ESSAYS ON TRADE OPENNESS AND AUTHORITARIAN SURVIVAL
By
Wen-Chin Wu

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Political Science
2012

ABSTRACT
ESSAYS ON TRADE OPENNESS AND AUTHORITARIAN SURVIVAL
By
Wen-Chin Wu
This dissertation consists of three essays that integrate theories in the fields of
international political economy and democratization to advance the understanding of
trade politics under dictatorships. The core argument is that dictators can utilize trade
policies to prolong their political survival. Thus, economic globalization inhibits, not
facilitates, democratization.
In the first essay, I address a substantive question: What factors motivate dictators to
expand the trade regime of their countries? Scholars have been focusing on the difference
in trade openness between democracies and dictatorships. They conclude that democracies
have more trade openness than do autocracies. However, these works fail to systematically
assess the high variation of trade openness among dictatorships. In addition, from the
perspective of modernization theory, trade-induced economic growth may facilitate
democratization in dictatorships. Accordingly, it looks like committing political suicide for
political leaders of some authoritarian countries, such as China and Vietnam, to
significantly engage their countries in the world economy. Based on the Heckscher–Ohlin
model of international trade and theories of democratic transitions, I argue that rising
inequality is a key determinant for dictators to expand the trade regimes. In other words,
inequality is a cause, rather than a consequence, of trade openness under dictatorships.

Building on the argument of the first essay, I expect that increase in trade openness not
only reduces inequality but also helps dictators strengthen their authoritarian rule. Thus,
my second essay focuses on the effects of signing preferential trade agreements (PTAs) on
inequality and authoritarian survival. In this chapter, I demonstrate that once a
dictatorship successfully signs PTAs with other countries, it can further reduce both
economic inequality and prospects of regime breakdown. This finding explains why the
some authoritarian countries, China in particular, become more resilient against
democratization after signing numerous PTAs.
The third essay investigates whether types of protectionism also differ by types of
authoritarianism. The current literature finds that single-party dictatorships, due to their
larger winning coalitions, are more open to trade with lower tariffs than other types of
authoritarian regimes, such as personalistic or military dictatorships. In this essay, I offer a
caveat to this research agenda by showing that larger sizes of winning coalition also result
in more complicated tariff schedules of single-party dictatorships. Those complicated tariff
schedules help dictators protect more members of their winning coalitions and consolidate
their authoritarian rule under single-party dictatorships.
By integrating dictators' strategic choice of trade policies into this analytic framework,
I conclude that trade liberalization leads to authoritarian consolidation. Overall, my
dissertation offers an economic anatomy of authoritarian survival to the ongoing research
agenda on dictatorships in the age of globalization.

To my late grandmother and late sister, who taught
me virtues of courage, kindness, and fortitude.

iv

ACKNOWLEDGMENTS

In the journey of researching and writing this dissertation, I have been fortunate to
receive considerable support and encouragement from many sources. First and foremost, I
would like to express my deepest gratitude to my advisor, Eric C. C. Chang. Throughout
my Ph.D. studies at Michigan State University (MSU), Eric has been ceaselessly supporting
me to become a political scientist with critical thinking and professional skills. His
mentoring always helps me find solutions to my personal and academic predicaments.
Interactions with Eric have been the most valuable assets to my career development.
My sincere gratefulness also goes to other members of my dissertation committee,
Cristina Bodea, Corwin D. Smidt, and Susan Chun Zhu. Their advices on substantive and
methodological issues inspire me to write my dissertation in a much better way. In addition,
I would like to thank Paul Abramson, Michael Bratton, Robert Franzese, Szu-yin Ho,
Daniel Lee, Tse-Kang Leng, Jih-wen Lin, Ting Liu, Carlos Pereira, and George Tsebelis for
their encouragement, guidance, and insights at different occasions. I also appreciate Tim
Büthe, Daniel Kono, Joseph Wright, and Jong-Sung You for their generosity of sharing data
with me. Furthermore, I thank Steve Kautz and Karen Battin for their timely help in every
administrative need. I am grateful for the financial support from the Afrobarometer Project,
Department of Political Science and Graduate School at MSU, and the Ministry of
Education in Taiwan.
Special thanks to my cohorts and friends. I enjoyed the casual and professional
conversations with Jenny Cairns, Maggie Chang, Tse-Hsin Chen, Seo-Yuan Choi, Luke
v

Chu, Tse-Min Fu, Sung-Min Han, Masaaki Higashijima, Hao-Yu Hsu, Alex Chienwu
Hsueh, Tung-Wu Hsueh, Shih-Hao Huang, Jason Kuo, Helen Lee, Thomas Lee, Dominique
Lewis, Bob Lupton, Chunho Park, Szu-Ning Ping, Francis Smart, Yung-Hsiu Tang, Po-Hsin
Tseng, Charles Chong-Han Wu, and Chia-Hsin Yeh. Without these people, my life at
Michigan would have been more humdrum and tougher.
I offer my most heartfelt appreciation to my family. My parents, Jin-Ben Wu and
Tsui-Chin Chen, supported me to finish my college education under stringent financial
conditions. During my graduate studies, my brother (Wen-Shen Wu) and sister-in-law
(I-Ling Chen) have taken over all family responsibilities that I am supposed to share.
Without the babysitting offered by my mother, mother-in-law (Mei-Jiou Soong), and
cousin (Chih-Ching Wu), I would have been unable to finish this dissertation before the
one-year-old birthday of my son, Infinity Chiung-Jing Wu. I owe the most to my beloved
wife and helpful colleague, Hsin-Hsin Pan. She endlessly supports me and spends most of
her time taking care of Infinity since he joined our family in September 2011. There is no
way for me to complete studies at MSU without her. Finally, I would like to dedicate this
dissertation to my late grandmother (Ju-Ying Huang Wu) and late sister (Pei-Ling Wu),
who taught me virtues of courage, kindness, and fortitude.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. viii
LIST OF FIGURES .............................................................................................................. ix
Chapter 1
1.1
1.2
1.3
1.4

Introduction .................................................................................................... 1
Three Puzzles about Trade Openness under Dictatorships .................... 4
The Arguments .................................................................................. 10
Outline of the Dissertation................................................................. 17
Contributions .................................................................................... 19

Chapter 2
2.1
2.2
2.3
2.4

Inequality and Trade Openness in Authoritarian Countries ........................... 22
Inequality, Trade Openness, and Authoritarian Survival .................... 28
Empirical Analysis ............................................................................. 35
Empirical Results ............................................................................... 41
Discussion and Conclusion ................................................................ 54

Chapter 3
3.1
3.2
3.3
3.4

Preferential Trade Agreements and Authoritarian Survival ............................ 57
Trade Openness and Regime Transition ............................................. 61
International Trade and Authoritarian Survival .................................. 65
Data, Operationalization, and Empirical Analysis .............................. 71
Discussion and Conclusions ............................................................... 81

Chapter 4
4.1
4.2
4.3
4.4

Disguised Protectionism under Authoritarianism .......................................... 85
Theories of Trade Liberalization under Dictatorships ......................... 91
Tariff Dispersion under Single-Party Dictatorships ............................. 95
Empirical Analysis ........................................................................... 101
Discussion and Conclusion .............................................................. 110

Chapter 5

Conclusion .................................................................................................. 113

APPENDIX … .................................................................................................................. 119
BIBLIOGRAPHY ............................................................................................................. 126

vii

LIST OF TABLES

Table 2.1

Determinants of Trade Openness under Dictatorships .................................. 43

Table 2.2

Determinants of Trade Openness under Dictatorships .................................. 46

Table 2.3

Effects of Inequality on Trade Openness ...................................................... 48

Table 3.1

Effects of Signing PTAs on Economic Inequality .......................................... 74

Table 3.2

Effects of Signing PTAs on Authoritarian Breakdown .................................. 79

Table 4.1

Authoritarianism, Single-Party Dictatorships, and Tariff Dispersion ........... 109

Table A.1

List of Dictatorships Analyzed in Chapter 2 .............................................. 120

Table A.2

Summary Statistics of Variables Used in Chapter 2 .................................... 121

Table A.3

List of Dictatorships Analyzed in Chapter 3 .............................................. 122

Table A.4

Summary Statistics of Variables Used in Chapter 3 .................................... 123

Table A.5

Geddes’ Coding of Authoritarian Regimes, 2000–2009 ............................. 124

Table A.6

Summary Statistics of Variables Used in Chapter 4 .................................... 125

viii

LIST OF FIGURES

Figure 1.1

Average Trade Volume by Regime Types, 1970–2008 .................................... 2

Figure 1.2

Average Trade Volume and Level of Democracy, 2000–2008......................... 3

Figure 1.3

Mean Tariff and Dispersion of Tariff under Dictatorships ........................... 10

Figure 2.1

Trade Openness among Democracies and Autocracies, 2000–2008 ............. 23

Figure 2.2

Predicted Probability of Trade Openness under Dictatorships ...................... 50

Figure 2.3

First Difference Measuring the Effect of Inequality on Trade Openness ........ 51

Figure 3.1

Number of Signed PTAs of Democracies and Dictatorships in 2000 ........... 58

Figure 3.2

Signing PTA and Differenced Predicted Probability of Regime Breakdown.. 80

Figure 4.1

Tariff Dispersion and Authoritarianism under Dictatorships, 2000–2009 ... 88

ix

CHAPTER 1
INTRODUCTION
1

This dissertation investigates trade politics under dictatorships. The literature has
developed theories to explain how different democratic institutions result in varieties of
trade policies (Grossman and Helpman 1996; Rogowski 1987), and how trade policies
differ between democracies and autocracies (Mansfield, Milner, and Rosendorff 2000). Yet,
we have limited understanding about the determinants of trade policymaking in
authoritarian countries. However, it is crucial for researchers to complete this missing piece,
because authoritarian countries, as Figure 1.1 illustrates, have been further involved in the
2

world economy in the age of globalization. Although democracies tend to trade more
than autocracies (Aidt and Gassebner 2010; Milner and Kubota 2005), Figure 1.1 suggests
that the average trade volume of autocracies began to increase since the early 1990s,
especially after China entered the World Trade Organization in late 2001.
Nevertheless, the surge of autocracies’ trade in the 21th century is not solely
contributed by China. In Figure 1.2, I plot each country’s average trade volume against its
3

average level of democracy during 2000 and 2008. As Figure 1.2 shows, many autocracies

1

Throughout this dissertation, I use dictatorship, autocracy, and authoritarian country
interchangeably.
2
In Figure 1.1, I use a binary variable to measure democracy and dictatorship based on
the dataset of Global Political Regimes constructed by Geddes, Wright, and Frantz (2012).
3
In Figure 1.2, I use a continuous variable to measure democracy and dictatorship. The
measure of level of democracy is taken from the Polity IV Project of Marshall and Jaggers
(2002) and normalized to an interval between 0 and 1, with 0 indicative of perfect
autocracy and 1 indicative of perfect democracy.
1

250
Trade (Current Billions USD)
50
100
150
200

Democracies

0

Dictatorships

1970

1990
2000
2010
Year
Sources: Correlates of War Project; Global Political Regimes (Geddes
et al. 2011)
Figure 1.1

1980

Average Trade Volume by Regime Types, 1970–2008

trade as much as their democratic counterparts, and some democracies even have lower
trade than other autocracies. In other words, although the literature concludes that
democracies, on average, trade more than autocracies, it is still unclear why autocracies
trade and some of them trade a lot.

2

Logged Trade Volume (Current Billions USD)
15
20
25
30

China

1
.75
.5
Level of Democracy
Sources: Correlates of War (COW) Project; Polity IV Project (Marshall
and Jaggers 2002)
0

Figure 1.2

.25

Average Trade Volume and Level of Democracy, 2000–2008

Inspired by the variation of trade openness among dictatorships, this dissertation
investigates the following three research questions: First, why do some authoritarian
leaders expand trade regimes of their countries while others do not? Second, what are the
consequences of trade openness on authoritarian politics? Third, does the pattern of
protectionism differ by types of authoritarianism? In this dissertation, I argue that the logic
of political survival can provide a theoretical framework to answer all of the three
questions. Specifically, authoritarian leaders can utilize trade policies, such as signing
preferential trade agreements or differentiating tariffs schedules, to secure their
incumbency. In this introductory chapter, I will elaborate on how to apply the logic of
3

political survival to analyze trade politics under dictatorships.

1.1

Three Puzzles about Trade Openness under Dictatorships
The main puzzle that motivates this dissertation is the variation of trade openness

across authoritarian countries. On the one hand, scholars argue that democracies are less
able to sustain protectionism than autocracies, because electoral competition makes
political leaders no longer able to rely on rents created by protectionism when financing
their incumbency. Instead, they have to expand the trade regimes of their countries to
benefit more citizens and garner more political support (Bueno de Mesquita et al. 2005;
Milner and Kubota 2005).
On the other hand, however, proponents of modernization theory contend that the
economic growth induced by trade openness may promote democratization (Lipset 1959),
because the newly rich middle class will demand more political rights to protect their
property from state intervention (Ansell and Samuels 2010; Moore 1966). Thus,
authoritarian leaders may intentionally isolate their countries from the world market and
trade less than democracies do (Aidt and Gassebner 2010; Banerji and Ghanem 1997).
However, there is an inconsistency between theory and evidence regarding trade
openness among authoritarian and democratic countries. Scholars do find that
democracies tend to trade more than their autocracies. Meanwhile, many dictatorships
also have high levels of trade openness. If more trade implies more surrender of
protectionism rents, and if trade-induced economic growth facilitates democratization,
then why do some authoritarian countries still trade a lot more than many democracies?
Obviously, the current literature fails to explain the huge variance of trade openness
4

among authoritarian countries. Therefore, we need a theory to answer the following
questions: Why do some authoritarian leaders choose to open their trade regime while
others do not? Are those authoritarian leaders choosing to open the trade regime not afraid
of being threatened by the forces demanding democratization? Or, do they have other
concerns that enforce them to open their trade regimes, regardless the threats of
democratization? I call this set of questions the trade openness puzzle.
Answers to the trade openness puzzle can not only resolve the inconsistencies between
theoretical implications and empirical evidence discussed above, but also help researchers
and policymaker to better understand the normative dimension of trade openness.
Specifically, theories of international trade have demonstrated that engaging in
international trade eliminates price distortion as well as deadweight loss generated by
protectionism, and thereby enhances the aggregate welfare of both trading countries
(Krugman and Obstfeld 2006). This effect of welfare enhancement is the main ground for
politicians and economists to promote free trade after the Second World War (Irwin 1996,
2009). Although some other scholars have forcefully criticized the “dark side” of free trade
(Stiglitz 2003), such as the deteriorated terms of trade and widened economic inequality in
developed countries, their criticisms mainly center around the distributional effects and
preconditions of free trade, not on free trade per se (Rodrik 1997).
More importantly, there is a normative reason to study trade politics under
dictatorships, because over one-third of the world’s population lives in authoritarian
countries (The Economist 2010). The move to free trade would generate significant
political and economic consequences for those citizens’ day-to-day life.
The trade openness puzzle is related to what scholars have been debating now:
5

Whether or not trade openness will promote democratization in authoritarian countries.
Empirical evidence is inconclusive regarding this question (Milner and Mukherjee 2009).
Take East Asian countries for example. One the one hand, some East Asian countries,
including Taiwan and South Korea, adopted export-oriented trade policy and
democratized after their economies prospered in the late 1980s. Both cases suggest that
trade openness may facilitate democratization. On the other hand, however, there are still
other authoritarian countries with extreme high trade openness in East Asia, including
China, Malaysia, and Singapore. Political scientists even find that the authoritarian control
in China and Malaysia becomes more resilient after both countries further engage their
economies in the world market (Nathan 2003; Pepinsky 2009).
In addition to East Asian countries, empirical results of large-N cross-national studies
are inconclusive on the relationship between trade openness and democratization. On the
one hand, some scholars find evidence to support the proposition that trade fosters
democratization (Eichengreen and Leblang 2008; Epstein et al. 2006). On the other hand,
however, others provide counterevidence to challenge their findings (Li and Reuveny 2003;
Teorell 2010). As Milner and Mukherjee (2009, 170) summarize, “More sophisticated
empirical work needs to be done to carefully evaluate the effect of trade openness on
democracy in developing countries.”
In this dissertation, I regard these inconsistent results about the relationship between
trade openness and democratization as the democratization puzzle. Similar to the trade
openness puzzle, there are both positive and normative reasons to investigate this
democratization puzzle. First, answers to the democratization puzzle can advance the
understanding of trade openness puzzle. If trade facilitates democratic transitions, then we
6

can explain why many authoritarian countries stay isolated from the world economy. By
contrast, if trade strengthens authoritarianism under dictatorships, then it becomes
theoretically intriguing that many dictatorships still embrace protectionism.
Second, the issue on whether trade facilitates or inhibits democratization matters to
ordinary citizens under dictatorships, since transitions to democracies usually enhanced
politicians’ accountability as well as citizens’ welfare, such as lower infant mortality rates
and better education opportunities (Gerring, Thacker, and Alfaro 2012; Stasavage 2005).
In other words, answers to the democratization puzzle give a moral foundation for
developed countries to promote free trade among developing and less developed countries.
In the next section, I will elaborate on how to deal with this democratization puzzle from
the theoretical perspective of political survival.
The third puzzle this dissertation intends to explore is the relationship between
protectionism and the typology of dictatorships. Scholars in the field of comparative
democratization have proposed many topologies to classify dictatorships, and Geddes’
typology is widely applied in the literature. Specifically, Geddes (1999a, 1999b) argues that
there are four types of dictatorships: Personalistic, military, monarchy, and single-party
dictatorships, Furthermore, she claims that each type of dictatorships generates its own
policy outcomes. For instance, Geddes (1999b, 135) finds that single-party dictatorships
usually have longer horizons of regime survival, because “their institutional structures
make it relatively easy for them to allow greater participation and popular influence on
policy without giving up their dominant role in the political system.” By contrast, other
scholars find that personalistic dictatorships engage in less efficient use of foreign aids as
well as more corruption (Chang and Golden 2010; Wright 2010), because personalistic
7

dictators have to maintain smaller sizes of constituencies and perceive shorter time
horizons of their incumbency.
Based on Geddes’ typology of dictatorships, many scholars have demonstrated that
single-party dictatorships are more open to trade with lower tariffs than other types of
dictatorships (Hankla and Kuthy 2012; Milner and Kubota 2005). The causal mechanism
is that political leaders under single-party dictatorships have to maintain larger winning
coalitions to sustain their incumbency. Since politicians are less able to finance their
incumbency with private goods when the sizes of their winning coalition increase, they
have to expand trade regimes to benefit more citizens. However, one remaining puzzle is
whether we can conclude that single-party dictatorships are less protectionism-oriented
given their higher trade openness and lower tariff rates. As Kono (2006) has forcefully
demonstrated, protectionism can be exercised in more sophisticated ways beyond mean
tariff levels when countries face a higher demand of free trade. Specifically, although
democracies have lower tariffs than dictatorships due to electoral competition, Kono (2006)
finds that they establish more non-tariff trade barriers that are more obscure to voters.
Since dictatorships generally rely more on protectionism to sustain their authoritarian rule
(Aidt and Gassebner 2010), it might be the case that single-party dictatorships just use
other forms of protectionism when they have to lower their mean tariffs.
Figure 1.3 further illustrates the issue of tariffs of dictatorships. I plot each
dictatorship’s mean tariff against its tariff dispersion during 2000 and 2009. As illustrated
in Figure 1.3, single-party dictatorships have lower mean tariffs, but at the same time their
tariff schedules seem to be more dispersed than other types of dictatorships. Although a
lower mean tariff is indicative of less protectionism, a higher level of tariff dispersion is
8

regarded as protectionism because it suggests that some products receive more protection
than others (Deardorff 2006). As a result, it is puzzling to observe that single-party
dictatorships have lower but more dispersed tariffs. I call these inconsistent trade policies
as the tariff puzzle.
The tariff puzzle is related to the first two puzzles I introduced in this section. It also
provides additional evidence to suggest that trade openness puzzle is intertwined with the
democratization puzzle. As single-party dictatorships are found to have longer time
horizon of regime survival with more trade openness, it is argued that regime longevity
contributes to trade openness under single-party dictatorships (Hankla and Kuthy 2012;
Milner and Kubota 2005). Both arguments imply that trade would eliminate, not facilitate,
the momentum of democratization under single-party dictatorships. Therefore, answers to
the tariff puzzle may further suggest that trade policy can be used to consolidate
politicians’ incumbency. I will elaborate on how the logic of political survival can explain
this tariff puzzle in the next section.

9

10
8
6
4
0

2

Mean Tariff

0

2

4

6

8

10

Tariff Dispersion
Single Party

Personalistic

Monarch

Military

Figure 1.3 Mean Tariff and Dispersion of Tariff under Dictatorships

1.2

The Arguments
The

three

puzzles

about

the

relationships

between

trade

openness

and

authoritarianism presented in the previous section are inter-connected. Specifically, they
can be explained under the logic of political survival in authoritarian regimes. In the past
two decades, scholars specializing in the field of comparative democratization have been
focusing on the ways how dictators maintain their dominant incumbency and achieve
“authoritarian consolidation.” (e.g., Brownlee 2007; Gandhi 2008b; Levitsky and Way
2010; Magaloni 2006). For example, one of the arguments is that the establishment of
quasi-democratic political institutions, legislatures and one-party dominance in particular,
10

helps authoritarian leaders not only signal credibility but also co-opt or collapse the
opposition forces. Thus, the democratization prospects in those countries are doomed as
those quasi-democratic institutions are established.
The argument that quasi-democratic institutions would consolidate authoritarian rule
sounds counterintuitive, but it is in fact reasonable after further investigation. Like their
counterparts in democracies, politicians in authoritarian countries have to get compliance
from citizens in order to stay in power (Tullock 1987). However, there are usually no
certain rules, such as regular elections, to replace top political leaders in authoritarian
countries. Thus, the main task of authoritarian leaders is to make sure that their ruling
status is solid enough to avoid coups, rebellions, or revolutions (Brownlee 2009; Tullock
1987). Put differently, the key to sustain political survival for authoritarian leaders is to
make their potential challengers’ costs of involving in opposition movements outweigh the
benefits of successful replacing the incumbent dictators (Bueno de Mesquita et al. 2005).
Wintrobe (1998) synthesizes this issue of political survival as the dictator’s choice
between repression and loyalty. On the one hand, dictators can rely on violence, such as
secret police or military, to deter and destroy potential challengers. On the other hand, they
can enhance citizens’ benefits of living under the authoritarian control by providing
patronage or public goods in exchange for their political support. Thus, the question is
why some authoritarian leaders can, while others cannot, make credible threats of
initiating violence and/or commitments of improving benefits in front of their citizens to
survive the challenges against their ruling status.
Scholars have been investigating the commitment problem of authoritarian survival
for decades (Myerson 2008; Olson 1993; Tullock 1987; Wintrobe 1998), and one
11

prevalent perspective focuses on the role of political institutions (Gandhi 2008b; Levitsky
and Way 2002; Svolik 2009). Overall, this institutionalist’s approach emphasizes on the
intention of authoritarian leaders to make credible commitment through establishing
political institutions, such as judiciaries, legislatures, or regular elections. With political
institutions, authoritarian leaders not only “bind their hands” to limit their capacity of
seizing citizens’ property (North and Weingast 1989), but also share their power as a way
to co-opt opposition forces in exchange for the minimum stability of the authoritarian
control (Gandhi and Przeworski 2006, 2007). Using Olson’s terminology, we can say some
authoritarian leaders establish institutions to become “stationary bandits” who protect
their citizens from being looted by other “roving bandits” (Olson 1993, 2000). Thus, those
dictators belonging to the type of stationary bandits can get the compliance of their citizens
and successfully dominate their authoritarian regimes.
Similar to the institutionalist’s argument discussed above, in this dissertation I argue
that trade policy is also included in the toolbox of dictators to sustain their political
survival. Thus, my answer to the trade openness puzzle is straightforward: Authoritarian
leaders choose to expand their trade regimes to ease the pressure of democratization
incurred by rising economic inequality. Scholars have been arguing that economic
inequality fosters the momentum of democratization. In particular, as the median voter is
the voter whose preference determines the results of election (Downs 1957), the median
voter may demand more redistribution from the rich people as income inequality increases
(Acemoglu and Robinson 2005; Boix 2003; Meltzer and Richard 1981). In other words,
higher economic inequality gives the poor more incentives to demand more redistribution,
resulting in democratic transitions that empower the poor median voters to set higher
12

taxes rates on their rich fellow citizens.
Although the role of median voter in authoritarian regimes is usually absent or weak
due to non-free or unfair elections (Schedler 2006), the impacts of economic inequality is
still a significant issue in authoritarian regimes, since high inequality implies “enough of a
redistributive reward to make revolutionary activity extremely profitable” (Schofield and
Levinson 2008, 269). One recent study also demonstrates that countries with high
inequality are more likely to engage in organized conflicts (e.g., civil wars and armed
violence), since the poor people expect that they can significantly benefit from
expropriating through violence (Boix 2008). Accordingly, to prevent inequality from
fostering anti-regime momentum, authoritarian leaders use either “carrots” to pacify or
“sticks” to suppress the poor people (Gallagher and Hanson 2009), and trade openness
belongs to the former category.
In this dissertation, I argue that increases in inequality compel authoritarian leaders to
consider expanding their trade regimes as an alternative to redistribution without taking
money from the rich. According to the Heckscher-Ohlin model, international trade will
benefit domestic abundant production factors, because the principle of comparative
advantage makes the production costs lower for the abundant factor (Krugman and
Obstfeld 2006). In addition, the owners of abundant factor will earn higher wages because
their products can be exported to other countries where the factor used in this product is
scarce. This effect is articulated in the well-known Stopler-Samuelson theorem (Stolper and
Samuelson 1941).
Based on the implications of Heckscher-Ohlin model and the Stopler-Samuelson
theorem, I develop the following theory to explain the trade openness puzzle. Since
13

authoritarian countries are usually abundant with poor and unskilled labors (Milner and
Kubota 2005), engaging in international trade brings wage premiums to the poor, alleviate
economic inequality, and then eases the pressure of democratization. In addition, while
direct income transfer may incur the backlash or capital flight of the rich (Boix 2003),
expanding trade regimes helps authoritarian leaders enrich the poor without taxing the
rich in their countries. As a result, trade openness is a better choice than taxation to
address rising economic inequality in authoritarian regimes, especially for those with
abundant labors.
In short, this dissertation focuses on how increases in inequality facilitate trade
openness that can stop inequality from rising. Previous studies mainly focus on the political
and economic consequences of inequality level, but they ignore the impacts of changes in
inequality. However, rising inequality may be more salient than high inequality for
authoritarian politics. Authoritarian leaders may be well-equipped to deal with opposition
forces when inequality is high, but when inequality is higher, it means that they need to
invest more to prevent potential uprisings. In short, increases in inequality change the
equilibrium conditions of the status quo and enforce the dictators to update their policy
tools to ensure their ruling status.
To sum up the discussions in this section so far, my hypothesis about the trade
openness puzzle can be formulated as follows:


Hypothesis 1: As income inequality increases, trade openness also increases in
authoritarian countries.

Based on Hypothesis 1, I derive another hypothesis to disentangle the democratization
puzzle. While the literature is inconclusive on whether trade will facilitate democratic
14

transitions, Hypothesis 1 implies that trade expansion would be beneficial to the
incumbency of political leaders for its alleviation of inequality. Nevertheless, it is difficult
to empirically test this implication of Hypothesis 1, since there is an issue of endogeneity
between trade and democratization (Milner and Mukherjee 2009).
Scholars have applied econometric techniques to deal with the endogeneity between
trade and democratization. For example, Eichengreen and Leblang (2008) estimate with
instrumental variable models to show that trade facilitates democratization. This finding is
further concurred by López-Córdova and Meissner (2008) when they use the generalized
method of moments (GMM) estimator. However, the empirical models estimated by Li and
Reuveny (2003) and Teorell (2010) suggest that more trade further impedes the prospects
of democratization. In addition, Ahlquist and Wibbels (2012) present their empirical
results derived from dynamic probit models and conclude that trade expansion does not
affect

democratic

transitions. Again,

those

inconclusive

findings

generate

the

democratization puzzle.
In this dissertation, I use signing preferential trade agreements (PTAs) as a proxy of
trade to disentangling the democratization puzzle. Note that my answers to the trade
openness puzzle suggest that trade policy is included in politicians’ toolbox to reduce
inequality and secure their incumbency. Since PTAs can further boost the trade volume of
their signatories (Baier and Bergstrand 2007), authoritarian leaders may intentionally sign
PTAs with other countries to prolong their political survival.
Based on this reasoning, the democratization puzzle can be explained by the following
causal mechanism. When the poor people’s demand for redistribution is met via gains from
trade induced by PTAs, authoritarian leaders can neutralize democratization forces and
15

secure their incumbency. Thus, I derive the following two hypotheses regarding the
democratization puzzle:


Hypothesis 2: Signing PTAs reduces economic inequality under dictatorships.



Hypothesis 3: Signing PTAs reduces the probability of authoritarian breakdown.

Strictly speaking, while the first two puzzles emphasize on determinants and
consequences of trade openness, the main focus of the tariff puzzle is protectionism under
dictatorships. Nevertheless, we can still apply the logic of political survival to explain the
tariff puzzle. In this dissertation, I propose the following causal mechanism to explain this
puzzle. According to the selectorate theory (Bueno de Mesquita et al. 2005), authoritarian
leaders subject to larger winning coalitions have to reduce the overall tariffs. However, it
does not mean that they will totally renounce protectionism and exclude those special
interests groups from their winning coalitions. Instead, they need to further differentiate
their protection on members of their winning coalitions. Thus, a dispersed tariff schedule is
just a result from differentiated responses of authoritarian leaders to members and
non-members of their winning coalitions.
Meanwhile, it should be emphasized that authoritarian leaders’ capacity to
differentiate their responses to special interests groups is also determined by their capacity
to exercise their authoritarianism. If a dictatorship is more capable of exercising
authoritarianism, defined by its capacity to repress citizens and the level of institutionalize
autocracy, to intervene civil society, it is more capable of levying taxes to maximize its
revenues. Since tariffs are taxes on imports, dictatorships with higher levels of
authoritarianism would be more able to set tariff schedules that consist of differentiated
and dispersed tariff rates.
16

To summarize my discussions on the tariff puzzle, in this dissertation I will test the
following two hypotheses regarding tariffs dispersion under dictatorships:


Hypothesis 4: Tariff schedules are more dispersed under single-party dictatorship
than under personalistic, military, or monarchy dictatorships.



Hypothesis 5: The level of tariff dispersion increases with the level of
authoritarianism.

1.3

Outline of the Dissertation
To empirically test the hypotheses proposed in the previous section, this dissertation is

organized as follows. Following this introductory chapter, Chapter 2 investigates how
authoritarian leaders employ trade openness as a response to rising inequality. Based on
the Heckscher-Ohlin model of international trade and models of democratic transitions, I
argue that unskilled laborers under authoritarian regimes can benefit from engaging in
international trade thus becoming more compliant to the authoritarian rule as their
countries integrate into the world economy. Accordingly, dictators in labor-abundant
countries expand trade to neutralize democratization threats initiated by rising inequality.
This argument is supported by the data of 80 odd authoritarian regimes from 1963 to
2003 and different model specifications. More importantly, the issue of endogeneity
between trade and inequality is also addressed by dynamic panel data and instrumental
regression models in this chapter. In short, the findings in Chapter 2 suggest that inequality
is a cause, rather than a consequence, of international trade in authoritarian regimes.
Chapter 3 explores consequences of signing preferential trade agreements (PTAs) in
authoritarian countries. I argue that dictators choose to sign PTAs as a means of
17

consolidating their authoritarian rule. Specifically, PTAs reduce economic inequality by
enriching unskilled poor laborers without taxing rich citizens under dictatorships.
Furthermore, the gains from trade introduced by PTAs help dictators acquire more
compliance from citizens and prolong their authoritarian rule. Thus, the threat of
democratization triggered by widening economic inequality is neutralized after signing
PTAs. With data from 88 authoritarian regimes from 1969 to 2000, I demonstrate that
signing PTAs reduces both economic inequality and the probability of authoritarian
breakdown, thereby contributing to ongoing debates on the effects of economic
globalization on democratization.
While the literature concludes that single-party dictatorships are more open to trade
with lower tariffs, Chapter 4 of this dissertation investigates why tariffs under single-party
dictatorships are more dispersed and complicated than other types of dictatorships. I argue
that when dictators have to maintain larger winning coalitions under single-party
dictatorships, they still need to satisfy the demand for protection from special interests
groups included in their winning coalitions. Accordingly, dictators under single-party
dictatorships set ad hoc tariffs in response to different members of their winning coalitions.
In addition, I argue that political leaders under dictatorships with more consolidated
authoritarianism, defined by their coercive capacity and the level of authoritarian
institutionalization, are more capable of designing tariff schedules to optimally collect
revenues from imports. Both arguments on the relationship between tariff dispersion and
authoritarian politics are supported by the panel data on 42 dictatorships from 2000 to
2009. The findings of this chapter contribute to the emerging research on authoritarian
trade politics in the age of globalization.
18

In Chapter 5, I discuss implications of main findings in this dissertation and offer
some concluding remarks. Overall, this dissertation offers an economic anatomy of trade
politics under dictatorships to studies on the authoritarian resilience in the age of
globalization.

1.4

Contributions
Based on the discussions in this chapter, this dissertation has the following

contributions to the literature of democratization and international political economy.
First, this dissertation can clarify the relationships among inequality, trade openness,
and authoritarian survival. The current debates over the relationship between trade
openness and democratization neglect the role of inequality. However, the democratization
literature argues that high inequality fosters the momentum of democratic transition, so it
is necessary to combine both fields of studies together to resolve the theoretical and
empirical inconsistencies about the impacts of trade openness on democratization.
This dissertation disaggregates this issue in two steps. On the one hand, it agrees that
inequality may result in political instability in authoritarian countries, so dictators have to
find some ways to release or oppress the pressure of redistribution. On the other hand,
however, it argues that the pressure of redistribution can be offset by means of trade
openness instead of democratization under dictatorships. Since the poor’s demand for
redistribution can be met through trade openness, politicians can use trade policies,
including signing PTAs, to consolidate their incumbency via distributing gains from trade
to the abundant poor laborers.
The theoretical argument that trade openness can be used to sustain political survival
19

is crucial for the current debates over the impacts of globalization on democratization. In
particular, it offers a novel perspective on the absence of democratic transitions in many
globalized authoritarian states. That is, involvement in economic globalization strengthens,
not weakens, authoritarianism with gains from trade, because citizens are benefited from
international trade and become more compliant to the state. Take China for example. The
authoritarian control in China seems to become more resilient against democratization as
China increases its involvement in the world economy after the early 1990s (Nathan 2003),
and its middles class becomes more conservative as Chinese economy grows (Unger 2006).
Thus, this dissertation challenges the implication of modernization theory elaborated by
Lipset (1959) as well as the theory of “democratization from the outside-in” proposed by
scholars of international relations (Pevehouse 2005).
Second, this dissertation contributes to the understanding about the pattern of
protectionism in different types of authoritarian regimes. While the literature concludes
that single-party dictatorships have lower mean tariffs due to their larger winning
coalitions, this dissertation presents a caveat. In particular, it demonstrates that
protectionism can be exercised via tariff complexity rather than tariff level under
single-party dictatorships. This finding complements to the research agenda on politicians’
choices of trade and non-trade barriers under different constraints of political institutions
(Grossman and Helpman 2005; Kono 2006; Rickard 2012).
In addition, the finding that single-party dictatorships have more complicated tariff
schedules contributes to studies investigating the longevity of single-party dictatorships.
While scholars argue that political elites under single-party dictatorships are more willing
to compromise with each other when facing political and economic crises (Geddes 1999b),
20

this dissertation points out an economic foundation for them to make compromises. Put
differently, a complicated tariff schedule can be regarded as a result of compromise among
political elites under dictatorships when they have to liberalize trade regimes, because
special interests groups can lobby different political elites of different factions in the parties
for tariffs that are more favorable to their industries.
To sum up, this dissertation contributes to the ongoing research agenda on the
authoritarian resilience of dictatorships in the age of economic globalization. It clarifies the
impacts of globalization on democratization by demonstrating that trade policies,
including trade liberalization and protectionism, can be included in dictators' toolbox to
secure their incumbency. Therefore, economic globalization will inhibit democratic
transitions under dictatorships.

21

CHAPTER 2
INEQUALITY AND TRADE OPENNESS IN
AUTHORITARIAN COUNTRIES
Economists and political scientists have developed numerous theories to analyze trade
politics within democracies. Yet, only a few pay careful attention to trade politics within
autocracies. This inadequacy is understandable for two reasons. First, autocracies are
generally more isolated from the world economy than democracies, so it is straightforward
to investigate why democracies trade more than autocracies instead of only focusing on
trade politics under dictatorships (e.g., Aidt and Gassebner 2010; Mansfield, Milner, and
Rosendorff 2002; Milner and Kubota 2005). Second, policy-making in autocracies is
usually exclusive for political elites, so it is difficult to empirically test theories aiming to
explain trade politics under dictatorships. However, as Figure 2.1 illustrates, many
autocracies are not only participating in the world economy in the age of globalization, but
4

also have higher levels of trade openness than many democracies. Even though recent
studies have concluded that democracies, on average, engage in more international trade
than autocracies (Aidt and Gassebner 2010; Milner and Kubota 2005), we still need to
explain why some authoritarian countries have higher trade openness than other
democracies and autocracies.

4

The most striking example is China. In 2009 China became the largest exporter of
merchandise in the world, and in 2010 it accounted for 10% and 9% of world
merchandise exports and imports, respectively.
22

200
150
100
50
0

Average Trade Openness

-10

-5

0

5

10

Average Polity Score
Note: Trade openness is measured by trade volume as a percentage
of GDP. Singapore is not shown in the figure (Trade openness: 417;
Polity score: -2.
Figure 2.1

Trade Openness among Democracies and Autocracies, 2000–2008

This is paper proposes a theory to analyze how rising economic inequality affects the
degree of trade openness in authoritarian countries. I argue that rising inequality causes
dictators in labor-abundant countries to increase the level of trade openness as a way to
pacify popular demand for redistribution. In other words, authoritarian leaders facing
rising inequality, which is associated with threats of losing incumbency, will open their
trade regime as redistribution to enrich the poor without taxing the rich. This argument is
supported by the data of 80 odd authoritarian regimes from 1963 to 2003 and different
model specifications.
The argument of this chapter is based on two lines of existing research. One includes
23

theories of democratic transitions developed by Boix (2003) and Acemoglu and Robinson
(2001, 2005). The other derives from the Heckscher-Ohlin model of international trade
(Heckscher and Ohlin 1991). Recent scholarship has noticed about the effect of inequality
on democratization. Extending the redistributive model developed by Meltzer and Richard
(1981), Boix (2003) conceptualizes democratization as a game of redistribution in which
non-rich citizens attempt to impose higher tax rates on their rich counterparts.
Nevertheless, as the poor's demand for redistribution increases with the level of inequality,
the rich are simultaneously more resistant to redistribution in that they do not want to pay
more taxes. Thus, when a society is more equal, rich citizens are more willing to accept
democratization. Otherwise, they will favor oppression on democratization forces.
Accordingly, Boix (2003) claims that democratic transition is more likely to occur when
inequality is lower.
However, Acemoglu and Robinson (2005) disagree with Boix's conclusion. On the
one hand, Acemoglu and Robinson concur with Boix that rich citizens have a strong
incentive to oppress poor citizens' revolutionary threats when inequality is high. On the
other hand, however, they claim that poor citizens do not have enough incentives to
initiate democratization when inequality is low. As a result, democratic transition is most
likely to occur when inequality is in the middle range. Although there is a disagreement on
the consequence of low inequality, both studies agree that higher inequality poses greater
pressures of redistribution and democratization on elites. To secure their authoritarian rule,
dictators need to either meet or suppress the popular demand for redistribution.

5

5

The Boix-Acemoglu-Robinson mechanism of democratic transitions is challenged by
24

The Heckscher-Ohlin model of international trade gives the other theoretical
foundation to my theory that trade openness can neutralize democratization threats. This
model predicts that countries will export goods that intensively use their abundant
production factor, and import products using production factors scarce to them
(Heckscher and Ohlin 1991). One important implication of this model is that owners of
the abundant factor will benefit while those of the scarce factor will suffer from
6

international trade (Stolper and Samuelson 1941). Given that engaging in international
trade creates wage differentials among owners of different production factors, Rogowski
(1989) argues that different political coalitions will emerge along with the cleavage of
production factors to support or resist protectionism. In particular, owners of abundant
production factors support free trade while owners of scarce factors oppose it. Based on
these insights, in this chapter I expect that labor-abundant dictatorships would be more
likely to participate in the world economy when their political leaders face more threats of
democratization implied by rising inequality.
The argument of this chapter is close to Ahlquist and Wibbels' recent study on the
relationship among trade, inequality, and democratization. Also based on the
Boix-Acemoglu-Robinson mechanism of democratic transition, Ahlquist and Wibbels
(2012) argue that countries with similar labor endowment face similar pressures of regime
breakdown when an exogenous shock hits the international economy, because political

other studies (e.g., Ahlquist and Wibbels 2012; Houle 2009). I will discuss them in the next
section.
6
For example, China is abundant with labor but scarce with capital, so it will export
labor-intensive textiles to and import capital-intensive airplanes from capital-abundant
countries, such the United States. Thus, the wages of laborers in China and the returns to
capital owners in the United States both increase after they trade with each other.
25

leaders of those countries would bear a similar burden of redistribution imposed by the
common shock. Thus, more involvement in international trade facilitates democratization
in labor-abundant dictatorships but not in labor-scarce ones. However, Ahlquist and
Wibbels (2012) do not find empirical support for this argument, casting a serious challenge
to distributive models of democratic transitions.
Nevertheless, the finding presented by Ahlquist and Wibbels (2012) is compatible
with the argument of this chapter. Specifically, Ahlquist and Wibbels (2012) provide a
“bottom-up” perspective to analyze how laborers' demand for redistribution facilitates
democratization via trade, whereas this chapter takes a “top-down” perspective by
focusing on political leaders' incentive to take advantage of international trade conditional
on the labor endowment of their countries. When facing more demand for redistribution,
dictators may be tempted to attract more external resources to ease their budget
constraints instead of relying on taxation. Thus, it is reasonable to observe that more
engagement in international trade does not facilitate democratic transitions in
labor-abundant dictatorships. Instead, we should expect that dictators facing rising
inequality will use trade openness in response to the poor's demand for redistribution.
This chapter contributes to the literature in two ways. First, it advances the
understanding of trade politics under dictatorships. Although there are numerous studies
investigating the political economy of trade policies, most of them either focus on
democracies (Goldberg and Maggi 1999; Grossman and Helpman 1994; Park and Jensen
2007; Verdier 1994), or compare trade policies between democracies and autocracies (Aidt
and Gassebner 2010; Mansfield, Milner, and Rosendorff 2000; Milner and Kubota 2005).
In brief, they analyze how the existence of electoral competition and special interests
26

determine the pattern of trade policies across different political regimes. However, by
definition authoritarian politics is mainly determined by non-democratic factors, such as
the preferences of top leaders and their inner circles, so the explanatory power of previous
studies is inadequate to account for the variation of trade openness among dictatorships.

7

This chapter offers a novel perspective by examining how the logic of political survival of
authoritarian leaders affects their decisions to participate in international trade.
Second, this chapter sheds new light on debates over the effects of trade openness on
democratization. Based on the modernization theory articulated by Lipset (1959), scholars
contend that trade openness, or economic globalization in a broader sense, facilitates
democratization after it induces economic growth (Eichengreen and Leblang 2008; Epstein
et al. 2006; López-Córdova and Meissner 2008). However, other scholars argue that
globalization may strengthen authoritarianism under dictatorships, because the state
apparatus becomes financed by external resources and more able to continue its
authoritarian control (O’Donnell 1988). This dissenting view is also supported by
empirical evidence indicating that trade openness does not facilitate democratic transitions
(Li and Reuveny 2003; Teorell 2010). Thus, this chapter adds a new perspective to this
view by demonstrating that trade openness is a strategic response of dictators to ease
inequality and to neutralize democratization threats.
The remainder of this chapter is organized as follows. In the next section, I elaborate
on how inequality affects dictators' decisions to expand trade. Section 2.3 proposes a

7

One of a few exceptions is Hankla and Kuthy (2012). They argue that dictatorships with
longer regime time horizons are more likely to implement trade liberalization. This chapter
complements to their finding by showing that trade openness may further prolongs
authoritarian survival.
27

research design to test the hypothesis of this chapter. In Section 2.4, I present the empirical
results estimated with different modeling techniques and measurements of inequality,
dictatorship, and trade openness. I also deal with the issue of endogeneity between
inequality and trade openness in this section. The empirical evidence in Section 2.4
supports the theory of this chapter. The final section discusses implications of this research
with concluding remarks.

2.1

Inequality, Trade Openness, and Authoritarian Survival
In this chapter, I argue that increasing the level of trade openness is a feasible policy

tool for dictators in labor-abundant countries to meet the demand for redistribution and to
sustain their incumbency. I develop my argument on the basis of two existing theories. The
first includes the Meltzer-Richard model of redistribution and its extensions to models of
democratic transitions (Acemoglu and Robinson 2001, 2005; Boix 2003); the second
includes the Heckscher-Ohlin Model of international trade and its implication to
distributive politics. I will briefly discuss those theories in turn.
The existing theoretical models focus on the role of the median voter in the
relationship between inequality and democratization, arguing that democratization is
triggered by the median voter's demand for redistribution. As it is intuitive to expect that
the demand for redistribution increases with the level of inequality, Meltzer and Richard
(1981) demonstrate that the median voter's preference determines the level of
redistribution. The summary of their model is that when the wage of the mean income
earner is higher than that of the median income earner, the latter would like to support
more taxes on those who are richer and redistribute those taxed money to themselves and
28

other poorer people. By contrast, if the median income earner's wage is higher than the
mean income earner's wage, the median income earner would oppose any redistribution
that reduces their welfare.
Based on the Metzler-Richard model, scholars conceptualize democratization as a
process of redistribution. They argue that economic inequality is one of the driving forces
of democratization that introduces a more comprehensive scheme of redistribution among
the populace (Acemoglu and Robinson 2001, 2005; Boix 2003). Although poor citizens
always prefer higher levels of redistribution in the form of higher taxes on the rich,
democratic transitions cannot occur without the support of the middle class. Since higher
economic inequality implies that the median voter (the median income earner in the
Metzler-Richard model) would benefit from redistribution, it imposes more pressure of
redistribution on the rich and political elites. If the elites fail to respond to the median
voter's demand for redistribution, then democratization will be triggered by revolution and
result in redistribution.

8

The rich citizens, however, have an incentive to resist redistribution and then to favor
suppression against the poor. With this reasoning, Boix (2003) posits a negative linear
relationship between inequality and prospects of democratization. He further points out
that democratization becomes more likely in two scenarios. First, when a society is more
equal, the rich are more willing to accept democratization, because their wealth loss would
be small after redistribution. Second, when the rich can move their assets (e.g., capital)

8

In other words, democracy is an institution in which the poor can ensure that the rich
can fulfill their promise of redistribution in exchange for the poor's self-restraint from
initiating revolution.
29

abroad to avoid taxation, they will be less likely to refuse redistribution implied by
democratization. Freeman and Quinn (2012) extend this thesis further. They find that as an
autocracy is more financially integrated into the world economy, the rich elites can
diversify their asset portfolios in the international market to leverage the risk of being
taxed and thus become less resistant to democratization.
Acemoglu and Robinson (2001) take a different perspective to analyze the
relationship between inequality and democratization. They propose an inverted U-shaped
relationship between two variables. On the one hand, rich citizens have strong incentives
to resist redistribution when inequality is high, so they are more willing to bear costs of
repression rather than accept redistribution. On the other hand, however, poor citizens
have little incentive to initiate revolution when inequality is low, because they will not
benefit much from post-revolution redistribution. Thus, democratization is most likely to
occur when inequality is in the middle level. This conclusion implies that increases in
inequality may facilitate democratization, unless dictators are either confident in their
coercive capacity to repress the demand for redistribution or willing to implement
redistributive policies via taxation.
However, the models of democratic transitions discussed above are challenged from
both theoretical and empirical perspectives. Theoretically, these models overemphasize the
role of the median voter in authoritarian regimes where elections are unfair, or even absent
(Schedler 2006). In addition, those models either ignore the problem of collective action
among the poor to initiate revolution (Houle 2009), or assume that the rich class is so
unified that there is no conflict of economic interests among them (Ansell and Samuels
2010). Empirically, their conclusions are drawn from either data with poor quality or case
30

studies, so the external validity of their arguments is subject to further debate (Alemán and
Yang 2011). Nevertheless, economic inequality is still a significant issue under
dictatorships in that high inequality implies “enough of a redistributive reward to make
revolutionary activity extremely profitable” (Schofield and Levinson 2008). Two recent
studies also demonstrate that countries with high inequality are more likely to incur
organized conflicts and killings, because the poor expects to benefit significantly from
expropriating the rich via violence (Boix 2008; Nepal, Bohara, and Gawande 2011).
Accordingly, to prevent inequality from fostering anti-regime momentum, authoritarian
leaders use either “carrots” to pacify or “sticks” to suppress the poor people (Gallagher
and Hanson 2009), trade openness being a key aspect of the former category in that it can
enrich the poor and alleviate inequality.
The second cornerstone of my argument is the Heckscher-Ohlin Model of
international trade. This model predicts that countries benefit from exporting goods
produced with their domestically abundant factors (Heckscher and Ohlin 1991). Stolper
and Samuelson (1941) make an important extension of this model by demonstrating that
the return to the owner of abundant factors will increase after trade expansion. The logic
here is intuitive: If a production factor is abundant, it makes production costs cheaper than
scarce factors. Thus, countries should utilize their abundant factors to produce goods, and
export those products to countries for whom those factors are scarce. Rogowski (1989)
extends this argument by demonstrating that political coalitions will emerge along with the
cleavage of production factors and take different positions on protectionism. That is,
owners of abundant factors favor free trade whereas owners of scarce factors embrace
protectionism. The implication of Rogowski's argument is critical: International trade has
31

redistributive effects that tend to cause political conflicts among different groups.
In addition to the formation of political coalitions, the Heckscher-Ohlin model implies
that abundant laborers in developing countries can benefit from trading with
capital-abundant developed countries (Baccini 2012; Mayer 1984). Since most
dictatorships happen to be developing or less developed countries, their abundant poor
laborers can earn higher income after trading with developed countries. With trade
expansion, dictators do not need to worry about either pressures of redistribution from the
poor or backlashes against redistribution from the rich. In other words, trade openness can
be regarded as a distributive effort without taxing the rich under labor-abundant
dictatorships, making it a better policy choice for dictators facing rising inequality.
Based on the implications of the Heckscher-Ohlin model, I extend the theoretical
focus in models of democratic transitions developed by Boix (2003) and Acemoglu and
Robinson (2001, 2005) from taxation to trade openness. Since trade expansion enriches
laborers in labor-abundant dictatorships and meets the popular demand for redistribution,
dictators will be tempted to expand trade when inequality is increasing.
It should be emphasized that the focus of this chapter is the increase in inequality
rather than the level of inequality. Previous studies mainly investigate the political and
economic consequences of inequality levels, but they ignore the impact of changes in
inequality. Rising inequality, however, is more salient than high inequality for authoritarian
politics. Dictators may be well-equipped to use force against opposition when inequality is
high, but as inequality is getting higher, they have to either strengthen their control to
prevent potential uprisings or share power to quell unrest. Put differently, increases in
inequality change the equilibrium conditions of the status quo and enforce dictators to
32

update their policy tools to maintain authoritarian control. Thus, we should consider
changes in inequality in addition to the level of inequality when analyzing how inequality
affects authoritarian politics.
The causal mechanism of this article emphasizes the interactive effects between trade
openness and inequality on democratic transition. It is similar to the argument elaborated
by Wibbels and Ahlquist (2012). To disentangle the issue of endogeneity between trade
openness and democratization, Wibbels and Ahlquist use the expansion of world trade as
an exogenous variable to test whether or not engaging in more trade facilitates
democratization in authoritarian regimes. They argue that if economic growth induced by
trade fosters democratization, more exposure to world trade would increase the
probability of democratic transitions in labor-abundant dictatorships. Also, it should be
expected that countries with similar factor endowments face similar pressures of
democratization when there is an exogenous shock to world trade. However, Wibbels and
Ahlquist (2012) find no evidence suggesting a positive effect of labor abundance or
international trade on democratization once the regional forces of democratization are
controlled in their empirical models.
At first glance, the finding of Wibbels and Ahlquist (2012) seems to reject the
argument of this chapter, since it implies that neither trade nor inequality has impacts on
democratization. However, their finding is theoretically compatible with the causal
mechanism proposed here. When dictators face rising inequality, their incentive to
maintain incumbency plays a crucial role. It is possible that trade openness is a strategic
response of dictators to ease their budget constraint of redistribution as inequality rises.
Thus, it is reasonable to observe no positive relationship between trade openness and
33

democratization after strategic reactions. Meanwhile, it can be expected that changes in
trade openness respond to changes in inequality under dictatorships.
One immediate objection to the central thesis of this article is that inequality is a
consequence, not a cause of trade openness. Numerous studies show that trade openness
contributes to increases in inequality (Goldberg and Pavcnik 2007). However, this claim is
far from complete, because other studies find that the impact of trade openness on
inequality is weak or even absent (Anderson 2005; Babones and Vonada 2009). Some
scholars offer an eclectic perspective on this debate, arguing that trade increases inequality
in

developed

countries

but

reduces

it

in

developing

countries

(International Monetary Fund 2008). I adopt this eclectic view in this chapter.
Furthermore, I employ both system generalized methods of moments (GMM) and
instrumental-variable regression models to deal with this issue of endogeneity in the
empirical section of this chapter. The estimation results demonstrate that rising inequality
causes further trade openness under dictatorships.

9

Another objection to the argument of this chapter is that dictators may rely on other
policy tools, such as military oppression and fiscal redistribution, to mitigate the pressure
the inequality. After all, authoritarian leaders by definition have more discretion to govern
their countries than their democratic counterparts for the lack of constraints on their
power (North and Weingast 1989). However, as forcefully pointed out by Boix (2003) and
Acemoglu and Robinson (2005), the rich class will not support political leaders' decision
to implement fiscal distribution by imposing higher taxes rates, so inequality may still

9

Following Higgins and Williamson (1999) and You and Khagram (2005), I use the ratio
between two age groups as an instrument of inequality in the empirical section.
34

prevails. On the other hand, it is costly for dictators to maintain a strong military to
oppress the poor's demand for redistribution. Furthermore, the existence of a strong
military, as argued by Acemoglu et al. (2010), creates a moral hazard problem in which
political leaders may not be able to control the military officers who can initiate military
coup. We should also keep in mind that a strong military itself does not eliminate
inequality but suppresses the demand for redistribution. Therefore, trade policy becomes a
more desirable policy choice than others for dictators to deal with rising inequality,
especially in labor-abundant autocracies.
Based on those discussions, in next sections I empirically test the following hypothesis:


Hypothesis 2.1: As income inequality increases, trade openness also increases
under labor-abundant dictatorships

2.2

Empirical Analysis
To test the hypothesis that increases in inequality facilitate more trade openness in

labor-abundant authoritarian regimes, I compile a dataset covering 80 odd authoritarian
regimes between 1963 and 2003.

10

The unit of analysis is country-year. The list of

authoritarian regimes covered by this chapter is reported in Appendix A.1.
The Dependent Variable. The dependent variable of this chapter is trade openness of
an authoritarian regime in a given year. I operationalize trade openness as one country's
yearly trade volume divided by its GDP (constant price). I multiply the fractional numbers

10

I use the data of authoritarian regimes constructed by Geddes, Wright, and Frantz
(2012). The dataset is available at http://dictators.la.psu.edu. (Accessed on May 31, 2012)
35

by 100 to rescale this variable into a percentage form. A higher value of this variable
implies that a country's economy has more exposure to the world market. The data of this
variable is taken from the Penn World Table (version 6.3) constructed by Heston et al.
(2009). Although this measurement of trade openness is the most commonly used in
literature, it is subject to critique (e.g., Ahlquist and Wibbels 2012; Pritchett 1996). For
example, this variable may just measure the “size” rather than the “openness” of one
country's trade regime. Thus, I also use other measurements of trade openness as
robustness checks.

11

Explanatory Variables. The first explanatory variable of this chapter is income
inequality (Inequality). I use the inequality measure of industrial payment constructed by
James Galbraith and his colleagues (Galbraith and Kum 2003; Galbraith 2012). While
there are other alternative measures of inequality, such as the one developed by Deininger
and Squire (1996), the quality of those Gini-based measurements is usually coarse and
problematic. For example, the widely-used Deininger-Squire Gini index and its extensions
are constructed on the basis of surveys, so the issue of missing data is usually serious,
especially in developing countries. To address the problem of missing values in the
Deininger and Squire's dataset, scholars take the average of inequality data over a certain
time period (Boix 2003; You and Khagram 2005) or impute the missing data (Solt 2009).
However, Galbraith (2012) points out that those methods “involves heroic guesswork,”
because sometimes the imputation is made on the basis of one observation on a country

11

I use the binary variable of trade openness constructed by Sachs and Warner (1995) and
updated by Wacziarg and Welch (2003). I also adopt the KOF index of economic
globalization developed by Dreher (2006). I will discuss the operationalization of those
alternative measurements in the next section.
36

over the whole period under study. Therefore, other scholars try to employ other measures
of inequality (e.g., Galbraith and Kum 2003; Houle 2009).
Using statistics of industrial wages surveyed by the United Nations Industrial
Development Organization (UN-IDO), the indicator of industrial pay inequality developed
by Galbraith and colleagues includes the industrial payment data in 156 countries from
1963 to 2003 (N = 3554, including autocracies and democracies). The inequality dataset
derived from the UN-IDO statistics provides a more complete time-series-cross-sectional
data structure than does the Gini index calculated by Deininger and Squire (1996), so we
do not need to impute the missing data of inequality as other studies do. More importantly,
from the theoretical perspective of this chapter, the industrial pay-inequality measure is
more preferable than the Gini index, because the expansion of trade have more direct
impacts to wages than household income measured by the Gini index.
The Heckscher-Ohlin model implies that wages in labor-abundant countries will
increase after trade expansion, so the second key explanatory variable of this chapter is
Labor Endowment. I use the variable of relative labor endowment calculated by Ahlquist
and Wibbels (2012, 455). This variable measures a country i's labor endowment with the
following formula: (Populationi / World Population) / (GDPi / World GDP) . A larger
value of this variable indicates a higher level of labor abundance, because it means that
country i uses more laborers to create the same output as other countries. Following
Ahlquist and Wibbels (2012), I take the logarithm of this variable to address skewness.
I include a set of control variables in my empirical models. First, I use GDP per capita
(GDPpc, in hundreds of US dollars) to account for the effect of economic development on
trade openness. The data of GDP per capita are taken from the Penn World Table (version
37

6.3). Second, I include two variables measuring a country's endowment of natural
resources, metals and oil in particular. The variables Metal and Oil measure one country's
total production of industrial and precious metals and oil divided by its population size,
respectively. Although it can be expected that countries with abundant natural resources
are more likely to export them to the world market, it can also be argued that those
resources-abundant countries are relatively closed to the world market due to the
underdevelopment of other industries. Thus, I am theoretical agnostic to the signs of both
variables in my empirical estimation. The data on natural resources are taken from Haber
and Menaldo (2011).
Third, I include the membership of GATT/WTO to control for the impact of
international trade agreements on trade openness. While some scholars argue that a
country's participation in the GATT/WTO boosts its trade volume (Goldstein, Rivers, and
Tomz 2007; M. H. Kim 2011), others disagree (Rose 2004). In line with the argument of
this chapter, I expect that the impact of joining GATT/WTO on trade openness is positive
among authoritarian regimes. The data on the GATT/WTO membership is taken from
Ulfelder (2011).
Fourth, I re-investigate how political institutions affect the level of trade openness
under dictatorships. Previous studies argue that single-party dictatorships, due to their
larger winning coalition sizes, have higher trade openness than other types of dictatorships,
including military, monarchy, and personalistic dictatorships (Hankla and Kuthy 2012;
Milner and Kubota 2005). Thus, I expect that single-party dictatorships are associated
with higher trade openness. The data on the typology of authoritarian regimes are taken
from Geddes (1999a).
38

Last, I control for the impacts of authoritarianism on trade openness. Existing studies
find that trade openness is positively correlated with the level of democracy (Aidt and
Gassebner 2010; Milner and Kubota 2005; Milner and Mukherjee 2009). Yet, the
argument of this chapter implies that dictators use trade policies to sustain their
authoritarian rule. Thus, I expect that more authoritarianism will be associated with more
trade openness.

12

I take the variable Autoc from the Policy IV project to measure

authoritarianism (Marshall and Jaggers 2002).

13

Appendix A.2 reports the descriptive statistics of variables used in this chapter.
Model Specification. Since the main interest of this research is how changes in
inequality conditional on labor endowment affect changes in trade openness under
dictatorships, I use error correction models (ECMs) to conduct empirical analysis. The
main advantage of this modeling strategy is to estimate how long it takes for the deviation
from the equilibrium, caused by changes of explanatory variables, to return to the
equilibrium path. In other words, ECMs are used to estimate both short-run and long-run
impacts of independent variables on the dependent variable (Boef and Keele 2008). From
the theoretical perspective of this chapter, it is appropriate to employ error correction
models, because my focus is how changes of inequality affect changes, rather than levels, of

12

This argument is related to the study of Hankla and Kuthy (2012). While they find that
dictatorships with longer regime time horizons are more likely to implement liberal trade
policies, this chapter argues that trade openness is a strategy to prolong authoritarian time
horizon.
13
The variable Autoc in the Polity IV project measures competition for the executive
recruitment, constraints on the executive power, and political participation. Thus, a higher
value of Autoc indicates a higher level of authoritarianism. Alternatively, since Autoc
measures how tightly political leaders maintain their authoritarian incumbency, it can be a
proxy of political time horizon of dictatorships.
39

trade openness under dictatorships.
Methodologically, ECMs take the first difference of the dependent variable and
regress it on the lagged dependent variable, lagged independent variables, and the first
differences of independent variables. In the general form of ECMs, the lagged variables
represent the long-run impacts and the first-differenced explanatory variables capture the
short-run impacts on the dependent variable. Formally, in this chapter I establish the
following error correction model to estimate the impact of inequality on trade openness in
authoritarian countries:

∆TradeOpen =a + b0TradeOpeni,t -1 + b1Inequalityi,t + b2∆Inequalityi,t +
b3LaborEndowmenti,t -1 + b4 ∆LaborEndowmenti,t +
b5(Inequalityi,t -1 × LaborEndowmenti,t -1) +
b6∆(Inequalityi,t × LaborEndowmenti,t ) +
b7GDPpci,t -1 + ∆b 8GDPpci,t +

(1)

b9Oili,t -1 + b10∆Oili,t + b11Metali,t -1 + b12∆Metali,t +
b13Authoritarianismi,t -1 + b14∆Authoritarianismi,t +
b15SinglePartyi,t -1 + b16GATT / WTOi,t -1 + ei,t
In equation (1), α is the constant term, the lagged variables capture the long-term
effects of explanatory variables, the first-differenced variables represent the short-term
effects of changes in independent variables, and εi,t is the disturbance term for country i
in year t. Based on the current literature and my discussions so far, I expect that the
short-term interactive effects between Inequality and LaborEndowment (i.e., β6 ) to be
positive. In other words, short-term increases of inequality conditional on labor
40

endowment are associated with more trade openness. To account for country-specific
heterogeneity and potential heteroskedasticity, I estimate ECMs with fixed effects models
and panel-corrected standard errors (Beck and Katz 1995). I also include five-year
dummies to control for common economic shocks across authoritarian regimes.

2.3

14

Empirical Results
The estimation results of error correction models are reported in Table 2.1. In brief,

the empirical results indicate that short-term increases in the interaction between
inequality and labor endowment also increase future levels of trade openness for
authoritarian countries in my sample. The result is robust across most models estimated in
Table 2.1, confirming the hypothesis of this chapter. I will discuss those models in detail.
First, I estimate a baseline model without considering the interaction effects. As
reported in Model 1, Inequality has neither short-term nor long-term effects on trade
openness. However, Labor Endowment has long-term negative effects, implying that
countries with larger population are less involved in international trade (Katzenstein 1985).
Model 2 includes two interaction terms between Inequality and Labor Endowment. The
sign of ∆(Inequality × LaborEndowment) is positive and statistically significant at 0.1
level. This result supports the hypothesis of this chapter. Since GDPpc is highly correlated

14

A preferred method is using year dummies in the estimation. However, including year
dummies will make the variance-covariance matrix highly singular and the panel-corrected
standard errors unavailable in Stata 11. Nevertheless, the main estimation results remain
the same if I use decade dummies or year dummies in fixed effect models with clustered
standard errors.
41

with Labor Endowment and insignificant in Model 1 and Model 2,

15

I follow the practice

of Ahlquist and Wibbels (2012) to drop GDPpc from estimation and report the results in
Model 3. The coefficient of the interaction term becomes slightly larger and significant at
0.05 level.
To check whether the results of Model 3 are sensitive to the choice of dataset on
dictatorships, I use another dataset on regime types constructed by Przeworski et al. (2000)
and updated by Cheibub, Gandhi, and Vreeland (2009). Similar to Geddes' dataset,
Cheibub et al. (2009) use a binary variable to code whether a country is autocratic or
democratic. Although their dataset covers fewer dictatorships, the statistical significance of
the interaction term ∆(Inequality × LaborEndowment) still holds in Model 4.
Some readers may wonder whether my argument about the effects of inequality on
trade openness can also be applied to democracies. As I mentioned in the previous section,
trade policymaking is subject to different causal mechanisms in democracies and
autocracies. Based on Model 4, I investigate cases of democracies in Model 5, in which the
sign of the interaction term becomes negative and statistically significant. This contrasting
result suggests that trade openness may not be a strategy in response to rising inequality in
democracies as it may be in autocracies.

15

The correlation between GDPpc and Labor Endowment is -0.73.
42

Table 2.1

Determinants of Trade Openness under Dictatorships
Model 1

Trade Opennesst-1
Inequalityt-1

△Inequality

Labor Endowment t-1
△Labor Endowment

(Inequality ×
Labor Endowment ) t-1
△(Inequality ×
Labor Endowment )
GDP pc t-1

Model 2

Model 3

Model 4

-0.281***
[0.082]

-0.281***
[0.082]

-0.280***
[0.082]

-0.278***
[0.085]

0.073
[0.156]
0.082
[0.138]

0.033
[0.202]
-0.422
[0.317]

-0.022
[0.187]
-0.531
[0.292]

-0.136
[0.220]
-0.501
[0.385]

-13.993***
[5.421]
11.238
[9.698]

-13.515**
[5.443]
8.052
[9.912]
-0.061

-11.642***
[4.391]
8.873
[9.873]
-0.004

-12.111***
[4.670]
8.254
[10.182]
0.032

[0.128]
0.403*
[0.242]

[0.119]
0.483**
[0.221]

[0.129]
0.483*
[0.254]

△GDPpc

-0.024
[0.018]
-0.005
[0.044]

-0.021
[0.019]
-0.011
[0.044]

Oil t-1

-0.017
[0.023]
-0.023
[0.025]

-0.027
[0.028]
-0.032
[0.027]

-0.039*
[0.023]
-0.044*
[0.026]

-0.044*
[0.024]
-0.044*
[0.026]

Metal t-1

0.569
[0.610]

0.603
[0.603]

0.773
[0.616]

0.728
[0.559]

2.888***
[0.879]

2.899***
[0.874]

3.007***
[0.869]

2.918***
[0.860]

0.695*
[0.417]
0.733*
[0.375]

0.699*
[0.418]
0.733*
[0.376]

0.741*
[0.412]
0.776**
[0.369]

0.670**
[0.306]
0.737**
[0.305]

△Oil

△Metal t-1

Authoritarianism t-1
△Authoritarianism

43

Table 2.1

(cont’d)

Model 1
Single Party t-1
GAT T /W T O t-1
Constant

Model 2

Model 3

Model 4

-1.843
[4.737]

-2.026
[4.734]

-1.761
[4.790]

-2.679
[4.668]

3.458**
[1.762]
16.972*
[9.709]
1173
81

3.496*
[1.807]
17.067*
[9.704]
1173
81

3.666**
[1.749]
13.001
[8.838]
1173
81

3.040*
[1.822]
15.370*
[8.356]
1135
78

N
No. of Countries
2
R
0.195
0.196
0.196
0.19
Note: All models include country fixed effects and five-year dummies (not shown).
Panel-corrected standard errors are included in brackets. The dependent variable in Model
1 to Model 4 is △Trade Openness. Model 4 analyzes the cases of dictatorships covered by
Cheibub et al (2009).
* p < 0.1, ** p < 0.05, *** p < 0.01.
Another reasonable doubt may be associated with the effects of Gini-based measure of
inequality on trade openness. Although I have explained why it is inappropriate to use
household income to measure inequality in this chapter, in Model 6 I use the Gini-based
measure of household income inequality constructed by Solt (2009). The result indicates
that household income inequality has no significant short-term effects on trade openness,
but it can further inhibit trade openness in the long run. Nevertheless, the interaction term

Inequality × LaborEndowment is positive and statistically significant, implying that
income inequality between households may facilitate trade openness in labor-abundant
countries in the long run.
The results of control variables are worth discussion, too. First, the role of natural
resources on trade openness is ambiguous. The estimation suggests that oil production has
negative impacts on trade openness while metal production has positive impacts. The
44

reason for this contrasting result is beyond the scope of this chapter, but it is consistent
with a recent study arguing that trade regime is more protected in oil-rich Middle East
countries where private sectors are less developed (Malik and Awadallah 2011). Second,
the membership of GATT/WTO, as argued by Goldstein, Rivers and Tomz (2007) and
expected by this chapter, facilitates more trade openness. However, unlike what scholars
have found (Hankla and Kuthy 2012; Milner and Kubota 2005), single-party dictatorships
do not have higher trade openness than their counterparts, while higher levels of
authoritarianism are associated with more trade openness.
One may argue that the dependent variable measuring the ratio between trade volume
and GDP is not a choice variable but an outcome variable to measure trade policies, so it is
inappropriate to use it as a measure of trade openness. Although it is common in literature
to use the trade-GDP ratio as an indicator of trade openness (e.g., Ghosh 2002; Milner and
Mukherjee 2009), I adopt two alternatives of trade openness. First, I use the KOF index of
economic globalization constructed by Dreher (2006). In addition to actual trade flows,
this variable also measures countries' restrictions on trade, such as mean tariff rates and
hidden import barriers. As a country with a huger value of this variable has a higher level
of trade openness, the results of Model 7 confirm that increases in inequality conditional
on labor endowment further facilitate trade expansion.

45

Table 2.2

Determinants of Trade Openness under Dictatorships
Mode 5

Trade Opennesst-1

Model 6

-0.145***
[0.038]

-0.105
[0.074]

Econ. Globalizationt-1

Model 7

Inequalityt-1

0.191*
[0.111]
0.248
[0.192]

-0.424**
[0.172]
0.022
[0.360]

-0.097***
[0.030]
0
[0.039]
-0.031
[0.054]

Labor Endowment t-1

-2.973
[1.979]
-6.503
[5.575]
-0.087

-17.603***
[5.032]
25.863
[18.204]
0.258**

-0.554
[0.534]
1.767
[1.450]
0.036

[0.103]
-0.248*
[0.131]

[0.117]
-0.098
[0.278]

[0.030]
0.081**
[0.034]

-0.039**
[0.017]
0.043
[0.030]

0.225
[0.201]
0.337
[0.308]

0.000
[0.005]
-0.002
[0.006]

-0.138
[0.099]

0.834**
[0.360]

0.141*
[0.076]

0.462*
[0.239]

3.262***
[0.747]

0.272**
[0.136]

Authoritarianism t-1

-0.253
[0.367]
-0.259
[0.367]

0.355
[0.297]
0.301
[0.300]

GAT T /W T O t-1

0.513
[1.087]

4.571***
[1.384]

△Inequality

△Labor Endowment

(Inequality ×
Labor Endowment ) t-1
△(Inequality ×
Labor Endowment )
Oil t-1
△Oil

Metal t-1

△Metal t-1

△Authoritarianism

46

-0.588
[0.412]

Table 2.2

Constant

(cont’d)

Model 5
2.032
[3.568]
1134
66

Model 6
26.270***
[9.316]
1197
79

Model 7
4.499***
[1.570]
942
69

N
No. of Countries
2
R
0.319
0.119
0.137
Note: All models include country fixed effects and five-year dummies (not shown).
Panel-corrected standard errors are included in brackets. The dependent variable in Model
5 and Model 6 is △Trade Openness. The dependent variable in Model 7 is Econ.
Globalization. Model 5 covers cases of democracies included in Geddes (1999). The
inequality variable in Model 6 is the household income inequality constructed by Solt
(2009); other models use the industrial pay inequality measure constructed by Galbraith
(2012).
* p < 0.1, ** p < 0.05, *** p < 0.01.
Alternatively, I use the measure of trade openness constructed by Sachs and Warner
(1995) and updated by Wacziarg and Welch (2003). This binary variable measures
whether or not a country's trade regime is open on the basis of its tariffs, non-tariffs
barriers, and black-market exchange rates. Since this variable is binary, I follow the
suggestion of Beck, Katz and Tucker (1998) to use the binary time-series-cross-section
(BTSCS) method that addresses the issue of time dependence. In particular, I include a
variable that counts the duration of time from when a country's opens itself up to trade
(Closed Years), and then create three time splines as the smooth function to deal with time
dependence (Spline1, Spline2, and Spline3). I use clustered standard errors to address
heteroskedasticity.

47

Table 2.3

Effects of Inequality on Trade Openness

Model 8
(BTSCS)
Inequality
Labor Endowment

Model 9
(BTSCS)

0.090**
[0.040]
1.110***
[0.402]

0.145**
[0.073]
1.512***
[0.569]
-0.050
[0.055]

0.049***
[0.009]
-0.197**
[0.082]
-0.037
[0.136]
-0.126*
[0.068]

0.053*** 0.039***
[0.009]
[0.010]
-0.198*** -0.125***
[0.074]
[0.025]
-0.055
0.017
[0.143]
[0.230]
-0.121*
-0.392***
[0.067]
[0.123]

Inequality × Labor
Endowment

Model 10
Mode 11
Mode 12
(BTSCS) (Sys. GMM) (Hausman-Taylor)
0.117
[0.091]
1.099
[0.896]
-0.036
[0.077]

Trade Opennesst-1
Trade Opennesst-1
GDPpc
Oil
Metal
Authoritarianism
GATT/WTO

-1.005
[0.995]
-66.213***
[24.202]
2.802**
[1.354]
0.061
[0.355]
0.157**
[0.071]

-0.058
[0.162]
-11.327***
[1.978]
0.244**
[0.116]
0.788
[0.033]
0.021
[0.035]

-0.104
[0.080]
1.754
[5.487]
-0.577
[0.769]
17.718
[12.755]

-0.029*
[0.015]
1.027***
[0.379]
0.377
[0.254]
4.421**
[1.957]
-61.080***
[21.620]
-70.773
[29.818]
1008
82
1962

Age40-59/Age15-69
Constant
N
No. of Countries
2

0.738
[0.741]
1210
73
323

0.211
[0.810]
1210
73
352

0.629
[1.121]
1236
70
496

83.721***
[24.543]
1295
85
36

χ
Note: The dependent variable in Model 8 to Model 10 is the binary Wacziarg-Welch index of
trade openness. The dependent variables in Model 11 and Model 12 are the trade-GDP ratio
and the KOF index of economic globalization, respectively. Model 10 covers only cases of
democracies. Robust standard errors clustered at the country-level are reported in brackets in
Model 8 to Model 11. The p-values of Arellano-bond test for AR(1) and AR(2) in first
difference are 0.383 and 0.805 in Model 11, respectively. The p-value of Hansen J-test is 0.509
in Model 11. The p-value of Sargan-Hansen test in Model 12 is 0.152. The following variables
are not shown in the table: Count for the years since a country’s open of its trade regime and
three time splines (Model 8 to Model 10), and five-year dummies (Model 11 and Model 12).
* p < 0.1, ** p < 0.05, *** p < 0.01.

48

Model 9 estimates the effects of Inequality and Labor Endowment without including
their interaction term. The results suggest that both Inequality and Labor Endowment
have positive impacts on trade openness. Yet, the interaction term between both variables
is statistically insignificant in Model 10. However, since the dependent variable is bounded
between 0 and 1 in logit models, the effect of one independent variable on the dependent
variable are not linearly addictive but conditional on values of other covariates.
Furthermore, the signs and statistical significance of interaction terms in logit models are
not reliable. Accordingly, scholars recommend to graph the effects of interaction terms in
nonlinear models (Berry, DeMeritt, and Esarey 2010; Brambor, Clark, and Golder 2006).
To further examine how changes in inequality conditional on labor endowment affect
trade openness, I use the Clarify package developed by King, Tomz, and Wittenberg (2000)
to simulate how the predicted probability for a dictatorship, conditional on different levels
of labor endowment, to open its trade regime changes as inequality increases. As Figure 2.2
illustrates, labor-abundant dictatorships are more likely to open their trade regimes, so the
effect of rising inequality is smaller in those countries than in labor-scarce dictatorships.

16

It has to be emphasized that the 90% confidence intervals are wider under labor-scarce
dictatorships, and they do not overlap with those of labor-abundant dictatorships unless
the inequality level is higher than the 8 (i.e., the 80th percentile).

16

I use the 25th and 75th percentile of Labor Endowment to represent labor-scarce and
labor-abundant countries, respectively. Other variables are set at their medians except the
following four variables generated by the BTSCS method: Closed Years, Spline1, Spline2,
and Spline3. To minimize the influence of time dependence in the simulation, I set the
values of those four variables at 1, -1, -0.88 and -0.54, respectively.
49

Predicted Prob. of Trade Openness
0
.25
.5
.75
1

Labor Abundant(Q75)

Labor Scarce(Q25)

1

7
10
Inequality
Note: Results are generated with Clarify in Stata 11 based on Model 9
in Table 2. All variables are controlled at their medians except Closed
Years and three time Splines (see footnote 16). Dash lines are 90%
confidence intervals. Labor endowment of country i is measured by
ln(i's Population/World Population)/( i'sGDP/World GDP).
Figure 2.2

4

Predicted Probability of Trade Openness under Dictatorships

50

.15
.1
.05
0

First Difference in Pr(Open)=
Pr(Open|Inequality=Q50) - Pr(Open|Inequality=Q25)

-1

0

1

2

Labor Endowment
Note: Results are generated with Clarify in Stata 11 based on Model
9 in Table 2. All variables are controlled at their medians except for
Closed Years and three time Splines (see footnote 16). Dash lines
are 90% confidence intervals. Labor endowment of country i is
measured by ln(i's Population/World Population)/( i's GDP/World GDP).

Figure 2.3

First Difference Measuring the Effect of Inequality on Trade Openness

Following the recommendation of Berry, DeMeritt, and Esarey (2010), I further
calculate the differences in predicted probabilities when inequality increases from its 25th
percentile to 75th percentile under dictatorships at different levels of labor endowment.
Figure 2.3 illustrates the results and suggests an inverted U-shaped relationship: Increases
in inequality have more impacts on trade openness when a dictatorship’s level of labor
abundance locates at the middle range. This result is reasonable, not only because the
dependent variable is bounded between 0 and 1, but also because countries with higher
labor endowment, as implied in Figure 2.3, are already more likely to open trade regimes.
51

I also check whether rises in inequality facilitate trade openness in democracies. As
Model 10 demonstrates, none of the variables associated with inequality or labor
endowment is statistically significant.

17

This result is similar to what we have seen in

Model 5 and suggests that trade openness in democracies may not be driven by inequality.
Although the evidence I presented so far indicates that rising inequality facilitates
trade openness under dictatorships, one reasonable challenge against this finding is the
endogeneity between trade openness and inequality. Numerous studies, as discussed in the
previous section, contend that trade liberalization increases inequality in both developed
and developing countries, so the causal direction may be reversed against the one specified
in this chapter. Yet, note that I also investigate cases of democracies in Model 5 and Model
10 in which there is no positive relationship between inequality and trade openness. Thus,
the endogeneity issue is not as prevalent as one may argue. Nevertheless, I estimate two
more models to deal with endogeneity.
First, I estimate a model using the system generalized methods of moments estimator
(system GMM). Theoretically, the system GMM models use both lagged and differenced
values of endogenous variables as instruments to deal with endogeneity, because those
instruments derived from previous periods are unlikely to be correlated with the error term
of the current period. Model 11 presents the results of a GMM model.

17

18

Since the

The results do not change if I drop Inequality × LaborEndowment from estimation.

18

I use lagged and differenced variables from the last 2 to 10 periods as instruments. I
also use robust standard errors to correct the bias generated from the two-step system
GMM estimation. The number of instruments is 42, while the number of estimated groups
is 85. The Hansen J-statistic is 23.19 (p-value = 0.51), suggesting that the used instruments
are exogenous. For other test statistics of Model 12, see Table 2.2.
52

interaction term Inequality × LaborEndowment is positive and statistically significant at
0.05 level, we can make the same inference that a dictatorship's increases in inequality
conditional on its labor endowment would facilitate more trade openness.
In addition to using lagged variables as instrumental variables with a system GMM
model, I also employ an instrumental variable regression model to deal with the issue of
endogeneity. While it is difficult to find a useful instrumental variable for inequality, in this
chapter I follow the suggestion of Higgins and Williamson (1999) and use a country's
mature cohort size relative to its adult population as an instrument of inequality. Higgins
and Williamson argue that a larger portion of mature adults aged between 40 to 59 years
old relative to the whole adult population aged between 15 and 69 years old implies lower
inequality. Their insight is that when those mature cohorts account for a larger portion of
the entire adult population than younger and older age groups (i.e., 15 to 39 years old and
60 to 69 years old, respectively), the inequality of wages in the labor market would be
smaller than the one with a larger portion of younger and older laborers. This variable is
also adopted by Leigh (2006) and You and Khagram (2005) as an instrument of inequality
when they analyzing the effects of inequality on fraternity and corruption, respectively.
With this instrument Age40-59/Age15-69,

19

I use the Hausman-Taylor estimator to

deal with endogeneity between inequality and trade openness. The reason for me to use
this modeling technique is that the variable Age40-59/Age15-69 is time-invarying in the
original dataset constructed by You and Khagram (2005), so the conventional regressional

19

Similar to LaborEndowment , I take the logarithm of this variable to deal with skewness.
The correlation between Age40-59/Age15-69 and industrial pay inequality is -0.40. The
correlation between Age40-59/Age15-69 and Trade-GDP ratio is -0.13.
53

models using instrumental variable with two-stage least squares will get incorrect results.
Hausman and Taylor (1981) propose an estimator to make time-invarying instruments
work in instrumental variable regression models.

20

Model 12 presents the estimation

results of Hausman-Taylor instrumental variable regression model, suggesting positive
influences of inequality, conditional on labor endowment, over trade openness. Note that
the p-value of Sargan-Hansen test of overidentifying restrictions is 0.47, indicating that
Age40-59/Age15-69 is a valid instrument for being uncorrelated with the error term.
Taken together, the hypothesis that dictators in labor-abundant countries respond to
rising inequality with more trade openness is supported by the empirical evidence and
different model specifications in this section.

2.4

Discussion and Conclusion
Based on the Heckscher-Ohlin model of international trade and models of democratic

transitions, this chapter argues that increases in economic inequality force dictators to use
trade openness as a way of neutralizing democratization threats. The causal mechanism is
that abundant unskilled laborers in under dictatorships not only benefit from engaging in
international trade, but also become more compliant to the authoritarian rule as inequality
reduces. This argument is supported by the data of 80 odd authoritarian regimes from
1963 to 2003.

20

The Hausman-Taylor estimator fits panel-data random-effects models in which some of
the covariates are correlated with the unobserved individual-level random effect, and none
of the covariates are correlated with the idiosyncratic error. For technical details, see Stata
Press (2009, 165-179).
54

The main finding of this chapter has both theoretical and policy implications. First,
while the current literature of democratic transitions investigates the impacts of inequality
on the prospect of democratization, this chapter pushes this research agenda a step further
by showing that dictators can use trade policies to alleviate the pressure of redistribution
and to nullify threats of democratization. This result sheds new light on the research
agenda of “authoritarian consolidation” in the age of globalization (Li and Reuveny 2009).
That is, globalization does not facilitate democratization but delays or even inhibits it. This
result contradicts to what proponents of modernization theory assert (Boix and Stokes
2003; Boix 2011; Epstein et al. 2006; Lipset 1959).
The main policy implication of this chapter is to reconsider the role of trade flow
between developed and developing countries. Many developed countries believe in the
modernization theory, expecting that trade-induced economic growth will facilitate
democratization in less developed dictatorships. Thus, those developed countries give
preferential trading treatment to less developed countries. Yet, the finding of this chapter
implies that engaging in more trade helps dictators sustain their authoritarian rule rather
than democratize. Policymakers should be aware of the unintended consequences of
authoritarian consolidation when using trade policies to coax dictators to improve the
levels of democracy in their countries.
Taking a broader view, my argument is in line with other scholars' work on the
relationship between government revenues and regime stability. Based on the selectorate
theory (Bueno de Mesquita et al. 2005), there is an emerging literature analyzing how
“unearned income” or “non-taxed revenues” stabilize authoritarian regimes (Bueno de
Mesquita and Smith 2009; Morrison 2009; Smith 2008). When facing the threat of
55

revolution, politicians can either increase the provision of public goods if their winning
coalition is large, or suppress the provision of public goods but compensate their core
supporters with private goods if their winning coalition is small. In either case, politicians
distribute available resources as unearned income to their supporters and consolidate their
incumbency. Additionally, Morrison (2009) finds that non-taxed revenues will help
governments tax elites less in democracies and spend more on welfare policies in
autocracies, respectively. The findings of these studies are consistent with the argument of
this chapter, because trade premiums enable dictators to increase the overall welfare of
their countries, especially that of laborers.
Known limitations of this chapter include poor inequality data as well as further
empirical estimation. Although I have tried to use a comprehensive inequality dataset to
conduct empirical analysis, some dictatorships are excluded from the analysis due to
missing data. One solution is to use other indicators as a proxy of inequality. For example,
Blaydes and Kayser (2011) use the consumption of calories as an alternative to inequality.
They find that individuals in democracies consume better calories in terms of quantity and
quality. Thus, the argument of this chapter can be tested with different proxies of
inequality in the future.
Second, although the question of whether or not trade openness facilitates democratic
transitions is still unsettled in the field of comparative political economy, this chapter
provides a new perspective on why some dictators choose to expand trade as a way to
consolidate their regimes. However, the effect of this policy tool is not empirically
evaluated in this chapter, and it needs to be carefully scrutinized in future research. In the
next chapter, I will elaborate on one possible direction of this research project.
56

CHAPTER 3
PREFERENTIAL TRADE AGREEMENTS AND
AUTHORITARIAN SURVIVAL
One emerging pattern in the international trading system is the growing number of
preferential trade agreements (PTAs) that allow signatories to give preferential treatments
of goods or services to each other. As of 2011, all member states of the World Trade
Organization except for Mongolia have signed at least one (WTO 2011). Not surprisingly,
an ongoing research agenda in the field of international political economy focuses on
probing the causes and consequences of PTAs signing (Baccini and Dür 2012; Büthe and
Milner 2008; Maggi and Rodríguez-Clare 2007; Schiff and Winters 1998).
Despite the extensive literature on PTAs, two puzzles continue to perplex scholars.
First, conventional wisdom holds that autocracies tend to isolate themselves from
international trade more than democracies since they rely more heavily on protectionism to
finance patron-client networks (Aidt and Gassebner 2010; Milner and Kubota 2005).
However, this perspective fails to explain why some authoritarian countries have signed
more PTAs than many democracies. As Figure 3.1 indicates, while democracies have signed
more PTAs than autocracies on average as of 2000, the medians of these two groups are
the same (i.e., 5 PTAs).

21

Additionally, it is paradoxical that trade-induced growth enriches

21

See the database of the World Trade Organization:
http://rtais.wto.org/UI/publicPreDefRepByCountry.aspx. (Accessed on August 1, 2011)
57

15

Russia

0

Number of Signed PTAs
5
10

Israel

Democracies
Figure 3.1

Dictatorships

Number of Signed PTAs of Democracies and Dictatorships in 2000

ordinary citizens but fails to facilitate

democratization

in many dictatorships that are

actively involved in international trade. Instead, some dictatorships, such as China and
Malaysia, become more resilient to the pressures of democratization even after their PTAs
induce more trade and economic growth (Nathan 2003; Pepinsky 2009).
This chapter provides a theory to disentangle these two puzzles. The key argument is
that dictators sign PTAs as a way of consolidating their authoritarian rule by collecting
support and compliance from citizens through gains from trade. Similar to the previous
chapter, I develop my argument on the basis of two well-known theories in political

58

economy: The Heckscher-Ohlin model of international trade and the Metzler-Richard
model of redistribution. According to the Heckscher-Ohlin model, countries gain
comparative advantages in trading goods produced with their relatively abundant
production factors. Since labor is usually more abundant to capital in most authoritarian
countries (e.g., China and Vietnam), trading with other countries increases wages as well as
the welfare of laborers under dictatorships.
On the other hand, the Metzler-Richard model posits that as the level of economic
inequality increases, the popular demand for redistribution also increases. Extending from
this insight, scholars have conceptualized democratization as a game of redistribution
(Acemoglu and Robinson 2001, 2005). The essence of this theory is that higher economic
inequality instigates more discontent among the poor and makes the alternative political
scheme of mass redistribution more attractive to them. Meanwhile, rich citizens have a
strong incentive to resist redistribution in which they have to pay more taxes. As a result,
rising economic inequality in authoritarian states represents a potential trigger for either
democratic regime transition or political oppression.
This chapter weaves together the insights from these perspectives and constructs a
unifying theoretical framework to account for the calculus of signing PTAs in authoritarian
regimes. I argue that when facing the pressure of democratic transition triggered by the
demand of redistribution, dictators would respond by signing PTAs to boost their trade
that enriches laborers in their countries. In other words, signing PTAs and engaging in
economic globalization help dictators ameliorate the pressure of redistribution from the

59

poor and neutralize threats of democratic transition.

22

Simply put, PTAs reduce economic

inequality and contribute to authoritarian survival.
To test both hypotheses, I compile a dataset that covers 88 authoritarian regimes from
1969 to 2000. I first investigate whether signing PTA ameliorates economic inequality
among dictatorships. Then I test whether signing PTAs reduces the probability of regime
collapse in authoritarian countries. The results suggest that signing PTAs reduces both
economic inequality and the probability of authoritarian breakdown. These findings are
robust across different model specifications, including using different measurements of the
PTA variable and controlling for other variables that are believed to be influential on
economic inequality and authoritarian breakdown, respectively.
My findings have significant implications for the literature of authoritarian politics
and international political economy. In particular, my theory offers a new perspective to
the unsettled debate over the effects of trade openness, or economic globalization in a
broader sense, on regime transition. One reason for the lack of consensus among scholars
on this issue can be attributed to the measurement issue. Conventionally, trade openness is
measured by the ratio between a country’s trade volume and its GDP. However, this
measurement has been criticized for only tapping into the size of trade flows and not trade
openness per se (Ahlquist and Wibbels 2012; Pritchett 1996). Other alternative
measurements, such as tariff rates or the level of non-tariff barriers, are also problematic

22

Although economic globalization usually refers to both trade openness and financial
integration into the world economy, due to the limited space of this chapter I mainly focus
on the role of trade openness in democratization. For the role of financial globalization, see
Li and Reuveny (2003), Eichengreen and Leblang (2008), and Milner and Mukherjee
(2009).
60

because of their poor quality in developing countries (Milner and Kubota 2005), making it
more difficult to perform cross-national analysis.
In this chapter, I take a different approach and focus on PTAs signing as the proxy of
trade openness. Intuitively, a larger number of PTAs implies more involvement in the world
economy and more surrender of rents generated from protectionism (Ornelas 2005).
Importantly, my argument challenges the theory of “democratization from the outside-in”
by showing that PTAs help authoritarian states pacify domestic discontent stemming from
the demand for redistribution and hence consolidate authoritarian rule.
The remainder of this chapter is organized as follows. Section 3.1 briefly reviews
studies about the effects of trade openness on regime transition. Section 3.2 proposes a
theory of PTA formation in the context of authoritarian countries. I discuss the research
design and present empirical analysis in Section 3.3. I discuss the implications of my
findings in the concluding section.

3.1

Trade Openness and Regime Transition
How international economy influences domestic politics has been a continuous

concern for many social scientists over the past decades. In the literature, one particularly
debated issue is the effect of trade openness on the stability of authoritarian regimes. On
the one hand, scholars suggest at least three causal mechanisms through which trade
openness would facilitate the regime transition for authoritarian countries.
First, modernization theory suggests that trade-induced economic growth fosters
democratization, because the enriched middle class would demand more political rights
and favor democratic transitions (Lipset 1959; Moore 1966). As Boix (2011, 827)
61

concludes in a recent study, economic “development has a causal effect on democracy.”
More specifically, economic growth makes the civil society become more educated and
stronger to demand for more democratic practices from the state. Since involvement in
international trade, including signing PTAs, induces economic growth (Hur and Park
2012), it can be expected that signing PTAs will facilitate democratic regime transitions.
Second, trade openness eliminates authoritarian leaders’ financial resources generated
from protectionism and then destabilizes their incumbency. Scholars have argued that
political leaders set trade barriers to collect rents from industries so as to sustain their
patron-client network and their incumbency (Grossman and Helpman 1994; Krueger 1974;
Milner and Kubota 2005; Ornelas 2005). Thus, dictators may be less able to secure their
incumbency and face more challenges of democratization forces as their countries engaged
in more international trade.
Third, the move to free trade may simultaneously open the door for democratization
forces from outside. The idea is that engaging in the world economy creates an
international linkage between dictatorships and the outside world, and such a linkage may
become an influential leverage for foreign countries to promote democratization in those
dictatorships via sanctions or other diplomatic tools. As Pevehouse (2002, 2005) have
demonstrated, regional international organizations with a sizeable portion of democratic
members may assimilate their authoritarian counterparts into democracies. Levitsky and
Way (2005, 29) echo this proposition of “democratization from the outside-in” and
contend that that the negotiations of North American Free Trade Agreement not only
exposes the Mexican government to more international scrutiny on its domestic politics
but also facilitates its political reform, resulting in the peaceful government turnover after
62

the 2000 presidential election.
On the other hand, however, a rivalry thesis argues that international trade impedes
the progress of regime transition for authoritarian states.

23

For instance, Rodrik (1997)

argues trade openness creates economic losers (at least in the short run), and governments
may fail to compensate them due to fiscal constraints or other political considerations.
Under such circumstances, the discontent instigated by economic globalization may result
in political instability that further dooms the prospect of democratization. As a
consequence, only countries that provide those globalization losers with enough welfare
transfers as “safety nets” can increase their levels of democracy when they participate in
the world economy (Rudra 2005).
In addition, trade openness enhances state capacity to counter democratization forces
in less developed or resource-abundant countries. In particular, the dependency theory
purports that authoritarian governments in less developed countries can collide with
multi-national companies headquartered in developed capitalist countries (O’Donnell
1988). Under such alliances, authoritarian leaders keep exploiting their citizens and share
gains from trade with those multi-national companies. As a result, even though those
developing countries are significantly involved in the world economy, they are unlikely to
experience regime transitions.
Empirical evidence is equally inconclusive about the effects of trade openness on
regime transition. For instance, while Boix (2011) shows trade openness promotes

23

Also see Li and Reuveny (2003) and Eichengreen and Leblang (2008) for two
comprehensive reviews.
63

democratization via economic development,

24

Ahlquist and Wibbels (2012) find no

relationship between trade openness and democratization once the regional cluster of
democratization is controlled in their empirical models. Other studies even find that trade
inhibits democratization (Li and Reuveny 2003; Teorell 2010).
This chapter attempts to reconcile this debate through a unique angle by using PTA
signing as an alternative measure of trade openness. This measurement is ideal for my
analysis for two reasons. First, Ornelas (2005) demonstrates that PTAs destruct rents of
protectionism enjoyed by import-competing industries, while Baier and Bergstrand (2007)
find that PTAs boost trade of their signatories. Thus, PTAs capture both trade activity and
trade barrier at the same time. Second, the data on PTAs, unlike other legal trade barriers,
are more transparent and widely available for most countries in the world, so researchers
are able to perform empirical analysis with more reliable data.
Using PTA signing as an alternative measure of trade openness, I show that
international trade results in authoritarian consolidation. The causal mechanism is that
dictatorships engage in international trade to reduce economic inequality and preempt
citizens’ demands for redistribution. In so doing, dictators are able to collect citizens’
compliance.
Taking a broader perspective, my argument is in line with the emerging literature on
how dictators utilize available policy tools to maintain their dominant incumbency and

24

When investigating the effects of economic development on democracy, Boix (2011)
follows the practice of Acemoglu et al. (2008) and uses a country's trade-share with other
countries as an instrument of GDP. However, while Acemoglu et al. (2008) find no causal
effects of income on democratization, Boix (2011) demonstrates that increases in income
facilitate democratic transition and contribute to democratic consolidation. Boix (2011)
also finds a declining marginal effect of income on the level of democracy.
64

then achieve “authoritarian consolidation.” (e.g., Brownlee 2007; Gandhi 2008b; Levitsky
and Way 2010; Magaloni 2006). One prevalent argument in the literature highlights how
the establishment of quasi-democratic political institutions, such as competitive elections
and binding legislatures, helps authoritarian leaders co-opt or collapse the opposition
forces. Specifically, by creating political institutions, authoritarian leaders bind their hands
to limit their capacity of seizing citizens’ property (North and Weingast 1989) and share
their power with opposition forces (Boix and Svolik 2010; Gandhi and Przeworski 2006,
2007; W. Kim and Gandhi 2010).

25

As a result, authoritarian regimes with binding

political institutions not only enjoy higher economic development but also have longer
regime horizon (Gandhi and Przeworski 2007; J. Wright 2008b).
Parallel to this insight, in the next section I argue that authoritarian leaders choose to
embrace international trade as a way to neutralize the threat of democratization and thus
to consolidate their rule. I will elaborate this argument in the next section.

3.2

International Trade and Authoritarian Survival
The central argument of this chapter is that signing PTAs helps dictators consolidate

their authoritarian rule because it introduces external resources to ease domestic demand
of redistribution. Accordingly, I expect that signing PTAs would reduce both economic
inequality and the probability of regime breakdown under dictatorships. In other words, I
will test the following two hypotheses:

25

Using Olson’s terminology, some authoritarian leaders establish political institutions to
become “stationary bandits” who protect their citizens from being looted by other “roving
bandits” (Olson 1993, 2000).
65



Hypothesis 3.1: Signing PTAs reduces economic inequality under dictatorships.



Hypothesis 3.2: Signing PTAs reduces probability of authoritarian breakdown.

I develop my argument on the basis of two existing theories. First, the
Heckscher-Ohlin model of international trade posits that countries engaging in
globalization can gain from trade by utilizing the comparative advantages and exporting
goods produced with their domestically abundant factors (Heckscher and Ohlin 1991).
Since the production costs of using abundant factors are lower, the wage premiums of
owners of those abundant factors will increase after trading with other countries where the
same production factors are scarce (Stolper and Samuelson 1941). Thus, workers in
labor-abundant countries, typically developing and authoritarian ones, can enjoy gains
from trade after their countries sign PTAs with other capital-abundant countries (Baccini
2012).

26

There are two political implications of the Heckscher-Ohlin-Stolper-Samuelson
mechanism. On the one hand, Rogowski (1989) argues that political coalitions will form
along with the cleavage of production factors in the area of trade politics. That is, owners
of abundant production factors support free trade while owners of scarce factors oppose it.
On the other hand, gains from trade expansion can further mitigate domestic conflicts
because it promotes economic development and increases the costs of using violence
(Barbieri and Reuveny 2005). Taken together, trade expansion fosters political stability
since governments can use gains from trade to collect compliance from ordinary citizens

26

For example, since China is endowed with abundant labor, the wages of Chinese
workers have increased significantly after China’s engagement in the world economy after
1980s. The effect of wage-enhancement is especially true for workers in urban areas and
the eastern coastal regions in China (Chen and Ravallion 2004).
66

(Adserà and Boix 2002; Rodrik 1998).
The second headstone of my theory is the Metlzer-Richard model of redistribution
and its extension to models of democratic transitions. In their seminal work, Metzler and
Richard (1981) demonstrate that the preference of the median voter determines the size of
government (i.e., redistribution). The intuition is that when the wage of mean income
earners is higher than that of median income earners, the latter would support higher levels
of tax rates on their richer fellow citizens and redistribute the taxed money so as to
increase their own welfare. Thus, the model predicts that the demand for redistribution
increases with the level of inequality.
While subsequent studies engage in a lengthy debate regarding whether or under what
conditions the Metzler-Richard model would hold, what is more pertinent to this chapter is
the literature on democratization and redistribution. Based on the Metzler-Richard model,
Acemoglu and Robinson (2001) conceptualize democratic transition as a solution to the
time-inconsistency problem involving redistribution. The idea is that when economic
inequality increases, dictators need to deal with the demand for redistribution from the
poor either via coercive repression or policy concession. However, when facing the poor's
threat of revolution dictators’ policy commitment of redistribution cannot be seen as
credible, since they can always renegade their promise after the revolutionary threat is
nullified. Under such circumstances, democratization solves this credibility issue because it
enables the median voter to dictate policies in the future. As a result, the occurrence of
democratic transition depends on how dictators and the rich compromise with the poor.

27

27

In their subsequent book, Acemoglu and Robinson (2005) propose an inverted
67

Based on those theories, I extend the theoretical insights from Acemoglu and
Robinson’s seminal work and examine the impact of globalization on the regime stability
of authoritarian regimes. I conceptualize trade openness, including signing PTAs, as an
alternative way to increase the welfare of the poor without taxing the rich. In other words,
engaging in international trade enriches laborers in authoritarian countries and hence
reduces income inequality and eases their demand for redistribution. Ultimately, signing
PTAs helps dictators consolidate their authoritarian regimes.

28

Note that I focus primarily on the impact that signing of PTAs has on authoritarian
breakdown, instead of focusing on the impact of PTAs on democratic transition as
investigated by previous studies. The difference between these two concepts is subtle but
crucial to our analysis. The theoretical models developed by Acemoglu and Robinson
(2001, 2005) and Boix (2003) assume that changes in inequality would initiate a regime’s
process of democratic transition. However, their models do not conclude that democratic

U-shaped relationship between inequality and democratic transition. On the one hand,
rich citizens have strong incentives to resist redistribution via costly repression when
inequality is high. On the other hand, poor citizens have few incentives to initiate
revolution when inequality is low because they will not benefit much from the
post-revolution redistribution. Therefore, according to Acemoglu and Robinson,
democratization is most likely to happen when the level of inequality is in the middle level.
By contrast, Boix (2003) posits a different perspective on the impacts of inequality on
democratization, arguing that transition to democracy becomes more likely when
inequality is low and asset mobility is high.
28
The underlying assumption of this argument is that trade openness ameliorates income
inequality. Although some studies find that trade openness increases inequality (for a
review, see Goldberg and Pavcnik 2007), others studies present the opposite evidence
(Anderson 2005; Babones and Vonada 2009). An eclectic perspective on this debate
suggests that trade openness increase inequality in developed countries but reduces it in
developing and less developing one. In the empirical section of this chapter, I demonstrate
that countries signing PTAs in the past 5 years have lower inequality than those without
any PTA signed in the past 5 years.
68

transition is the only equilibrium. In particular, the rich facing inequality can either resist
the demand for redistribution by suppressing the poor’s revolution threats or accept
democratization that enhances the poor’s welfare. When the rich decide to repress the
poor’s revolutionary threats, the current regime will experience political instability rather
than democratic transition. If neither the rich nor the poor can dominate the other group,
the confrontation between two groups may collapse the regime. Yet, the type of subsequent
regime is not identified in the models established by Acemoglu and Robinson (2005) and
Boix (2003). According to other studies, the military may intervene in these conflicts and
take over the government (Acemoglu, Ticchi, and Vindigni 2010; Svolik 2012).
Put differently, authoritarian breakdown is a broader concept than democratic
transition. As Brownlee (2009, 522) succinctly points out, an authoritarian regime
experiences democratic transition when it is “succeeded by a sustained period, at least
minimally, democratic rule.” By contrast, it may be also the case that the collapsed
authoritarian regime is replaced by another authoritarian government, military junta in
particular. From our theoretical perspective, inequality intensifies the pressures to
democratize that lead to the collapse of an authoritarian government, but it does not
ensure the succession and continuation of a democratic regime because the new political
coalition may again fail to respond to popular demands for redistribution (Houle 2009).
Based on this differentiation, our theoretical interest focuses on how PTAs prolong
authoritarian survival rather than how they facilitate democratic transition.
Before proceeding to the empirical sections, I should emphasize that my argument is
parallel to a recent study by Hollyer and Rosendorff (2012), which holds that signing PTAs
reduces the uncertainty of trade policies and favors politicians’ incumbency. They
69

convincingly show that the effect of PTAs signing on political survival is stronger in
democratic than in authoritarian countries, since politicians in democracies are subject to
electoral accountability and hence have more incentives to signal policy certainty.
My theory differs from that made by Hollyer and Rosendorff (2012) in two distinct
aspects. First, I only focus on cases of dictatorships, because I believe that authoritarian
leaders’ logic of political survival is driven by distributional effects rather than policy
certainty or electoral competition (Bueno de Mesquita et al. 2005; Acemoglu and
Robinson 2001, 2005; Boix 2003). As Hollyer, Rosendorff, and Vreeland (2011)
demonstrate elsewhere, economic policies in dictatorships are less transparent than in
democracies, where political leaders under the pressure of electoral competition need to
make their policies more transparent and credible for voters. Meanwhile, authoritarian
leaders are less constrained by electoral competition (Schedler 2006), so policy certainty
does not weigh as heavily in dictators’ calculus of political survival as it does in democratic
rulers’. In other words, although politicians in both regimes types have similar incentives to
sign PTAs to prolong their incumbency, they are subject to different constraints and need
to be analyzed separately.
Second, the outcome of interest in this chapter is regime survival, not the incumbency
of individual politicians. This difference results from my theoretical concern about the
distributional effects of PTAs on the overall regime stability. While it is theoretically
important to investigate how PTAs affect the incumbency of individual dictators, I regard
it as a difficult empirical task due to the heterogeneity of political leaders in authoritarian
countries. As I have argued, PTAs have distributive effects across different groups within a
political regime, and the process for those effects to fully realize may not be limited within
70

one or two terms in office. Thus, I take a more conservative step to focus on regime
survival instead of individual incumbency under dictatorships.

3.3

Data, Operationalization, and Empirical Analysis
To test my hypotheses, I compile a dataset that includes 1777 observations of 88

authoritarian regimes from 1969 to 2000. The Appendix A.3 presents the list of
authoritarian regimes analyzed in this chapter.
-

PTAs and Inequality
I first test whether signing PTAs reduces economic inequality that imposes the

pressure of democratic transition on authoritarian leaders. In this chapter, I use the Gini
index of Standardized World Income Inequality Database established by Solt (2009) as the
dependent variable.
I construct the key independent variable, PTA signing, from the dataset by Büthe and
Milner (2008). Specifically, I create five dummy variables to indicate whether a country
signs at least one PTA in the past one to five years. This operationalization of signing PTA
is more conservative and preferable than using the cumulative number of PTAs in a given
year as coded in Büthe and Milner’s dataset, because it avoids imposing a strict assumption
that each PTA has an equal effect on the outcome of interest. For example, it may be
problematic to claim that the North American Free Trade Agreement has the same effect as
the Mexican-Bolivia Free Trade Agreement on the authoritarian survival in Mexico.
In addition, I control for a set of variables that are found influential on inequality,
including economic development, external resources, and the size of winning coalition.
Economic Development. The level of economic development is perhaps the most
71

influential factor for income inequality. The well-known “Kuznets curve” argues that as
economy develops, inequality rises first but declines in the long run. Kuznets (1955) also
suggests that the process of urbanization reduces inequality because it enhances the wealth
of low-income groups and lessens the rural-urban income disparity. Thus, I include Urban
Population, Logged GDP per capita and Economic Growth in my empirical models. The
data on those variables are taken from the World Development Indicators (WTO 2010).
External Resources. I take into account the effects of other economic resources on
inequality, such as foreign direct investment (FDI) and trade openness. Scholars have found
that FDI increases inequality, because it introduces technological upgrading and distorts
the income distribution between skilled and unskilled laborers (Basu and Guariglia 2007;
Beyer, Rojas, and Vergara 1999). On the other hand, however, trade openness enriches the
low-income unskilled groups for the creation of employments and wage premiums in less
developed countries (Babones and Vonada 2009; Reuveny and Li 2003) and thereby
reduces inequality. Based on my theory, I expect that trade openness would have negative
effects on inequality reduction.

29

Oil Revenues. The literature suggests that oil revenues may be associated with lower
inequality. For example, Ross (2007) claims that resource-rich countries can reduce
inequality via better allocation of their resources. Morrison (2009) further demonstrates
that oil revenues lesson the fiscal burden of dictators when they increase social spending
that can mitigate inequality. Thus, I control for the effects oil revenues on inequality. I use
Harber and Menaldo’s (2011) data on a country’s overall revenue from oil production

29

The data on FDI and trade openness are taken from UNCTAD and WDI, respectively.
Both variables are measured by a country’s total volumes divide d by its GDP.
72

divided by its population. I take the natural logarithm of this variable to address skewness.
Size of Winning Coalition. According to Malesky, Abrami, and Zheng (2011),
dictatorships with larger winning coalitions tend to have lower inequality in authoritarian
countries, since politicians subject to larger winning coalitions need to implement more
welfare transfers to members of their winning coalitions and then engender more equality.
Thus, I control for the effects of winning coalition in my empirical models. The data of
winning coalition in this chapter is taken from Bueno de Mesquita et al. (2005).
Appendix A.4 reports the summary statistics of variables used in this chapter.
Model Specification. To estimate whether signing PTAs reduces inequality, I apply
OLS regression models with panel-corrected standard errors (Beck and Katz 1995).
Following the suggestion Plumper, Troeger, and Manow (2005), I correct for the first-order
autocorrelation (AR1) in my empirical estimation. To avoid reversed causality, I lag all
independent variables for one period
Empirical Results. Table 3.1 reports the empirical results. As shown in Table 3.1, the
PTAs variables have negative and significant coefficients across all models, indicating that
countries signing PTAs in the past 5 years will have lower inequality. In addition, the effect
of PTAs on reduction of inequality seems to increase as time lapses.

30

This finding

supports my theory that signing PTAs helps political leaders reduce inequality under
dictatorships. Meanwhile, I also find that higher levels of urbanization, more oil revenues,
and larger sizes of winning coalition are associated with lower inequality, while more FDI
and GDP per capita are associated with more inequality.

30

However, the dummy variable of PTA becomes statistically insignificant when we
consider a duration of six years.
73

Table 3.1

Dummy of Signed PTAs
(Past 1 Year)
Dummy of Signed PTAs
(Past 2 Years)
Dummy of Signed PTAs
(Past 3 Years)
Dummy of Signed PTAs
(Past 4 Years)
Dummy of Signed PTAs
(Past 5 Years)
Urban Population
Logged GDP per capita
GDP Growth
FDI
Trade Openness
Logged Oil Income
per capita
Winning Coalition
Constant

Effects of Signing PTAs on Economic Inequality
Mode 1
-0.290*
[0.167]

Model 2

Mode 3

Model 4

Model 5

-0.481**
[0.216]
-0.710**
[0.287]
-0.673**
[0.299]

-0.132***
[0.050]
2.966**
[1.370]
-0.019
[0.020]
0.182**
[0.093]
-0.007
[0.008]
-0.467***
[0.176]
-1.023
[0.793]
28.057***
[8.859]
775
68
0.969

-0.111**
-0.073
-0.113***
[0.045]
[0.045]
[0.042]
2.611**
0.514
2.495*
[1.261]
[1.480]
[1.354]
-0.021
-0.008
-0.018
[0.021]
[0.022]
[0.021]
0.198*
0.177
0.186
[0.119]
[0.113]
[0.115]
-0.01
-0.01
-0.009
[0.006]
[0.007]
[0.007]
-0.441***
-0.143
-0.389**
[0.151]
[0.176]
[0.151]
-1.606** -2.485** -1.540**
[0.683]
[0.975]
[0.746]
30.063*** 44.237*** 30.876***
[8.255]
[9.674]
[9.041]
755
740
730
68
68
67
0.967
0.964
0.96

-0.773***
[0.192]
-0.109***
[0.040]
2.981**
[1.249]
-0.027
[0.020]
0.186*
[0.111]
-0.006
[0.007]
-0.445**
[0.173]
-1.318*
[0.770]
26.496***
[8.265]
724
67
0.962

N
No. of Countries
2
R
Note: The dependent variable is Gini index taken from Solt (2009). All right-hand side
variables are lagged for one period. Panel-corrected standard errors are included in
brackets.
* p<0.1, ** p<0.05, *** p<0.01.

74

-

PTAs and Authoritarian Breakdown
I further test the second hypothesis on the relationship between PTAs and authoritarian

breakdown. The dependent variable, authoritarian breakdown, is taken from Geddes’
dataset on dictatorships. Geddes (1999a) uses a binary variable to indicate whether or not
an authoritarian regime collapses in a given year.

31

I limit my sample within authoritarian

countries and drop countries that have polity scores larger than 6.

32

I retain the same PTA variables specified in Table 3.1, and I include the following
variables to control for the potential confounding effects on regime breakdown: Logged
GDP per capita, GDP Growth, Military Size, Logged Oil Income per capita, Regional
Democratization, Conflict, and Single-Party.
Economic Development. To control for impacts of economic development on
democratic transition, I incorporate both GDP per capita and GDP growth in my empirical
models. The data of both variables are taken from the World Development Indicators
(WTO 2010). Following the standard practice, I take the natural logarithm of the GDP per
capita to address the skewness.
Size of Military. The coercive capacity of the state, as argued by Albertus and Menaldo
(2012), is an important factor determining the survival of authoritarian regimes. Intuitively,
if the state has stronger capacity to counter opposition forces, the regime is less likely to

31

If an authoritarian regime prevails in a given year, it enters into the next year of
observation and coded as 0 in the dataset. If this regime collapses, it is coded as 1 and
exits the observation process in the following year. However, if this regime is just replaced
by another authoritarian regime, it stays in the dataset with a new time count starting at 1
to record the age of this newly established authoritarian regime.
32
I also use a polity score of 0 as the threshold of democracy and dictatorship. The key
results reported in this chapter remain the same.
75

collapse. I include the same variable measuring state capacity used in their empirical model,
and I expect authoritarian regimes with larger military sizes to be less likely to collapse.

33

Natural Resources. A recent debate involves whether natural resources have fueled
authoritarianism and delayed democratization (Harber and Menaldo 2011; Ross 1999). In
this chapter, I view it as an empirical issue. I use Harber and Menaldo’s (2011) data on a
country’s overall revenue from oil divided by its population to measure the abundance of
natural resources. I take the natural logarithm of this variable to address skewness.
Regional Democratization. It is often argued that democratization has a “domino
effect” among neighboring countries (Leeson and Dean 2009; Pevehouse 2005). The
political unrests in the Middle East in 2011 appear to confirm this argument. Thus, I
incorporate the variable used by Harber and Menaldo (2011) to measure the intensity of
democracies for a given country’s region. Their variable Regional Democratization
calculates “the percentage of democracies in a country’s geographic-cultural region.” I
expect this variable to be positive.
Conflict. In this research, two competing theories have motivated us to control for the
number of conflicts an authoritarian regime is involved in. On the one hand, more conflicts
against the government imply higher propensity of a regime to collapse, since it belies the
regimes’ incapacity to control those conflicts. On the other hand, however, it is highly
possible that the emergence of conflicts will lead to the growth of military size,
consequently contributing to authoritarian consolidation. Thus, I am theoretical agnostic

33

Specifically, I use the data on military size collected by the Correlates of War Project
Version 4.0 (Singer 1988), and take the natural log of the number of military personnel per
1000 persons in a country.
76

regarding effects of conflicts. The variable Conflict in my models measures how many
armed conflicts an authoritarian government is involved in a given year. The data of this
variable are taken from the Uppsala Conflict Data Program (Gleditsch et al. 2002).
Single Party. According to Geddes (1999a, 1999b), single-party authoritarian regimes
are more resistant to regime breakdown and enjoy longer time horizons, not only because
their political elites are more likely to compromise with each other, but also because the
opposition forces are more likely to be co-opted. By contrast, military or personalistic
regimes are more vulnerable to the split among elites and have higher probability of
experiencing regime breakdowns. I control for the effect of authoritarian regime types in
my empirical models by including a dummy variable of single-party dictatorships.
Model Specification. In this chapter, the dependent variable is a binary variable of
regime survival, and the observations are time-series-cross-sectional for 88 authoritarian
regimes from 1969 to 2000. Thus, I apply binary time-series-cross-sectional (BTSCS)
models to conduct empirical analysis. According to Beck et al. (1998), BTSCS models are
grouped duration models that can better deal with temporal dependence than the standard
logit models. Specifically, the BTSCS models use time splines to “trace out the path of
duration dependence” for the dependent variable (Beck et al. 1998, 1270). Thus, I include
the age of an authoritarian regime with three time splines of it in my models. Also, I lag all
explanatory variables with one period to mitigate the problem of reversed causality. I also
use clustered standard errors to account for unit heterogeneity and heteroskedasticity.
Empirical Results. I report the estimation results in Table 3.2. In brief, my theoretical
conjecture regarding the effect of signing PTAs on authoritarian breakdown is well
supported by the empirical evidence. In Models 6 to Model 10, I investigate how signing
77

PTAs in the past one to five years affects authoritarian breakdown. The results indicate
that signing PTAs prolongs the survival of authoritarian regimes in the next five years, but
those newly signed PTAs have no statistically significant effects on regime breakdown of
dictatorships after the sixth year.
The results on control variables are broadly consistent with prior studies. First, the
signs of Logged GDP per capita and GDP Growth are negative, but only the latter are
statistically significant in all models. The negative and statically significant coefficient of
GDP Growth provides support for my theory. Since signing PTAs creates economic growth
(Wacziarg and Welch 2003), dictators can enjoy more compliance from their citizens if
they can keep the economy growing via PTA formation. Furthermore, my models agree on
what Geddes (1999a, 1999b) has already pointed out: Single-party authoritarian regimes
are less likely to collapse due to elite compromise and cooptation of opposition forces.
Empirical results also suggest that the state’s coercive capacity to counter the
anti-regime momentum is crucial for authoritarian breakdown. The variable Logged
Military Size has negative signs that are statistically significant in all models, indicating
that a stronger authoritarian regime is better equipped to resist threats of regime
breakdown. Meanwhile, the positive and statistically significant signs of variables Conflict
and Regional Democratization imply that dictatorships beleaguered by more violent
conflicts or democratization forces are more likely to collapse.

78

Table 3.2

Dummy of Signed PTAs
(Past 1 Year)
Dummy of Signed PTAs
(Past 2 Years)
Dummy of Signed PTAs
(Past 3 Years)
Dummy of Signed PTAs
(Past 4 Years)
Dummy of Signed PTAs
(Past 5 Years)
Logged Military Size

Effects of Signing PTAs on Authoritarian Breakdown
Model 6
-1.043*
[0.576]

Model 7

Model 8

Model 9

Mode 10

-1.183**
[0.460]
-0.782**
[0.368]
-0.830**
[0.351]

-0.536*
[0.286]
-0.334**
-0.278*
-0.278**
-0.16
-0.191
[0.143]
[0.145]
[0.139]
[0.143]
[0.142]
Logged GDP per capita
-0.109
-0.201
-0.235
-0.312
-0.247
[0.216]
[0.212]
[0.210]
[0.219]
[0.213]
GDP Growth
-0.044*** -0.046*** -0.042** -0.044*** -0.040**
[0.016]
[0.018]
[0.017]
[0.017]
[0.017]
Logged Oil Income
-0.127
-0.112
-0.116
-0.108
-0.134*
per capita
[0.080]
[0.076]
[0.075]
[0.077]
[0.081]
Regional Democratization
0.023*** 0.023*** 0.021*** 0.022*** 0.020**
[0.008]
[0.008]
[0.008]
[0.008]
[0.008]
Conflict
0.339**
0.331**
0.322**
0.314**
0.350**
[0.135]
[0.143]
[0.142]
[0.146]
[0.138]
Single-Party
-0.797*** -0.834*** -0.713** -0.745** -0.690**
[0.309]
[0.317]
[0.304]
[0.305]
[0.302]
Constant
-4.333**
-3.190*
-2.88
-1.671
-2.253
[1.867]
[1.851]
[1.793]
[1.913]
[1.911]
N
1275
1231
1196
1170
1155
No. of Countries
88
88
88
87
86
Log Pseudolikelihood
-195.193
-188.092 -191.092 -180.056 -181.511
Note: All right-hand side variables are lagged for one period. Time count of authoritarian
regime duration and three time splines are not reported. Clustered standard errors are
included in brackets.
* p<0.1, ** p<0.05, *** p<0.01

79

.02
0
-.02
-.04
-.06
-.08

Difference in Predicted Probability

1

2

3

4

5

6

7

Number of Years Elapsed
Note: Predicted robabilities with 90% confidence intervals are simulated
on the basis of Model 6 to Model 10 with Clarify in Stata 11. Variables
are controlled at their means, except for Single-Party, which is controlled
at 1 in every simulation.
Figure 3.2

Signing PTA and Differenced Predicted Probability of Regime Breakdown

To better capture the effect of signing PTAs on authoritarian survival, I use the
statistical package Clarify developed by King, Tomz, and Wittenberg (2000) to simulate
the predicted probabilities of regime breakdown based on the estimation of Model 6 to
Model 10. I further estimate two models to investigate effects of PTAs in the sixth and
seventh year after they are signed. I simulate and calculate the differences in predicted
probabilities between signing and not signing PTAs for countries in my sample.

34

Figure

3.2 illustrates the first-differences in predicted probabilities with their 90% confidence

34

The variables are controlled at their means, except for Single Party (set at 1).
80

intervals, suggesting that the effects of signing PTAs on prolonging authoritarian survival
last for five years.

3.4

Discussion and Conclusions
In this chapter, I investigate the impacts of signing preferential trade agreements (PTAs)

on authoritarian survival. Based on the Heckscher-Ohlin model of international trade and
the Metzler-Richard model of redistribution, I argue that signing PTAs helps dictators
enrich poor citizens without taxing rich ones in their countries. As a result, signing PTAs
reduces economic inequality in authoritarian regimes, and consequently neutralize the
pressure of democratization and hence prolong the duration of authoritarian regimes. My
argument is supported by the data collected from 88 authoritarian regimes during 1969
and 2000 after controlling for political and economic variables that may affect economic
inequality and/or authoritarian breakdown.
My findings have several implications for the current literature. First, I offer a new
perspective to the debates over whether economic globalization facilitates democratization.
While scholars use trade inflow or membership in international organization as the
measurements of globalization, I use PTA signing as an alternative measurement of
economic globalization. With this measurement of PTA formation, I find that signing PTAs
helps dictators maintain their authoritarian rule because PTAs ease their budget
constraints of redistribution and gather compliance among citizens with gains from trade.
Second, my findings contribute to the understanding of authoritarian consolidation
(Göbel 2011), especially for cases like China and other countries with significant
involvement in the world economy. As modernization theory fails to explain why so many
81

dictatorships like China do not collapse as their economies expand through their
engagement in the world economy, I provide a plausible explanation to this puzzle from
the logic of authoritarian survival. Specifically, authoritarian rule can be strengthened
when dictators are financed by external resources, such as trade premiums. As a result,
dictators have incentives to sign PTAs with foreign countries as a way to enrich their poor
citizens without taxing the rich in their countries.
Taking a broader view, my theory is in line with other scholars’ works on the
relationship between government revenues and regime stability. Based on the selectorate
theory (Bueno de Mesquita et al 2005), Smith (2008) argues that unearned income
collected by the government, such as natural resources and foreign aid, can help political
leaders to counter revolutionary threats. When facing the threat of revolution, politicians
can either increase the provision of public goods if their winning coalition is large, or
suppress the provision of public goods but compensate their core supporters with private
goods, if their winning coalition is small. In either case, politicians distribute available
resources as unearned income to their supporters and consolidate their incumbency.
Morrison (2009) makes a similar argument. He finds that non-taxed government revenues,
such as natural resources, foreign aids, and government borrowing, have significant effects
on regime stability in both democracies and autocracies. Specifically, those non-taxed
revenues help governments tax elites less in democracies and spend more on welfare
policies in autocracies, respectively. In other words, non-taxed revenues ease the
government’s budget constraints when implementing redistributive policies. The findings
of these studies are consistent with my theory, because trade premium enables dictators to
increase the overall welfare of their countries, especially that of laborers.
82

Additionally, the findings in this chapter cast a doubt on modernization theory
elaborated by Lipset (1959). Proponents of modernization theory contend that economic
development promotes democratization, because the newly enriched middle class will
demand more political rights to protect their property from state intervention (Moore
1966). However, the enriched middle class may not necessarily be opposed to authoritarian
rule as their countries become more prosperous. Instead, they are highly likely to be more
conservative in that they do not want the process of democratic transition, associated with
political instability and social unrest, to damage their current quality of life and benefits.
One striking example is China. Although the Chinese economy grows exponentially, its
middle class has become more conservative and supportive of the authoritarian status quo
(Unger 2006; T. Wright 2010). Based on my theory, one can argue that due to foreseeable
political instability, citizens may not want to sacrifice their current trade premiums created
by PTAs in exchange for future redistribution associated with costly democratic transition.
Nevertheless, my theory subjects to one immediate objection: If signing PTAs can
stabilize authoritarian rule, why do we not observe that autocracies, on average, sign more
PTAs than democracies? While a comprehensive investigation on why some authoritarian
countries sign more PTAs than others is beyond the scope of this chapter, here I propose
two plausible explanations. First, not every dictatorship faces the same levels of income
inequality and democratization pressure. Meanwhile, as I have discussed previously, some
dictators are better equipped with resources, such as oil or foreign aids, to lessen the
pressure of redistribution. Due to the difference in economic endowments, dictators in
authoritarian countries have different incentives to rely on PTAs to secure their
incumbency.
83

Second, although it is beneficial for dictators to sign PTAs, not every dictator can
successfully implement this strategy. A dictator faces two constraints when he wants to
expand the trade regime of his country via PTAs. Internationally, autocracies are less
credible than democracies, so they may not be able to induce their trade partners to sign
PTAs (Mansfield, Milner, and Rosendorff 2002). Domestically, although signing PTAs can
introduce new resources to meet the domestic demand for redistribution, it also implies
that dictators have to give up existing rents and resources at hand that finance their
patron-client network (Ades and Di Tella 1999; Kurer 1993). As Keefer (2007)
compellingly argues, politicians in young democracies are more prone to provide targeted
goods via their patron-client network, because they are not able to make credible
commitment of providing public goods. Therefore, even though trade openness, as a sort of
public goods, relieves the pressure of redistribution and helps dictators create more
compliance among citizens, signing PTAs is not equivalently attractive to every dictator.
Nevertheless, my theory still prevails: Once a dictatorship has successfully signed PTAs
with other countries, its probability of experiencing regime breakdown will decrease.

84

CHAPTER 4
DISGUISED PROTECTIONISM UNDER
AUTHORITARIANISM
In the literature, scholars have drawn two tentative conclusions on trade politics of
authoritarian countries. First, politicians are more likely to liberalize trade when their
authoritarian control loosens or when their countries democratize (Baccini 2012; Frye and
Mansfield 2003, 2004; Milner and Kubota 2005; Milner and Mukherjee 2009). Second,
politicians under single-party dictatorships are more likely to adopt liberal trade policies
than their counterparts under personalistic, military, or monarchy dictatorships, not only
because they need to maintain larger winning coalitions via free trade, but also because
they have longer time horizons of their incumbency to invest in public goods that generate
long-term revenues (Bueno de Mesquita et al. 2005;Hankla and Kuthy 2012; Milner and
Kubota 2005).
However, both arguments underestimate the role of special interests groups in trade
politics (Grossman and Helpman 1994, 2001). On the one hand, special interests groups
are usually privileged by protectionism and included in politicians' winning coalitions, so
they have a strong incentive to resist trade liberalization when politicians introduce it. On
the other hand, politicians under single-party dictatorships need to maintain larger
winning coalitions, but it does not necessarily imply that special interests groups will be
totally excluded due to their demand for protectionism. Taken together, what is missing in
the literature is a theory to explain how dictators interact with special interests groups
when they make trade policies.
In this chapter, I argue that although single-party dictatorships are more open to trade
85

with lower tariffs, they also develop a more sophisticated form of protectionism to meet
the demand of special interests groups. In particular, conventional wisdom holds that
dictators under single-party dictatorships need to liberalize trade as a sort of public goods
to benefit more members of their winning coalitions that sustain their incumbency. Yet, this
chapter demonstrates that tariff rates under single-party dictatorships are more
particularly designed on each category of imported products and thus more dispersed than
those under other types of dictatorships, including personalistic, military, and monarchy
dictatorships. More specifically, politicians under single-party dictatorships are more likely
to set different tariff rates on different kinds of products, even though those products are
similar to each other. Besides, I argue that tariffs are more dispersed under dictatorships
with higher levels of authoritarianism, defined by the state capacity to repress citizens and
the level of institutionalized autocracy, since authoritarianism enables dictators to collect
revenues in more efficient ways. Thus, protectionism is exercised via the complexity rather
than the level of tariffs under single-party dictatorships or under dictatorships with higher
levels of authoritarianism.
The case of Malaysia provides a typical example of the coexistence of low but
dispersed tariff rates. Malaysia has been politically dominated by the National Front since
1974. While Malaysia has reduced its tariff rates from 17.5% in 1988 to 9.5% in 2002, at
the same time its tariff dispersion increased from 91% to 210% (Athukorala 2005).
According to Athukorala and Wai-Heng (2007), one main source of tariff dispersion in
Malaysia comes from the protection on automotive industry. In particular, the Malaysian
government established the national automobile corporation Proton in 1983 and gave it
enormous protection since then, including duty exemptions and different tariffs on parts of
86

foreign cars (Jomo 1994; WTO 2002).

35

As a consequence, the overall trend of trade

liberalization does not prevent the Malaysian government from protecting its automotive
industry. Instead, trade liberalization contributes to the complication of tariffs on
automotive imports.
Figure 4.1 further illustrates the core argument of this chapter. In Figure 4.1, I plot the
averages of tariff dispersion and authoritarianism for dictatorships from 2000 to 2009. 36 I
also draw two simple regression lines with 95% confidence intervals for single-party and
other dictatorships, respectively. As shown in Figure 4.1, tariff dispersion increases with
the level of authoritarianism under single-party dictatorships. However, a similar pattern
does not exist among other types of authoritarian regimes.
The contrast of tariff dispersion between single-party and other dictatorships in Figure
4.1 presents the following puzzle that motivates this chapter: If single-party dictatorships
are more open to trade with lower tariffs (Hankla and Kuthy 2012; Milner and Kubota
2005), why are their tariff schedules more dispersed than other dictatorships? According to
trade economists, a dispersed tariff schedule usually results from a more complicated tariff
schedule and causes more trade distortion, because exporters and importers have an
incentive to reclassify their products into other categories subject to lower tariffs (Daly and

35

The political connection between Proton and the Malaysian government even made the
latter stopped buying imported cars as official vehicles but purchases from Proton during
the financial crisis in 1997 (Pepinsky 2009, 129–130).
36
The measurement of tariff dispersion is taken from the Economic Freedom of the World
Project (Gwartney, Lawson, and J. Hall 2011). The measurement of authoritarianism is
taken from the variable Autoc of the Polity IV Project (Marshall and Jaggers 2002). As I
shall discuss operationalization of variables later in the empirical section, a higher value of
Autoc indicates more institutionalized authoritarian rules.
87

10
8
6
4

Other regimes

0

2

Tariff Dispersion

Single-party

0

2
Single Party

4
Authoritarianism
Personalistic

8

6
Monarch

Military

Note: Vertical bars represent 95% confidence intervals. The measure
of authoritarianism is taken from the variable Autoc of the Polity IV
Project (Marshall and Jaggers 2002).
Figure 4.1

Tariff Dispersion and Authoritarianism under Dictatorships, 2000–2009

Kuwahara 1998; Deardorff 2006). Thus, a dispersed tariff schedule, like a high tariff
rate, is not only associated with rent-seeking behavior and corruption (Gatti 1999;
Krueger 1974), but also regarded as an alternative form of protectionism in the literature
(Gwartney, Lawson, and J. Hall 2011; OECD 2000; Oliva 2000).

37

Additionally, Ehrlich

(2011) finds that tariff complexity will impede trade volume in developed countries. This
chapter aims to explain the seemingly inconsistent trade policies under single-party

37

Thus, some economists suggest developing countries using a uniform tariff rate to
lower the administrative costs and corruption of custom officials (Gatti 1999; Panagariya
and Rodrik 1993).
88

dictatorships.
In this chapter, I propose two causal mechanisms linking authoritarianism to dispersed
tariff schedules. First, when political leaders have to maintain larger winning coalitions
under single-party dictatorships, it does not mean that they will exclude all special interests
groups from their winning coalitions. Instead, they need to differentiate which special
interests groups should be included in and which ones be excluded from their winning
coalitions. Thus, dictators will selectively respond to the demand for protectionism or free
trade, resulting in different and dispersed tariffs on imported goods. Second, when a
dictatorship has institutionalized its authoritarian rule, its political leaders are more
capable of balancing forces between protectionism and free trade to stay in power. They
are also more capable of collecting taxes with a more dispersed and complicated tariff
schedule. Both arguments are supported by the data collected from 42 dictatorships during
2000 and 2009.
This chapter contributes to the literature in two ways. First, it advances the
understanding about trade politics under dictatorships. While scholars have found that
single-party dictatorships are more open to trade with lower tariffs (Hankla and Kuthy
2012; Milner and Kubota 2005), this chapter offers a caveat to those who might conclude
that single-party dictatorships are less protectionist than other dictatorships. This finding is
similar to what Kono (2006) finds about the difference in the pattern of protectionism
between democracies and dictatorships: Although democracies have lower tariffs, they also
have higher non-tariff barriers (NTBs) than dictatorships. According to Kono (2006, 369),
these NTBs enable democracies to achieve “optimal obfuscation that allows politicians to
protect their markets while maintaining a veneer of liberalization.” This chapter
89

demonstrates how similar economic statecraft is adopted by authoritarian leaders.
Second, this chapter also contributes to studies on authoritarian consolidation in the
age of globalization. As modernization theory predicts that economic growth will facilitate
democratization (Boix and Stokes 2003; Boix 2011; Epstein et al. 2006; Lipset 1959), it is
puzzling that many dictatorships, especially those under single-party rules, are still resilient
to democratic transition when they their economy prospers. What is more puzzling is that
political leaders under some dictatorships intentionally establish institutions or adopt
policies that generate economic growth (Gandhi 2008a, 2008b; J. Wright 2008a). One
prevalent approach to analyze both puzzles focuses on authoritarian institutionalization.
Specifically, dictators endeavor to institutionalize their authoritarian rule and consolidate
their incumbency in two ways. For one thing, dictators share some powers with opposition
forces within political institutions in exchange for their support (Gandhi and Przeworski
2007; Svolik 2009). For the other, with those binding institutions both incumbent and
opposition can enjoy higher economic growth that will further garner more compliance
with the authoritarian status quo among citizens (Gallagher and Hanson 2007, 2009). This
chapter extends this research agenda a step further, showing how adeptly setting tariff
schedules can be a feasible policy tool for dictators to build their winning coalitions and
secure their incumbency.
Before proceeding, I should emphasize that the argument of this chapter is highly
relevant to the “access point theory” elaborated by Ehrlich (2007, 2011). According to
Ehrlich, more access points to the policymaking process in a political system will result in
more policy bias and policy complexity, because special interests groups have more access
that “will enable them to push policy more in their favor” (Ehrlich 2011, 12). According to
90

Ehrlich (2011), example of access points include federalism, bicameralism, number of
parties, and number of district sizes. Since special interests groups are usually more capable
of overcoming the collective action problem (Olson 1971), they are more likely to take
advantage of those access points to influence policy outcomes. Thus, Ehrlich (2007, 2011)
argues that more access points would result in higher levels of protectionism in the area of
trade policy, including higher tariffs and more complicated tariff schedules. I will elaborate
on how to apply the access points theory to analyze trade politics under dictatorships in
Section 4.3.
The rest of this chapter is organized as follows. Next section briefly reviews theories
of trade liberalization under dictatorships. In Section 4.3, I elaborate on an argument to
explain why tariffs rates are lower but more dispersed under single-party dictatorships. In
the empirical section, I discuss the operationalization of key variables and perform a series
of statistical analysis to test the hypotheses regarding trade politics under dictatorships.
The final section draws some implications of my findings and offers concluding remarks.

4.1

Theories of Trade Liberalization under Dictatorships
The current literature develops two approaches to analyze trade politics under

dictatorships. The first approach, elaborated by Milner and Kubota (2005), applies the
selectorate theory to investigate the pattern of trade liberalization in both democratic and
autocratic regimes. According to Bueno de Mesquita et al. (2005), political leaders'
incentives to invest in public goods, such as government expenditure or free trade, are
determined by two groups of citizens: Selectorate (S) and winning coalition (W). The
91

selectorate is the group of residents within a country who have the power to express a
preference over the selection of leadership; while W is the set of individuals within S whose
support is essential for a leader to remain in power. Bueno de Mesquita et al. (2005) argue
that the W/S ratio in a polity matters to many policy outcomes. Specifically, the size of W
determines how many private goods a political leader can exclusively provide for members
of his winning coalition, so the significance of private goods decreases as W increases.
Hence, one important implication of the selectorate theory is that political leaders subject
to larger winning coalitions are more likely to provide public goods instead of private
goods.
Based on the selectorate theory, Milner and Kubota (2005, 115) argue that “the
optimal level of protectionism for political leaders is a declining function of the size of the
winning coalition.” In other words, political leaders under democracies, which usually
require politicians to maintain larger winning coalitions than do dictatorships, have to
adopt free trade policies as a sort of public goods to benefit more members of their
coalitions.

38

This proposition that democracies trade more than autocracies is reaffirmed

by a recent work of Aidt and Gassebner (2010). Additionally, Milner and Kubota (2005)

38

Strictly speaking, whether free trade can be regarded as a sort of public goods is open
to debate. From the perspective of international trade theories, free trade does not benefit
everyone, since import-competing industries or owner of scarce production factors would
suffer. Thus, free trade does not fit to the definition of public goods (i.e., indivisible and
non-exclusive). The main reason to justify that free trade is a public goods is that “[i]t
ensures consumers -- and everyone is a consumer -- that they can have access to quality
goods at competitive prices rather than be limited to access to government-propped up
industries that are likely to be inefficient ... and over priced relative to the world market
price for comparable quality.” (Bueno de Mesquita et al. 2005, 196). However, this
justification does not consider the distributional effects of trade on the society (Hiscox
2002; Rogowski 1989).
92

find that single-party dictatorships, also due to their larger winning coalitions, have highest
level of trade openness with lowest tariff barriers than those of other types of dictatorships.
This finding is consistent with a recent study using different model specifications and
measurements of winning coalition in authoritarian countries (Hankla and Kuthy 2012).
Taken together, politicians subject to larger winning coalitions are more likely to adopt
more liberal trade policies that induce more trade flows.
Other studies demonstrate that the democratization process forces politicians to adopt
liberal trade policies, as it usually enlarges the sizes of winning coalitions and fragmentizes
political authority. For example, Frye and Mansfield (2003, 2004) analyze cases of
post-communist countries and find that nascent democracies are more likely liberalize their
trade regime as their political powers are more fragmented among elites. They also find
that trade liberalization is more likely to occur after elections. Baccini (2012) extends this
argument a step further. He argues that newly democratized countries are more likely to
form PTAs with developed countries, but not with developing or less developed ones. The
causal mechanism behind this phenomenon is straightforward: Newly democratized
countries want to stabilize their trade relationship with developed countries as a way to
gain higher welfare of their citizens, especially the welfare of the median voter (Mayer
1984). At the same time, they still compete with other developing countries in terms of
production and market access to developed countries. Thus, nascent democracies are less
likely to sign PTAs with other developing countries (Baccini 2012). Those studies provide a
dynamic perspective to explain why democracies adopt more liberal trade policies than do
autocracies.
The second approach to analyze trade politics under dictatorships focuses on how
93

dictators' perceived time horizons of their incumbency affect their trade policies. Clague et
al. (1996) are among the first generation of scholars who argue that political leaders' time
horizons affect their incentives to invest in better policies, including provision of
infrastructure and protection of property rights. Simply put, if politicians expect to stay in
power for a longer period, they are more likely to invest in better policies that generate
more revenues for them (Levi 1988; Olson 1993). This theoretical perspective is adopted
by recent studies on how dictatorships with longer regime time horizons make long-term
policies, including the allocation of foreign aid (J. Wright 2008b), the pattern of health
policy (Dionne 2011), and involvement in corruption (Chang and Golden 2010).
In the area of trade policies, Hankla and Kuthy (2012) find that single- and
multi-party dictatorships, due to their regime stability and longevity, are more likely to
implement liberal trade policies. The reason for those party-dictatorships to embrace trade
openness is intuitive: Trade openness will foster long-term economic development that
benefits both politicians and their citizens, and only those dictators with longer time
horizons of their incumbency would choose to implement free trade. Otherwise, they will
keep relying on protectionism to finance their authoritarian rule.
To sum up, the current literature reaches two tentative conclusions about trade
politics under dictatorships. First, dictatorships trade less than democracies. Second,
single-party dictatorships have higher trade openness with lower tariff rates than other
types of dictatorships, including personalistic, military, and monarchy dictatorships.
Nevertheless, in the next section I will discuss one important dimension ignored by the
current literature: The role of special interests groups.

94

4.2

Tariff Dispersion under Single-Party Dictatorships
It is well-documented that special interests groups are influential over trade policies in

democracies, because they are more organized (Davis 2003; Olson 1971), informative
(Milner 1997), or endowed with resources with which they can influence politicians'
decisions (Grossman and Helpman 1994, 1996; R. L. Hall and Deardorff 2006). In other
words, this interest-group perspective regards protectionism as a result of lobbying that
distorts policy outcomes in favor of special interests (Grossman and Helpman 2001).
The leverage of special interests groups can be further constrained or magnified by
political institutions, such as electoral rules or party systems. Take studies on trade politics
for example. Scholars have found that democracies using majoritarian electoral rules are
more protectionist in both tariff and non-tariff barriers than their counterparts using
proportional rules, because majoritarian rules usually create more incentives for legislators
to protect special interests from their districts (Evans 2009; Grossman and Helpman 2005;
Rickard 2012). Similarly, Nielson (2003) argues that presidentialism is less associated with
protectionism than parliamentarianism, because the President has to get the majority
support in the entire national district.
Ehrlich (2007; 2011) proposes a general theoretical framework, the access points
theory, to encompass studies analyzing how political institutions affect trade politics. First,
he argues that the political institutions can be conceptualized as “access points” to
policymaking process. Second, he claims that a political system with more access points,
such as the number of parties or the size of electoral districts, generates more policy bias
and complexity, because special interests groups have more opportunities to penetrate into
95

to policymaking process via those access points. Put differently, policies are more
particularly designed when there are more access points within a political system. One
important implication of the access points theory to trade politics is that countries with
more access points have higher and more complicated tariff rates (Ehrlich 2011).
While Ehrlich provides a convincing theory to explain how different establishments of
political institutions generate access points that distort policy outcomes, he does not apply
the access point theory to dictatorships with a belief that access points play a different role
in non-democracies. First, political systems under dictatorships are either closed or less
institutionalized, so they might not have clear access points with which special interests can
penetrate into policymaking process. Second, even if there are access points under
dictatorships, many authoritarian leaders restrict interest groups from either organizing or
becoming influential over their governments. Nevertheless, Ehrlich (2011, 184-185) argues
that studies on political institutions under dictatorships, such as the establishment of
legislature or opposition parties (Gandhi 2008b), can shed light on the application of
access points theory to non-democracies. This chapter concurs with this insight and applies
the access points theory to analyze the complexity of trade policy under dictatorships. In
particular, I argue that single-party dictatorships, due to their larger winning coalitions,
have more access points that result in more dispersed and complicated tariffs schedule.
Although a summary measure of access points under dictatorships does not exist in
the literature, I follow Ehrlich's practice of using political party as a proxy of access points
when analyzing trade politics under dictatorships. Specifically, Ehrlich (2007; 2011)
focuses on how different establishments of political institutions create access points in
democracies, and he finds that the number of access points increases with the number of
96

parties but decreases with party discipline in democracies. Based on this party-based
operationalization of access points, in this chapter I argue that dictatorships governed by a
single party have more access points than other types of dictatorships, including military,
monarchy, and personalistic regimes.
My claim that single-party dictatorships have more access points is supported by two
existing theories that are discussed in the previous section: The selectorate theory
elaborated by Bueno de Mesquita et al. (2005) and its extension to the typology of
dictatorships proposed by Geddes (1999a, 1999b), respectively. For one thing, the
selectorate theory contends that political leaders have to maintain winning coalitions that
can sustain their incumbency. When the size of winning coalition increases, political leaders
are less able to rely on private goods to finance their winning coalitions. Instead, they have
to increase the provision of public goods to induce more individuals to stay in their
winning coalitions. For the other, scholars have found that single-party dictatorships, due
to their larger winning coalitions and longer regime time horizons, are more open to trade
with lower tariffs than other types of dictatorships.
Although larger winning coalitions and longer regime time horizons of single-party
dictatorships lead to more provision of public goods, including free trade (Bueno de
Mesquita et al. 2005; Hankla and Kuthy 2012), it does not mean that special interests
groups will be totally excluded from dictators' winning coalitions. As forcefully argued by
Levi (1988, 11–17), “[p]olicies are the outcome of an exchange between the ruler and the
various groups who compose the polity,” and “[r]ulers whose power resources diminish
will either have to offer more in exchange or give up some of their ends.” In other words,
dictators cannot rule alone but seek for support from other elites or social groups. Based
97

on this insight, I expect that larger winning coalitions require authoritarian leaders to meet
the demand of special interests groups in a more selectively way that leads to different
policy outcomes.
In the trade polity area, a larger winning coalition not only enforces dictators to give
up collecting high tariffs from imported goods in a uniform way, but also induces them to
differentiate their trade policies in response to members and non-members of their ruling
coalitions. Specifically, dictators can exclude special interests groups that have less
contribution to their incumbency. Meanwhile, the special interests groups could increase
their “bid” on protectionism or free trade in exchange for dictators' extension of favor.
Since some special interests groups can still be privileged by dictators while most others
cannot, the overall tariff schedules will have lower means but become more dispersed and
complicated under dictatorships with larger winning coalitions, especially under
single-party dictatorships.
Based on my discussions so far, the first hypothesis of this chapter is as follows:


Hypothesis 4.1: Tariff schedules are more dispersed under single-party
dictatorship than under personalistic, military, or monarchy dictatorships.

One may immediately criticize this argument by claiming that the role of special
interests is insignificant under dictatorships, where the political systems are more closed
and dictators usually disallow influential interest groups to exist. However, many studies
have found that special interest politics matters to policy outcomes in authoritarian
countries as well. Pepinsky (2008), for example, demonstrates how the economic interests
of politicians' ruling coalitions result in different policy responses to economic crisis in
Indonesia and Malaysia, respectively. During the Asian Financial Crisis of 1997-1999, the
98

Indonesian government was supported by the owners of fixed and mobile capitals and
responded to the crisis by floating its currency as well as tightening its macroeconomic
policy. Those adjustment policies favor capital owners but hurt laborers. By contrast, the
Malaysian government is supported by laborers and owners of fixed assets, so it took
expansionary macroeconomic policy and imposed capital controls to favor them.
Therefore, special interests groups can have significant leverage over policy choices under
dictatorships. Another example is China's exchange rate policy. Steinberg and Shih (2012)
find that interest groups in the export sector play a significant role in Chinese
government's choice of maintaining a undervalued currency that significantly favors
Chinese exports. Thus, the key question is not whether there are special interests groups
under dictatorships, but how to identify their leverage over authoritarian leaders'
decision-making.
In addition to the size of winning coalition, the level of authoritarianism affects tariff
dispersion under dictatorships as well. Here, the concept authoritarianism includes two
dimensions. First, it refers to the level of institutionalization of authoritarian rule, such as
the competitiveness of executive power. Second, authoritarianism is also indicative of state
capacity to collect revenues as well as oppress citizens. In other words, a higher level of
authoritarianism under a dictatorship implies that the state has more control over the
society and more capacity to collect as well as allocate revenues. For example,
Escribà‐Folch (2008) shows that dictatorships collect more taxes when they can
institutionalize their authoritarian rule, because these institutionalized autocracies induce
more legitimacy, credibility and compliance among citizens. Less institutionalized
dictatorships, by contrast, have to tax on international trade more because it is easier for
99

their politicians to collect revenues from both imports and exports than from personal
income. Based on this reasoning, one can further infer that more institutionalized
dictatorships are more capable of imposing differentiated tax rates on imports to maximize
their revenues. Accordingly, a higher level of authoritarianism implied by more
institutionalization of autocracy may result in a more dispersed tariff schedule.
Gehlbach (2008) offers another theoretical perspective to explain how a higher level
of authoritarianism will result in more dispersed tariffs. While previous literature focuses
on how the strength of special interests contributes to different level of taxes, Gehlbach
(2008) argues that it is the “taxability” among different economic sectors that determines
politicians' decision to set taxes. Since not all economic activities can be taxed by the state,
politicians will support those economic sectors and industries with public resources when
those are easier to be taxed in the future. In other words, the differences in tariff rates
among imported products can be attributed to their heterogeneous “taxability” from
politicians' perspective on taxation. Although Gehlbach (2008) does not investigate
whether politicians will impose higher or lower tariffs on imported goods,

39

one

reasonable application of his argument to this chapter is that the taxability of imported
goods also depends on the level of authoritarianism. Specifically, higher levels of
authoritarianism increase politicians' capacity to efficiently impose optimal tariffs on
different imported goods, making the tariff rates more complicated and dispersed.
To summarize my discussions on the relationship between authoritarianism and tariff
39

While it is intuitive to argue that a higher “taxability” of an imported goods will be
taxed more, it does not imply an imported goods with lower taxability will be taxed less. A
politician may still impose an extreme high tariff rate on the hard-to-tax goods in favor of
domestic producers.
100

dispersion, the second hypothesis of this chapter is as follows:


Hypothesis 4.2: The level of tariff dispersion increases with the level of
authoritarianism.

4.3

Empirical Analysis
Data. To test the hypotheses of this chapter, I compile a dataset that includes 42

dictatorships during 2000 and 2009. I use the dataset of authoritarian regimes constructed
by Geddes (1999) and updated by Wright (2009). The unit-of-analysis is country-year. It
should be clarified that the data quality on authoritarian regimes are usually poor, and the
data used in this chapter are no exception. Among 610 observations of 75 dictatorships in
Geddes' dataset, I can only find data available for about 300 observations of 42
dictatorships. Nevertheless, most missing values come from countries with relatively closed
and backward economy where economic data are not publicly available (e.g., Cambodia
and North Korea). In addition, about 40% of the missing data are from single-party
dictatorship, so those missing values are slightly balanced between single-party
dictatorships and other types of dictatorships combined. Therefore, the issue of missing
data should not significantly invalid my analysis.
Appendix A.5 reports the list of dictatorships and their types included in this chapter.
Dependent Variable. The dependent variable in this chapter is Tariff Dispersion. While
there are many ways to measure tariff dispersion, one common and intuitive way is to
calculate the standard deviation of a country's tariff rates (Cooper 1964; Daly and
Kuwahara 1998; Gatti 1999). A higher standard deviation of tariffs implies that the overall
101

tariff rates are more dispersed around the mean tariff. The data on this variable are taken
from the annual reports of the Economic Freedom of the World (EFW) Project
(Gwartney, Lawson, and Hall 2011).

40

To mitigate the leverage of outliers with extremely

high standard deviations of tariff rates, this variable is normalized from 0 to 10. As the
EFW Project uses higher numbers to indicate more economic freedom (i.e., lower standard
deviation of tariffs), I reorder the variable and use higher numbers to represent higher tariff
dispersion.
Independent Variables. Based on the two hypotheses specified in the previous section,
there are two key explanatory variables in this research, single-party dictatorships and
authoritarianism. To test the Hypothesis 4.1, I include a dummy variable Single Party to
represent single-party dictatorships in Geddes' dataset. Thus, the baseline group in the
empirical estimation would include military, monarchy, and personalistic dictatorships.
To test Hypothesis 4.2, I operationalize authoritarianism from three different
theoretical perspectives. First, I use the variable Institutionalized Autocracy included in the
Polity IV Project as a proxy of authoritarianism. In particular, this variable is constructed
additively on the basis of five criteria that evaluate how autocratic a country is in terms of
political competition and political participation (Marshall and Jaggers 2002).

41

Thus, a

higher value of Institutionalized Autocracy not only refers to a higher level of authoritarian

40
41

Available at http://www.freetheworld.com/index.html (Access on March 1, 2012).

The five criteria include (1) Competitiveness of executive recruitment, (2) Openness of
executive recruitment, (3) Constraints on chief executive, (4) Regulation of participation,
and (5) Competitiveness of participation. A higher value of this variable indicates more
institutionalized authoritarian rules with less political participation and competition, and
thereby a higher level of authoritarianism.
102

control in an autocracy, but also implies that authoritarian leaders are more capable of
setting tariffs rates in response to forces of protectionism and free trade within their
winning coalition. I expect that the variable Institutionalized Autocracy have positive
impacts on tariffs dispersion.
Second, I use regime durability as another measure of authoritarianism. Theoretically,
if an authoritarian regime is stable with a long duration of survival, it has strong durability
to resist threats of regime breakdown. Meanwhile, regime durability, as Smith (2004, 2005)
argued, may result from successful manifestation of authoritarianism, such as the
utilization of natural resources and the establishment of parties. In other words, the level of
authoritarianism should base on the level of regime durability under dictatorships. It
should be clarified, however, that regime duration is not linearly related to regime
durability, because a regime is fragile when it is just established or when it is on the edge of
breakdown. I include a year count of regime duration and its squared term to estimate how
regime durability affects tariff dispersion via exercises of authoritarianism. If the inverted
2

U-shaped relationship does exist, I expect that the signs of Durability and Durability are
positive and negative, respectively. The data on regime duration of dictatorships are taken
from the Polity IV Project (Marshall and Jaggers 2002).
The third measure of authoritarianism in this chapter is the size of military. Military
troops represent the coercive capacity of state to suppress its opposition forces. Albertus
and Menaldo (2012, 152) find that dictatorships with larger military troops are more
resistant to democratization, because “a large security force is a function of a strong state,
which has a better ability to generate resources, defend its borders, and maintain its
monopoly on the use of violence.” Following their insight, I use the size of military to
103

measure the level of authoritarianism. I take the natural logarithm of the number of
military personnel plus 1 among every 1000 persons in a country.

42

I expect that a larger

value of this variable Military Size is indicative of a higher level of tariff dispersion. The
data on military size, measured by is taken from the Correlates of War (COW) Project
Version 4.0 (Singer 1988).
Control Variables. I include a set of variables in the empirical analysis. First, I control
for the effect of mean tariff rates. From the statistical perspective, a tariff schedule is
dispersed when there are outliers in the distribution of tariff rates.

43

Thus, it is

recommended in the literature to control for the mean tariff when researchers investigate
tariff dispersion (Ehrlich 2011; Mikic and Gilbert 2009). The variable of Mean Tariff is
taken from the EFW Project and normalized from 0 to 10, with a higher value indicating a
higher mean tariff.
Second, conceptually tariffs are taxes on imported products, so a country may have a
more differentiated and dispersed tariff schedule if it imports a lot of from other countries.
Accordingly, I control for the total amount of imports (in US dollars) of each dictatorship. I
take the natural logarithm of this variable to address its skewness. The data on Logged
Import are taken from the World Development Indicators (World Bank 2011).
Third, I control GDP per capita in my models for two reasons. One the one hand, a
country with higher economic development may have a more differentiated tariff schedule,

42

The qualitative results do not change in my empirical estimation if I add 0.1 instead of
1 to the original variable in the COW Project.
43
The existence of outliers may not invalid the argument of this chapter, since those
outliers may result from a large number of access points favoring certain special interests
groups.
104

since its citizens may be more able to purchase foreign goods. On the other hand, a
developed economy is tend to use other trade barriers rather than tariff level to protect its
market (Kono 2006). Thus, I am theoretically agnostic to the effect of economic
development on tariff dispersion. The data on GDP per capita are taken from the World
Development Indicators (World Bank 2011). I take the natural logarithm of it to address
skewness.
Appendix A.6 reports the summary statistics of each variable used in this chapter.
Model Specification. Since the dataset of this research includes the data from 42
dictatorships during 2000 and 2009, the pooled time-series cross-sectional analysis usually
violates the basic assumptions of ordinary least squares (OLS) estimation and results in
non-spherical errors. In this chapter, I follow the advice of Beck and Katz (1995) to use
OLS regression with panel-corrected standard errors (PCSEs). In addition, changes of
tariffs are usually slow in most countries (O’Reilly 2005), so I correct for first-order serial
correlation (AR1) when conducting empirical analysis. To mitigate the issue of reversed
causality, I lag all right-hand-side variables for one period.
Empirical Results. Table 4.1 presents the estimation of OLS regression with PCSEs.
Model 1 investigates the effects of authoritarian institutionalization and establishment of a
single party on tariff dispersion under dictatorships. The results support both hypotheses
that tariff schedules are more dispersed under dictatorships governed by a single party and
by higher levels of authoritarianism. This finding remains unchanged when I use Durability,
2

Durability , and Military Size to measure authoritarianism in Model 2 and Model 3,
respectively. However, when I estimate a full model with all measures of authoritarianism
105

and a dummy variable of single-party dictatorships, the statistical significance of
Authoritarian Institutionalization collapses. This statistical insignificance may result from
the collinearity between Authoritarian Institutionalization and Durability, since the
correlation between these two variables is 0.48.
The control variables are worth some discussions, too. The coefficient of Mean Tariff
has an expected sign across all models. That is, higher mean tariffs are associated with
higher tariffs dispersion under dictatorships, suggesting that protectionism can be
exercised in both the mean level and the overall dispersion of tariffs under dictatorships.
Second, the statistically significant signs of Logged Import in Model 1 and Model 2
indicate that a higher volume of import results in more dispersed tariffs. However, the
statistical significance of Logged Import collapses after Military Size is controlled for. The
negative signs of Logged GDPpc in the empirical models suggest that as a country has
higher economic development, it will rely less on tariff barriers. This result is consistent
with the finding of Kono (2006) but contradictory to that of Ehrlich (2011). I leave
detailed examination on this difference for future work.
In addition to the dataset of dictatorships constructed by Geddes and coauthors, I use
another dataset of dictatorships established by Cheibub et al. (2009). In their dataset,
Cheibub et al. construct a variable to measure how many parties exist outside the regime
front. There are three categories of this variable: No party, One party or multiple parties
that belong to regime front, and Multiple Parties. Based on this variable, I construct three
dummy variables for each category and use “No party” as a baseline category in my
empirical estimation.
Model 5 presents the empirical results and suggests that one-party dictatorships have
106

more dispersed tariffs than dictatorships without any parties. Nevertheless, it is interesting
that tariff schedule of multi-party dictatorships are less dispersed than those of their
one-party counterparts. At first glance, this result seems to invalidate the theory of this
chapter. However, it is reasonable to get this result from the theoretical perspective of this
chapter. Dictatorships may allow the formation of other parties to either generate
legitimacy or to split opposition forces (Gandhi 2008). Once opposition parties cannot
form a united front to overthrow the regime, a dictator does not need to maintain a large
winning coalition to keep his incumbency. Put differently, the existence of multiple parties
outside the authoritarian government splits elites and dilutes the influence of special
interests groups, making tariffs less complicated.
Readers may argue that the dependent variable used in Model 1 to Model 5 is not an
ideal measure of tariff complexity, because the value of standard deviation is sensitive to
outliers or extreme values. It is likely that a country has a simple but more dispersed tariff
schedule than others (Ehrlich 2011; Kono 2007). For example, a country may have just
two of tariff rates, 10% and 90%, and its tariff dispersion is larger than another country
with complicated tariff rates that are centering around 50%. Although tariff dispersion
cannot perfectly reflect tariff complexity in this simple example, a tariff schedule with a
large variance may result from extraordinary influences of some special interests groups
lobbying for privileges.
Nevertheless, I use Kono's data on tariff particularism to conduct additional
robustness checks. Based on the tariff data of Trade Analysis and Information System
constructed by the United Nations, Kono (2007) calculates the concentration of tariff rates
applied to imported products in 113 countries during the 1990s. If a country uses only a
107

few tariff rates to cover many categories of imported goods, it implies that its tariff
schedule is less complicated than the one with more tariff rates on the same amount of
imported goods. In other words, imposing different tariffs on the same coverage of
imported goods symbolizes “particularism” on specific products, and this particularism
may result from lobbies of special interests groups. While Kono measures tariff
particularism with an interval between 0 and 1, I multiply it by 10 to make it consistent
with the variable of tariff dispersion used in the previous models.

44

Since Kono's dataset of tariff particularism contains a lot of missing values for
authoritarian countries,

45

I use OLS with robust standard errors clustered at the country

level to conduct empirical analysis. In Model 6, I use Geddes and coauthors’ sample of
authoritarian regimes. The results are similar to what I have derived from the previous
models: Dictatorships ruled by a single party or large military have more dispersed tariff
schedules. In Model 7, I use the dataset of dictatorships constructed by Cheibub et al.
(2009) and get similar findings, with the exception that the variables on regime durability
become statistically insignificant but still bear the expected signs.
To summarize, the empirical evidence suggests that a dictatorship ruled by a single
party or by a higher level of authoritarianism has a more dispersed and ad hoc tariff
schedule.

44

The correlation between the standard deviation of tariff and tariff particularism is 0.57
as of 1995.
45
Kono's dataset include 295 yearly observations of 113 countries from 1990 to 2000.
However, only 85 out of 295 are collected from dictatorships.
108

Table 4.1

Authoritarianism, Single-Party Dictatorships, and Tariff Dispersion

Model 1
Mean Tariff
0.315**
[0.152]
Logged Import
0.318*
[0.168]
Logged GDPpc -0.354**
[0.174]
Institutionalized 0.112**
Autocracy
[0.056]
Durability

Model 2 Model 3 Model 4
0.333** 0.338** 0.376**
[0.156] [0.162] [0.151]
0.417** -0.021
-0.003
[0.194] [0.230] [0.232]
-0.546*** -0.057
-0.272
[0.180] [0.206] [0.218]
0.029
[0.073]
0.113***
0.109***
[0.038]

Durability

2

[0.042]

-0.002***

Single Party

[0.001]
0.436** 0.468**
[0.194] [0.217]
1.077*** 0.898* 1.149*** 0.964**
[0.403] [0.468] [0.413] [0.460]

One Party
Multi-Party
Constant

-3.462
[3.703]

-5.186
[4.443]

Model 6 Model 7
0.129*** 0.125***
[0.036] [0.036]
-0.038
0.101
[0.383] [0.520]
-0.054
0.039
[0.388] [0.463]

0.077**

0.023

-0.048

[0.033]

[0.051]

[0.053]

-0.002*** -0.002** -0.001*

[0.001]

Military Size

Model 5
0.309**
[0.158]
0.016
[0.246]
0.044
[0.262]

2.272
[4.748]

2.323
[4.827]

[0.001]
0.162
[0.214]

3.210**
[1.402]
0.591
[1.276]
1.145
[5.385]

[0.001]
0.813*
[0.430]
0.949*
[0.505]

1.886
[6.789]

0.000
[0.001]
0.218
[0.475]

3.284**
[1.389]
1.189
[1.426]
-0.383
[9.954]

N
278
288
256
248
285
74
78
No. of
42
43
42
42
48
37
40
Countries
2
0.291
0.301
0.302
0.325
0.267
0.603
0.465
R
Note: Panel-corrected standard errors are included in brackets in Model 1 to Model 5.
Model 6 and Model 7 report OLS estimates with clustered standard errors. The dependent
variable in Model 1 to Model 5 is the standard deviation of tariffs. The dependent variable
in Model 6 and Model 7 is the index of tariff particularism constructed by Kono (2007).
All independent variables are lagged for one year except in Model 6 and Model 7. The data
on numbers of parties under dictatorships in Model 5 and Model 7 are taken from Cheibub,
Gandhi, Vreeland (2009).
* p < 0.1, ** p < 0.05, *** p < 0.01.
109

4.4

Discussion and Conclusion
This chapter investigates the impacts of party organization and authoritarianism on

tariff dispersion under dictatorships. While the previous literature concludes that
dictatorships ruled by a single party have lower tariffs rates, this chapter demonstrates that
they also have more dispersed tariff schedules for two reasons. First, a larger winning
coalition of a single-party dictatorship implies more access points for special interests
groups to penetrate into the policymaking process. Second, political leaders under
single-party dictatorships have to selectively respond to the demand for protectionism
within their winning coalitions. In addition, this chapter finds that political leaders under
dictatorships with higher levels of authoritarianism are more capable of balancing forces of
protectionism and free trade as well as taxing international trade in a more efficient way.
As a result, tariff schedules are more dispersed and complicated under dictatorships that
are governed by a single party or more institutionalized and consolidated. This argument is
confirmed by the data of 42 authoritarian countries during 2000 and 2009.
With these findings, this chapter makes two contributions to the literature of
authoritarian politics as well as international political economy. First, it advances the
understanding about the determinants of trade policies under dictatorships. While the
current literature finds that authoritarianism under dictatorships result in higher tariffs,
this chapter further demonstrates that the consolidation of authoritarianism also enables
dictators to collect revenues more efficiently from international trade with more dispersed
and complicated tariff schedules.
Second, as the current literature emphasizes that larger winning coalitions of
110

single-party dictatorships have higher trade openness with lower tariff rates, their political
leaders are not necessarily more willing to wholeheartedly renounce protectionism. Instead,
this chapter demonstrates that they may become adepter to implement more complicated
trade policies that can better balance forces of protectionism and free trade of their
winning coalitions. Accordingly, this chapter clarifies the varieties of protectionism under
dictatorships. As the early literature attributes protectionism to influences of special
interests (Grossman and Helpman 1994; Grossman and Helpman 2001), some studies
have found that the forms of protectionism depend on other political variables. For
example, Kono (2006) has convincingly demonstrated that democracies tend to use
non-tariff trade barriers to protect their domestic industries, because voters are less capable
of detecting qualitative non-tariff trade barriers as their preference is free trade. This
chapter investigates how special interests politics affects the degree of tariff dispersion
when mean tariffs are reduced under single-party dictatorships. The finding of this chapter
suggests that politicians in autocracies may also rely on sophisticated forms of
protectionism as their counterparts in democracies do.
Taking a broader view, this chapter contributes to an emerging research agenda on
how political institutionalization affects authoritarian survival (Gandhi 2008; Gandhi and
Przeworski 2007; Wright 2008). The tentative consensus of those studies is that some
dictators establish binding political institutions, such as parties or legislatures, to
consolidate their authoritarian incumbency. Based on those studies, this chapter
demonstrates one unintended economic consequence of such political institutionalization:
A more dispersed tariff schedule. Additionally, this chapter is relevant to what Keefer and
coauthor find about clientelism in young democracies (Keefer and Vlaicu 2008; Keefer
111

2007). That is, political leaders in nascent democracies are less able to make credible
commitment to citizens after democratization, so they will return to rely on previous
patron-client network to sustain their incumbency. This chapter provides a parallel
explanation about how political leaders under single-party dictatorships maintain their
patron-client network via tariff schedules to stay in power.
Nevertheless, this chapter suffers from some limitations. First, the issue of missing
data prevents it from including more authoritarian countries into the analysis. About 32
authoritarian regimes are completely excluded from the empirical analysis of this chapter
due to data unavailability. Since dictatorships are less transparent than democracies in
terms of economic statistics (Hollyer, Rosendorff, and Vreeland 2011), future research
should elaborate on addressing this issue when analyzing trade politics of dictatorships.
Second, this chapter does not directly measure the numbers of access points under
different types of dictatorships. While in this chapter I argue that single-party dictatorships
have more access points due to larger winning coalitions, it does not mean that the role of
access points is nonexistent or inconsequential in other types of dictatorships. Thus, future
research may focus on how to exactly measure access points in authoritarian regimes.
Third, although I argue that tariff dispersion results from the interaction between
special interests groups and dictators, the impact of tariff dispersion on authoritarian
survival is not directly analyzed in this chapter. One possible extension of this argument is
that more dispersed tariff schedules will prolong the political survival of authoritarian
leaders, since they are more capable of responding to the demand of members in their
winning coalitions. However, one should keep the issue of endogeneity in mind when
pursuing this research idea.
112

CHAPTER 5
CONCLUSION
This dissertation investigates trade politics under dictatorships. Inspired the variation
of trade openness under dictatorships, I argue that political leaders in authoritarian
regimes can use trade policies to secure their incumbency, because engaging in
international trade creates additional external revenues for politicians to benefit poor
citizens without taxing the rich ones. Meanwhile, I argue that some authoritarian leaders
can go beyond the level of tariffs to the complexity of tariffs when they intend to protect
their domestic industries.
The theoretical foundations of this dissertation include the Heckscher–Ohlin model of
international trade, models of democratic transitions, and the logic of political survival.
For one thing, rising inequality instigates political instability and may facilitate regime
transitions, so political leaders in authoritarian countries have an incentive to mitigate
inequality by engaging their countries in the world market to benefit abundant unskilled
poor laborers. Accordingly, trade liberalization neutralizes threats of democratic transition
and prolongs political survival of dictators. For the other, although political leaders under
single-party dictatorships are more open to trade with lower tariffs, they also establish
more complicated tariff schedules to efficiently collect revenues when financing their
winning coalitions.
Based on those theoretical arguments, the first main finding of this dissertation is that
as inequality increases, dictatorships with abundant labor forces will expand their trade
regimes. As tested in a series of empirical models in Chapter 2, this finding is not driven by
113

the rival thesis that trade openness results in inequality. In particular, increases in inequality
do not promote further trade openness under democracies. Also, the results of instrumental
variables model suggest that rising inequality facilitates trade openness under dictatorships.
In other words, trade openness is a consequence, rather than a cause, of inequality in
authoritarian regimes.
Second, I find that trade openness would further reduce inequality under dictatorships.
In Chapter 3, I use signing preferential trade agreements (PTAs) as a proxy of trade
openness, because it is a better measurement to capture both trade volume and trade
policies at the same time. On the one hand, PTAs boost the trade flows of their signatories,
so it can be a proxy of trade openness. On the other hand, PTAs are indicative of trade
liberalization, since they eliminate the rents generated by protectionism. With this
operationalization of trade openness, in Chapter 3 I demonstrate that signing PTAs would
reduce inequality under dictatorships. At the same time, I find that a dictatorship is less
likely to collapse once it can successfully sign PTAs with other countries. Both findings of
Chapter 3 provide additional support for the theoretical argument and empirical evidence
of Chapter 2: Trade liberalization can be an intentional policy choice of dictators to reduce
inequality and sustain their authoritarian incumbency.
In additional to trade liberalization, this dissertation investigates protectionism and
finds that political leaders under single-party dictatorships use a more sophisticated form
of protectionism to sustain their winning coalitions. In particular, as the literature suggests
that single-party dictatorships are more open to trade with lower tariffs, I find that their
tariff schedules are more dispersed and complicated. Although political leaders subject to
larger winning coalitions have to liberalize trade regimes to benefit more citizens, I
114

demonstrate in Chapter 4 that larger winning coalitions implies more access points for
special interests groups to lobby for protection and result in more ad hoc tariffs rates.
Meanwhile, I find that dictatorships with higher levels of authoritarianism, defined by their
capacity of repression and level of autocracy, are more capable of setting complicated tariff
schedules to collect revenues from imports in a more efficient way.
Due to those findings, the main contribution of this dissertation is to advance the
understanding of trade politics under dictatorships. Although scholars have been
developing theories to analyze how different democratic institutions would results in the
varieties of trade policies, studies on trade politics under dictatorships are still
underdeveloped. This dissertation explores politics of trade liberalization and
protectionism in authoritarian regimes, arguing that dictators can use trade policies,
including both trade liberalization and protectionism, to prolong their incumbency.
Taking a broader view, this dissertation sheds new light on the debates over the
relationship between globalization and democratization. As the literature is inconclusive
on the impacts of globalization on democratization (Eichengreen and Leblang 2008; Li and
Reuveny 2009; Milner and Mukherjee 2009), the findings of this dissertation suggest that
engaging in the world economy helps dictator consolidate their authoritarian rule, because
gains from trade enrich the abundant labor forces in authoritarian regimes and ameliorate
income distribution. As a result, the threat of regime breakdown is neutralized after
inequality is reduced.
In addition, this dissertation demonstrates that politics of trade liberalization under
dictatorships may be more complicated than conventional wisdom posits. When facing the

115

demand of free trade, authoritarian leaders may develop more sophisticated ways to keep
exercising protectionism that can buttress their incumbency.
Based on these findings, there are some possible extensions for further research. First,
the quality of data used in this dissertation is far from perfect, since many dictatorships are
so isolated from the world market that their economic data, such as trade or inequality, are
usually unavailable. Researchers may elaborate more on new ways to measure or
operationalize trade as well as inequality. For example, Blaydes and Kayser (2011) use the
consumption of calories as an alternative measure of inequality across countries. Similarly,
Ahlquist and Wibbels (2012) operationalize a country’s trade openness with its trade
volume as the ratio of world trade. Thus, future research can seek for other alternative
measurements to investigate the relations between trade openness, inequality, and
democratic transitions.
Second, while this dissertation finds that tariff schedules are more dispersed and
complicated under single-party dictatorships, it does not investigate the effects of tariff
complexity on authoritarian survival. One possible extension of arguments raised in
Chapter 4 is that tariff complexity, as well as other non-tariff trade barriers, will contribute
to authoritarian survival, because dictators can collect more revenues to garner more
support for their winning coalition via disguised protectionism. Nevertheless, one should
keep the issue of endogeneity in mind when pursing this research agenda, because the
finding of Chapter 4 suggests that dictatorships with higher levels of authoritarianism are
more able to differentiate tariff rates.
Third, future research may adopt other trade models to investigate the relationship
between trade openness and democratization. One of the main theoretical foundations of
116

this dissertation is the Heckscher-Ohlin model of international trade. There are two key
assumptions of this model. For one thing, the factors of production, such as capital or labor,
are immobile between countries trading with each other. For the other, the technology of
production is identical between countries. Although the first assumption to a large extent
can still hold for most dictatorships, the second assumption may not be satisfied in cases of
dictatorships. It has been argued that the technological difference among developing
countries will result in inequality among skilled and unskilled laborers in both developed
and developing countries (Zhu and Trefler 2005), so how to apply this revised Ricadian
international technology difference model of international trade to the literature of
democratic transitions would be an intellectually rewarding task to pursue.
Fourth, recent studies have noticed about the impacts of financial openness on
democratic transitions (Acemoglu and Robinson 2005; Boix and Stokes 2003; Freeman
and Quinn 2012). The main finding of this research agenda is that dictatorships are more
likely to democratize as they are more financially integrated into the world market,
because the rich citizens have less incentive to resist democratization as they can move their
assets abroad to avoid taxation. This conclusion is in tension with this dissertation that
suggests trade openness will inhibit democratization. One possible solution to this
inconsistency is to investigate how the logic of political survival affects dictators’ choice
between trade openness and financial integration. While trade openness would benefit
owners of abundant factors in authoritarian states (i.e., labor), financial openness usually
favor rich citizens. Thus, dictators’ choice between two related policy tools may depend on
the composition of their winning coalitions.

117

Last, a unresolved puzzle of this dissertation is the existence of those dictatorships that
are highly isolated from the world economy, such as North Korea or Myanmar. According
to the theoretical arguments of this dissertation, trade liberalization could be a feasible
policy choice for dictators, but empirically it is not adopted by all dictatorships. Although I
have already offered some explanations in this dissertation, such as factor scarcity or the
lack of credibility to negotiate free trade with other countries, the issue on how those
internationally isolated dictatorships sustain their authoritarian rule stills awaits future
research.

118

APPENDIX

119

Table A.1
Afghanistan
Albania
Algeria
Armenia
Azerbaijan
Bangladesh
Benin
Bolivia
Botswana
Bulgaria
Burkina Faso
Burundi
Cameroon
Cen African Rep
Chile
China
Congo-Brz
Cuba
Dominican Rep
Ecuador
Egypt
El Salvador
Eritrea
Gabon
Gambia
Ghana
Greece
Guatemala
Haiti

List of Dictatorships Analyzed in Chapter 2
Honduras
Hungary
Indonesia
Iran
Iraq
Ivory Coast
Jordan
Kenya
Korea South
Kuwait
Kyrgyzstan
Liberia
Libya
Madagascar
Malawi
Malaysia
Mauritania
Mexico
Mongolia
Morocco
Mozambique
Namibia
Nepal
Nicaragua
Nigeria
Oman
Pakistan
Panama
Paraguay

120

Peru
Philippines
Poland
Romania
Russia
Rwanda
Saudi Arabia
Senegal
Sierra Leone
Singapore
Somalia
South Africa
Spain
Sri Lanka
Sudan
Syria
Taiwan
Tanzania
Thailand
Togo
Tunisia
Turkey
UAE
Uganda
Uruguay
Zambia
Zimbabwe

Table A.2

Summary Statistics of Variables Used in Chapter 2

Variable
Trade Openness (Trade/GDP)
Trade Openness (Wacziarg and
Welch)
Economic Globalization
Inequality (Industrial Pay)
Inequality (Gini)
GDP per capita (Hundreds USD)
Labor Endowment
Inequality × Labor Endowment
Oil per capita (Hundreds USD)
Metal per capita (Hundreds
USD)
GATT/WTO
Authoritarianism
Single-Party Dictatorship
Age40-59/Age15-69
Closed Years
Spline1
Spline2
Spline3

N
1362

Mean
72.23

Std. Dev.
60.31

Min.
2.26

Max.
622.63

1315

0.30

0.46

0

1

1087
1458
907
1362
1385
1385
1409

39.29
5.97
45.41
59.07
0.64
4.17
7.77

18.02
5.92
11.26
94.88
0.85
9.18
40.51

8.68
0.10
17.59
5.07
-2.10
-47.90
0

95.15
102.57
75.12
978.13
2.47
160.27
631.84

1434

0.63

1.86

0

14.30

1458
1449
1458
1409
1315
1315
1315
1315

0.60
5.89
0.53
-1.39
11.74
-6001.86
-2192.88
-2991.13

121

0.49
0
2.67
0
0.50
0
0.13
-2
10.49
0
9936.05
-64000
2801.98 -13320.12
4608.91 -25080.46

1
10
1
-1.01
40
0
0
0

Table A.3
Algeria
Angola
Argentina
Armenia
Azerbaijan
Bangladesh
Belarus
Benin
Bolivia
Brazil
Burkina Faso
Burundi
Cambodia
Cameroon
Central African Rep.
Chad
Chile
China
Congo
Congo Dem. Rep.
Coted'Ivoire
Dominican Republic
Ecuador
Egypt
ElSalvador
Eritrea
Ethiopia
Gabon
Gambia
Georgia

List of Dictatorships Analyzed in Chapter 3
Ghana
Guatemala
Guinea
Guinea Bissau
Haiti
Honduras
Hungary
Indonesia
Iran
Jordan
Kazakhstan
Kenya
Kuwait
Kyrgyz Rep.
Laos
Lesotho
Madagascar
Malawi
Malaysia
Mali
Mauritania
Mexico
Morocco
Mozambique
Namibia
Nepal
Nicaragua
Niger
Nigeria
Oman

122

Pakistan
Panama
Paraguay
Peru
Philippines
Russia
Rwanda
Saudi Arabia
Senegal
Sierra Leone
Singapore
South Africa
South Korea
SriLanka
Sudan
Syria
Tajikistan
Tanzania
Thailand
Togo
Tunisia
Turkmenistan
Uganda
United Arab Emirates
Uruguay
Uzbekistan
Zambia
Zimbabwe

Table A.4

Summary Statistics of Variables Used in Chapter 3

Variable
Authoritarian Breakdown
Dummy of signed PTA in the
past 1 year
Dummy of signed PTA in the
past 2 year
Dummy of signed PTA in the
past 3 year
Dummy of signed PTA in the
past 4 year
Dummy of signed PTA in the
past 5 year
Dummy of signed PTA in the
past 6 year
Logged Military Size
Logged GDP
GDP Growth
Logged Oil Income per capita
Regional Democratization
Political Instability
Single Party
Urban Population
FDI
Trade
Authoritarian Duration
Spline1
Spline2
Spline3

N
1777

Mean
0.042

Std. Dev.
0.200

Min.
0

Max.
1

1674

0.103

0.304

0

1

1586

0.192

0.394

0

1

1512

0.277

0.448

0

1

1448

0.356

0.479

0

1

1392

0.432

0.496

0

1

1343

0.503

0.500

0

1

1773
1459
1777
1756
1759
1763
1777
1777
1705
1701
1777
1777
1777
1777

-5.682
1.011
-8.042
-2.778
7.649
0.910
6.084
10.623
3.708
6.387
-50.248
35.224
2.093
2.878
0
11.054
13.748
12.937
0
84.211
0.328
0.631
0
4
0.444
0.497
0
1
37.862
21.611
2.372
100
1.475
3.136
-25.782
39.875
61.107
32.248
6.320
220.407
20.737
29.803
1
259
-2490.090 3951.058 -20135.84
0
-5782.495 9951.521 -52307.55
0
-8131.804 15831.040 -90396.910
0

123

Table A.5
Country
Algeria
Angola
Armenia
Azerbaijan
Botswana
Burkina Faso
Cameroon
Cen African Rep
Chad
China
Congo/Zaire
Congo-Brz
Egypt
Gabon
Haiti
Iran
Ivory Coast
Jordan
Kazakhstan
Kenya
Kuwait
Kyrgyzstan

Geddes’ Coding of Authoritarian Regimes, 2000–2009
Regime Type
Military
Party
Personal
Personal
Party
Personal
Personal
Personal
Personal
Party
Personal
Personal
Party
Party
Personal
Party
Personal
Monarch
Personal
Party
Monarch
Personal

Country
Malaysia
Mauritania
Morocco
Mozambique
Namibia
Nepal
Oman
Pakistan
Russia
Rwanda
Singapore
Syria
Tanzania
Togo
Tunisia
UAE
Uganda
Venezuela
Vietnam
Zambia
Zimbabwe

124

Regime Type
Party
Personal
Monarch
Party
Party
Monarch
Monarch
Military
Personal
Party
Party
Party
Party
Personal
Party
Monarch
Personal
Personal
Party
Party
Party

Table A.6
Variable
Tariff Dispersion
Tariff Particularism
Mean Tariff
Logged Import
Logged GDPpc
Authoritarianism
Durability
Durability
Military Size
Single Party
One Party
Multi-Party

Summary Statistics of Variables Used in Chapter 4
N
274
78
285
285
285
272
285
285
285
244
285
285

Mean
4.703
6.580
2.674
27.333
7.303
3.526
16.295
536.723
3.811
0.488
0.102
0.926

125

Std. Dev.
2.457
2.530
1.323
1.727
1.302
2.589
16.497
833.813
1.690
0.501
0.303
0.262

Min.
0
0
0
23.745
4.745
0
0
0
0
0
0
0

Max.
10
9.508
9.300
32.271
10.771
9
58
3364
7.941
1
1
1

BIBLIOGRAPHY

126

BIBLIOGRAPHY

Acemoglu, Daron, Simon Johnson, James A. Robinson, and Pierre Yared. 2008. “Income
and Democracy.” American Economic Review 98(3): 808–842.
Acemoglu, Daron, and James A. Robinson. 2001. “A Theory of Political Transitions.”
American Economic Review 91(4): 938–963.
———. 2005. Economic Origins of Dictatorship and Democracy. New York: Cambridge
University Press.
Acemoglu, Daron, Davide Ticchi, and Andrea Vindigni. 2010. “A Theory of Military
Dictatorships.” American Economic Journal: Macroeconomics 2(1): 1–42.
Ades, Alberto, and Rafael di Tella. 1999. “Rents, Competition, and Corruption.”
American Economic Review 89(4): 982–993.
Adserà, Alícia, and Carles Boix. 2002. “Trade, Democracy, and the Size of the Public
Sector: The Political Underpinnings of Openness.” International Organization
56(2): 229–262.
Ahlquist, John S, and Erik Wibbels. 2012. “Riding the Wave: World Trade and
Factor‐Based Models of Democratization.” American Journal of Political Science
56(2): 447–464.
Aidt, Toke S., and Martin Gassebner. 2010. “Do Autocratic States Trade Less?” World
Bank Economic Review 24(1): 38–76.
Albertus, Michael, and Victor Menaldo. 2012. “Coercive Capacity and the Prospects for
Democratization.” Comparative Politics 44(2): 151–169.
Alemán, José, and David D. Yang. 2011. “A Duration Analysis of Democratic Transitions
and Authoritarian Backslides.” Comparative Political Studies 44(9): 1123 –1151.
Anderson, Edward. 2005. “Openness and Inequality in Developing Countries: A Review
of Theory and Recent Evidence.” World Development 33(7): 1045–1063.
Ansell, Ben, and David Samuels. 2010. “Inequality and Democratization: A Contractarian
Approach.” Comparative Political Studies 43(12): 1543 –1574.
Athukorala, Prema-chandra. 2005. “Trade Policy in Malaysia: Liberalization Process,
Structure of Protection, and Reform Agenda.” ASEAN Economic Bulletin 22(1):
19–34.
127

Athukorala, Prema-chandra, and Loke Wai-Heng. 2007. “Distortions to Agricultural
Incentives in Malaysia.’’ World Bank Agricultural Distortions Working Paper No.
48476.
Babones, Salvatore J., and Dorian C. Vonada. 2009. “Trade Globalization and National
Income Inequality - Are They Related?” Journal of Sociology 45(1): 5–30.
Baccini, Leonardo. 2012. “Democratization and Trade Policy: An Empirical Analysis of
Developing Countries.” European Journal of International Relations 18(3):
455–479.
Baccini, Leonardo, and Andreas Dür. 2012. “The New Regionalism and Policy
Interdependence.” British Journal of Political Science 42(1): 57–79.
Baier, Scott L., and Jeffrey H. Bergstrand. 2007. “Do Free Trade Agreements Actually
Increase Members’ International Trade?” Journal of International Economics 71(1):
72–95.
Banerji, Arup, and Hafez Ghanem. 1997. “Does the Type of Political Regime Matter for
Trade and Labor Market Policies?” World Bank Economic Review 11(1): 171–194.
Barbieri, Katherine, and Rafael Reuveny. 2005. “Economic Globalization and Civil War.”
Journal of Politics 67(4): 1228–1247.
Basu, Parantap, and Alessandra Guariglia. 2007. “Foreign Direct Investment, Inequality,
and Growth.” Journal of Macroeconomics 29(4): 824–839.
Beck, Nathaniel, and Jonathan N. Katz. 1995. “What to do (and not to do) with
Time-Series Cross-Section Data.” American Political Science Review 89(3):
634–647.
Beck, Nathaniel, Jonathan N. Katz, and Richard Tucker. 1998. “Taking Time Seriously:
Time-Series-Cross-Section Analysis with a Binary Dependent Variable.” American
Journal of Political Science 42(4): 1260–1288.
Berry, William D, Jacqueline H. R DeMeritt, and Justin Esarey. 2010. “Testing for
Interaction in Binary Logit and Probit Models: Is a Product Term Essential?”
American Journal of Political Science 54(1): 248–266.
Beyer, Harald, Patricio Rojas, and Rodrigo Vergara. 1999. “Trade Liberalization and
Wage Inequality.” Journal of Development Economics 59(1): 103–123.
Blaydes, Lisa, and Mark Andreas Kayser. 2011. “Counting Calories: Democracy and
Distribution in the Developing World1.” International Studies Quarterly 55(4):
887–908.
Boef, Suzanna De, and Luke Keele. 2008. “Taking Time Seriously.” American Journal of
128

Political Science 52(1): 184–200.
Boix, Carles. 2003. Democracy and Redistribution. New York: Cambridge University
Press.
———. 2011. “Democracy, Development, and the International System.” American
Political Science Review 105(4): 809–828.
———. 2008. “Economic Roots of Civil Wars and Revolutions in the Contemporary
World.” World Politics 60(3): 390–437.
Boix, Carles, and Susan C. Stokes. 2003. “Endogenous Democratization.” World Politics
55(4): 517–549.
Boix, Carles, and Milan Svolik. 2010. “The Foundations of Limited Authoritarian
Government: Institutions and Power-Sharing in Dictatorships.” Working Paper.
Brambor, Thomas, William Roberts Clark, and Matt Golder. 2006. “Understanding
Interaction Models: Improving Empirical Analyses.” Political Analysis 14(1):
63–82.
Brownlee, Jason. 2007. Authoritarianism in an Age of Democratization. New York:
Cambridge University Press.
———. 2009. “Portents of Pluralism: How Hybrid Regimes Affect Democratic
Transitions.” American Journal of Political Science 53(3): 515–532.
Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow.
2005. The Logic of Political Survival. Cambridge: MIT Press.
Büthe, Tim, and Helen V Milner. 2008. “The Politics of Foreign Direct Investment into
Developing Countries: Increasing FDI through International Trade Agreements?”
American Journal of Political Science 52(4): 741–762.
Chang, Eric, and Miriam A. Golden. 2010. “Sources of Corruption in Authoritarian
Regimes.” Social Science Quarterly 91(1): 1–20.
Cheibub, José Antonio, Jennifer Gandhi, and James Raymond Vreeland. 2009.
“Democracy and Dictatorship Revisited.” Public Choice 143(1-2): 67–101.
Chen, Shaohua, and Martin Ravallion. 2004. “Welfare Impacts of China’s Accession to
the World Trade Organization.” World Bank Economic Review 18(1): 29–57.
Clague, Christopher, Philip Keefer, Stephen Knack, and Mancur Olson. 1996. “Property
and Contract Rights in Autocracies and Democracies.” Journal of Economic
Growth 1(2): 243–276.
129

Cooper, Richard N. 1964. “Tariff Dispersion and Trade Negotiations.” Journal of
Political Economy 72(6): 597–603.
Daly, Michael, and Hiroaki Kuwahara. 1998. “The Impact of the Uruguay Round on
Tariff and Non-Tariff Barriers to Trade in the ‘Quad’.” World Economy 21(2):
207–234.
Davis, Christina L. 2003. Food Fights Over Free Trade: How International Institutions
Promote Agricultural Trade Liberalization. Princeton: Princeton University Press.
Deardorff, Alan V. 2006. Terms of Trade: Glossary of International Economics.
Hackensack: World Scientific.
Deininger, Klaus, and Lyn Squire. 1996. “A New Data Set Measuring Income Inequality.”
World Bank Economic Review 10(3): 565 –591.
Dionne, Kim Yi. 2011. “The Role of Executive Time Horizons in State Response to AIDS
in Africa.” Comparative Political Studies 44(1): 55 –77.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper.
Dreher, Axel. 2006. “Does Globalization Affect Growth? Evidence from a New Index of
Globalization.” Applied Economics 38(10): 1091–1110.
Ehrlich, Sean D. 2011. Access Points: An Institutional Theory of Policy Bias and Policy
Complexity. New York: Oxford University Press.
Ehrlich, Sean D. 2007. “Access to Protection: Domestic Institutions and Trade Policy in
Democracies.” International Organization 61(3): 571–605.
Eichengreen, Barry, and David Leblang. 2008. “Democracy and Globalization.”
Economics & Politics 20(3): 289–334.
Epstein, David L. et al. 2006. “Democratic Transitions.” American Journal of Political
Science 50(3): 551–569.
Escribà-Folch, Abel. 2008. “Do Authoritarian Institutions Mobilize
Cooperation?” Constitutional Political Economy 20(1): 71–93.

Economic

Evans, Carolyn L. 2009. “A Protectionist Bias in Majoritarian Politics: An Empirical
Investigation.” Economics & Politics 21(2): 278–307.
Freeman, John R., and Dennis P. Quinn. 2012. “The Economic Origins of Democracy
Reconsidered.” American Political Science Review 106(1): 58–80.
Frye, Timothy, and Edward D Mansfield. 2003. “Fragmenting Protection: The Political
Economy of Trade Policy in the Post-Communist World.” British Journal of
130

Political Science 33(4): 635–657.
Frye, Timothy, and Edward D. Mansfield. 2004. “Timing is Everything.” Comparative
Political Studies 37(4): 371–398.
Galbraith, James K. 2012. Inequality and Instability: A Study of the World Economy Just
Before the Great Crisis. New York: Oxford University Press.
Galbraith, James K., and Hyunsub Kum. 2003. “Inequality and Economic Growth: A
Global View Based on Measures of Pay.” CESifo Economic Studies 49(4):
527–556.
Gallagher, Mary, and Jonathan K. Hanson. 2007. “``Power Tool or Dull Blade? Resilient
Autocracy and the Selectorate Theory.’’ Presented at the conference Why
Communism Didn’t Collapse: Understanding Regime Resilience in China, Vietnam,
Laos, North Korea, and Cuba, Dartmouth College, Hannover, NY.”
———. 2009. “Coalitions, Carrots, and Sticks: Economic Inequality and Authoritarian
States.” PS: Political Science & Politics 42(4): 667–672.
Gandhi, Jennifer. 2008a. “Dictatorial Institutions and their Impact on Economic Growth.”
European Journal of Sociology 49(1): 3–30.
———. 2008b. Political Institutions Under Dictatorship. New York: Cambridge
University Press.
Gandhi, Jennifer, and Adam Przeworski. 2007. “Authoritarian Institutions and the
Survival of Autocrats.” Comparative Political Studies 40(11): 1279–1301.
———. 2006. “Cooperation, Cooptation, and Rebellion under Dictatorships.” Economics
and Politics 18(1): 1–26.
Gatti, Roberta. 1999. “Corruption and Trade Tariffs, or a Case for Uniform Tariffs.”
World Bank Policy Research Working Paper No. 2216.
Geddes, Barbara. 1999a. “Authoritarian Breakdown: Empirical Test of a Game Theoretic
Argument.” Paper Presented at the Annual Meeting of the American Political
Science Association, Atlanta, September 2–5.
———. 1999b. “What Do We Know About Democratization After Twenty Years?”
Annual Review of Political Science 2: 115–144.
Geddes, Barbara, Joseph Wright, and Erica Frantz. 2012. “Authoritarian Regimes Data —
Authoritarian Regimes.” http://dictators.la.psu.edu/ (Accessed September 4, 2012).
Gehlbach, Scott. 2008. Representation Through Taxation: Revenue, Politics, and
Development in Postcommunist States. New York: Cambridge University Press.
131

Gerring, John, Strom C. Thacker, and Rodrigo Alfaro. 2012. “Democracy and Human
Development.” Journal of Politics 74(1): 1–17.
Ghosh, Atish R. 2002. Exchange Rate Regimes: Choices and Consequences. Cambridge:
MIT Press.
Gleditsch, Nils Petter et al. 2002. “Armed Conflict 1946-2001: A New Dataset.” Journal
of Peace Research 39(5): 615–637.
Göbel, Christian. 2011. “Authoritarian Consolidation.” European Political Science 10(2):
176–190.
Goldberg, Pinelopi Koujianou, and Giovanni Maggi. 1999. “Protection for Sale: An
Empirical Investigation.” American Economic Review 89(5): 1135–1155.
Goldberg, Pinelopi Koujianou, and Nina Pavcnik. 2007. “Distributional Effects of
Globalization in Developing Countries.” Journal of Economic Literature 45(1):
39–82.
Goldstein, Judith L., Douglas Rivers, and Michael Tomz. 2007. “Institutions in
International Relations: Understanding the Effects of the GATT and the WTO on
World Trade.” International Organization 61(1): 37–67.
Grossman, Gene M., and Elhanan Helpman. 2005. “A Protectionist Bias in Majoritarian
Politics.” Quarterly Journal of Economics 120(4): 1239–1282.
———. 1996. “Electoral Competition and Special Interest Politics.” Review of Economic
Studies 63(2): 265–286.
———. 1994. “Protection for Sale.” American Economic Review 84(4): 833–850.
———. 2001. Special Interest Politics. Cambridge, Mass: MIT.
Gwartney, James, Robert Lawson, and Joshua Hall. 2011. Economic Freedom of the
World—2011 Annual Report. Vancouver: Fraser Institute.
Haber, Stephen H, and Victor Menaldo. 2011. “Do Natural Resources Fuel
Authoritarianism? A Reappraisal of the Resource Curse.” American Political
Science Review 105(1): 1–26.
Hall, Richard L., and Alan V. Deardorff. 2006. “Lobbying as Legislative Subsidy.”
American Political Science Review 100(1): 69–84.
Hankla, Charles, and Daniel Kuthy. 2012. “Economic Liberalism in Illiberal Regimes:
Authoritarian Variation and the Political Economy of Trade.” International Studies
Quarterly Forthcoming.
132

Hausman, Jerry A., and William E. Taylor. 1981. “Panel Data and Unobservable
Individual Effects.” Econometrica 49(6): 1377–1398.
Heckscher, Eli F, and Bertil Gotthard Ohlin. 1991. Heckscher-Ohlin Trade Theory.
Cambridge: MIT Press.
Heston, Alan, Robert Summers, and Bettina Aten. 2009. “Penn World Table Version 6.3.”
Center for International Comparisons of Production, Income and Prices at the
University of Pennsylvania’.”
Higgins, Matthew, and Jeffrey G. Williamson. 1999. “Explaining Inequality the World
Round: Cohort Size, Kuznets Curves, and Openness.” National Bureau of
Economic Research Working Paper No. 7224.
Hiscox, Michael J. 2002. International Trade and Political Conflict: Commerce,
Coalitions, and Mobility. Princeton: Princeton University Press.
Hollyer, James R., and B. Peter Rosendorff. 2012. “Leadership Survival, Regime Type,
Policy Uncertainty, and PTA Accession.” International Studies Quarterly
Forthcoming.
Hollyer, James R., B. Peter Rosendorff, and James Raymond Vreeland. 2011. “Democracy
and Transparency.” Journal of Politics 73(4): 1191–1205.
Houle,

Christian. 2009. “Inequality and Democracy: Why Inequality Harms
Consolidation but Does Not Affect Democratization.” World Politics 61(4):
589–622.

Hur, Jung, and Cheolbeom Park. 2012. “Do Free Trade Agreements Increase Economic
Growth of the Member Countries?” World Development 40(7): 1283–1294.
International Monetary Fund. 2008. World Economic Outlook 2007: Globalization and
Inequality. Washington, D.C.: International Monetary Fund.
Irwin, Douglas A. 1996. Against the Tide: An Intellectual History of Free Trade. Princeton:
Princeton University Press.
———. 2009. Free Trade Under Fire. 3rd ed. Princeton: Princeton University Press.
Jomo, Kwame Sundaram. 1994. “The Proton Saga: Malaysian Car, Mitsubishi Gain.” In
Japan and Malaysian Development: In the Shadow of the Rising Sun, ed. Kwame
Sundaram Jomo. London: Routledge, pp. 263–290.
Katzenstein, Peter J. 1985. Small States in World Markets: Industrial Policy in Europe.
Ithaca: Cornell University Press.
Keefer, Philip. 2007. “Clientelism, Credibility, and the Policy Choices of Young
133

Democracies.” American Journal of Political Science 51(4): 804–821.
Keefer, Philip, and Razvan Vlaicu. 2008. “Democracy, Credibility, and Clientelism.”
Journal of Law, Economics, and Organization 24(2): 371–406.
Kim, Myeong Hwan. 2011. “Do We Really Know That the WTO Increases Trade?
Revisited.” Global Economy Journal 11(2): 1–21.
Kim, Wonik, and Jennifer Gandhi. 2010. “Coopting Workers under Dictatorship.”
Journal of Politics 72(3): 646–658.
King, Gary, Michael Tomz, and Jason Wittenberg. 2000. “Making the Most of Statistical
Analyses: Improving Interpretation and Presentation.” American Journal of
Political Science 44(2): 347–361.
Kono, Daniel Y. 2007. “‘One Policy for Each or One for All? A New Measure of
Trade-Policy Particularism.’ Paper presented at the 2007 Annual Meeting of the
American Political Science Association, Chicago.”
———. 2006. “Optimal Obfuscation: Democracy and Trade Policy Transparency.”
American Political Science Review 100(3): 369–384.
Krueger, Anne O. 1974. “The Political Economy of the Rent-Seeking Society.” American
Economic Review 64(3): 291–303.
Krugman, Paul R., and Mauricen Obstfeld. 2006. International Economics: Theory and
Policy. 6th ed. Massachusetts: Addison-Wesley.
Kurer, Oskar. 1993. “Clientelism, corruption, and the allocation of resources.” Public
Choice 77(2): 259–273.
Leeson, Peter T, and Andrea M Dean. 2009. “The Democratic Domino Theory: An
Empirical Investigation.” American Journal of Political Science 53(3): 533–551.
Leigh, Andrew. 2006. “Does Equality Lead to Fraternity?” Economics Letters 93(1):
121–125.
Levi, Margaret. 1988. Of Rule and Revenue. Berkeley: University of California Press.
Levitsky, Steven, and Lucan A Way. 2010. Competitive Authoritarianism: Hybrid Regimes
after the Cold War. New York: Cambridge University Press.
Levitsky, Steven, and Lucan A. Way. 2005. “International Linkage and Democratization.”
Journal of Democracy 16(3): 20–34.
———. 2002. “The Rise of Competitive Authoritarianism.” Journal of Democracy 13(2):
51–65.
134

Li, Quan, and Rafael Reuveny. 2009. Democracy and Economic Openness in an
Interconnected System: Complex Transformations. New York: Cambridge
University Press.
———. 2003. “Economic Globalization and Democracy: An Empirical Analysis.” British
Journal of Political Science 33(1): 29–54.
Lipset, Seymour Martin. 1959. “Some Social Requisites of Democracy: Economic
Development and Political Legitimacy.” American Political Science Review 53(1):
69–105.
López-Córdova, J. Ernesto, and Christopher M. Meissner. 2008. “The Impact of
International Trade on Democracy: A Long-Run Perspective.” World Politics 60(4):
539–575.
Magaloni, Beatriz. 2006. Voting for Autocracy: Hegemonic Party Survival and Its Demise
in Mexico. New York: Cambridge University Press.
Maggi, Giovanni, and Andrés Rodríguez-Clare. 2007. “A Political-Economy Theory of
Trade Agreements.” American Economic Review 97(4): 1374–1406.
Malesky, Edmund, Regina Abrami, and Yu Zheng. 2011. “Institutions and Inequality in
Single-Party Regimes: A Comparative Analysis of Vietnam and China.”
Comparative Politics 43(4): 401–420.
Malik, Adeel, and Bassem Awadallah. 2011. “The Economics of the Arab Spring.” CSAE
Working Paper Series No. 2011–23.
Mansfield, Edward D., Helen V. Milner, and B. Peter Rosendorff. 2000. “Free to Trade:
Democracies, Autocracies, and International Trade.” American Political Science
Review 94(2): 305–321.
———. 2002. “Why Democracies Cooperate More: Electoral Control and International
Trade Agreements.” International Organization 56(3): 477–513.
Marshall, Monty, and Keith Jaggers. 2002. Polity IV Project: Political Regime
Characteristics and Transitions, 1800-2002. College Park: Center for International
Development and Conflict Management, University of Maryland.
Mayer, Wolfgang. 1984. “Endogenous Tariff Formation.” American Economic Review
74(5): 970–985.
Meltzer, Allan H., and Scott F. Richard. 1981. “A Rational Theory of the Size of
Government.” Journal of Political Economy 89(5): 914–927.
Mikic, Mia, and John Gilbert. 2009. Trade Statistics in Policymaking—A Handbook of
Commonly Used Trade Indices And Indicators. New York: United Nations.
135

Milner, Helen V. 1997. Interests, Institutions, and Information: Domestic Politics and
International Relations. Princeton: Princeton University Press.
Milner, Helen V., and Keiko Kubota. 2005. “Why the Move to Free Trade? Democracy
and Trade Policy in the Developing Countries.” International Organization 59(1):
107–143.
Milner, Helen V., and Bumba Mukherjee. 2009. “Democratization and Economic
Globalization.” Annual Review of Political Science 12(1): 163–181.
Moore, Barrington. 1966. Social Origins of Dictatorship and Democracy; Lord and
Peasant in the Making of the Modern World. Boston: Beacon Press.
Morrison, Kevin M. 2009. “Oil, Nontax Revenue, and the Redistributional Foundations
of Regime Stability.” International Organization 63(1): 107–138.
Myerson, Roger. 2008. “The Autocrat’s Credibility Problem and Foundations of the
Constitutional State.” American Political Science Review 102(1): 125–139.
Nathan, Andrew J. 2003. “Authoritarian Resilience.” Journal of Democracy 14(1): 6–17.
Nepal, Mani, Alok K Bohara, and Kishore Gawande. 2011. “More Inequality, More
Killings: The Maoist Insurgency in Nepal.” American Journal of Political Science
55(4): 886–906.
Nielson, Daniel L. 2003. “Supplying Trade Reform: Political Institutions and
Liberalization in Middle-Income Presidential Democracies.” American Journal of
Political Science 47(3): 470–491.
North, Douglass C., and Barry R. Weingast. 1989. “Constitutions and Commitment: The
Evolution of Institutions Governing Public Choice in Seventeenth-Century
England.” The Journal of Economic History 49(4): 803–832.
O’Donnell, Guillermo. 1988. Bureaucratic Authoritarianism: Argentina 1966-1973 in
Comparative Perspective. Berkeley: University of California Press.
O’Reilly, Robert F. 2005. “Veto Points, Veto Players, and International Trade Policy.”
Comparative Political Studies 38(6): 652 –675.
OECD. 2000. Agricultural Policies in Emerging and Transition Economies 2000. Paris:
OECD Publishing.
Oliva, Maria-Angels. 2000. “Estimation of Trade Protection in Middle East and North
African Countries.” IMF Working Paper No. 00/27.
Olson, Mancur. 1993. “Dictatorship, Democracy, and Development.” American Political
Science Review 87(3): 567–576.
136

———. 2000. Power and Prosperity: Outgrowing Communist and Capitalist
Dictatorships. New York: Basic Books.
———. 1971. The Logic of Collective Action; Public Goods and the Theory of Groups.
Cambridge, Mass: Harvard University Press.
Ornelas, Emanuel. 2005. “Rent Destruction and the Political Viability of Free Trade
Agreements.” Quarterly Journal of Economics 120(4): 1475–1506.
Panagariya, Arvind, and Dani Rodrik. 1993. “Political-Economy Arguments for a Uniform
Tariff.” International Economic Review 34(3): 685–703.
Park, Jong Hee, and Nathan Jensen. 2007. “Electoral Competition and Agricultural
Support in OECD Countries.” American Journal of Political Science 51(2):
314–329.
Pepinsky, Thomas B. 2009. Economic Crises and the Breakdown of Authoritarian
Regimes: Indonesia and Malaysia in Comparative Perspective. New York:
Cambridge University Press.
Pepinsky, Thomas B. 2008. “Capital Mobility and Coalitional Politics: Authoritarian
Regimes and Economic Adjustment in Southeast Asia.” World Politics 60(3):
438–474.
Pevehouse, Jon C. 2005. Democracy from Above: Regional Organizations and
Democratization. New York: Cambridge University Press.
Pevehouse, Jon C. 2002. “Democracy from the Outside-In? International Organizations
and Democratization.” International Organization 56(3): 515–549.
Plümper, Thomas, Vera E Troeger, and Philip Manow. 2005. “Panel Data Analysis in
Comparative Politics: Linking Method to Theory.” European Journal of Political
Research 44(2): 327–354.
Pritchett, Lant. 1996. “Measuring Outward Orientation in LDCs: Can It Be Done?”
Journal of Development Economics 49(2): 307–335.
Przeworski, Adam, Michael E. Alvarez, Jose Antonio Cheibub, and Fernando Limongi.
2000. Democracy and Development: Political Institutions and Well-Being in the
World, 1950-1990. New York: Cambridge University Press.
Reuveny, Rafael, and Quan Li. 2003. “Economic Openness, Democracy, and Income
Inequality: An Empirical Analysis.” Comparative Political Studies 36(5): 575–601.
Rickard, Stephanie J. 2012. “A Non-Tariff Protectionist Bias in Majoritarian Politics:
Government Subsidies and Electoral Institutions.” International Studies Quarterly.
Forthcoming.
137

Rodrik, Dani. 1997. Has Globalization Gone Too Far? Washington, D.C.: Institute for
International Economics.
———. 1998. “Why Do More Open Economies Have Bigger Governments?” Journal of
Political Economy 106(5): 997–1032.
Rogowski, Ronald. 1989. Commerce and Coalitions: How Trade Affects Domestic
Political Alignments. Princeton: Princeton University Press.
———. 1987. “Trade and the Variety of Democratic Institutions.” International
Organization 41(2): 203–223.
Rose, Andrew K. 2004. “Do WTO Members Have More Liberal Trade Policy?” Journal
of International Economics 63(2): 209–235.
Ross, Michael L. 2007. “How Mineral-Rich States Can Reduce Inequality?” In Escaping
The Resource Curse, eds. Macartan Humphreys, Jeffrey Sachs, and Joseph E.
Stiglitz. New York: Columbia University Press, pp. 173–193.
Ross, Michael Lewin. 1999. “The Political Economy of the Resource Curse.” World
Politics 51(2): 297–322.
Rudra, Nita. 2005. “Globalization and the Strengthening of Democracy in the Developing
World.” American Journal of Political Science 49(4): 704–730.
Sachs, Jeffrey D., and Andrew Warner. 1995. “Economic Reform and the Process of
Global Integration.” Brookings Papers on Economic Activity 1995(1): 1–118.
Schedler, Andreas, ed. 2006. Electoral Authoritarianism: The Dynamics of Unfree
Competition. Boulder, Colo: L. Rienner Publishers, Inc.
Schiff, Maurice, and L. Alan Winters. 1998. “Regional Integration as Diplomacy.” World
Bank Economic Review 12(2): 271–295.
Schofield, Norman, and Micah Levinson. 2008. “Modeling Authoritarian Regimes.”
Politics, Philosophy & Economics 7(3): 243 –283.
Simon Kuznets. 1955. “Economic Growth and Income Inequality.” American Economic
Review 45(1): 1–28.
Singer, J.
David. 1988. “Reconstructing the Correlates of War Dataset on Material
Capabilities of States, 1816–1985.” International Interactions 14(2): 115–132.
Smith, Alastair. 2008. “The Perils of Unearned Income.” Journal of Politics 70(3):
780–793.
Smith, Benjamin. 2005. “Life of the Party: The Origins of Regime Breakdown and
138

Persistence under Single-Party Rule.” World Politics 57(3): 421.
———. 2004. “Oil Wealth and Regime Survival in the Developing World, 1960-1999.”
American Journal of Political Science 48(2): 232–246.
Solt, Frederick. 2009. “Standardizing the World Income Inequality Database.” Social
Science Quarterly 90(2): 231–242.
Stasavage, David. 2005. “Democracy and Education Spending in Africa.” American
Journal of Political Science 49(2): 343–358.
Stata Press. 2009. Longitudinal-Data/Panel-Data Reference Manual: Release 11. College
Station: StataCorp LP.
Steinberg, David A, and Victor C Shih. 2012. “Interest Group Influence in Authoritarian
States: The Political Determinants of Chinese Exchange Rate Policy.” Comparative
Political Studies Forthcoming.
Stiglitz, Joseph E. 2003. Globalization and Its Discontents. New York: W.W. Norton.
Stolper, Wolfgang F., and Paul A. Samuelson. 1941. “Protection and Real Wages.” Review
of Economic Studies 9(1): 58–73.
Svolik, Milan. 2009. “Power Sharing and Leadership Dynamics in Authoritarian
Regimes.” American Journal of Political Science 53(2): 477–494.
———. 2012. The Politics of Authoritarian Rule. Cambridge: Cambridge University Press.
Teorell, Jan. 2010. Determinants of Democratization: Explaining Regime Change in the
World, 1972-2006. New York: Cambridge University Press.
The Economist. 2010. “Democracy in Retreat.” Global Forecasting Service. http://gfs.eiu.
com/Article.aspx?articleType=wif&articleId=407 (Accessed January 8, 2011).
Tullock, Gordon. 1987. Autocracy. Dordrecht: Kluwer Academic Publishers.
Ulfelder, Jay. 2011. “Country Memberships in Selected Intergovernmental Organizations
and Accession to Selected Regional and Global Treaty Regimes: Global,
Country-Year Format, 1955-2010.” Inter-university Consortium for Political and
Social Research (ICPSR). http://dx.doi.org/10.3886/ICPSR30541.v1 (Accessed June
3, 2011).
Unger, Jonathan. 2006. “China’s Conservative Middle Class.” Far Eastern Economic
Review 169(3): 27–41.
Verdier, Daniel. 1994. Democracy and International Trade: Britain, France, and the
United States, 1860-1990. Princeton, N.J: Princeton University Press.
139

Wacziarg, Romain, and Karen Horn Welch. 2003. “Trade Liberalization and Growth:
New Evidence.” National Bureau of Economic Research Working Paper Series No.
10152.
Wintrobe, Ronald. 1998. The Political Economy of Dictatorship. Cambridge, UK:
Cambridge University Press.
World Bank. 2011. World Development Indicators 2011. Washington, D.C.: World Bank.
Wright, Joseph. 2010. “Aid Effectiveness and the Politics of Personalism.” Comparative
Political Studies 43(6): 735 –762.
———. 2008a. “Do Authoritarian Institutions Constrain? How Legislatures Affect
Economic Growth and Investment.” American Journal of Political Science 52(2):
322–343.
———. 2009. “How Foreign Aid Can Foster Democratization in Authoritarian Regimes.”
American Journal of Political Science 53(3): 552–571.
———. 2008b. “To Invest or Insure? How Authoritarian Time Horizons Impact Foreign
Aid Effectiveness.” Comparative Political Studies 41(7): 971–1000.
Wright, Teresa. 2010. Accepting Authoritarianism: State-Society Relations in China’s
Reform Era. Stanford, Calif: Stanford University Press.
WTO. 2010. International Trade Statistics 2010. Geneva: World Trade Organization.
———. 2002. Trade Policy Review: Malaysia 2001. Geneva: World Trade Organization.
———. 2011. World Trade Report 2011. Geneva: World Trade Organization.
You, Jong-Sung, and Sanjeev Khagram. 2005. “A Comparative Study of Inequality and
Corruption.” American Sociological Review 70(1): 136–157.
Zhu, Susan Chun, and Daniel Trefler. 2005. “Trade and Inequality in Developing
Countries: A General Equilibrium Analysis.” Journal of International Economics
65(1): 21–48.

140