1 '3
57.251:

m.»
.1 ti“!
1?; I
I... at
.F

.2 ..
019....

53:1: 99...“. .
3.5% 5m
‘2 at... 1-" .-
1! inc;- v 4 -
«wiggawmu
.. a. 5.
:I v.

. .5.»M

55.11..
.1311 5.

IR!
II\ 3.}. .I,
33‘“:
.r.‘ \
.3...

.n

1

wk}. “.15.?” 131:.
.I.. s .35
2 at . b A:
It)!
a
OI
(50!.
.. :0
i955.» L,
5119:"? .
v.
~11...

I r
1. tr.
1.:
351.67. 1
.2.

 

u. . .1.

"|‘av pl. . ’0.“

7.5. .. I . ,.

55...... 5.. . .55:
.- I

§,tfl1814

 

 

 

if.

 

, .
infinfi

 

. .dl. ,3: 11.51.:
3‘..." 93.2".qu :3 I a 5...:

 

 

Wm LIBRARY
" . Michigan State
200‘” l _ University

 

 

 

This is to certify that the
dissertation entitled

INSTITUTIONS, PARTISAN STATUS, AND CITIZENS’
POLITICAL SUPPORT IN EMERGING DEMOCRACIES

presented by

WONBIN CHO

has been accepted towards fulfillment
of the requirements for the

Ph. D. degree in Political Science

 

 

 

Major Professor’s Signature

m/zs/os

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

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
SEP 2 9 2008
04 2n? 'n a

 

'

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/C|RCIDateDue.indd-p.1

 

INSTITUTIONS, PARTISAN STATUS, AND CITIZENS’ POLITICAL
SUPPORT IN EMERGING DEMOCRACIES

BY

WONBIN CHO

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Political Science

2005

ABSTRACT

INSTITUTIONS, PARTISAN STATUS, AND CITIZENS’ POLITICAL
SUPPORT IN EMERGING DEMOCRACIES

By

Wonbin Cho

Do political institutions affect political support? This dissertation takes
political institutions into account to explain citizens’ political support in emerging
democracies. By combining individual-level survey data collected from 35
developing countries in sub-Saharan Africa, Latin America, and Asia, with
national-level data on political institutions, this study demonstrates party systems
have a significant impact on popular evaluations of political system performance.
Popular satisfaction with democracy is increased in countries that have greater
effective number of parliamentary parties (ENPP), where citizens are likely to have
more options to choose from. In addition, a person’s status as a part of a political
majority or minority is strongly associated with his or her satisfaction with
democracy. The gap in people’s view of how a given political system works is
significantly large between people whose party is in government (winners) and
those in opposition (losers). Winners are more likely to show higher levels
satisfaction with democracy than losers.

The multilevel model, moreover, shows that party systems mediate the
relationship between the partisan status (winners or losers) on one hand and
popular satisfaction with democracy on the other hand. Greater ENPP is likely to

mitigate the negative view of losers about the way democracy works and, in turn,

decrease the gaps in popular satisfaction with democracy between winners and
losers in emerging democracies.

Finally, using Aftobarometer data collected before and after the 2002
electoral reform in Lesotho, this study explores how electoral system changes from
the first-past-the-post (FPTP) to the mixed-member-proportional (MMP) system
influence citizens’ attitudes toward political systems. I find both direct and indirect
effects. In the aggregate, Lesotho’s electoral reform from a majoritarian to a mixed
electoral system is directly associated with increased levels of citizen support for
the country’s state and regime. Importantly, however, formal institutions have only
indirect effects at the individual level. Increasing the proportionality in translating
votes into seat share narrows the difference in popular satisfaction with democracy
between winners and losers. Because of the reform, opposition parties were able to
win 40 seats in the National Assembly in 2002 elections and, in turn, their
supporters’ levels of satisfaction with democracy are increased.

This study sheds light on the general issue of whether institutions matter in
emerging democracies. Constitutional debates - whether about electoral systems,
executive-legislative relations, or unitarism versus federalism — have risen on the
policy agenda to become very important issues in emerging democracies. If
institutional arrangements influence political support, as this research finds, this
may have important consequences for constitutional debates. The challenge for
further research is to consider these issues in terms of the democratization process.
If institutional designs can strengthen popular satisfaction with democracy this

may provide significant lessons for the process of democratization.

Copyright by
WONBIN CHO
2005

To All my family who kept faith and hope in me.

ACKNOWLEDGEMENTS

A dissertation is a collective project. I owe thanks to many people who
contributed to the completion of this dissertation. First, and the most important
person is my dear wife, Hee-Young Kim. Without her love, trust, patience, and
invaluable sacrifice, this dissertation would not have been done. My two daughters,
Yoon and Jin, have always encouraged my work with their smiles. Both my family
and Hee-Young’s family have also played a major role cheering me up until I
arrived at this point. I especially want to acknowledge my parents, Eek-Jee Cho
and Mae-J a Moon, who raised me to foster scholarship and pursue this doctoral
degree.

I would also like to acknowledge and thank my committee members
(Michael Bratton, Mark Jones, Eric Chang, and Kimberly Maier) who gave me
valuable advice and guidance during all of the stages of this process. Without their
help along the way, this dissertation would not have been accomplished. In
particular, I would like to acknowledge the help and support of my advisor,
Michael Bratton, who first introduced me to the African Politics by offering me an
opportunity to work with him for the Aftobarometer. He gave me not only advice
on dissertation substance but also on scholastic attitude. Now he is my role model
in academia and will remain a mentor-colleague to me. I sincerely hope that I will
be given opportunity to replicate my experience with younger graduate students in
the future. Mark Jones also encouraged my professional development and sparked
my interest in political institutions. Eric Chang and Kimberly Maier provided

important assistance and helped me hone my methodological skills. Other scholars

vi

have been instrumental in building my idea and developing my approach to the
dissertation, including Burt Monroe, Brian Silver, and Carolyn Logan.

I would like to thank Doh C. Shin from the New Korea Barometer, Marta
Lagos from the Latinobarometer, and numerous individuals from the Taiwan
Election and Democratization Study ('I’EDS) for allowing me to use their survey
data sets for the dissertation. Those data sets are obviously necessary components
of this dissertation. I must extend special thanks to Paloma Bauer. With her
suggestions and skilled editing, this dissertation has been nicely completed.

Finally, this dissertation would have been obviously more difficult, if not
impossible, without the valuable assistance of numerous individuals in Lesotho. I
would like to acknowledge the help of numerous political and administrative
leaders in Lesotho. All of them were extremely generous with their time and
willingness to answer so many questions. I would also like to acknowledge the
assistance of individuals at the Sechaba Consultants who helped me organize my
research in Lesotho. Most importantly, I would like to thank my Basotho friend,

Tlotliso Leboela.

vii

TABLE OF CONTENTS

LIST OF TABLES ..................................................................... x
LIST OF FIGURES ................................................................... xi
LIST OF ABBREVIATIONS ........................................................ xii
CHAPTER 1: Introduction ............................................................ 1
CHAPTER 2: Theory ................................................................... 7
2.1 Introduction ......................................................................... 7
2.2 Cultural Explanation ............................................................... 7
2.3 Government Performance ......................................................... 9
2.4 Institutional Explanations ........................................................ 11
2.4.1 Party Systems ............................................................... 14
2.4.2 Electoral Systems ........................................................... 18
2.4.3 Presidentialism vs. Parliamentarism ..................................... 21
2.4.4 Unitarism vs. Federalism .................................................. 25
2.5 Partisan Status and Political Institutions ....................................... 28
2.5.1 Partisan Status: Winner, Non-partisan, and Loser ..................... 28
2.5.2 Partisan Status in Different Political Institutions ....................... 32
CHAPTER 3: A Model of Political Support ....................................... 39
3.1 General Research Strategy ....................................................... 39
3.2 Multilevel Analysis ............................................................... 40
3.3 A Model of Political Support .................................................... 42
3.3.1 Dependent Variable ......................................................... 42
3.3.2 Independent Variables ...................................................... 47
3.3.3 Control Variables ............................................................ 53
3.3.4 Demographic Variables .................................................... 55
CAHPTER 4: Individual-Level Analyses .......................................... 62
4.1 The Effects of Partisan Status on Satisfaction with Democracy
at the Aggregate Level ........................................................... 62
4.2 Multivariate Single-Country Models ........................................... 66
CAHPTER 5: Multilevel Analyses .................................................. 86
5.1 Political Institutions and Satisfaction with Democracy ...................... 86
5.2 Analysis of Variance .............................................................. 87
5.3 Multilevel Model of Satisfaction with Democracy ........................... 89

CHAPTER 6: Electoral Systems, Partisan Status, and Citizens’

Political Support in Lesotho ......................................... 108
6.1 Introduction ....................................................................... 108
6.2 Background ....................................................................... l 11
6.3 Electoral Systems and Citizen Attitudes ...................................... 1 15

viii

6.4 Partisan Status: Winners, Losers, and Non-Partisans ....................... 117

6.5 Partisan Status in Different Electoral Systems ............................... 120

6.6 A Model of Political Support ................................................. 122

6.6.1 Dependent Variables ..................................................... 122

6.6.2 Independent Variables .................................................... 124

6.6.3 Control Variables .......................................................... 126

6.6.4 Demographic Variables .................................................. 127

6.7 Results ............................................................................. 127
6.7.1 The Effects of Partisan status on Satisfaction with Democracy

at the Aggregate Level .................................................... 127

6.7.2 Multivariate Models ....................................................... 129

6.7.3 Effects of Electoral Systems on Political Support .................... 132

6.8 Conclusion ........................................................................ 139

CHAPTER 7: Conclusion ........................................................... 152

APPENDIX A ......................................................................... 158

APPENDIX B ......................................................................... 159

REFERENCES ........................................................................ 160

ix

LIST OF TABLES

Table 3.1 Political and Economic Indicators in Countries under Study ......... 59

Table 4.1 Effects of Individual-Level Variables on Satisfaction with

Democracy .................................................................. 76
Table 4.2 Probability of Satisfaction with Democracy by Country ............. 84
Table 5.1 One-way ANOVA Model ............................................... 102
Table 5.2 Individual and National-Level Predictors of Satisfaction with

Democracy ................................................................ 103
Table 5.3 Model with Cross-Level Interaction ................................... 105
Table 6.1 Vote and Seat Shares in 1998 and 2002 Elections ................... 147

Table 6.2 Effects of Individual Characteristics on Satisfaction with
Democracy and Trust in Political Institutions in 2000 and 2003. . . 148

Table 6.3 Effects of Electoral Systems on Satisfaction with Democracy:
Pooled Model Estimates .................................................. 150

Table 6.4 Effects of Electoral Systems on Trust in Political Institutions:
Pooled Model Estimates .................................................. 151

LIST OF FIGURES

Figure 2.1 Hypothesized Satisfaction with Democracy in Different

Party Systems .............................................................. 37
Figure 2.2 Hypothesized Satisfaction with Democracy in Different

Electoral Systems ......................................................... 38
Figure 3.1 Satisfaction with Democracy, by Country .............................. 56
Figure 3.2 Satisfaction with Democracy, by Continent ............................ 58
Figure 4.1 Satisfaction with Democracy and Partisan Status ..................... 73
Figure 5.1 Predicted Probability of Popular Satisfaction with

Democracy by Partisan Status and ENPP ............................. 107
Figure 6.1 Hypothesized Satisfaction with Democracy in Different

Electoral Systems ......................................................... 144
Figure 6.2 Satisfaction with Democracy among Winners,

Non-partisans, and Losers .............................................. 145

.146

Figure 6.3 Predicted Values of Popular Satisfaction with Democracy. . . . . . ..

xi

AD
BAC
BN P
BCP
CCM
CN
COPEI
CSES
ENPP

IPA
LCD
LLA
LPC
LWP
MAS
MCP
MDC
MFP
MMM
MMP
MPD
MVR
NIP
N PP
NRM
PAICV
PAN
PFD
PR
PRD
PRI

PVEM
SMD

SN TV
SSD
TEDS
UDF
ZANU-PF

LIST OF ABBREVIATIONS

Democratic Action
Barutoland African Congress
Basotho National Party
Basutoland Congress Party
Chama Cha Mapinduzi
Convergencia Nacional
Organized Political Committee for Independent Election
Comparative Study of Electoral Systems
Effective Number of Parliamentary Paries
First-Past-the-Post
Index of Disproportionality
Interim Political Authority
Lesotho Congress Democrats
Lesotho Liberation Army
Lesotho People’s Congress
Lesotho Workers Party
Movimiento al Socialismo
Malawi Congress Party
Movement for Democratic Change
Marematlou Freedom Party
Mixed-Member—Majoritarian
Mixed-Member-Proportional
Movement for Democracy
Movimiento V Republica
National Independent Party
National Progressive Party
National Resistance Movement
African Party for the Independence of Cape Verde
Partido Accion Nacional
Popular Front for Democracy
Proportional Representation
Partido de la Revolucion Democratica
Partido Revolucionario Institucional
Partido del Trabajo
Partido Verde Ecologista de Mexico
Single Member District
Single Non-Transferable Vote
Single-Seat District
Taiwan’s Election and Democratization Study
United Democratic Front
Zimbabwe African National Union-Patriotic Front

xii

Chapter 1: Introduction

Do political institutions affect popular satisfaction with democracy? This
research takes political institutions into account to explain citizens’ political support in
emerging democracies. By combining cross-national survey data collected from 35
developing countries in sub-Saharan Africa, Latin America, and Asia, with data on
political institutions in which people live, this study finds both constitutional
arrangements and party systems have a significant impact on popular evaluations of
political systems performance. In addition, a person’s status as part of a political
majority or minority is strongly associated with his or her satisfaction with democracy.
Moreover, party systems mediate the relationship between a person’s partisan status and
his or her satisfaction with the way democracy works.

Democracy thrives on popular support and withers in its absence (Easton 1965).
Since the mid of 19705, we have witnessed an unprecedented and unanticipated wave of
democratization around the world (Huntington 1991). With the collapse of communism,
moreover, democracy has reached every region of the world for the first time in history.
Most democratic theorists would argue that the success or failure of democratization is
in large part a function of the support built among citizens. Citizens’ support for the
regime has proven to be a necessary condition for the regime’s legitimacy. It is further
asserted that legitimacy is strongly influenced by the performance of the regime (Lipset
1959). Political systems that are ineffective in meeting public expectations over long

periods of time can lose their legitimacy. Stable or increasing levels of popular support

facilitate the consolidation of democracy, whereas declining levels of support
undermine democracy and can even lead to its collapse.

Political support is a multi-dimensional concept. Easton (1965) distinguished
between support for the community, the regime, and the authorities, while other
scholars expand the classification into a fivefold framework: support for the political
community; support for regime principles or democracy as an ideal form of
government; evaluations of the regime’s performance; support for regime institutions;
and support for political actors (Norris 1999). This study focuses on evaluations of
regime performance, measuring citizens’ approval of how democratic political systems
function in practice.

The underlying dynamics of a regime’s performance approval in a fledgling
democracy are potentially very different from those of an established democracy. In
established democracies in which the structure and character of political institutions are
constant within countries over extended period time, scholars emphasize the importance
of policy performance, especially economic outcomes (Przeworski, Alvarez, Cheibub,
and Limongi 2000). In addition, the nature of the regime does not change when an
election results in defeat for the incumbent government and new party takes office.
However, in an emerging democracy there can be competition between regimes (Rose,
Mishler, and Haerpfer 1998). Free elections give undemocratic as well as democratic
parties the opportunity to compete for office. Some citizens may still demand the return
of the old or the introduction of a new undemocratic regime. A new democracy has yet
to institutionalize or even introduce all the features of an established democracy. In a

new democracy, newly adopted political institutions can matter as much as policy

outputs, and political outputs can matter as much as economic performance. Emerging
democracies, especially those in Sub-Saharan Africa, Latin America and Central and
Eastern Europe, are suffering from limited public support for regime performance
because they have to cope with simultaneous political and economic transitions (Bratton
and Mattes 2001; Rose, Mishler and Maerpfer 1998; Lagos 2001 ). In these cases,
popular satisfaction with democracy depends not only on political performance but also
on economic performance. This research attempts to understand how citizens in
emerging democracies evaluate their political systems’ performance.

There are substantial cross-national variations in citizens’ evaluations of how
democracy works. For instance, Western European countries present relatively high and
stable levels of popular satisfaction with democracy, while Latin American countries
express lower levels of public evaluations of regime performance. Within the sub-
Saharan African countries, Cape Verde and Nigeria show the lowest levels of the
satisfaction with democracy. Why are there such major differences between countries?
What accounts for the differences in political support across countries?

Three different theories attempt to explain significant cross-national variations
in popular approvals of regime performance by focusing on cultural values, government
performance, and political institutions. Democratic political institutions have relatively
stable features that serve both to shape citizens’ political experiences and to develop
attitudes about the functions of the political system (Powell 1982, 1989). How
democratic institutions treat citizens is very important for their development. This study

focuses mainly on variance of constitutional arrangements, for example presidentialism

and parliamentarism, unitarism and federalism, a wide variety of party systems, and
electoral systems.

Since the early 19905, constitutional reforms have been the subject of debates in
many democracies. Such debates have already led to a number of important reforms.
Among established democracies, New Zealand reformed its electoral system from a
First-Past-the-Post (FPT P) to a Mixed-Member-Proportional (MMP) system. Japan also
changed it Single Non-Transferable Vote (SNTV) to a Mixed-Member-Majoritarian
(MMM) system. In Latin America, Bolivia, Ecuador, Mexico, and Venezuela undertook
large-scale electoral reform in the 19905, as did Lesotho and South Korea elsewhere in
the world.

These debates about constitutional reforms concern the alleged effects of a given
reform on the stability and endurance of democracy. Scholars assert that political
institutions should be key in the efforts to expand democracy and reduce violent
conflicts in divided societies where we observe sharp ethnic, religious, national, and/or
linguistic cleavages (Reynolds 2002). The debates have focused only on macro levels,
for example national or district levels. However, political institutions by themselves do
not directly define the life of a regime. A regime’s survival is also affected by citizen’s
perceptions toward the performance of the constitutional arrangement they live in. The
same constitutional arrangement does not necessarily produce the same popular
approval levels of regime performance in a different country. Citizens and groups in a
particular country have conflicting preferences over political, economic, and social
issues. Political institutions aggregate these preferences into specific political outcomes,

and these in turn induce public policy decisions in different issues. Surprisingly, few

political scientists have addressed the question of constitutional effects on citizens’
evaluations of a regime’s performance (Anderson and Guillory 1997; Norris 1999;
Farrell and McAllister 2003; Rohrschneider 2002), even though analyzing the effects of
alternative constitutions has long been a main research topic in political science. The
goal of this study is to understand the effects of constitutional arrangements on popular
satisfaction with political systems in fledgling democracies.

The literature on political support discusses at least two different levels of
analysis. First, a large number of studies concentrate on the individual level; these
studies assess the nature and determinants of inter-individual differences in attitudes
toward political systems (e.g., Anderson and Guillory 1997). Second, a sizable literature
documents the nature of cross-national differences in political support (e.g., Wei] 1989).
Considering these two levels of analysis together, what emerges is a multilevel data
structure: individuals are nested within countries. This research includes both individual
level survey data and national characteristic data from 35 emerging democracies. 1
model this data structure through a two-level hierarchical analysis.

This dissertation is organized as follows. Chapter Two reviews existing
literature related to political support and provides the theoretical framework for the
study. In Chapter Three, I detail the methods used to carry out this research. This
project is divided into two separate inquiries. I created a multi-level data set to allow me
to explore my research questions across 35 emerging democracies. Secondly, I conduct
a cross-time analysis of Lesotho to explore my research question in much greater detail.
Chapter Three includes a thorough discussion of the operationalization of key variables

in the study.

Using Ordered Probit Analysis, I test the various explanations for satisfaction
with democracy across 35 countries in Chapter Four. The results presented in Chapter
Four support my hypotheses concerning the expected impact of partisan status (majority
or minority) on popular satisfaction with democracy in emerging democracies. Equally
important, the results show the importance of citizens’ evaluations of economic
performance to explain popular approvals of regime performance.

Using Multilevel Analysis, I demonstrate how political institutions affect
popular satisfaction with democracy and how they mediate the relationship between
partisan status (minority in politics) and satisfaction with democracy in Chapter Five. I
find that political institutions have significant effects on citizens’ evaluations of regime
performance and party systems in particular mediate the relationship between partisan
status and popular satisfaction with democracy. A Greater Effective Number of
Parliamentary Parties is likely to reduce the gap in the satisfaction between political
majority and minority in fledgling democracies.

In Chapter Six, using an analysis over time in Lesotho, I show how electoral
system changes from FPTP to MMP systems influence popular satisfaction with
democracy. Increasing the proportionality in translating votes to seat share narrows the
difference in popular satisfaction with democracy between citizens whose party is in
government and those whose party is in opposition. Finally, Chapter Seven summarizes

the main findings of the study and suggests some implications.

Chapter 2. Theory

2.1 Introduction

Various theories have tried to explain significant cross-national variations in
political support, incorporating social, cultural and economic variables. Political support
is a complicated research question, since it is determined by a variety of different
factors, including political, social, economic, institutional and historical factors. Yet,
many scholars have focused primarily on either cultural characteristics or government
performance (e. g. Putnam 1993; Inglehart 1997; Clarke, Dutt, and Kornberg 1993).
However, the explanatory power of those variables remains uncertain particularly in
emerging democracies.

This chapter will review the relevant literature and assess the sufficiency of
existing theories in explaining political support in developing countries and propose

several alternative explanations including political institutions.

2.2 Cultural Explanation

One can identify three separate schools of thought that seek to explain cross-
national variation in political support for a regime. First, theories of cultural values
emphasize deep-rooted social and political values in each country. A significant number
of studies argue that social capital (or interpersonal trust) plays a crucial role in
democracy (Putnam 1993; Norris 1999, Warren 1999). The theory of social capital
emphasizes the importance of social trust for civic engagement. Social capital is
regarded as a strong determinant of, or influence upon, political support of various

kinds, including support for political community, confidence in institutions, and trust in

political leaders. With survey evidence of Europe, however, Newton (1999: 185) shows
that “there is not a close or consistent association between social and political trust,
between social trust and political behaviour, or between activity in voluntary
associations and political attitudes of trust and confidence.” In other words, social
capital is not necessarily translated into political trust. He argues that political trust
should be more a product of political rather than social factors.

Using three decades of time—series data, Inglehart (1977, 1990, and 1997) shows
an intergenerational shift toward Post-material values. Since members of the Post-
material generation give high priority to protecting freedom of speech and to
participation in making important government decisions, this trend should bring
growing mass demands for democratization. Post-materialist theory emphasizes that the
modernization process has undermined support for traditional, hierarchical institutions
and authoritarian values, producing a crisis of confidence in government, but this should
not be understood as a crisis of confidence in democracy per se.

Gibson (1998) has argued that tolerance of out-groups is essential to democracy:
civil liberties and legitimate opposition require tolerance and forbearance toward groups
with whom one disagrees and dislikes. Inglehart (2003) finds that the “Survival vs. Self-
expressing values” syndrome including tolerance of out-groups, Postmaterialist values,
political activism, interpersonal trust, and subjective well-being has a very strong
association with democratization. However, it is not clear whether democratic
institutions give rise to “self-expression values” or whether higher levels of “self-

expression values” increase a society’s prospects for democracy. Muller and Seligson

(1994) argue that interpersonal trust appears to be a product of democracy rather than a
cause of it.

These theories are most powerful when explaining persistent and long-term
differences between nations and diffuse support, such as feelings of national identity.
However, this research accounts for cross-national differences on the attitudes towards

political systems.

2.3 Government Performance

Second, theories of government performance commonly focus on how regime
support relates to government performance, particularly concerning the economy (W eil
1989; Weatherford 1984, 1987; Clarke, Dutt, and Kornberg 1993; Lockerbie 1993;
Anderson 1995; Weisberg 1996; McAllister 1999; Miller and Listhaug 1999). Studies
that examine the connection of system outputs with political support typically accept
that the two are related because “the government is assumed to possess the tools and
abilities to solve social problems” (W eatherford 1984: 189). Government takes a credit
when performance is strong, and the blame when it is weak. Most of the studies have
focused on the relationship between aggregate levels of economic growth,
unemployment, or inflation, as well as individual- level retrospective and prospective
evaluations of the economy, and satisfaction with governance in North America and
Western Europe. A significant number of studies empirically demonstrate that economic
performance is related to system support (Weatherford 1984, 1987; Clarke, Dutt, and

Kornberg 1993). Related evidence suggests that public perceptions of both personal and

national economic conditions are related to system support (Anderson and Guillory
1997; Citrin and Green 1986; Lockerbie 1993; McAllister 1999).

Nevertheless, the links between economic performance and satisfaction with
government are not clear (Lawrence 1997). First, there is no consensus how to measure
economic performance. For example, do measures of production or productivity
accurately capture improvement in citizen’s living standards? Second, there is
uncertainty about how performance is perceived and appraised. Do people care about
their absolute living standards, their relative living standards, or the rate at which these
improve? Finally, people could well respond to poor economic performance by
dissatisfaction with their current government, but not with political system per se. This
distinction is more likely to occur in the established democracies where citizens draw a
clear distingction between the operation of the system as a whole, and the activities of a
particular party’s government. In emerging democracies, which often do not enjoy
widespread popular support, or where support is unevenly distributed across social
groups, such distinctions are more likely to be blurred.

The empirical analyses do not show any strong or consistent evidence.
McAllister (1999) finds and concludes that there is a modest yet consistent relationship
between support for political institutions and subjective economic evaluations, but
institutional support seems unaffected by objective indicators of economic performance
such as GDP and unemployment. Neither Steenberg and Jones (2002) nor
Rohrschneider (2002) find any significant relationship between an objective economic
performance indicator at the national-level and “mass support for an EU—wide

government.”

10

If government performance is defined in purely economic terms, then this
perspective seems to provide a poor fit for cross-country analysis. This theory becomes
more convincing if non-economic aspects of government performance are included in
the model. Miller and Listhaug (1999) argue that we need to broaden our evidence to
include citizens’ perceptions of the fairness of policy outcomes, as well as their
expectations of government.

The empirical studies on public opinion in emerging democracies show that
politics have more influence on citizens’ attitudes towards democratic systems than
economics (Evans and Whitefield I995; Linz and Stepan 1996; Rose, Mishler and
Haerpfer 1998; Shin 1999; Bratton, Mattes and Boadi 2005). After democratic
transitions in new democracies, popular support for the new political regime would be
maintained by providing political goods such as political order, civil rights, and good
governance, in spite of governments’ poor economic performance. Using three rounds
of Afrobarometer data set collected in Nigeria from 2000 to 2003, Bratton and Lewis
(2005: 34) find that “the delivery of political goods is not only central to the formation
of mass opinion about democracy but also durable over time, at least in the medium

term following regime transitions.”

2.4 Institutional Explanations

Finally, institutional theories seek to understand the attitudes of citizens towards
the political system within their broader constitutional context (Anderson and Guillory

1997; Norris 1999; Farrell and McAllister 2003; Rohrschneider 2002). These studies

11

focus on the linkages between representative elites and public opinion provided by the
intermediary institutions of political parties, interest groups and electoral systems.

People form attitudes about politics in systemic structures whose institutional
contexts mediate preference and define the choices available (Powell 1982, 1989).
There are many ways in which a democracy can be organized and run. Current
democracies show a variety of political institutions, for example, presidentialism and
parliamentarism, two-party and multiparty systems, majoritarian and proportional
representation electoral systems, and federal and unitary systems. Among these political
institutional arrangements, two patterns can be seen: majoritarianism and
proportionalism (Lijphart 1999; Powell 2000). The essence of the majoritarian model of
democracy is government by and for the majority of the people. The majoritarian model
concentrates political power in the hands of a majority — often even merely a plurality
instead of majority. The key of the proportional model, on the contrary, is government
by and for as many people as possible. The proportional model is characterized by its
inclusiveness, bargaining process, and compromise, whereas the majoritarian model is
exclusive, competitive, and adversarial.

These two quite different visions of democracy lead to distinct kinds of
relationships between citizens and government officials (Powell 2000). In the
majoritarian model, citizens are more suspicious of the autonomy of elected
representatives, less concerned about minorities, more desirous of seeing that
government officials are clearly accountable to voters. However, in the proportional
model, citizens are more suspicious of majorities (especially those created by elections),

less worried about the autonomy of policymakers as long as citizens have had a role in

12

selecting them, and less worried about negotiated inaction. Thus, the different political
institutional arrangements in which citizens live result in different attitudes towards the
government.

An extensive literature has explored the consequences of alternative institutional
designs, for example, the influence of presidentialism or parliamentarism on the
stability of democracy (Powell 1982; Lijphart 1992; Linz and Valenzuela 1994). The
definitional distinction between presidential and parliamentary regimes is the political
independence or interdependence of the legislative and the executive branches (Linz
1994; Shugart and Carey 1992). Linz (1994) criticizes the stability of presidentialism to
sustain democratic regime, arguing that presidential systems have no mechanism for the
resolution of conflicts between the executive and the legislative, while parliamentary
systems have such a mechanism.

Stepan and Skach (1993) focus on how key differences between pure
presidentialism and pure parliamentarism affect democratic performance, which is
operationalized using a variety of measures. In addition to democratic survival, the
authors compare the relationship between institutional design and the following
variables: the number of effective parties; Vanhanen’s democracy prediction residuals;
a scale of political rights; the susceptibility to military coups; the likelihood of
legislative majorities; and cabinet stability. Across all these measures, pure
parliamentary regimes perform better than pure presidential regimes.

Shugart and Carey (1992) compare up to forty-six different countries at different
time periods in order to examine a range of important questions about presidential

systems and democratic performance. They are interested in key aspects of presidential

l3

democracies, including the elections of the president and assemblies, the formation of
cabinets, the legislative power of the president, and the ways in which electoral systems
produce different party systems. In addition, focusing on sixteen Latin American
countries, Jones (1995) demonstrates that successful performance among presidential
democracies relies on the degree to which the electoral system produces legislative
support for the president.

However, few studies have looked systematically at the relationship between
institutional arrangements and public support for the political system (Anderson and
Guillory 1997; Norris 1999; Farrell and McAllister 2003). Democratic political
institutions have relatively stable features that serve both to shape citizens’ political
experiences and to develop attitudes about the functions of the political system. How
democratic institutions treat citizens is very important for their development. In
following sections, I define and describe the various categories of key political
institutions, including party systems, electoral systems, presidentialism and

parliamentarism, and unitarism and federalism.

2.4.1 Party Systems

Political parties and the party systems they form constitute the major channels of
interest aggregation and citizen input in both established and emerging democracies.
They are the vehicles through which political elites supply policy alternatives, and they
constitute the major route for citizens to organize the demand for such alternatives
(Klingemann, Hofferbert, and Budge 1994). Parties also play a crucial role in the nature

of democratic governance because they help legitimize the state by mediating between

14

citizens and the state. Parties have taken on the roles as simultaneous agents both of the
state and its citizens (Mair 1997). In new democracies, especially those emerging from
single party regimes, parties may also continue to be top-down agents of political
mobilization (Bratton, Mattes, and Gyimah-Boadi 2005).

While virtually all democracies have political parties that compete for office,
political systems differ in a number of important ways with regard to how they go about
channeling inputs or providing policy alternatives. Moreover, the way in which political
institutions condition the formation, functioning, and development of political parties
and party systems varies as well. While there is an extensive literature linking electoral
systems and the party systems, few researchers have investigated their link with how
citizens feel about the way the political system works.

Party systems grant citizens greater or fewer choices. My assumption in this
study is that citizens’ satisfaction with the way democracy works should be higher in
the political systems where they feel that their preferences are represented in the
political arena by a party which reflects their interests. Multi-party systems may
alleviate dissatisfaction with political system because these systems provide more
options for the citizens. In theory, the issue differences and ideological distances
between parties should exhibit greater variance and be clearer in a multi-party than a
two-party system. In the latter, each party attempts to resemble its opponent as closely
as possible (Downs 1957). It may be a critical issue for citizens in a two-party system
whether any meaningful differences exist between the two parties. If citizen discontent

with the inability of the parties to provide policies for social problems persists over an

15

extended period of time, this dissatisfaction may become generalized to other political
institutions and the political system as a whole.

A multi-party system offers more possibilities that at least some party will
represent one’s interests, thereby moderating the overall growth of dissatisfaction with
the political system. In a multi-party system the individual merely has more options
when selecting a preferred party. In a multi—party system, especially one with a coalition
government, citizens may consider the extent to which they are represented by coalition
blocs rather than their own party or any particular party.

Miller and Listhaug (1990) argue that a smaller number of parties in a system is
correlated with popular dissatisfaction with democratic governance. Based on
comparable survey data from Norway, Sweden, and the United States, they find that
structural differences in the party systems of the three countries are critical in
determining the levels of popular satisfaction with democracy. For example, the authors
show that “political discontent in Norway was reduced because new parties provided
the disaffected with a means of representation, thus channeling dissatisfaction back into
electoral arena” (Miller and Listhaug 1990: 357). In the inflexible party systems of the
United States and Sweden, on the contrary, the results demonstrate that popular support
for the regime was declining “because many people failed to see any of the parties as a
viable alternative” (Miller and Listhaug 1990: 357).

In contrast to the research by Miller and Listhaug (1990), however, Weil’s
(1989) cross-national study of system support in western democracies finds that party
system fractionalization - that is, an extremely large number of parties — is

systematically associated with lower levels of regime support. Citizens’ dissatisfaction

l6

with political institutions may result from “an unresponsive structure of opposition, for
example, a party system that generates extremism, or a governing coalition that does not
reflect voters’ preferences” (W eil 1989: 683). In extremely polarized or fractionalized
party systems, at least one extremist or antisystem party is in quasi-permanent
opposition. Those extremist parties are sufficiently unacceptable to others that they
cannot form alternative coalitions, but they are strong enough to block alternative
coalitions that do not include themselves. Weimar Germany and Fourth-Republic
France may be examples of extremely fractionalized party systems. Weil empirically
demonstrates that fractionalized party systems lead to a decline of popular confidence in
political institutions with survey data of six Western countries.

J ackman ( 1987) argue that because citizens’ vote is not directly connected with
government formations in the multiparty system, people are less likely to turn out to
vote in the multiparty system. With the data from 19 industrial countries, he empirically
shows that multiparty system lowers voter turnout. Norris (1999) also finds that citizens
in two-party or moderate multi-party systems exhibit higher levels of confidence in
political institutions than those with both fragmented party systems and predominant
one-party systems.

Using directly comparable survey evidence from about 20 democracies collected
between 1993 and 1995, Anderson (1998) examines whether and to what extent party
system performance explains differences in levels of satisfaction with the political
system. The results demonstrate that there is no significant correlation between party
systems and popular satisfaction with democracy at the aggregate level. His study does

not include any individual level analysis. Anderson reports the results of correlation

l7

analysis with the aggregate data set, while the data set under his study have two
different levels, individual and national levels.

While those works present important hypotheses about the relationship between
party systems and popular satisfaction with democracy, they do not show consistent
empirical results. It may be not only because they are using different measurements of
party systems but also because their methods to explore the impacts of different party
systems on popular satisfaction with democracy do not represent the data structure in
their study. Most of the works demonstrate the correlation analyses between the two
variables with aggregate data, and some of them ignore the characteristics of the two-
level data set by treating the national-level data as an individual-level data. This study
examines whether, and how, cross-national differences in party system affect citizen’s
evaluations of their country’s political system performance with a multilevel analysis. I
argue that the greater effective number of parliamentary party (ENPP) should induce

higher levels of popular satisfaction with the way democracy works (HI)

2.4.2 Electoral Systems

In addition to political parties, the electoral system is another important
mechanism connecting citizens and the state. A large number of studies have classified
the types of electoral systems and explored the relationship between electoral system
and party system (Duverger, 1954; Rae 1967; Katz 1980 Taagepera and Shugart 1989;
Lijphart 1994; Cox 1997; Reynolds and Reilly 1997). Electoral systems based on
sin gle-member districts tend to produce two parties while systems with proportional

representation (PR) tend to produce multiple parties (Duverger 1954). Later scholars,

18

however, have considered the possibility that social cleavage and electoral structures
may interact (Sartori 1968; Rae 1971; Riker 1982; Cox 1997). Duverger’s institutional
claims are conditioned by the nature of social cleavages. Ordeshook and Shvetsova
(1994) find that increasing the proportionality of an electoral system in a homogeneous
society does not proliferate the number of parties, whereas it does in a heterogeneous
society.

But, few studies look at the impacts of different electoral systems on popular
satisfaction with the political system (Anderson and Guillory 1997; Farrell and
McAllister 2003; Norris 1999). Literature on electoral systems focuses on the degree of
proportionality or disproportionality in the translation of votes into seats and on their
effects on the number of parties. Lijphart (1994; 1999) introduces three main types of
electoral formulas and a large number of subtypes within each of these: majoritarian
formulas (including plurality, two-ballot systems, and alternative vote as the main
subtypes), proportional system (classified further into largest remainders, highest
averages, mixed member proportional formulas, and single transferable vote formulas),
and semi-proportional systems (such as the limited vote, the cumulative vote, and
parallel plurality-PR).

Many scholars who study political institutions would agree that the electoral
system is one of the primary factors that constrains the development and nature of party
system in a country. Some of the scholars, however, argue that social cleavages also
condition the party system (Lipset and Rokkan 1967). In addition to those macro-
factors, different electoral systems induce different electoral strategies from both voters

and political elites (Cox 1997). In his study of strategic voting, acting to reduce the

19

number of competitors, Cox (1997) finds that strategic voting fades out in multimember
districts when the district magnitude is greater than five, as it gets harder for both voters
and political elites to get information from each other. Thus, electoral systems may
constrain both representativeness and policy making processes in a country and affect
how citizens evaluate the way democracy works.

Proportional systems are more representative than majoritarian and semi-
proportional systems because they facilitate the representation of all relevant societal
and ethnic groups. Lijphart (1999) shows that, although no PR system is perfectly
proportional and although there is a great deal of variation within the PR family, PR
systems tend to be considerably less disproportional than plurality and majority
systems. More proportional electoral system may make stronger linkages between
political parties and voters and lead to citizens’ higher levels of satisfaction with
democracy.

Using the Comparative Study of Electoral Systems (CSES) data set of thirty
countries, Farrell and McAllister (2003) examine the direct impact of the electoral
system, specifically ballot structure, on levels of voter satisfaction with democracy.
Because preferential electoral systems give voters more choice in the electoral act, it is
argued that such electoral systems produce closer links between voters and politicians,
which necessarily encourage a politics of accommodation between politicians. The
authors conclude that these electoral systems generate higher levels of voter satisfaction
with democracy. Their study empirically finds that electoral systems directly affect
levels of voter satisfaction with democracy. Particularly, the levels of disproportionality

and assembly size appear to be strong predictors among the national level measures of

20

electoral systems. Farrell and McAllister also show that political attitudes mediate the
relationship between electoral system and voter satisfaction with democracy. For
example, the disproportionality of the electoral system has an impact on popular
perception of electoral fairness and via that to satisfaction with democracy.

Based on aggregate data collected in 19 European democracies, Anderson
(1998) finds that there is a strong correlation between electoral systems and popular
satisfaction with democracy, indicating that the more proportional the electoral system,
the higher the level of satisfaction. However, Norris (1999), in a study including both
individual-level survey and national-level data collected in 25 democracies, finds that
majoritarian electoral systems produce higher levels of confidence in political
institutions than proportional electoral systems. Whether and how electoral systems can
affect levels of popular satisfaction with democracy is still unresolved. This study
examines whether more proportional electoral systems could increase levels of popular

satisfaction with democracy across emerging democracies (H2).

2.4.3 Presidentialism vs. Parliamentarism

Does the institutional locus of power within the regime affect citizens’
satisfaction with democracy? Within political systems there is a major distinction
between parliamentary and presidential systems. The merits of these systems have been
widely debated (Lijphart 1992; Linz and Valenzuela 1994; Shugart and Carey 1992).
Linz (1994) claims that presidentialism with a weak legislature has the advantages of
executive stability, greater popular control, and more limited government, but the

disadvantages of executive-legislative deadlock, temporal rigidity and winner-take-all

21

government. Parliamentarism, on the contrary, is claimed to reverse the pros and cons.
For example, a parliamentary election might produce an absolute majority for a
particular party, but more normally it gives representation to a number of parties. This
means that the prime minister will be much more aware of the demands of different
groups and much more concerned about retaining their support. Linz (1994) thus argues
that presidential systems inevitably produce divided governments, deadlocks,
institutional paralysis, and, ultimately, the breakdown of democratic regimes.

Lijphart (1999: 117-124) identifies three crucial differences between
parliamentary and presidential systems of government. First, parliamentary systems
have “collective or collegial” executives, whereas presidential systems have one-person,
“non-collegial” executives. While the prime minister’s position in the cabinet can vary
from preeminence to virtual equality with the other ministers, there is always a
relatively high degree of collegiality in decision-making. On the contrary, the members
of presidential cabinets are mere advisers and subordinates to the president. That is, the
most important decisions in presidential systems can be made by the president with or
without, or even against, the advice of the cabinet.

Shugart and Carey (1992), on the other hand, argue that all presidential regimes
are not alike and that certain types of electoral and party systems may exacerbate the
problems of presidentialism. They suggest that “the basic principles of presidential
government may be worth preserving” (Shugart and Carey 1992: 54). Presidentialism
offers the advantage of direct accountability between voters and an executive, while
parliamentarism allows the assembly to remove the executive by shifting coalitions. The

checks and balances inherent in presidentialism are also conductive to compromising

22

when majoritarian rule does not work. While the independence of the assembly
generates the problem of immobilism, it prevents one party from ruling alone.

Empirical results on the difference between presidential and parliamentary
regimes have not been consistent yet. Przeworski, Alvarez, Cheibub, and Limongi
(2000) provide evidence in favour of a parliamentary system. Using data from both
OECD and non-OECD countries during 1950-90, they find that presidential systems
tend to last considerably shorter than parliamentary systems. Their empirical results also
show that this difference is not due to the wealth of countries in which these institutions
were observed, to their economic performance, or to the social conditions under which
they emerged. Neither is it due to any of the political conditions under which they
functioned.

Other studies that focus on the developing world produce somewhat different
results. Power and Gasiorowski (1997) suggest that debates over presidential and
parliamentary systems for the consolidation of democracy have missed the forest for the
trees. Existing empirical works (Mainwaring 1993; Stepan and Skach 1993) exclude
young democracies, although much of the choice of institutions literature is obviously
studying the survivability of emerging democracies. Presidentialism is largely a
developing country phenomenon: The only advanced industrial democracy using pure
presidentialism is the United States. In their study of 56 transitions to democracy in
developing countries, Power and Gasiorowski (1997) find no significant difference in
the rate at which presidential and parliamentary regimes survive as democracies.

Cheibub (2002) supports the preceding argument by seeking to identify

mechanisms that underlie the supposed greater fragility of presidential systems. In a

23

study of all presidential and mixed1 regimes between 1946 and 1996, he finds that
neither divided government nor deadlock affect negatively the longevity of presidential
regimes or the survival in office of presidents and their parties.

Based on the arguments of Linz and Lijphart, Norris (1999) empirically shows
that people who live in parliamentary systems have slightly higher levels of institutional
confidence than those who live in presidential systems, while the difference is not
significant. After more than a decade of theoretical and empirical debates, the
fundamental question of “Presidential or Parliamentary Democracy: Does it Make a
Difference?” (Linz 1994) has not been conclusively resolved.

Persson, Roland and Tabellini (1997) identify (a) conflict of interest and (b) the
necessity of joint agreement as essential features in the achievement of good
governance. The competing bodies should keep each other honest, while being
individually accountable to voters. Presidential systems have a deeper separation of
powers and more direct accountability to the voters, compared to parliamentary systems
(Lintz 1994; Lijphart 1992; Shugart and Carey 1992; Tabellini 2000). While the latter
invites collusion among politicians, the former induces conflict and competition among
them. Thus, overall rents to politicians are likely to be larger in parliamentary systems.
The theory predicts that presidential democracies have smaller governments, fewer
rents, less public goods and more narrowly targeted redistribution, compared to
parliamentary systems (Persson, Roland, and Tabellini 2000). Since voters are assumed

to want rents as small as possible, while incumbent politicians want them opposite,

 

Mixed regimes are those systems rn which governments respond to both legislative assemblies and
elected presidents are mixed.

24

presidential systems are likely to produce higher levels of popular satisfaction with

democracy than parliamentary systems (H3).

2.4.4 Unitarism vs. Federalism

World-wide only 24 out of 192 states have a federal system (Derbyshire and
Derbyshire 1996), although many more have adopted some degree of regional
decentralization. Riker (1975, 101) defines that “federalism is a political organization in
which the activities of government are divided between regional governments and a
central government in such a way that each kind of government has some activities on
which it makes final decisions.” In contrast with Riker’s definition in terms of a
guaranteed division of power, Elazar (1997, 239) prefers focusing on the
“noncentralization” of power: he describes federalism as “the fundamental distribution
of power among multiple centers . .. , not the devolution of powers from a single center
or down a pyramid.” In addition to these primary characteristics, political scientists
identify several secondary characteristics of federalism: in particular, a bicarneral
legislature with a strong federal chamber to represent the constituent region and a
written constitution that is difficult to amend (Lijphart 1999).

Riker (1975) stresses three factors present in the US. form of federalism that he
claims to be true for federalism in general. First, Riker assumes that every long—standin g
federation is the result of a bargain whereby previously sovereign polities agree to give
up part of their sovereignty in order to pool their resources to increase their collective
security and to achieve other goals, including economic ones. Second, he assumes that

one of the goals of federalism is to protect individual rights against threats from the

25

central government through a number of institutional devices, such as a bicameral
legislature. Finally, as a result of the federal bargain that created the United States, each
of the states was provided with the same constitutional competences.

However, although these three points are a reasonably accurate depiction of the
political structures and normative values associated with US. federalism, most
democratic countries that have adopted federal systems have chosen not to follow the
US. model (Amoretti and Bermeo 2004). Indeed, American-style federalism embodied
some features that would be very inappropriate for many developing countries,
especially multi-ethnic polities. Most of the multi-ethnic democracies, for example, are
constitutionally asymmetrical: in order to hold the multi-ethnic polity together, they
assign different linguistic, cultural, and legal competences to different subnational units.

Federal institutions create, or at least reinforce, localistic politics: strong
MP/constituency bonds and greater policy authority at local levels. Politicians are apt to
be more aware of local conditions and more concerned with local-level results.
Federalism is suggested to institute a closer connection between government units and
the citizens they are intended to serve.

Federal institutions also create multiple veto points, thus raising the threshold
level of agreement necessary for reaching any political decision. Within federal systems,
decision must gain the consent of both houses within the legislature and local or state
governments. Proponents of federalism argue that the existence of multiple veto points
establishes a protective barrier against the passage of legislation representing special or
local interests (Lijphart 1999). It should also lead to greater consensus, since each group

is forced to seek the co-operation of at least most other groups.

26

In a unitary system, on the contrary, the state is one and indivisible, meaning
that the central government exercises authority over the people directly. Unitary
institutions lead to a more encompassing vision of public policy, where the logic of
issue and constituency definition is national rather than local (Gerring and Thacker
2004). The more encompassing these policy issues are, the less susceptible to special
interests and personalistic pressures they become (Crepaz 1996; Olson 1982). Local
governments encompass small and homogeneous social groups, almost by definition,
and are apt to be managed by close-knit power structures. In the absence of rival entities
there may be no effective way to counter official malfeasance. In addition, it is often
difficult to transfer corrupt officials out of local governments where they enjoy
protection and political support if the polity itself is highly decentralized. It may be
more difficult to find competent replacements since highly trained professionals often
resist taking up posts in the periphery.

By contrast, national bureaucracies are large and interdepartmental transfers
tend to be frequent. It is more difficult to maintain clientelistic networks under such
circumstances (Wade 1985). Anti-corruption programs are more often the results of
efforts by centralizing elites, whose orientation is national rather than local, whose
training is extensive and professionally oriented, and whose self-identified political
mission is to unify the nation rather than to represent local or particularistic interests.
Thus, the quality of public service is likely to decline as one move from centre to
periphery (Huntington 1968; Tanzi 1996). Using the data of 125 countries, Gem'ng and
Thacker (2004) find that centralized constitutions help reduce levels of political

corruption. The larger the setting, the more a policy is likely to rely upon universalistic

27

norms, ones conductive to general interests. Norris (1999) also finds that institutional
confidence is likely to be greater in unitary state than federal state. Thus, we should
expect higher levels of popular satisfaction with democracy in a polity that is unitary

(H4).

2.5 Partisan status and Political Institutions
2.5.1 Partisan status: Winner, Non-partisan and Loser

Democracy is about winning and losing within a context of set rules: “Since the
struggle for political office is bound to create winners and losers, this necessarily
generates ambivalent attitudes towards authorities on the part of the losers” (Kaase and
Newton, 1995: 60). As Ginsberg and Weissberg (1978) point out, every election
represents a test and potentially a threat to support for the political regime. Electoral
conflicts may strain public acceptance of legal and institutional processes. At the
simplest level, if people feel that the rules of the game allow the party they endorse to
be elected to power, they are more likely to feel that representative institutions are
responsive to their needs so that they can trust the political system, and consequently,
they are inclined to be satisfied with the govemment’s performance (Lambert, et al.,
1986) and with the way the system works (Kornberg and Clarke, 1994; Nadeau and
Blais, 1993). On the other hand, those whose preferred party loses the election are more
likely to feel that their voice is excluded from the decision-making process, producing
dissatisfaction with political institutions. A growing number of studies empirically show
that citizens whose party is in government (winners) are predisposed to evaluate the

govemment’s performance positively and to be more supportive of the political system

28

than those whose party is in opposition (Anderson and Guillory 1997; Anderson and
LoTempio 2002; Anderson and Tverdova 2003; Ginsberg and Weissberg 1978; Nadeau
and Blais 1993; Norris 1999).

In a study of eleven European democracies, Anderson and Guillory (1997) find
that respondents who voted for the winning party or parties are more likely to be
satisfied with how well democracy works in their country than respondents who voted
for the losing party or parties. As a result of elections, democracy produces winners and
losers. Because the political system is more favorable for people whose party is in
power, they argue, losers whose party is in opposition express lower levels of
satisfaction with the way democracy works than do winners.

Anderson and LoTempio (2002) examine the effect of voting for the winners
and losers of presidential and congressional elections on political trust. On the basis of
survey and electoral data for 1972 and 1996, they argue and demonstrate empirically
that presidential winner-loser status systematically affect citizens’ trust in government.
They find that Americans’ trust in government is strongly affected by the presidential
election, but not the congressional election. For example, voters for the losers of the
presidential election show lower levels of trust. Moreover, they find that voting for the
congressional winners does not attenuate this effect. Political trust is highest among
voters who voted either for both the presidential and congressional winners or the
presidential winner and congressional losers; trust is lowest among those who voted for
both the presidential and congressional losers or congressional winners and the

presidential loser.

29

Using 25 countries’ survey data, Norris (1999) finds that people who endorsed
the ruling party or parties exhibit significantly higher levels of confidence in political
institutions than do those who backed one of the opposition parties, which is consistent
with other works.

The differences in levels of political support between winner and loser in
emerging democracies are wider than those in developed democracies. While citizens
who live in Belgium, Germany, Luxembourg, and Denmark show relatively narrow
differences (less than 20 percentage points) in satisfaction with democracy between
winner and losers, those in young democracies such as Portugal, Spain, and Greece,
express significantly large differences of more than 30 percentage points (Fuchs,
Guidorossi, and Svensson 1995). Citizens living under the former regimes do not have
long enough experience with democracy. They have been watching just a few regular
multiparty elections and only some of them have experienced the ruling party
alternations in their country. In sub-Saharan Africa, most parties that won founding
elections are still in power (Van de Walle 2003). Those losers in this area have lower
expectations for their supporting party to win an election and more dissatisfaction with
the political system.

Yet losers play a crucial role in the stability and continued functioning of
political systems (Anderson, Blais, Bowler, Donova, and Listhaug 2005). Their regime
support has greater impact on the stability of the regime than winners’ regime support,
because losers are more likely to reject the rules of the game that are not responsive to
their interests and to look for alternatives assuring their needs. The disposition of losers

toward the result of not only founding but also regular subsequent elections is critical to

30

democratization. Anderson et al. (2005) empirically show that losing has a more
significant impact after the second rather than first loss.

The experience of losing in emerging democracies has more significant effects
on citizens’ attitudes towards political systems than the experience in established
democracies. While citizens in the latter have become accustomed to the experience of
losing as the fortunes of parties shift over time, those in the former will not have
experienced winning and losing. Those losers in new democracies need more time to
learn how to lose gracefully by alternations in power through consecutive elections. In
addition, in a new democracy, being the minority and majority can be expected to weigh
more heavily because citizens are less sure when and whether that there will be another
opportunity to determine who has the power to rule and who does not. To achieve a
consensus about the new rules of the game after a democratic transition, it is necessary
that losers will not be permanently excluded from political power, but will have
subsequent chances to contest and win elections. We saw several cases, like Cameroon
and Kenya, in which an attempt to marginalize the opposition from any share of power
was followed by a period of political conflict (Bratton and Van de Walle 1997).

Because democracies are inherently designed to create winners and losers
through a series of elections, there must be a tension between those in the majority and
those in the minority in democratic political life. That is, even when they perform
satisfactorily, democracies, by their very nature, face the potential for instability due to
the different preferences that those in the political majority and minority have for
legitimizing the system. I therefore expect that, relative to those in the political majority

(the winners), (HS-a) citizens who are in the political minority (the losers) should have

31

a diminished stake in the system. As a consequence, they are likely to express more
negative views of the political system’s performance. In addition, we are able to observe
a significant number of citizens who support neither ruling nor opposition parties in
emerging democracies. For non—partisans, this study expect that their levels of political

support should be somewhere between winners and losers (HS-b).

2.5.2 Partisan Status in Different Political Institutions

The distinctions between winners and losers may be more complex than those
discussed in the previous section. Different political institutions may also affect
citizens’ attitudes by providing incentives or disincentives for certain behavioral
patterns in different ways. Winning and losing have different meanings in different
political systems. Anderson and Guillory’s study (1997) shows that the differences
between winners and losers in consensual democracies are significantly smaller than in
majoritarian democracies. Using cross-sectional survey data for eleven European
democracies together with data on the type of democracy in which individuals live,
Anderson and Guillory explain cross-national differences in public satisfaction with
political system as the product of the type of democracy (majoritarian or consensual)
and the distribution of winners and losers in each country. They demonstrate that the
nature of representative democratic institutions mediate the relationship between a
citizen’s status (winner or loser) and his or her satisfaction with the way democracy
works.

Combining public opinion surveys from twelve West European citizens with

rating of nation’s institutional quality, Rohrschneider (2002) explains how political

32

institutions mediate the relationship between popular perception of under-representation
and mass support for the EU-wide government. Citizens would be likely to support a
system that provides a fair articulation of one’s interests. In particular, when national
institutions are corrupt, or bureaucracies are ineffective, people can be less certain that
their interests are considered properly. Empirical results of the study show that
representational perception is a strong predictor of the EU support and the effect of the
perception significantly varies across twelve countries. And, the quality of national
institutions explains why the influence of perceived under-representation on EU support
varies cross-nationally. Rohrschneider (2002: 470) demonstrates that “citizens are
especially likely to penalize the EU for not representing them when they reside in
nations with superior institutions.” He also empirically shows that there is no significant
direct effect of the institutional quality indicator on EU support.

Multiple political parties are able to represent a multidimensional citizens
preference configuration in the legislature (Powell 2000). Multiparty systems can easily
accommodate more issue dimensions than two-party or one-party dominant systems,
because issue dimensions have been defined in terms of differences between instead of
within parties. Lijphart (1999) finds that the effective number of parties in thirty—six
democracies is significantly and positively associated with the number of issue
dimensions. Citizens would have more chance to choose a party closely representing
their interest about any given issue in multiparty systems. Since two-party systems are
built on electoral systems discouraging to third parties, on the other hand, people whose
policy preference deviates from those of the two dominant parties such as non-partisans

will easily become alienated. They cannot find any political party for their own interest

33

among the established parties. In addition, any new party is unlikely to endure in two-
party systems. This increases the levels of political marginalization among people
whose party is not in government.

Opposition parties are one of the two main components of legislative
representation in a democracy. For meaningful representation in policy-making
processes, opposition parties representing minority’s interest should have some
significant role in the processes. The greater the opposition’s size in the legislature, the
greater, generally, its ability to challenge government policy proposals, offer
amendments and alternatives, criticize the implementation of policies by the
bureaucratic agencies. In a minority government, for example, the influence of
opposition parties is significant (Powell 2000). Minority governments must negotiate
with the opposition parties. In two- or one-dominant party systems, however, opposition
parties may do little more than use the legislature for mobilizing public opinion. They
have an indirect, limited impact on government decision-making. Under such a system
citizens whose party is not in power are less likely to see the ruling party as providing
viable alternatives to their own party. If they are dissatisfied with their own party but
see nowhere else to go, they may become increasingly cynical about the system as a
whole.

I therefore argue that the nature of representation in party systems systematically
mediates the relationship between individual’s partisan status (as a majority “winner” or
minority “loser”) and his or her political support. Greater ENPP is likely to decrease
differences in popular satisfaction with democracy both between winners and losers and

between winner and non-partisans. It is expected that (H6-a) losers (or non-partisans)

34

who live under greater ENPP should express higher levels of satisfaction with the way
democracy works than those who live under smaller ENPP. Figure 2.1 describes this
hypothesized interactive effect. The relationships revealed among party systems,
partisan status, and popular satisfaction with democracy will shed light on the question
concerning the effects of institutions on citizen attitudes toward how well their political

systems work.

Figure 2.1 Hypothesized Popular Satisfaction with Democracy in Different Party

Systems

Moreover, electoral systems may also play mediating roles. Proportional
electoral systems are held to produce more equitable outcomes between political
minority and majority and to encourage wider social group representation than
majoritarian electoral systems (Lijphart 1999). These features of PR systems are likely
to have a greater impact on losers than winners. Adopting a PR system increases the
possibility to build an alliance among political minorities and induces protest parties
into current political system that enhances political minority groups’ perceptions of
system legitimacy. Thus, minority groups feel that they are included and represented
within the structures of government which make decisions affecting their interests. It
should lead to increase their levels of satisfaction with democracy.

Majoritarian electoral systems, on the other hand, can reinforce the gap in
popular satisfaction with democracy between winners and losers. Winners literally take

all with maximum a half of votes. Majoritarian electoral systems create an incentive for

35

winners (political majority) to bring more pork to their region or ethnic group at the
sacrifice of the numerically smaller population groups who lose elections. This
substantive exclusion increases the degree to which a minority group feels that it is
excluded from the system. This strengthens losers’ dissatisfaction with democracy and
threatens the stability of emerging democracies.

I therefore argue that the nature of representation in electoral systems
systematically mediates the relationship between individual’s partisan status (as a
majority “winner” or minority “loser”) and his or her political support. If an electoral
system increases representation of political minority groups and causes people to have
more favorable evaluations of political processes, we expect to see a shift particularly in
political minority’s attitudes that reflects greater satisfaction with democracy. The more
proportional electoral systems decrease differences in satisfaction with democracy both
between winners and losers and between winner and non-partisans. It is expected that
(H6-b) losers (or non-partisans) who live under proportional electoral systems should
express higher levels of political support than those who live under non-proportional
electoral systems. Figure 2.2 describes this hypothesized interactive effect. The
relationships revealed among political systems, partisan status, and political support will
shed light on the question concerning the effects of institutions on citizen attitudes

toward how well their political systems work.

Figure 2.2 Hypothesized Popular Satisfaction with Democracy in Different Electoral
Systems

36

853A— .ES—SEaF—am he aux—:32 958.5—

 

flamed

 

€36.89an

it 95:53

 

 

Bo:

menace—don—
5v:
neuuummumm

emf

386$ .53.— anoth E 35.8259 5MB :euoflmzam cosmmofioaml flu 0.53.”—

37

956% 3.38.”.—

RcoEomoi SWEEOFE

 

flowed

 

mnemuumméo Z

J @8555

 

 

Bod

3.28599
:33
:euoeumuem

as:

255$ fiasco—m Eur—DEG E menace—En :33 553.35% gamma—533* Na charm

38

Chapter 3. A Model of Political Support

3.1 General Research Strategy

1 adopt two research strategies in my efforts to explain different levels of
political support. In this project, I utilize a cross-national data set composed of 35
fledgling democracies. This approach is complemented by a case study of Lesotho,
which allows for an insight over time. This combined approach offers different
methodological, empirical, and theoretical benefits, which are discussed further in this
chapter.

Lijphart (1971) called comparing many countries using quantitative analysis the
“statistical” method. Comparing many countries most closely approximates the
experimental method of science, since it is particularly suited to quantitative analysis
through measurement and analysis of aggregate data. The extensive coverage of
countries allows for stronger inferences and theory-building, since a given relationship
can be demonstrated to exist with a greater degree of certainty. Despite the advantages
of comparing many countries, there are some distinct disadvantages, including the
availability of data and the validity of measures (Landman 2000).

Single-country studies that merely describe or interpret political phenomena
have been referred to as “interpretative” (Lijphart 1971: 691). These types of studies are
not comparative but are useful for comparison purely for their information. Single-
country studies are also useful for generating hypotheses on theories that have yet been
fully specified. My study of Lesotho, however, does more than this. This study makes
use of a quasi-experimental comparison across time in Lesotho (before and after

institutional reform) to analyze the impact of political institutions to promote political

39

support among citizens. In this chapter, I introduce the multilevel analysis used to
conduct this research and discuss the operationalization of key variables with

hypotheses to be tested.

3.2 Multilevel Analysis

This research makes use of data measured at multiple levels, such as individual—
level survey data from the Afrobarometer, Koreabarometer, Latinobarometer, and the
Taiwan Election and Democratization Study, and national-level contextual data.
Respondents are the level 1 units nested within nations, which comprise the level 2
units. Thus, this research combines multiple levels of analysis into a single
comprehensive model by specifying predictors at different levels. The goal of multilevel
analysis is to account for variance in a dependent variable that is measured at the lowest
level of analysis by considering information from all levels of analysis.

This study adopts contextual analysis of political behavior. Multilevel analysis
allows researchers to explore causal heterogeneity (Western 1998). By specifying cross-
level interactions, it is possible to determine whether the effect of lower-level predictors
is conditioned or moderated by hi gher-level predictors. In other words, is there a single,
uniform relationship or does relationship vary across higher levels of analysis? The
contextual analysis of political behavior assumes that environmental factors interact
with individual factors to shape political behavior.

Contextual analysis focuses on general concepts applicable across many
countries. It looks at general political attributes in many countries, for example whether
the constitution is unitary or federal, or the system presidential versus parliamentary. In

a contextual analysis, individuals’ attitudes and behaviors are assumed to be affected by

40

their surroundings. People form attitudes about politics in systemic contexts whose
institutional structures mediate preferences, define the choices available, and provide
citizens with opportunities to be heard in the political process. This type of analysis
goes along with the objective of comparative analysis to “reduce proper names to
explanatory variables” (Przeworski, 1987: 38f). Explanatory variables are abstractions
describing a context that can in principle be found in more than one country. Since
political institutions are relatively stable phenomena, the most appropriate evidence
concerns a cross-national comparison, which maximizes the variance in institutional
arrangements.

Multilevel analysis can provide a test of the generalizability of findings
(Steenbergen and Jones 2002). Because it allows researchers to explore causal
heterogeneity, multilevel analysis can contribute to these generalizability tests. If the
contextual units are randomly sampled, as multilevel methods typically assume,
multilevel analysis may have the added benefit that it helps overcome case selection
problems that often plague comparative research (King, Keohane, and Verba 1994).

In addition to those substantive merits, there is also an important statistical
motivation for explicitly accounting for multilevel data structure. Specifically, ignoring
the multilevel character of data carries significant statistical costs in the form of
possibly incorrect standard errors and inflated Type I error rates ( Raudenbush and Bryk
2000; Steenbergen and Jones 2002). In the context of multilevel data structures, the
correlation between observations within contexts is referred to as intra-class or cluster
correlation. In most cases this correlation is non-zero, and it causes the estimated

standard errors to be too low and the test-statistics too high if the intra—class correlation

41

is not accounted for in the statistical model. As a result, predictors appear to have a
significant effect when in fact they do not. By using a multilevel analysis, I try to
examine the influence of different political institutions on the level of political support

across emerging democracies.

3.3 A Model of Political Support
3.3.1 Dependent Variable: Satisfaction with Democracy

The concept of political support, in its general meaning, “refers to the way in
which a person evaluatively orients himself to a [political] object through his
attitude” (Easton 19752436). Political support is a multi-dimensional phenomenon:
political support for the community, regime principles, regime performance, regime
institutions, and political actors (Norris 1999). This study focuses attention on regime
performance, measuring support for how democratic political systems function in

practice. It is commonly measured by “satisfaction with the way democracy works in

[the country].”2 In other words, the question is concerned with how democracy
functions in practice as opposed to the ideal.

The concept of political support on which my analysis is based requires only a
“low level of generalization” (Fuchs, Guidorossi, and Svensson 1995, p. 330; see also
Anderson and Guillory 1997, p. 70). The item of “satisfaction with democracy” is not
intended to include any information on political leaders, parties, or policies. Therefore,
this item dismisses the possibility to tap support for authorities. Fuchs (1993, p. 240)

argues that because the question on “satisfaction with democracy” asks about the

 

The question asks, “On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all
satisfied with the way democracy works in [country] ?”

42

functioning of democracy, the item “refers to the informal structure of the regime; in the
generalization hierarchy it is between attitudes in respect to the formal structure and
those in respect to the authorities. In support of this claim, Fuchs shows that the item of
“satisfaction with democracy” correlated with items that tap attitudes regarding the
formal structure (0.49) and incumbent authorities (0.50).

There has been a debate regarding what dimension or dimensions of political
support the measure of “satisfaction with democracy” represents. A number of scholars
criticize it for its ambiguity, as it can tap multiple dimensions of political support, for
example both support for democracy as a value and satisfaction with the incumbent
government (Norris 1999; Canache et a1. 2001; Linde and Ekman 2003). Instead, Norris
develops an institutional confidence scale combining confidence in five major political
and civic institutions, including parliament, the civil service, the legal system, the police
and the army to measure support for regime institutions.

The question on “satisfaction with democracy,” however, has most commonly
been used as a measure of system support (e. g., Anderson and Guillory 1997; Farrel and
McAllister 2003; Fuchs 1999; Fuchs, Guidorossi, and Svensson 1995; Harmel and
Robertson 1986; Klingermann 1999; Lockerbie 1993; Morlini and Tarchi 1996; Toka
I995; Weil1989). While they criticize that the item of “satisfaction with democracy”
represents multiple levels of political support, Canache et a1. (2001) show that it is
consistently and significantly related to system support in 17 Latin American countries.
In addition, because this item has been included in various survey data, it fits very well
for cross-country analysis. Using the Afrobarometer Round 1 data set, including 12 sub-

Saharan African countries, Bratton, Mattes, and Gyimah-Boadi (2005) find that popular

43

satisfaction with democracy has significant correlations with the following three
dimensions: support for democratic regime, trust in state institutions, and evaluation of
government performance. Moreover, the evaluation of government performance has a
consistently powerful relationship with popular satisfaction with democracy across 12
countries.

In addition, a significant number of studies find that satisfaction with democracy
is clearly an indicator of actual system support and not conterrninous with support for
the incumbent government (Clarke and Kornberg 1992; Kornberg and Clarke 1994).
Fuchs examines the validity of the satisfaction with democracy indicator and finds that
the results constitute “a successful validation of the indicator as a measuring instrument
for a generalized attitude towards the political system on the legitimacy dimension”
(1993: 242). This study uses the item of “satisfaction with democracy” as a dependent
variable to measure the levels of popular satisfaction with democracy across 35
developing countries because it refers to a certain institutionalized form of democracy
in a particular country. This research is interested in understanding citizen’s responses

to the process of democratic governance.

Figure 3.1 Satisfaction with Democracy, by country

Combining the percentage of respondents saying they are “very” or “somewhat
satisfied” with the way democracy works, Figure 3.1 shows the distribution of
satisfaction with democratic performance across the countries investigated in this study.
Satisfaction ranged from 83 percent in Kenya in 2003 to 18 percent in Peru in 1998,

with a mean of 49 percent, and a standard deviation of 17.2. Taking the 50 percent

threshold as a criterion, then a majority of people in 16 of the 35 countries reported that
they were very or somewhat satisfied with the way democracy works in their country.

In 16 sub-Saharan African countries on the whole, more than half of all
respondents (59 percent) said they were satisfied with democracy in the period between
mid 2002 and mid 2003. An absolute majority was satisfied in 12 out of the 16
countries under study. In Cape Verde and Nigeria, only a third of the respondents were
satisfied with democracy. For Nigerians, popular satisfaction with democracy shows
significant change from 84 percent in 2000 to only 34 percent in 2003 (Afrobarometer
Network 2004). The high level of satisfaction in Nigeria 2000 might reflect a
honeymoon period for the restoration of civilian rule, but the extremely low level in
2003 showed that the new political system failed to reach to Nigerian expectation with
real progress.

Unexpectedly, the Afrobarometer mean score is comparable with the
Eurobarometer mean score on the same item in 1996 (56 percent). However, this high
level of satisfaction with democracy among sub-Saharan African countries can be traced
back to the late 19903 when almost 60 percent of the respondents said they were
satisfied with democracy (Bratton, Mattes, and Gyimah-Boadi 2005).

In Latin America on the whole, only 38 percent of the respondents are satisfied
with the way democracy works in 1998. Only in Costa Rica (61 percent) Guatemala (57
percent) and Uruguay (69 percent) were majorities of the respondents satisfied with
democracy. In Brazil, Colombia, Chile, Mexico, Nicaragua, Paraguay, and Peru, no

more than a third of the public is satisfied. Since 1996, the regional figure has never

45

exceeded 41 percent. In 2000, only 37 percent of the respondents said that they were
satisfied with the way democracy works (Lagos 2001).

The item of satisfaction with democracy is less stable than the item of support

for democracy.3 For example, popular support for democracy in Nigeria has dropped
from 82 percent in 2000 to 67 percent in 2003, while Nigerians’ satisfaction with

democracy has fallen by 50 percentage points (Afrobarometer Network 2004).
Figure 3.2 Satisfaction with Democracy, by Continent

Figure 3.2 shows the distribution of satisfaction with democracy across three
regions. In both sub-Saharan Africa and Latin America, almost one-third of respondents
belong to extremes. On a 4-point scale, only 16 percent of sub-Saharan Africans said
that they were very satisfied with democracy in their countries, and 12 percent of Latin
Americans. Only 6 percent of the public was very satisfied with democracy in Asia,
including South Korea and Taiwan. In a complementary manner, 16 percent of Latin
Americans reported that they were very dissatisfied, and 15 percent of sub-Saharan
Africans. Only 8 percent of the respondents were very dissatisfied in South Korea and
Taiwan. Many citizens in both sub-Saharan Africa and Asia were moderately satisfied

with democracy, while the majority in Latin America were moderately dissatisfied.

 

3 A standard question about support for democracy asks respondents whether democracy is always
preferable to any other form of government, whether there are certain circumstances in which non-
democratic form of government might be preferable, or whether the form of government really does not
matter to a person like themselves.

46

3.3.2 Independent Variables

3.3.2.1 Party system: Effective Number of Parliamentary Parties4

H I : ENPP and satisfaction with democracy should have a positive relationship.

In this study I test whether various institutional characteristics are indeed
systematically associated with levels of popular satisfaction with democracy in
emerging democracies. First, I measure party system fragmentation with an index
developed by Laakso and Taagapera (1979), which measures the effective number of

parties (N) in a system as follows:

1

n

ZS?

i=1

 

N:

in which s,- is the proportion of seats of the i-th party (Lijphart, 1999: 68). This
formula contains information about the number and relative size of the parties in the
system. It helps to differentiate between two- and multi-party systems, but also offers a
more subtle measure than simply counting the number of parties that gain representation

or receive votes.

. . 5
3.3.2.2 Electoral System: Index of Disproportionality (I.D.)
H2: More proportional electoral systems should produce higher levels of popular

satisfaction with democracy.

 

4 . . .
Primary sources: Nohlen, Krennerich, and Thubaut (1999); LeDuce, Nremr, and Norm (2002); Election
around the World available at www.clectionworld.org

5 Primary sources: Nohlen, Krennerich, and Thubaut (1999); Election results around the World available
at www.clectionworld.org '

47

Second, electoral systems are measured by the degree of disproportionality in
the translation of votes into seats. This study uses an index developed by Michael

Gallagher ( 1991) as follows:

 

. . 1
Gallagher 5 least-squares index = [72(121- -Si)

in which v, is vote percentage and s, is seat percentage. The higher the figure the greater

the disproportionality produced by the electoral systems. This factor tests whether or
how increasing proportionality of electoral systems effects on citizens’ attitude toward

how well the system works.

3.3.2.3 Presidentialism
H3: Presidential systems should produce higher levels of popular satisfaction with

democracy than parliamentary systems

Third, presidentialism refers to a system where policy-making power is divided
between two directly elected bodies, the legislature and the president. Parliamentarism,
by contrast, is a system of government in which the executive is chosen by, and
responsible to, the legislature. These two concepts represent poles at each end of a
continuum, which may be conceptualized along two dimensions: (a) the degree of
separation between president and parliament, and (b) the relative power of the two

players (the more power the president possesses the more “presidential” is the resulting

48

system). I create three categories: 1, parliamentary; 2, semi-presidential; 3,

. . 6
presrdentral.

3.3.2.4 Unitarism
H4: Unitary systems should produce higher levels of popular satisfaction with

democracy than Federal systems.

Fourth, this research hypothesizes that a unitary state structure, as opposed to a
federal one,should lead to higher levels of popular satisfaction with democracy, all other
things being equal. By unitarism this study refers to a political system where the
national government is sovereign relative to its territorial units (if any). Following
Gerring and Thacker’s (2004) work, I conceptualize unitarism as a continuum that
varies along two dimensions: (a) the degree of separation (independent) between
national and subnational units, and (if any separation at all) (b) the relative power of the
two players (the more power the center possesses, the more unitary the system). Of the
many institutional factors that may determine variation along these dimensions, two
predominate: federalism and bicarneralism.

Federalism refers to a permanent and highly institutionalized sharing of
responsibilities between a national authority and semi-autonomous regional units. Since
this sharing of responsibilities takes a variety of forms and is not always formally
prescribed, this study constructs three coding categories: 1, non-federal; 2, semi-federal

(where there are elective legislatures at the regional level but in which constitutional

 

6 Primary sources: Delury (1999); Derbyshire and Derbyshire (1996).

49

sovereignty is still reserved to the national government); 3, federal (elective regional

legislatures plus constitutional recognition of subnational authority).
Bicameralism refers to the sharing of policy-making power between two
chambers at the national level. Since, like federalism, bicameralism is often a matter of

degrees, I code cases into three categories: 0, unicameral (no or weak upper house); 1,

weak bicameral (upper house has some effective veto power); and 2, strong bicameral.8

I construct the unitarism variable by adding the scores of each country together
on these two components (federalism and bicameralism) and reversing the scale, thus
creating a scale from 1 to 5. In a fully unitary state, federalism has no constitutional
standing, no independently elected territorial legislature, and no specific policy
purviews reserved for them. The opposite of a unitary state is a federal, bicameral one.
In this study, I generally refer to this case as “federal”, although all constitutionally

federal states are not necessarily bicameral.

3.3.2.5 Partisan status: Winner, Loser and Non—partisan
H5-a: Loser would express lower levels of satisfaction with democracy than winner.

HS-b: Non-partisan should show lower levels of satisfaction with democracy than

winner.

Finally, this study classifies respondents as being a loser or non-loser with the

help of an Afrobarometer survey question that asks which party, if any, the respondent

 

7 Principal sources: Derbyshire and Derbyshire (1996); Watts (1997).
8 Principal sources: Patterson and Mughan (1999); Tsebelis and Money (1997).

50

feel close to or a Latinobarometer survey question that asks which party the respondent

would vote for.9 This research then combines these responses with information about
the party or parties that won the most recent presidential, legislative, or general
elections to identify loser. I created a dummy variable of loser. If the respondent felt
close to or would vote for an opposition party, I coded that individual as “1” in the loser
dummy variable. Otherwise I coded a respondent as “0.” In addition, survey data shows
that there is a significant number of respondents who did not feel close to any party,
refused to answer the question of party affiliation or said “don’t know.” Thus, I created
another dummy variable for non-partisan. Those two dummy variables show the
differences of popular satisfaction with democracy both between winner and loser and

between winner and non-partisan.

3.3.2.6 Partisan status and Party system
H6: Loser (or non-partisan) who lives in more ENPP party systems should express

higher levels of satisfaction with the way democracy works than loser (or non-partisan)

in less ENPP party.

To examine whether party system would mediate the relationship between
minority-majority and popular satisfaction with democracy, I include two interaction
terms - loser*ENPP and non-partisan*ENPP - from two different levels, partisan status
from individual level and ENPP from country level. The coefficient of the two

interaction terms would show whether and how party systems explain the different

 

9 The Taiwan’s Elections and Democratization Study (2003) and the Koreabarometer (2001) have the
same question as the Afrobarometer has. ‘

51

degrees of the relationship between partisan status and popular satisfaction with
democracy across emerging democracies. I expect that the coefficients of the two
interaction term should be positive. The greater the ENPP is, the smaller the differences
in popular satisfaction with democracy both between winner and loser and between

winner and non-partisan are.

3.3.2.7 Partisan status and Electoral System
H7: Loser (or non-partisan) who lives in more proportional electoral systems should
present higher levels of satisfaction with democracy than loser (or non-partisan) who

lives in less proportional electoral systems.

To examine whether and how electoral systems mediate the relationship
between partisan status and popular satisfaction with democracy I create two interaction
terms, i.e., loser * ID. and non-partisan*I.D. in the model. The coefficient of the two
interaction terms measures the effect of different electoral systems on the relationship
between partisan status and popular satisfaction with democracy across countries in this
study. It is expected that both interaction terms should have a negative coefficient.
Increasing disproportionality of electoral systems should induce greater differences in
popular satisfaction with political systems both between winners and losers and between
winners and non-partisans.

Partisan status (winner or loser) is the product of electoral contests among
political parties. Citizen’s vote for a party or a candidate is collected and counted by a

certain electoral rule on which members of a polity have agreed. Thus, we can expect

52

that party and electoral systems could have a mediate effect the relationship between
partisan status and popular satisfaction with democracy. However, either the
relationship between executive and legislative branches or the relationship between
national and regional units does not have any close connection with partisan status. This
study looks at neither the interaction between partisan status and presidentialism nor the

interaction between partisan status and unitarism.

3.3.3 Control Variables

3.3.3.1 Democratization: Freedom House Index

For the purpose of this study, I adopt the Freedom House Index10 to control for
the different degrees of democratization across the 35 emerging democracies in the
model. The mean combined scores for political rights and civil liberties for the survey
year of each country is used in the analysis. Across countries it would be expected that
the existence of widespread political rights and civil liberties should be associated with

higher levels of public satisfaction with democratic governance.

3.3.3.2 Economic Development: GDP per capita in 2001 (US$)11
To control for the level of economic development across the countries under this
study, I introduce GDP per capita in 2001 (natural log in PPP US$) in the model. Lipset

(1959) observes that democracy is related to economic development. Because the

 

10 Freedom House, “Freedom around the World,” 2001-2002. www.freedomhouse.com.
H UNDP. The UNDP Human Development Report 2003 (New York: Oxford University Press).

53

intensity of distributional conflicts is lower at higher income levels, citizens are likely to

show higher levels of satisfaction with democracy.

3.3.3.3 Sub-Saharan Africa and Asia: two dummy variables.

There are strong reasons to believe that geography might also correlate
systematically with popular satisfaction with democracy. To capture geography, this
study uses two dummy variables for continental location. They refer to countries in sub-
Saharan Africa (sub-Saharan Africa) and eastern Asian (Asia). The base group of
countries thus consists of countries in Latin America. Table 3.1 lists country-level data

included in the study.

Table 3.1 Political and Economic Indicators in the countries under study

3.3.3.4 Interest in Politics

Many political scientists have demonstrated that interest in politics is related to
political support (Almond and Verba, 1965; Anderson and Guillory, 1997; Weatherford,
1991). Citizens who understand the political process and believe that their participation
can influence policymaking are likely to take a more optimistic view of democratic
governance. In other studies, political interest has been taken as a direct measure of
cognitive competence which stimulates a critical skepticism towards political system
(Listhaug and Wiberg 1995). The survey question used here asked respondents how
interested they are in public affairs. I expect to find that those who are more interested

in politics are also more likely to be satisfied with the way democracy works.

54

3.3.3.5 Economic Performance Evaluations: sociotropic and egocentric evaluations
Citizen evaluations of system outputs can be one of the most important factors
shaping the reputation of political institutions and of the political system as a whole
(W eatherford, 1987). In particular, several studies suggest that economic performance,
as well as public perceptions of both personal and national economic conditions, are
related to system support (Clarke, Dutt, and Kornberg, 1993; Kornberg and Clarke,
1992; Listhaung and Wiberg, 1995 ; Weatherford, 1984, 1987). For this study, I use
responses to two survey questions that asked respondents to evaluate national and
personal economic conditions, that is, to form sociotropic and egocentric evaluations of
economic performance. I expect that those who evaluate economic performance

positively should be more satisfied with the way democracy works.

3.3.4 Demographic Variables: education, gender, and age
I control for the usual socio-demographic variables: education, gender, and age.
Coding procedures and descriptive statistics for the individual-level variables used in

this study are shown in Appendix A.

55

Figure 3.1 Satisfaction with Democracy, by Country (Africa and Asia)

 

 

°°.
W
\O
N.
\O
\D
”l
00
V0
9
.—
h
(\I
M
l\

 

 

50.6 48.7 48.2

 

 

56

V 0 0T 94. 59 cow» 00 OxU
V 00+9£Veo 9%. “WC/0 9@0 00 @VA. $0 900 ’70? 0000 09460: 09 A09 90 othO
44 as», 559 4a.. as 9/ 5% 500% 524 e? e op so. 9% so aw

 

 

 

 

 

 

 

$05848. used 3550 .3 iofiooEoQ 5.3 :ouoflmuwm Acoscucouv mm Semi

57

:5 «a 82 .98.» 82 bum—mm .95.»

:38 :22 B
52: EotoE< .53-

52: gfimbsm I

 

 

 

 

 

 

 

 

 

 

E25200 .3 SomhooEoQ FEB cowoflmumm Nd “mama

58

 

 

 

 

 

 

 

 

 

 

 

omwd m Bastian Nwmd o~.~ Wm 3598
8% m 3.528.; 5:. CS 2. 52.2% _
ow~.m v 328285 wood 3% Wm 83:
ca; m 320285 $5. wfim Wm BBQ
82 m 3:52.85 NE? SN Wm amazon
8:. m 335285 93. 84 ON «552
omvd m §u=o§=3m Bad 2 .m m .N 9:83
oom.m N 380285-38 Sim cmd od «Sm—oi
986 _ m fizaoflmofiéuom em; and ed 88M 5.5m
2w m RucuEmoE cow; _ _m._ ed :2);
cad v RucoEmoE 2 fin S .N QN «SEC
on: m 55.5.5; m2: NE 3 «$58
coed _ 3.5285 RNN S.m m; .sta
83.5 m 3528:. ~93 m3 2 82 s89
wwmé m 335235488 086 :3 m. _ 533.
Ema m EEoEmui-§om «.23 SN 2 oEo> 0&9
8a.: _ 3.5.5.5; 2.3 Ba 2 8.5. 58m
A 58
mm: man: tag 523$
SE3 .89.. 02.0...
.8. .50 a "V 525...: 521.2853; 52353.35 Ezm 5282". .5560

 

:35 .m5.... 85.58 5 33885 ufigsm Ba .85....— 3 ass...

59

 

 

 

on. . 3.523... NS... 2.. o... 2.0%
can m 3.528... .2... w; Wm .3335.
92.... . 3.523... amnm cod m.m 8.5.2.
8.... m 3:5...on one.» 3.... Wm «Havana
8.: N 3.52.... 3.3 3.5 mm .3528
com... . 3.528... 3.3. m2. m.m ESE
9.... m 3.528... 63 3.. nm BESS...

0% v 3.523... mad. .2 m.m £5.58...

chm m 3.5285 :23 $9. Wm .332.
cam... . N EscoEmoE 896 onm o.m museum}.

93 m 3.528... ammd N..m ed 99.3.
chew . 3.528... m2: Ed Wm 233:8/
enhm m 3.523E Rod. mad Wm «:85».
amtm m Bustier. to... and Wm 2.3.8.2
8.3 m 3.528... mafia own Om 32.6.5:
coma m 3.523... .3... and Wm 8.2.28 .m
.Foom

mm: mam. cam.» .623.
3.13 .85... 3:0:
.8. .30 .m u. 52.2.5 5a.»..m .8..._o.. b._..=o...o..o.%.a anzm 588... .3..on

 

 

 

 

 

 

 

 

.25 3...... $5.55 a. 238...... 9.588.. .2... .82....— ..§=....so. 3 2......

60

£3315... .355 3.020 9.03 €380... 82.82....— Hzouaz .2.ng E d

 

 

 

 

 

 

 

 

 

 

mwofi mm a... moms wo.m ad 55:.
owm.m m 3.538... mood mo. . o6 cananEN
8.... m 3.525... a .. ca 3.5...
Em... m 3.338... 3nd mm... m... =5...
own m 3:538... man... .2 o... «53,
Scam
am: an... .59. .625.
233 .805 3:0:
.3 no... a u. 52.8.5 523$ .82.... 3:22:8235 Ezm £83... 3.5.8

 

 

......m 5...... 5......50 2 535...... .388... 2... .52.... .52....8. 3. 2......

61

Chapter 4. Individual-Level Analysis

4.1 The Effects of Partisan Status on Satisfaction with Democracy at the Aggregate
Level

In emerging democracies, political parties may be the most important and the
only collective actors linking citizens with the political decision-making processes.
The manner in which they link citizens depends decisively on whether the party is in
or out of government. This difference may also have an impact on citizens’
subjective perceptions. Citizens whose party is in power (winners) certainly have the
feeling that their interests are considered in political decision-making, in contrast to
citizens whose party is in opposition (losers). Moreover, supporters whose party is in
power are likely to see the functioning of democracy in their country in a more
positive light than those whose party is not in government.

Are losers less satisfied than winners with the way democracy works? Figure
4-1 offers some preliminary evidence at the aggregate level. To examine differences
in satisfaction between losers and winners, Icompare satisfaction with the way
democracy works between these two groups across 34 countries, ordered according
to the gap in popular satisfaction with democracy between winners and losers from

the largest to the smallest.

Figure 4—1 Satisfaction with Democracy and Partisan Status

Figure 4-1 clearly illustrates that there are differences in satisfaction between
winners and losers, regardless of the general level of satisfaction with democracy.
The results are consistent with my expectation that winners should be more satisfied

with democracy than losers. Winners show higher levels of satisfaction with

62

democracy than do losers in 33 of the 34 countries. While there is a certain gap
between winners and losers, a significant number of losers are satisfied with
democracy. In 11 countries 50 percent and more of opposition party supporters are
satisfied with democracy. Although there is variation across countries in the strength
of the relationship between partisan status and popular satisfaction with democracy,
the differences exist also within every country examined. For instance, there is a
strikingly large difference in popular satisfaction with democracy between winners
and losers in the 5 countries: between 30 and 40 percentage points in Malawi,
Nicaragua, and Tanzania, 47 percentage points in Uganda, and 58 percentage points
in Zimbabwe.

The largest gap in popular satisfaction with democracy between citizens
whose party is in government and those whose party is in opposition in Zimbabwe
reflects the present political crisis in the country (Makumbe 2002). People in
Zimbabwe had rarely been able to make an electoral choice until the 2000 elections,
when the new party Movement for Democratic Change (MDC) posed a serious
challenge to the ruling Zimbabwe African National Union-Patriotic Front (ZANU-
PF), although Zimbabwe is one of the very few African states that have had a
continuous multi-party system since majority rule in 1980 (Laakso 2002). Elections
did not threaten the position of the ruling party, but rather confirmed this power in a

ritualistic manner. In the parliamentary elections of June 2000, however, the MDC

won almost half of the contested seats in the legislature. 12 The presidential elections
of 2002 in which President Mugabe from the ruling ZANU-PF won with 56 percent
against Tsvangirai from the MDC with 42 percent, increased the conflict between

government and opposition parties. The elections were reportedly marred by

 

12 MDC took 57 out of the 120 contested seats and the ruling ZANU-PF won 62 seats.

63

disenfrenchisement of urban voters, violent intimidation against opposition
supporters, intimidation of the independent press and other irregularities. The MDC
alleged that Mugabe’s victory was due to a number of irregularities and petitioned
the High Court to declare the election invalid.

In Malawi, the United Democratic Front (UDF) has won the last three
presidential elections, since its leader Bakili Muluzi replaced Dr. Hastings Kamuzu
Banda, the leader of the Malawi Congress Party (MCP), as the head of state in the
1994 presidential elections. While there is no formal rule to hinder conduct of
opposition parties after the transition to multi-party system, Malawi has been mostly
ruled by a defacto one-party government (Englund 2002). The UDF government
has retained the strong executive powers of the state president and promoted the
exclusiveness of power through neo-patrimonial networks. Thus, opposition parties
and their supporters express higher levels of dissatisfaction with democracy. In
addition, people in Tanzania have experienced similar political processes. Although
Tanzania has held multi-party elections since 1995, the Chama Cha Mapinduzi
(CCM) won last two presidential elections and still is defacto in power.

Thus, the big differences in popular satisfaction with democracy between
those two political groups come from the fact that supporters of opposition parties
have no positive experiences with the political systems over the past years. In sub-
Saharan Africa, in particular, citizens have been experiencing only several regular
multiparty elections and only a few have observed the government party alternation
in their country. People with lower expectations for their party to win an election are
likely to show lower levels of satisfaction with democracy.

On the other hand, the experience with relatively smooth electoral turnovers

of ruling parties leads to smaller gap in popular satisfaction with between winners

64

and losers in emerging democracies. In Kenya, 80 percent of opposition party
supporters said they were satisfied with the way democracy works. The timing of the
Afrobarometer survey was unusual in Kenya in 2003 (Wolf, Logan, and Owiti 2004).
The survey was conducted only eight months after the first electoral transfer of
power in Kenya’s history, overlapping with some researchers refer to as a
“honeymoon period” of indeterminate length, during which citizens show some
patience with the continuance of inherited problems (Aslund 1994; Balcerowicz
1995). As Figure 3-1 shows, 83 percent of the respondents said they were satisfied
with democracy. The very high levels of satisfaction with democracy among the
population at large suggests that satisfaction is high among the supporters of both
ruling and opposition parties.

In Mexico, losers express even higher levels of satisfaction with democracy
than winners. The 1997 legislative elections led to the first experience of divided
government in Mexican politics since the 19208 (Weldon 1997). The polls failed the
ruling Partido Revolucionario Institucional (PR1). With just under 40 percent of the
effective vote (the total vote for all parties above the 2 percent legal threshold), the
PRI won 239 of 500 seats in the Mexican Chamber of Deputies. The Partido Accion
Nacional (PAN) and the Partido de la Revolucion Democratica (PRD) divided the
opposition vote evenly. The former won 121 and the latter 125 seats. Two small
parties, the Partido Verde Ecologista de Mexico (PVEM) and the Partido del
Trabajo (PT), won 15 seats. Those respondents, supporters of opposition parties,
might be pleased with the fact that their parties formed a majority in the Chamber of
Deputies for the first time and had an influence on policy-making processes. In 2000
Vincent Fox Quesada from the PAN won the presidential election with 43.4% of

valid votes.

65

In most countries under this study, it is possible to see a difference in the
level of satisfaction with democracy between supporters of ruling and opposition
parties. In one-third of the countries, a majority of losers are also satisfied with the
functioning of democracy in their country. It suggests that there might be a

legitimacy problem in the other two-thirds of emerging democracies under study.

4.2 Multivariate Single-Country Models

Next, I examine the performance of the individual-level variables in a series
of multivariate single-country models. Linear Regression Models are commonly
used for this type of analysis. However, when a variable is ordinal — a common

outcome in a survey analysis — while its categories can be ranked from low to high,

. . . l3 . .
the distances between adjacent categories are unknown. Since the Linear
Regression Model includes the implicit assumption that the intervals are in fact

equal, it may not be the best tool for evaluating survey results. Instead, I use Ordered

Probit Analysis (Long, 1997: 114-147) to obtain maximum likelihood estimates.14

Table 4-1 shows the results.

Table 4-1 Effects of Individual-Level Variables on Satisfaction with Democracy in
34 countries

I find that the effects of partisan status — loser and non-partisan - remain

even when I control for a number of other variables. The effects are strong and

 

l3 Researchers often treat ordinal dependent variables as if they were interval, numbering the

dependent categories sequentially and using the Linear Regression Model, which assumes that the
intervals between adjacent categories are equal. For example, the distance between strongly agreeing
and agreeing is assumed to be the same as the distance between agreeing and being neutral on the
survey statement. However, both McKelvey and Zavoina (1975: 117) and Winship and Mare (1984:
521-523) give examples where regression of an ordinal outcome provides misleading results.

14 There was no significant difference between the Ordered Probit model and the OLS one.

66

consistent with the expected direction. The relationship holds in 26 of 34 countries,
indicating that losers are significantly less satisfied with the way democracy works
than are winners. The results also show that non-partisans express significant lower
levels of satisfaction with democracy than do winners in 29 of 34 countries. The
effects of loser are strongest in Uganda and Zimbabwe among sub-Saharan African
countries and in Chile and Peru of Latin America. In Cape Verde, Costa Rica,
Mexico, and Venezuela, the effects of political minority are negative but are not
statistically significant. In Mozambique, Honduras, and Paraguay, the effects are in
the incorrect direction, but they do not even approach statistical significance.

In Uganda, the significantly large difference in popular satisfaction with
democracy between winners and losers is not necessarily surprising given the fact
that the ruling National Resistance Movement (NRM) established the present regime
and has ruled the country for the past twenty years. Political parties are permitted to
exist but are forbidden from electoral activities. While Uganda’s referendum 2000
approved that Ugandans gave credit to President Museveni and the movement
system for their achievements in government, opposition parties have been
criticizing the system and advocating multiparty electoral campaigns.

There is an obvious regional cleavage between the north and the three
southern regions in Uganda (Logan, Muwanga, Sentamu, and Bratton 2003). In
addition, people’s party affiliation is strongly associated with region. According to
the Afrobarometer Round 2 data set collected in Uganda, 24 percent of the citizens
whose party is in opposition are from the north, while 58 percent of ruling NRM
party supporters are from the west, President Museveni’s home area. In the 2001
relatively contested presidential elections, President Museveni from the NRM was

selected to a second five-year term with 69.3 percent of the total vote, compared to

67

only 27.8 percent for Besigye. Logan et al. (2003) show that the differences between
winners and losers - which authors describe as “insiders” and “outsiders” with
reference to the patronage system -- emerge sharply on issues that are fundamental
to the practice and consolidation of democracy, for example, evaluations of the
nature and performance of the political system as a whole. This evidence confirms
my results indicating that there is a large gap in popular satisfaction with democracy
between winners and losers in Uganda.

Cape Verde is one of few African countries that have experienced a renewed
change of government after successful founding elections. In January and February
2001, Cape Verde held its third parliamentary and presidential elections since its
democratic transition in 1990. The result was that the African Party for the
Independence of Cape Verde (PAICV), which had been ousted from office by the
Movement for Democracy (MPD) in 1991, regained a majority in the National
Assembly and saw its candidate win the presidency (Meyns 2002). The National
Electoral Committee confirmed the victory of Pedro Pires as the new President by
. just 12 votes with 49.43 percent against the 49.42 percent of his opponent, Carlos
Veiga (MPD).

Cape Verde is moving toward a two-party system, with the PAICV
occupying the space left of center and the MPD to the right of center (Meyns 2002).
The MPD has been pursuing policies of reducing state involvement in the national
economy and encouraging greater privatization. The party enjoys the support of
business interests, including those from Portugal, as well as the backing of the
influential Catholic hierarchy. The PAICV, on the other hand, has been criticizing
the MPD’s neoliberalism and established itself as a social democratic party. The tiny

margin of victory in the presidential election result and the alternation of

68

government would predict that the losers’ effect on popular satisfaction with
democracy should not be significant in Cape Verde. The empirical results show that
there is no significant gap in satisfaction with democracy between supporters of
ruling and opposition parties.

During the 19905 Venezuela had experienced big changes in party system
(Penfold-Becerra 2004), due to the rapid decline in the dominance of Democratic
Action (AD) and the Organized Political Committee for Independent Election
(COPEI), coupled with an increasing fragmentation of the party system as a whole.
The effective number of parties more than doubled from 2.24 to 4.8. In December
1993, a coalition of Movimiento al Socialismo (MAS), Convergencia Nacional (CN)
and other small parties elected former president and COPEI founder Rafael Caldera
to the presidency. Although Caldera’s decrees, which included substantial public
intervention in the financial system, temporarily increased his popularity in 1994,
the failure of the govemment’s reforms to improve living standards led to
widespread attacks on the president, Congress, and the party system after 1995
(Romero 1997). In December 1998, the ex-coup plotter Hugo Chaves from the
Movimiento V Republica (MVR) was elected president with 56.2 percent of valid
votes, the largest margin in Venezuelan history. Because the Latinobarometer survey
took place during Venezuela’s transitional period, we expect no meaningful
difference between political majority and minority. The results in Table 4-1 clearly
show that partisan status has no significant effect on popular satisfaction with
democracy in Venezuela in 1998.

Economic performance evaluations also play an important role in satisfaction
with democracy. The coefficients for national economic performance evaluation are

consistently significant and are in the expected direction in most countries (31 out of

69

34 countries). The results show that citizens who evaluate their nation’s economic
performance positively also are more likely to show higher levels of satisfaction
with democracy. In Ghana, Lesotho, and Guatemala, the effects of respondents’
assessments of national economic conditions on satisfaction with democracy are not
statistically significant. In addition, the effects of egocentric assessments are
statistically significant in two-thirds (21) of the countries under study. Some
countries, i.e., Nigeria, Zimbabwe, Argentina, and Chile, show that the effect of
sociotropic assessments is greater than egocentric ones, while others show that the
latter is much stronger than the former, for example, South Africa, Kenya, and Costa
Rica. The strongest sociotropic effects are in El Salvador, Brazil, Nicaragua, and
Nigeria. Egocentric evaluations are strongest in Nicaragua, Venezuela, Mexico, and
Panama. In both Lesotho and Guatemala, neither of the two economic performance
evaluations have any significant effect on popular satisfaction with the way
democracy works.

Political interest was a statistically significant and positive predictor of

satisfaction with the way democracy works in only 17 of 33 countries.15 These
results suggest that most politically informed and involved citizens develop an
allegiance to democratic performance, thus challenging the thesis that cognitive
competence breeds dissatisfaction with political system. In Mozambique, however,
people who are more interested in politics show lower levels of satisfaction with
democracy.

Demographic factors had no consistent effects in countries under this study
on popular satisfaction with democracy. An overwhelming majority of the

coefficients are nowhere close to being statistically meaningful. In addition, the

 

15 Taiwan’s Election and Democratization Study (T EDS) 2003 survey questionnaire does not include
the item of interest in politics.

70

signs of coefficients switch frequently from country to country, indicating a lack of
cross-national consistency in the effects of these variables. For example, the results
show that older people are significantly more satisfied with the way democracy

works than younger people in only 14 out of 34 countries.

Table 4-2 Probability of Satisfaction with Democracy by Country

To ease the interpretation of the results, I compute the predicted probabilities
of “very satisfied with democracy” for each country separately in Table 4-2.
Predicted probabilities are given for the ‘average’ female respondent; all variables
other than loser, non-partisan, and female, are held constant at the mean and the
predicted probabilities calculated for both winner and loser. The substantive results
of these analyses are summarized in Table 4-2. The results provide strong additional
support for the hypothesis of partisan status. Changing from being a loser to a
winner has a substantively meaningful effect on the likelihood of an individual being
“very” satisfied with the way democracy works in 31 out of the 34 countries under
study. Across these 31 countries, the average increase in the likelihood of being
“very” satisfied with democracy as one moves from being a loser to a winner is 8.3
percent. In some countries, the increase is significantly larger: in Ghana, Malawi,
Tanzania, and Uganda, winners are about 20 percentage points more likely to be
satisfied with democracy than losers. We can see the marginal increase of the
probability of satisfaction with democracy as one moves from a loser to a winner in
Mexico (0.3 percentage point), Brazil (0.7 percentage point), Panama (1.5
percentage point) and Cape Verde (1.8 percentage point). In short, the results
provide significantly strong support for my hypotheses’ predictions, even controlling

for demographic variables. While the differences in popular satisfaction with

71

democracy between winner and loser are various across countries, citizens whose
party is in government are likely to show higher levels of satisfaction with

democracy than those whose party is in opposition.

72

 

 

€83 I
$353 I

 

 

 

 

 

 

 

 

magnum :85an ER mow—8:8: 5T: gauflmuam fiv 3:“:

73

 

 

£83 I
$2.53 I

 

 

 

 

 

 

 

 

 

2.23m Simian ER 5.29255 :33 nouoflmzam €255.58 fiv PEER

74

 

 

£83 I
32:55 I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

253m SE25— UEW mangoes—on 5:» 553.35% €o::=:o3 “.v 9.53m

75

488.
689
m8. :-

88.9
5.5—. P.01

889
288.8

88.9
.838

88.9
:83

:89
28:8

889
85.8

88.9
:53.

:89
I! Fmo. PI

mucmm:

88.8-
38.9
~88-

689
«SN? P .o:

88.9
:88

889

$98

83.9
88.8

:89
2.48.8

88.9
:88

889
com PN.C|

a: :9
.5.m55.oi

NEGNCNP

88.8
89.9
88.

88.9
48.8-

:89
~88-

88.9
$8.8

$89
:88

88.9
.888

88.9
~88

98.9
:88-

33.9
.«me.or

8:? 58m

84. F
88.9
88

88.9
echoAun

88.9
:88

:39
88.8

98.9
:83

88.9
:488

88.9
.888

:89
.888-

88.9
..8:.o-

$592

84.8
88.9
88.8-

:89
:m : 8

88.9
.88

:89
~58

88.9
:88

68.9
.888

68.9
:8 F :8

63.9
:89?

:89
chNVAur

5582

Bad.
68.9
88.8-

68.9
:3 to-

:89
~88

889
2.48.8.

:89
888

68.9
.828

:89
.88

68.9
.888-

88.9
: :88-

558.2

~28
38.9
88.8-

:89
88.8

88.9
.88-

:89
88.8-

:39
88.8

68.9
:88

:89
~88

689
how Fm.o|

:89
aum0®.oi

050mm...

83.
88.9
~88-

$8.9
~88

88.9
88.8

88.9
88.8-

83.9
:8:.o

:89
~88

:89
88.8

88.9
«omfifigun

a: :9
o¢©@©.°r

9.sz

89¢
:89
~88-

68.9
~58-

88.9
.88-

88.9
Rod

88.9
88.8

88.9
:88

68.9
88.8

88.9
acmwm.0r

:89
camNMAUi

mcmzmaom

Q 2285:

E 288::

Cosmo-6m

om<

o_mEom

oocmctorma
o_Eocooo .mcoflon.

oocmEOtoa
2:558 .98sz

mason c_ 692:.

cmemaéoz

ammo;
833...;
33.536:—
32:60

 

35.500 3 5 580058 5.3 coco—333m co ao_nu_._u> .25..-.32305 .0 3025 _..v 035...

 

76

 

.3. v a n 3.. .2. v a n * 8885823 5 88 80:8 33:85 88388.8. 858m 88020 8.82
9.8 88 88 8- ~88: 88: 88 8» ~88 z
888 888 888 888 ~88 888 888 3.88 :88 m 88:88
82.9 88.9 82.9 :89 :89 :89 88.9 88.9 :89
88.: ~8~.: 88.: R8 88.: ~88 48.: 88.: :8: :9 29.8.5
:89 :89 :89 $8.9 38.9 :89 88.9 88.9 88.9

77

 

888 888 888 :88 m :88- 88.: ~88 8-8- 9 8.882:
889 8:9 889 :89 88:9 889 38.9 88.9
88- 88. 888- 88. :- 8:8. :- 8:8 888- 8:. :- :9 8.885:
68.9 88.9 88.9 38.9 :89 68.9 :89 :89
:~88- 888 :88- :88- :88. 888 888- :8:8- 8888
88.9 :89 88.9 88.9 88.9 88.9 88.9 88.9
:88- :888 ~88 :88 .488 .888 888 ~88 8<
:89 38.9 88.9 :89 :89 88.9 88.9 :89
888 :88- :88- :k8- 888- 888- 8:88- :88- 8.858“.
:89 88.9 68.9 88.9 88.9 88.9 38.9 38.9
388 .888 888 888 .838 :88- :88 :88 88.588
250.508 atom-En—
889 88.9 88.9 :89 88.9 88.9 88.9 :89
.888 :88 :88 :8 :8 .188 2.2.8 :38 :28 88828
259.000 .950sz
88.9 88.9 88.9 88.9 :89 88.9 :89 88.9
: :8 .888 :88 :888- .888 :888 88 888- 88.8 s 8.2:.
88.9 88.9 88.9 :89 88.9 88.9 88.9 :89
: :88- :8 8- .888. :88- .888- 88- :88- :88- 88:88.82
88.9 889 8:9 :59 88.9 88.9 :8 :.9 88.9
.8 8- :8 :8- :888- 888 .8: 8- 888- :88. :- : :88- 88..
835:5»
Eve—592::—
w_>__om 2sz98: .mdfiocom 3998835. .983. oEo> oqmo anmnEfi mBEmN bfiaoo

 

35.500 3 E 8202.50 525 cozoflazom ca «030...; 35.7.3220... :0 33:5 32.52.03 :6 03¢...

78

 

.3. v m u in .2. v a n * 8805:0qu E 28 88.8 982.8% 8838.5: 858n— 82090 .632
we? womm owm Nvm mm 5. X: F NR mmop Z
mmod owe-o vmod :56 one-o awe-o mmod mac-o m 25me
$888 38.9 €888 38.3 888 A-~8V 68.2 $888
omoé o 5.? own-F hm fir who... 5 3d man-F 5...? 8v 20:82.:-
889 888 888 88.9 $8.3 $8.3 $8.8 88.3

79

 

8-.~ F P~. P ~88- 88.~ ~8 I 8; ~88 ~ 58 ~88 8 8.852,:
629 829 A-~9 89 :89 689 A-~9 889 $89
m~m8 5.8- ~88- 888 88- 88 88- 88- 88 3 888.5
689 $89 :89 :89 889 889 689 889 $89
:888- 888 :28- 888- 888 88- ..~88 ~88- ~88 8:888
A~89 889 $89 889 A~89 $89 A~89 A~89 $89
88 :888 88 ~88 .888 :888 £88 88- :88.8 8<
£89 389 :89 $89 689 £89 889 €89 A~89
888- ..-8 .~88- 888- ~88 :88 :88 888 888- 286“.
A-89 $89 :89 889 $89 :89 :89 :89 889
:88 88 ~88- .~88 ~88 :88 :88 :888 :88 88828
26958 858?.
889 $89 889 889 689 £89 :89 889 689
:88 :88 888- :888 :88 :88 :888 :38 :88 85.88
£60.58 .80sz
:89 889 $89 689 689 889 689 689 889
:888 .88 :8 8 :88 :-8 :88 :888 :88 :-8 8:8 s 8.2:.
689 889 $89 $89 889 889 889 689 :89
~88- :88- :38- :88- ..88- : 88- E88- : 3.8- 88- 888-82
889 8.9 :29 389 889 :89 $89 :89 :89
~88- 88 :88- :88- :28- ..-~8- 888- .888- :88- .88..
8833...;
«51:095..—
02me 8.59.0: «.mEmumao 588280. 58me 220 moi Smoo 85:80.00 :85 89550
E

 

«0.5550 en 5 39.00200 5.! 2030.83.83 :0 «038:; 35.7.3232... .0 8035 €352.63 fie 03a...

80

 

.8. V a n .2. .2. v a u * 8885883 5 88 888 98:85 83:85.. 850.5 88080 .882
8- 89 88 82 8-~ 8-~ ~8~ ~8- 88 z
888 888 88 ~38 88 ~88 888 888 88 m 8888
£89 6:9 859 E~9 €29 €29 6-9 €89 8.9
8~.~ ~€~.~ 888 88 8 8 88 88.. 8 *8 ~88 A9 8.882fi
8:9 €-9 $89 889 888 8.9 a-9 889 8:9

81

 

E8 888- «88; 8888 888 53 mm: «31 Q 2289,:
$29 6&9 2889 :29 589 A889 389 $889
883- 58.7 8888 8858- 8888- 888.8 8888 v88 E 228::
A839 639 $889 $39 6889 639 A389 $889
:8888. $88- :888- :888- 838- 838- 2,838. $88. 8:838
8889 8889 A889 :89 3889 6889 A889 @889
:858 .388 :388 :888 8888 588 :888 ~88 8?
:89 8889 639 639 :89 $889 :89 8889
858- «888 8888 ::.~8 388 :mt8- :88. 888.9 8.858”.
$889 $889 8889 3889 A839 639 839 $89
:828 «88 :828 «38 :828 :898 :828 :«88 8:858:88
359.com .958me
839 8889 8889 3889 $889 6889 #889 E89
:828 :828 2.8888 :Em8 :83 :888 :38 2.8888 888E858
o_Eocoom .95sz
6889 $889 $89 8889 $889 589 8889
8888- 858- :«98 .888- :38 2.88.8 8888 85.8 c_ .8925
889 :89 38.9 $889 839 339 A889 :89
§¢mNN.OI ¢tmmm.ou FOCAVI ti FN F.Cu ¢¢8V.OI aim? F.C limo F.°| timcm.o: CNwZLonP—OZ
8889 $889 829 8889 $889 6889 A889 329
ii§.03 ifiommfivu rm P.0I «imNV.ol «5005.0: on —..o VQO.°| famQQHVu LQwCJ
83825»
8:86.595:—
cmém... 893. .w m_m=~wcm> >398: Ema. 230798.... 8.88ch mamfimgz B2500

 

8055.00 3 5 380088 5.? 2030523....” :0 338...; _u>o.._._u:u_>_u:_ .0 300.5 30:58:03 he 03:

82

 

.3. v a u i. .3. v a n * 8885883 E 88 88.8 88985 flax—88¢. 838m @8280 .832
83 82 8- 8- 88— 8: 8:: 8: z
88 «~88 888 ~88 888 888 888 888 m 8:88
389 E~9 $89 :29 8&9 ~89 c-9 8&9
8~.~ 53 828 85.? $8; ~8~ E8 888 E u_o;mm.fi
:29 $89 A~8~9 $89 859 889 E~9 889

83

8885883 E 82.8 8.888% .882

 

 

 

 

 

:89 859 689 889 689 859 $89 659 8289

888 888 888 88 88 88 88 88 ~88 88..
889 8189 A289 889 A289 8289 A289 889 859

8 58 88 58 ~88 88 88 E8 88 88 2852,
:85 m_>__om mczcmmi .mmocow 03858835. 883. 88> amo anmnEfi mBEmN #5550
£89 889 a 89 889 889 8:89 A-89 889 a 289

88 ~88 88 888 88 88 38 ~28 8:8 288..
8 S9 A-89 a 89 :89 :89 :89 889 :89 A-89

v88 88 88 ~88 888 88 888 888 88 2852,
macaw: 9:888... moE< 538 £592 39:82 {$8.82 05088.. 8:20 mewsflom ”.5550

 

$560 3 58858 53 5.6888 3 83828 ~8 can»

84

8885883 5 88.8 888% .882

 

 

 

 

 

889 €89 889 889 :89 889 859 €89

88 ~38 82 8 88 -88 888 888 888 88..
€59 889 889 859 389 889 :59 :89

~88 88 2.8 888 888 888 888 88 2852,
8588. 893. :58 8.82.883 8:985 28.... 8:38.“ mEmcmm 858.822 .#5550
$89 €89 €289 889 889 889 €~89 €89

888 88 88 88 €88 858 888 888 288..
$89 €289 €289 889 859 €89 389 8:89

88 88 88 88 888 888 288 B8 2853
02x22 88:680.... 8.8.8.2830 8823 m. 5828. 8.30 SE 8800 856200 25550

 

E88 3 328.58 5? 5:88:88 3 28382.. 62.5an ~.e 28.2

85

Chapter 5. Multilevel Analyses

5.] Political Institutions and Satisfaction with Democracy

As seen in chapter 4, individual-level results show clear support of my
hypotheses. Partisan status has a significant effect on popular satisfaction with
democracy. Both losers and non-partisans show lower levels of satisfaction with
democracy than winners. What remains to be tested, however, is whether political
institutions uniformly and directly affect popular satisfaction with democracy or
whether they mediate the effect of partisan status on popular satisfaction with
democracy. For instance, do party systems or electoral systems help explain why
citizens whose party is in opposition in Zimbabwe, for example, are about three times as
likely (b = -l .073) than those in Brazil (b = -0.364) to be dissatisfied with the way
democracy works?

This study combines variables of political institution with individual-level data
in order to model their joint effect on popular satisfaction with democracy. Specifically,
the political institutions used are ENPP for party systems, the index of
disproportionality for electoral systems, presidentialism, and unitarism. I also include
both the Freedom House index and a nation’s GDP per capita, because prior research
argues that those two factors have a strong influence on popular trust in political
institutions (Norris 1999).

The data in this research have a multilevel structure in which respondents are
nested within countries. Because the model merges data from two different levels, OLS
creates a number of statistical problems, such as nonconstant variance and clustering.

Failure to recognize the hierarchical nature of the data can lead to underestimating the

86

standard errors for country-level variables (Jones and Steenbergen 2002). This research
therefore estimates the model using a software developed by Raudenbush and Bryk
(2002) for multilevel models (HLM version 5).

Several specific issues may pose problems for inference because of the
multilevel nature of the data. First, the intercepts may vary across countries; failure to
control for this may result in biased estimates. Specifically, if intercepts are variable, we
may be overestimating the effect of political institutions on popular satisfaction with
democracy. A second concern is that individual-level variables may have unequal
slopes across countries. In this case, a pooled estimator may be biased for each
particular country. A third concern relates to the robustness of the inferences based on
potentially inefficient standard errors resulting from potential clustering (Zorn 2001).
To deal with these issues, multilevel modeling techniques allow for estimating varying
intercepts and slopes, produce asymptotically efficient standard errors, and provide for a
direct estimation of variance components at each level of the model.

Below, I show the coefficients of interest (constants and independent variables),
as well as the variance components at each level of the data (individual- and country-

level).

5.2 Analysis of Variance
To determine, first, whether there is significant variation in popular satisfaction
with the way democracy works at the individual and country levels, I estimated an

AN OVA model that decomposes the variance in the dependent variable, where

87

Satisfactionij = ,300 + 50,- + 8:7 . (1)

In this model, flOO is the overall mean of popular satisfaction with democracy (i.e., the

weighted mean across individuals and countries). The sources of cross-national

variation, which cause a particular country to deviate from the grand mean, are

contained in 60]. . Finally, Eij captures inter-individual differences. The variances of

these different sources of variation are given by TOO and 0' 2, respectively.

Table 5-1 One-way ANOVA Model

Table 5-1 shows the Maximum Likelihood estimates of the overall mean and the
variance components. The country-level variance component is statistically significant,
suggesting that there is significant variance in popular satisfaction with democracy
between countries. This is evidence that the multilevel character of the data should not
be ignored.

In order to obtain a better sense of the importance of the two different levels of
analysis, Raudenbush and Bryk (2002) suggest researchers to calculate the ratio of each

variance component to the total variance in popular satisfaction with democracy, which

is equal to (700 + 0' 2). For example, the ratio of TOO over the total variance indicates

the importance of the country-level of analysis. Results of the ANOVA model show that
country—level variance is proportionally much smaller than individual-level variance.

About 12% of the variance in popular satisfaction with the way democracy works is

88

between countries.16 Given that the data are measured at the individual level, this is not
surprising. The ratio also means the estimates of the intra-cluster at the country-level
(Raudenbush and Bryk 2002). Thus, at the country level the results show an intra-
cluster correlation of .115. This is a sizable positive correlation that suggesting a fair
amount of clustering at the country-level. Clearly, to ignore these sources of variance is
to miss out on important aspects of popular satisfaction with democracy. This could
result in both incorrect statistical inference and improper substantive conclusions about

the satisfaction with democracy (Steenbergen and Jones 2002).

5.3 Multilevel Model of Satisfaction with Democracy
The results of the ANOVA model indicate clearly that there is significant
variation in popular satisfaction with democracy at the country-level analysis. How do

we account for popular satisfaction with the way democracy works in the country?

Multilevel analysis requires the specification of following equations:

SdtleClCthflij = 610]. + aleoserij + aszopartyl-j + a3jlntepolij

+a'4jNatecoij + aSjPerecoij + anFemaleU

+a7 jAgeij + angducij + 8,}. (1)

 

16 11.5 = 0.099/(0.099+0.760) This calculation is based on the ratio of country-level variance component
relative to the total variance component in satisfaction with democracy (Raudenbush and Bryk 2002).

89

(10]. = ,600 + flmENPPj + flOZDisproportionalityj + ,603Presidentialismj

+,605F HI 1 + ,BO6GDPpcj + ,607Africaj + flogAsiaj + 60]. (2)
a1]. = 1610 +1611ENPPJ' +,612Dispr0p0rtionalityj + 51]. (3)
a2]. = ,620 + flZIENPPj +fl21Dispr0porti0nalityj + 52]. (4)

Satisfactionij = 1600 + flmENPPj + ,BOZDisproportionalityj
+flO3Presidentialismj + ,BMUnitarismj + flOSFHIj + flmGDPPCj
+1607 Africa j + flog/‘31.“; + fllOLoserb. + flzoNOPartyl-j + fl301ntpollj
+fi40Natecol-j + flSOPerecq-j + £60]: emaleij + flmAgeij + ,680Educij
+ :61 IENPPJ. x Loserb. + flZIDisproportionalityj x Losery.
+1621ENPPJ- X N Opartyij + ,BZZDisproportionalityj x Nopartyij

+50]. + 51].L0ser}j + (52].N0partyij + 81-]- (5)

The first equation models the influence of individual-level predictors on popular
satisfaction with the way democracy works. The j subscript for the loser coefficient
indicates that it is expected to vary across the j nations. Equation (2) models the
influence of ENPP, electoral system disproportionality, presidentialism, unitarism,
democratization, and GDP per capita on satisfaction with democracy. In order to control,
equation (2) also includes two dummy variables for Africa and Asia. In HLM, this

effect is modeled through the influence of macro-level factors on the intercepts from

90

equation (1). Equation (3) models that the effect of loser, which is given by all. in

equation (1), varies as a function of ENPP and Disproportionality. I do not expect the

two political institutions to explain all of the cross-country variation in 0:11.. Equation
(4) models that the effect of non-partisan, which is given by dz]. in equation (1), varies

as a function of ENPP and Disproportionality. My model also does not expect the two

variables to explain all of the cross-country variation in (£21.. The interactions in

equation (5) are captured by estimating the effect of political institutions such as ENPP
and Disproportionality on both loser (people in political minority) and non-partisan
coefficients from equation (1). They show the mediating effects of political institutions
on the relationship between partisan status and popular satisfaction with democracy. I
expect two interaction terms with ENPP to have a positive effect and the other two
interaction terms with Disproportionality to have a negative sign. Note that equations
(2), (3), and (4) contain an error term; unlike OLS, multilevel estimation techniques do
not assume that a contextual variable completely determines the coefficient from the

individual-level variables(J ones and Steenbergen 2002). Thus, the model in equation (5)

implies variance components 0' 2 for 81.1., 700 for 50]., 2'“ for 51]. , and 2'22 for (52]..

Table 5-2 Individual and Country-level Predictors of Satisfaction with Democracy

Table 5-2 shows the results of two models estimating the direct effects of
political institutions and partisan status (minority or non-partisan) on popular

satisfaction with the way democracy works. The second column in Table 5-2 gives the

91

ML estimates of the fixed effects and the variance components of this multilevel model.

This model is a significant improvement over the one-way AN OVA model:

12 = 3747.584, df = 16, p < .05, suggesting that at least a subset of the predictors has

effects that are different from 0.

At the individual-level, the results show a strong effect of both loser and non-
partisan on popular satisfaction with democracy. People in political minority
significantly express lower levels of satisfaction with democracy than those in political
majority (winner). The overall performance of the individual-level predictors is
significantly powerful except for the effect of female on satisfaction with democracy.
Economic performance evaluations play an important role in satisfaction with
democracy. The coefficients for both sociotropic and egocentric evaluations of the
economy are statistically significant and are in the expected direction. Citizens who
evaluate their national and personal economic performances positively are likely to
express higher levels of satisfaction with democracy. Interest in politics is a statistically
significant and positive predictor of satisfaction with the way democracy works. Of the
control variables, the coefficient of age is statistically significant and positive; education
has a significant and negative effect on popular satisfaction with democracy.

As a group, the individual-level predictors explain small parts of the individual
level variance in popular satisfaction with democracy. It is possible to assess this
explanatory power by comparing the individual-level variance component in Table 5-1
to the same component in Table 5-2 (see Bryk and Raudenbush 2002). The difference in
these variance components is 0.760 — 0.7077 = 0.0523. Relative to the size of the

AN OVA variance component, this is a reduction of 0.0523 / 0.760 = 0.069. Thus, the

92

individual-level variance components explain about 7 percent of the individual-level
variance in popular satisfaction with democracy.

At the national level, results show that political institutions have a strong power
to explain popular satisfaction with democracy. Party system measured by ENPP has a
significant and positive effect on satisfaction. The greater ENPP is likely to increase the
level of popular satisfaction with democracy. Citizens in countries with a greater ENPP
are more satisfied with democracy than citizens in countries with a smaller ENPP. I
interpret this to mean that multi-party systems increase popular satisfaction with
democracy because these systems offer more alternatives for citizens to choose.

In emerging democracies, political parties are one of the most important and
influential organizations linking people’s interests to decision-making processes. Due to
the democratic transitions, people have been able to vote for a party closer to their needs
in different aspects, for example, economy, ethnicity, region, and so on. For citizens in
new democracies, it is very critical to evaluate their political system performance
whether the system allows a party representing their interests to run for elections and to
win seats in a parliament.

The coefficient of presidentialism is statistically significant at the level of .1 and
positive in the model. People who live in presidential systems are likely to express
higher levels of satisfaction with democracy than those who live in parliamentary
systems. This result is not consistent with Norris’ finding that institutional confidence is
“slightly higher in parliamentary than in presidential systems” (1999: 232). Presidential
systems have more direct accountability to voters than parliamentary systems. In

particular, choosing a preferable candidate in presidential elections is still recognized as

93

an achievement of democratic transition among citizens in emerging democracies. They
are not ready to delegate their right to directly vote for a presidential candidate to their
representatives or party leaders. For example, people in Brazil decisively defeated a
1993 referendum on shifting presidentialism to parliamentarism, while many political
scientists have been recommending the latter for political development in new
democracies.

Unitarism is also a significant and positive predictor of popular satisfaction with
democracy. Across emerging democracies, people who live in unitary systems are likely
to show higher levels of satisfaction with democracy than those who live in federal
systems. This is consistent with Norris’ finding that institutional confidence is greater in
countries that have a unitary system than federal one. Among emerging democracies,
many countries are trying to unify the nation and increase the prospects of govemability.
Developing countries have limited financial and human resources to deliver to local
governments. It is easier to maintain patron-client networks in local-level, because local
governments include smaller and more homogeneous groups than national governments,
by definition. In emerging democracies, in turn, federal systems does not seem to be
successful to induce positive attitudes towards political systems performance from
citizens.

However, electoral systems have no significant effect on popular satisfaction
with democracy. In emerging democracies, the degree of the proportionality in the
translation of votes into seats in assemblies does not have any impact on popular
evaluations of political systems performance. This contradicts both Norris’s (1999)

finding that electoral systems directly affect institutional confidence and Farrell and

94

McAllister’s (2003) finding that electoral systems have a direct effect on satisfaction
with democracy. While existing two studies on influence of electoral systems on
citizens’ attitude towards political systems focus more on established democracies than
on new democracies, this project is closely looking at the latter. Citizens in emerging
democracies have been shortly experienced multi-party elections after their democratic
transitions. For them, whether those elections have been held freely and fairly is critical
to evaluate new political systems in transition.

The effect of Freedom House Index is statistically significant and is in expected
direction (Norris 1999). Citizens who live in countries where there are higher degrees of
civil liberties and human rights show higher levels of satisfaction with democracy.

Finally, the dummy variable for sub-Saharan Africa has a powerful effect on
satisfaction with democracy. The result indicates that people who live in sub-Saharan
Africa are likely to show higher levels of satisfaction with democracy than those who
live in Latin America. There is no significant difference in the satisfaction between
people in Eastern Asia and people in Latin America. These regional dummies
essentially are a measure of our ignorance. They do not specify what it is about being a
sub-Saharan African country or an Asian country that increase popular satisfaction with
democracy. The dummy variable of sub-Saharan Africa just identify such effects,
indicating that important variables remain that my model has not captured, which
contribute to differentiating levels of popular satisfaction with democracy from those in
Latin America.

The differences in levels of popular satisfaction with democracy between sub-

Saharan African and Latin America can be explained by how long citizens have been

95

experienced the new political systems after their democratic transition. As I mentioned
earlier, popular satisfaction with democracy is a volatile attitude. Comparing with Latin
American country, on average, the majority of sun-Saharan African countries have
relatively short experiences of multiparty elections after’democratic transition. In a few
of them, for example, Kenya and Senegal, alternations in government by multiparty
elections strongly increase positive evaluations of political systems performance among

citizens. National level predictors as a set account for about 60 percent of the cross-

country variance in popular satisfaction with democracy.1

Importantly, the direct effect of electoral system is insignificant at the .05 level.
This is inconsistent with existing studies (Norris 1999; Farrell and McAllister 2003;
Anderson et al. 2005). Those works use OLS estimation. It is useful at this point to
consider the implications of ignoring the multilevel data structure. Specifically, what
inferences would we draw if we were to ignore the multilevel character of the
satisfaction with democracy data? To answer this question I estimate an OLS regression
model on the data. The last column in Table 5-2 presents the results from this analysis.
With the exception of the variable of female, all predictors are significant. The
differences are particularly significant at the national-level of analysis. All of the
national-level predictors are statistically significant at least at the .1 level in the OLS
regression analysis, while only five predictors out of eight are at the .1 level in the
multilevel analysis. In the OLS model, the disproportionality is statistically significant
at the .1 level and is a positive predictor of popular satisfaction with democracy. This

result is consistent with Norris’ (1999) finding with OLS estimation that majoritarian

 

l . . . .
7 The relative change in the national-level variance component is (.097 - .039) / .097 = .598.

96

electoral systems produce higher levels of institutional confidence than proportional
ones. Also note that GDP per capita does not pass the conventional level of statistical
significance in the multilevel analysis. This result is consistent with a few works using
multilevel analysis (Steenbergen and Jones 2002; Rohrschneider 2002; Anderson and
Tverdova 2003). Clearly, the inferences drawn from the regression analysis are different
than those drawn from the multilevel analysis. These patterns indicate that previous
analyses, which do not account for the multilevel nature of such data, overestimate the
impact of these country-level characteristics on popular satisfaction with democracy.
Steenbergen and Jones (2002) point out that these differences arise because the OLS
analysis ignores the clustering of the data assuming that those individual-level
observations are independent in the multilevel data set.

Now I closely look at the effect of partisan status, both loser and non-partisan,
by accounting for possible “causal heterogeneity” (Western 1995; Steenbergen and
Jones 2002). Up to this point I have focused on the simple additive effect of political
institutions on popular satisfaction with democracy. I also hypothesized (H6 and 7),
however, that the negative relationship between popular satisfaction with democracy
and partisan status (loser or non-partisan) should be moderated by multiparty systems
and proportional electoral systems. In short, an individuals’ partisan status and his or
her institutional context should interact in shaping his or her satisfaction with
democracy. By specifying cross-level interactions in Equation (5), it is possible to
determine whether the lower-level predictors’ effect on the dependent variable of

popular satisfaction with democracy is conditioned or moderated by hi gher-level

97

predictors. In other words, is there a single, uniform relationship or does the relationship
vary across higher levels of analysis?

In Table 5-2, both loser and non-partisan have a fixed effect which is the same
analysis as Norris (1999) has. However, my hypothesis suggests that the effect of
partisan status varies across nations as a function of party or electoral system. Does
ENPP or the disproportionality in changing votes into seats help explain the influence
of political minority status (loser or non-partisan) on cross-national variations in

popular satisfaction with the way democracy works?

Table 5-3 Model with Cross-Level Interaction

Table 5-3 provides the ML estimates of the multilevel model in Equation (5).
The results are in part consistent with my predictions. Party systems significantly
mediate the relationship between partisan status and popular satisfaction with
democracy. The cross-level interaction between ENPP and loser is positive and
statistically significant at the .05 level. The positive sign denotes that the negative effect
of loser on popular satisfaction with democracy is moderated by greater ENPP. The
result indicates that citizens in political minority are likely to express higher levels of
satisfaction with the way democracy works when they live in a country with greater
ENPP than smaller ENPP. Another cross-level interaction between ENPP and non-
partisan is also positive and statistically significant. It demonstrates that while citizens
do not support any existing political party, they are likely to show higher levels of

satisfaction with democracy when they have political system with greater ENPP than

98

smaller EN PP. I findthat greater ENPP alleviates the negative effects of belonging to a
political minority on popular satisfaction with the way democracy works.

However, electoral systems have no significant mediating effect on the
relationship between partisan status and popular satisfaction with democracy. The two
interaction terms with the disproportionality are not statistically significant. The results
are not consistent with Anderson and Guillory’s (1997) work. Using eleven developed
countries, they show that losers in systems that are more proportional express higher
levels of satisfaction with democracy than losers do in systems with majoritarian
characteristics.

In emerging democracies, whether elections are fairly and freely conducted is
still a prominent issue among political elites as well as ordinary people, while citizens in
developed countries are concerned less about free and fair elections than about how
electoral systems count their vote. As Norris (1999) shows, democratization is
significantly associated with people’s confidence in political institutions. Emerging
democracies still need more confidence in electoral systems from their ordinary people.

Note also that the model in Table 5-3 is a significant improvement over the

multilevel analysis in Table 5-2: 12 = 251.004 , df = 4 , p < .05, suggesting that at

least a subset of the predictors has effects that are different from 0.

Figure 5-1 Predicted Probability of Popular Satisfaction with Democracy by Partisan
status and ENPP

99

In order to estimate the magnitudes of the interactive effect between partisan
status and party system, I employ the Table 5-3 multilevel model to examine the
consequences of variation in both partisan status and ENPP variables while holding
other independent variables constant. In Figure 5-1 I examine the effect of ENPP on the
likelihood of popular satisfaction with democracy. Figure 5-1 clearly shows a mediating
effect of party system between partisan status and popular satisfaction with democracy.
Greater EN PP narrows the gap in popular satisfaction with democracy both between
winners and losers and between winners and non-partisans.

As the ENPP increases, the largest party’s seat share in the assembly decreases.
The ruling party needs coalition with other parties to pass bills with a majority in the
assembly. Thus, opposition parties can play more significant roles on decision-making
in political system with greater ENPP. That is, it may be the case that greater number of
people feel that their voice will be heard by the government. The political systems with
greater ENPP produce generally fewer losers. Losers who live under the party systems
with greater ENPP are likely to be more satisfied with the policy position of the
government and, in turn, to express less negative evaluations on political systems
performance.

In emerging democracies, the results show, it is critical for citizens to evaluate
their political systems performance whether political systems have a party closely
representing a given subgroup’s interests or not and whether the party has a significant
influence on government decision-makings. Citizens had a limited mechanism linking
themselves with the state under authoritarian governments (or one party system). They

were not able to choose a governing party, while some authoritarian governments held

100

formal elections. Those elections had worked as a way for authoritarian governments to
demonstrate their legitimacy. The democratic transition has introduced multiparty
elections to political elites and ordinary people. Those multiparty elections offer many
Opportunities both for elites to organize a political group, i.e., political party, and for
ordinary people to express their interests and ideas by elections. Because of multiparty
elections, people in political minority are able to have a vice in the government political
and economic performance. Citizens in new democracies are still pleased with the new
regime as a product of democratic transition. When they evaluate current political
systems performance, people seem not to consider long-term effects of political systems.
For example, an extremely fractionalized party system could increase the probability of
regime collapse.

In sum, this study finds that some political institutions (e.g., party systems,
presidentialism, and unitarism) have significant effects on popular view about the way
democracy works in fledgling democracies. Greater ENPP increases popular
satisfaction with democracy. Particularly, citizens in countries with greater ENPP are
likely to show higher levels of satisfaction with democracy than those in countries with
smaller ENPP or a dominant party. Presidentialism and unitarism also have a positive
impact on popular evaluation of political system performance. In addition, party
systems mediate the relationship between citizens’ partisan status and their satisfaction
with democracy. Greater ENPP is likely to alleviate the negative attitudes of people in
political minority toward political system and reduce the difference in popular
satisfaction with democracy both between winners and losers and between winners and

non-partisans.

101

Table 5-1 . One-way ANOVA Model

 

 

 

 

Fixed Effect Coefficient s.e. P-value

Average country mean (floo) 2.392 0.044 0.000
Variance

Random Effect Component df Chi-square P-value

Country mean (Too) 0.097 31 6877.599 0.000

Individual effect (0'2) 0.76

- 2 Log x Likelihood 132462.073

 

102

Table 5-2. Individual and National-Level Predictors of Satisfaction with

 

 

Democracy
Multilevel Estimate OLS Estimate
Country-level predictors:
ENPP 0.052“ 0.040"
(0.023) (0.003)
Disproportionality 0.003 0.002*
(0.007) (0.001 )
Presidentialism 0.129* 0.121 **
(0.072) (0.010)
Unitarism 0.073" 0.074"
(0.020) (0.003)
Freedom House Index -0.100** -0.094**
(0.038) (0.005)
Log of GDP per capita 0.060 0.057“
(0.049) (0.007)
sub-Saharan Africa 0.342" 0.313"
(0.107) (0.016)
Asia 0.196 0.188"
(0.1 67) (0.023)
Individual-[e vel Predictors:
Loser (minority) -O.260** -0.251 **
(0.011) (0.011)
Non-partisan -0.192** -0.201 **
(0.009) (0.009)
Interest in Politics 0.067" 0.067"
(0.004) (0.004)
National Economic Performance 0.140“ 0.144"
(0.004) (0.004)
Personal Economic Performance 0.078“ 0.082"
(0.005) (0.005)
Female 0.009 0.010
(0.007) (0.008)
Age 0.002“ 0.003"
(0.001) (0.001)

103

Education “0.043”, -0051 H

(0.005) (0.005)
Constant 0.696 0.749"
(0.568) (0.086)
Variance of Random Effects:
Country-level 0.0385
Individual-level 0.7077
- 2 x Log Likelihood 128717.043 131176.688
R2 0.134
F 501.089

 

Note: Multilevel table entries are maximum likelihood estimates with estimated standard errors
in parentheses. Regression table entries are OLS estimates with estimated standard errors in
parentheses. * = p < .10, ** = p < .05.

104

Table 5-3. Model with Cross-Level Interaction

 

Country-level predictors:
ENPP

Disproportionality

Presidentialism

Unitarism

Freedom House Index

Log of GDP per capita

sub-Saharan Africa

Asia

Individual-level Predictors:

Loser (minority)

Non-partisan

Interest in Politics

National Economic Performance

Personal Economic Performance

Female

Age

Multilevel Estimate

105

0.031
(0.024)

-0.002
(0.005)

0.064
(0.039)

0.079**
(0.017)

-0.106**
(0.033)

0.021
(0.039)

0.239"
(0.100)

0137*
(0.074)

-0.420**
(0.095)

-0.300**
(0.059)

0.067"
(0.009)

0.137"
(0.012)

0.078"
(0.006)

0.009
(0.011)

0.002"
(0.001)

Education -0.039**

(0.009)
Constant 1.312"
(0.403)
ENPP x Loser 0.036"
(0.014)
ENPP x Non-partisan 0.024“
(0.008)
Disproportionality x Loser 0.003
(0.006)
Disproportionality x Non-partisan 0.003
(0.003)
Variance of Random Effects:
Country-level 0.0400
Loser 0.0358
Non-partisan 0.014
Individual-level 0.7025
- 2 x Log Likelihood 128466.039

 

Note: Maximum likelihood estimates. Standard errors are in parentheses.
*=p<.10, **=p<.05.

106

 

iomd

Aaeioowaa un uouoegsnes

 

107

 

 

 

280..
mcmwfimaéoz ll .
90:52, 1 - - - reg

 

 

n52m— ccm mafim 835mm 3 \AoEooEoQ 53> cofluoflmcmm 333m mo 559305 3860*“ .Tm oSwE

Chapter 6. Electoral Systems, Partisan status, and Citizens’
Political Support in Lesotho

6.1 Introduction

Do electoral institutions affect citizens’ political support? The previous chapter
suggests that the proportionality of electoral systems has no effect on popular
satisfaction with democracy across emerging democracies. However, existing works
studying the impact of electoral systems on citizens’ attitudes towards political systems
show that there is a significant effect, while some of them use improper estimates to
their data structure. To prove this anomaly, this chapter tries an alternative research
strategy, switching from an analysis with a pooled cross-national data set to an
investigation with a cross-time data set in one country. Studying one country across
time provides researchers an experimental environment to hold many things constant
but subjects to be investigated.

On the basis of survey data from Lesotho collected by Afrobarometer'8 in both
2000 and 2003, this study examines the impacts of electoral system changes on citizens’
political support. I closely look at changes in citizens’ attitudes towards political
systems “prior to” and “after” electoral system changes in Lesotho. The change of
electoral rules in Lesotho in the context of “before” and “after” public opinion surveys
offers a rare opportunity for a natural experiment to test the effect of electoral systems

on citizens’ attitudes. Further, this study also shows that the nature of representation in

 

18 . . . . .
The Afrobarometer consists of research partners 1n 16 African countries, and rs coordinated by the

Institute for Democracy in South Africa, the Centre for Democratic Development in Ghana, and the
Department of Political Science at Michigan State University. See www.afrobarometer.org.

108

electoral systems mediates the relationship between a person’s partisan status as part of
the political minority or majority and his or her orientation to democracy.

Since the third wave of democratization started, there has been a growing
interest in researching the institutional factors underlying citizens’ attitude toward
democratic systems (Anderson and Guillory 1997; Norris 1999; Farrell and McAllister
2003). Of all the institutions that may matter, electoral systems are a central institutional
design issue for a new polity to resolve (Taagepera and Shugart 1989; Horowitz 1991;
Lijphart 1995; Sisk and Reynolds 1998; Reynolds 1999 and 2002; Taagepera 2002).
However, we have just a few works to explain how an electoral system actually makes
an impact on citizens’ attitudes toward political systems (Anderson and Guillory 1997;
Norris 1999; Farrell and McAllister 2003). These works use cross-national survey data
to test the role of proportional representation electoral systems in increasing popular
satisfaction with democracy. Those empirical results are not consistent yet. Some
studies show a positive relationship between proportional representation and popular
satisfaction with democracy (Anderson 1998; Farrell and McAllister 2003), but others
demonstrate less proportional institutions increase public trust in state institutions
(Norris 1999). This research tries to arbitrate these debates.

Africa could not escape from the third wave of democratization (Huntington
1991; Bratton and Van de Walle 1997). Between 1989 and 1996, twenty-nine of the
forty-eight sub-Saharan African countries moved from non-democratic to multiparty
competitive political systems (Reynolds 1999). Lesotho is one of the countries that
experienced a democratic transition in the early 19905 and has since held several

multiparty elections. Lesotho transformed its electoral system from a First-Past-the-Post

109

(FPT P) to Mixed-Member Proportional (MMP) system before the 2002 elections,
becoming the first parliamentary elections to adopt a MMP system in Africa (Elklit
2000). Because of the element of proportional representation of the MMP system,
opposition parties have taken 40 seats in the parliament for the first time since Lesotho
held multiparty elections in 1993.

What evidence is there that electoral systems matter for political support? I find
that people whose party is in opposition are likely to show lower levels of political
support than those whose party is in government. While the nature of the electoral
system does not have any uniform effects on public attitudes, it mediates the
relationship between a person’s status as a part of the political minority (loser) or
majority (winner) and his or her attitudes toward the regime. Increasing the
proportionality of electoral system decreases differences in popular satisfaction with
democracy both between winners and losers and between winners and non-partisans in
Lesotho. Both losers and non-partisans who were disadvantaged by a previous
majoritarian FPT P system are likely to express more positive attitudes toward political
systems after electoral reforms increasing proportionality in the translations of votes
into seats in parliament take place.

The chapter proceeds as follows. First, I begin with an historical description of
Lesotho’s parliamentary elections and the electoral results of both the 1998 and 2002
elections. Second, I review the literature on electoral systems and individual partisan
status- minority or majority- to derive a number of testable hypotheses about the
relationships among these two major factors and individual political support in Lesotho.

I then develop a model of citizen political support that includes electoral systems and

110

individual-level variables, such as partisan status, economic performance evaluations,
and interest in politics. Thereafter, I present empirical tests of the hypotheses utilizing
survey data from Lesotho in both 2000 and 2003 with methods of analysis that are
appropriate to the data. Finally, I offer some conclusions based on the results estimated

here.

6.2 Background

Lesotho gained independence from Great Britain in 1966. Since this date, the

government must be appointed through elections according to constitutional terms.19
The electoral system set out in the 1966 Constitution was a First-Past-the-Post (FPTP),
single member district (SMD) system inherited from the departing colonial power.
Perhaps as a result of this system, opposition parties were unable to take as many
parliamentary seats as votes received in elections. In both 1993 and 1998 parliamentary
elections, they did not gain any seat with significant number of votes but one. The
results of every general election held in the country since independence have been
rejected by the losing parties (Southall and Fox 1999; Makoa 2002; Fox and Southall
2002; Rule 2000; Southall 2003). For example, the Basutoland Congress Party (BCP)
attempted to overturn the results of the pre-independence election of 1965 in the courts.
In addition, the then ruling Basotho National Party (BNP) declined to step down from
power when it lost the election of 1970.

From 1970 to 1986, the BNP led by Chief Leabua Jonathan, maintained its

control of the state with repressive force. The BCP, backed by the apartheid government

 

19 Both the 1966 Independent Constitution and the Constitution of Lesotho which came into effect in
1993.

111

in South Africa, established the Lesotho Liberation Army (LLA) to contest BNP’s
control. The BNP government was ousted in a coup in 1986, and the military ruled until
1993. However, the military was unable to resist both internal and external pressures for
democratization which raised quickly after the start of the transition process in South
Africa in early 1990. The military government allowed the BCP leadership to return
from exile and to participate in the 1993 elections. It led to a dramatic win for the BCP
with taking all 65 seats in the National Assembly, perhaps because Lesotho people
punished the BNP for the years of dictatorship.

In the 1990s after the democratic transition in Lesotho, the opposition parties
still found themselves unable to trust the electoral results, while international and NGO
monitors endorsed those electoral processes as free and fair. Because Lesotho is an
linguistically homogeneous society (apart from small Xhosa- and Ndebele-speaking
minorities, all inhabitants speak Sesotho as a first language) and support for the major
political party consistently spreads across most constituencies, the FPTP system
produced extremely unbalanced election results in both 1993 and 1998. For instance,
the BCP took all 65 parliamentary seats with just 74.7% of the total votes in 1993.
Opposition parties therefore dismissed the BCP’s victory in the election as rigged, and
subsequently backed the king when, in 1994, he dismissed the new government
(Weisfelder 1999; Rule 2000).

Moreover, in 1998 the Lesotho Congress of Democrats (LCD) took 79 out of the
increased number of 80 seats with just 60.5% of the total vote (Southall and Fox 1999).
Despite winning a combined total of 40% of the votes, the opposition parties were again

left without representation in the parliament. The results were unbelievable to

112

opposition party leaders. They were expecting a close election in which no party would
gain a parliamentary majority, because the LCD split from the then ruling BCP right
before the election. Mokhehle, the then Prime Minster, lost control of his BCP
machinery and created the LCD. Because it enjoyed majority support in Parliament, the
LCD became the ruling party. While international and local observers estimated that the
result of the election reflected the will of the electorate (Southall and Fox 1999), those
losing parties repudiated the results. The rejection of the results by the opposition
parties led to a major crisis that undermined the capacity of the LCD to rule. Opposition
supporters burned down the central business district of the capital, Maseru. This crisis
led South Africa and Botswana to use military force to intervene in the name of
defending democracy and restore order.

The Interim Political Authority (IPA), composed of two members of all political
parties which had participated in 1998 election, was established on the advice of the
intervening powers. Protracted negotiations between the IPA and the government

resulted in agreement on a Mixed-Member—Proportional electoral system (MMP)

(Southall 2003).20 In addition to 80 existing constituency seats to be elected by FPTP,
40 PR seats were allocated among parties, with PR being applied by total national list
election votes cast divided by 120 to determine a quota per seat. The overall PR formula

is Hare quota with no formal threshold providing for the highest level of proportionality

(Lijphart 1995).

 

There are two broad subtypes in mixed-member electoral systems, which they called mixed-member
majoritarian (MMM) and mixed—member proportional (MMP) (Shugart and Wattenberg 2003). In the
MMM systems, there is no linkage between single-seat district tiers and PR list tier in the allocation of
seats to parties. In the MMP systems, however, the number of seats a party takes from its lists is
determined partially by the number of seats it has own from the single-seat district tier.

113

On 25 May 2002 Lesotho became the first African country to test the MMP
electoral model in a parliamentary election (Elklit 2000). 1,085 candidates from 18
parties, alongside 32 independents, were registered to contest the 80 FPTP
constituencies. Sixteen parties presented full list of a total of 770 candidates for election
via proportional representation (Southall 2003). The LCD again made virtually clean
sweep of the constituencies, taking 77 out of the 78 seats21 contested on election day
with 57.7% of the vote. It also took a very similar proportion of the PR vote (54.9%).

Table 6.1 summarizes both vote and seat shares in the elections of 1998 and
2002. The percentages show a much better correspondence between vote shares and seat
shares in 2002 (the first MMP election) than was the case in 1998 (under the old FPTP
system). An election’s Index of Disproportionality (ID) measures the degree to which
the distribution of parliamentary seats among parties diverges from the distribution of
votes, with zero representing a perfectly proportional outcome (Gallagher 1999). While
ID scores typically ranges from near zero to 20, the ID score for Lesotho in 1998
election was 31.5. The May 2002 election produced an ID of 8.2, a fourfold decrease. It
clearly shows an increase in proportionality from the previous FPTP electoral system to

the current MMP system.

Table 6.1 Vote and Seat Shares in 1998 and 2002 Elections

For the first time in Lesotho electoral history, the losing opposition parties

accepted the results. Since multiparty elections were reintroduced in 1993, opposition

parties had rarely taken any seat in the parliament. The new MMP electoral system

 

2‘ Constituency elections were postponed in Mt Moorosi and Hlotse owing to the deaths of candidates.

114

allowed opposition parties to take all 40 PR seats in the parliament. Thus, MMP
performed well in the 2002 election, as the level of disproportionality between vote and
seat shares declined dramatically compared to previous elections. This study
investigates whether, and how, this increase in pr0portionality of electoral systems

affects citizens’ political support in Lesotho.

6.3 Electoral Systems and Citizen Attitudes

The electoral system is an important mechanism connecting citizens and the
state. Electoral systems are conventionally divided into two categories, majoritarian and
proportional (Lijphart 1995). Majoritarian systems usually employ single-seat districts
with a plurality rule and tend to give greater representation to parties that receive the
most votes. Proportional systems employ multi-seat districts, usually with party lists,
and typically produce parliamentary representation that largely mirrors the vote shares
of multiple parties.

Recently, however, there has been a marked tendency around the world to mix
these two principles of electoral system design. Many newly adopted electoral systems,
including those in long established democracies such as Italy, Japan, New Zealand, and
Venezuela, have entailed various hybrids of the majoritarian and proportional principles
(Shugart and Wattenberg 2003). Mixed-member systems employ two tiers: single-seat
district (SSDs) and PR list. In the prototype of a mixed-member system, half of the
seats in a legislative chamber are elected in single-seat districts while the other half are
elected from party lists allocated by proportional representation. Lesotho’s MMP

electoral system has 80 seats for SSDs tier and 40 seats for PR list tier.

115

The literature on electoral systems focuses on the degree of proportionality or
disproportionality in the translation of votes into seats and on their effects on the
number of parties. Lijphart (1999) shows that, although no PR system is perfectly
proportional and although there is a great deal of variation within the PR family, PR
systems tend to be considerably less disproportional than plurality and majority
systems. Proportional systems are more representative than majoritarian and semi-
proportional systems because they facilitate the expression of interests by all relevant
societal and ethnic groups.

While a large literature has classified the types of electoral systems and explored
the relationship between electoral system and party system (Rae 1967; Katz 1980
Taagepera and Shugart 1989; Lijphart 1995; Cox 1997; Reynolds and Reilly 1997), few
studies look at the influence of different electoral systems on citizens’ attitudes toward
political systems. Based on aggregate data collected in 19 European democracies,
Anderson (1998) finds that there is a strong correlation between electoral systems and
aggregate popular satisfaction with democracy, indicating that the more proportional the
electoral system, the higher the level of satisfaction. Farrell and McAllister (2004) also
demonstrate that electoral systems directly affect levels of voter satisfaction with
democracy. Particularly, the disproportionality of electoral systems is a significant and
negative predictor of voter satisfaction with democracy. With survey data set collected
in 25 democracies, however, Norris (1999) finds that majoritarian electoral systems
produce higher levels of confidence in political institutions than do proportional
electoral systems. Thus, the questions of whether and how electoral systems can affect

levels of popular satisfaction with democracy remains unresolved. Maybe this

116

discrepancy is due to the different ways in which these authors measure political
support, with Anderson and others using satisfaction with democracy and Norris using
trust in institutions. This study explicitly tests this possibility.

Lesotho’s electoral reform offers a clear example of the effects of adding a list
tier (40 PR seats). There has been a great increase in the number of parties physically
present in the National Assembly, from an average of one in the 1993-1998 period to
ten after the new MMP election in 2002. Because of the list tier and especially where
voters cast two votes, minor parties were able to gain representation. Those small
parties could act as genuinely national parties by focusing campaigns on gaining votes
throughout the country to maximize their list representation. Given the absence of any
formal mechanism for effective representation of small parties in the pre-existing FPTP
system, the result of 40 seats of opposition parties in the parliament demonstrates a clear
treatment of the role of list tier in facilitating the expression of minority interests. In this
paper I look at whether increasing the proportionality of the electoral system could

induce higher levels of popular satisfaction with the way democracy in Lesotho (H1).

6.4 Partisan status and Attitudes: Winners, Losers, and Non-partisans

Democracy is about winning and losing within a context of given rules: “Since
the struggle for political office is bound to create winners and losers, this necessarily
generates ambivalent attitudes towards authorities on the part of the losers” (Kaase and
Newton, 1995: 60). At the simplest level, if people feel that the rules of the game allow
the party they endorse to be elected to power, they are more likely to feel that

representative institutions are responsive to their needs so that they can trust the

117

political system, and consequently, they are inclined to be satisfied with the
government’s performance (Lambert, er al., 1986) and with the way the system works
(Kornberg and Clarke, 1994; Nadeau and Blais, 1993). On the other hand, those whose
preferred party loses the election are more likely to feel that their voice is excluded from
the decision-making process, producing dissatisfaction with political institutions. In
their study of 11 European democracies, Anderson and Guillory (1997) find that
respondents who voted for the winning party or parties were more likely to be satisfied
with how well democracy works in their country than respondents who voted for losing
parties.

Yet losers play a crucial role in political development. Their regime support has
greater impact on the stability of the regime than winners’ regime support, because
losers are more likely to reject the rules of game that are not responsive to their interests
and to look for an alternative assuring their needs. The disposition of the losers toward
the result of not only founding but also regular subsequent elections is critical to
democratization. To achieve a consensus about the new rules of game after a democratic
transition, it is necessary that losers will not be permanently excluded from political
power, but will have subsequent chances to contest and win elections. We saw several
cases, like Cameroon and Kenya, in which an attempt to marginalize the opposition
from any share of power was followed by a period of political conflict (Bratton and Van
de Walle 1997).

The differences in levels of political support between winner and loser in
emerging democracies are wider than those in developed democracies (Fuchs,

Guidorossi, and Svensson 1995). While citizens who live in Belgium, Germany,

118

Luxembourg, and Denmark show relatively narrow differences (less than 20 percentage
points) in satisfaction with democracy between winner and losers, those in young
democracies such as Portugal, Spain, and Greece, express significantly large differences
of more than 30 percentage points. Citizens living under the latter regimes do not have
long enough experience of democracy. They have been watching just a few regular
multiparty elections and only some of them have experienced the ruling party
alternations in their country. In sub-Saharan Africa, most parties that won founding
elections are still in power (Van de Walle 2003). Those losers in this area have lower
expectations for their supporting party to win an election and more dissatisfaction with
the political system.

Because democracies are inherently designed to create winners and losers
through a series of elections, there must be a tension between those in the majority and
those in the minority in democratic political life. That is, even when they perform
satisfactorily, democracies, by their very nature, face the potential for instability due to
the different preferences that those in the political majority and minority have for
legitimizing the system. I therefore expect that, relative to those in the political majority
(the winners), (HZ-a) citizens who are in the political minority (the losers) should have
a diminished stake in the system. As a consequence, they are likely to express more
negative views of the political system’s performance. In addition, we are able to observe
a significant number of citizens who support neither ruling nor opposition parties in
emerging democracies. For non-partisans, this study expect that their levels of political

support should be somewhere between those for winners and losers (HZ-b).

119

6.5 Partisan status in Different Electoral Systems

The distinctions between winners and losers may be more complex than as we
see in previous section. Different electoral systems may also affect citizens’ attitudes by
providing incentives or disincentives for certain behavioral patterns in different ways.
Winning and losing have different meanings in different electoral systems. Proportional
electoral systems are advocated with producing more equitable outcomes between
political minority and majority and encouraging wider social group representation than
majoritarian electoral systems (Lijphart 1999). These features of the PR systems have a
greater impact on losers than winners. Adopting a PR system increases the possibility to
build an alliance among political minorities and induces protest parties into current
political system that enhances political minority groups’ perceptions of system
legitimacy. Thus, minority groups feel that they are included and represented within the
structures of government which make decisions affecting their interests. It leads to
increase their levels of satisfaction with democracy. Anderson and Guillory’s study
(1997) shows the general pattern that in consensual democracies the differences
between winners and losers are significantly smaller than in majoritarian democracies.

The neopatrimonialism pervasive across sub-Saharan Africa (Bratton and Van
de Walle 1997) creates a divergent incentive structure for winners and losers through
clientelism. Clientelism leads to increase state intervention in the economy. A large
number of economic policies are motivated at least in part by the economic rents
provided to the clients. In the elections, candidates commonly make a promise that they
will provide better roads and social services to people who vote for them. Citizens often

fear and realistically so that they will be deprived of these services if they do not vote

120

for the winning candidate. In Lesotho, the party identification card of the ruling LCD is

often assumed as a ticket for ordinary people to gain resources from the government.

The high stakes of neopatrimonialism are reinforced by majoritarian electoral

arrangements, which are more common in sub-Saharan Africa rather than PR system.22
Winners literally take all with maximum a half of votes. Majoritarian electoral systems
create an incentive for winners (political majority) to bring more pork to their region or
ethnic group at the sacrifice of the numerically smaller population groups who lose
elections. This substantive exclusion increases the degree to which a minority group
feels that it is excluded from the system. This strengthens losers’ dissatisfaction with
democracy and threatens the stability of emerging democracies.

I therefore argue that the nature of representation in electoral systems
systematically mediates the relationship between individual’s partisan status (“winner”
or “loser”) and his or her political support. If the new MMP system increases
representation of previously excluded groups and causes people to have more favorable
evaluations of political processes, I expect to see a shift particularly in political
minority’s attitudes from 2000 to 2003 that reflects greater satisfaction with democracy
in Lesotho. The more proportional electoral systems decrease differences in popular
satisfaction with democracy both between winners and losers and between winner and
non-partisans. It is expected that (H3) losers (and, to a lesser extent, non-partisans) who

live under proportional electoral systems should express higher levels of political

 

22 No more than one-quarter of the countries of the region - including Angola, Benin, Cape Verde,
Equatorial Guinea, Guinea-Bissau, Liberia, Mozambique, Namibia, Sierra Leone, and South Africa -
have adopted various forms of PR.

121

support than those who live under non-proportional electoral systems. Figure 6.1
describes this hypothesized interactive effect. The relationships revealed among
electoral systems, partisan status, and political support will shed light on the question
concerning the effects of institutions on citizen attitudes toward how well their political

systems work.

Figure 6.1 Hypothesized Popular Satisfaction with Democracy in Different Electoral
Systems

6.6 A Model of Political Support

6.6.1 Dependent Variables: Satisfaction with Democracy and Trust in Political
Institutions.

To assess levels of political support, this study focuses attention on both regime
satisfaction and trust in political institutions from the Afrobaromenter Round 1 (2000)
and Round 2 (2003) data. The former is commonly measured by “satisfaction with the
way democracy works in [the country].” Consistent with Anderson and Guillory’s
(1997) approach, the question is concerned with how democracy functions in practice as
opposed to the ideal.

A significant number of scholars criticize the concept of satisfaction with
democracy for its ambiguity, as it can tap multiple dimensions of political support, for
example both support for democracy as a value and satisfaction with the incumbent
government (Norris 1999; Canache et al. 2001; Linde and Ekman 2003). Using
Afrobarometer Round 1 data set, including 12 sub-Saharan African countries, Bratton,

Mattes, and Gyimah-Boadi (2005) find that popular satisfaction with democracy has

122

significant correlations with all three dimensions such as support for democratic regime,
trust in state institutions, and evaluation of government performance. Moreover, the
evaluation of government performance has a consistently powerful relationship with
popular satisfaction with democracy across 12 countries. While they have a relatively
short experience of new democratic systems and rare alternations of government,
Afrobarometer respondents consistently put more emphases on how an elected
government functions in practice when they report on satisfaction with the regime of
democracy.

Another dependent variable for analysis concerns trust in political institutions.
Following Norris (1999), I combined trust in five major institutions, including the Prime
Minister, the National Assembly, the army, the police, and the courts. These items refer
to the confidence that citizens possess regarding the major institutions of the state.
Responses to these five items are highly inter-correlated. Factor analysis extracts two
factors (Eigenvalues are 2.38 for factor 1 and 1.19 for factor 2), which cumulatively
explain 71.4 percent of the common variance. The scale has acceptable alpha reliability
coefficient (Cronbach’s Alpha = .72; n = 2377), suggesting that these five items
reflected a general orientation towards political institutions. In terms of testing
competing theories, this chapter will particularly look at whether any institutional

effects are due to the way in which political support is measured.

123

6.6.2 Independent Variables
Electoral Systems (Index of Dispr0portionality): To measure the disproportionality of
electoral systems, this study uses an index developed by Michael Gallagher (1991) as

follows:

 

. . 1
Gallagher s least-squares rndex = \/[§-Z (Vi - Si)

in which v, is vote percentage and Si is seat percentage. The higher the figure the

greater the disproportionality produced by the electoral systems. Lesotho has
experienced electoral system change from FPTP system to MMP system with adding 40
PR seats into the pre-existing electoral system. It has decreased the disproportionality of
electoral system from 31.5 in 1998 to 8.2 in 2002. I expect disproportionality to have a
negative effect on both popular satisfaction with democracy and trust in political
institutions. Citizens who live under MMP system should show higher levels of
satisfaction with democracy or higher levels of trust in political institutions than do

those who live under FPTP system.

Partisan Status (Winners, lasers and non-partisans): The Afrobarometer survey data set
from Lesotho classifies respondents as being a winner, loser, or non-partisan with the

help of a survey question that asks which party, if any, the respondent feels close to.

124

The data set then combines these respondents with information about the party or
parties that did not win the most recent parliamentary elections to identify loser. This
study treats as non-partisans all respondents who did not feel close to any particular

political party, said “don’t know” or refused to answer the question of party

affiliation.23 Using winner as excluded base category, I created two dummy variables
for both loser and non-partisan. Two variables show the differences in popular
satisfaction with democracy both between winners and losers and between winners and
non-partisans. I expect that both losers and non-partisans would have lower levels of

satisfaction with democratic system or trust in political institutions than winners.

Two Interaction Terms between Partisan status and Electoral systems: To examine
whether and how electoral systems mediate the relationship between partisan status and
citizens’ political support I add two interaction terms -- loser * ID. and non-
partisan*I.D. -- into the models. I expecte that both interaction terms should have a
negative coefficient. Increasing disproportionality of electoral systems should induce
greater differences in popular satisfaction with political systems both between winners
and losers and between winners and non-partisans. I expect that losers marginalized by
previous majoritarian FPTP system should show higher levels of satisfaction with the
way democracy works or higher levels of trust in political institutions after adopting a
new MMP system. This study expects the electoral system change should have a similar

effect on non-partisans’ political support in Lesotho.

 

23 In the sample, 82 respondents (3.5%) refused to answer the question. I also run the same models
without those respondents, but no significant change in the empirical results.

125

6.6.3 Control Variables

Interest in Politics: Many political scientists have demonstrated that interest in politics
is related to political support (Almond and Verba 1965; Anderson and Guillory 1997;
Weatherford 1991). Citizens who understand the political process and believe that their
participation can influence policymaking are likely to take a more optimistic view of
democratic governance. The Afrobarometer asked respondents how interested they are
in public affairs. For the analysis, I created a dummy variable coded as one including
“very interested” and “ somewhat interested.” The other responses are coded as zero. I
expect that those who are more interested in politics are also more likely to show higher

levels of political support.

Economic Performance Evaluations (Sociotropic and egocentric evaluations): Citizen
evaluations of government performance can be one of the most important factors
shaping the reputation of political institutions and of the political system as a whole
(Weatherford 1987). In particular, several studies suggest that economic performance,
as well as public perceptions of both personal and national economic conditions, are
related to system support (Clark, Dutt, and Kornberg 1993; Kornberg and Clarke 1992;
Listhaung and Wiberg 1995 ; Weatherford 1987). For this study, I use responses to two
Afrobarometer questions that asked respondents to evaluate national and personal
economic conditions, that is, to form sociotropic and egocentric evaluations of
economic performance separately. This study expects that those who evaluate economic

performance positively should show higher levels of political support.

126

6.6.4 Demographic Variables
I control for the usual socio-demographic variables: education, gender, and age.
Coding procedures and descriptive statistics for the individual-level variables used in

this study are shown in Appendix B.

6.7 Results

6.7 .1 The Effects of Partisan status on Satisfaction with Democracy at the Aggregate
Level

At first glance, aggregate levels of political support appear to increase over time.
Between 2000 and 2003, the proportion of Lesotho’s citizens expressing satisfaction

with democracy rose from 38% to 49%, a significant difference (Gay and Mattes

2004).24 Aftobarometer Round 2 demonstrates that Lesotho is rather unusual in this
regard, since popular satisfaction with democracy has been significantly decreasing in
most African countries (The Afrobarometer Network 2004). In only three countries

(Ghana, Lesotho, and Namibia) do respondents show higher levels of satisfaction with

democracy in 2003 than 2000.25 Similarly with institutional trust: Lesotho was one of
only three African countries (the other two were Malawi and Mali) where this attitude is
trending upward. And the statistically significant l4-polint increase in institutional trust

in Lesotho (from 40 percent to 54 percent average) was wider than for any other African

 

24 . . . . .
Chi-square = 48.280, p < .001. “Satisfaction” combines respondents who are “very” or “fairly”
satisfied.

25 The increase was 17 percentage points in Ghana and 5 percentage points in Namibia. The change in
Ghana was inflated, however by a change in question wording between Afrobarometer Round 1 and 2.
Thus, I am confident that satisfaction with democracy increased more in Lesotho than in any other
African country.

127

county studied. So far, therefore, the results for Lesotho are consistent with the
expectation of H1.

Are winners more satisfied than losers with the way democracy works in
Lesotho? What about non-partisans? To examine differences in satisfaction between
winners, non-partisans, and losers, I compare satisfaction with democracy between

these three groups across two different years.

Figure 6.2 Satisfaction with Democracy among Winners, Non-partisans,
and Losers

Combining the percentage of respondents saying they are “very” or“ fairly
satisfied” with the way democracy works in Lesotho, Figure 6.2 shows the distribution
of satisfaction with democracy for winners, non-partisans, and losers under two
different electoral systems (FPTP and MP systems). The results reported in Figure 6.2
provide largely consistent support for my hypotheses. There is a clear gap in the level of
satisfaction with the democratic system between winners and losers in each electoral
system. Non-partisans’ levels of satisfaction are somewhere between those for winners

and losers. Losers show a doubling of satisfaction with democracy from 23 to 46

26 .
percent after electoral system changes. Whrle the vote losers cast under the old
electoral system was not an effective choice to show their will, the vote under the MMP
system has “real” effects on representing their interests in parliament with 41 seats of

opposition parties. Thus, we infer that electoral reform induced opposition party

 

26 t: -2.31, p < .05.

128

supporters to significantly upgrade their assessments of the performance of Lesotho’s
democracy.

On the contrary, there is an obvious, but relatively small decrease in winners’

satisfaction with democracy from 75 to 67 percent.27 Winners under the MMP system
express lower levels of satisfaction than do winners under the FPTP system. Many LCD
supporters complained about the 2002 election results, because their party took no PR
seat with 54.9 percent of the PR vote. It was the first time for Lesotho people to
experience the MMP electoral system. Ordinary people need more time to fully
understand the new electoral system. Non-partisans did not show any changes in
satisfaction with democracy between two different electoral systems. As Figure 6.2
shows, the difference in the level of satisfaction with democracy between winners and
losers decreased from 52 to 21 percentage points as Lesotho changed the electoral
system from FTPT to MMP system. This suggests that the MMP system reduced the
proportion of voters who cast ineffective votes and incorporate marginalized political

minorities into the political system.

6.7.2 Multivariate Models
Next, I examine the performance of the individual-level variables in a series of

multivariate models. I use Ordinary Least Squares (OLS) estimators in Table 6.2.

Table 6.2. Effects of Individual Characteristics on Satisfaction with Democracy and
Trust in Political Institutions in 2000 and 2003

 

27t=-4.51,p<.001..

129

I find that the effect of loser (or minority) remains even when I control for a
number of other influences, such as interest in politics, economic performance
evaluations, and demographic factors. The effect is strong and consistent with the
expected direction. Those coefficients for loser are consistently significant and negative
in all four models of Table 6.2, indicating that citizens with preferences for opposition
parties are almost always significantly less satisfied with the way democracy works than
those with preferences for the ruling LCD. Losers have similar attitudes in terms of trust
in political institutions. They significantly show a lower level of trust in political
institutions than winners. Non-partisans express significantly lower levels of both
satisfaction with democracy and trust in political institutions than do winners. Those
coefficients for non-partisans are consistently significant and negative in all four models.

The absolute value of loser’s coefficient is greater than non-partisan’s
coefficient in the model of satisfaction with democracy, indicating that non-partisans’
levels of satisfaction are likely to be somewhere in between those for winners and losers.
In the model of trust in political institutions, however, the results are not consistent
between two different electoral systems. While the difference in popular trust in
political institutions between winners and losers is likely to be greater than the
difference between winners and non-partisans under FI‘PT electoral system, the former
is likely to be smaller than the latter under the new MMP system. Because the new
electoral system allowed opposition parties to take a significant number of seats (41) in
the parliament, citizens whose party is in opposition might show relatively higher levels

of trust in political institutions than those who do not support any established party.

130

Interest in politics is a statistically significant and positive predictor of trust in
political institutions. Citizens who are more interested in politics are likely to show
higher levels of trust in Lesotho political institutions. While interest in politics is a
consistently powerful predictor of the trust in both 2000 and 2003, the degree of its
effects relatively decreased after citizens experienced electoral system changes from
FPTP to MMP system. This evidence indicates that introducing a proportional aspect
into Lesotho electoral system reduces the differences in popular trust in political
institutions between citizens who are interested in politics and those who are not. In
2000, it was a big issue among political elites what kinds of electoral system should be
introduced in Lesotho after the 1998 political crisis. The discussion about a new
electoral system among party leaders might draw ordinary citizens’ attention into
politics. However, after they agreed upon introducing the MMP system and held the
2002 elections under the new electoral system, political elites might not have any more
interesting political issue to attract people’s attention to politics.

The role of economic performance evaluations in both satisfaction with
democracy and trust in political institutions is not clear in the result. The coefficients for
national evaluation of the economy are consistently significant and are in the expected
direction in 2000. Citizens who reported that the national economic condition was good
were likely to express higher levels of both satisfaction with democracy and trust in
political institutions. In 2003, however, neither national nor personal evaluations of the
economy are statistically significant to explain both satisfaction with democracy and
trust in political institutions. Lesotho’s economic status is typical for an African

country: its Gross National Income per capita of $470 mirrors the sub-Saharan African

131

average of $451 (World Bank 2004:5). People are largely dissatisfied with government
performance on recent economic liberalization to attract foreign investment. In
Afrobarometer Round 2, 82 percent of the respondents said that national economic
condition was either “bad” or “very bad,” and 86 percent reported their own personal
living conditions this way as well (Gay and Mattes 2004). This extremely unbalanced
distribution may lead those effects of two economic performance evaluations on both
popular satisfaction with democracy and trust in political institutions to being marginal
in both 2003 models.

Apparently, however, the positive effects of institutional reform were sufficient
to wash out the effects of economic pessimism on mass political satisfaction and
institutional trust, at least during the post-reform honeymoon of 2003. Moreover, since
there was no meaningful change in the country’s economic growth rate during the

period of this study, I am confident that other events or trends -— like a jump in the

. . . , . . . 28
country’s economic fortunes - do not explain the publlc 5 improved political mood.

6.7.3 Effects of Electoral systems on Political Support

I next test whether the electoral system reform from FPTP to MMP, increasing
fairness in the transfer of votes to seats, has generalized effects on citizens’ attitudes
toward political systems with the pooled sample. I then test effects of individual
partisan status on popular satisfaction with democracy and trust in political institutions
after controlling the effects of electoral system with other variables. Finally, I look at

whether electoral systems mediate the relationship between citizens’ partisan status and

 

28 Indeed, the growth rate dropped from 2.8 to 2.4 percent between 2002 and 2003 (World Bank 2005).

132

their attitudes toward democratic performance. I assume that political minorities
disadvantaged by majoritarian rules will have relatively more positive attitudes toward
the political system after a transition from a majoritarian system to proportional
representation. To test whether losers who live under proportional electoral system
(MMP system) exhibit higher levels of satisfaction and trust than do those who live
under non-proportional electoral system (FPT P system), I estimate the pooled OLS
model with an interaction term between loser and the Index of Disproportionality (ID).

Table 6.3 shows the results of the pooled OLS models of Satisfaction with democracy.

Table 6.3. Effects of Electoral Systems on Satisfaction with Democracy:
Pooled Model Estimates

Electoral system change had no direct, generalized effects on citizens’
satisfaction with democracy. The coefficient on ID is not statistically significant in
Model 1. The result is not consistent with any of previous studies of the direct impacts
of political institutions on citizens’ attitudes (Anderson 1999; Farrell and McAllister
2003). This study shows that the electoral reform to resolve the long lasting conflict
between ruling and opposition parties does not have any uniform impacts on levels of
popular satisfaction with democracy in Lesotho, because, at the same time that losers
and non-partisans saw benefits, winners perceived drawbacks. And, in the 2002 election
and the 2003 survey sample, winners greatly outnumbered losers (see Figure 6.2). The
slight drop-off in democratic satisfaction among 655 randomly selected winners
interviewed in 2003 was sufficient to offset the doubling of democratic satisfaction

among 208 randomly selected losers. Those LCD supporters (winners) were not

133

satisfied with the electoral results of the new MMP system, because they were

accustomed to winner-take-all system (FPTP). When I interviewed LCD party

leaders,29 they said that after the 2003 parliamentary elections those ordinary party
supporters complained how opposition parties took the 40 seats without no win in all
constituencies but one and why the LCD didn’t receive any seat from the PR tier.

However, party leaders’ evaluation of the electoral reform was more positive than the

mass opinion of ordinary citizens.30 Regardless of their party affiliation, all party
leaders with whom I conducted an interview said that the 2002 elections produced more
acceptable results than the 1998 elections because of the new MMP system and they
were more satisfied With democracy after electoral reforms because the electoral system
changes brought both peace and stability in Lesotho politics. In addition, those
opposition party leaders emphasized that the future electoral system should be a pure
(100 percent) PR system.

While there is no uniform shift in popular satisfaction with democracy, thus, I
must emphasize that the results of Model 1 do not suggest that the change in electoral
systems did not have any effect on citizens’ satisfaction with democracy. Rather this
finding suggests that electoral system changes had different effects on different political
groups’ attitudes toward democratic government. In Model 2, I include two dummy
variables of partisan status, one for losers and the other for non-partisans.

Being an electoral loser is a strong determinant of satisfaction with democracy.

Opposition party supporters are likely to express lower levels of satisfaction with

 

29 Interview with Sephiri E. Motanyane, deputy speaker of the National Assembly; M.S.Lehata, MP;
R.F.Shea, whip of the LCD, December 2004.

Interviews with 10 party leaders from various parties in the National Assembly, December 2004.

134

democracy than are ruling party supporters. Non-partisan is a statistically significant
and negative predictor of popular satisfaction with democracy. Moreover, evaluation of
national economic performance is consistently significant and positively related to
satisfaction with the way democracy works. So far, the individual-level analyses have
provided support for my argument that both losers and non-partisans have
systematically different attitudes about the way democracy works from winners in
Lesotho, even when controlling for evaluations of personal and national economic
performances, interest in politics, and a number of demographic factors.

Because both losers and non-partisans are less likely to be satisfied with the way
democracy works than winners, and because, as shown in Figure 6.2 and Table 6.2,
there are significant differences in the level of the satisfaction between losers under
FPTP system and those under MMP system, I expect that the differences in satisfaction
with democracy between winners, non-partisans, and losers can be explained by an
electoral system difference (FPTP/MMP system). Those people most disadvantaged by
the previous FPT P system are expected to show more positive attitudes toward the
political system after the MMP election in 2002. The results are shown in Model 3,
which includes two interaction terms between [D on the one hand and losers and non—
partisans on the other.

I find that the electoral system mediates the relationship between partisan status
and popular satisfaction with the way democracy works. The coefficient of the
interaction term between loser and ID is statistically significant and negative in Model 3.
More proportionality in electoral systems is likely to alleviate losers’ dissatisfaction

with democracy. Conversely, less proportional (or more majoritarian) electoral systems

135

increase differences of satisfaction with the way democracy works between winners and
losers. Losers might feel disadvantaged or marginalized by the majoritarian FPTP
system that failed to offer them credible chances for representing their preferences. For

substantive interpretation, I calculate the simulated expected values of popular

satisfaction with democracy.31 As Figure 6.3 shows, the expected values for losers
increase from 1.96 to 2.20 after the electoral reform, while the values of winners

decrease from 3.17 to 2.79. The results are consistent with H3.

Figure 6.3. Predicted Probabilities of Popular Satisfaction with Democracy

Electoral systems have a similar impact on non-partisans’ attitude toward
democratic system. The coefficient of the interaction term between non-partisan and ID
is statistically significant and is in the negative direction. More proportional electoral
systems decrease differences of satisfaction with democracy between winners and non-
partisans. The results show that the gaps in popular satisfaction with democracy both
between winners and losers and between winners and non-partisans are reduced
following the adoption of MMP system. The MMP system mitigates the winner-take-all
characteristics of the previous electoral system with 40 PR seats for opposition parties.
It enhances citizens’ perceptions of system legitimacy by incorporating opposition
parties into the political system. Otherwise, opposition parties and their supporters

could be severe threats to the survival of an emerging democracy. In this respect, the

 

31 Expected values are the mean of the 1000 simulated expected values generated by Clarify 2.1 (Tomz,

Wittenberg, and King 2003).

136

results from Lesotho are consistent with the argument about the advantages of

consensual institutions made by Anderson and Guillory (1997).

In Table 6.4, finally, I run the same model with the different dependent variable
of trust in political institutions as I do in Table 6.3 in order to look at whether electoral
reforms have the same effects on citizens’ trust in political institutions as they does on

popular satisfaction with democracy.

Table 6.4. Effects of Electoral Systems on Trust in Political Institutions:
Pooled Model Estimates

Table 6.4 report results for the same model with trust in political institutions as
the dependent variable. In this case, electoral system changes have effects on citizen’s
trust in political institutions in general. The ID coefficient is statistically significant and
negative in Model 1. This implies that the increased proportionality resulting from
electoral reforms is likely to boost citizens’ levels of trust in political institutions. Note,
however, that these effects disappear when partisan status and interaction terms are
entered into the equation. For this reason, I offer interpretations of only the fullest set of
the results, that is, Model 3.

The results show that popular trust in political institutions arises from multiple
sources. Interest in politics is a statistically significant and positive predictor of trust in
political institutions. Citizens who are more interested in politics are likely to show
higher levels of trust in political institutions. Both national and personal evaluations of
the economy play an important role in trust in political institutions. Their coefficients

are statistically significant and positive all across the models. Assessments of national

137

economic condition are more influential than assessments of personal economic
condition.

Gender and age variables are consistently significant in three models. The
results indicate that the younger respondents are, the higher levels of trust in political
institutions they express. Since the new MMP system is more complex than the old
FPTP system, younger generation could be more confident about the democratic system
than older generation. Women are likely to show lower levels of trust in political
institutions than are men in Lesotho.

My main concern, however, is with electoral systems and their interactions with
partisan status. The effects of both loser and non-partisan — which are statistically
significant and in the expected (negative) direction — again survive controls for
extraneous factors. And, the difference in trust in political institutions between losers
and winners is greater than the difference between non-partisans and winners.

The electoral system change has a mediate effect on the relationship between
non-partisans and their trust in political institutions. In Model 3, an interaction term
between non-partisans and ID is a statistically significant and positive predictor of trust
in political institutions. The less proportional electoral systems are, the higher non-
partisans’ levels of trust in political institutions are. While there is an overall increase in
popular trust in political institutions in Model 1, citizens with no preference for parties
are likely to express lower levels of the trust after the adoption of MMP system.
Electoral reforms increased differences in levels of trust between non-partisans and
winners, but did not have any impact on the differences between losers and winners. I

calculate the simulated expected values of citizens’ trust in political institutions. While

138

the expected values of the trust for both winners and losers increase after the electoral

system change, the expected values for non-partisans decrease from 5.80 to 5.39.32 In
other words, this group considers that political institutions are more trustworthy under
an FPTP electoral system, which is consistent with the argument about the advantages
of majoritarian institutions made by Norris (1999).

Electoral systems mediate the relationship between an individual’s partisan
status (winner, non-partisan, or loser) and his or her attitude toward political systems.
This study shows that more proportional electoral systems are likely to induce narrower
gaps of popular satisfaction with democracy between losers and winners as well as
between non-partisans and winners. On the other hand, more proportional electoral
systems are likely to increase differences in popular trust in political institutions
between winners and non-partisans. The results suggest that various electoral systems
affect citizens’ attitudes by providing incentives or disincentives for certain behavioral

patterns in various manners.

6.8 Conclusion

Electoral system changes have significant effects on citizens’ attitudes toward
their political system. These effects are not uniform as existing studies on institutional
effects on citizens’ political support demonstrate. I have found that winning and losing
have distinctive meanings under different electoral systems. PR rules boost positive

attitudes toward political system among political minorities who were disadvantaged or

 

32
The expected values for winners increase from 7.04 to 7.48 and the values for losers change from 5.33
to 5.71.

139

marginalized by winner-take-all electoral systems. After electoral reforms, both losers
and non-partisans are likely to show higher levels of satisfaction with democracy than
before. On the contrary, electoral reforms have negative effects on popular satisfaction
with democracy among winners who benefited by previous majoritarian rules. Electoral
reforms also affect citizens’ trust in political institutions. PR rules increase levels of the
trust among winners and losers, while non-partisans are likely to show lower levels of
trust in political institutions after adding 40 PR seats to existing 80 district seats.

Lesotho is one of sub-Saharan African countries plagued by neopatrimonialism,
in which resources are distributed to clients in return for political loyalty. Because
private-sector economies remain underdeveloped, government executes its monopoly
power to collect and distribute public resources. With no seat in the National Assembly,
moreover, there is no way for opposition parties to influence and check government
decision making, while they constantly receive a certain number of votes in elections.
Those votes for opposition parties are not effective at all to elect representatives in the
National Assembly. The ruling party enjoys and reinforces neopatrimonial network to
secure winning next elections. Access to government resources is often critical to
economic and political survival for both political elites and ordinary voters. People
often are scared to be left away from the clientelistic network if they do not support the
ruling party.

After electoral reforms, however, opposition parties have been able to watch
what and how the LCD government is doing for people in Lesotho, because of their 41

seats in the National Assembly. The National Assembly has five sessional select

140

committees in which all chairs are from the opposition parties.33 While the 41 seats are
obviously small relative to 79 seats of the ruling LCD in the parliament, opposition
parties are at least able to bring important issues to the public, for example, how the
government executes the budget. Thus, people are more likely to acknowledge what the
government is doing and have an opinion on given government policies. Citizens whose
party is in opposition begin to feel their party has an influence on policy-making in the
government only after their representatives take seats in parliament. Prior to the
electoral reforms, opposition parties ought to be outside of the National Assembly and
their supporters could challenge the legitimacy of the government. It led to an unstable
regime as an emerging democracy in particular. As this study shows, however, after the
reforms enhancing fairness in the transfer of votes to seats and increasing effective
parties take place, both losers and non-partisans are likely to display higher levels of
satisfaction with the way democracy works than before.

This study suggests that political reforms should focus on political institutions
marginalizing or alienating certain groups of citizens in political systems. Increasing
proportionality in the translation of votes into seats in parliament, for example, leads to
increases in positive attitudes about political systems among people in the political
minority. Electoral system changes are likely to enhance people’s engagement in
democracy and contribute to the democratic consolidation of a regime.

We have observed just a few cases of alternation of ruling party in sub-Saharan
Africa. People in the political minority are likely to remain marginalized by both formal

and informal institutions and they feel that political systems are not representing their

 

Busrness Commlttee, House Committee, Committee on Standing Orders, Public Accounts Committee,
Staff Committee, Committee of Privileges, and HIV/AIDS Committee.

141

interests. The gaps in political support between losers and winners are likely to grow in
this area. Because of the relatively short experience of democratic institutions, however,
emerging democracies necessarily need losers’ consent and support for the democratic
system to survive and develop for a long time. This study shows that in addition to their
minority status in politics, losers are less likely to be satisfied with the way democracy
works if there are few opportunities to influence a decision-making process, such as no
parliamentary seat for opposition parties. For democratic stabilization, therefore, this
study suggests that institutional reforms allowing the political minority to gain more
access to the decision-making process in emerging democracies.

The final caveat concerns the object of inquiry in studies of political support.
Leading analysts in this field have not adopted uniform constructs and measurements of
this important dependent variable. Norris, who studies trust in state institutions, may not
be fully justified in counter-posing her research results against Anderson and Guillory,
who are concerned with popular satisfaction with the regime of democracy. The only
way to resolve the debate over the relative advantages of majoritarian versus consensual
institutions is with reference to the same dependent variables. Fortunately, the
Afrobarometer measures both satisfaction with democracy and trust in political
institutions. In this chapter, I have shown that proportional rules boost citizen political
support when this concept is measured as satisfaction with democracy. But I have also
found trace evidence among non-partisans in Lesotho that disproportional (i.e.
majoritarian) rules are more conductive to political support, at least when this concept si

measured in terms of institutional trust.

142

At minimum, then, analysts should explicitly identify the referents of political
support. If the object of study is democratization, then satisfaction with democracy is
clearly the preferable indicator; but if one is interested in the legitimacy of the state,
then it is appropriate to measure trust in institutions. Divergent approaches and findings
in the current research have substantive implications, including what we see as the
relative sophistication of citizens in new democracies. I have shown that the citizens of
Lesotho, especially those who previously lacked representation, report that electoral
reforms improve the quality of their country’s democracy. At the same time, however,
these same citizens doubt that a change in electoral institutions will alter the day—to-day
operations of the executive agencies of the state. They recognize that, even where
electoral democracy and parliamentary representation have been attained, power still

resides with leaders who are driven by formal and informal norms of “winner take all.”

143

2:89AM €985

Notowtomoam

:ELSYQQE

 

Emma..—

 

mgmmgméoz

J

£0253

 

 

23

xomeooEoD

firs
couommmuam

as:

mEonxm 12885 “:80th E xomeooEoQ HEB :ouofimumm wommmofiomxm 4.0 8sz

144

 

_ m8 we. 5 MR 8N me u z _

 

20:53 23:89:02 £83

 

 

 

 

 

 

 

€83 use .mcamfitaméoz £255» macaw honaoofiofl HEB :ouofimumm «6 Earn

145

 

 

883+ m§m_§m-=oz.r . EEEBIOI

 

 

moom ooom

 

 

 

 

 

 

 

 

 

 

 

 

xomeooEoQ 5MB couoflmuam 5395 no mesa; “682on m6 shamr—

o;

m4

QN

Wm

o.m

Wm

ed

146

 

 

 

m F F 8F 09 ow 8 F ooF ES...

Nd m. Fm n:

o o a. F o o m 38:88::on

F md 50 .m.: .m.: .m.: mZZ

F m.o F. F .9: 6.: .m.: Em

F no 3 .9: .9: .9: .55

m 3 ma .9: 6.: .9: 93

m N6 m.m .9: .m.: .m.: E2

m N6 Qm .9: .m.: .9: On:

F wd NF 0 o 0F :32

m mN Em C o 09 mum

FN wNF YNN F m. F mém mZm

R Qmo mém mm mdm Now DOA

A65 Baum Aoev anew Aosv 38> A65 33m 36v 3QO 33 88> moEwm
N08 89

 

26:85 88 Be 32 5 69% now an 96> .3 29¢

147

 

 

 

38.9 88.9 88.9 98.9
wood. :83. mod mood mood omod omod mood
68.9 :89 $89 88.9
mood- oood- S to- Snood- o Foo- Food- mood mood
$2.9 38.9 88.9 38.9
mood- zomod- Food- lowed- oood- oFod- mood- mood-
$9.9 $3.9 Aooo9 :89
mod omod ooFo ...ooo.o mood Koo oood mood
$8.9 38.9 $8.9 $8.9
oood oopd mono Smood mood mood mono £33
:59 63.9 83.9 98.9
onod ...ooo.. 5d too: Food 83 mod mood
98.9 88.9 $8.9 $8.9
QVNAVI ceaONm. Pi mm F.01 iteONm. Fl FON.OI aaio Fm.ol N FMAU: fiiiQGN.°l
$3.9 68.9 98.9 :39
FG —..Ot «ea—mm“. P1 am F.01 «ceN PF. Ft ma P.0n neegm.°l 000.01 eaema F. F1
90m .9000 090m .500 20m .9000 .2000
$2.9 doom €an ooom $2.9 doom $.59 ooom

 

 

 

£03355 303—09 5 3.3.2.

hodooofion— FEB :0zoflmcdm

8:835

03..

2 H 29:03 30:00

Docgotom

250:00m 3:039:

8:588:92
9:853 3:232
82.09 E 382:2
58:89:02

Abtofiev .304

o_nm_:m> Eon—.3005

 

MOON fig OOON Cm wGOflBH—mfiu Roam—Cam a: awn—HR. USN

>080080Q £23 :0uofimuom :0 85008830 3:03:03 90 3009mm .m0 033.

148

.5. v Q... .m0. v Q... ”00.5-02: 0.0 $00. Emoficofi .2. 000050050 :_ 0.0 20:0 0.00080 ”00.00500 08030.0. 7.04 .0255 ”0:02

 

mm:
cord

80.9
...8os

N? F N00 0K 2
m0 F .0 000.0 «0 F0 mm
88.9 3&9 :89
1,000.0 Emma/N Enood 25800

149

Table 6.3. Effects of Electoral Systems on Satisfaction with

Democracy: Pooled Model Estimates

 

 

 

Model 1 Model 2 Model 3
Independent
Variable Coef. Beta Coef. Beta Coet. Beta
Index of
Disproportionality 0.002 0.018 0.007‘" 0.066 0.016‘" 0.159
(0.003) (0.002) (0.004)
Loser (minority) 0840*" -0.269 -0.370"** -0.119
(0.076) (0.140)
Non-partisan
-0.620”* -0.250 0388"" -0.157
(0.063) (0.1 19)
Loser * ID.
0027*“ -0.191
(0.007)
Non-partisan * 1D.
0013“ -0.139
(0.005)
Interest in Politics 0.188" 0.057 0.100 0.031 0.089 0.027
(0.081) (0.078) (0.078)
National Economic
Performance 0116*“ 0.106 0.103'“ 0.094 0.101 '** 0.092
(0.027) (0.026) (0.002)
Personal Economic
Performance 0.044 0.037 0.051 0.043 0.057” 0.048
(0.029) (0.028) (0.028)
Gender
(female = 1) -0.043 -0.018 -0.032 -0.013 -0.030 -0.013
(0.057) (0.054) (0.054)
Age 0.001 0.018 0.001 0.015 0.001 0.012
(0.002) (0.002) (0.002)
Education 0010 -0.013 0.028 0.039 0.025 0.034
(0.018) (0.018) (0.017)
Constant 2.084'“ 2100“" 2203*"
(0.157) (0.139) (0.156)
Adj. R2 0.014 0.097 0.105
N 1771 1771 1771

 

Note: Ordinary Least Squares estimates; standard errors are in parentheses.
All significant tests are two-tailed: **p < .05, ***p < .01.

150

Table 6.4. Effects of Electoral Systems on Trust in Political

Institutions: Pooled Model Estimates

 

Independent
Variable

Index of
Disproportionality

Loser (minority)

Non-partisan

Loser * ID

Non-partisan * ID

Interest in Politics

National Economic
Performance

Personal Economic
Performance

Gender
(female = 1)

Age

Education

Constant

Adj. R2
N

 

 

Model 1 Model 2 Model 3
Coef. Beta Coef. Beta Coef. Beta
0018” -0.054 -0.005 -0.015 -0.019 -0.059
(0.007) (0.007) (0.010)

-1.781‘“ -0.177 -1.802"* -0.179
(0.216 (0.409)
-1.643“* -0.210 2408*“ -0.307
(0.175) (0.343)
0.003 0.007
(0.018)
0.037“ 0.128
(0.015)
1.360'" 0.140 1.116‘“ 0.115 1.135“ 0.117
(0.207) (0.205) (0.205)
0.290‘” 0.083 0.268'“ 0.077 0264*" 0.075
(0.074) (0.072) (0.072)
0.223‘" 0.059 0.242'“ 0.064 0.247'” 0.065
(0.080) (0.078) (0.072)
0630*" -0.083 0591‘“ -0.078 0603*" -0.079
(0.156) (0.152) (0.151)
0015'" -0.070 -0.016"*' -0.074 0016*“ -0.074
(0.005) (0.005) (0.005)
0138"" -0.057 -0.056 -0.023 -0.052 -0.022
(0.052) (0.051) (0.051)
6.270‘" 6.951'" 7.179'"
(0.428) (0.425) (0.439)
0.052 0.096 0.098
2321 2321 2321

 

Note: Ordinary Least Squares estimates; standard errors are in parentheses.
All significant tests are two-tailed: **p < .05, ***p < .01.

151

 

Chapter 7: Conclusion

This study provides clear support for the argument that political institutions,
especially party systems, presidentialism or parliamentarism, and unitarism or
federalism, strongly determine how individuals in emerging democracies view the way
democracy works in their country at the national level. My analyses show that popular
satisfaction with democracy is increased in countries that have greater ENPP, where
citizens are likely to have more options to choose from. People in presidential systems
are likely to show higher levels of popular satisfaction with democracy than those in
parliamentary systems. Unitarism also increases levels of popular satisfaction with
democracy across new democracies.

Those findings that both presidentialism and unitarism have positive effects on
popular satisfaction with democracy are contrasted with consociationalists’ points of
view. Many consociatioinalists recommend parliamentarism and federalism for new
democracies with divided society. A large number of existing studies on political
institutions have focused on established democracies rather than emerging democracies,
while one of the most important goals of such studies is the democratic consolidation
for the latter. On the contrary, this study pays direct attention to both political
institutions and popular Opinions in fledgling democracies. The results show that people
in presidential and unitary systems are likely to show higher levels of satisfaction with
democracy than those in parliamentary and federal systems. This finding suggests that
there exist differences in both how political institutions work and how citizens

understand and evaluate the performance of political systems between established

152

 

democracies and new democracies. Citizens who live in new democracy have a
relatively shorter experience with a new regime and, in turn, are less confident with
political institutions than those live in established democracies.

This cross-national analysis of the impact of electoral system on popular
satisfaction with democracy raises questions about the role electoral institutions play in
both emerging and established democracies. The results of this study show that electoral
systems, measured by disproportionality between votes and seats in parliament, have no
significant direct effects on citizens’ views about how a democratic system works. This
finding suggests that citizens in emerging democracies still need more time to learn
about an electoral system in order to evaluate how the system counts their vote. The
question of whether elections were free and fair is still the most critical issue among
people in a new democracy when evaluating a regime performance.

At individual level, a citizen’s partisan status has a clear effect on his or her
satisfaction with democracy. The gap in people’s views of how political systems works
is significantly large between those in a political majority and those in a minority across
new democracies. The results confirm that people whose party is in government are
more likely to show higher levels of satisfaction with democracy. In contrast, both
losers and non-partisans are less likely to believe that the political system is responsive
to their interests.

Losers’ evaluation of political system performance, however, is critical for the
consolidation of political institutions in emerging democracies and, in turn, ultimately
for the survival of the regime itself. Without losers’ consent and support for the system,

those democracies hardly sustain their democratic consolidation process. Thus,

153

 

narrowing the gaps in popular satisfaction with democracy between winners and losers
needs to be taken as a serious issue in emerging democracy to continue the functioning
of the new regime as well as its life.

In addition, the results show that party systems mediate the relationship between
citizens’ partisan status and their satisfaction with democracy. Greater ENPP is likely to
mitigate the negative view of political minority (losers and non-partisans) about the way
democracy works and, in turn, decrease the differences in popular satisfaction with
democracy both between winners and losers and between winners and non-partisans.

The empirical results support the new institutionalists’ argument that individual
behaviour is constrained and channeled by the institutional context. This research
observed the variations in losers’ satisfaction with democracy across countries. It has
been shown that party systems explain differences across nations. Greater EN PP in
political systems means that no party can be a dominant one in a parliament and that
parties need agreements from each other in order to pass laws. Opposition parties in a
polity with greater ENPP have more chances to influence the decision-making process
than those in a polity with smaller ENPP. Thus, losers in greater ENPP are likely to
express higher levels of popular satisfaction with democracy than those losers living in
a smaller ENPP or in a one party dominant system.

Finally, this study demonstrates the effects of electoral system changes on
citizens’ political support in Lesotho. Particularly, electoral reform introducing PR seats
in the National Assembly shows increased popular satisfaction with democracy and
trust in political institutions. Because of the reform, opposition parties have been able to

win 40 seats in parliament since democratic transition in 1993. Replacing FPTP with

154

 

MP is likely to narrow the gaps in popular satisfaction with democracy both between
winners and losers and between winners and non-partisans in Lesotho.

This study sheds light on the general issue of whether institutions matter in
emerging democracies. Constitutional debates — whether about electoral systems,
executive-legislative relations, or unitarism versus federalism — have risen on the policy
agenda to become very important issues in emerging democracies. If institutional
arrangements influence political support, as this research finds, this may have important
consequences for constitutional debates. The challenge for further research is to
consider these issues in terms of the democratization process. If institutional designs
can strengthen popular satisfaction with democracy this may provide significant lessons
for the process of democratization.

This study can also be extended to consider how political institutions manage
ethnic conflict. New democracies have been establishing a set of political institutions in
order not only to structure political competition but also to accommodate the incentives
of different groups. Electoral rules, for example, are expected to manage the conflict
and to contribute to political stability. Many arguments have been developed about this
issue, but few studies have empirically tested these arguments. Future study will benefit
from exploring whether and how political institutions mediate the relationship between
people’s ethnicity and their political support.

This study clearly shows how important formal institutions are in shaping
people’s views about their political system performance. This finding suggests that
scholars should pay attention to both formal and informal institutions in order to

improve institutional analyses. In emerging democracies, constitutions and state

155

 

institutions coexist with informal institutions, for example, clientelism. corruption, and
neo-patrimonialism. Those informal institutions are argued to have a significant effect
on shaping people’s incentives and expectations in developing countries in particular.
For political scientists, nevertheless, identifying informal institutions and measuring
their effects in order to compare across nations are still challenging issues. This study
demonstrates that formal institutions should be included to explain people’s attitudes
toward political systems.

More generally, this study explains popular attitudes toward political system by
examining the contingent effects of an individual-level variable (political majority-
minority status) and a contextual variable (various political institutions). By integrating
both micro- and macro-level explanations of political support, this analysis enhances
our understanding of how citizens form their attitudes toward the democratic systems in
countries with different political institutions.

In methodology, it is important to emphasize that a statistical model must reflect
the nature of the data being examined. For example, this study works with two-level
data: respondents are nested within nations. Respondents in a given nation are not truly
independent; they are clustered and duplicate one another to some extent. Ignoring this
feature of multilevel data set is likely to lead to incorrect standards errors. The results
demonstrate that OLS estimation is likely to underestimate the standard errors for
national-level variables. When we test a cross-level hypothesis, such as whether the
effect of lower-level predictors is conditioned or moderated by higher-level predictors,
multi-level analysis is necessary. It allows researchers to estimate a model with varying

intercepts and slopes. Thus, we can see whether there is a single, uniform causal effect

156

 

or whether the causal effect varies across higher levels of analysis. This study shows
that the effect of partisan status on popular satisfaction with democracy significantly

varies across nations and this variance is likely to be explained by party systems.

157

 

 

APPENDIX A

Coding

Satisfaction with Democracy. “Overall, how satisfied are you with the way democracy
works in your country?” Response categories included: 1 = “not at all satisfied”; 2 =
“not very satisfied”; 3 = “fairly satisfied”; 4 = “very satisfied.”

Partisan Status. “Which party do you usually think of yourself as close to?” If party
choice matches with a government party (winner), if it matches with an opposition party
(loser), otherwise (non-partisan).

National Economic Performance. “In general, how would you describe the present
economic condition of your country?” 1: “Very bad”; 2 = “Fairly bad”; 3 = “Neither
good nor ba ”; 4 = “Fairly good”; 5 = “Very good.”

Personal Economic Performance. “In general, how would you describe your own
present living conditions?” 1: “Very bad”; 2 = “Fairly bad”; 3 = “Neither good nor
bad”; 4 = “Fairly good”; 5 = “Very good.”

Interest in Politics. “How interested are you in politics and government?” 1 = “Not at
all interested”; 2 = “Not very interested ”; 3 = “Somewhat interested”; 4 = “Very
interested.”

Gender. 0 = Male; 1 = Female.

Education. 1 = No formal schooling; 2 = Primary school; 3 = Middle school; 4 =
College or university; 5 = Post-graduate.

Descriptive Statistics

 

 

Variables Obs. Mean Std. Dev. Min Max
Satisfaction with Democracy 59610 2.43 0.926 1 4
Loser 63663 0.22 0.41 3 0 1
Non-partisan 63663 0.47 0.499 0 1
Interest in Politics 60524 2.45 1.1 12 0 1
National Economic Performance 62821 2.48 1.056 1 5
Personal Economic Performance 63299 2.82 0.993 1 5
Gender 63108 0.50 0.500 0 1
Age 62965 37.80 1 5.085 1 6 1 05
Education 62376 2.74 0.91 5 1 5

 

158

APPENDIX B

Coding

Satisfaction with Democracy. “Overall, how satisfied are you with the way democracy
works in Lesotho?” Response categories included: 1 = “not at all satisfied”; 2 = “not
very satisfied”; 3 = “fairly satisfied”; 4 = “very satisfied.”

Trust in Political Institutions. “How much do you trust each of the following, or haven’t
you heard enough about them to say?” 0 = “not at all”; 1 = “a little bit”; “2 = “a lot”; 3
= “a very great deal.” The combined institutional trust scale ranges from O to 15.

Partisan Status. “Which party do you usually think of yourself as close to?” If party
choice matches with a government party (winner), if it matches with an opposition party
(loser), otherwise (non-partisan).

National Economic Performance. “In general, how would you describe the present
economic condition of Lesotho?” 1: “Very bad”; 2 = “Fairly bad”; 3 = “Neither good
nor bad”; 4 = “Fairly good”; 5 = “Very good.”

Personal Economic Performance. “In general, how would you describe your own
present living conditions?” 1= “Very bad”; 2 = “Fairly bad”; 3 = “Neither good nor
ba ”; 4 = “Fairly good”; 5 = “Very good.”

Interest in Politics. “How interested are you in politics and government?” 0 = “not
interested”; I = “somewhat interested ”; 2 = “very interested.” For the analysis, I
create a dummy variable coded as “1” including “Very interested” and “Somewhat
interested.” The other two responses are coded as “0.”

Descriptive Statistics

 

 

Variables Obs. Mean Std. Dev. Min Max
Satisfaction with Democracy 1808 2.589 1.183 1 4
Trust in Political Institutions 2377 6.653 3.817 0 15
Loser 2377 0.171 0.377 0 1
Non-partisan 2377 0.380 0.486 0 1
Interest in Politics 2377 0.806 0.396 0 1
National Economic Performance 2377 1.863 1.088 1 5
Personal Economic Performance 2377 2.006 1.008 1 5
Gender 2377 0.497 0.5 0 1
Age 2377 43.025 17.376 18 92
Education 2377 2.659 1 .593 0 9

 

159

REFERENCES

Afrobarometer Netwok. 2004. “Afrobarometer Round 2: Compendium of Comparative
Results from a 15-country Survey.” Afrombaromerter Working Paper No.
34.

Almond, Gabriel A., and Sidney Verba. 1963. The Civic Culture: Political Attitudes
and Democracy in Five Nations. Princeton: Princeton University Press.

Amoretti, Ugo, and Nancy Bermeo (ed.). 2004. Federalism and Territorial Cleavages.
Baltimore: Johns Hopkins University Press.

Anderson, Christopher J. 1995. Blaming the Govemment: Citizens and Economy in Five
European Democracies. New York: M. E. Sharpe.

Anderson, Christopher J. 1998. “Parties, Party Systems, and Satisfaction with
Democratic Performance in The New Europe,” Political Studies 46: 572-
588.

Anderson, Christopher J ., and Christine A. Guillory. 1997. “Political Institutions and

Satisfaction with Democracy,” American Political Science Review 91
(March): 66-81.

Anderson, Christopher J., and Andrew LoTempio. 2002. “Winning, Losing, and
Political Trust in America,” British Journal of Political Science 32
(2):335-351.

Anderson, Christopher, and Yuliya V. Tverdova. 2003. “Corruption, Political
Allegiances, and Attitudes Toward Government in Contemporary
Democracies.” American Journal of Political Science 47: 91-109

Anderson, Christopher, Andre Blais, Shaun Bowler, Todd Donovan, and Ola Listaug.
2005. Losers’ Consent: Elections and Democratic Legitimacy. New York:
Oxford University Press.

Aslund, A. 1994. “The Care for Radical Reform.” Journal of Democracy 5: 63-74.
Balcerowicz, L. 1995. “Understanding Postcmmunist Transitions.” In Economic Reform

and Democracy, ed. L. Diamond and M.F. Plattner. Baltimore: Johns
Hopkins University Press.

Bowler, Shaun, and Todd Donovan. 2002. “Democracy, Institutions, and Attitudes

about Citizen Influence on Government,” British Journal of Political
Science 32 (2): 371-390.

160

Bratton, Michael, and Robert Mattes. 2001. “Support for Democracy in Africa: Intrinsic
or Instrumental.” British Journal of Political Science 43: 447-474.

Bratton, Michael, Robert Mattes, and E. Gyimah-Boadi. 2005. Public Opinion,
Democracy, and Market Reform in Africa. New York: Cambridge
University Press.

Bratton, Michael, and Van de Walle. 1997. Democratic Experiments in Africa. New
York: Cambridge University Press.

Canache, Damarys, Jeffry J. Mondak, and Mitchell A. Seligson. 2001. “Meaning and
Measurement in Cross-National Research on Satisfaction with
Democracy.” Public Opinion Quarterly 65: 506-528.

Cheibub, Jose Antonio. “Presidentialism and Democratic Performance.” In Andrew
Reynolds (ed.) The Architecture of Democracy. New York: Oxford
Univerity Press.

Clarke, Harold. D. and A. Kornberg. 1992. “Support for the Canadian Federal
Progressive Conservative Party since 1988: the Impact of economic
evaluations and economic issues,” Canadian Journal of Political Science

25: 29-53.

Clarke, Harold D., Nitish Dutt, and Allan Kornberg. 1993. “The Political Economy of
Attitudes Toward Polity and Society in Western European Democracies.”
Journal of Politics 55(4): 998- 1021 .

Cox, Gary W. 1997. Making Votes Count: Strategic Coordination in the World ’s
Electoral Systems. Cambridge: Cambridge University Press.

Crepaz, Markus M. 1996. “Constitutional Structures and Regime Performance in 18
Industrialized Democracies.” European Journal of Political Research 29:
87-104.

Delury, George. Eds. 1999. World Encyclopedia of Political Systems and Parties. New
York: Facts on File.

Derbyshier, Denis, and Ian Derbyshier. 1996. Political Systems of the World. New
York: St. Martin’s Press.

Dhal, R.A. 1989. Democracy and its Critics. New Haven: Yale University Press.

Easton, David. 1965. A Systems Analysis of Political Life. New York: Wiley.

161

 

Easton, David. 1975. “A Re-Assessment of the Concept of Political Support,” British
Journal of Political Science 5: 435-457.

Elklit, J. 2002. “Lesotho 2002: Africa’s First MMP Elections” Journal of African
Elections. Vol. 1: 1-10.

Elazar, Daniel J. 1997. “Contrasting Unitary and Federal Systems.” International
Political Science Review 18: 237-251.

Englund, Harri. 2002. “Winning Elections, Losing Legitimacy: Multi-partyism and the
Neopartimonial State in Malawi.” In Multi-party Elections in Africa, ed.
Michael Cowen and Liisa Laakso. New York: Palgrave.

Evan, Geoffrey, and Stephen Whitefield. 1995. “The Politics and Economics of
Democratic Commitment: Support for Democracy in Transition Societies.”
British Journal of Political Science 25: 485-514.

Farrell, David, and Ian McAllister. 2003. “Voter Satisfaction and Electoral Systems:
Does Preferential Voting in Candidate-Centered Systems Make a
Difference?” Paper presented to the Australasian Political Studies
Association Conference 2003.

Fox, R. and Roger Southall. 2002. “Adapting to Electoral System Change: Voters in
Lesotho 2002.” Journal of African Elections. Vol. 2: 86-96.

Fuchs, Dieter. 1993. “Trends of Political Support.” In Political Culture in Germany ed.
D. Berg-Schlosser and R. Rytlewski. New York: St. Martin’s

Fuchs, Dieter. 1999. “The Democratic Culture of Unified Germany.” In Critical
Citizens. ed. Pippa Norris. New York: Oxford University Press.

Fuchs, Dieter, Giovanna Guidorossi, and Palle Svensson. 1995. “Support for the
Democratic System.” In Hans-Dieter Klingemann and Dieter Fuchs, eds.,
Citizens and the State. New York: Oxford University Press.

Gay, John, and Robert Mattes. 2004. “The State of Democracy in Lesotho: A report on
the 2003 Afrobarometer Survey.” Afrobarometer Working Paper No. 32.

Avaiable at http://www.afrobarometerorgz

Gerrin g, John, and Strorn C. Thacker. 2004. “Political Institutions and Corruption: The
Role of Unitarism and Parliamentarism.” British Journal of Political
Science 34: 295-330.

Gibson, James L. 1998. “A Sober Second Thought: An Experiment in Persuading
Russians to Tolerate.” American Journal of Political Science 42: 819-850.

162

 

Golden, Miriam. 2003. “Electoral Connections: The Effects of the Personal Vote on
Political Patronage, Bureaucracy and Legislation in Postwar Italy.” British
Journal of Political Science 33: 189-212.

Hall, Peter A., and Rosemary C. R. Taylor. 1996. “Political Science and the Three New
Institutionalisms.” Political Studies XLIV: 936-957.

Harmmel, Robert, and John D. Robertson. 1986. “Government Stability and Regime
Support: A Cross-National Analysis.” Journal of Politics 48: 1029-1040.

Holmberg, Soren. 1999. “Down and Down We Go: Political Trust in Sweden.” In
Critical Citizens. ed. Pippa Norris. New York: Oxford University Press.

Horowitz, Donald. 1991. A Democratic South Africa? : Constitutional Engineering in a
Divided Society. Berkeley: University California Press.

Huntington, Samuel. 1968. Political Order in Changing Societies. New Haven: Yale
University Press.

Huntington, Samuel. 1991. The Third Wave: Democratization in the Late Twentieth
Century. Norman: University of Oklahoma Press.

Immergut, Ellen M. 1998. “The Theoretical Core of the New Institutionalism.” Politics
& Society 26: 5-34.

Inglehart, Ronald. 1977. The Silent Revolution: Changing Values and Political Styles
Among Western Public. Princeton: Princeton University Press.

In glehart, Ronald. 1990. Culture Shift in Advanced Industrial Society. Princeton:
Princeton University Press.

Inglehart, Ronald. 1997. Modernization and Postrnodemization: Cultural, Economic,

and Political Change in 43 Societies. Princeton: Princeton University
Press.

Inglehart, Ronald. 2003. “How Solid is Mass Support for Democracy — And How Can
WE Measure It?” Political Science and Politics 36: 51-57.

Jackman, Robert W. 1987. “Political Institutions and Voter Turnout in the Industrial
Democracies.” American Political Science Review 81 (2): 405-424.

Jones, Mark. 1995. Electoral Laws and the Survival of Presidential Democracies. Notre
Dame: University of N otre Dame Press.

Kasse, Max, and Kenneth Newton. 1995. Beliefs in Government. New York: Oxford
University Press.

163

 

Katz, Richard. 1980. A Theory of Parties and Electoral Systems. Baltimore: Johns
Hopkins University Press.

King, Gary, Robert O. Keohane, and Sidney Verba. .1994. Designing Social Inquiry:
Scientific Inference in Qualitative Research. Princeton: Princeton
University Press.

Klingermann, Hans-Dieter. 1999. “Mapping Political Support in the 19903: A Global
Analysis.” In Critical Citizens. ed. Pippa Norris. New York: Oxford
University Press.

Klingemann, Hans-Dieter and Dieter Fuchs. ed. 1995. Citizens and the State. New
York: Oxford University Press.

Kornberg, A. and H. D. Clarke. 1994. “Beliefs about democracy and satisfaction with
democratic government: the Canadian Case.” Political Research Quarterly
47: 537-563.

Lagos, Marta. 2001. “How People View Democracy: Between Stability and Crisis in
Latin America.” Journal of Democracy 12: 137-145.

Laakso, Liisa. 2002. “When Elections are Just a Formality: Rural-Urban Dynamics in
the Dominant-Party System of Zimbabwe.” In Multi-party Elections in
Africa, ed. Michael Cowen and Liisa Laakso. New York: Palgrave.

Lambert, Robert D., James E. Curtis, Steven D. Brown, and Barry J. Kay. 1986.
“Effects of Identification with Governing Parties on Feeling of Political
Efficacy and Trust.” Canadian Journal of Political Science 19
(December): 705-728.

Landman, Todd. 2000. Issues and Methods in Comparative Politics. New York:
Routledge.

LeDuce, Lawrence, Richard G. Niemi, and Pippa Norris. Eds. 2002. Comparing
Democracies 2. Thousand Oaks: Sage.

Lijphart, Arend. 1971. “Comparative Politics and Comparative Method.” American
Political Science Review 39: 523-541.

Lijphart, Arend. Eds. 1992. Parliamentary versus Presidential Government. Oxford:
Oxford University Press.

Lijphart, Arend. 1994. Electoral Systems and Party Systems: A Study of Twenty—Seven
Democracies, 1945-1990. Oxford: Oxford University Press.

164

 

Lijphart, Arend. 1999. Patterns of Democracy: Government Forms and Performance in
Thirty-Six Countries. New Haven: Yale University Press.

Linde, J. and J. Ekman. 2003. “Satisfaction with Democracy: A Note on a Frequently
Used Indicator in Comparative Politics.” European Journal of Political
Research 42: 391-408.

Linz, Juan J. 1994. “Presidential or Parliamentary Democracy: Does it Make a
Difference?” In The Failure of Presidential Democracy. ed. Juan J. Linz,
and Arturo Valenzuela. Baltimore: Johns Hopkins University Press.

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

Linz, Juan J ., and Alfred Stepan. 1996. Problems of Democratic Transition and
Consolidation: Southern Europe, South America, and Post-Communist
Europe. Baltimore: Johns Hopkins University Press.

Lipset, Seymour Martin. 1959. “Some Social Requisities of Democracy: Economic
Development and Political Legitimacy.” American Political Science
Review 53(1): 69-105.

Listhaug, Ola, and Matti Weiberg. 1995. “Confidence in Political and Private
Institutions.” In Hans-Dieter Klingemann and Dieter Fuchs, eds., Citizens
and the State. New York: Oxford University Press.

Lockerbie, B. 1993. “Economic Dissatisfaction and Political Alienation in Western
Europe.” European Journal of Political Research 23: 281-293.

Logan, Carolyn, Nansozi Muwanga, Robert Sentamu, and Michael Bratton. 2003.
“Insiders and Outsiders: Varying Perceptions of Democracy and
Governance in Uganda.” Afrobarometer Working Paper No. 27.

Long, J. Scott. 1997. Regression Models for Categorical and Limited Dependent
Variables. California: SAGE Publications.

Mainwaring, Scott. 1993. Presidentialism, Multipartism, and Democracy: The Difficult
Combination. Comparative Political Studies 26: 198-228.

Makoa, Francis K. 2002. Elections, Election Outcomes and Electoral Politics in
Lesotho. Pretoria: Africa Institute of South Africa.

Makumbe, John. 2002. “Zimbabwe’s Hijacked Election.” Journal of Democracy 13: 87-
102.

165

Marks, Gary, and Larry Diamond. Eds. 1992. Reexamining Democracy: Essays in
Honor of Seymour Martin Lipset. Newbury Park: Sage Publication.

Mattes, Robert, and Michael Bratton. 2003. “Learning About Democracy in Africa:
Awarness, Performance, and Experience.” Afrobarometer Paper.

McAllister, Ian. 1999. “The Economic Performance of Governments.” In Critical
Citizens. ed. Pippa Norris. New York: Oxford University Press.

McKelvey, RD, and W. Zavoina. 1975. “A Statistical Model for the Analysis of
Ordinal Level Dependent Variables.” Journal of Mathematical Sociology
4: 103-120.

Meyns, Peter. 2002. “Cape Verde: An African Exception.” Journal of Democracy 13:
153- 165.

Miller, A.H., and O. Listhaug. 1990. “Political Parties and Confidence in Government:

A comparison of Norway, Sweden and the United States.” British Journal
of Political Science 29: 357-386.

Miller, Arthur, and Ola Listhaug. 1999. “Political Performance and Institutional Trust.”
In Critical Citizens. ed. Pippa Norris. New York: Oxford University Press.

Mishler, William, and Richard Rose. 2001. “What Are The Origins of Political Trust?
Testing Institutional and Cultural Theories in Post-Communist Societies.”
Comparative Political Studies 34: 30-62.

Morlino, Leonardo, and Marco Tarchi. 1996. “The Dissatisfied Society: The Roots of

Political Change in Italy.” European Journal of Political Research 30: 41-
63.

Nadeau, Richard, and Andre Blais. 1993. “Accepting the Election Outcome: The Effect
of Participation on Losers’ Consent.” British Journal of Political Science

23 (October): 553-563.

Nohlen, Deiter, Michael Krennerich, and Bernhard Thinaut. Eds. 1999. Elections in
Africa. New York: Oxford University Press.

Norris, Pippa, 1999. “Institutional Explanations for Political Support.” In Critical
Citizens. ed. Pippa Norris. New York: Oxford University Press.

Olsen, Mancur. 1982. The Rise and Decline of Nations. New Haven: Yale University
Press.

166

Ordeshook, Peter, and Olga Shvetsova. 1994. “Ethnic Heterogeneity, District
Magnitude, and the Number of Parties.” American Journal of Political
Science 38: 100-123.

Patterson, Samuel, and Anthony Mugham. Eds. 1999. Senates: Bicameralism in the
Contemporary World. Columbus: Ohio State University Press.

Penfold-Becerra, Michael. 2004. “Electoral Dynamics and Decentralization in
Venezuela” In Decentralization and Democracy in Latin America, ed.
Alfred P. Montero and David J. Samuels. Notre Dame: University of Notre
Dame Press.

Persson, T., G. Roland, and G. Tabellini. 1997. “Separation of Powers and Political
Accountability.” Quarterly Journal of Economics 112: 310-327.

Persson, T., G. Roland, and G. Tabellini. 2000. “Comparative Politics and Public
Finance.” Journal of Political Economy 108: 1121-1161.

Powell, G. Bingham. 1982. Contemporary Democracies: Participation, Stability and
Violence. Cambridge: Cambridge University Press.

Powell, G. Bingham. 1989. “Constitutional Design and Citizen Electoral Control.”
Journal of Theoretical Politics 1 (April): 107-130.

Powell, G. Bingham. 2000. Elections as Instruments of Democracy. New Haven: Yale
University Press.

Power, Timothy J ., and Mark J. Gasiorowski. 1997. “Institutional Design and
Democratic Consolidation in the Third World. Comparative Political
Studies30: 123-155.

Przeworski, Adam. 1987. “Methods of Cross-National Research, 1970-1983,” in M.
Dierkes, H.N.Weiler and A.B.Antal, eds., Comparative Policy Research.
Aldershot: Gower.

Przeworski, Adam, Michael E. Alvarez, Jose Antonio Cheibub, and Fernando Limongi.
2000. Democracy and Development: Political Institutions and Well-Being
in the World, 1 950-1 990. Cambridge: Cambridge University Press.

Putnam, Robert D. 1993. Making Democracy Work: Civic Traditions in Modern Italy.
Princeton: Princeton University Press.

Rae, Douglas W. 1967. The Political Consequences of Electoral Laws. New Haven:
Yale University Press.

167

Raudenbush, Stephen W., and Anthony S. Bryk. 2002. Hierarchical Linear Models 2nd.
Thousand Oaks: Sage Publications, Inc.

Reynolds, Andrew. 1999. Electoral Systems and Democratization in Southern Africa.
New York: Oxford University Press.

Reynolds, Andrew. Eds. 2002. The Architecture of Democracy. New York: Oxford
University Press.

Reynodls, Andrew, and Ben Reilly. 1997. The International IDEA Handbook of
Electoral System Design. Stockholm: International Institute for Democracy
and Electoral Assistance.

Riker, William H. 1975. “Federalism.” In Fred J. Greenstein and Nelso Polsby (ed.)
Handbook of Political Science. Vol.5, Governmental Institutions and
Processes. Reading: Addison-Wesley.

Rohrchneider, Robert. 2002. “The Democracy Deficit and Mass Support for an EU-
Wide Government.” American Journal of Political Science 46: 463—475.

Romero, Anibal. 1997. “Rearranging the Deck Chairs on the Titanic: The Agony of
Democracy in Venezuela.” Latin American Research Review 32: 7-36.

Rose, Richard, William Mishler, and Christian Haerpfer. 1998. Democracy and Its
Alternatives. Baltimore: Johns Hopkins University Press.

Rule, Stephen P. 2000. Electoral Territoriality in Southern Africa. Aldershot: Ashgate.
Sartori, Giovanni. 1968. “Political Development and Political Engineering.” In John D.

Montgomery and Albert 0. Hirschman, eds., Public Policy. Cambridge:
Cambridge University Press.

Shin, Doh Chull. 1999. Mass Politics and Culture in Democratizing Korea. New York:
Cambridge University Press.

Shugart, Matthew Soberg, and John M. Carey. 1992. Presidents and Assemblies:
Constitutional Design and Electoral Dynamics. Cambridge: Cambridge
University Press.

Shugart, Matthew, and Martin P. Wattenberg (ed.). 2003. Mixed-Member Electoral
Systems: The Best of Both Worlds? New York: Oxford University Press.

Sisk, Timothy D., and Andrew Reynolds (ed.). 1998. Elections and Conflict
Management in Africa. Washington: United States Institute of Peace Press.

168

Southall, Roger. 2003. “An Unlikely Success: South Africa and Lesotho’s Election of
2002.” The Journal of Modern African Studies. Vol. 41: 269-296.

Southall, Roger, and Roddy Fox. 1999. “Lesotho’s General Election of 1998: Rigged
or De Riguer?” The Journal of Modern African Studies 37 (4): 669-696.

Steenbergen, Marco R., and Bradford S. Jones. 2002. “Modeling Multilevel Data
Structures.” American Journal of Political Science 46(Jan.): 218-237.

Stepan, Alfred, and Cindy Skach. 1993. Constitutional Frameworks and Democratic
Consolidation: Parliamentarism versus Presidentialism. World Politics 46:
1-22.

Taagepera, Rein. 2002. “Designing Electoral Rules and Waiting for an Electoral System
to Evolve.” In Andrew Reynolds (ed.), The Architecture of Democracy:
Constitutional Design, Conflict Management, and Democracy. New York:
Oxford University Press.

Taagepera, Rein, and Mattew Soberg Shugart. 1989. Seats and Votes: The Effects and
Determinants of Electoral Systems. New Haven: Yale University Press.

Tabellini, Guido. 2000. “Constitutional Determinants of Government Spending.”
Innocenzo Gasparini Institute for Economic Research, Bocconi University.
Working Paper.

Tanzi, Vito. 1996. “Fiscal Federalism and Efficiency: A Review of Some Efficiency
and Macroeconomic Aspects.” In M. Bruno and B. Pleskovic (ed.), Annual

World Bank Conference on Development Economics 1995. Washington:
World Bank.

The Afrobarometer Network. 2004. “Afrobarometer Round 2: Compendium of Results
from 15-country Survey. Afrobarometer Working Paper No. 34,
www.afrobarometer.org.

Toka, Gabor. 1995. “Political Support in East-Central Europe.” In Citizens and the
State, ed. Hans-Dieter Klingemann and Dieter Fuchs. New York: Oxford
University Press.

Tomz, Michael, Jason Wittenberg, and Gary King. 2003. CLARIFY: Software for
Interpreting and Presenting Statistical Results. Version 2.1. Stanford
University, University of Wisconsin, and Harvard University. January 5.

Avaiable at http://gkingharvardedu/

Tsebelis, George, and Jeannette Money. 1997. Bicameralism. Cambridge: Cambridge
University Press.

169

Van de Walle, Nicolas. 2003. “Presidentialism and Clientelism in Africa’s emerging
party systems.” Journal of Mordem African Studies 41: 297-321.

Wade, Robert. 1985. “The Market for Public Office: Why the Indian State Is Not Better
at Development.” World Development 13: 467-497.

Warren, Mark, ed. 1999. Democracy and Trust. Cambridge: Cambridge University
Press.

Watts, Ronald. 1997. Comparing Federal Systems in the I990s. Montreal: McGill-
Queen’s University Press.

Weatherford, Stephen M. 1984. “Economic ‘Stagflation’ and Public Support for the
Political System.” British Journal of Political Science 14: 187-205.

Weatherford, Stephen M. 1987. “How Does Government Performance Influence
Political Support?” Political Behavior 9: 5-28.

Weatherford, M. Stephen. 1991. “Mapping the Ties that Bind: Legitimacy,

Representation and Alienation.” Western Political Quarterly 44 (J une):
251-276.

Weil, Frederick. 1989. “The Sources and Structure of Legitimation in Western
Democraceis: A Consolidated Model Tested with Time-Serises Data in Six
Countries Since World War H,” American Sociological Review 54 (Oct.):
687-706.

Weisberg, Jacob. 1996. In Defence of Govemment: The Fall and Rise of Public Trust.
New York: Scribner.

Weisfelder, Richard. 1999. “Why Fair Elections Get Fouled in Lesotho: A Search for
Solutions.” The Journal of African Policy Studies 5: 109-132.

Weldon, J.A. 1997. “The Political Sources of Presidentialismo in Mexico.” In
Presidentialism andDemocracy in Latin America, ed. S. Mainwaring and
M.S.Shugart. Cambridge: Cambridge University Press.

Western, Bruce. 1998. “Causal Heterogeneity in Comparative Research: A Bayesian
Hierarchical Modeling Approach.” American Journal of Political Science
42: 1233-1259.

Winship, C., and R. D. Mare. 1984. “Regression Models with Ordinal Variables.”
American Sociological Review 49: 512-525.

170

 

Wolf, Thomas, Carolyn Logan, and Jeremiah Owiti. 2004. “A New Dawn? Popular
Optimism in Kenya After the transition.” Afrobarometer Working Paper
No. 33.

World Bank. 2004. African Development Indicators, 2004. Washington, DC: World
Bank.

World Bank. 2005. World Development Indicators, 2005. Washington, DC: World
Bank.

Zorn, Christopher J .W. 2001. “Generalized Estimating Equation Models for Correlated

Data: A Review with Applications.” American Journal of Political Science
45: 470—90.

171

   

l«illiiiiilliiililliili