wellington: Victoria University Press.

Bowler, Shaun, Todd Donovan, and David Brockington. 2003. Electoral reform and
minority representation: local experiments with alternative elections. Columbus:
Ohio State University Press.

Box-Steffensmeier, Janet M., and Bradford S. Jones. 2004. Event history modeling: a
guide for social scientists. Cambridge; New York: Cambridge University Press.

Budge, Ian, Hans-Dieter Klingemann, Andrea Volkens, Judith Bara, and Eric
Tanenbaum. Mapping Policy Preferences: Estimates for Parties, Electors, and
Governments 1945-1998. Oxford University Press 2001 [cited.

Carter, J. Scott, and Casey A. Borch. 2005. "Assessing the Effects of Urbanism and
Regionalism on Gender-Role Attitudes, 1974-1998." Sociological Inquiry 75
(4):548-63.

Centre for Prison Studies International. 2006. "World Prison Brief ."

Chang, Eric C. C., and Miriam A. Golden. 2006. "Electoral Systems, District Magnitude
and Corruption." British Journal of Political Science 37 (01):] 15-3 7.

Chamock, David, and Peter Ellis. 2004. "Postmaterialism and Postmodemization in
Australian electoral politics." Electoral Studies 23 (1):45-72.

Cox, Gary W. 1990. "Centripetal and Centrifugal Incentives in Electoral Systems."
American Journal of Political Science 34 (4):903-35.

. 1997. Making votes count: strategic coordination in the world's electoral
systems. Cambridge; New York: Cambridge University Press.

 

Curran, James, and Myung-Jin Park. 2000. De-Westernizing Media Studies. London;
New York: Routledge.

Dahl, Robert Alan. 1971 . Polyarchy; participation and opposition. New Haven: Yale
University Press.

106

Davis, D. W., and C. Davenport. 1999. "Assessing the Validity of the Postmaterialism
Index." The American Political Science Review 93 (3):649-64.

Davis, Darren W., Kathleen M. Dowley, and Brian D. Silver. 1999. "Postmaterialism in

World Societies: Is It Really a Value Dimension?" American Journal of Political
Science 43 (3):935-62.

De Dreu, Carsten K., and Michael A. West. 2001. "Minority dissent and team innovation:
the importance of participation in decision making." J Appl Psychol 86 (6):]191-
201.

Delli Carpini, Michael X., and Scott Keeter. 1996. What Americans know about politics
and why it matters. New Haven, Conn.: Yale University Press.

Dobbin, Frank, Beth Simmons, and Geoffrey Garrett. 2007. "The Global Diffusion of
Public Policies: Social Construction, Coercion, Competition, or Learning?"
Annual Review of Sociology 33 :1—24.

Dolowitz, David, and David Marsh. 1996. "Who learns what from whom: a review of the
policy transfer literature." Political Studies 44 (2):343-57.

Dow, Jay K. 2001. "A comparative spatial analysis of majoritarian and proportional
elections." Electoral Studies 20 (1): 109.

Downs, Anthony. 195 7. An economic theory of democracy. New York: Harper.
Druckman, James N. 2004. "Political Preference Formation: Competition, Deliberation,

and the (Ir)relevance of Framing Effects." American Political Science Review 98
(4):67l-86.

Duverger, Maurice. 1954. Political parties, their organization and activity in the modern
state. London, New York: Methuen; Wiley.

European Values Study Group and World Values Survey Association. 2004.
"EUROPEAN AND WORLD VALUES SURVEYS INTEGRATED DATA
FILE, 1999-2002, RELEASE I [Computer ﬁle]. 2nd ICPSR version. Cologne,
Germany: Zentralarchiv fur Empirische Sozialforschung (ZA)/Tilburg,
Netherlands: Tilburg University/Amsterdam, Netherlands: Netherlands Institute
for Scientiﬁc Information Services (NIWI)/Madrid, Spain: Analisis Sociologicos
Economicos y Politicos (ASEP) and JD Systems (JDS)/Ann Arbor, MI: Inter-
university Consortium for Political and Social Research producers], 2004.
Cologne, Germany: Zentralarchiv ﬁrr Empirische Sozialforschung (ZA)/Madrid,
Spain: Analisis Sociologicos Economicos y Politicos (ASEP) and JD Systems
(J DS)/Ann Arbor, MI: Inter-university Consortium for Political and Social
Research [distributors]."

107

Fish, M. Steven, and Robin S. Brooks. 2004. "Does Diversity Hurt Democracy?" Journal
of Democracy 15 (1).

F redriksson, Per G., and Daniel L. Millimet. 2004. "Electoral rules and environmental
policy." Economics Letters 84 (2):237-44.

Gerring, John, Strom C. Thacker, and Carola Moreno. 2005. "Centripetal Democratic
Governance: A Theory and Global Inquiry." American Political Science Review
99 (4):567—81.

Gibbons, Matthew. 2003. "Institutional Change and Political Cleavages in New Zealand."
Paper presented to the European Consortium for Political Research. Edinburgh,
Scotland

Gilens, Martin. 2000. Why Americans hate welfare: race, media, and the politics of
antipoverty policy. Pbk. ed. Chicago: University of Chicago Press.

 

. 2001. "Political Ignorance and Collective Policy Preferences." The American
Political Science Review 95 (2):379-96.

Gordon, Stacy B., and Gary M. Segura. 1997. "Cross-National Variation in the Political
Sophistication of Individuals: Capability or Choice?" Journal of Politics 59
(1)1126-47.

Grofman, Bernard. 2004. "Downs and Two-Party Convergence." Annual Review of
Political Science (7):25-46.

Hackman, J. Richard, and Charles G. Morris. 1975. "Group tasks, group interaction
process, and group performance effectiveness: A review and proposed
integration." Advances in experimental social psychology 8:45-99.

Hallin, Daniel C., and Paolo Mancini. 2004. Comparing Media Systems: Three Models of
Media and Politics. Cambridge, UK; New York: Cambridge University Press.

Hayward, Janine, and Chris Rudd. 2000. "Metropolitan Newspapers and the Election." In
Left Turn: The New Zealand General Election of 1 999, ed. J. Boston, S. Church,
S. Levine, E. McLeay and N. S. Roberts. Wellington: Victoria University Press.

Heidenheimer, Arnold J. 1973. "The Politics of Public Education, Health and Welfare in
the USA and Western Europe: How Growth and Reform Potentials Have
Differed." British Journal of Political Science 3 (3)2315-40.

Hoffman, L. Richard, and Norman R. F. Maier. 1961. "Quality and acceptance of
problem solutions by members of homogeneous and heterogeneous groups."
Journal of Abnormal Social Psychology 62 (2):401-7.

108

Holland, John H., Keith James Holyoak, Richard E. Nisbett, and Paul Thagard. 1989.
"Induction processes of inference, learning, and discovery." Cambridge,
Massachusetts: MIT Press.

Hong, Lu, and Scott E. Page. 2004. "Groups of Diverse Problem Solvers Can Outperform
Groups of High-Ability Problem Solvers." Proceedings of the National Academy
of Sciences of the United States of America 101 (46).

Huber, Evelyne, and John D. Stephens. 2001. Development and crisis of the welfare
state: parties and policies in global markets. Chicago: University of Chicago
Press.

ICVS International Working Group, Anna Alvazzi del Frate, Jan J .M. van Dijk, John van
Kesteren, and Pat Mayhew. 2003. INTERNATIONAL CRIME VIC TIMIZA TION
SURVEY (1C VS), 1989-2000 [Computer ﬁle]. ICPSR version. Netherlands:
University of Leiden/Turin, Italy: United Nations Interregional Crime and Justice
Research Institute (UNICRI) [producers], 2002. Ann Arbor, MI: Inter-university
Consortium for Political and Social Research [distributor], 2003.

lnglehart, Ronald. 1990. Culture shift in advanced industrial society. Princeton, NJ:
Princeton University Press.

 

. 1997. Modernization and postmodernization: cultural, economic. and political
change in 43 societies. Princeton, NJ: Princeton University Press.

 

. 2000. "Globalization and Postmodern Values." Washington Quarterly 23
(1):215-28.

lnglehart, Ronald, and Christian Welzel. 2005. Modernization, cultural change, and
democracy: the human development sequence. New York: Cambridge University

Press.

Inkeles, Alex. 1983. Exploring Individual Modernity. New York: Columbia University
Press.

Inkeles, Alex, and David Smith. 1974. Becoming Modern: Individual Changes in Six
Developing Societies. Cambridge, MA: Harvard University Press.

Inter-parliamentary Union. 1997. "Men and Women in Politics: Democracy Still in the
Making: A World Comparative Study." Geneva: Inter-parliamentary Union, 1997.

Janis, Irving L. 1972. Groupthink: Psychological Studies of Policy Decisions and
Fiascoes. Boston, MA: Houghton Mifﬂin.

109

Jehn, Karen A., Gregory B. Northcraft, and Margaret A. Neale. 1999. "Why Differences
Make a Difference: A Field Study of Diversity, Conﬂict, and Performance in
Workgroups." Administrative Science Quarterly 44 (4):741-63.

Karp, Jeffrey A., Jack Vowles, Susan A. Banducci, and Todd Donovan. 2002. "Strategic
voting, party activity, and candidate effects: testing explanations for split voting
in New Zealand's new mixed system." Electoral Studies 21 (1): 1-22.

Katz, Richard S. 1997. Democracy and elections. New York: Oxford University Press.

Kheng, Yeoh Kok. 2001. "Towards an Index of Ethnic Fractionalization." In FEA
Working Paper No. 2001-3.

Kitschelt, Herbert P. 1986. "Political Opportunity Structures and Political Protest: Anti-

Nuclear Movements in Four Democracies." British Journal of Political Science
16 (1):57-85.

 

. 1988. "Left-Libertarian Parties: Explaining Innovation in Competitive Party
Systems." World Politics: A Quarterly Journal of International Relations 40
(2): 194-234.

Klingemann, Hans-Dieter, Andrea Volkens, Judith L. Bara, Ian Budge, and Michael D.
McDonald. 2006. Mapping Policy Preferences 11: Estimates for Parties, Electors,
and Governments in Eastern Europe, European Union, and OECD 1990-2003.
New York: Oxford University Press.

Leblang, David, and Steve Chan. 2003. "Explaining Wars Fought by Established
Democracies: Do Institutional Constraints Matter?" Political Research Quarterly
56 (4):385-400.

Lemer, Daniel. 1958. The Passing of Traditional Society: Modernizing the Middle East.
New York: Free Press.

Lijphart, Arend. 1969. "Consociational Democracy." World Politics 21 (2).

 

. 1977. Democracy in plural societies : a comparative exploration. New Haven:
Yale University Press.

 

. 1984. Democracies: Patterns of Majoritarian and Consensus Government in
Twenty-One Countries. New Haven: Yale University Press.

 

-. 1994. Electoral systems and party systems: a study of twenty-seven democracies,
I 945-1990. Oxford; New York: Oxford University Press.

110

 

. 1999. Patterns of democracy: government forms and performance in thirty-six
countries. New Haven: Yale University Press.

Lin, Tse-min, James M. Enelow, and Han Dorussen. 1999. "Equilibrium in
Multicandidate Probabilistic Spatial Voting." Public Choice 98 (1-2):59-82.

Linz, Juan J ., and Arturo Valenzuela. 1994. The Failure of presidential democracy.
Baltimore: Johns Hopkins University Press.

Maslow, Abraham H. 1943. "A Theory of Human Motivation." Psychological Review
50:370-96.

McGregor, Judy, Susan Fountaine, and Margie Comrie. 2000. "From Contest to Consent:
The Impact of Public Journalism on New Zealand Campaign Coverage." Political
Communication 13:133-48.

Milliken, Frances J ., Caroline A. Bartel, and Terri R. Kurtzberg. 2003. "Diversity and
Creativity in Work Groups: A Dynamic Perspective on the Affective and
Cognitive Processes That Link Diversity and Performance." In Group Creativity:
Innovation Through Collaboration, ed. P. B. Paulus and B. A. Nij stad. New York:
Oxford University Press.

Monroe, Alan D. 1983. "American Party Platforms and Public Opinion." American
Journal of Political Science 27 (l):27—42.

Nemeth, Charlan J. 1986. "Differential contributions of majority and minority inﬂuence."
Psychological Review 93 (l):23-32.

Nemeth, Charlan Jeanne, and Brendan Nemeth-Brown. 2003. "Better than Individuals?

The Potential Beneﬁts of Dissent and Diversity." In Group Creativity: Innovation
Through Collaboration, ed. P. B. Paulus and B. A. Nijstad. New York: Oxford
University Press.

Niemi, Richard G., Guy Whitten, and Mark N. Franklin. 1992. "Constituency
Characteristics, Individual Characteristics and Tactical Voting in the 1987 British
General Election." British Journal of Political Science 22 (2):229-40.

Nisbett, Richard E., and Lee Ross. 1980. Human inference: strategies and shortcomings
of social judgment. Englewood Cliffs, NJ: Prentice-Hall.

Norris, Pippa. 2004. Electoral engineering: voting rules and political behavior.
Cambridge, UK; New York: Cambridge University Press.

Page, Benjamin 1., and Robert Y. Shapiro. 1983. "Effects of Public Opinion on Policy"
The American Political Science Review 77 (l): 175- 90.

111

Page, Scott E. 2007. The Difference: How the Power of Diversity Creates Better Groups,
Firms, Schools, and Societies. Princeton and Oxford: Princeton University Press.

Pereira, Carlos, and Salomon Orellana. 2008. "Hybrid Political Institutions and
Govemability: The Budgetary Process in Brazil." Working Paper.

Persson, Torsten, and Guido Enrico Tabellini. 2003. The economic effects of
constitutions. Cambridge, Mass: MIT Press.

Pilot Environmental Performance Index. 2002. An Initiative of the Global Leaders of
Tomorrow Environment Task Force, World Economic Forum.
wvwvciesincolumbia.edit/indicatorst'ESI.

Polzer, Jeffrey T., Laurie P. Milton, and William B. Swann, Jr. 2002. "Capitalizing on

Diversity: Interpersonal Congruence in Small Work Groups." Administrative
Science Quarterly 47 (2):296-324.

Poole, Keith T., and Howard Rosenthal. 1984. "US. Presidential Elections 1968-80: A
Spatial Analysis." American Journal of Political Science 28 (2):282-312.

Popper, Karl Raimund. 1971. The open society and its enemies: Vol. 2: The high tide of
prophecy .' Hegel, Marx, and the aftermath. 5th ed. Princeton, NJ: Princeton
University Press.

Powell, G. Bingham. 2000. Elections as instruments of democracy: majoritarian and
proportional visions. New Haven, CT: Yale University Press.

Przeworski, Adam. 2000. Democracy and development: political institutions and well-
being in the world, 1950-1990. Cambridge; New York: Cambridge University
Press.

Rabinowitz, George, and Stuart Elaine Macdonald. 1989. "A Directional Theory of Issue
Voting." The American Political Science Review 83 (1)193-121.

Rae, Douglas W. 1971. The political consequences of electoral laws. Rev. ed. New
Haven: Yale University Press.

Raudenbush, Stephen W., and Anthony S. Bryk. 2002. Hierarchical linear models:
applications and data analysis methods. 2nd ed. Thousand Oaks: Sage
Publications.

Rokeach, Milton. 1960. The open and closed mind. New York: Basic Books.
Senge, Peter M. 2006. Theﬁfth discipline: the art and practice of the learning
organization. Rev. and updated. ed. New York: Doubleday/Currency.

112

Shugart, Matthew Soberg, and John M. Carey. 1992. Presidents and assemblies:
constitutional design and electoral dvnamics. Cambridge [England]; New York:
Cambridge University Press.

Shvetsova, Olga. 2003. "Endogenous Selection of Institutions and Their Exogenous
Effects." Constitutional Political Economy 14 (3):]91.

Stepan, Alfred, and Cindy Skach. 1993. "Constitutional Frameworks and Democratic
Consolidation: Parliamentarianism versus Presidentialism." World Politics 46
(1)21-22.

Sweden 'is beating Kyoto emissions targets '. 2007. The Local: Sweden's News in English;
18 December 2007; available from http://www.theloeal.se/9444/20071218/:
Internet; accessed 15 September 2008.

 

Taagepera, Rein, and Matthew Soberg Shugart. 1989. Seats and votes: the effects and
determinants of electoral systems. New Haven: Yale University Press.

The Pew Center on the States. 2008. "One in 100: Behind Bars in America 2008."
http://uvwv.pewcenteronthestates.org/repon detail.aspx?id=35904.

Vowles, Jack, and Peter Aimer. 1993. Voters’ Vengeance: The 1990 Election in New
Zealand and the Fate of the Fourth Labour Government. Auckland: Auckland
University Press.

Vowles, Jack, Peter Aimer, Susan Banducci, and Jeffrey Karp. 1998. Voters' Victory?
New Zealand’s First Election Under Proportional Representation. Auckland:
Auckland University Press.

Vowles, Jack, Peter Aimer, Jeffrey Karp, Susan Banducci, Raymond Miller, and Ann
Sullivan. 2002. Proportional Representation on Trial: The 1999 Election in New
Zealand and the Fate of MMP. Auckland: Auckland University Press.

Wejnert, Barbara. 2002. "Integrating Models of Diffusion of Innovations: A Conceptual
Framework." Annual Review of Sociology 28:297-326.

Westholm, Anders, and Richard G. Niemi. 1992. "Political Institutions and Political
Socialization: A Cross-National Study." Comparative Politics 25 (1)225-41.

Wirth, Louis. 1942. "Urbanism as a Way of Life." The American Journal of Sociology 44
(1)11-24.

113

llIlliillllIlllllllilllllillllllllllillilllllllllll