Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
This page intentionally left blank
July 5, 2008
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
Often dismissed as window dressing, nominally democratic institutions,
such as legislatures and political parties, play an important role in nondemocratic regimes. In a comprehensive cross-national study of all nondemocratic states from 1946 to 2002 that examines the political uses
of these institutions by dictators, Jennifer Gandhi finds that legislative
and partisan institutions are an important component in the operation
and survival of authoritarian regimes. She examines how and why these
institutions are useful to dictatorships in maintaining power. In their
efforts to neutralize threats to their power and to solicit cooperation
from society, autocratic leaders use these institutions to organize concessions to potential opposition. The use of legislatures and parties to
co-opt opposition results in significant institutional effects on policies
and outcomes under dictatorship.
Jennifer Gandhi is an assistant professor in the Department of Political Science at Emory University. She received her PhD in comparative
politics and political economy from New York University in 2004. Her
research interests include the political institutions dictators use to maintain power and the strategies the opposition may adopt in response. Her
work has appeared in Economics and Politics and Comparative Political Studies.
i
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
ii
July 5, 2008
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
Political Institutions under
Dictatorship
JENNIFER GANDHI
Emory University
iii
July 5, 2008
14:38
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo
Cambridge University Press
The Edinburgh Building, Cambridge CB2 8RU, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521897952
© Jennifer Gandhi 2008
This publication is in copyright. Subject to statutory exception and to the
provision of relevant collective licensing agreements, no reproduction of any part
may take place without the written permission of Cambridge University Press.
First published in print format 2008
ISBN-13
978-0-511-43693-2
eBook (EBL)
ISBN-13
978-0-521-89795-2
hardback
Cambridge University Press has no responsibility for the persistence or accuracy
of urls for external or third-party internet websites referred to in this publication,
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Permission to reprint material is gratefully acknowledged. Portions of Chapter 3
appear in Gandhi, Jennifer, and Adam Przeworski, “Cooperation, Cooptation,
and Rebellion under Dictatorship,” 2006, Economics and Politics 18, 1: 1–26;
portions of Chapter 5 appear in Gandhi, Jennifer, “Dictatorial Institutions and
Their Impact on Economic Growth,” forthcoming, European Journal of
Sociology; and portions of Chapter 6 appear in Gandhi, Jennifer, and Adam
Przeworski, “Authoritarian Institutions and the Survival of Autocrats,” 2007,
Comparative Political Studies 40, 11: 1279–1301.
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
To my parents and Neil
v
July 5, 2008
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
vi
July 5, 2008
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Contents
List of Tables and Figures
page xi
Acknowledgments
xiii
Introduction
0.1 The Argument
0.2 The Study of Institutions in Dictatorships
0.3 Plan of the Book
xv
xvii
xix
xxi
1
2
The World of Dictatorial Institutions
1.1 Introduction
1.2 What Is Dictatorship?
1.2.1 Historical Usage
1.2.2 Contemporary Controversies
1.3 Who Are the Dictators?
1.3.1 Monarchs
1.3.2 Military Dictators
1.3.3 Civilian Dictators
1.3.4 Operationalization of Dictatorial Types
1.4 Nominally Democratic Institutions
1.4.1 Legislatures
1.4.2 Political Parties
1.5 Conclusion
Three Illustrative Cases
2.1 Introduction
2.2 Kuwait: Strength without Institutions
2.2.1 Sabah Consolidation of Power
2.2.2 Merchants and the Push for Institutions
1
1
2
3
7
12
21
25
29
31
34
34
36
39
42
42
44
44
46
vii
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
Contents
viii
2.3
2.4
2.5
3
4
July 5, 2008
2.2.3 The Majlis Movement and the Regime’s
Response
Morocco: Survival with Institutions
2.3.1 The Rise of Monarchical and Nationalist
Forces
2.3.2 Managing State-Building and Opposition
2.3.3 The Battle over Institutions
Ecuador: The Perils of Noninstitutionalization
2.4.1 Prelude to Military Rule
2.4.2 The Military’s Agenda
2.4.3 The Rise of Opposition
2.4.4 Underinstitutionalization and Its Consequences
Conclusion
Use of Institutions to Co-opt
3.1 Introduction
3.2 Dictatorial Concessions
3.3 Dictatorial Institutions
3.4 Actors and Preferences
3.5 Timing
3.6 Results: Three Equilibria
3.6.1 Cooperation Equilibrium
3.6.2 Co-optation Equilibrium
3.6.3 Turmoil Equilibrium
3.7 Predicting Dictatorial Institutions
3.7.1 Data
3.8 Conclusion
3.9 Appendix
3.9.1 Solutions
3.9.2 Comparative Statics
Institutions and Policies under Dictatorship
4.1 Introduction
4.2 Which Policies?
4.3 Civil Liberties
4.3.1 Data
4.3.2 Effect of Institutions
4.4 Military Expenditures
4.4.1 Data
4.5 Social Spending
4.5.1 Data
4.5.2 Effect of Institutions
4.6 Conclusion
49
52
52
55
58
61
61
64
66
68
71
73
73
74
77
82
86
87
87
88
90
92
92
100
101
101
105
107
107
110
116
117
122
126
127
132
133
135
137
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
Contents
5
July 5, 2008
ix
139
139
142
146
147
148
149
150
151
152
7
Institutions and Outcomes under Dictatorship
5.1 Introduction
5.2 Institutional Mechanisms for Economic Development
5.3 Alternative Theoretical Frameworks
5.3.1 Mobilization Power of Single-Party Regimes
5.3.2 State Autonomy
5.3.3 Importance of Noninstitutionalized Factors
5.4 Institutions and Economic Growth
5.4.1 Model Specification
5.4.2 Data
5.4.3 Effect of Institutions: Results from Random
Effects Models
5.4.4 Effect of Institutions: Results from Heckman
Selection Model
5.4.5 Conclusion
Institutions and the Survival of Dictators
6.1 Introduction
6.2 Managing Political Survival
6.3 Impact of Institutions on Survival
6.3.1 Event History Analysis
6.3.2 Data
6.3.3 Impact of Institutions on Survival
6.4 Conclusion
Conclusion
8
Codebook of Variables
189
6
154
157
160
163
163
165
169
169
171
175
177
180
Bibliography
195
Author Index
Subject Index
215
221
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
x
July 5, 2008
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
June 25, 2008
List of Tables and Figures
tables
1.1.
1.2.
1.3.
1.4.
1.5.
1.6.
3.1.
3.2.
3.3.
4.1.
4.2.
4.3.
4.4.
4.5.
4.6.
4.7.
4.8.
4.9.
5.1.
5.2.
Countries under Dictatorship, 1946–2002
page 13
Replacement of Dictators by Type
20
Postwar Monarchs and Their Exit from Power
24
Types of Dictators by Region
34
Legislatures under Dictatorship by Method of Selection
35
Number of Dictatorial Political Parties
39
Institutions by Type of Dictator
94
Descriptive Statistics of Variables in Institutions Analysis
97
Explaining Dictatorial Institutions
98
Transition Probabilities of Institutional Arrangements
113
Descriptive Statistics of Variables in Civil Liberties
Analysis
122
Effect of Institutions on Civil Liberties
123
Predicted Probabilities for Speech Rights
124
Predicted Probabilities of Workers’ Rights Conditional
on Institutions
125
Descriptive Statistics of Variables in Military Spending
Analysis
128
Effect of Institutions on Military Expenditures
130
Descriptive Statistics of Variables in Social Spending
Analysis
134
Effect of Institutions on Social Spending
135
Descriptive Statistics of Variables in Economic Growth
Analysis
155
Effect of Institutions on Economic Growth
156
xi
9:14
Gandhi_Revised_Dec07.tex
xii
cuus282-gandhi.cls
978 0 521 89795 2
June 25, 2008
List of Tables and Figures
5.3.
6.1.
6.2.
6.3.
Heckman Second-Stage Growth Model
Replacement of Dictators by Type
Life Table of Dictators in the Sample
Descriptive Statistics of Variables in Event History
Analysis
6.4. Effect of Institutions on Political Survival of Dictators
159
166
172
175
177
figures
1.1.
1.2.
1.3.
3.1.
3.2.
Dictatorships in the world, 1946–2002
Dictatorial legislatures by year and method of selection
Number of dictatorial parties by year
Logic of dictatorial concessions
“Cooperation” policy concessions, γc , under different
degrees of policy polarization, θ 2
3.3. “Co-optation” policy concessions, γz , under different
degrees of policy polarization, θ 2
3.4. Three equilibria in (q, −L) space
11
37
40
86
88
89
91
9:14
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Acknowledgments
My intellectual debts are many, but first and foremost are my dues
to Adam Przeworski. From him, I have learned much, but the lessons
that stay with me are the ones he taught by example: to aspire to
ask big questions and to embrace the process of retooling in order to
answer those questions. I thank others from whom I have learned
valuable lessons: Bernard Manin for his carefulness, John Roemer
for his clarity, and Stephen Holmes for his boldness. In addition, my
thanks to Leonard Wantchekon, Youssef Cohen, Jonathan Nagler,
and Kanchan Chandra for their encouragement at various stages in
the process.
My time in New York was enriched by a cohort in graduate school
that spanned many years and countries. I thank the members of the
Thursday group who provided much criticism, made easier to swallow
with much laughter: Suzy Fry, Matt Golder, Sona Nadenichek Golder,
Wonik Kim, Covadonga Meseguer, and Sebastian Saiegh. Wonik and
Sebas, in particular, were loyal companions who provided advice that
made me a better scholar and, often, a better person. I am grateful
to Carmela Lutmar and Jeff Lax for their support. José Cheibub and
Jim Vreeland deserve special thanks for their suggestions on my work
and their counsel on becoming a scholar in one’s own right. Finally, I
consider myself lucky to have been part of the melting pot that came
together in the Department of Politics at New York University: My
thanks to Despina Alexiadou, Tamar Asadurian, Dulce Manzano, and
Julio Rı́os Figueroa.
xiii
14:38
Gandhi_Revised_Dec07.tex
xiv
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Acknowledgments
Since graduate school, the political science department at Emory
University has provided a supportive environment. I thank members
of the Political Economy Workshop and, in addition, Dan Reiter, Tom
Remington, and Jeff Staton for their feedback on my work. For their
encouragement and advice, I am in debt to Cliff Carrubba, David
Davis, Rick Doner, and Alex Hicks. My thanks to Amy Liu and Jeffrey Kucik for their research assistance. In addition, the Stanford Comparative Politics Workshop, Jason Brownlee, Bill Keech, and Beatriz
Magaloni provided helpful feedback on either parts of or the whole
manuscript. Finally, my thanks to Eric Crahan at Cambridge University Press for his heartening interest in the project and his steady
guidance in its completion.
Through this experience, I learned that writing a book is at times a
gratifying experience and, at other times, an excruciating one. I owe
many thanks to Michael Owens, Maisha Fisher, Tavishi Bhasin, Leslie
Davis, and Ravish Bhasin for helping me remember that the tempest
should stay in the teapot. I also thank Joanne Fox-Przeworski for her
endless hospitality and kindness.
Life in New York was a formative moment in no small part due
to three people without whom my life would have been all that
much poorer. A constant companion since graduate school, Melissa
Schwartzberg always has shown a generosity of spirit in all things that
helped me complete this book and so much more. In addition, I thank
Carol Hsu and Carissa Montgomery, the best of friends who always
forgave my absences and allowed me to pick up our friendship wherever it left off. I am grateful to the three of them for their enduring
support and for making me feel at home whenever I am back in the
city.
Finally, I thank Neil for his patience with an itinerant sister and his
willingness to close the gap as we grow older. His sincerity and sense
of humor help me keep things in perspective. In addition, heartfelt
thanks to my parents, Josephine and Niranjan, for encouraging us to
do the best we can in all things and for indulging me my intellectual
curiosities. This book is dedicated to the three of them.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Introduction
Why do nondemocratic rulers govern with democratic institutions,
such as legislatures and political parties? One view is that these institutions under dictatorship are mere shams. Scholars and policy-makers
alike have pronounced the irrelevance of formal institutions under
dictatorship. In discussing the role of political institutions on regime
change, Gasiorowski (1995: 883) writes: “Huntington’s (1968) argument about the importance of institutionalization also applies under
authoritarian regimes, but consociationalism, party system structure,
electoral rules, and the type of executive system are largely irrelevant
and therefore presumably have little effect. . . . ” A USAID report (n.d.:
1), in describing communist regimes in Eastern Europe and Central
Asia, states more bluntly: “Elections were a sham. Parliaments had no
real power. Basic democratic freedoms – free speech, the freedom to
assembly and organize, the right to form independent parties did not
exist.” The conclusion is clear: nominally democratic institutions constitute mere window dressing that dictators can point to as evidence
of their democratic credentials.
Yet those who encourage the formation of these institutions in
the interests of promoting democracy imply another view. As Jeane
Kirkpatrick (1979: 37) observed: “democratic governments have come
into being slowly, after extended prior experience with more limited
forms of participation during which leaders have reluctantly grown
accustomed to tolerating dissent and opposition. . . . ” As a result, nongovernmental organizations, such as the National Democracy Initiative
xv
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
xvi
July 5, 2008
Introduction
for International Affairs (NDI), provide countries with “. . . assistance
in building their democratic structures. These include: national legislatures and local governments that function with openness and competence; broad-based political parties that are vehicles for public policy
debates; and nonpartisan civic organizations that promote democratic
values and citizen participation.”1
In this view, semiautonomous parties should provide political leaders and followers the opportunity to learn and practice the “civic
culture.” Electoral contests and legislative debates should enable opposition forces to make progress, even if incremental, in liberalizing
the regime. The hope, of course, is that liberalization sets the stage
for democratic transitions, even if with disappointment; we have witnessed enough instances in which this has not been the case. But if
this anticipated sequence of events motivates encouragement of elections, parties, and assemblies under nondemocratic rule, then it must
be the case that we believe these institutions serve as more than mere
ornamentation. Mere drapery cannot sow the seeds of destruction of
dictatorships.
The variation in dictatorial institutions is immense. During the post–
World War II period, the proportion of nondemocratic regimes with
legislatures varies from 60 to 88 percent. Legislatures are ubiquitous
in party dictatorships, but less so under military rule and monarchy.
More heterogeneity exists in the number of political parties tolerated
by authoritarian regimes, whether they are allowed simply to legally
exist or also to obtain seats within the legislature. The share of dictatorships in any given year that has banned parties ranges from 8 to 25
percent. Although the majority of nondemocracies have allowed for
multiple political parties to exist (58 percent, counting by country-year
observations), only in half of these cases are parties other than the one
organized by the regime permitted to obtain seats within a legislative
body.
The aggregate patterns are a reflection of some infamous examples. Communist dictatorships always have been organized around
the regime party or a front in which auxiliary parties are forced to
join alongside in an assembly. Lenin’s invention was copied by other
authoritarian incumbents, such as William Tubman of Liberia and
1
http://www.ndi.org/about/about.asp.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
Introduction
978 0 521 89795 2
July 5, 2008
xvii
Rafael Trujillo of the Dominican Republic, even if they did not import the ideological trappings. Other nondemocratic rulers, such as
Mohammed Mahathir in Malaysia and Anastasio Samoza in
Nicaragua, allowed for the formation of multiple parties that participated within legislatures. Still other dictators, such as Idi Amin in
Uganda and Augusto Pinochet in Chile, banned legislatures and parties
upon seizing power. Incumbents also may change their institutional arrangements like musical chairs. King Hussein, for example, closed and
reopened the Jordanian parliament four times. What becomes apparent is that the institutional variation in dictatorships is bounded by
neither geographic nor temporal considerations.
The two views of nominally democratic institutions under dictatorship, then, appear contradictory. One says that legislatures and
political parties are nothing more than window dressing with little expected effects for policies or outcomes, whereas, in another view, these
institutions are meaningful precursors for greater liberalization if not
more dramatic democratic change. Yet neither view can account for
the variation in the behavior of authoritarian rulers and their institutional choices. If these institutions are costless window dressing that
provides reputational benefits on the international stage, why do not
all dictators have them? In turn, if these institutions have the potential
to undermine autocratic rule, why would any incumbent create or tolerate them? Whether nondemocratic rulers should either promote or
shun these institutions, their behaviors should be consistent.
If parties do not compete and legislatures do not represent under
dictatorship, what is the purpose of these institutions? Are there systematic reasons for why some nondemocratic rulers govern with these
institutions, whereas others do not? Furthermore, if institutions are
the product of conscious choices, do they have effects on policies and
outcomes under dictatorship?
0.1
the argument
Dictators face two basic problems of governance. First, as rulers who
hold power without the legitimacy of having been chosen by their
citizens, they must prevent attempts to undermine their legitimacy and
usurp power. In other words, they must thwart challenges to their
rule. Second, autocrats also must solicit the cooperation of those they
14:38
Gandhi_Revised_Dec07.tex
xviii
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Introduction
rule. Even if their interests lie only in accumulating wealth and power,
incumbents will have more to amass if their countries are affluent and
orderly. Internal prosperity can be generated only if citizens contribute
their capital and their labor to productive activities. Autocrats, in other
words, need compliance and cooperation.
Yet the severity of the problems of ensuring compliance and generating cooperation vary across authoritarian incumbents. The degree to
which they face serious threats to their rule depends on the strength of
the potential opposition within society. Incumbents have more to fear
from a united, broad-based resistance movement. When the opposition
is weak – whether due to an unpopular ideology or collective action
problems – rulers have less need to manage outside groups. Similarly,
the degree to which dictators must solicit cooperation from citizens
to generate prosperity and rents for themselves differs. If rulers have
access to external sources of revenue, for example, they may rely less
on the cooperation of domestic groups for the creation of wealth.
To both thwart rebellion and solicit cooperation, dictators must
make concessions to outside groups. Concessions may come in the
form of rents; the dictator may agree to distribute some of his spoils
to certain segments of society as a solution to these two problems
of governance. Yet the potential opposition may demand more, and
incumbents may have to make policy concessions as well.
To organize policy compromises, dictators need nominally democratic institutions. Legislatures and parties serve as a forum in which
the regime and opposition can announce their policy preferences and
forge agreements. For the potential opposition, assemblies and parties provide an institutionalized channel through which they can affect
decision-making even if in limited policy realms. For incumbents, these
institutions are a way in which opposition demands can be contained
and answered without appearing weak. If authoritarian leaders face
a weak opposition and need little cooperation, they will not need to
make concessions and, therefore, will not need institutions. But if they
must impede opposition mobilization and solicit outside cooperation,
rulers may need to make policy concessions, in which case they need
institutions to organize these compromises.
As a forum through which dictators can make policy concessions,
nominally democratic institutions are instruments of co-optation. As
such, they determine the way in which political life is organized in
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
Introduction
July 5, 2008
xix
dictatorships and, consequently, affect the policies and outcomes that
are produced. Legislatures and parties facilitate policy concessions that
result in policy differences across differently organized authoritarian
regimes. Variations in policy translate into differences in economic
outcomes. But assuming that incumbents are able to observe with
some accuracy the conditions that dictate the choice to institutionalize and then choose their institutions as a strategic response to these
conditions, we should observe no significant differences in tenure on
the basis of institutions. These are the claims to be elaborated and
empirically assessed for all post–World War II dictatorships. Considered together, they not only demonstrate that institutions have effects
in dictatorships but also account for the institutional variation across
nondemocratic regimes.
0.2
the study of institutions in dictatorships
The focus on institutions that long has pervaded the study of democracies is now resurgent in the study of dictatorships. Traditional classifications have recognized – even if implicitly – that dictatorial regimes
differ in their organization and bases of support. Arendt (1951) and
Friedrich and Brzezinski (1965) highlighted the features of totalitarianism, whereas Linz (1970: 254) argued for distinguishing authoritarian
regimes because they have “. . . distinctive ways in which they resolve
problems common to all political systems. . . . ” Because institutions are
precisely those procedures and structures by which actors try to resolve
a variety of political problems, we can understand Linz’s distinction
as one founded on institutionalist criteria.
Moving beyond the distinction between totalitarian and authoritarian regimes, comparative politics scholars have identified a number
of important types of nondemocratic regimes. The communist totalitarian state eventually evolved into posttotalitarianism (Linz and
Stepan 1996), whereas various forms of personalist rule have been
identified as sultanism (Chehabi and Linz 1998) or neopatrimonialism
(Bratton and Van de Walle 1997). The prevalence of military regimes
throughout the developing world stimulated the study of these regimes:
their emergence and their organization (e.g., Barros 2002, Finer 1988,
Nordlinger 1977, Stepan 1971) as well as their effects on policies and
outcomes (e.g., Biglaiser 2002, Remmer 1978, Zuk and Thompson
14:38
Gandhi_Revised_Dec07.tex
xx
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Introduction
1982). O’Donnell (1979) identified an important subtype of military
regimes in bureaucratic-authoritarianism (see also Collier 1980, Im
1987). Even monarchy, as a subtype of nondemocratic regimes, persists in the contemporary world.
Within these categorizations is an understanding that institutions
matter under nondemocratic rule, but which institutions matter depends on the subtype of nondemocratic regime. The literature on
bureaucratic-authoritarianism and military regimes in general focuses
on the importance of the armed forces, whereas the study of monarchies emphasizes their dynastic structure (Herb 1999). Studies of personalist regimes focus on the nature of executive power that allows for
dictatorial leaders to exert tight control. Although these works emphasize important institutional features for each nondemocratic subtype,
none of them focus on the role of nominally democratic institutions,
such as legislatures and political parties. These institutions are assumed
to play a marginal role in these types of nondemocratic regimes.
The exception is the voluminous literature that surrounds the singleparty state. Many initial studies of regime parties focused on their description and categorization, providing useful intuitions about the origins and functions of regime parties (Collier 1982, Huntington 1970,
Michels 1949, Tucker 1961, Zolberg 1969). More recent work, such as
that by Geddes (1999), Slater (2003), Smith (2005), Magaloni (2006),
and Brownlee (2007), builds on their insights, making prominent again
the study of hegemonic or dominant parties. From this work, we have
acquired a better understanding of the origins of regime parties and
their maintenance, especially through their combination with other
institutions such as elections.2
In tandem with theoretical development and the accumulation
of evidence from specific countries and regions, the compilation of
2
Elections under dictatorship are another nominally democratic institution under significant inquiry. Hermet et al. (1978) examine the institution in detail as do more
recent works that investigate “hybrid regimes” (Diamond 2002), “competitive authoritarianism” (Levitsky and Way 2002), and “electoral authoritarianism” (Schedler
2006), as well as how nondemocratic incumbents shape electoral rules (Lust-Okar
and Jamal 2002), perpetuate electoral fraud (Lehoucq 2003, Schedler 2002), and manipulate the economy (Blaydes 2006, Magaloni 2006) to win electoral contests and
remain in power. Although these elections are another example of nominally democratic institutions under dictatorship, I do not cover them here because they may serve
different roles from those of dictatorial legislatures and parties.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
Introduction
978 0 521 89795 2
July 5, 2008
xxi
cross-national data on nondemocratic states allows us to determine
whether more general statements about the genesis, functioning, and
effects of these regimes are supported by evidence. In this regard, Geddes’ (1999) categorization of personalist, party, and military regimes
and her use of this classification to examine theories regarding the
survival of dictatorships and the likelihood of democratic transitions
have been path breaking. Her collection of data on dictatorships allows for cross-national empirical tests that often better fit our theories
than those based on older regime measures, such as Polity or Freedom
House. The result has been a burgeoning of quantitative research on
the effects of dictatorial types and institutions on outcomes such as
war (Lai and Slater 2006, Peceny et al. 2002), repression (Vreeland
2008), and economic development (Wright 2008).
This work follows contemporary trends in both the emphasis on
institutions and the use of various methods to examine the institutionalist account. Yet the argument advanced here and the data used
to assess it differ from previous work in a number of important respects. First, a common assumption is that rents are the only means by
which dictators build political coalitions. Spoils certainly constitute a
significant share of dictatorial concessions, but in this account, policy
compromises take center stage because there is no reason to believe that
policy is not an important second dimension over which the potential
opposition and incumbents may want to bargain. Second, although
the idea that rulers trade concessions for broadened political support
is not new, the claim that dictatorial legislatures and parties play a
significant role in this exchange is novel. As discussed earlier, these
institutions frequently have been dismissed as insignificant window
dressing. Finally, the claims about the emergence and effects of these
institutions are assessed using new cross-national time-series data on
the legislative and partisan arrangements of all nondemocratic regimes
from 1946 to 2002.
0.3
plan of the book
Chapter 1 commences with a brief sketch of our historical understanding of dictatorship, demonstrating how an institutionalized form
of rule in ancient Rome devolved into contemporary forms of dictatorship that frequently are thought to operate without institutional
14:38
Gandhi_Revised_Dec07.tex
xxii
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Introduction
constraints. Yet looking more carefully, we see that in reality dictatorships vary in their institutional structures as illustrated by the
different types of dictatorships (e.g., military, civilian, and monarchy)
and their nominally democratic institutions (e.g., legislatures and parties). In this opening chapter, I provide an overview of this variation
using data from 1946 to 2002 on all dictatorships around the world.
Because ultimately the goal is to understand the role of nominally
democratic institutions in dictatorships, the focus in this chapter is
on providing a systematic picture of the variation in legislative and
partisan arrangements.
After a description of the institutional forms under dictatorship,
the following two chapters are intended to provide an explanation for
the variance. In other words, the question to be answered is: What
accounts for the differences in legislative and partisan arrangements
across dictatorships? I argue that dictators face two basic problems of
governance: first, the need to obtain cooperation from some segments
of society and, second, the need to neutralize potential opposition.
Dictators can solve these problems by using nominally democratic institutions to share the spoils of power and to make policy concessions.
Policy compromises, in particular, require an institutional forum in
which demands can be revealed and agreements can be hammered out.
Chapter 2 uses three case studies – the ruling family of Kuwait, the
monarchy of Morocco after independence, and the military dictatorship of General Rodriguez Lara in Ecuador – to illustrate the logic
and plausibility of these arguments. Even though the cases are very
different – in historical, cultural, and political contexts – they demonstrate the logic of institutionalization. The cases are not intended as
tests of the theory but simply as illustrations of the plausibility of these
arguments.
The intuition provided from the cases is used to construct the formal argument elaborated in Chapter 3. The model relates institutional
strategies of dictators to the conditions they face, predicting that the
number of legislative parties should increase in the dictator’s need for
cooperation and in the strength of the opposition. A statistical test of
this prediction for all the countries for which the requisite data are
available between 1946 and 2002 shows this is the case. When dictators need to build support within society, they use legislatures and
parties as instruments of co-optation.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
Introduction
978 0 521 89795 2
July 5, 2008
xxiii
For dictatorial institutions to effectively encapsulate the potential
opposition, they must offer groups within society some real decisionmaking power even if in very limited policy realms. Without the hope
of policy concessions, the potential opposition has few incentives to
participate in nominally democratic institutions. In addition, the crafting of policy compromises requires an arena in which negotiations
can occur and deals can be hammered out. Legislatures and parties
under dictatorships serve this purpose. As a consequence, institutionalized dictatorships should exhibit differences in policies from their
noninstitutionalized counterparts. Chapter 4 provides a quantitative
analysis of this observable implication derived from the theory elaborated in the previous chapters. An examination of both civil liberties
and government spending for all dictatorships during the postwar period shows that institutions have an effect on government policies
about which citizens can form reasonably unified preferences. As such,
institutionalized dictatorships are forced to institute more liberal policies regarding citizens’ rights to speak freely and to organize collectively as workers and to spend less on the military. Yet the effect
of institutions on other types of spending is mixed, likely due to the
heterogeneity of preferences citizens may have over distributive goods.
If institutions influence policies under dictatorship, do they also
have an impact on outcomes? In Chapter 5, I take up this question,
looking specifically at the impact of dictatorial institutions on economic growth. In previous chapters, I argue that legislatures and parties help dictators build their bases of support in part by allowing
for some policy concessions to the potential opposition. As a result,
these institutions foster greater cooperation between the regime and
outside groups, reducing the potential for political instability. In addition, institutions serve as a conduit of information between the two
sides. Finally, the willingness of rulers to play the institutional game
indicates a measure of policy predictability. For all of these reasons –
political stability, greater information, and policy predictability – institutionalized dictatorships are expected to have higher economic
growth than noninstitutionalized regimes. A statistical analysis of all
dictatorships during the postwar period shows that institutions, in fact,
have a positive effect on economic growth.
The last observable implication concerns the political survival of
autocrats. In Chapters 2 and 3, I argue that authoritarian rulers choose
14:38
Gandhi_Revised_Dec07.tex
xxiv
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Introduction
to operate with legislatures and parties only when conditions dictate
that they must. As a result, depending on how badly dictators need
cooperation and the strength of the potential opposition, their degree
of institutionalization varies. Yet because dictators formulate their
institutional strategies as a best response to the conditions they face,
those rulers who choose to rule with institutions should not survive
significantly longer than those who govern without them. Chapter 6
provides details of this argument along with an event history analysis
of the 558 dictators of the postwar period. The results confirm that
nominally democratic institutions do not have a statistically significant
impact on their survival in power.
The book closes with a brief conclusion that summarizes the arguments and findings of the previous chapters and addresses whether
the presence of legislatures and parties in dictatorships renders these
regimes “more democratic.” I argue against such a view because these
institutions are instruments of co-optation for authoritarian regimes,
offering little in the way of representation and accountability to participants and ordinary citizens. In addition, because dictators retain the
power to alter and eliminate assemblies and parties, institutionalized
dictatorships remain closer in spirit to their noninstitutionalized counterparts than to democracies. Nominally democratic institutions under
dictatorship do matter but in ways that differ from their counterparts
in democracies. This distinction has implications for our understanding
of these regimes and for scholars and policy-makers who would encourage the creation of these institutions for the purpose of facilitating
democratic transitions.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
1
The World of Dictatorial Institutions
1.1
introduction
The contrast between democracy and dictatorship – in structures, policies, and performance – has been the object of intense scrutiny. Yet
little consensus exists over the definition of regime type. What cases
qualify as “democracies”? Which cases constitute the universe of “dictatorships”? The latter question is easy to answer when we encounter
the ferocity of a Joseph Stalin or a Pol Pot. No one would quarrel with
labeling their regimes as dictatorships. But other regimes are more
controversial. For almost seven decades, a new president in Mexico
was elected every six years. Nevertheless, the same party’s candidate
always won. Or consider Singapore, where Lee Kuan Yew crushed
political competition for over thirty years. Yet continuous measures
of regime type rate him somewhere in between “most autocratic” and
“most democratic.”
Part of the problem is that dictatorial rulers are quite inventive
in how they organize their rule. Decision-making power is concentrated in everything from juntas to politburos to family councils, for
example. Yet the institutional inventiveness of dictators is most apparent when they govern with nominally democratic institutions, such
as legislatures and political parties. Dictators frequently govern with
legislatures, some of which have formal law-making powers, whereas
others serve only to “advise and counsel.” Membership to assemblies
1
14:38
Gandhi_Revised_Dec07.tex
2
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
may be by appointment or by election. In elections, candidates may use
party labels or may be forced to run as independents. Party identification may mean little if political space is monopolized by a single party.
But many dictatorships allow for multiple parties, picking and choosing the types of parties to ban. Of course, some dictators rule without
any such institutions. The institutional diversity, however, makes it
difficult to identify a consistent set of criteria by which to define and
classify dictatorships.
The other reason for the confusion is that the historical usage of the
term dictatorship originated in ancient Rome, where it was identified
with very clear and specific institutional traits – in sharp contrast to
contemporary usage. Over time, however, the understanding of what
constitutes dictatorship evolved due to political manipulations of the
term. As a result, a regime type that originally was well-defined by its
rules became known as a regime type characterized by the absence of
rules.
What are dictatorships? Who are dictators? In what ways do they
organize their rule using nominally democratic institutions? The answers to these questions are mired in contemporary controversies. To
grasp the source of these disagreements, it is necessary to track the
historical understanding of this regime type. This chapter, then, begins
with a brief account of this historical evolution as a means to understanding contemporary debates over what constitutes dictatorship.
I adopt a minimalist definition of dictatorship to identify the post–
World War II sample of cases to study. To impose a minimal amount
of meaningful order on the dictatorial zoo, I classify dictators into three
types: monarchical, military, and civilian. Finally, the chapter shows
the institutional diversity of dictatorships in the various ways they
combine nominally democratic legislatures and political parties. From
this chapter, which delineates the universe of cases for analysis and
highlights the empirical patterns to be explained, a systematic study of
the emergence and effects of dictatorial institutions can proceed.
1.2
what is dictatorship?
Defining dictatorship should be simple: it is obviously the opposite
of democracy. At least, titles such as The Social Origins of Dictatorship and Democracy and The Economic Origins of Dictatorship and
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
3
Democracy would have us believe.1 Yet defining dictatorship is not
a trivial matter. Its conception has evolved from an institutional device used in ancient Rome to a system of rule that in modern times is
frequently associated with the absence of institutions and constraints.
The transformation of its meaning was the result of several distinct
moments when the original term was contorted and twisted for political ends. The result, by the mid-twentieth century, was a negative
definition of dictatorship that defined this form of rule by the absence of attributes associated with democracy. Yet the neglect of institutional forms in nondemocratic regimes is not justified, as will be
demonstrated by a description of the post–World War II sample of
dictatorships. Authoritarian regimes vary widely in their institutional
arrangements, and the task for the remainder of the book will be to
examine the reasons for and the effects of this variance.
1.2.1 Historical Usage
In contemporary usage, the terms tyranny and dictatorship maintain
close association. Yet this was not always the case. Although tyranny
was recognized as a type of regime since Aristotle, it initially was not
linked to the concept of dictatorship. For one, the term dictatorship,
originating in ancient Rome, postdates Aristotle. For two, in its original conception, dictatorship had a very distinct and specific meaning:
rule by a leader who was selected by the Consul in Rome to govern
during periods of emergency when external war or internal rebellion
threatened the existence of the polity.2 The term of the dictator was
to last no more than six months, and he could not remain in power
after the Consul that appointed him stepped down.3 During his term,
the dictator was authorized to use whatever power was deemed necessary to deal with the crisis at hand with the goal of restoring the old
constitutional order.
Within such a concise description of the institution are several aspects worthy of highlighting. First, regular institutions of the state, such
1
2
3
Moore’s (1966) classic work has been followed by Acemoglu and Robinson (2006).
The following description of dictatorship during the Roman period relies on the account of Nicolet (2004).
Initially, limits on the term of the dictator were unspecified; he was to abdicate power
as soon as the task for which he was appointed was completed.
14:38
Gandhi_Revised_Dec07.tex
4
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
as magistrates or the Senate, determined whether the situation at hand
required the nomination of a dictator for a resolution. Yet those who
decided on the necessity of a dictatorship were not allowed to nominate
themselves for the position. Second, the position of dictator was explicitly designed to be occupied by one man; collective leadership might
stymie attempts to resolve the crisis. Third, the dictatorship had a
broad range of power but not the authority to abolish other state institutions. Fourth, the dictator was never chosen by the people.4 Finally,
the goal of dictatorial rule was restoration of the old political order.
Within these institutional parameters, seventy-six dictatorships existed in Rome from 501 to 202 b.c. The majority of dictatorships
during this period were engaged in either military campaigns against
foreign powers or attempts at domestic reconciliation but not in repression of sedition. As a consequence, dictatorship was not associated
with brutal or repressive rule.
Sulla, a Roman general who refused to accept his dismissal and
went on to march on Rome, revived the institution of dictatorship
in 82 b.c. in an attempt to legitimate his rule. Significant differences,
however, existed between traditional dictatorships and Sulla’s regime.
For one, because Sulla obtained power only after his armies conquered
Rome and massacred his enemies, his regime marks the first time that
a dictatorship was established by the use of military force. The excessive brutality that Sulla used to neutralize his opponents led to an
association of dictatorship with terror. Moreover, in contrast to past
dictators whose tenures were understood to be limited given the nature
of the problem to be addressed, Sulla’s dictatorship involved the complete placement of power – military, executive, legislative, judicial –
in one man to remake the political order. The notion of dictatorship
to restore a previous political order was over.
Oddly enough, however, Sulla still adhered to the term limitations
attached to the title of dictator. In fact, after a short period, he stepped
down from power and returned to private life. It was not until January
44 b.c., when Caesar accepted the title of “dictator for life,” that the
temporariness of dictatorial power was abandoned.5
4
5
With one exception in 211 b.c. Yet the decision to allow for popular participation
was very controversial within the Senate. See Nicolet (2004) for more details.
Caesar initially had accepted dictatorial terms of one and ten years (Nicolet 2004).
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
5
That the concept of dictatorship was already corrupted by Sulla
and Caesar was often forgotten by later advocates. In The Discourses,
Machiavelli (Book I, Chapter XXXIV; 1950: 203) lauded the Roman
invention, observing: “ . . . truly, of all the institutions of Rome, this
one deserves to be counted amongst those to which she was most
indebted for her greatness and dominion.” The reason was simply
because processes of collective decision-making and even laws themselves may be inflexible and, hence, ill-suited instruments for resolving
crises. In a similar vein, Rousseau (Book IV, Chapter VI; 1987: 217)
remarked approvingly, “ . . . a supreme leader is named who silences all
the laws and briefly suspends the sovereign authority. In such a case,
the general will is not in doubt, and it is evident that the first intention
of the people is that the state should not perish.” For defenders of both
absolutist and liberal states, then, dictatorship constituted a solution
due to its decisiveness. But its temporary nature was equally important
for otherwise, “ . . . once the pressing need has passed, the dictatorship
becomes tyrannical or needless” (Book IV, Chapter VI; 1987: 218).
Aside from these mentions, however, dictatorship received scant
reference. In the nineteenth century, for example, the term was used
during only two periods: between 1789 and 1815 in reference to France
and briefly after 1852 to denote the Second Empire. What is noteworthy about the use of the term in the former period is how dictatorship
no longer referred to rule by just one man but to rule by a group. In
October 1793, the French National Convention suspended the constitution of the same year and established a provisional government
that served as a dictatorship of a revolutionary group: the Committee
of Public Health. Another feature of the original Roman conception –
rule by one man – was changed yet again.
It was only a matter of time, then, that the term dictatorship was
used to refer to not just a group but to an entire class. The term
resurfaced after 1917, when Lenin and his comrades used the phrase
“dictatorship of the proletariat” in self-congratulatory terms. Yet in
just a few years, the term dictatorship was imbued with pejorative connotations as liberal opponents of the Fascist Italian and Nazi regimes
used the label to describe what they were fighting against: “a highly
oppressive and arbitrary form of rule, established by force or intimidation, enabling a person or group to monopolize political power without
any constitutional limits, thus destroying representative government,
14:38
Gandhi_Revised_Dec07.tex
6
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
political rights, and any organized opposition” (Baehr and Richter
2004: 25). Following these trends, the Socialist International did its
own about-face in 1933, using “dictatorship” as a negative description of the Soviet regime as well. In any case, the employment of the
term in reference to both the self-proclaimed dictatorship of the proletariat in Russia and the fascist dictatorship in Italy were deviations
from the original Roman conception. The fascists were never committed to any temporary notion of power; and although the dictatorship
of the proletariat was to be temporary in nature, as a fundamental step
in the transformation from capitalism to socialism, it would obviously
not aim for restoration of the old order.
Events of the interwar period were important in that they precipitated an attempt to save the original notion of dictatorship.6 Attempting to save the notion of exceptional power and to counter the Bolshevik use of the term, Carl Schmitt distinguished commissarial from
sovereign dictatorship. Commissarial dictatorship conforms to the original Roman concept of dictatorship. With sovereign dictatorship, Schmitt collapses the distinction between normal and exceptional times,
claiming that the dictator has authority to restore the preconstitutional
will of the people, even if it means altering the constitution itself. At
the time, Schmitt was writing to justify giving dictatorial powers to
Germany’s reichspresident to deal with escalating economic and social crises. Schmitt’s conception of sovereign dictatorship is important
because it cements an important alignment of theory and practice in
the understanding of the term: this type of dictatorship may be neither
temporary nor restorative of the prior constitutional order.
The positive connotation of dictatorship, however, was never to
take hold. As Baehr and Richter (2004: 26) observe: “well into the
1940s, in liberal, constitutional states, dictatorship continued to be
used as the polar opposite of democracy in countless books, as well as
in political discourse.” During and immediately after the war, because
democracy was thought to embody all that was good, its antithesis, by
definition, was negative.
The emphasis on the distinction between democracy and dictatorship is similarly a twentieth-century phenomenon. Regime distinctions historically have been threefold since Aristotle first distinguished
6
This discussion is from McCormick (2004).
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
7
regimes by number of rulers. The distinction among monarchy, aristocracy, and democracy collapsed in two ways. Machiavelli was the
first to distinguish between government of the one versus government
of the assembly (whether of the few or the many), thereby setting
monarchy apart from other forms of regime. Bobbio (1989) attributes
the second collapse of the threefold regime distinction to Hans Kelsen,
who claimed distinctions on the basis of number to be superficial.
Instead, Kelsen proposed distinguishing on the presence or absence
of political freedom: “Politically free is he who is subject to a legal
order in the creation of which he participated” (Kelsen 1945: 284).
The crucial distinction, then, is between autonomy and heteronomy:
democratic forms of government are those in which laws are made
by the same people to whom they apply (i.e., autonomous norms),
whereas in autocratic states, lawmakers differ from those to whom
the law is addressed (i.e., heteronomous norms). As a result, “it is
now more precise to distinguish between two types of constitution,
instead of three: democracy and autocracy” (Quoted in Bobbio 1989:
137).
With this dichotomy, contemporary focus fell on the task of defining democracy, leaving dictatorship as the residual category, defined
only in terms of what it is not. Dictatorships are regimes without
competitive elections, without rule of law, without political and civil
rights, without regular alternation in power. These attributes may well
characterize dictatorships relative to democracies, but such definitions
emphasizing the relative absence of traits also masks significant heterogeneity in the organization of these regimes.
1.2.2 Contemporary Controversies
Dictatorships are defined here as regimes in which rulers acquire power
by means other than competitive elections.7 Leaders may come to
power by a coup d’état, a palace putsch, or a revolution. They may
take power themselves or be installed by military or foreign powers. The critical distinction is that they do not accede to power by
7
Except when the ruler first entered power by election and then consolidated his power
at the expense of democracy. In these cases (e.g., Marcos in the Philippines, Park in
South Korea, Fujimori in Peru), the leader’s reign is considered a dictatorship from
the beginning of his elected term.
14:38
Gandhi_Revised_Dec07.tex
8
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
a “competitive struggle for people’s votes” (Schumpeter 1976). This
conception of dictatorship, although not an advancement toward a
positive definition, is useful for both practical and conceptual reasons.
These reasons are discussed in turn.
The above definition of dictatorship is a minimal one, focusing on
the procedural rather than substantive aspects of the regime type. The
purpose of a minimalist definition is primarily for analytical clarity.
Definitions of regime type that incorporate a number of attributes
suffer from a number of problems. First, and most important, a multiplicity of attributes makes verifying causal connections more difficult.
Consider, for example, Linz’s (1970) four defining elements of authoritarianism: limited political pluralism, distinctive mentalities rather
than a guiding ideology, little political mobilization, and a leader who
exercises power predictably even if within ill-defined limits. If authoritarianism is defined on the basis of these four dimensions and we were
to find a relationship between authoritarianism and economic development, what then should we conclude about the causal story behind
the observed relationship? Is it the limited political mobilization or
the predictable leadership or some combination of these factors that
causes the observed patterns? With dictatorships distinguished in this
way, we hardly can say.
Second, broad definitions can entail substantive notions that either
generate tautologies or limit the applicability of the concept produced.
Evans’s (1989) distinction between “developmentalist” and “predatory” states already hints at the type of development outcomes they
will generate. Not surprisingly, bureaucratic-authoritarian regimes will
pursue exclusionary policies because they are defined, in part, by the
fact that this is what they do (O’Donnell 1979). Even if we were to
discover that these regimes, in fact, pursue inclusive policies, we would
be able to do nothing more than comment on the validity of the label’s
criteria.
Third, although not a problem inherent to broad-based definitions,
a strong correlation seems to exist between the number of attributes
and the amorphousness of these concepts. “Distinctive mentalities,”
for example, is difficult to measure or even identify.
Finally, by appending more and more attributes to a definition
of dictatorship, we run the risk of creating an empty set or, at the
very least, neglecting the most important distinctions among regimes.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
9
Mexico, for example, in spite of nearly a century of the Institutional
Revolutionary Party (PRI) dominance, was thought to be a case of a
dictatorship with qualifications. Yet the variety of dictatorships based
on other dimensions (e.g., freedoms and rights, civilian control of the
military) should not cloud the central distinction that is common to all
dictatorships and sets them apart from democracies – the absence of
competitive elections.
The reasons for using minimalist criteria do not constitute a justification for the substance of the criteria. In fact, the focus on elections
as the distinguishing feature between democracy and dictatorship is
not uncontroversial.8 Given a minimalist approach, then, why focus
on elections?
First, the focus on elections is compatible with most of the theoretical issues that animate empirical research on political regimes.
The prospect of acceding to power by regular, contested elections is
thought to produce incentives for political actors that are different
from those produced by irregular, nonelective methods of selection.
Different incentives will lead to different behavior (e.g., policies) that
then should produce different outcomes. Consider the impact of regime
type on economic development, for example. According to arguments
made famous by Locke and later the framers of the U.S. constitution, democracy was believed to be detrimental to political order and
economic development because elections without suffrage restrictions
would enable the poor to elect demagogues who would seize and redistribute the assets of the propertied classes.9 More recent arguments,
in contrast, claim that dictatorships are bad for development because
without electoral constraints, dictators are free to extract rents and
substitute the provision of private goods for public ones. In either case,
elections are the reason why political actors are expected to behave differently and produce different outcomes in democracy and dictatorship.10
8
9
10
For some background on this debate, see Cheibub and Gandhi (2004), Collier and
Adcock (1999), Diamond (2002), and Munck and Verkuilen (2002).
The argument was repackaged in the twentieth century with the fear that democratic
governments would capitulate to demands for consumption (rather than investment)
made by workers who, of course, were also voters (de Schweinitz 1964, O’Donnell
1979).
Elections also feature prominently in arguments about the relationship between
regime type and education (Brown 1999, Brown and Hunter 2004, Habibi 1994),
14:38
Gandhi_Revised_Dec07.tex
10
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
Second, distinguishing regime type on the basis of elections reminds
us that even if dictators have other nominally democratic institutions,
such as legislatures and parties, they are still dictators. Political life
is organized fundamentally differently in systems in which leaders are
chosen by competitive elections and in systems where they are not. As
Przeworski (1991) explains, democracy is distinct from dictatorship
as a political system in which no actor can control outcomes with
certainty, either by altering chances ex ante or overturning outcomes
ex post. The most tangible signifier of this uncertainty are competitive
elections because the result is “an instruction what the winners and the
losers should and should not do: the winners should move into a White
or Pink House or perhaps even a palacio . . . The losers should not move
into the House and should accept getting not more than whatever is
left” (Przeworski 1999: 45). This point is revisited and elaborated in
the Conclusion because it has implications for how we think about
dictatorial parties and legislatures and whether their presence makes
some authoritarian regimes “more democratic” than others.
I use Przeworski et al.’s (2000) dichotomous classification that distinguishes regimes by electoral criteria. The resulting sample encompasses 140 countries that experienced dictatorship at some point during the period 1946 to 2002, or 4,607 country-year observations. In
Figure 1.1, all countries that have been ruled by dictators at some point
during the postwar period are shaded. During the 1970s, 75 percent
of all countries were dictatorships. By the mid-1990s, that percentage
had fallen to around 50 percent.
In most cases, the identification of dictatorships is uncontroversial
in that the most prominent cross-national regime classifications agree.
The correlations among Freedom House, Polity, and Przeworski et al.
(2000) range from 88 to 95 percent. Even though Przeworski et al.
(2000) is a dichotomous measure, its correlation with both Freedom
House and Polity are high because these continuous measures have bimodal distributions. Once the easy cases at the extremes of the distribution (e.g., North Korea, Iran, Sweden, and Great Britain) are excluded,
however, the correlations become significantly weaker: for example,
only 0.75 between Polity and Freedom House. The difficulty appears
economic reforms (Haggard and Kaufman 1995, Hellman 1998), interstate war
(Fearon 1994, Reiter and Stam 1998), and many other things.
14:38
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
figure 1.1. Dictatorships in the world, 1946–2002.
Gandhi_Revised_Dec07.tex
11
14:38
Gandhi_Revised_Dec07.tex
12
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
with cases such as Mexico, Botswana, Malaysia, and other countries
that populate the middle of the distribution. Przeworski et al. (2000),
however, determine the regime type of countries on the basis of clear
and observable criteria that are consistent with the minimalist, procedural definition of regimes used here.11 Table 1.1 provides the list of
countries and years of dictatorial spells by region.
During the 1946–2002 period, the sub-Saharan African region constitutes the largest portion of the sample with over 1,800 country-year
units. Eighty-seven percent of these country-years were under dictatorship. The proportion of observations under nondemocratic rule is
similarly high in the Middle East/North Africa (87 percent), Eastern
Europe/former Soviet Union (74 percent), and Asia (68 percent). The
Latin American sample has a smaller proportion of dictatorial countryyears (38 percent) due to numerous transitions between democratic
and authoritarian rule. Finally, only 7 percent of Western European
country-years are included in the dictatorial sample because authoritarianism was confined to southern Europe during this fifty-six-year
period.
Within regions, high variance in number of dictatorial spells and
in duration of authoritarian rule also exists among countries. Some
countries, such as Jordan and the People’s Republic of China, have
been dictatorships since their respective dates of independence. Other
countries have had smaller, but multiple, authoritarian spells, such as
Argentina with four periods or Ghana with three. Still other countries
have had very limited experience with dictatorship. During the postwar
period, Costa Rica, for example, was a dictatorship for only 18 months
after José Figueres’ seizure of power in May 1948. The average length
of a dictatorial spell, however, is long (40 years).12
1.3
who are the dictators?
After determining the sample of dictatorial regimes, the next step
is to identify the effective head of government. Effective heads of
11
12
A regime is a dictatorship if it fails to meet at least one of the following four rules:
(1) the chief executive must be elected; (2) the legislature must be elected; (3) there
must be more than one political party competing in elections; and (4) an alternation
in power must have taken place. For more details, see Przeworski et al. (2000).
Lengths of spells are not left-hand censored.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
13
table 1.1. Countries under Dictatorship, 1946–2002 (in Alphabetical
Order by Region)
Country
Africa
Angola
Benin
Botswana
Burkina Faso
Burundi
Cameroon
Cape Verde Island
Central African Republic
Chad
Comoro Island
Congo
Djibouti
Equatorial Guinea
Eritrea
Ethiopia
Gabon
Gambia
Ghana
Guinea
Guinea-Bissau
Ivory Coast
Kenya
Lesotho
Liberia
Madagascar
Malawi
Mali
Mauritania
Mozambique
Niger
Nigeria
Rwanda
Sao Tome and Principe
Senegal
Seychelles
Sierra Leone
Somalia
Somaliland
South Africa
Years under Dictatorship
1975–2002
1960–1990
1966–2002
1960–2002
1962–92; 1996–2002
1960–2002
1975–1990
1960–1992
1960–2002
1975–1989; 1995–2002
1960–1991; 1997–2002
1977–2002
1968–2002
1993–2002
1946–2002
1960–2002
1965–2002
1957–1968; 1972–1978; 1981–1992
1958–2002
1974–1999
1960–1999
1963–1997
1966–1992
1946–2002
1960–1992
1964–1993
1960–1991
1960–2002
1975–2002
1960–1992; 1996–1999
1966–1978; 1983–98
1962–2002
1975–1990
1960–1999
1976–2002
1967–1995; 1997
1969–2002
1991–2002
1946–1993
(continued)
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
14
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
table 1.1 (continued)
Country
Sudan
Swaziland
Tanzania
Togo
Uganda
Zaire
Zambia
Zimbabwe
Years under Dictatorship
1958–1964; 1969–1985; 1989–2002
1968–2002
1961–2002
1960–2002
1962–1979; 1985–2002
1960–2002
1964–1990
1965–2002
Middle East
Algeria
Bahrain
Egypt
Iran
Iraq
Jordan
Kuwait
Lebanon
Libya
Morocco
Oman
Qatar
Saudi Arabia
Syria
Tunisia
Turkey
United Arab Emirates
Yemen
Yemen Arab Republic
Yemen PDR
1962–2002
1971–2002
1946–2002
1946–2002
1946–2002
1946–2002
1961–2002
1975–2002
1952–2002
1956–2002
1951–2002
1971–2002
1946–2002
1946–2002
1956–2002
1946–1960; 1980–1982
1971–2002
1990–2002
1967–1989
1967–1989
Asia
Bangladesh
Bhutan
Brunei
Cambodia
China
Fiji
Indonesia
Korea, North
Korea, South
Laos
Malaysia
Maldives
Mongolia
1971–1990
1971–2002
1984–2002
1953–2002
1946–2002
1970–2002
1946–1998
1948–2002
1948–1959; 1961–1987
1959–2002
1957–2002
1965–2002
1946–1991
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
15
table 1.1 (continued)
Country
Myanmar
Nepal
Pakistan
Philippines
Samoa
Singapore
Sri Lanka
Taiwan
Thailand
Tonga
Vietnam
South and Central America
Argentina
Bolivia
Brazil
Chile
Colombia
Costa Rica
Cuba
Dominican Republic
Ecuador
El Salvador
Grenada
Guatemala
Guyana
Haiti
Honduras
Mexico
Nicaragua
Panama
Paraguay
Peru
Suriname
Uruguay
Venezuela
Eastern Europe and former
Soviet Union
Afghanistan
Albania
Azerbaijan
Belarus
Years under Dictatorship
1958–1959; 1962–2002
1946–1990; 2002
1956–1971; 1977–1987; 1999–2002
1965–1985
1962–2002
1965–2002
1977–1988
1949–1995
1946–1974; 1976–1982; 1991
1970–2002
1976–2002
1955–1957; 1962; 1966–1972;
1976–1982
1946–1978; 1980–1981
1964–1978
1973–1989
1949–1957
1948
1952–2002
1946–1965
1946–1947; 1963–1978; 2000–2002
1946–1983
1979–1983
1954–1957; 1963–1965; 1982–1985
1966–1991
1946–1993
1946–1956; 1963–1970; 1972–1981
1946–1999
1946–1983
1946–1948; 1951; 1968–1988
1946–2002
1948–1955; 1962; 1968–1979;
1990–2000
1980–1987; 1990
1973–1984
1948–1958
1946–2002
1946–1991
1991–2002
1991–2002
(continued)
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
16
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
table 1.1 (continued)
Country
Bosnia-Herzegovina
Bulgaria
Czechoslovakia
Georgia
Germany, East
Hungary
Kazakhstan
Kyrgyzstan
Moldova
Poland
Romania
Serbia and Montenegro
Tajikistan
Turkmenistan
Uzbekistan
U.S.S.R.
Yugoslavia
Western Europe
Cyprus
Greece
Portugal
Spain
Years under Dictatorship
1991–2002
1946–1989
1946–1989
1991–2002
1949–1989
1946–1989
1991–2002
1991–2002
1991–1995
1946–1988
1946–1989
1991–2002
1991–2002
1991–2002
1991–2002
1946–1990
1946–1990
1960–1982
1967–1973
1946–1975
1946–1976
government constitute the focus because we are concerned with the
effects of institutional features of dictatorial regimes on policies and
outcomes, and decisions about the regime’s institutional features and
its policies are made only by those who truly have authority. Under
democracies, identifying the effective head of government is usually
straightforward: prime ministers lead parliamentary regimes, whereas
presidents are at the helm in presidential systems.13 Under dictatorship,
the task is considerably more difficult. Why?
One reason is that in dictatorships, the head of government goes by
many different titles. Under dictatorship, the chief executive often bears
the title of “president.” In monarchies, the dictator is known as the
13
Even in democracies, exceptions do exist. The effective head of government in France,
for example, depends on the partisan identities of the president and the prime minister.
Or in South Africa, the effective head of government bears the title of “president”
even though he is head of a parliamentary system.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
17
“king” or the “emir.” But other dictators have assumed more creative
titles: “Chairman of the State Law and Order Restoration Council,”14
“Leader of the Revolution,”15 or simply “spiritual leader.”16
Semantics aside, identifying the effective head of government within
a dictatorship is difficult because dictatorships frequently have multiple
executive figures – both nominal and effective. In communist regimes,
for example, a prime minister, a chairman of the council of state (i.e., a
president), and a general-secretary of the Communist Party nominally
constitute the authorities, yet the prime minister and president are not
effectively in charge. An extreme example makes the point: Liu Shaoqi
remained president of communist China even as he was dragged from
his home by Red Guards and forced to make a “self-criticism” during
the Cultural Revolution (Short 2000, Salisbury 1992).17 In communist
regimes, in fact, the general-secretary of the party is the effective head
of government.
Even more difficult are noncommunist cases in which the nominal
and effective heads differ because an éminence grise lurks behind the
scenes. The military may be in charge, but unwilling to assume power
directly. In this case, a civilian is propped up as president, as occurred in
Algeria under Abdelaziz Bouteflika, in Uruguay under Juan Bordaberry
Arocena and Aparisio Mendez Manfredini, and in Fiji under Ratu
Sir Kamisese Mara. In addition, there are rare cases in which a longlasting dictator rotates nominal heads of government even as he retains
ultimate power. Rafael Trujillo was the effective head even though he
made his son, Hector, president of the Dominican Republic in 1952.
Similarly, Anastacio Samoza Garcia ruled Nicaragua for nineteen years
but allowed Victor Roman y Reyes to assume the presidency for a
three-year interlude in the 1940s.
Finally, for some countries, their political histories are largely about
fighting over who is the effective head of government. From its independence to 1970, politics in Cambodia largely centered around the
struggles between the monarch and the prime minister over how much
14
15
16
17
The official title of the leader of the Burmese junta since 1981.
Muammar al-Quaddafi’s official title since 1979, adopted ten years after seizing
power.
The title “rahbar” was first adopted by Khomeini in 1980. The position is distinct
from that of president.
Mao stripped Liu of his presidential position before having him killed in 1969.
14:38
Gandhi_Revised_Dec07.tex
18
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
power each of them should have. Pakistan has witnessed a similar type
of political tug-o-war between the president and the prime minister. In
these cases, the nominal head of government may not be the effective
head of government.
Given the diversity in leadership arrangements, constructing rules
by which the effective head of government can be identified becomes
especially crucial in avoiding ad hoc judgments. Here the effective
head of government is identified as (1) general-secretaries of the communist party in communist dictatorships, except in the case of Deng
Xiaoping in China;18 (2) king or presidents in noncommunist dictatorships, except in the cases of Singapore, Malaysia, Cambodia, Laos,
and Myanmar, where the effective head sometimes bears the title of
“prime minister,” and (3) another individual or the institutionalized
military if sources agree that the nominal head of government is not
effectively in charge.19
The first two rules are relatively uncontroversial and successfully
classify the vast majority of the cases within the postwar sample. In
approximately 270 country-year observations, the effective head of
government differs from the nominal one for the reasons discussed
earlier. The classification of these cases depends on the relative subjectivity of the sources employed, but I have used a consistent set of
sources to capture these hard cases.
Dictators do not appear one per year so that each one will neatly
fit the data matrix of country-year observations. Political instability
is often found in dictatorships so that in any given year, there may
be multiple changes in the effective head of government.20 In 100
country-year observations, the effective head changes more than once.
These heads in the middle of the year may be in power for as short a
period as five days, as in the case of Leon Cantave in Haiti. No matter
how short their tenures, their appearance is recorded because ex ante
we could not have known how long they would remain in power.21
18
19
20
21
Deng never held the official title of general-secretary.
These sources are primarily Banks (various years) and Lentz (1994).
One interesting conceptual question that is not tackled here is whether we can properly talk about a “regime” when the chief executive changes so often. Noteworthy
cases include Comoros in 1995, which experienced four changes in heads, or Bolivia
in 1979, when power changed hands three times. Such instability, however, occurs
in only a small handful of cases.
For this same reason, “interim” and “acting” leaders who survive for longer than one
year also are included in the sample (in contrast to Bienen and Van de Walle (1991).
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
19
Because the unit of observation is country-year, however, the effective
head to consider is that of December 31.
Finally, excluded are dictatorships that were ruled under collective
leadership: Yugoslavia after the death of Marshall Tito in 1980 until the disintegration of the state in 1990, Bosnia-Herzegovina from
1998 to 2002, and Somalia under the five-chairmen National Salvation Council from 1997 to 1999. Under collective leadership, it is
impossible to assign responsibility for decisions to a particular leader,
and for this reason, these observations are not included.
The sample of dictators, then, includes 558 dictators in 140 countries from 1946 to 2002. Some countries, such as Equatorial Guinea
and Oman, have undergone a single leadership change, whereas other
states have witnessed much more executive volatility. Haiti, for example, from 1946 until the transition to democracy in 1994, was governed
by 19 chief executives with some lasting in power for mere days.
Among the dictators, themselves, their backgrounds and tenures
vary widely. Ernesto Zedillo, a Western-educated economist and last
of a long line of Mexican leaders who came to power under more or
less orderly rules of succession under the PRI, seems to share little with
Iran’s Ayatollah Khomeini, a religious cleric who led a revolution to
unseat a king. Similarly, Ferdinand Marcos, who initially won elections
and later clung to power in the Philippines for two decades, seems
incomparable to Pol Pot, who presided over the Khmer Rouge’s killing
fields in Cambodia for three years. Even dictators want to believe they
are unique. As Imelda Marcos, wife of Ferdinand and first lady of the
Philippines, quipped: “At least, when they opened my closets, they
found shoes, not skeletons.”22 These examples highlight the diversity
of dictators and raise the question: Is it possible to impose order on
such a heterogeneous group of rulers?
I distinguish here on the basis of three types of dictators: monarchs,
military rulers, and civilian dictators. Why these particular types? For
one, monarchy historically has been treated as a separate regime type.
Moreover, among other nondemocratic rulers, those who originate
from the armed forces are often treated as distinct from their civilian
counterparts. Tradition, in and of itself, is not a sufficient reason to
22
One cannot distinguish ex ante the cases in which heads attempted to consolidate
power and failed from cases in which heads did not try to do so.
Quoted from Diaz (2004).
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
20
table 1.2. Replacement of Dictators by Type
Type of Successor
Type of Current
Dictator
Monarch
Military
Civilian
Democrat
Total
Monarch
Military
Civilian
24
0
2
6
100
60
4
46
158
1
58
33
35
204
253
Total
26
166
208
92
492
rely on such distinctions. Yet there are important substantive reasons
for distinguishing these types.
First, these distinctions signify the unique types of threats to dictators and the institutional methods by which they deal with them and
organize their rule. Although discontent with dictatorship can come
from any segment of society, members of the ruling elite usually constitute the first major threat to the dictator. Dictators, in fact, are
frequently deposed by a fellow member of the regime. As a member
of the ruling elite, the usurper is in a privileged position to acquire the
force and support he needs to successfully depose the incumbent. So it
is not surprising that, as Table 1.2 shows, when dictators are replaced
by other dictators, they usually fall to a ruler of the same type.
To mitigate the threat posed by elites, dictators frequently establish
inner sanctums where real decisions are made and potential rivals are
kept under close scrutiny. The setup of these inner sanctums depends
on the type of dictator because the source of the threat and the means
by which the dictator can deal with it depends on his type. As a result,
monarchs rely on family and kin networks along with consultative
councils; military rulers contain key rivals from the armed forces within
juntas; and civilian dictators usually create a smaller body within a
regime party – a political bureau – to co-opt potential rivals. Because
real decision-making power lies within these small institutions, they
generally indicate to whom dictators may be responsible and how
power is organized within the regime. These elite institutions produce
different incentives and constraints on dictators that, in turn, should
have an impact on their decisions and performance.
As will be shown in the following chapter, one of the ways in which
these distinctions will reveal themselves as important is in their impact
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
21
on the decision of dictators to govern with other types of institutions –
namely legislatures and political parties. Moreover, because the focus
of this study is on the institutional structure of dictatorships, it seems
fitting to incorporate another institutional distinction among these
regimes. Finally, distinction of these types is possible through clear
observable criteria. The justification and criteria for identifying each
type of dictator is discussed in succeeding sections.
1.3.1 Monarchs
Monarchy has always been treated as a type apart from other forms
of nondemocratic rule. One reason for this separate treatment may be
that early regime classifications distinguished on the basis of number of
those who rule. Aristotle, for example, distinguished monarchy as rule
by one (as opposed to the few or the many) to promote the good life.
As mentioned earlier, Machiavelli was among those who perpetuated
the distinction between monarchy and other forms of government even
as regime classifications became increasingly dichotomous.
Historical accounts of absolutist monarchical power only cemented
the distinction, yet they also pointed to the inherent drawbacks of rule
by one. Freedom from all constraints opened up the possibility of myopic, unpredictable, and unstable behavior on the part of monarchs
that was detrimental to their interests along with those of their subjects. When French kings, for example, continued to engage in various
forms of expropriation to fund wars and lavish lifestyles, their subjects
suffered. Eventually, his subjects modified their behavior in response
to the state’s predatory actions.
As a result, the virtues of binding or constraining one’s own powers
became apparent. Jean Bodin, a staunch royalist of the sixteenth century, counselled his king to share power with others and accept limits
on his own will, if only to secure obedience. Bodin observed: “The
less the power of the sovereign is (the true marks of majesty thereunto still reserved) the more it is assured” (quoted in Holmes 1995:
115, Bodin, Book IV, Chapter 6: 517).23 Hence, it is in the king’s
23
While liberal thinkers espoused constraints on absolute sovereigns, Bodin’s thoughts
on the subject are notable not only because they predate liberal thought but also
because they are deployed for proroyalist purposes. On the importance of Bodin on
this point, see Holmes (1995).
14:38
Gandhi_Revised_Dec07.tex
22
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
interest to accept various constraints on his power: laws of nature, laws
of succession, immemorial customs, a prohibition on taxation without consultation, independent magistracies, and even parliamentary
prerogatives.
It is questionable, however, whether European monarchs took to
heart the lessons of Bodin and others. Although kings established
constitutions along with parliaments and judicial bodies, the extent
to which they intended these institutions to serve as a form of selfbinding is unclear. The reason for the ambiguity lies in the fact that
although the British and French kings may have learned the hard way
(by being overthrown and beheaded) that they would have to cede real
power to these institutions, most monarchs fought to preserve their
discretionary powers in the face of parliamentary challenges.24 Parliamentary responsibility developed only after several years of wrangling
between monarchs and assemblies and often was codified only much
later (von Beyme 2000). As the early European examples illustrate,
for at least a notable period, constitutions did not necessarily imply
constraints on the sovereign or constitutionalism.
Contemporary monarchs in the Middle East evinced the same resistance to constraints on their authority. Faced with a variety of fiscal crises or political independence, monarchs in the region tried to
rationalize their rule by establishing constitutions, parliaments, and
political parties. The 1925 Iraqi constitution, for example, allowed the
king to choose the prime minister but clearly established ministerial
responsibility to the parliament. Similar provisions were included in
the Kuwaiti constitution and the 1923 Egyptian constitution; under
the latter document, the opposition Wafd Party was able to garner impressive parliamentary majorities. Yet like their early European counterparts, these monarchs had no intention of relinquishing authority
to their institutional creations. Unlike their predecessors, they had the
brute power to insure that such devolution of power would not happen. In Egypt, the monarch fundamentally violated the constitution
of 1923 three times in its first seven years of operation (Brown 2002:
39). Similarly, “[O]n paper, Iraq’s parliament was as powerful as any
that has existed in the Arab world. Yet on no occasion did a single
24
North and Thomas (1973) and other literature in political science of this ilk overstate
the generality of the British and French cases.
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
23
government, or even a single minister, fall because of a parliamentary
vote of no confidence” (Brown 2002: 45).
The price for these monarchs, however, seems to have been accepting constraints of a different kind. Rather than constitutions and
assemblies, constraints came in the form of dynastic families. Rulers
of contemporary monarchies had to accept kin as holders of important government positions, as legitimate recipients of state revenues,
and as participants in decision-making at all levels. Khalifa in Qatar,
for example, reshuffled his cabinet in 1992 so that his closest family
members would serve as his ministers: his sons were Ministers of Defense, Finance and Petroleum, Interior, and Finance, Economy, and
Trade; his grandson was in charge of State for Defense Affairs; and his
nephews were in Public Health and Islamic Affairs (Herb 1999: 123).
Similarly, in Saudi Arabia, King Fahd appointed his six full brothers –
with Fahd, nicknamed the “Sudairi Seven” – to major posts in the
government.
Family members play a crucial role in making decisions on important matters, as seen with the issue of succession. In Kuwait, for example, succession alternates between two branches of the Sabah family,
but “the most basic rule of the succession is that family ‘elects’ the
ruler by consensus, based on the perception by family leaders of their
own best interests” (Herb 1999: 80). In Oman, the next in line must
be a male descendant from the al-Said family but must also be chosen
by a family council. Saudi succession became resolved by a more consensual process after King Faysal established the Higher Committee of
Princes as an advisory council to the king on issues of succession. The
Committee’s composition was designed to rally the entire family, and
the Committee was even given the authority to supervise the succession
in the event of Faysal’s death (Bligh 1984: 88).
Certainly rules of succession have not always been followed. As
the list of post–World War II monarchs and how they left power in
Table 1.3 shows, natural death is one way in which monarchs leave
power but so is assassination or deposal at the hands of other family
members.
Yet even when the rules of succession are broken, it typically happens with the blessing of key family members. Faysal did not depose
Sa’ud, his father’s designated heir, until he received the support of
other Saudi princes. That support was quickly extended due to Sa’ud’s
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
24
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
table 1.3. Postwar Monarchs and Their Exit from Power (Alphabetical
by Country within Each Category)
Exit from Power
Monarch (Country: Year of Exit)
Natural death
Isa (Bahrain: 1999)
Dorji (Bhutan: 1972)
Hussein (Jordan: 1999)
Abdullah (Kuwait: 1965)
Sabah (Kuwait: 1977)
Muhammad V (Morocco: 1961)
Hassan II (Morocco: 1999)
Tribhuwan (Nepal: 1955)
Mahendra (Nepal: 1955)
Tamasese (Samoa: 1963)
Ibn Saud (Saudi Arabia: 1982)
Sobhuza II (Swaziland: 1982)
Resignation
Abdullah (Iraq: 1953)
Naif (Jordan: 1951)
Dzeliwe (Swaziland: 1986)
Ntombi (Swaziland: 1986)
Assassination
Abdullah (Jordan: 1951)
Faysal (Saudi Arabia: 1975)
Deposal by family member
Zahir Shah (Afganistan: 1973)
Mwambusta IV (Burundi: 1966)
Talal (Jordan: 1952)
Said (Oman: 1970)
Ahmad (Qatar: 1972)
Khalifah (Qatar: 1995)
Saud (Saudi Arabia: 1964)
Deposal by outsiders (including
military or civilian dictators and
transitions to democracy)
Ntare V (Burundi: 1966)
Sihanouk (Cambodia: 1955)
Faruq I (Egypt: 1952)
Haile Selassie (Ethiopia: 1974)
Reza Pahlavi (Iran: 1979)
Faisal II (Iraq: 1958)
Idris I (Libya: 1969)
Farid Didi (Maldives: 1986)
Birendra (Nepal: 1991)
Note: Table from Gandhi and Schwartzberg (2004).
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
25
incompetence. Nevertheless, Faysal had to obtain it. While vacationing in Switzerland, Khalifa of Qatar was deposed in 1995 by his own
son, who had rallied the support of his kin. The main threat to monarchs, then, originates from family members who are the only people
legitimately qualified to succeed them.
Bodin’s original point – that constraints can be empowering and
necessary for political survival – may not have been entirely lost on
monarchs. Herb (1999) convincingly argues that in the Middle East,
only those monarchs who successfully transformed from traditional
monarchies based on the king’s supremacy to dynastic monarchies in
which the family as a whole is the locus of power have been able to
survive. The installation of a dynastic structure saved monarchies in
Bahrain, Kuwait, Qatar, and Saudi Arabia from threats of revolution
and continue to make regime change unlikely (even apart from the
effects of oil). In contrast, monarchies that failed to adapt and maintained absolute monarchism without familial constraints, as in Egypt,
Iraq, and Libya, fell to revolutionary dictatorships. The transformation to dynastic monarchy meant that kings would no longer be as
powerful vis-à-vis their family members. But given the alternative that
came with the failure to transform – the complete end of monarchical
rule – it was a price worth paying for most monarchs. Monarchs are
not constrained by parliaments or courts but rather by family factions
and kin networks. Hence, rule by one has become rule by family.
1.3.2 Military Dictators
As Sulla’s march on Rome shows, the use of armies and force in seizing
power has a long history. In fact, all extralegal seizures of power
require force even if the threat is implicit. In addition, nonmilitary
leaders sometimes employ just as much violence as the military in
usurping power. Consequently, the use of violence or coup d’états in
taking power is not what distinguishes military dictatorship from other
forms of nondemocratic rule.
What constitutes military rule is the fact that the armed forces are the
institution through which rulers govern. Just as contemporary monarchs must co-opt family members and, in turn, use them to consolidate
their rule, military dictators must neutralize the threats posed by their
closest colleagues and harness their cooperation to govern.
14:38
Gandhi_Revised_Dec07.tex
26
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
Within military regimes, then, the junta is the locus of decisionmaking power. For generals who take power on behalf of the institutionalized military, their juntas typically are small and include heads of
the various service branches. For members, sometimes lower ranked,
who seize power in a factional coup, their juntas tend to be larger based
on their need to attract members to their cause. After the 1966 coup
in Ghana, for example, the selection of junta members was extended
in stages as the need to consolidate support within the armed forces
arose. Immediately after the coup, the National Redemption Council
announced six members; the following day, it expanded to nine only
to be followed by the addition of two more members four days later.
Still two weeks later, six additional officers were added (Welch 1974:
136–138). Indeed, Finer (1988: 260) reports that in contrast to most
Latin American juntas that are composed of the three or four heads of
the service branches, by the 1980s, juntas outside the region, organized
by middle-ranking officers, on average had eleven members.
The preexisting organizational structure of the armed forces can be
harnessed for the purposes of governance. In Indonesia, for example, a
fifth of the parliament’s seats were reserved for members of the armed
forces, and a soldier was stationed in each of Indonesia’s thousands of
villages, serving as the military’s representative (Brooker 1995). The
Argentine Processo is perhaps the most striking use of military hierarchy in governance. Prior to the 1976 coup, the four service branches
agreed to a detailed power-sharing agreement in an attempt to prevent
dominance by the army. According to the agreement, legislation underwent review by various subcommittees within each service branch
before it was considered by members of the junta. In this way, decisions
reflected a deliberative process within the armed forces that incorporated the opinions of its most important constituencies (Fontana 1987).
Military dictatorship as a regime type, then, cannot have existed
prior to the professionalization of the armed forces that occurred at
varying times throughout the developing world. In Latin America, for
example, after independence from Spain, it was more proper to speak
of armed factions led by caudillos than modern military organizations.
The establishment of a modern military began as early as 1840 in Brazil
and as late as 1896 in Peru with the creation of academies for officer
training (Rouquié 1987). The introduction of European instruction
(which in some cases lasted until the end of World War II) and of
14:38
Gandhi_Revised_Dec07.tex
cuus282-gandhi.cls
978 0 521 89795 2
The World of Dictatorial Institutions
July 5, 2008
27
compulsory military service led to the creation of “new” armies by the
early 1900s.
Professionalization of the military created an institution distinct
from the rest of society. During the era of caudillos in Latin America, a
career in the military was unattractive. As a result, recruitment became
more and more internalized, which in turn increased the separation of
the military from civilian elites and contributed to their esprit de corps
under formal military education. As Rouquié (1987: 104) observes,
the end result was that “The combination of isolation from society
as a whole and the cohesion and prestige of the group produces a
haughty, closed quality in military life, a proud withdrawal into the
institutions that limits one’s horizon at the same time that it produces a
consciousness of having an important role to play. The professionalized
officers of the new armies direct their loyalties to the institution of the
military while believing they are serving the state.”
The professionalization of the military conformed to the plans of
civilian elites who wanted to curtail the power of caudillos and to
rationalize the state’s monopoly on violence. At the same time, civilian elites wanted to be able to call on the military as an arbiter or a
problem-solver during times of crisis. As the poder moderador, “the
military is repeatedly called into politics to be the moderator of political activity, but is denied the right systematically to attempt to direct
changes within the political system. The military task in the moderator model is essentially the conservative task of systems maintenance”
(Stepan 1971: 63). Concretely, this meant that when the military intervened, it was restricted to the removal of the chief executive and
to handing power to alternative civilian forces. As a temporary form
of rule designed to restore the old political order, the military came
to serve the functions of dictatorship, as understood according to the
original Roman conception.
Unforeseen by the civilian elites who advocated military professionalization, however, was that the creation of an autonomous military
above civilian parties provided it with the means to intervene on its
own in politics. No matter how extensive their interventions increasingly became, the military continued to perceive them as “a sort of
internal adjustment. The military coup d’état, then, is only an intervention of the state within itself, producing a rupture in which by a sort
of metonymy the part is taken for the whole” (Rouquié 1987: 104).
14:38
Gandhi_Revised_Dec07.tex
28
cuus282-gandhi.cls
978 0 521 89795 2
July 5, 2008
Political Institutions under Dictatorship
Military intervention that exceeded the norm of poder moderador soon
became common.
Latin America’s experience of distinguishing between the soldier
and the citizen, however, was not universal. In many developing countries, professionalization of the military occurred to some extent during
the colonial period as European rulers sought to reinforce their control over distant lands. After independence, the armed forces, having
already been molded as a profession, became “a model of development and a microcosm of the state with tentacles reaching, through
compulsory military service, to all groups in society” (Khuri and Obermeyer 1974: 62). The socialization process stressed the integration of
the military into society rather than its maintenance as an institution
apart. If members of the armed forces were not just soldiers, but also
citizens, they would lead the way in helping to modernize the societies
they shared with their fellow compatriots. In this task, the military
was uniquely qualified to lead because it was the first sector to utilize
imported technology (i.e., earlier than industry or agriculture), to employ meritocratic standards in recruitment and promotion, to regularly
provide for the basic needs of its members (e.g., literacy, food, shelter),
and to offer programs associated with the modern welfare state (e.g.,
insurance, pensions, family benefits). The armed forces came to be seen
as leaders in modernization that would transform traditional societies
into well-organized ones that could provide for citizens as well as the
military did for its own.
Seen as either separate or integral to civilian society, as either temporary solutions to crisis or long-term forces for change, the armed
forces came to occupy a special role in which they could claim to stand
for the �…
Purchase answer to see full
attachment