ﬂurry“.

,.
.

Jada”:
. .3: 1?...

“2.!

H51,

 

QAL ..

3:212?
i.“ _

c 3. L4 .Hufwu. I. kiﬁvzf .. V {uu‘nnpﬂr I V . H . . “TL...“ . w.........w»rx..¥£rr ‘ 1 ‘ , . ;».mv..u..r|u.‘n.rh

 

"”f: 12515:

no;

This is to certify that the

dissertation entitled

ESSAYS ON INTERDEPENDENCE, INSTITUTIONS, AND
INTERNATIONAL CONFLICT

presented by

Mark Anthony Souva

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Political Science

3M4?

Major prof or \3

Date flu“?! ¥ i 2430 L]

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0‘ 12771

 

_ ——4—'

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DaleDue.p65-p.15

 

ESSAYS ON INTERDEPENDENCE, INSTITUTIONS, AND INTERNATIONAL
CONFLICT

By

Mark Anthony Souva

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Political Science

2002

ABSTRACT

ESSAYS ON INTERDEPENDENCE, INSTITUTIONS, AND INTERNATIONAL
CONFLICT

By

Mark Anthony Souva

How does economic interdependence inﬂuence the prospects for peace between
nations? While much has been written on the nexus between trade and conﬂict,
we lack a comprehensive argument and analysis of the relationship between
different types of trade, foreign direct investment, and international conﬂict. I
develop an analytical model and empirically test hypotheses from the model on
the relationship between strategic and non-strategic commodities, foreign direct
investment and militarized conﬂict. I argue that trade interdependence does not
always promote peace between nations. The effects of interdependence are
conditional on the type of goods traded and the political relationship between
states. Speciﬁcally, non-strategic trade interdependence and foreign direct
investment promote peace, while strategic trade interdependence between
nations lacking political afﬁnity enhances the likelihood of dyadic conﬂict. This
dissertation also extends the interdependence argument to state level behavior
by linking economic dependence to the initiation and targeting of militarized
interstate disputes. I argue that militarized states dependent on importing
strategic commodities are more likely than other states’ to initiate militarized
disputes. Similarly, weak states with strategic resources are more likely to be

targets of militarized disputes. I also ﬁnd support for a monadic democratic

peace. Democracies, in general, are more peaceful than other nations, even
when controlling for economic development. In a ﬁnal essay, I argue that
assessing the similarity of states’ political and economic institutions is central to
evaluating their satisfaction with the status quo, and, in turn, the likelihood of
conﬂict between nations. In addition to the previously confirmed paciﬁc effects of
political institutional similarity, | ﬁnd that states with similar economic institutions

are less likely to ﬁght each other.

Dedicated to

my parents, Henry and Alice, and my brother, Mike.

ACKNOWLEDGMENTS

In the process of writing this dissertation, I greatly beneﬁted from the
comments, advice, and support of many people. I could not have asked for a
better committee. Scott Gates introduced me to the economic interdependence
literature and encouraged me to pursue this topic. More importantly, he taught
me to look for theoretical puzzles and to analyze them. He guided my work with
constructive feedback and supported me throughout. Perhaps no one person
has inﬂuenced my mindset more than Dave Rohde. Dave provided my
introduction to rational choice theory, and, by example, emphasized the
importance and fun of thinking theoretically. He is the best logician of which I
know. Jim Granato and Bill Reed equipped me with a full and versatile
quantitative toolbox, and always emphasized the importance of theoretically
guided empirical tests. Last, but not least, I owe much to Doug Lemke. He
provided copious amounts of constructive feedback on every part of each essay.
This dissertation would be much worse without his help. He also introduced me
to the wonderful world of Sherlock Holmes and the concept of silent dogs! The
remaining ﬂaws in this dissertation are solely mine, and reﬂect my stubbornness
in not more closely following the advice of my committee.

While the aforementioned people have had the most inﬂuence on this
work, my academic training, and mindset, many other people played a role in
strengthening this dissertation. Brandon Prins always took time to answer

questions and comment on drafts of this work. Dave Lektzian provided many

helpful comments over many cups of coffee. Both Brandon and Dave have been
invaluable resources and friends. As a fellow in the Political Institutions and
Public Choice program, | beneﬁted greatly from the knowledge and camaraderie
of the other members: Bryan Marshall, Jamie Carson, Charles Finocchiaro,
Edurado Leoni, Mike Crespin, and Ryan Van der Wielen. They not only provided
a productive atmosphere in which to work, but made going to the ofﬁce
enjoyable. I also owe much to Jamie and Chuck for going out of their way in
helping me with the legwork and ﬁnal preparation of this manuscript. They have
done much to teach me the meaning of generosity. Eric Hoffman read the entire
manuscript and provided copious editorial comments. He was also a valuable
listening ear when I needed to vent dissertation frustrations. One could not ask
for a better friend.

Finally, I would like to thank a number of other people whose friendship
supported me throughout my graduate school days: Cathy, Chris, Carrie, Indridi,
Justin, Kirk, Mike, Shane, and Scott, who always provided a great story and a
good laugh. Most of all, I am indebted to my parents and brother for their love
and encouragement. Without their support, this dissertation would never have

been written.

vi

1'-

TABLE OF CONTENTS

List of Tables ........................................................................................ ix
Introduction .......................................................................................... 1

Chapter 2: Disaggregated Trade, Foreign Direct Investment, and International

Conﬂict .............................................................................................. 1 1
Literature Review ............................................................................. 13
Liberal Arguments ....................................................................... 14
Realist Arguments ....................................................................... 15
Empirical Research ..................................................................... 18
The Model and Hypotheses ............................................................... 21
Research Design ............................................................................. 33
Population of Cases and Dependent Variable .................................... 33
Independent Variables .................................................................. 35
Empirical Results ............................................................................. 42
Robustness and Sensitivity ............................................................ 46
Conclusion ..................................................................................... 50

Chapter 3: Economic Dependence and Militarized Interstate Dispute Initiation

and Targeting, 1970-1992 ...................................................................... 55
Commercial Dependence and Interstate Conflict .................................... 57
A Modiﬁed Realist Theory .................................................................. 64
The Directionality of Conﬂict ............................................................... 68
Hypotheses .................................................................................... 71
Research Design ............................................................................. 72

vii

Dependent Variables ................................................................... 74

Independent Variables ................................................................. 75
Empirical Results ............................................................................. 79
Initiation of Militarized Interstate Disputes ......................................... 79
Targets of Militarized Interstate Disputes .......................................... 82
Robustness and Sensitivity ............................................................ 83
Conclusion ..................................................................................... 85
Chapter 4: Institutional Similarity and Interstate Conﬂict .............................. 88
Satisfaction and Interstate Conﬂict ...................................................... 91
Domestic Institutions As A Source of Foreign Policy Preferences ............... 92
Institutional Similarity, Satisfaction, and Interstate Conﬂict ........................ 96
Dyadic Satisfaction vs. Systemic Satisfaction ........................................ 98
Measures of Satisfaction Based on Institutional Similarity ....................... 100
Political Institutions .................................................................... 1 03
Economic Institutions ................................................................. 104
Empirical Model ............................................................................. 108
Empirical Results ........................................................................... 1 12
Dyadic Satisfaction, Systemic Satisfaction, And Types of Conﬂict ............ 115
Conclusion .................................................................................... 118
Chapter 5: Conclusion ........................................................................ 121
Economic Dependence, Institutional Similarity, And State Sponsored
Terrorism ..................................................................................... 128
Future Research ............................................................................ 130
Appendix A ....................................................................................... 133

viii

Appendix B ....................................................................................... 147
Appendix C ....................................................................................... 154

Bibliography ...................................................................................... 156

ix

LIST OF TABLES

Table 2.1: NBER SITC Trade Categories ................................................ 133
Table 2.2: Total Dyadic Trade Interdependence and Militarized Conﬂict, 1970-
1992 ................................................................................................ 134
Table 2.3: Disaggregated Trade, Foreign Direct Investment and Militarized
Conﬂict, 1970-1992 ............................................................................. 135
Table 2.4: Conditional Trade Interdependence Model, 1970-1992 ................ 136

Table 2.5: Conditional Trade Interdependence and Fatal Militarized Disputes,
1970-1992 ........................................................................................ 137

Table 2.6: Conditional Trade Interdependence, FDI and Fatal Militarized
Disputes, 1970-1992 ........................................................................... 138

Table 3.1: Economic Dependence and the Initiation of Militarized Interstate
Disputes: An Event Count Analysis ........................................................ 139

Table 3.2: Developed Democracies, Economic Dependence, and the Initiation of
Militarized Interstate Disputes: An Event Count Analysis ............................ 140

Table 3.3: Economic Dependence and Targets of Militarized Interstate Disputes:
An Event Count Analysis ...................................................................... 141

Table 3.4: Initiation of Militarized Disputes With An Alternative Measure of
Militarization ...................................................................................... 142

Table 4.1: Dyadic Institutional Similarity and Militarized Interstate Conﬂict, 1982-
1992 ................................................................................................ 143

Table 4.2: Substantive Impact of Dyadic Satisfaction on Militarized Dispute
Onset ............................................................................................... 144

Table 4.3: Systemic Institutional Similarity and Militarized Interstate Conﬂict,
1982-1992 ........................................................................................ 145

Table 4.4: Institutional Similarity and Type of Militarized Interstate Dispute, 1982-
1992 ................................................................................................ 146

 

Introduction

“The globalization system,” writes Thomas Friedman, “has one
overarching characteristic: integration” (Friedman, 1999: 111). More speciﬁcally,
“globalization” refers to two speciﬁc developments in the post-World War II
international environment. The ﬁrst development is an increase in economic
interconnectedness between states. Economic interdependence has increased
because the system leader, the United States, has established a market oriented
economic system. Interdependence has also increased because of advances in
the communication and transportation industries. Technological improvements in
the areas of telecommunications, aviation, and shipping have lowered the cost of
conducting trade with ﬁrms in other states. The second development is the
spread of liberal political and economic institutions. By establishing democratic
political institutions and economic institutions that protect private property, states
are able to take full advantage of international markets. The adoption of liberal
political and economic institutions is also encouraged by the United States. With
the fall of the Soviet Union ﬁrmly ensconcing the United States as the dominant
state in the world and the already tangible beneﬁts of economic
interconnectedness, the process of globalization is likely to continue.

While much has been written on globalization, in general, and the
relationship between trade and conﬂict, we lack a comprehensive argument and
analysis of the relationship between different types of trade, foreign direct

investment, and international conﬂict. We also lack a full analysis of the

 

relationship between political and economic institutions and interstate dispute
onset. I address these issues in a series of three essays (Chapters 2, 3, and 4).
While each essay is self-contained, a focus on globalization and interstate
conﬂict unites them. In chapter 2, I advance a conditional theory of economic
interdependence. I argue that the effects of trade interdependence on the
likelihood of interstate conﬂict vary with the type of commodity traded and the
political relationship between countries. Trade in non-strategic commodities
contributes to peace between states. However, trade in strategic commodities
between states lacking political afﬁnity increases the likelihood of conﬂict,
whereas trade in strategic commodities between states with similar foreign policy
preferences strengthens pacific motivations. Further, I contend that foreign direct
investment contributes to international peace and stability. In chapter 3, I apply
these arguments to state level behavior. I argue that, in general, economic
dependence on international trade and investment makes states less likely to
initiate militarized disputes, but dependence on trade makes states more likely to
become targets of disputes. These relationships are also conditional. The
pacifying inﬂuence of trade dependence varies with the commodities on which a
state is economically dependent and its level of militarization. Military states
dependent on importing strategic commodities are more likely than other states
to initiate militarized interstates. Relatedly, producers of strategic commodities
are more likely to be attacked by other states. In chapter 4, I argue that the
similarity of economic and political institutions in a dyad contributes to

satisfaction with the status quo. As a result, dyads with similar institutions are

 

 

less likely to experience militarized conﬂict. In short, this dissertation examines
the nexus between globalization and international conﬂict. In the next section, I
provide an overview of each chapter and preview the empirical ﬁndings.

In chapter 2, I develop an analytical model of economic interdependence
and international conﬂict that disaggregates dyadic trade into strategic and non-
strategic commodities and includes foreign direct investment. This analytical
model addresses two extant weaknesses in the literature on economic
interdependence and conﬂict. First, current research tends to treat all trade the
same. No distinction is made between trade in strategic commodities like oil and
non-strategic commodities like textiles. Yet, existing theoretical arguments
suggest the need for disaggregating trade into strategic and non-strategic
commodities. Liberal theories, for example, argue that interdependence
enhances the prospects of peace between nations by increasing the opportunity
costs of conﬂict, strengthening normative ties, and providing an additional means
to signal clearly one’s intentions. Realist theories counter that all types of trade
interdependence are not the same. Interdependence in non-strategic
commodities likely has no effect on the likelihood of conflict between states.
However, interdependence in strategic commodities is likely to increase the risk
of dyadic conﬂict. Similarly, McMilIian (1997: 53) writes: “If trade in strategic
goods creates conﬂict, it would support the realist hypothesis, but it would not
necessarily undermine the arguments and expectations of liberalism if trade in
other types of goods tends to inhibit conﬂict.” Strategic and non-strategic trade

are likely to have different effects on conﬂict behavior because they differentially

 

affect a state’s vulnerability. Since states aim to reduce their vulnerability on
others, strategic trade interdependence may increase the likelihood of conﬂict
between states. More importantly, models and empirical tests that do not
disaggregate trade into strategic and non-strategic commodities are not modeling
and testing the appropriate arguments.

Other researchers have made similar observations. Morrow, Siverson,
and Taberas state that if we are to express conﬁdence in our empirical ﬁndings
on trade and conﬂict, it is necessary to use “lower levels of analysis than
aggregate trade ﬂows” (1998: 659). Perhaps most importantly, Reuvany and
Kang (1998), in the only research to study disaggregated trade, ﬁnds that the
trade and conflict relationship depends on the goods traded. Given this variation
across goods, he notes that there is a “need to use disaggregated trade data in
future research” (Reuvany, 1999: 37).

A second weakness in the current literature on economic interdependence
and conﬂict is that economic interdependence is deﬁned exclusively in terms of
trade. While trade is an important feature of economic interdependence,
globalization has increased the ﬂow of foreign direct investment more than trade.
The Economist notes that “foreign investment in the past three decades has risen
faster than trade and world output” (Economist, June 18, 1998). Similarly, the
United Nations Conference on Trade and Development ﬁnds that “foreign direct
investment jumped 27 percent in 1999 to reach an all-time high” (Foreign Policy,
2001 ). Studies of economic interdependence that ignore foreign direct

investment are likely to have biased results due to an omitted variable problem.

From the analytical model, I derive propositions on the relationship
between strategic trade, non-strategic trade, foreign investment and interstate
conﬂict. In brief, I argue that political leaders want to retain their hold on power.
To stay in power, leaders need to balance domestic welfare and international
security goals, for a failure in either area can reduce a leader’s tenure in ofﬁce.
International commerce contributes to domestic wealth, giving political leaders an
incentive to engage in trade. International conﬂict, however, increases the risk
and costs of economic transactions. As such, international conﬂict reduces the
amount of trade between nations. Dyads with the most trade, then, face the
greatest costs from interstate conﬂict. However, dependence on strategic
commodities from a hostile regime enhances the vulnerability of a leader’s hold
on power. To protect their positions in power, leaders strive to reduce their
vulnerability making strategic trade interdependence with adversarial states a
source of conﬂict. Foreign direct investment, on the other hand, always reduces
the likelihood of conﬂict between states. Like trade, foreign direct investment
enhances domestic wealth. However, unlike strategic trade it does not enhance
vulnerability nor does it provide an incentive for conquest. While trade can be
shut off quickly, foreign investments typically take the form of ﬁxed assets in
another country. In addition, foreign direct investment is a form of “extracting”
wealth from another nation; thus, high levels of foreign direct investment also
reduce the potential beneﬁts of conquest.

I empirically test these hypotheses by examining the onset of militarized

interstate disputes for all dyads in the international system over the period 1970

 

to 1992. The empirical evidence supports the propositions of the analytical
model. Non-strategic trade interdependence and foreign direct investment
reduce the likelihood of a militarized interstate dispute occurring. Strategic trade
interdependence, however, increases the likelihood of conﬂict between states
lacking political afﬁnity, but strengthens peaceful motivations between states with
similar preferences. The arguments and ﬁndings concerning the conditional
effect of strategic trade and the paciﬁc effects of foreign direct investment are
especially noteworthy. The central question countering arguments linking
economic interdependence with peace is the outbreak of World War I. This
research helps address the apparently anomalous case of the First World War.
First, economic interdependence does not guarantee peace, but it does increase
the prospects for peace. Second, the arguments advanced here suggest that we
should have expected conﬂict between European nations prior to World War I
given their high level of strategic trade interdependence and lack of political
afﬁnity. Third, while the major European powers did exhibit high levels of trade
interdependence, there was little foreign investment between them. The growth
of foreign investment is what sets modern economic interdependence apart from
previous forms of interdependence.

In chapter 3, I draw on the work of Richard Rosecrance (1986, 1999) and
argue that nations choose one of two paths to promoting domestic wealth:
territorial aggrandizement or commercial transactions. States centered on
political-military interests emphasize territorial aggrandizement and pursue

militarily aggressive foreign policies. In contrast, states focused on commercial

 

interests emphasize exchange and pursue more peaceful foreign policies. As
states focus on exchange, they become more economically dependent on
international commerce in terms of both trade and investment. In turn, economic
dependence increases the opportunity costs of conﬂict, thus encouraging state
leaders to pursue more peaceful foreign policies.

As before, these arguments concerning the peaceful effects of economic
engagement are conditional on the type of goods traded. Militarized states
importing strategic commodities are more likely than other states to initiate
militarized disputes. Similarly, less developed states exporting strategic
commodities are more likely to be targets of militarized disputes. These
arguments concerning the differential and conditional effects of trade are
important modiﬁcations to extant arguments that simply suggest all trade
contributes to peace in all circumstances.

The theoretical arguments and empirical tests in chapter 3 emphasize
states, instead of dyads. This is important since many arguments on economic
dependence and conﬂict emphasize state level behavior. The difﬁculty with state
level arguments on international conﬂict is that conﬂict requires two or more
actors. To properly evaluate these hypotheses, I examine initiated and targeted,
rather than participation in, militarized interstate disputes from 1970 to 1992. The
empirical evidence indicates that trade dependence on non-strategic
commodities and foreign direct investment reduces the initiation of militarized
disputes, whereas militarized states dependent on importing strategic

commodities increases the likelihood of aggressive foreign policies. Likewise,

resource rich poor states are more likely than other states to be the targets of
militarized disputes. Perhaps most important, I ﬁnd that foreign direct investment
reduces both the likelihood of a state initiating a dispute and the likelihood that a
state will become the target of a militarized dispute.

I also ﬁnd signiﬁcant differences between developed democracies and
developed autocracies. Speciﬁcally, developed democracies are less likely to
initiate disputes, though developed nations, in general, are more likely to initiate
disputes. Put differently, this analysis uncovers a monadic democratic peace in
the post-World War II period. After controlling for development, interdependence,
and global reach, democracies are still less likely than non-democracies to
initiate militarized interstate disputes.

In chapter 4, I turn attention to the second effect of globalization, the
spread of liberal institutions. The research question under investigation is
whether or not similar political and economic institutions reduce the likelihood of
conﬂict between states? I argue that foreign policy preferences are a reflection
of a state’s institutions. Thus, states with similar political and economic
institutions are likely to have similar foreign policy preferences. When states
share similar preferences, they are satisﬁed with the status quo. An additional
question arises at this point. Is satisfaction best conceived at the dyadic or the
systemic level? I create measures of satisfaction for both levels and the let the
data speak.

After conducting an empirical analysis of economic and political

institutional similarity and the onset of militarized interstate disputes, I conclude

 

that institutional similarity, at both the dyadic and systemic levels of analysis,
reduces the likelihood of conﬂict. As expected, dyads with similar institutions
have less to ﬁght over, so they ﬁght less. I also examine the relationship
between economic and political institutional similarity and different types of
conﬂict. Does institutional similarity reduce the likelihood of all types of
militarized disputes or are some types of disputes more likely with institutional
similarity than other types of disputes? I ﬁnd that the strongest influence of
institutional similarity, whether in its political or economic form, is on reducing the
likelihood of disputes over regime changes. I also ﬁnd that economic institutional
similarity tends to reduce the likelihood of territorial disputes more than political
similarity. This last ﬁnding is positive news given the adoption of market
institutions around the world, especially in China. Standard correlates of war
might suggest the United States and China are heading for conflict over Taiwan,
a territorial dispute. As China becomes more economically similar to the United
States, however, this research suggests that military conﬂict becomes less likely.
Political pundits like Thomas Friedman note that globalization is “shaping
the domestic politics, economic policies, and foreign relations of virtually every
country” (Friedman, 1999: 110). Further, Friedman contends that globalization is
a positive development in terms of contributing to peace between nations. In
contrast, I argue that the effects of globalization are less straightforward. Foreign
direct investment, non-strategic trade, and economic institutional similarity
contribute to peace between nations. Each of these forces contributes to peace

because they align the foreign policy preferences of states. Yet, when

preferences are dissimilar and a source of tension exists, such as trade in
strategic commodities, economic interdependence may increase the likelihood of
conﬂict between states. In summary, this dissertation ﬁlls important gaps in the

extant research on globalization and international conﬂict.

10

Chapter 2

Disaggregated Trade, Foreign Direct Investment, and Interstate Conﬂict

Since the end of World War II, and especially in the last twenty years, the
world has become increasingly economically interdependent. This growing
economic interdependence between nations is often referred to as globalization.
For theorists of international relations and foreign policymakers, a pressing
question about globalization concerns its effects on international conﬂict and
cooperation. Will increased economic interdependence contribute to peace
between nations or will it exacerbate tensions and contribute to increased
hostility and conflict?

Some international relations scholars argue that interdependence
promotes peace. Oneal and Russett, for example, claim that “policymakers
avoid the use of force against states with which they engage in economically
important trade” (Oneal and Russett, 1999a: 4-5). They argue that conﬂict
disrupts trade; therefore, trade promotes peace by increasing the costs of
conflict. Other scholars argue that economic interdependence enhances the
likelihood of conﬂict between states. Kenneth Waltz, for instance, writes that
“interdependence hastens the occasion for war” (1979: 138). For Waltz, anything
that increases vulnerability increases the prospects for militarized conﬂict, and
extensive economic linkages increase a state’s vulnerability by making it more
dependent on other states. From Waltz’s realist perspective, all trade is not the

same; the various types of trade do not generate equal amounts of vulnerability.

11

 

Rather, “dependence on the importation of strategic goods increases the
likelihood of conﬂict, since countries tend to pursue aggressive expansionist
policies to ensure the supply of such goods” (Uchitel, 1993).

I argue that to more accurately evaluate the effects of economic
interdependence on international conﬂict it is necessary to both disaggregate
trade into strategic and non-strategic categories and to consider the effects of
foreign direct investment (FDI). Over-aggregating trade is both theoretically and
empirically unsound.1 Theoretically, it does not afford an accurate assessment of
the realist argument. Empirical evidence also indicates that the relationship
between trade and conﬂict differs across commodities (Reuvany and Kang,
1998). Economic interdependence also involves more than trade alone; thus,
studies that omit foreign direct investment may have biased results. While trade
ﬂows have increased threefold in the last thirty years, foreign direct investment
has increased sixfold (Economist, June 18, 1998). Further, with FDI a state is
already “extracting” economic beneﬁts from another state, so FDI decreases the
potential beneﬁts of conquest. Despite their importance, we lack a theoretical
model of economic interdependence and conﬂict that includes strategic trade,
non-strategic trade, and foreign direct investment. We also lack an empirical
analysis of the relationship between different types of trade, foreign direct
investment, and interstate conﬂict.

In the next section, I review the literature on economic interdependence

and conﬂict. This review indicates that previous studies have over-aggregated

 

' Other scholars have also called for the disaggregation of trade (see McMillian, I997; Reuvany, 1999; and
King and Zeng, 200]).

12

 

trade and neglected the impact of foreign direct investments. In the third section,
I incorporate strategic trade, non-strategic trade, and foreign direct investment
into an analytical model of economic interdependence and conflict. This model is
an extension of Polachek’s (1980) path-breaking work in this research area.
From the model, I derive hypotheses that non-strategic trade and foreign direct
investment should reduce conﬂict between nations, while the effect of strategic
trade on the likelihood of conﬂict between nations is conditional on the political
afﬁnity of the two nations. In the fourth section of the paper, I present an
empirical model to evaluate the hypotheses both from the analytical model of
section three and from contending explanations. In the ﬁfth section of the paper,
I present the results of the empirical analyses. This is the ﬁrst large-scale
empirical study of disaggregated trade and conﬂict and the ﬁrst study to analyze
both different types of trade and foreign direct investment. The ﬁndings support
the hypotheses of the analytical model, and are robust to a number of alternative
speciﬁcations. In a ﬁnal section, I discuss implications of this research and

suggest avenues for future research.

Literature Review

This section reviews the extant literature on interdependence and conflict
in order to illuminate the conceptual relationships between interdependence and
conﬂict and reveal the areas that need further attention. Theoretical explanations
linking economic interdependence to international conﬂict fall into two categories:
liberal and realist. Liberal explanations emphasize the normative and

constraining forces associated with interdependence, and conclude that these

13

forces enhance the prospects for peace between nations. Realist explanations
emphasize the vulnerability created by interdependence, and conclude that in a

self-help, anarchic system such vulnerability fosters insecurity and conﬂict.

Liberal Arguments

Liberal theories linking economic interdependence to international conﬂict
emphasize connectedness and costs. “Free trade,” according to David Hume, “is
the vital principle by which the nations of the earth are to become united in one

harmonious whole.”2

According to John Stuart Mill, commerce produces
harmony and renders “war obsolete, by strengthening and multiplying the
personal interests which are in natural opposition to it” (Mill, 1848: 582).
Other classical political philosophers like the Baron de Montesquieu
emphasize the costs of disrupting trade as peace enhancing.
“Peace is the natural effect of trade. Two nations who trafﬁc with each
other become reciprocally dependent; for if one has an interest in buying,
the other has interest in selling; and thus their union is founded on their
mutual necessities” (de Montesquieu, 1900: 316)
Relatedly, Norman Angell argues that trade produces greater beneﬁts than
conﬂict. “The great danger of the modern world is not absolute shortage, but
dislocation of the process of exchange, by which alone the fruits of the earth can
be made available for human consumption” (1912: 31).
Modern theorists also emphasize the costs of trade. Anchoring his

analysis in economic theory, Polachek (1980) argues that states want to

maximize their social welfare. Since conflict disrupts trade, it reduces a nation’s

 

2 Quoted in Burton, 1946: 521.

14

welfare and should be avoided. “If conﬂict leads to a cessation or at least a
diminution of trade (perhaps through tariffs or quotas), then countries with the
greatest gains from trade face the highest costs of potentially lost trade and
hence engage in the least conflict and the most cooperation” (Polachek, Robst,
Chang, 1999: 405). Rosecrance (1986) notes that the relationship between trade
and conﬂict is dependent not only on the utility of trade but also on the utility of
war. According to Rosecrance, the choice between conﬂict and war “depends
upon the cost and beneﬁt of waging war on the one hand and engaging in trade
on the other. The greater the restraints on trade and the fewer its likely beneﬁts,
the more willing nations have been to seek to improve their position through

military force” (1986: 31 ).

Realist Arguments

Realist explanations revolve around the nature of the international system
and the goal of states. For realists, the international system is anarchic, and
therefore self-help. The primary goal of states is security (or power). To
determine the effect of intervening variables like economic interdependence, one
examines how they impact security in such an environment. In general, realists
conclude that in an anarchic environment “interdependence hastens the occasion
for war” (Waltz, 1979: 138). Interdependence is harmful as it compromises self-
reliance, thus making a nation more vulnerable. Vulnerability, in turn, generates
insecurity thereby pushing leaders into taking aggressive actions to alleviate their

vulnerability.

15

The anarchic nature of international politics and the pursuit of security are
not the only reasons economic interdependence enhances vulnerability. In an
anarchic environment, nations must focus not only on their security, but also on
relative gains. “Relative gains sensitivity is affected by the political-military
relationship between the nations involved, the offense-defense balance, and
system structure” (Libennan, 1996: 148). While interdependence via trade may
always produce absolute gains, it may also lead to a relative loss. In Gowa’s
terms (1989) such a loss is a negative security externality. The gains from trade
may be used for any purpose, thus trade “increases the potential military power
of any country that engages in it” (Hirschman in Baldwin, 1985: 211). As a result,
nations should be more likely to trade with allies than adversaries.

Trade may not only produce a security externality, it may also lead to
conﬂict by enhancing rivalry and enmity between nations. The driving force
behind international trade is comparative advantage, which encourages a certain
amount of specialization in order to maximize proﬁt. Whether or not it is
economically accurate, policymakers typically perceive international economics
as a competition between states. When trade is viewed as a competition
between states, it can lead to interstate rivalry and even militarized conﬂict.
Prestowitz (1988) and Reich (1991), for example, use the idea that economics is
a competition between states to advocate government subsidization of industries
they view as “strategic,” such as semiconductor manufacturers. In a self-help
system where nations are concerned about security, an industry is perceived as

having a multiplier effects, and a premium is placed on national champions, then

16

“trade conﬂict bodes ill for peace, since highly interdependent nations might be
tempted to regain lost markets and resources by force” (Liberman, 1996: 148).

Whether or not nations trade is a choice, and when evaluating any choice
we must consider the alternatives. Anderton, Anderton, and Carter (1999) note
that economic exchange may bring beneﬁts, but so does appropriation. In fact, if
leaders believe appropriation will produce greater beneﬁts than exchange, then
nations will pursue expansionistic policies. “The promise of capturing economic
beneﬁts from conquered territory historically has been a signiﬁcant motivating
force for war,” writes Brooks; such “wars of conquest still occur, as Iraq’s recent
invasion of Kuwait amply demonstrates” (Brooks, 1999: 646). Expressing the
view of economic nationalists, Rotte (1997) argues that “war can be proﬁtable,
since, by territorial expansion, it can not only secure new markets, resources,
and commercial supremacy, but also contribute to a country’s industrialization
and production potential” (Rotte, 1997: 10).

The strategy of appropriation is most likely to occur when strategic goods
are involved. “A nation’s economy is particularly dependent on imports of goods
for which demand is highly inelastic and domestic production is extremely
inefﬁcient, especially those that have multiplier effects on the whole economy”
(Liberman, 1996: 154). For Liberman, when strategic goods are involved,
nations are especially apt to focus on relative gains, thus increasing the
likelihood of conﬂict. Put differently, since trade in strategic goods creates
greater vulnerability than trade in non-strategic goods, this former type of trade

increases the likelihood of conﬂict.

l7

Empirical Research

In recent years, scholars have conducted a number of empirical tests of
the economic interdependence, conceived almost exclusively in terms of trade,
and conﬂict relationship. Polachek’s seminal 1980 article is one of the ﬁrst
empirical tests of the relationship between trade and interstate conﬂict. He
examines thirty states over a ten year period and ﬁnds that trade reduces
conﬂict. In subsequent work, Polachek and his associates have expanded the ..
empirical analysis to about a hundred states over a thirty year time period (1992;
Polachek, Robst, and Chang 1999; Gasiorowski, 1986). In each of these
analyses, the central dependent variable is net conﬂict, which is a function of acts
of cooperation and conﬂict, while the central independent variable is dyadic trade
as a percentage of gross domestic product. All of these studies ﬁnd that higher
levels of trade are associated with lower levels of conﬂict.

Domke (1988) conducts a monadic analysis and examines a much longer
period, 1871-1975, than the Polachek research group. He also uses a different
dependent variable. Polachek and Gasiorowski use the COPDAB database to
construct an index of cooperation, whereas Domke examines the number of wars
a country participated in. He also ﬁnds that trade promotes peace. Mansﬁeld
(1994) conducts a systemic investigation of the relationship between trade and
conﬂict and ﬁnds that the amount of trade in the international system is inversely
related to the amount of war in the system.

In recent years, several researchers have conducted large-N dyadic

analyses of economic interdependence and conﬂict. In a series of articles on the

18

 

liberal peace, Oneal and Russett and their colleagues assess the relationship
between dyadic trade as a percentage of gross domestic product and militarized
interstate disputes (1997, 1999a, 1999b). They consistently ﬁnd that trade
decreases the likelihood of a militarized interstate dispute occurring.

The ﬁndings of Oneal and Russett have not gone unchallenged. Barbieri
(1996) examines all dyads between 1870 and 1938, uses militarized interstate
disputes as her dependent variable and ﬁnds that trade does not promote peace.
Similarly, Beck, Katz, and Tucker (1998) have called into question the early
ﬁndings of Oneal and Russett. They argue that Oneal and Russett’s research
suffers from a methodological problem, namely that in large cross-section time-
series analyses one needs to control for temporal dependence. When they
replicate Oneal and Russett’s work, while controlling for temporal dependence
between observations, they ﬁnd that trade is no longer associated with peace.

A third group of researchers emphasize the potential endogeneity of both
trade and conﬂict. Reuvany and Kang (1998) ﬁnd a reciprocal relationship
between conﬂict and bilateral trade. In some of the sixteen dyads they analyze,
conﬂict causes trade, while in other dyads, trade causes conﬂict. More
importantly, Reuvany and Kang (1998) is the only study to date that analyzes
disaggregated trade, and they find that different types of goods have different
relationships to conﬂict. Gartzke, Li, and Boehmer (2000) also argue that trade
and conﬂict are both endogenous. One important implication of this argument is
the relationship between trade and peace. If dyads that should be trading are not

trading, then we should expect to see more conﬂict between them. Indeed,

19

Reuvany and Kang find that higher levels of trade are associated with lower
levels of conﬂict. Gartzke, Li, and Boehmer (2000) is also signiﬁcant because
they include foreign direct investment under the rubric of economic
interdependence. In fact, they ﬁnd that foreign direct investment has a greater
pacifying influence than dyadic trade.

This brief review of the literature reveals three important points. First, the
effects of trade on international conﬂict are not clear. Some scholars argue that
trade promotes peace, while other scholars suggest trade promotes conﬂict.
Reuvany contends that the competing empirical claims regarding the relationship
between trade and conﬂict may be the result of omitted variables (1999: 30).

Second, except in one study, researchers have treated all trade as if there
were no differences between commodities. Given realist arguments regarding
the importance of particular types of goods, this is an important limitation to
previous studies. “In addition, despite the extant research that emphasizes the
nature of the goods being traded. . .we still have little knowledge about whether
trade in certain types of goods has a signiﬁcant impact on conﬂict. If trade in
strategic goods creates conﬂict, it would support the realist hypothesis, but it
would not necessarily undermine the arguments and expectations of liberalism if
trade in other types of goods tends to inhibit conflict” (McMillan, 1997: 53).
Reuvany also notes that not considering variation in the type of goods traded is a
central weakness in the current research program.

Third, almost all previous research has narrowly operationalized economic

interdependence as trade, ignoring the importance of foreign direct investment. I

20

argue that this over-aggregation of trade and the omission of foreign direct
investment is both unnecessary and results in an inaccurate empirical
examination of the relevant theories. In the next section I develop a theory of
economic interdependence and interstate conﬂict that incorporates both trade in

different types of goods and foreign direct investment.

The Model and Hypotheses

While the focus of this research is on the political relationship between
states, the primary factors driving international trade and investment are
economic. To this end, I posit an expected utility model grounded in economic
theory.

First, I assume state leaders are rational, unitary actors. These
assumptions are common to both the liberal and realist perspectives. The
assumption of rationality means leaders have preferences over alternative
outcomes, they can compare these preferences, and order them in a transitive
manner. Put differently, the assumption of instrumental rationality means leaders
have goals and they attempt to attain these goals.3 Treating states composed of
many bureaucracies and people as unitary actors is a plausible assumption when
the focus is foreign policy decision-making. “In these situations, most, if not all,

modern states formally confer on a single individual the choice to invoke force"

 

3 In the realm of foreign policy, where the stakes of a decision are high, this assumption is highly plausible.
In addition, the instrumental rationality assumption “is the only general assumption of decision making
available. Any other assumption requires detailed, actor-specific information to make behavioral
predictions” (Lake, 1999: 40). For a more thorough discussion of this assumption, see Bueno de Mesquita,
1999, and as it applies to states, see Bendor and Hammond, 1992.

21

(Bueno de Mesquita and Lalman, 1992: 26). Plus, this assumption does not
preclude competing domestic interests (Huth, 1996: 36).4

Given that ﬁrms, not states, conduct trade and investment, it may seem
problematic to employ a unitary actor assumption. Governments, however,
greatly inﬂuence the amount and type of business transactions that occur. First,
governments establish the regulatory environment in which ﬁrms operate, and by
extension the amount of business that is conducted. For instance, more
commercial activity will occur in an environment that protects property rights than
in an environment with non-secure property rights. In economic terms, greater
protection of property rights lowers transaction costs, thereby facilitating the
exchange of goods and capital. Second, governments may promote particular
industries, or even ﬁrms. Governments promote speciﬁc industries by providing
subsidies and tax breaks, and they do this for the political purpose of enhancing
their hold on power. Subsidies and tax breaks, then, are another way that
governments alter the costs of conducting business, and therefore inﬂuence the
amount and type of trade that occurs both within and between states. In
summary, while the state is rarely an economic agent, when examining
transactions aggregated at the state level it is useful to abstract to the state.

Next, I assume state leaders desire to remain in ofﬁce (Bueno de
Mesquita and Siverson, 1995; Bueno de Mesquita, Morrow, Siverson, and Smith,
1999). This assumption does not mean leaders only have a single goal. Rather,

it suggests leaders are aware of the beneﬁts of ofﬁce, and the best way, typically,

 

4 The seminal defense of the unitary actor assumption is Bueno de Mesquita, 1981: 20-23. Bueno de
Mesquita and Lalman (1992: 25-30) and Huth (1996: 35-36) also provide extensive discussions of the
unitary actor assumption.

22

to attain their other policy goals is to remain in ofﬁce. This assumption also
highlights the importance of domestic political factors. Normally, when a leader
loses political ofﬁce it is the result of domestic forces. It is clear how this applies
to democratic political systems, where elections determine one’s time in ofﬁce.
The fall from power of Mikhail Gorbachev, however, shows its applicability in
non-democratic political systems. This “ofﬁce” assumption, then, anchors the
argument in domestic politics.

Next, I assume that to stay in ofﬁce leaders maximize a mix of domestic
welfare and international security. Put differently, politicians optimize domestic
welfare and international security in order to counter the two types of forces,
internal and external, that may remove a leader from power. Leaders maximize
domestic welfare as it is the most effective way to ward off domestic opponents.
Lewis Beck (1988) has found extensive empirical support for the proposition that
economic prosperity enhances a leader’s tenure in ofﬁce. It is worth noting that
the wealth maximization impulse applies to both autocratic and democratic
regimes and it does not require the equal distribution of income. Whether a
leader bestows the state’s wealth on a small number of key supporters or
establishes a system in which the majority of people in a society beneﬁts is
irrelevant. Rather, the wealthier a state, in absolute terms, the more secure are
the political leaders.

However, the maximization of domestic welfare is not done in a vacuum.
Since external forces may also remove a leader from ofﬁce, security

considerations inﬂuence the pursuit of welfare. While economic interdependence

23

may always produce an absolute economic gain, it may also lead to a relative
security loss. This may occur because the gains from trade may be used for any
purpose; thus trade “increases the potential military power of any country that
engages in it” (Hirschman in Baldwin, 1985: 211). Because of security
externalities (Gowa, 1989), nations should be more likely to trade with allies than
adversaries.

Based on Polachek (1980), I deﬁne an actor’s political welfare function as
W(C, 2), where C represents total goods and services consumption and Z
represents conﬂict toward a particular country. Further, I deﬁne C as a function

of domestic consumption, investment, exports and imports. Speciﬁcally,

C=q-I-Xs-Xn+Md+Me (1)
Where q represents domestic consumption, (I) denotes foreign direct investment
in another nation, X8 and Xn signify strategic exports and non-strategic exports,
and Md and Ma stand for strategic and non-strategic imports. Essentially,
equation 1 is the standard national income identity for an open economy, where
C approximates a nation’s gross national product (GNP). In turn, GNP measures
the productivity, or wealth, of a nation. GNP works on the principle that all
production must go someplace, so it measures how output is used. In most
countries, domestic consumption, q, consumes most output. In an open
economy, some output is sold to other countries, therefore exports are
subtracted from the total goods and services consumption in a state. Similarly,
output invested in another country, (I), is also subtracted from the home state’s

consumption. Finally, some consumption originates from another country, thus,

24

imports are added to a state’s goods and services consumption. This view of an
actor’s political welfare function speciﬁcally incorporates both domestic and
international factors. Leaders may suffer politically for purely domestic reasons
(e.g. weak domestic consumption) or from a combination of domestic and
international factors (eg. minimal investment in the economy or decreases in
international trade).

This depiction of consumption differs from Polachek’s (1980) in two ways.
First, I include a term, (I), to capture the effects of foreign direct investment. The
inclusion of this variable is motivated by both economic and international
relations theory. Economic interdependence between states is not limited to
trade. It also involves foreign investment. “The point about FDI is that it is far
more than mere “capital”: it is a uniquely potent bundle of capital, contacts, and
managerial and technological knowledge. It is the cutting edge of globalisation”
(Economist, February 24, 2001).

Foreign investment may occur in one of two forms: portfolio investment or
foreign direct investment. “Portfolio investments are holdings of stocks and
bonds designed to earn dividends and interest rather than exercise control over
the use of foreign facilities” (Kenen, 2000: 163). “Direct investments create,
extend, or facilitate control over productive facilities in other countries. They are
the building blocks of multinational enterprises. . (Kenen, 2000: 280). The
difference in control over resources makes foreign direct investments a greater

source of economic interdependence. Because it is a central part of economic

25

interdependence, I include foreign direct Investment in the maximization
calculations of states.

In addition, foreign direct investment modiﬁes the willingness of a state to
engage in conﬂict with another state. Brooks, for instance, writes: “In general,
as a state is increasingly able to rely on multinational corporations (MNCs) to
secure needed external resources and supplies, the overall willingness of that
state to engage in conquest should decrease” (Brooks, 1999: 666). Indeed, it is
interesting to note that although there were large portfolio investments between
the major opposing European powers prior to World War I, the amount of direct
investments was small.

The other difference between Polachek’s model and the one presented
here is that l disaggregate international trade into strategic and non-strategic
categories. Both liberal and realist motivations encourage the disaggregation of
trade, yet very little work has been done in this area. In the liberal perspective,
dependence on trade creates a constraint on conﬂict. If some types of goods are
more important, for example strategic goods, then they should create a greater
dependence and thus decrease the likelihood of conﬂict. Reallsts also argue that
some types of goods are more important than other types of goods. However,
the vulnerability attached to dependence on strategic goods promotes aggressive
foreign policies. Similarly, the argument of economic nationalists is that conﬂict
may be beneﬁcial if it involves regaining lost markets in strategic commodities.
Finally, Reuvany (1999) has found that “the gains from trade vary across goods,”

suggesting an empirical foundation for disaggregating trade.

26

I also argue that interstate conﬂict decreases foreign direct investment and
trade, thus making investment and trade functions of conﬂict. This assumption
follows from two arguments. First, nations engaged in militarized conﬂict are
unlikely to trade with or invest in the enemy.5 Second, conﬂict increases the risk,
and thus the cost, of commercial transactions; as a result, proﬁts decrease so
ﬁrms will be less likely to pursue those markets. If these arguments are correct,
then there should be less foreign investment in politically unstable areas, where
the risk of conducting business transactions is high. Evidence on the patterns of
foreign investment throughout the world supports this contention (see eg. Sobel,
1999).

In addition, I assume the political welfare function is separable in C and 2
(Wm = 0),6 and that there is a positive, but diminishing marginal utility of
consumption (Wc > 0 and Woo < 0) and conﬂict (Wz > 0 and WH < 0). The
assumption of a diminishing marginal utility implies that the gains from either
consumption or conﬂict decrease as one achieves higher levels of consumption
or conﬂict. For example, if a person is very thirsty, the utility of the ﬁrst glass of
water they consume is greater than the utility of the second glass of water.

Finally, all actors are subject to a balance-of-payments constraint.
Inclusion of a balance-of-payments constraint highlights the ever present reality
of limited resources, and it makes a state’s domestic welfare subject to market

forces. Put differently, without a balance of payments constraint, commercial

 

5 Barbieri and Levy (1999) ﬁnd that nations at war may not completely cut their trade ties, though their
analysis does suggest that conﬂict reduces trade.

6 Polachek, Robst, and Chang (1999) provide a clear discussion of this assumption. As they note, it is
“primarily for ease of exposition” (ﬁt 4).

27

actors in one state would not have to pay commercial agents in another state. I

assume the following constraint:

R, + PAxs + Pxn — Pmd — PAme = 0 (2)
Because a state’s balance of payments depend on revenue from investment and
the prices of goods, not the quantity, equation 2 contains revenue and price
terms, eg. R, represents the rate of return on a foreign investment, Pxn
represents the price value of non-strategic exports. Essentially, this means a
state’s current account plus investment equals zero. Further, in light of the
international security goal, it is necessary to modify the cost of strategic exports
and imports by political factors. In other words, strategic goods involve both an
economic and a political cost.

Given these conditions, an actor chooses a particular level of conﬂict, 2,
that maximizes W (C, 2) subject to the balance of payments constraint.
Therefore, an actor maximizes the following Lagrangian (L):

L = W(Z; q—l-Xs-Xn + M, + M) + MR, + Pxn + PAxs- Pmd- PAme)

(3)

Differentiating L with respect to conﬂict, Z, yields a set of ﬁrst order conditions
(FOCs) for optimal conﬂict.

01/52 = w'(z,- q — I- Xs — Xn + Md + Me) + MRI/(z) + X’nPn ’(z) + X’sPAs’(z)
- M’de’(z)- M’ePAe’(z)] (4.1)

ova/1 = RI (2) + XnPn (z) + XsPAs (z) — Mde (z) - MePAe (z)
(4.2)

In equilibrium, the marginal gains from conﬂict equal the marginal costs of

conﬂict. Rearranging equation 4.1 shows the marginal gains from conﬂict are

28

W71, while the marginal costs of conﬂict are [M an’(z) + M’ePAe ’(z) - Rl’(z) -
X ’,,Pn ’(z) - ’SPAs’(z)]. Keeping in mind that Pd ’and Pe’are positive, and RI’,
Pn’, and Ps’are negative, it is clear that an increase in imports, exports, or
investment increases the marginal costs of conﬂict, which in turn decreases the
gains from conﬂict.

Taking the total differential of the F005 yields a set of second order

Iltl
It

conditions, from which we can derive comparative static equilibria regarding the
relationship between economic interdependence and conﬂict. First, I analyze the
effect of a change in non-strategic exports on conﬂict by taking the total
differential of the FOCs and using Cramer’s rule on the resulting system of
equations to solve for dZ/an.

dZ/an = (5)
-[Pn (Z)]/[I o‘R/o‘Z (q— I-Xs-Xn + Md + Me) +Xs o‘Ps/o‘Z (q- I—Xs-Xn +
Md+ Me) +Xn Bonk? (q— I—Xs-Xn + Md+ Me) — Md and/52 (q- I—Xs—Xn
+ Md+ Me)— Me o‘Me/o‘Z (q- I—Xs-Xn + Md+ Me)]

To determine the effect of an increase in non-strategic exports, one needs to ﬁnd
the sign of this fraction. Prices are always positive, but multiplying this by a
negative makes the numerator negative. Assuming a well-behaved utility
function, then the Hessian must be negative deﬁnite; thus, the principle minors
alternate sign making the above second order conditions produce a positive

denominator. The whole term is then negative. Thus,

dZ/an < 0.

Substantively, this means conﬂict decreases as non-strategic exports

increase. Conﬂict decreases because an increase in non-strategic exports

29

 

produces a welfare gain. Using the same procedure, one can analyze the effect
of a change in strategic exports, non-strategic imports, strategic imports, and
foreign direct investment on conﬂict behavior. The results are as follows (see
Appendix A for the derivations). States with the highest amounts of non-strategic
trade, imports and exports, and investment have the most to lose from conﬂict;
thus, high amounts of non-strategic trade reduce the likelihood of conﬂict
occurring between states. The story is different with strategic commerce. With
non-strategic commerce, welfare considerations dominate, but with strategic
commerce both welfare and security considerations are important. Because high
amounts of strategic trade with a likely adversary include a political cost, there is
an incentive for conﬂict not present when the trading partner shares similar
foreign policy views. Speciﬁcally, the comparative static derivations produce the
following hypotheses:

Hypothesis 1: Dyads with high amounts of non-strategic trade are less likely to
experience conﬂict than dyads with smaller amounts of non-strategic trade.

Hypothesis 2: Dyads with close political afﬁnity and high amounts of strategic
trade are less likely to experience conﬂict than dyads lacking political afﬁnity with
high amounts of strategic trade.
Hypothesis 3: Dyads with high amounts of foreign direct investment are less
likely to experience conﬂict than dyads with smaller amounts of foreign direct
investment.

To summarize, the above premises lead to the following expectations on
economic interdependence and military conﬂict. First, non-strategic trade
interdependence should decrease the likelihood of conﬂict in a dyad. Non-

strategic trade reduces the likelihood of conﬂict by increasing the cost of conﬂict.

The principle of comparative advantage informs that trade increases a state’s

30

wealth, so trade is economically beneﬁcial. However, ﬂows of goods between
states are inﬂuenced by a variety of factors, call these transaction costs, that
inﬂuence the potential beneﬁts of commercial exchange. Governments have an
important inﬂuence over transaction costs. Militarized conﬂict, which is a
government policy, increases the transaction costs of conducting commercial
transactions, and thereby decreases the ﬂow of goods; thus, states with the most
non-strategic trade have the most to lose from conﬂict. If a sharp reduction of
trade occurs, then a political leader’s hold on power is weakened by the loss of
state wealth. As a result, states with high amounts of non-strategic trade will aim
to avoid costly conﬂict.

The story is different with strategic commerce. Trade in strategic
commodities enhances the likelihood of conﬂict between states. Strategic and
non-strategic trade interdependence have different effects because the question
of whether or not trade interdependence promotes peace is tied to the question
of whether or not conquest pays. Based on the premises above, it is clear that
the expected utility of conflict varies based on the amount of trade, the type of
trade, and the political affinity between the trading states. As discussed earlier,
the disruption of non-strategic trade imposes a cost, thus lowering the utility of
conﬂict and reducing the likelihood of conﬂict in a dyad. A disruption of strategic
trade also imposes an economic cost, but this economic cost may be outweighed
by political beneﬁts. The political beneﬁts of conﬂict with a strategic trading
partner may outweigh the economic costs from the disruption of trade when the

two states have dissimilar foreign policy preferences. Dissimilar foreign policy

31

preferences generate a potential security beneﬁt that offsets the economic loss of
trade. lmportantly, the security beneﬁt applies to both sides. The beneﬁt for the
importing state is to reduce the uncertainty of supply of the strategic
commodities. If conﬂict creates a change in the political relationship either
through conquest or a change in the exporting state’s policies that reduces the
importing state’s vulnerability to disruption of the strategic goods, then the utility
of conﬂict will be positive. Similarly, an exporting state may also desire conﬂict
with a trading partner. While the exporting state beneﬁts from the trade, it needs
to bear in mind security externalities. If its trading partner has very different
foreign policy preferences and it diverts the gains from trade to its military, then
the exporting state may also want to curtail trade. In other words, the conﬂict
beneﬁt for the exporting state is to keep a potential adversary weak.

Finally, I expect that foreign direct investment promotes peace between
nations. Foreign direct investment is a source for peace between nations for two
reasons. First, foreign direct investment brings capital, financial and human, into
a country, so like trade it enhances a state’s economy and generates wealth.
Since conﬂict reduces foreign direct investment, it will rarely pay, in terms of
enhancing a leader’s hold on power, to disrupt investment. Second, foreign
direct investment promotes peace by reducing the beneﬁts of conquest.
Conquest is a method of extracting resources from a territory. But, foreign direct
investment also permits an extraction of resources, and it enables the local
population to beneﬁt. In the limit, foreign direct investment eliminates any

additional beneﬁts that conquest may provide.

32

Research Design

The primary focus of this research is to investigate the relationship
between economic interdependence and militarized conﬂict between states. To
analyze the propositions of the analytical model, I specify the following

regression model.

MID Onset = a + 81 Non-Strategic Trade + (32 Strategic Trade + [33 FDI + [34
Political Afﬁnity + 85 Strategic Trade*PoIitical Afﬁnity + [36 Democracy + 87

Power Parity + 88 Allies + 89 Distance + 810 Contiguity + (311 Major Powers + a

This general regression model includes the central theoretical variables (trade
and FDI), a number of control variables, and a dependent variable that taps into
international conﬂict. Each variable will be described in turn. (See Appendix B

for a summary of the variables, operationalizations, and data sources.)

Population of Cases and Dependent Variable

While most research examining the relationship between interdependence
and conﬂict examines politically relevant dyads, I choose to examine all dyads.
Politically relevant dyads are dyads in which both states are contiguous or when
one state in the dyad is a major power. Maoz and Russett (1993) argue that the
set of politically relevant dyads is an appropriate baseline for empirical analyses
as this is the set most likely to experience conﬂict. Non-politically relevant dyads,
such as Belize and Sri Lanka, have little opportunity for conﬂict; therefore, they

should excluded. Lemke and Reed (2001) point out that an examination of

33

 

politically relevant dyads may enhance measurement error. After a thorough
analysis of the MID dataset, they conclude that “relevant dyad status is an
imperfect indicator of the opportunity for conﬂict. To the extent that the
opportunity for conﬂict is an explicit or implicit part of our investigations, analysis
of relevant dyads introduces measurement error” (Lemke and Reed, 2001: 132).7
Further, in the period under analysis advances in transportation make it very
easy for non-politically relevant states to trade with each other. Therefore, to
reduce problems of selection bias and measurement error, I analyze all dyads.

The dependent variable is whether or not a dyad experienced a militarized
interstate dispute (MID) in a particular year,8 as such the unit of analysis is the
dyad year. I drop from the analysis dyad-years in which a dispute is ongoing.
Although Oneal and Russett (1999) have argued that the decision to continue a
dispute is similar to the decision to begin a dispute, it is likely these two decisions
are different. Decisions to continue a dispute must weigh sunk costs and the
state of the campaign, while decisions to begin a dispute face greater
uncertainty. 1 also drop from the analysis “joiner” states. “Joiners” are states
drawn into a dispute either through alliance ties or as in an effort to aid one side
of the dispute.

Data limitations on commodity trade and investment data restrict the
empirical domain to the period 1970-1992. The above decision rules produce a
dataset with 291,541 observations. Because of missing data, however, the

empirical analyses only include 104,052 observations when the trade variables

 

7 They ﬁnd that about 10% of all disputes occur in non-politically relevant dyads.
8 See Jones, Bremer, and Singer (1996) for a discussion of the MID data set.

34

are included and 64,922 observations when foreign direct investment is included.
While Oneal and Russett (1997, 1999a, 1999b) assume dyads with missing trade
data did not have any trade, I do not set values of missing trade or foreign direct
investment equal to zero. When missing values are set equal to zero (analyses
not reported here), the interdependence variables are always statistically

signiﬁcant.

Independent Variables

Unlike previous research, I emphasize the importance of distinguishing
between strategic and non-strategic commodities. However, the distinction
between strategic and non-strategic goods is not entirely clear. Arguing from a
comparative advantage perspective, David Baldwin says that strategic goods are
“anything that is needed to pursue a given strategy and that is relatively
inefﬁcient to produce at home” (Baldwin, 1985: 215). In other words, all trade is
strategic trade. Although all trade is beneﬁcial to a nation’s economy, some
commodities are more important than other commodities in contributing to a
nation’s security and welfare. Forland argues that “it must be emphasized that
not every item is strategic even if it is produced inefﬁciently: it has to be
supportive of the pursued strategy” (Forland, 1991: 197). Thus, NATO’s
Coordinating Committee (COCOM) on exports to the Soviet Union did not include
Coca-Cola as a strategic good, though it was inefﬁcient to produce in the Soviet
Union. Citing a Truman Administration report of 1948, Forland notes that the

West’s export control lists included goods of both direct and indirect military

35

signiﬁcance, and “commodities “the denial of which would affect strategic sectors
of the economy of the Soviet bloc’ or ‘of considerable importance to the industrial
potential of Eastern European countries” (Forland, 1991: 198). In general, then,
strategic goods are products that significantly affect a nation’s military power or
economic health.

Based on Forland, one may classify food, minerals, iron and steel, basic
manufactures, and fuels as strategic commodities. Each of these products
signiﬁcantly affects a nation’s economy, and in particular its military strength. An
army cannot ﬁght effectively without food, fuel, small- and heavy-arms. Reuvany
and Kang (1998) conduct a Granger causality analysis of disaggregated trade
and conﬂict, and ﬁnd that conﬂict between states signiﬁcantly inﬂuences trade in
minerals, fuels, iron and steel, and basic manufactures. They conclude that
“such goods may now be viewed as strategic” (1998: 597). Based on their
ﬁndings, they also argue that trade in food, beverages, tobacco, machines, and
transport equipment may be considered “less strategic” (1998: 597). It is likely
that these latter goods are “less strategic” because there are more substitutes for
them. Because of substitutes, for instance, it is very difﬁcult to starve a nation
into submission. While this should not be considered a deﬁnitive statement on
what types of goods are strategic and non-strategic, it does provide theoretical
and empirical support for a particular classiﬁcation.

To measure strategic and non-strategic trade, it is ﬁrst necessary to
identify the value of trade in different commodity groups. Using data from the

National Bureau for Economic Research (NBER), l categorize all dyadic trade

36

into their single digit SITC categories.9 These categories include minerals, fuels,
basic manufacturers, etc (see Table 1 for a complete listing of the categories).
Based on the arguments above, I operationalize strategic trade as trade in SITC
categories 2, 3, 6, and 8. Non-strategic trade includes all other SITC categories.
NBER commodity trade covers the period 1970-1992.

Large values of trade may or may not be important to a nation’s economy,
depending on the size of the economy. Therefore, I measure both strategic and
non-strategic trade relative to a state’s gross domestic product.10 Then, I follow
Oneal and Russett’s weak link argument and only include in the regression
model the lower of the two trade-to-GDP ratios.

Militarized conﬂict also reduces F DI, thus increasing the costs of conﬂict.
In turn, high levels of F Dl should deter nations from ﬁghting. Dyadic foreign
direct investment data is unavailable for a large number of countries. As a
substitute, I include a variable that measures a state’s overall foreign direct
investment relative to its gross domestic product. I also follow the weak link
procedure for this variable. FDI data comes from the World Development
Indicators produced by the World Bank. This data is available for the period

1 970-1 992.

 

9 SITC stands for Standard lntemational Trade Classiﬁcation, and is the United Nations system for
classifying all trade between nations. SITC codes begin with ten general categories and become more
speciﬁc. For instance, SITC code 0 represents trade in food and live animals. SITC code 02 more
speciﬁcally identiﬁes trade in dairy products and birds’ eggs. SITC code 022 identiﬁes trade in milk and
cream and milk products other than butter or cheese. I group trade according to the 10 general SITC
categories.

'0 The data on Gross Domestic Product comes from Kristian Gleditsch (2000). Gleditsch’s dataset is based
on the Penn World Tables but is more comprehensive. In particular, it corrects a systematic problem in
previous datasets by including data for a number of developing and socialist countries. To the best of my
knowledge, this is the ﬁrst study on interdependence and conﬂict to use this dataset.

37

Next, I include a set of control variables that are also hypothesized to
affect the opportunity or willingness for conﬂict. Contiguity, distance, power
parity, and anarchy are the central factors affecting the opportunity for conﬂict.
Consider distance. As the distance separating two states increases, it is more
difﬁcult to mount a successful military campaign (Bueno de Mesquita, 1981;
Lemke, 1995). Larger distances make it more difﬁcult both to get troops to the
battleﬁeld and to re-supply troops. Separation, then, makes it more difﬁcult for
conﬂict to occur. Greater distance between countries may also have an indirect
effect of decreasing the willingness for conﬂict. If two states are far apart, they
may have minimal interaction, and thus little to ﬁght over. I operationalize
distance as the great circle distance between capital, or major, cities.11

Power transition theory has also identiﬁed power parity as an important
factor affecting the opportunity for conﬂict (Organski and Kugler, 1980; Kugler
and Lemke, 1996). As the imbalance of power in a dyad increases, the
probability of victory in a dispute for the weaker nation decreases. Unless the
stakes are very high, the weak state will conclude that conﬂict is not an option as
it has little probability of winning. Put differently, power parity provides an
opportunity for conﬂict because it increases the potential beneﬁts relative to the
costs. If there is not power parity, the costs of conﬂict for the weaker side greatly
outweigh the beneﬁts, thus, they are expected to give in to the demands of the
stronger state. If there is parity, then both sides have a chance to beneﬁt;
therefore, as power parity increases the potential benefits relative to the costs

increase and conﬂict becomes more likely. Insofar as this argument is accurate,

 

” Data for this variable comes from Bennett and Stam’s (2000) EUGene program, v 2.101.

38

then conﬂict should be more likely to occur under conditions of power parity then
under conditions of a power imbalance. Numerous empirical studies support this
hypothesis (see Kugler and Lemke, 1996; and Kugler and Lemke, 2000). I
operationalize power parity as the ratio of the larger state’s value on the
Correlates of War Composite Capabilities index over the smaller state’s value on
this index.12 Thus, greater values indicate less power parity.

In addition to opportunity, for nations to have a conﬂict, there must be
something to ﬁght over, that is there must be a willingness to ﬁght. Factors that
affect a nation’s willingness to ﬁght another nation include foreign and domestic
policy preference similarity, territorial issues, religious and ethnic similarity.
What’s important about the willingness condition is that it means that the
probability of conﬂict is not the same in all dyads (Gartzke, 1998, 2000; Werner,
2000; Gowa and Farber, 1997). Research on the democratic peace supports this
contention. Lemke and Reed (1996: 145) ﬁnd that “States satisﬁed with the
status quo desire no changes to the international order, and thus have nothing
over which to ﬁght.” In expected utility terms, as preference afﬁnity decreases,
the beneﬁts of conﬂict increase relative to the costs. Werner (2000), for instance,
argues that one of the primary beneﬁts that a state can derive from conﬂict is the
re-structuring of another state’s foreign policy preferences. By achieving greater
preference similarity, states enhance both their security and potential for
economic beneﬁts. Therefore, close political afﬁnity takes away one of the larger

potential beneﬁts from conﬂict.

 

'2 Data for this variable comes from Bennett and Stam’s EUGene program.

39

To measure preference afﬁnity, I include a measure of the similarity of
alliance ties in a dyad. Dyads with similar alliance portfolios are assumed to
have similar foreign policy preferences and thus less to ﬁght over. Perhaps more
importantly for the present research is that the effect of strategic trade is
conditional on the political afﬁnity in the dyad. As discussed in hypothesis two,
strategic trade between states with close afﬁnity is expected to be pacifying,
while strategic trade in dyads lacking afﬁnity is expected to contribute to conﬂict.
This measure of political afﬁnity, S, draws on the work of Bueno de Mesquita
(1981) and Signorino and Ritter (1999), and is based on the similarity of dyadic
alliance portfolios. States with similar alliance portfolios are assumed to have
similar foreign policy preferences. The variable ranges from zero to one, with
higher values representing greater afﬁnity.

Many studies have found that joint democracy reduces the likelihood of
conﬂict between states. Lemke and Reed (1996) and Werner (2000) suggest
this is the result of democracies having similar preferences. States with similar
domestic political institutions “are less likely to have certain types of
disagreements in the ﬁrst place” (Werner, 2000: 4). I use the democracy index in
the Polity lV (Marshall and Jaggers, 2000) dataset to operationalize democracy.
This index ranges from 0 to 10, with higher values indicating higher levels of
democracy. I also follow the weak link procedure for including this concept in the
regression model. Alliances also tap into preference afﬁnity; they indicate a
similar preference on at least one foreign policy issue. Accordingly, I include a

variable for alliances. Alliance is a dummy variable that equals one when a dyad

4O

shares a defense pact, non-aggression pact, or entente, zero othenlvise. Alliance
data comes from the COW project via Bennett and Stam’s EUGENE program.

Finally, I include a variable measuring whether or not a Major Power is a
member of the dyad. Major powers have both greater capabilities and, typically,
greater commitments around the world. As a result, they engage in more
militarized conﬂicts than non-major powers. I use the Correlates of War dataset
to identify major powers. For this period, the major powers are the United States,
the Soviet Union/Russia, France, Britain, China, and after 1990 Germany and
Japan. If at least one of these states is in a dyad, the variable takes on a value
of 1, 0 otherwise.

Statistically, I analyze the above regression model using a quasi-likelihood
method, speciﬁcally a general estimating equation (GEE).13 GEE models are
especially useful for estimating time-series, cross-sectional data as they allow
one to specify a variety of within group correlation structures for the panels.

Beck, Katz, and Tucker (1998) suggest that observations in time-series cross-
sectional data are likely to exhibit temporal dependence, and that failure to
control for temporal dependence may lead to inaccurate inferences. As a
remedy, they propose the use of temporal dummy variables. Although this ﬁx is
better than ignoring the issue, a better alternative is to directly incorporate one’s
“knowledge regarding within-unit interdependence through speciﬁcation of the

working correlation matrix” (Zorn, 2001: 474). To this end, I run a GEE model

 

'3 See Liang and Zeger (1986) and Zorn (200 I) for more technical discussions of general estimating
equations.

41

using a logistic link function and specifying an AR(1) correlation structure.14 In
addition, I employ robust standard errors to control for cross-sectional
heterogeneity in the data.

The GEE approach is also useful when the substantive focus is on
“making comparisons across groups” (Zorn, 2001: 475). Unlike a regular logistic
model, GEE models calculate population averages. As the goal of this research
is to understand whether or not dyads with high levels of economic
interdependence are more or less likely to engage in militarized conﬂict as
opposed to explaining the likelihood of a particular dyad for engaging in conflict,

a GEE model is the appropriate choice.

Empirical Results

To empirically evaluate the hypotheses from the analytical model and
contending arguments, I present results from ﬁve empirical models. The ﬁrst
model examines total dyadic trade and serves as a baseline for comparisons with
prior research. Models two and three evaluate the hypotheses on non-strategic
trade, strategic trade, and FDI. Models four and ﬁve examine the effects of
economic interdependence on an alternative operationalization of the dependent
variable. The empirical analysis provides strong support for the hypotheses.

I ﬁrst run an analysis using aggregate dyadic trade for all dyads over the
period 1970-1992. This is the operationalization of economic interdependence

used by Oneal and Russett (1997, 1999a, 1999b) and Polachek (1980). As

 

'4 Using a similar data set Oneal and Russett ( 199%) and Zorn (2001) also specify an AR(1) correlation
structure.

42

such, it serves as a baseline model for comparing this research to previous
research. The results of Model 1 are consistent with previous studies examining
dyadic trade interdependence (Oneal and Russett, 1997, 1999a, 1999b). Dyadic
trade interdependence reduces the likelihood of a militarized interstate dispute.
The trade interdependence variable, however, is not statistically signiﬁcant at
conventional levels (p-value = 1.5). This is probably a result of the particular time
period under investigation. The control variables also perform as expected.
Democratic institutions, power preponderance, and distance all reduce the
likelihood of a dyad having a militarized dispute, while contiguous and major
power dyads are more likely to experience militarized disputes. Equally
important, the constant term is statistically signiﬁcant and negative, indicating
most dyads do not experience militarized disputes. The allies variable, however,
is not statistically significant. Further exploration of the data suggests this is
likely the result of a number of militarized disputes between the United States
and a variety of Latin American countries with whom the United States was
allied. Overall, the results of Model 1 suggest no major structural changes
occurred after 1970 to make the 1970 to 1992 period vastly different from the
1950 to 1992 period that Oneal and Russett examine. This should enhance our
conﬁdence in the generalizability of the results.

The variables in a logistic regression model have a non-linear relationship
with the dependent variable; as a result, it takes some care in interpreting the
substantive impact of the variables. To provide a sense of the substantive

impact of the central variables, I calculated ﬁrst differences. A ﬁrst difference

43

expresses the change in the probability of event occurrence between two values
of an independent variable, while holding all other variables constant. For
instance, after calculating a baseline model where the interval level variables are
placed at their mean value and the dichotomous variables are held at zero, I
recalculated the probability of event occurrence by adjusting the interdependence
and, for comparison purposes, the democracy and power parity variables one
standard deviation. The difference between these probabilities is one measure
of the substantive influence of a variable. The third column of Model 1 shows
that a one standard deviation increase in total trade interdependence leads to
nearly a 16% reduction in the likelihood of the onset of a militarized interstate
dispute. This is nearly the same effect as increasing the democratic
characteristics of the least democratic state in the dyad, which reduces the
likelihood of conﬂict by about 22%.

Next, I disaggregate trade into strategic and non-strategic commodities to
examine their independent effects on the outbreak of militarized disputes. Recall
the reason for disaggregating trade. Guided by realist arguments, I expect that
different types of trade have different effects on the likelihood of conﬂict between
nations. On the other hand, liberal arguments expect all types of trade to exert a
pacifying effect. Disaggregating trade, then, affords a critical test between these
competing arguments concerning economic interdependence.

The results from Model 2 indicate non-strategic trade interdependence
exercises a negative inﬂuence on the occurrence of militarized interstate

disputes. This effect, however, is not statistically signiﬁcant. Strategic trade

44

interdependence, however, exercises a statistically signiﬁcant and positive
inﬂuence on the occurrence of MIDs! Contrary to liberal arguments, trade does
not always lead to peace. Greater strategic trade interdependence increases the
likelihood of conﬂict between nations. As expected, then, different types of trade
differentially inﬂuence the likelihood of conﬂict between states.

While non-strategic trade interdependence is not statistically signiﬁcant,
it’s substantive inﬂuence is meaningful. A one standard deviation increase in
non-strategic trade interdependence reduces the likelihood of conﬂict by about
15%. Similarly, a one standard deviation increase in strategic trade
interdependence increases the likelihood of a MID by about 15%. An equally
important ﬁnding concerns the inﬂuence of foreign direct investment. As
expected, higher levels of FDI make conﬂict less likely to occur, with a one
standard deviation increase in foreign investment reducing the likelihood of a
MID by about 10%. The rest of the variables mirror their behavior in Model 1.

While the results of Model 2 indicate that different types of trade have
different effects on conﬂict onset, Model 2 does not fully evaluate the arguments
presented earlier. The argument advanced here is that the effect of strategic
trade interdependence is conditional on the political afﬁnity in the dyad. Model 3
offers an evaluation of this argument.

As anticipated, strategic trade between states with similar foreign policy
preferences reduces the likelihood of militarized conﬂict. Indeed, an increase in
strategic trade and afﬁnity reduces the likelihood of conﬂict onset by about 15%

(see column 3 of Model 3). On the other hand, strategic trade between states

45

lacking afﬁnity increases the likelihood of conﬂict. This effect is especially strong.
In dyads with different foreign policy preferences, a one standard deviation
increase in strategic trade interdependence increases the likelihood of conﬂict by
46%. These results support hypothesis two. Further, in this more fully specified
model foreign direct investment is still statistically signiﬁcant and reduces the
likelihood of conﬂict between states. As before, non-strategic trade
interdependence is statistically insigniﬁcant. Non-strategic trade
interdependence, however, continues to exert a substantively meaningful
inﬂuence on conﬂict onset. A one standard deviation increase in non-strategic
trade interdependence reduces the likelihood of conﬂict by about 12%. In Model
3, the democracy variable is still not statistically signiﬁcant at conventional levels,
though as indicated in column 3 an increase in democracy does lead to a

substantial reduction in the likelihood of conﬂict onset.

Robustness and Sensitivity

To enhance conﬁdence in the ﬁndings, I conducted a number of
robustness checks. The robustness analysis consists primarily in analyzing
models using alternative operationalizations of the central variables. First,
instead of the weak link assumption, l constructed joint measures of both
strategic and non-strategic trade. This measure directly incorporates the trade
dependence of both nations. For each category of trade, I multiplied state A’s
trade dependence by state B’s trade dependence. Then, I took the square root

of this value and divided by the largest value of trade dependence so that the

46

resulting score ranges from 0 to 1. The results remain the same. As the
economic importance of strategic trade increases in the dyad, conﬂict becomes
less likely. Similarly, the greater the amount of non-strategic trade in a dyad, the
less likely the dyad is to experience militarized conﬂict. Foreign direct investment
also remains statistically signiﬁcant and negative. F DI is still operationalized via
the weak link procedure as dyadic investment data is unavailable.

Next, I operationalized strategic trade only as trade in minerals and fuels,
SITC category 3 goods. The idea behind this measure is that strategic goods
must not simply be valuable, but capturable. Now, of course, any product, once
produced, is appropriable, yet unless a single or small number of units of a
product is enough to turn the tide in a military campaign it is of little use that the
end product can be appropriated. For example, appropriating a few tanks is of
little use to a country engaged in war. Yet if a nation is able to appropriate the
technology and parts to produce tanks, then conquest may turn the tide of the
war. In this sense, factories may be of limited strategic value as they are of little
use without the knowledge and skills on how to run them. Commodities that are
strategic in this sense are primary products like oil and other natural resources.

Commodities that are easily rendered useless like factories or service
industries are much less appealing to foreign conquest than primary resources.
If an opponent is willing, factories can rather easily be blown up or have their
production limited due to mechanical breakdowns. Natural resources may also
be rendered useless, though it is more difﬁcult to do so. Iraq’s “scorched earth”

withdrawal from Kuwait in 1991, for instance, only temporarily limited the value of

47

Kuwait’s oil wells. On the other hand, anecdotal evidence suggests that nations
go to war to acquire natural resources. Germany and Japan, for instance, both
pursued militarily aggressive foreign policies to secure a supply of oil for their
economies.

The empirical results remain the same when strategic trade is
operationalized as only including minerals and fuels. Strategic trade between
dyads lacking political afﬁnity increases the likelihood of conﬂict between these
trading states, while strategic trade between states with similar political
preferences reduces the likelihood of conﬂict. Non-strategic trade
interdependence is still insigniﬁcant. Finally, FDI is still statisitically signiﬁcant
and negative.

In a ﬁnal set of robustness analyses, I examine the effects of economic
interdependence on an alternative dependent variable. Rather than analyzing all
Mle, Model 4 only analyzes MIDs with at least one casualty. Analyzing fatal
MIDs tests whether or not economic interdependence only affects the likelihood
of low level disputes, or if it also inﬂuences more serious disputes.

The results from Model 4 are similar to the earlier ﬁndings. Non-strategic
trade interdependence reduces the likelihood of a dyad experiencing a fatal
militarized interstate dispute.15 While the non-strategic trade interdependence
variable is not statistically signiﬁcant, it’s substantive inﬂuence is impressive. A

one standard deviation increase in non-strategic trade interdependence reduces

 

'5 If strategic trade is used in place of total dyadic trade, then trade interdependence is not statistically
signiﬁcant at the .05 level, though it is at the .10 level. Further, if missing values of strategic trade are set
equal to zero, as is the procedure followed by Oneal and Russett, then strategic trade is signiﬁcant and
negatively related to dispute onset.

48

the likelihood of a fatal militarized interstate dispute by about 37%; this is almost
twice the impact of increasing the level of democracy in the least democratic
state in the dyad. More importantly, the strategic trade interdependence and
strategic trade-afﬁnity interaction variables are both statistically signiﬁcant and
their signs are in the hypothesized direction. An increase in strategic trade
interdependence increases the likelihood of a fatal dispute by 126%! This
suggests that strategic commodities are associated with efforts to conquer
territory. As before, strategic trade does not always increase the likelihood of a
dispute. Dyads with close political afﬁnity and high levels of strategic trade are
less likely to experience a fatal dispute.

Model 5 continues to examine the inﬂuence of the disaggregated trade
variables on the outbreak of fatal militarized disputes, but also incorporates
foreign direct investment. F DI is statistically signiﬁcant and negative, meaning
higher levels of foreign direct investment reduce the likelihood that a fatal
militarized dispute will occur. When FDI is included, none of the trade
interdependence variables are statistically signiﬁcant, though the sign of each
variable is in the expected direction. Nevertheless, the trade interdependence
variables continue to exert a signiﬁcant substantive inﬂuence. In fact, non-
strategic trade, strategic trade, and the strategic trade-afﬁnity interaction

variables all inﬂuence the outbreak of a fatal dispute more than FDI.

49

 

Conclusion

In conclusion, this research presents an analytical model of economic
interdependence and interstate conﬂict. Unlike previous research on economic
interdependence, the model incorporates different types of trade and foreign
direct investment. The analytical model reveals that the relationship between
trade interdependence and peace is conditional on the type of goods traded and
the political relationship in the dyad. Speciﬁcally, non-strategic trade
interdependence is expected to contribute to peace between nations. Similarly,
strategic trade interdependence between states with similar foreign policy
orientations reduces the likelihood of conﬂict. Contrary to the liberal argument on
economic interdependence, however, I argue that trade interdependence does
not always pacify. Strategic trade between states with dissimilar foreign policy
preferences contributes to conﬂict. In addition, this research takes the liberal
argument a step further by demonstrating the relationship between foreign direct
investment and international conflict. Models not including FDI do not fully
analyze the central concept of economic interdependence.

The second part of this research empirically examines hypotheses from
both the analytical model and competing arguments. Consistent with the
theoretical model, I ﬁnd that all trade is not the same and that the political
relationship in a dyad inﬂuences the effect of trade. Non-strategic trade
interdependence contributes to peace between states, but strategic trade

interdependence enhances the likelihood of conﬂict between states. The ﬁnding

50

that strategic trade between states lacking afﬁnity contributes to conﬂict is
especially important as it casts doubt on the standard liberal argument.

Moreover, I ﬁnd that foreign direct investment is a signiﬁcant inﬂuence on
the outbreak of militarized disputes between states. In line with the analytical
model, greater levels of foreign direct investment reduce the likelihood of a
dispute occurring. This ﬁnding is especially important as it answers a central
criticism of interdependence theories. As is well known, the major European
powers had a high level of trade interdependence during the ﬁrst decade and a
half of the twentieth century, yet this trade interdependence did not prevent
World War I from occurring. These same European states, however, did not
have high levels of foreign direct investment either with each other or in general;
thus, the nature of the economic interdependence was both one-dimensional and
limited. These same states now exhibit greater levels of foreign direct
investment. What has changed then between 1914 and the present is not only
an increase in the number of democratic regimes in Europe, but also an increase
in economic interdependence chieﬂy through foreign direct investment. Finally, it
is important to note that the empirical results are robust to a number of
alternative speciﬁcations. The ability of the argument to withstand a wide array
of tests should enhance our confidence in the ﬁndings.

The policy implications of this research are subtle. States interested in
promoting peace and prosperity should encourage greater economic
interdependence between allies. Economic interdependence fosters peace

because it enhances prosperity, which is a central influence on the ability of a

51

political leader to remain in power. Economic interdependence also fosters
peace because it builds common interests. With interdependence, dyads have a
shared goal in assuring stability and the free ﬂow of goods. Economic
interdependence, however, is not a panacea. In fact, interdependence in
strategic goods may exacerbate tensions and contribute to conﬂict between
states. The reason for the difference between strategic and non-strategic trade
is that the latter enhances a state’s vulnerability, thus providing an additional
incentive for conﬂict. Political leaders, then, should pursue different trade
policies with different states.

Finally, this analysis suggests a number of avenues for future research.
Future research should examine the inﬂuence of different types of trade and FDI
at additional levels of analysis. In particular, classical arguments on
interdependence focused on the nation-state, yet little research has been
conducted on the relationship between interdependence and the proclivity of a
state to experience conﬂict. Plus, the more tests a theory can pass, especially at
different levels of analysis, the more conﬁdence we can have in the argument.
Future research should also focus on reﬁning the measures offered here and
applying these concepts in a more strategic context. For instance, a true dyadic
measure of foreign direct investment would permit a more accurate assessment
of the theory. In addition, future research should consider the effects of different
types of trade and foreign direct investment at different stages of conﬂict.
Perhaps some types of trade are more important than other types at inhibiting the

onset of conﬂict but not the escalation of a dispute. In examining both conﬂict

52

onset and escalation, we can more clearly distinguish between interest-based

interdependence theories and signaling-based interdependence theories.

53

Congratulations on reaching the half-way point in this dissertation.

©

54

Chapter 3

Economic Dependence and Militarized Interstate Dispute Initiation and
Targeting, 1970-1992

Does economic dependence promote peace? Realists argue either that
trade is “low politics” and does not greatly affect relations between states or that
it creates asymmetric gains and dependencies, which increase the likelihood of
conﬂict. In particular, nations economically dependent on importing strategic
resources are vulnerable and likely to choose a policy of military expansion to
alleviate this vulnerability. In contrast, liberals argue that trade creates both
domestic and international constraints on the use of force, making nations more
peaceful. In the trade promotes peace arguments, the causal mechanism takes
one of two forms. Some scholars suggest trade promotes peace because
commercial activity develops peaceful norms for dispute settlement. Others
argue that trade promotes peace by increasing the opportunity cost of conﬂict,
thus deterring nations from military expansion.

I argue that the economic dependence and conflict nexus is more complex
than previous arguments and research suggest. In particular, I address three
ﬂaws in the extant research. First, previous research treats economic
dependence only as trade. In contrast, I argue that modern international
economic dependence is characterized as much by foreign investment as it is by
the level of trade ties. Second, current research does not distinguish between
the type of commodities traded; trade in oil has the same inﬂuence as trade in

textiles. In contrast, I argue that it is necessary to differentiate states based on

55

the types of goods they import and export. Because earlier research did not
distinguish between the types of goods traded, the effect of trade was always the
same, either it increased or decreased the likelihood of conﬂict. Third, previous
research operationalizes conﬂict simply as participation in a militarized dispute. I
argue that participation is a ﬂawed operationalization of conﬂict when the unit of
analysis is the nation-state. Instead, it is necessary to distinguish between
initiators and targets of disputes.

After distinguishing between the types of goods traded and whether a
state is an initiator or target of a dispute, I make the following hypotheses. First,
trade in non-strategic commodities promotes peace. Imports of non-strategic
commodities enhance a state’s wealth and increase the costs of conﬂict, thus
making a state less likely to initiate conﬂict. Next, the effect of importing strategic
commodities depends on a state’s degree of militarization. Militarized states
importing strategic commodities have the ability to alleviate their economic
vulnerability, making them more likely to initiate conﬂict than other states.
Similarly, economically underdeveloped states exporting strategic commodities
are more likely to be the targets of militarized disputes, while strong states
exporting strategic commodities are less likely to be targets of militarized
disputes. Finally, I hypothesize that foreign direct investment promotes peace.
Foreign investment depends on a stable, secure environment; nations with high
levels of foreign investment avoid military conﬂicts in order to continue drawing in
investment. In addition, foreign investment decreases the beneﬁts of conquest.

With foreign investment, a nation is already extracting wealth from a territory, so

56

there is less of an incentive for conquest. In summary, I argue that the effect of
trade on conﬂict varies with the type of goods traded and is conditional on a
state’s military power and economic development.

The paper proceeds as follows. The ﬁrst section develops a modiﬁed
realist theory of economic dependence and militarized conﬂict. Where previous
research simply focuses on the beneﬁts of economic exchange, I argue that it is
necessary to view the beneﬁts of exchange in light of a state’s military power and
economic development. The next section tests hypotheses from the theory
against the empirical record using an event-count speciﬁcation. This empirical
analysis of economic dependence and conﬂict contributes to existing empirical
research in the “trade and conﬂict” research program with its explicit focus on the
monadic level of analysis and in its distinction between initiators and targets of
militarized disputes. The empirical record strongly supports the theoretical
argument. The effects of economic dependence vary with the type of good
traded and are conditional on a state’s military power and economic
development. Further, and as expected, foreign direct investment reduces the
likelihood of a state either initiating a dispute or being the target of a dispute. The

paper concludes with a discussion of the implications of this research.

Commercial Dependence and Interstate Conﬂict

Following WWII, reductions in tariff rates helped facilitate large absolute

welfare gains by increasing both global trade and foreign direct investment. The

57

volume of global trade in goods and services is today over 16 times larger than in
1950, “while the world’s total output is only ﬁve-and-a-half times as big. The ratio
of world exports to GDP has climbed from 7% to 15%” (Economist, November 8,
1997, 85). Equally important, “foreign investment in the past three decades has
risen faster than trade and world output” (Economist, June 18, 1998). Recently,
foreign investment has become the most important economic linkage between
states. “Between 1986 and 1990 FDI outﬂows grew at an average annual rate of
28 percent and cumulative FDI stocks at a rate of 20 percent a year compared
with a growth rate of world exports of 14 percent” (Dicken 1998: 42).

The above accurately describes economic integration at a global level, yet
it does not discuss the causes or effects of this phenomenon. While it is
important to understand the causes of globalization, the rest of this essay
investigates the effects of this phenomenon on world politics. Previous research
on economic dependence and conﬂict emphasizes four factors: costs, norms,
vulnerability, and signaling.

Economic dependence increases the costs of conﬂict. As a result,
Polachek (1980), Polachek, Robst, and Chang (1999), and Oneal and Russett
(1997, 1999a, 1999b) argue that economic dependence promotes peace.

Military conﬂict clearly reduces the willingness of states to engage in commercial
activity. As a result, “countries with the greatest gains from trade face the
highest costs of potentially lost trade and hence engage in the least conﬂict and
the most cooperation (Polachek, Robst, Chang, 1999: 405).” Oneal and Russett

make a similar argument: “Fearful of the domestic political consequences of

58

losing the beneﬁts of trade, policymakers avoid the use of force against states
with which they engage in economically important trade” (1999b: 4-5). While the
arguments of Polachek and his colleagues and Oneal and Russett emphasize
dyadic trade, they are compatible with arguments linking commerce and peace at
both a monadic and systemic level as well.

Domke advances a similar theoretical argument as Polachek (1980) and
Oneal and Russett (1999a, 1999b), though he emphasizes the nation-state as
the appropriate referent. “Foreign trade,” he writes, “produces a constraint on
decisions for war through the growth of international, domestic, and
governmental forces with a stake in open and unfettered foreign dealings [thus]
the relevant indicator of foreign trade would measure an economy’s involvement
in trade” (Domke, 1988: 118). In a system level investigation of the relationship
between trade and conﬂict, Mansﬁeld (1994) ﬁnds an inverse relationship
between trade and war. As global trade increases, nations have less incentive to
pursue policies of conquest. In all of these arguments, trade promotes peace
because it promotes wealth and increases the costs of conﬂict.

A second way in which global commerce constrains decision-makers from
pursuing militarily aggressive foreign policies is by promoting peaceful interests
within a state. In liberal systems the use of force is especially costly because
commercial transactions develop interest groups to protect and enhance their
economic interests; these interest groups inevitably pressure policymakers to
avoid breaking political and economic relationships (Rosecrance, 1986). Trade,

then, reduces the likelihood of interstate conﬂict by helping to form interest

59

groups that have a ﬁnancial stake in maintaining a stable and cooperative
political environment. Rational elites undoubtedly recognize the potential
domestic costs for ignoring these important constituency groups by destabilizing
commercial ties.

Peaceful norms are a second factor emphasized in the trade and conﬂict
research program. In addition to altering the costs associated with conﬂict,
economic integration fosters political and social ties between nations. These ties
promote the development of peaceful ways to address problems. As part of the
growth in global economic dependence, formal and informal rules and
procedures have developed to help guide states in their decision making and
political bargaining (Keohane and Nye 1989). Deutsch et al. (1957) offered a
similar insight. Complex interdependence helps mitigate violent conﬂict by
fostering a sense of community or shared identity. Indeed, economic
dependence "cements bonds of friendship” by enabling governments to more
effectively discern mutual interests (Barbieri and Schneider 1999, 387). The end
result, according to Buzan, Little, and Jones (1993), may be an attenuation in the
structural effects of anarchy.

While the costs of disrupted trade and norms associated with conﬂict
resolution suggest trade dependence promotes peace, a third factor,
vulnerability, emphasized in previous research exercises a countervailing force.
Despite obvious absolute gains from increased global economic activity,
commercial dependence may bode ill for peace between nations if it enhances a

state’s vulnerability. This perspective emphasizes the anarchic nature of the

60

lntemational system and goal of state security. Because the international system
is anarchic and self-help (Waltz, 1979), states often emphasize relative gains.
While all states may gain in the long-run from an open international economic
system, some states will certainly, if only for a short period of time, gain more
than others. These relative gains produce a security externality (Gowa, 1989). A
security externality occurs when a state uses the gains from trade to build its
military rather than for more peaceful purposes. In other words, since trade
produces gains, “trade increases the potential military power of any country that
engages in it” (Hirschman in Baldwin, 1985: 211). As a result, short-temr
asymmetries in wealth accumulation compel states to distrust the positive effects
produced by expanding trade and ﬁnancial ties. In short, trade is not a panacea
for the world’s ills; in fact, it may increase a nation’s vulnerability and increase the
probability of conﬂict (Waltz, 1979).

A fourth factor emphasized in the trade and conﬂict research program is
the informational potential of trade. In the informational perspective, uncertainty
is central to conﬂict, a factor directly addressed by commercial exchange. The
result of international trade, according to Hegre (2000, 5), is "improved
communication between the inhabitants of the trading states. This reduces the
chances of misunderstanding and helps to build institutions for the peaceful
resolution of conﬂic ." Therefore, if uncertainty and deception contribute to the
collapse of negotiations, then economic dependence can help state leaders
anticipate the intentions and reactions of adversaries, reduce blufﬁng, and

prevent bargaining breakdowns. Similarly, Keohane (1984, 245) alleges that

61

international systems with institutional structures that produce valuable
information are less conflict prone than international systems where these
information providing institutions are absent. Trade and foreign investment, then,
may function as a medium for information exchange, allowing states to
demonstrate resolve without resorting to military violence (Gartzke, Li, and
Boehmer, 2001).

Morrow (1999), however, questions the informational content of trade.
The informational argument says trade can serve as a measure of a state's
resolve for war. That is, "the expectation is that resolve declines as trade
increases, making war less attractive" (485). Yet, if we consider a dyadic
measure of resolve to be zero sum, then a decrease in one side's willingness to
ﬁght necessarily leads to an increase in the other side's. The result, Morrow
(1999, 488) writes, is that "trade ﬂows. ..have an indeterminate effect on the
initiation and escalation of international conﬂict."

Overall, from the primary theoretical arguments, the relationship between
economic dependence and interstate conﬂict is unclear. Empirical research
assessing the effect of trade on conflict is also ambiguous. Oneal and Russett
(1997, 1999a, 1999b), for instance, find that democratic political institutions and
economic interdependence constrain the use of force between nations. “Thus,”
they write, "interdependence and democracy contribute to what we have called
the “liberal peace” (1999: 2). Dorussen (1999) also ﬁnds that trade tends to

reduce interstate conflict, although the size of the state system additionally plays

62

a role, with increases in the number of nations reducing the pacifying effects of
trade ties.

Other research has cast doubts on the paciﬁc beneﬁts of commercial
engagement. Barbieri ﬁnds that “extensive economic interdependence increases
the likelihood that dyads engage in militarized dispute” (1996: 42). Beck, Katz,
and Tucker (1998) argue that in large cross section time series analyses one
needs to control for temporal dependence. When they replicate Oneal and
Russett’s 1997 work, while controlling for temporal dependence between
observations, they ﬁnd that trade is no longer associated with peace. For similar
reasons, Domke’s (1988) empirical ﬁndings on the paciﬁc beneﬁts of economic
dependence may be considered inconclusive. In addition to not controlling for
temporal dependence, his study does not include control variables. Similarly,
Benoit (1996) analyzes the liberal peace at the monadic level. He ﬁnds a
positive relationship between democracy and peace, but no relationship between
economic dependence and peace.

It is likely that the inconsistent empirical ﬁndings, at least with regard to
the nation-state, which is the focus of this research, result from an overly general
theoretical argument, a broad operationalization of the central theoretical
concept, and the omission of important control variables. Further and unlike
previous research, I argue that studies of a nation-state’s involvement in military
conﬂict need to distinguish between initiated and targeted conﬂict involvement as

opposed simply being involved in a conﬂict.

63

A Modiﬁed Realist Theory

I contend that the effects of economic dependence must be viewed in a
cost-beneﬁt framework that emphasizes differences across commodities. First,
all commodities are not the same. The vulnerability created from importing
strategic commodities such as oil is greater than the vulnerability created from
importing non-strategic commodities like textiles. Strategic commodities
generate greater vulnerability because they have fewer substitutes and exercise
multiplier effects on a state’s economy and security. Next, it is important to view
the effects of economic dependence in a full framework that incorporates both
the utility of an outcome and the probability of the outcome in order to compare
the utility of different circumstances. For instance, the utility of alleviating a
vulnerability may be the same for two states, but if one state has greater military
strength then the probability of being able to alleviate the vulnerability is different.
In turn, our expectations should be different.

Before analyzing the effects of an independent variable, such as trade or
investment, it is beneficial to specify the premises on which an argument is
based. I assume nation-states are unitary, rational actors. What is perhaps most
signiﬁcant about the assumptions of states as unitary actors and rationality is the
acceptance of them among both realist and liberal theorists.16 While multiple
actors inﬂuence a decision, on major foreign policy issues a single actor, the

leader of the state, ultimately makes the decision. Moreover, calling states

 

‘6 Some research in the realist tradition employing these assumptions include Krasner (1978), Waltz (1979),
and Gilpin (1981). Some research in the liberal tradition emphasizing these assumptions includes Polachek
(I980), Rosecrance (1986), and Oneal and Russett (1997, 1999a). For more thorough discussions of these
two assumptions see Bueno de Mesquita (1981) and Huth (1996).

64

unitary actors does not mean domestic politics are unimportant to foreign policy
decisions. Rather, a political leader’s calculus is still based, in part, on domestic
factors. Similarly, the assumption of rationality means state leaders have
preferences and they work to attain them. In more conceptual terms, rationality
implies that state leaders can connect their preferences in a transitive fashion.

Next, I assume states maximize a mix of security and wealth. Realist
scholars, for instance, posit that the anarchical nature of the international system
requires states to focus, ﬁrst and foremost, on security (Waltz, 1979). The
reason for focusing on security is to protect one’s sovereignty and secure one’s
hold on power. Bueno de Mesquita, Morrow, and Zorick (1997), for example,
have found that leaders who lose an international conﬂict are often removed from
ofﬁce. Thus, promoting national security is important if a leader is to retain
political ofﬁce. States focus on wealth for similar reasons. Wealth maximization
helps a leader remain in ofﬁce by providing the means for warding off internal
and external threats. In democracies, it is often said that people vote with their
pocketbook. In prosperous times a leader is more likely to be re-elected. In non-
democracies, wealth, manifested as economic develop, decreases dissatisfaction
with the state and the concomitant likelihood of revolution and civil war. Further,
wealth enables a state to build a strong military,17 thus decreasing the likelihood
of attack.

States pursue security and wealth via territorial acquisition and
commercial development. Unlike Rosecrance (1986), I suggest the two methods

are not mutually exclusive. In terms of wealth, territory is important primarily for

 

'7 Indeed, one of the most common measures of power is a state’s gross domestic product.

65

raw materials. All economies need raw materials to function. If a state does not
possess necessary raw materials, then it has to obtain them either through
exchange or conquest. Economic exchange is generally a more cost effective
method for acquiring raw materials than military conﬂict.

Economic integration produces more wealth than conquest as a result of
economic specialization and greater investment in the economy. Comparative
advantage leads states to specialize in goods that generate the highest proﬁts,
producing an efﬁcient and profitable business enterprise. Moreover, as states
develop economically, increasing returns to scale make intra-industry trade more
valuable.18 Because of increases in intra-industry trade since World War II,
industrialized countries have experienced a rate of trade growth twice that of their
GNP (Helpman and Krugman, 1985: 159). Plus, when a nation does not have to
invest in large amounts of military weapons for the purpose of conquest, it can
invest in physical capital and research and development; thus, making an
economy more productive (Rosecrance, 1986: 155-157).

All trade, however, is not the same. Resources are a central factor of
production; thus, “a nation’s economy is particularly dependent on imports of
goods for which demand is highly inelastic and domestic production is extremely
inefﬁcient, especially those that have multiplier effects on the whole economy”
(Liberman, 1996: 154). Because of their importance to a state’s economy,
resource starved states have an incentive to pursue policies of conquest.

Conquest pays. Conquest “can be proﬁtable, since, by territorial expansion, it

 

‘8 Helpman and Krugman (1985) note that for industrialized countries, most trade is intra-industry, while
developed and undeveloped dyads primarily conduct inter-sectoral trade.

66

can not only secure new markets, resources, and commercial supremacy, but
also contribute to a country’s industrialization and production potential.” (Rotte,
1997: 10). The expansionist policies of Germany and Japan in the ﬁrst half of the
twentieth century are consistent with this perspective.19

The differences in vulnerabilities across commodities lead to different
incentives for exchange and conﬂict depending on the type of good at issue.
States importing and exporting non-strategic commodities have no incentive to
disrupt their ﬂow as a disruption of trade is a cost. Similarly, a disruption of
strategic commerce is also a cost. However, reducing vulnerability of supply of
strategic goods is a beneﬁt. In other words, states dependent on importing
strategic goods are vulnerable to a disruption of supply, and such a disruption
may greatly impact their security and welfare. As such, states importing strategic
commodities have an incentive to either pursue conquest or at least engage in
military conﬂict as a means for ensuring the supply of these goods.

Next, international commercial transactions involve more than trade; they
also involve investment. Indeed, for Rosecrance (1999) foreign direct investment
(FDI) is the essence of virtualization, a phenomenon diminishing the importance
of national borders. “Virtualization is based on the growing importance of capital
and capital ﬂows (particularly foreign direct investment) in the world economy”

(Rosecrance, 1999: 36).

 

'9 Liberrnan (1993) also argues that conquest may not be as costly as Rosecrance (I986) maintains.
Liberrnan (I993: 138) notes that “modernization increases the efﬁciency of coercion because it centralizes
control over coercive resources, facilitates the quick deployment of this power over expansive regions, and
gives hostage societies more to lose from resistance." In contrast, Rosecrance (I986: 37) argues that “the
rise of mobilized populations, the spread of guerrilla insurgency, and the growing consciousness of ethnic
nationalism would make future conquest more difﬁcult than it was in the nineteenth century or even in
World War 11.”

67

F DI is also important as it distinguishes the current era of economic
activity from international economic dependence at the turn of the twentieth
century. “In World Wars I and II highly developed industrial nations battled over
the possession of land. They needed oil, coal, iron, and even sources of food”
(Rosecrance, 1999: 30). FDI, however, lowers the incentives for conquest as
investments “create, extend, or facilitate control over productive facilities in other
countries” (Kenen, 2000: 280). “In general, as a state is increasingly able to rely
on MNCs to secure needed external resources and supplies, the overall
willingness of that state to engage in conquest should decrease” (Brooks, 1999:
666). In the 1930s, for instance, Japanese foreign policy emphasized conquest
as a means for acquiring natural resources. Since the end of the Second World
War, however, Japanese F DI has “overwhelmingly” focused on natural resources

(Dicken, 1998: 55).

The Directionality of Conﬂict

Research on the democratic peace suggests that democratic states may
appear equally as conﬂict prone as non-democratic states as a result of being
frequently targeted by unconstrained autocratic regimes (see Rousseau, et al.
1993). If so, then democracy would be insigniﬁcant as an explanatory variable in
monadic analyses. According to Maoz and Abdolali (1989, 6-7), “politically free
states may be more likely targets of international violence than non-free states. A
political elite of a non-free state may calculate that its chances of getting its way

in a dispute are higher if it picks on a politically free state.” Similar arguments

68

apply to economic dependence. High levels of trade may reduce conﬂict
between nations, but do high levels of trade reduce a nation’s overall propensity
to initiate militarized conﬂict?

The directionality of conﬂict is important for drawing accurate inferences
regarding the determinants of foreign policy decision-making. Chan (1997: 68),
for example, has insisted that “even though the role of initiator of violence does
not necessarily mean the country in question is the aggressor in a particular
conﬂict, it is still the most important discriminating indicator for examining the
democratic peace proposition.” Fearon (1994), as well, has suggested that
initiation is a salient distinction to make when investigating conﬂict proneness. To
the extent we believe certain factors, like trade, investment, and democratic
regimes, inﬂuence the use of military force, then analyses need to pay more
attention to directionality.

States with higher audience costs are less likely to initiate militarized
disputes than states with lower audience costs. Audience costs are domestic
political costs (Fearon, 1994). As audience costs increase, a state is able to
more clearly signal its intentions and resolve, reducing the likelihood of a
militarized dispute due to uncertainty. Fearon argues that democracies face
higher audience costs for foreign policy failures because democratic leaders are
more accountable to their publics (1994). According to Partell and Palmer,
“Fearon’s hypothesis is monadic. The audience cost hypothesis should apply to
all high-audience-cost states, regardless of their opponent’s domestic political

structure” (1999: 390). The higher costs connected to foreign policy failure make

69

democratic states less likely to initiate militarized conﬂicts than non-democratic
states.

In the same way that democratic regimes deter a state from initiating
conﬂict, trade in non-strategic commodities and foreign investment also reduce
the likelihood of a state initiating a dispute. States with high levels of investment
and non-strategic trade are more likely to resolve conﬂicts of interest through
diplomatic bargaining because commercial interest groups, which suffer from
breakdowns in political relations, require government elites to avoid incurring the
economic costs that result from military confrontation. Greater commercial
dependence, then, also increases a state’s audience costs, making the state less
likely to initiate militarized conflicts.

A disruption of strategic trade also creates a cost, yet there is a greater
audience beneﬁt from alleviating, even attempting to alleviate, a strategic
vulnerability. The mechanics of this beneﬁt are apparent in George W. Bush’s
goal to open the Alaskan National Wildlife frontier to oil exploration. Even though
this new drilling is not expected to produce signiﬁcant amounts of oil, the public
supports the goal of reducing vulnerability.

Similar arguments explain which states are more likely to be targeted by
military aggression. As before, the greater a state’s level of non-strategic trade
and foreign investment, the greater are the costs from conﬂict. In turn, states
with the most non-strategic trade and investment have the most to lose from

conﬂict so they work harder to avoid conﬂict. The informational properties of

70

trade again complement the cost arguments. Since trade facilitates the sending
of signals, it reduces the likelihood of unintended conﬂict.

Strategic trade is again different. States exporting strategic commodities
have something other states want and may be able to obtain via conﬂict. If a
state exporting strategic commodities cannot defend itself, then it is more likely

than other states to be a target of military aggression.

Hypotheses

In summary form, the above premises lead to the following expectations.
Political leaders aim to retain ofﬁce, as such they compare the costs and beneﬁts
of choices. Conﬂict disrupts non-strategic trade, thereby increasing the costs of
conﬂict, and deterring states from initiating conﬂict. Moreover, non-strategic
trade does not promise any benefits from conquest, so states exporting non-
strategic goods are less likely than other states to be targets of military conﬂict.
On the other hand, states’ security motivation provides an incentive for a state to
alleviate strategic trade vulnerability. Conﬂict also disrupts strategic commerce,
creating a cost, but the beneﬁts of reducing strategic trade vulnerability outweigh
the costs of conﬂict if the probability of success is reasonable. In other words,
militarized states importing strategic commodities have an incentive to reduce
their vulnerability and the means to do so. Therefore, militarized states
dependent on importing strategic commodities are more likely than other states
to initiate militarized conﬂict. Similarly, weak states exporting strategic
commodities are a tempting target. Finally, military conﬂict also disrupts foreign

direct investment. Plus, states with large amounts of foreign direct investment

71

are not appealing targets because a potential attacker is already investing in the
potential target taking away an incentive for conﬂict, and even if a potential
attacker is not investing in the target, as a recipient of foreign direct investment
the potential target has wealthy allies. Thus, high levels of foreign direct
investment reduce the likelihood a state will initiate conﬂict and reduce the
likelihood a state will be a target of military aggression.

In summary form, we have the following hypotheses.
Hypothesis 1a: The more economically dependent a state is on non-strategic
trade, the less likely it is to initiate militarized conﬂict.

Hypothesis 1b: The more economically dependent a state is on non-strategic
trade, the less likely it is to be a target of militarized conﬂict.

Hypothesis 2a: Militarized states economically dependent on importing strategic
goods are more likely than other states to initiate militarized conﬂict.

Hypothesis 2b: Economically underdeveloped states with strategic commodities
are more likely to be a target of militarized conﬂict.

Hypothesis 3a: The more economically dependent a state is on foreign direct
investment, the less likely it is to initiate militarized conﬂict.

Hypothesis 3b: The more economically dependent a state is on foreign direct
investment, the less likely it is to be a target of militarized conﬂict.

Research Design

To properly test any argument, it is necessary to determine the process
that generates the observed data, which in this case is involvement in militarized
interstate disputes. The two dependent variables of interest are the number of
conﬂicts (Mle) initiated and the number of times a state was the target of

military aggression. Both dependent variables have important characteristics for

72

modeling purposes. First, they cannot be less than zero. Second, they must
always take on an integer value; you cannot be involved in a partial war or a
partial MID. Given the nature of the data generating process, non-normally
distributed errors and a dependent variable that is only positive, discrete and
unbounded, I employ an event count model.20

In modeling event counts, it is important to consider possible contagion
effects. International conﬂict, for instance, may beget a contagion effect where
the number of events at time t + 1 is inﬂuenced by the number of events at time t.
Further, international conﬂict is very unlikely to occur at a constant rate; a variety
of circumstances are likely to change the probability of militarized conﬂict over
time intervals. Thus, when the dependent variable is a MID, the data generating
process is likely to produce overdispersion in the data. Overdispersion occurs
when the variance of the expected events is greater than the expected value of
the events, V(Yi) > E(Yi). To address this potential statistical problem, I employ a
Negative Binomial model, which treats the rate of event occurrence, lambda, as a
random variable following a gamma distribution, thereby accounting for the
potential lack of independence between events. This model also adds an
additional variance parameter that allows for V(Y,) 2 E(Y,). It should be noted
that as this extra parameter, sigma squared, approaches one the negative
binomial approximates the Poisson.21

Although a Negative Binomial event count addresses overdispersion and

heterogeneity in the data, it may not address problems of serial correlation. As

 

20 Of the 856 country years involving the initiation of a MID, 23% initiate more than one MID in a year.
2' King (1989) and Benoit (1996) provide thorough discussions of the parameterization of the Poisson and
Negative Binomial regression models.

73

the data is pooled, it is possible that values of some of the variables in one period
are conditional on values in a prior period. One way to address this issue is with
a general estimating equation (GEE). The GEE method is quasi-likelihood and
emphasizes a population-averaged approach to estimation. In a recent review of
GEE models, Zorn (2001: 475) notes that population-averaged models are
“valuable for making comparisons across groups or subpopulations.” Since the
substantive focus of this research is on the general propensity of a state to
initiate and reciprocate militarized interstate disputes, a population-averaged
approach, such as GEE, is the most appropriate statistical method. In addition to
harmonizing one’s statistical model to the substantive questions under
investigation, the GEE approach allows the modeler to specify the within group
correlation structure for the panels. The nature of the data generating process
suggests the potential for serial correlation, so I specify an AR (1) correlation
structure.22 I also lag all of the independent variables one year to gain more
conﬁdence in the causal relationship between the independent and dependent

variables.

Dependent Variables

This paper analyzes the causes of interstate military conﬂict. I measure
this concept by focusing on militarized interstate disputes (MIDs). Mle
encompass wars, but also include less extreme forms of conﬂict, including

threats and displays of force and conﬂicts with less than 1000 casualties (Jones,

 

22 For more detailed discussion of GEE models, see Zorn (2001) and Liang and Zeger (I986). Oneal and
Russett (199% and 1999b) also employ a GEE model with an AR (1) correlation structure.

74

Bremer, and Singer, 1996). Additionally, this paper emphasizes the initiation and
targeting of MIDs. This research follows Leeds and Davis (1997) and Prins
(2001) and deﬁnes initiators from the MID dataset based on the Side-A and
Originator codings. An initiator is a state that is involved on the very ﬁrst day of
hostilities and is considered the ﬁrst to militarize the dispute. The target is the

non-initiating state in a dispute.

Independent Variables

The primary predictor variables of interest concern the level of a state’s
trade and investment activity. I use commodity trade data from the National
Bureau of Economic Research (NBER). Using the NBER data, | distinguish
between exports and imports and classify all trade into one of ten Standard
International Trade Classiﬁcation (SITC) categories. Based on Reuvany and
Kang (1998), I operationalize strategic imports as trade in SITC category 3—
mineral fuels. Non-strategic imports encompass all other types of trade, that is
trade in all other SITC categories. The argument does not simply emphasize
those nations that have large absolute amounts of trade, but rather those states
whose economies most depend on trade. The emphasis on domestic welfare
requires that one examine trade in relationship to its overall impact on the
economy. Therefore, I divide the value of trade (strategic imports and exports,

non-strategic imports and exports) by the state’s gross domestic product.23 The

 

23 Data for Gross Domestic Product comes from Gleditsch (2000).

75

NBER data is only available for the 1970 to 1992 period; this variable is also
lagged one year.

Foreign direct investment is also an integral part of economic
dependence. Like trade, foreign direct investment needs to be measured relative
to the size of the economy. The variable foreign investment measures a state’s
“net inﬂows of investment to acquire a lasting management interest (10 percent
or more of voting stock) in an enterprise operating in an economy other than that
of the investor” (World Bank, 1997). Data for this variable comes from the World
Bank’s World Development Indicators (1997).

It is more difﬁcult to deﬁne militarization. The essence of the concept is
that a state is primed for military aggression. Raw military expenditures are not a
satisfactory measure as they are biased toward wealthier states. Japan, for
example, spends about one percent of its gross domestic product on defense.24
Because the Japanese economy is so large, this is a signiﬁcant amount, yet few
people would call Japan a military state. Instead, I create a measure of
militarization based on the number of military personnel in a state. A measure
based on military personnel better approximates the concept of militarization than
a measure based on military expenditures because personnel are essential for
conquering territory. Since larger nations, in terms of both population and
territory, are likely to have more people in their military, I take a ratio of military
personnel divided by total population.25 Data on military personnel and total

population come from the Correlates of War Composite Capabilities (COWCAP)

 

24 CIA World Factbook: http://www.cia.gov
25 Benoit (1996) uses a similar measure.

76

index.26 (See Appendix A for a list of the most militarized states at the beginning
of each of the last ﬁve decades.)

To the extent that global economic involvement is part of a broader liberal
peace, it is important to control for the effects of regime type. While the dyadic
evidence indicates that democracy reduces conﬂict propensities, the relationship
between democracy and peace at the monadic level remains unclear. Maoz and
Abdolali (1989), for instance, ﬁnd that democracies engage in conﬂict as much as
other regime types. However, Benoit (1996) observes that democracies ﬁght
fewer wars than other regime types, although the relationship between
democracy and war becomes insigniﬁcant when control variables are added.
Leeds and Davis (1999) examine non-militarized international conﬂict at the
nation-state level and ﬁnd that democracy does contribute to more cooperation
and less conﬂict. I use the Polity IV democracy index to measure the level of
democracy in a state; this variable ranges from zero to ten, with higher values
indicating higher levels of democracy (Marshall and Jaggers, 2000).

Previous research has also uncovered relationships between economic
development, major power status and international conﬂict. Recent research
suggests that the democratic peace may really be an economic development
peace. Hegre (2000: 7), for instance, argues that “the liberal peace hypothesis
may depend in part on the structure of the economies of the states in question."
Similarly, Mousseau (2000: 473) writes that “the democratic peace may be

limited to the prosperous market nations." Rosecrance’s theory also emphasizes

 

26 I use Bennett and Stam’s (2000) EUGene program (version 2.101) to generate the data on military
personnel and population. I also use EUGene to generate the data for economic development and major
power status.

77

the importance of domestic welfare. The effect of both democracy and economic
openness, then, may be contingent on the level of a state’s economic
development. The impact of development itself is unclear. Development may be
associated with democracies, suggesting that it leads to a peaceful foreign
policy. But, underdevelopment may also contribute to peace. Underdeveloped
states simply do not have the means to attack other states, and given their poor
ﬁnancial status there is not much that a potential conqueror can gain. To
measure economic development I use the energy consumption variable from the
COWCAP index, divide it by the state’s population, and then take the natural log.
This is the same measure of economic development utilized by both Benoit
(1996) and Hegre (2000).27

Finally, it is necessary to consider the role major powers play in
international society. Major powers have interests and responsibilities that span
the globe. They also are frequently called upon to settle international issues or
broker peace agreements. Major powers, then, have greater opportunities and
typically a greater willingness than other states to be involved in conﬂict. I use
the COW coding to determine which states are major powers. In the post-World
War II period, the COW coding includes the United States, the Soviet

Union/Russia, France, Britain, China, and after 1990 Germany and Japan.

 

27 Mousseau (2000) operationalizes economic development as GDP per capita. Hegre (2000) shows that
GDP per capita and energy consumption per capita are highly correlated. The advantage of energy
consumption is that it has fewer missing values.

78

Empirical Results

Initiation of Militarized Interstate Disputes

An empirical analysis of the initiation of militarized interstate disputes
(Mle) for all members of the international system between 1970 and 1992
reveals strong support for the theoretical hypotheses. I begin with a baseline
model that measures the inﬂuence of a state’s import dependence in both
strategic and non-strategic commodities, foreign direct investment, militarization,
level of democracy, economic development, and major power status on the
number of initiated militarized disputes per year (see Model 1a in Table 1). As
indicated in Model 13 in Table 1, both trade dependence variables are
statistically signiﬁcant and negative. The more a state’s economy is dependent
on international trade, whether it be in strategic or non-strategic goods, the less
likely it is to opt for military conﬂict against other states. However, the pacifying
inﬂuence of trade dependence only extends so far. As indicated by the strategic
trade-militarization interaction variable, militarized states importing strategic
goods are more likely to initiate a MID. These results support hypotheses 1a and
1b. Trade, in general, is a constraint on military aggression. But, strategic trade
dependence is a vulnerability states aim to alleviate, and will do so if they have
the means. Consistent with the arguments of this paper, foreign direct
investment also exerts a statistically signiﬁcant and negative inﬂuence on the
initiation of Mle. The more a state is open to foreign investment, and thus
economically dependent on other states, the less likely is it to initiate a militarized

dispute.

79

The empirical results for Model 1a also conﬁrm the pacifying inﬂuence of
democratic institutions. Not only are democracies unlikely to enter into a MID
against other democracies, democracies are also less likely to initiate Mle in
general. That is, the results for Model 1a reveal a monadic democratic peace in
the post-WWII time period. Democracies are less likely to initiate disputes than
non-democracies, even after controlling for the important effects of development,
trade, and global reach.

Given their commitments around the world and responsibility to protect
allies and client states, I also hypothesized that major powers would be likely to
initiate more militarized interstate disputes than other states. The empirical
analysis supports this conclusion.

Although statistical signiﬁcance is important, empirical analyses may ﬁnd
signiﬁcant variables that exhibit little or no substantive impact. Substantive
signiﬁcance is especially important to policy-makers who need to allocate scarce
resources in the most efﬁcient manner. One useful way to assess the
substantive impact of a variable is by calculating changes in the relative risk of an
event occurring. Relative risk measures the change in the risk of event
occurrence given one set of values of the explanatory variables relative to the
probability of event occurrence given a baseline set of values. As noted by King
and Zeng (2001), relative risks are commonly reported in the popular media (e.g.,
ﬁve servings of fruits and vegetables a day will reduce the risk of colon cancer by
ﬁfty percent). Accordingly, I create a baseline model of event occurrence where

all the interval level variables are set at their mean value and the dichotomous

80

variables are set at zero. Then, I assess the change in relative risk by increasing
(and decreasing) the interval covariates by one standard deviation and the
dummy variables from zero to one.

Calculating changes in the relative risk of event occurrence shows that
trade dependence exerts a large substantive impact on the rate of MID initiation
(see Column 2 of Table 1). For instance, a one standard deviation increase in
non-strategic import dependence leads to a 65% decrease in the rate of MID
initiation. Similarly, a one standard deviation increase in strategic import
dependence leads to a 65% decrease in the risk of initiating a MID. However,
militarized states that increases its strategic import dependence is 29% more
likely to initiate a MID.

Model 1 also indicates that the other components of the liberal peace
argument, foreign direct investment and democratic institutions, exercise an
important inﬂuence on the likelihood of a state initiating a militarized dispute. If a
state were to increase its level of foreign direct investment by one standard
deviation over the mean for all states, it would decrease the likelihood of that
state initiating a dispute by 48%. An increase in the democratic institutions and
culture of a state, on the other hand, only decreases the risk of initiating a MID by
26%. Finally, the level of militarization in a state is also an important inﬂuence on
MID initiation. Increasing the level of militarization in a state increases the risk of
MID initiation by 28 °/o.

Perhaps the most interesting finding in Model 1a is that economic

development is positively related to the initiation of Mle. Other research has

81

found that economic development contributes to peace (Hegre, 2000; Mousseau,
2000). And, Mousseau (2000: 473) writes that “economic development may
foster the values and preferences that lead to both democratic consolidation and
democratic peace.” To more thoroughly examine these arguments, I interact
democracy with economic development.

Including the democracy-development interaction term in the analysis
(Model 2) makes the story increasingly clear. The interaction of development
and democracy is statistically signiﬁcant and negative. The democracy variable
is still signiﬁcant and negative, indicating that the pacifying inﬂuence of
democracy is not dependent upon economic development. On the other hand,
non-democratic, developed states are more likely to initiate MIDs. Development
without democracy seems to encourage militarily aggressive foreign policies.

Further, all other variables in Model 2 retain their sign and statistical signiﬁcance.

Targets of Militarized Interstate Disputes

In this section I empirically evaluate the factors inﬂuencing the likelihood of
a state being targeted by military aggression. Model 3 presents the results of
these analyses. As before, I determine the substantive impact of the variables by
calculating changes in the relative risk of event occurrence given a change in the
covariates. Changes in the relative risk of being the target of military aggression
are presented in column 3 of Model 3.

As expected, higher levels of strategic exports increase the likelihood that

a state will become the target of a militarized dispute (see Model 3). In fact,

82

strategic export dependence increases the likelihood of being targeted by about
26%. Importantly, this only applies to underdeveloped states. Wealthy states
with high levels of strategic exports are less likely to be targets of military
aggression. Put differently, states with high amounts of strategic resources have
something other states want, and if they are weak, they are tempting targets for
military aggression. Consistent with hypothesis 3b, the empirical results indicate
that higher levels of foreign direct investment reduce the likelihood a state will be
a target of military aggression. A one standard deviation increase in a state’s
level of foreign direct investment decreases the likelihood of it being the target of
a militarized dispute by about 22%. This illustrates an important difference
between trade and investment. Where trade may or may not contribute to peace,
investment, by increasing the costs of conﬂict and reducing the incentives for
conquest, foreign direct investment always contributes to peace.

The empirical results also reveal that democracies are neither more nor
less likely to be targets of military aggression. Insofar as democracies rarely
enter into military conflict with other democracies, the insigniﬁcance of this
variable suggests democracies may be perceived as weak by non-democracies.
While it is often said that the best way to promote peace is to prepare for war, I
ﬁnd that the more militarized a state, the more likely it is to be a target of military
aggression. Increasing a state’s level of militarization increases the likelihood it

will become the target of a militarized dispute by 16%.

Robustness and Sensitivity Analysis

83

In this section, I analyze the sensitivity of the empirical models to
alternative operationalizations of central variables. The arguments about a
state’s military orientation emphasize conquest as a wealth generating
mechanism. Because the seizure and occupation of land are at the heart of this
perspective, the military state variable is the proportion of military personnel to
total population. Nonetheless, for a state’s military to be most effective, it needs
sophisticated weaponry. To take into account this additional aspect, I created a
military index variable based on both military personnel and expenditures. For
the military personnel part of the index, I determined the ratio of each state’s
military personnel per capita over the highest military personnel per capita in the
international system for each year. I calculated the military expenditures portion
of the index in a similar manner. Then, I added the two components. To make
the variable range between zero and one, I divided the result by two.

Running the above analyses with the military state index variable does not
change the results (see Table 3.4).28 Like the military personnel variable, the
military index variable is always signiﬁcant and positive. The trade and
investment dependence variables remain statistically signiﬁcant and negative,
and, perhaps most importantly, the strategic trade-militarization interaction
variable is signiﬁcant and positive. Thus, none of the central theoretical variables
are affected by this alternative speciﬁcation.

It is not uncommon to ﬁnd researchers employing a dummy variable for

democratic institutions, suggesting they are either present or not present. This is

 

28 The correlation between the military state index variable and the militarization variable based only on
military personnel is .83.

84

inspired by arguments underscoring the importance of mature and stable
democratic norms and regimes. In fact, democratic peace arguments emphasize
an either/or situation with respect to democratic institutions. To examine if this
alternative speciﬁcation affects the results, I re-ran every model using a
democratic dummy variable. The variable equals 1 if a nation registered 6 or

greater on the Polity IV democracy index. All results remain the same.

Conclusion

One aim of this research is to develop an argument on economic
dependence and international conﬂict. Contrary to previous research, I argue
that the relationship between trade and conflict varies with the type of goods
traded and is conditional on a state’s power. Non-strategic trade dependence
increases the costs of conflict and contributes to peace. However, states
dependent on importing strategic goods have an incentive to reduce their
vulnerability, and will do so if they have the power. In other words, dependence
on importing strategic goods does not contribute to peace if a state also has a
large military. Moreover, I posit that previous arguments have omitted an
important part of economic dependence: foreign direct investment. Economic
dependence on foreign direct investment is similar to trade dependence in that it
raises the costs of potential conﬂict. States with investment do not want to
jeopardize their investments. Investment, however, differs from strategic trade in
that it does not provide an incentive for attack. Foreign direct investment, then,

always contributes to peace.

85

A second goal of this research is to appropriately test this theoretical
argument. For both Immanuel Kant (1795) and Richard Rosecrance (1986,
1999), and contrary to most research analyzing their arguments, the effects of
economic dependence are not limited to dyadic relationships. Correspondingly, I
evaluate the inﬂuence of economic dependence on interstate conﬂict at a
monadic level of analysis. Further, previous studies analyzing monadic
arguments have over-aggregated the dependent variable. By deﬁnition, conﬂict
involves at least two actors. To evaluate a monadic argument, then, it is
necessary to distinguish between participation in interstate conﬂict and the
initiation and targeting of interstate disputes. Failure to parse the dependent
variable in this manner may lead to biased results. More generally, this research
contributes to our understanding of foreign policy decision-making by
distinguishing between initiators and defenders in international hostilities.

I ﬁnd that states economically dependent on importing strategic goods are
more likely to initiate disputes than other states, if they also have a significant
military establishment. Relatedly, I ﬁnd that states with strategic resources are
more likely to be targets of military aggression if they are weak. In short, states
will ﬁght over strategic commodities. The pacifying inﬂuence of trade, then, only
extends to non-strategic commodities. Research that does not disaggregate
trade misses this important distinction. I also ﬁnd that higher levels of FDI reduce
the likelihood of a state initiating and reciprocating militarized disputes. Perhaps
the best way to build peace, then, is to promote globalization. Finally, these

ﬁndings are robust to alternative model speciﬁcations.

86

This research also contributes to the democratic peace research program
by conducting the ﬁrst large-scale examination of the effects of democratic
regimes on the initiation and targeting of disputes. Democracies are not only
unlikely to engage other democracies in militarized disputes, but are also more
peaceful in general. However, democratic institutions do not reduce the
likelihood that a state will become the target of a dispute. Complementing the
research of Hegre (2000) and Mousseau (2000), this research ﬁnds a synergistic
relationship between democracy and development. Developed democracies are
less likely to initiate disputes than developed non-democracies. This implies that

development without democratic institutions is potentially dangerous.

87

Chapter 4

Institutional Similarity and Interstate Conﬂict

In 1989, Francis Fukuyama proclaimed an end to history.29 The end to
which F ukuyama is referring is the end of ideological conﬂict between nation-
states. For Fukuyama, the collapse of the Soviet empire and the concomitant
spread of liberal political and economic institutions ushers in a new age of
international peace. While much has been written about Fukuyama’s
controversial thesis, we lack a comprehensive analysis of the argument. At its
core, the end of history thesis is about political and economic institutions. But,
what is the relationship between political and economic institutions and interstate
conﬂict? If two states have different political institutions, is conﬂict more likely to
occur between them? If two states have different economic institutions, is
conﬂict more likely to occur between them? If so, why?

I argue that institutional similarity is an important inﬂuence on the
likelihood of interstate conﬂict between states. The central tenet of this
argument, in line with Fukuyama (1989, 1992), is that a state’s political and
economic institutions are a reﬂection of its ideology. Further, ideology inﬂuences
preferences. Therefore, states with similar ideology, owing to similar institutions,
are likely to share similar foreign policy preferences. Gartzke (1998, 2000) notes
that a necessary condition for interstate conﬂict is a difference in preferences; to
the extent that states share similar preferences, they have less to ﬁght over and

are more likely to be at peace with each other. Power transition theorists

 

29 F ukuyama’s original article appeared in The National Interest (1989). He later expanded the argument
into a book, The End of History and the Last Man (1992).

88

advance a similar argument (Organski and Kugler, 1981; Kugler and Lemke,
1996; Lemke and Reed, 1996). In power transition theory, the two necessary
conditions for great power conﬂict are a power transition and dissatisfaction with
the status quo. Dissatisfaction clearly implies some difference of preferences.
Moreover, recent work in the power transition research program suggests that
dissatisfaction is inﬂuenced by the similarity of domestic institutions in two states
(Werner and Lemke, 1997).

Despite its theoretical relevance, little research has analyzed the
relationship between domestic institutions and dissatisfaction. Gartzke notes that
“most current explanations for the democratic peace argue implicitly that the
factors that motivate conﬂict are not substantially less common between
democracies than among other states. It is assumed that democracies are about
as likely to disagree” (1998: 3). Gartzke does not examine the link between
domestic institutions and satisfaction, suggesting that “joint democracy is not the
primary contributor to national preferences” (1998: 11). Nonetheless, I contend
that institutions are important determinants of preferences and satisfaction.
Kugler and Lemke also suggest that “It might well be possible to create a valid
indicator of status quo evaluations by calculating the degree to Which states are
domestically similar to the dominant state” (Kugler and Lemke, 2000: 154).

“The problem lies in identifying what speciﬁc domestic institutions matter”
(Kugler and Lemke, 2000: 154). Drawing on Fukuyama, I argue that both the
political and economic institutions in a state are important inﬂuences on the

state’s foreign policy preferences. A second problem lies in identifying the

89

appropriate reference point for assessing satisfaction with the status quo.
Werner (2000) advances a dyadic preference similarity argument. Fukuyama’s
argument also seems to emphasize dyadic institutional similarity and satisfaction.
Power transition theory, however, emphasizes satisfaction with the systemic
status quo. The appropriate measure, then, should assess the similarity of
institutions between the dominant state and each individual state in a dyad. To
more fully address these issues, I create four new measures of satisfaction. I
construct separate measures of satisfaction based on the dyadic similarity of
political institutions and economic institutions. I also construct measures of
satisfaction that incorporates the similarity of each state’s political and economic
institutions with the dominant state.

I also investigate whether the types of conﬂict are the same for each type
of status quo satisfaction. Does dyadic satisfaction reduce one type of conﬂict
and systemic satisfaction reduce another type of conﬂict? Do measures of
satisfaction based on economic institutions affect different dispute types than
measures of satisfaction based on political institutions?

The paper proceeds as follows. In the next section, I develop the
arguments connecting domestic institutions to satisfaction with the status quo
and interstate conﬂict. These arguments emphasize that both political and
economic institutions are central to satisfaction. The argument leads to
hypotheses for both economic and political institutions and both the dyadic and
systemic levels of analysis. In the third section, I present two new measures of

status quo satisfaction based on domestic political and economic institutions.

90

 

Cr

I8

The fourth section of the paper tests the two hypotheses. The fifth section
presents an analysis of the relationship between satisfaction based on
institutional similarity and types of militarized disputes. In the ﬁnal section, I

discuss implications of this research and avenues for future research.

Satisfaction and Interstate Conﬂict

Why do nations ﬁght?

Two sets of factors inﬂuence decisions regarding international conﬂict:
opportunity and willingness. In more speciﬁc terms, for conﬂict to occur, states
must have both the means and the desire to attack another state; thus,
opportunity and willingness are both necessary conditions for the outbreak of
conﬂict (Most and Starr, 1989; Gartzke, 1998). The opportunity and willingness
structure is related to the other primary framework for studying international
relations, namely expected utility. Expected utility is a function of the probability
and utility associated with winning and losing in a conﬂict. In this perspective,
probability is akin to opportunity and utility is similar to willingness.

Most research has focused on the opportunity dimension of conﬂict. This
is attributable to the inﬂuence of realism, which assumes there is always a
motivation for conﬂict so there is little reason to investigate the willingness for
conﬂict. While I will brieﬂy discuss factors related to opportunity and include
them in the empirical model presented in section four, the primary focus of this

research is exploring the willingness dimension of conﬂict.

91

To say that willingness is a necessary condition for conﬂict is to say that
there must be something to ﬁght over. Further, there will only be something to
ﬁght over when actors hold different preferences on an issue(s). One task, then,
for students of international relations is to identify the sources of foreign policy
preferences. Relatedly, if one actor is dissatisﬁed with the status quo, then there
is potential for conﬂict. Lemke and Reed (1996: 145) argue that “States satisﬁed
with the status quo desire no changes to the international order, and thus have
nothing over which to ﬁght.” A second task, then, is to identify what makes states

satisﬁed with the status quo, and how to measure such satisfaction.

Domestic Institutions As A Source of Foreign Policy Preferences

Domestic institutions are a reﬂection of a state’s ideology, which in turn
strongly influence a state’s foreign policy preferences. Domestic institutions
reﬂect a state’s ideology because they are a function of the group in power.
Bueno de Mesquita, Morrow, Siverson, and Smith (1999) argue that states with
large winning coalitions, i.e. democratic states, provide a more extensive set of
public goods. On the other hand, states with small winning coalitions provide
fewer public goods. Differences in domestic institutions also inﬂuence a state’s
views of human rights. The United States, a democratic state, for example, has
pushed for a set of universal human rights that includes freedom of religion.
China, on the other hand, does not include freedom of religion in its view of
human rights. Assuming there are universal human rights, which both the United

States and China do, then the difference on what is a human right and what is

92

 

not is likely a reﬂection of the domestic institutions in a state. Finally, the change
of government in Iran in 1979 led to a dramatic change in that state’s foreign
policy preferences. The shift from a secular autocracy to a strict religious
oligarchy immediately disrupted Iran’s alliances with secular states like the
United States.

Because domestic institutions inﬂuence a state’s foreign policy
preferences, the similarity of domestic institutions between states affects a
dyad’s satisfaction with the status quo. Research on the relationship between
democratic institutions and interstate conﬂict supports this contention. As is well
known, two democracies rarely, if ever, engage in extensive military conﬂict.

One explanation of the democratic peace phenomenon is that democracies

share common norms of dispute reconciliation so they are able to resolve severe
conﬂicts.30 A second explanation of the democratic peace emphasizes the role
of democratic institutions in communicating resolve.“1 A third explanation,
however, suggests that the democratic peace is the result of democracies having
nothing to ﬁght over.32 “It is thus satisfaction with the status quo that accounts for
the lack of wars between democracies in the past two centuries (Lemke and
Reed, 1996: 160).” Further, institutional similarity and satisfaction is not limited

to democracies. Werner (2000: 344) ﬁnds that “politically similar states are less

 

3° See Maoz and Russett (1993) for a discussion of the normative explanation. For a broader and survey of
the democratic peace research program, see Russett and Starr (2000).

3' Most game theoretic analyses emphasize the institutional explanation of the democratic peace. See for
example, Bueno de Mesquita, Morrow, Siverson, and Smith (1999) and Schultz (1999). Maoz and Russett
( 1993) also discuss the institutional explanation.

32 In addition to Lemke and Reed (1996), Gartzke advances this argument (1998, 1999).

93

 

likely to be engaged in conﬂict with each other than are politically disparate
states.”

In assessing foreign policy preferences and satisfaction, previous
research has focused exclusively on political institutions and relationships. World
politics, however, is not only about political relationships, but also economic
relationships. Indeed, ideology is as much a function of a state’s economic
institutions as it is political institutions in a state. The historian Richard Pipes
argues that private property rights are the basis for the establishment of liberal
political systems: “property . . . provides the key to the emergence of political
and legal institutions that guarantee liberty” (1999: xii). Similarly, Fukuyama
argues that “In its economic manifestation, liberalism is the recognition of the
rights of free economic activity and economic exchange based on private
property and markets” (1992: 44). Accordingly, I contend that a central
dimension along which to distinguish between states foreign policy preferences is
the degree to which states protect private property.

To the extent that the protection of private property inﬂuences a state’s
foreign policy preferences, then dyadic satisfaction depends as much on the
similarity of economic institutions as it does on the similarity of political
institutions. Brawley (1993), for instance, argues that international politics is as
much about economic rules as it is about political rules. When a new state
becomes dominant, it establishes a particular economic order to maximize its
power. According to Brawley, the type of economic order established by the

dominant state is a function of both its economic factor endowments (e.g. land,

94

labor, and capital) and political system. What is especially important about his
argument is that different political systems may favor similar economic policies
depending on their factor endowments. For instance, a labor-abundant
republican regime and a labor-abundant autocratic regime will both favor free
trade policies for the labor intensive sector of the economy. To understand

satisfaction with the status quo, then, it is necessary to consider both political and

_,-1

economic institutional arrangements. Similarly, in the context of power transition

theory, Lemke and Reed note that for two states to be satisﬁed they “would have

E~¢

to possess similar internal economic composition as well as similar regime types”
(1996:146)

Including economic institutions in an analysis of satisfaction permits useful
distinctions among states not previously possible. Olson (1996), for instance,
ﬁnds that the central difference between economically developed states and
economically underdeveloped states is the extent that they protect private
property. Similarly, Sobel (1999) ﬁnds that differences in the protection of
property rights across states is a signiﬁcant inﬂuence on the amount of capital 3
state can attract. Additional distinctions based on protection of property rights
are also possible. For instance, the United States and Sweden are both
democracies. However, we can distinguish clear differences between the
economic institutions in the United States and Sweden. Further, an increasing
number of states are adopting free market economic institutions but retaining
authoritarian political institutions. Singapore and China are the most prominent

examples. If we only consider political institutions, we are likely to expect conﬂict

95

between the United States and China. But, if we consider the growing similarity
in economic institutions, we may not expect conﬂict to occur. This distinction has

not been addressed in previous research.

Institutional Similarity, Satisfaction and Interstate Conﬂict

Institutional similarity enhances satisfaction and reduces conﬂict by
diminishing areas of disagreement and reducing the beneﬁts of conﬂict. First,
institutional similarity improves mutual satisfaction by taking away a central issue
of interstate conﬂict: ideological disagreement. Historians of international conﬂict
note that ideological issues have become a growing source of interstate conﬂict
(Holsti, 1991). Ideological concerns usually revolve around both the treatment of
individuals within a state and the composition of a state’s government (Werner,
2000). For example, a key source of contention between the United States and
China is the issue of universal human rights. The United States has pushed for a
set of universal human rights that includes freedom of religion. China, on the
other hand, does not include freedom of religion in its view of human rights. This
ideological difference, which is a reﬂection of different domestic institutions,
between the United States and China is clearly a source of conﬂict. Because
ideology is a function of a state’s domestic institutions, the composition of a
state’s government may also be a source for interstate conﬂict. The fear of
changes to state governments prompted the Soviet Union to intervene in
Hungary in 1956 and in Czechoslovakia in 1968.

Second, institutional similarity strengthens satisfaction and reduces the

likelihood of conﬂict by reducing the beneﬁts of conquest. Bueno de Mesquita

96

 

(1981) argues that the beneﬁt of conﬂict is a function of the differences between
states foreign policy preferences. As preference afﬁnity decreases, the beneﬁts
of conﬂict increase relative to the costs. Werner (2000) also notes that one of the
primary beneﬁts a state can derive from conﬂict is the re-structuring of another
state’s foreign policy preferences. In short, as institutional similarity reduces the
beneﬁts of conquest, states have less incentive to ﬁght.

The central factors affecting the opportunity dimension of conﬂict are
distance and power parity.33 Consider distance. As the distance separating two
states increases, it is more difﬁcult to mount a successful military campaign
(Bueno de Mesquita, 1981; Lemke, 1995). Larger distances make it more
difﬁcult both to get troops to the battleﬁeld and to re-supply troops. Separation,
then, makes it more difﬁcult for conﬂict to occur. Greater distance between
countries may also have an indirect effect of decreasing the willingness for
conﬂict. If two states are far apart, they may have minimal interaction, and thus
little to fight over.

Power transition theory has also identiﬁed power parity as an important
factor affecting the opportunity for conﬂict (Organski and Kugler, 1981; Kugler
and Lemke, 1996). As the imbalance of power in a dyad increases, the
probability of victory in a dispute for the weaker nation decreases. Unless the
stakes are very high, the weak state will conclude that conﬂict is not an option as

it has little probability of winning. Put differently, power parity provides an

 

33 While contiguity is also a major inﬂuence on the opportunity for conﬂict, 1 consider it a subset of
distance and do not consider it separately. Contiguous dyads have minimal distance between them. Realist
theories emphasize the importance of anarchy as creating an opportunity for conﬂict. While the
international system may be anarchic, a homogeneous structure cannot exercise any leverage in explaining
why some dyads experience conﬂict and others do not.

97

 

opportunity for conﬂict because it increases the potential beneﬁts relative to the
costs. If there is not power parity, the costs of conﬂict for the weaker side greatly
outweigh the beneﬁts, and they are expected to give in to the demands of the
stronger state. If there is parity, then both sides have a chance to beneﬁt;
therefore, as power parity increases the potential beneﬁts relative to the costs
increase and conﬂict becomes more likely. Insofar as this argument is accurate,
then conﬂict should be more likely to occur under conditions of power parity then
under conditions of a power imbalance. Numerous empirical studies support this

hypothesis (see Kugler and Lemke, 1996, 2000).

Dyadic Satisfaction vs. Systemic Satisfaction

The impact of preference similarity on conﬂict revolves around the status
quo. However, previous research has measured the status quo at two different
levels of analysis: the dyad and the system. Most research on the democratic
peace focuses on the dyad as the appropriate level and unit of analysis. The
research question in the democratic peace research program is whether or not
two democracies are less likely than other dyads to experience interstate conﬂict.
Similarly, Werner’s (2000) research analyzes the effects of dyadic political
similarity on interstate conﬂict. The argument is straightforward and leads to my
ﬁrst set of hypotheses. The hypotheses distinguish between political and
economic institutions in order to examine the independent effects of each
Hypothesis 1a: Dyads with similar political institutions will have similar

preferences, thus, they will be satisﬁed with each other and experience less
conﬂict than dyads with dissimilar political institutions.

98

Hypothesis 1b: Dyads with similar economic institutions will have similar
preferences, thus, they will be satisﬁed with each other and experience less
conﬂict than dyads with dissimilar economic institutions.

Power transition theory also underscores the importance of satisfaction
with the status quo as a central determinant of interstate conﬂict. The focus of
power transition arguments, however, is not the dyadic status quo, but the
systemic status quo. Power transition arguments emphasize a hierarchical
international system, where the rules of the system are established by the
dominant state. States maximize their gains when they are domestically
organized in a way similar to the dominant state. Relatedly, Douglass North
(1981) and Richard Rosecrance (1999) note that shared ideology is important for
obtaining the beneﬁts of a market-based system. Market systems depend on
trust, and similar institutions enhance trust between actors. An individual state,
then, is satisﬁed with the systemic status quo when its domestic institutions are
similar to the domestic institutions of the dominant state. It is not entirely clear
how satisfaction with the systemic status quo affects dyadic behavior. If both
states in a dyad are satisﬁed with the systemic status quo, then conﬂict is not
likely to occur. However, if two states are satisﬁed with each other, but not with
the systemic leader, conﬂict expectations in this dyad are unclear. As a
minimum, we can state the following hypothesis.

Hypothesis 2a: Dyads in which both states have political institutions similar to
the dominant state should experience less conﬂict than other dyads.

Hypothesis 2b: Dyads in which both states have economic institutions similar to
the dominant state should experience less conﬂict than other dyads.

99

 

Measures of Satisfaction Based on Institutional Similarity

A central question for students of international conﬂict is to determine the
appropriate reference for satisfaction with the status quo. Clearly, the status quo
is a function of time and point-of-view. At present, Israeli control over Jerusalem
is the status quo on this contentious issue. One hundred years ago, however,
Jerusalem was controlled by the British. This research does not address status
quo issues concerning time and point-of-view. Instead, the focus here is on the
appropriate level of analysis for assessing satisfaction with the status quo. Does
dyadic satisfaction have a different inﬂuence on conﬂict behavior than systemic
satisfaction? In the next section, I discuss previous measures of both dyadic and
systemic satisfaction. Then, I offer a new measure based on domestic
institutions.

Previous research has measured a dyad’s willingness for military conﬂict,
or satisfaction with the status quo, in four ways: alliance portfolio similarity,
United Nations voting records, military build-ups, and money market discount
rates. Bueno de Mesquita (1981) created a measure of dyadic afﬁnity based on
the similarity of two states alliances portfolios?4 The idea is that states with
similar alliances have similar foreign policy preferences, and therefore less to
ﬁght over. Two issues limit the utility of using alliance portfolios to measure
foreign policy preferences. First, alliances are strategic decisions. As such,
alliances only indicate speciﬁcally revealed preferences. Generally, an alliance is

\

3 . . . , . . . . . . ,
4 Signorino and Ritter s (1999) S-index IS a statistical improvement over Bueno de Mesquita 5 measure.
conceptually, the measures are the same.

100

 

sometimes made to signal to a potential adversary one’s commitment to a third
party. If this commitment is clear, then an alliance may not be necessary. The
United States does not have a formal alliance with Israel, yet all available
evidence indicates that the United States is committed to the defense of Israel.
The United States and Israel also seem to share similar foreign policy views,
despite the absence of a formal alliance. Moreover, if a state is not directly

threatened by another state, then it may not enter into any alliances, or if a state

feels it is secure without an alliance, then it may not enter into any alliances.
New Zealand is an example of the ﬁrst situation. New Zealand is no longer a
member of the ANZUS alliance, in part because it does not face any immediate
threats. Yet, it is not clear that the United States and New Zealand share vastly
different foreign policy preferences. Switzerland is an example of the second
situation. Switzerland is a neutral country, yet on most issues the United States
and Switzerland share similar views. Second, Gartzke (1998) argues that during
the Cold War, states did not vary their alliance patterns, limiting the leverage of
this measure.

Gartzke (1998) created a measure of dyadic afﬁnity based on the similarity
of states’ United Nations voting records. This measure is similar to Bueno de
Mesquita’s (1981) alliance portfolio measure; the more two states vote alike, the
less willingness they have for military conﬂict. Many United Nations votes,
however, are on symbolic issues, such as condemning terrorism. Such votes are

unlikely to reveal signiﬁcant foreign policy preferences. Additionally, United

101

Nations votes on which there is a lack of unanimity tend to be on a single issue:
the Middle East conflict between Israel and the Palestinians.

Werner and Kugler (1996) present a measure of dissatisfaction based on
military buildups. While Werner and Kugler focus on challenger states and
satisfaction with the systemic status quo, the idea of military buildups easily
extends to dyads. Dyads engaging in arms races may be considered dissatisﬁed
with each other. However, it is much less clear that the lack of an arms race
implies satisfaction with the status quo.

Finally, Bueno de Mesquita (1990) uses money market discount rates to
measure Germany’s evaluation of the systemic status quo in the late 19th
century. He argues that “a rising discount rate for a nation’s money reﬂects a
broad base of declining conﬁdence in that nation” (Bueno de Mesquita, 1990:
42). Lemke and Reed (1996) note that this deﬁnition of satisfaction may result in
the dominant state being dissatisﬁed with the status quo, but, by deﬁnition, the
dominant must be satisﬁed. To address the shortcomings in previous measures,
I offer a new measure of foreign policy preferences.

I argue that a state’s foreign policy preferences are driven, to a large
extent, by the nature of its political and economic institutions. For instance,
authoritarian political institutions and extensive state control over the economy
and property rights characterize communist political systems. On the other hand,
democratic political systems and extensive protection of private property rights
characterize liberal political systems. The Cold War ideological conﬂict is a

reﬂection of this difference in political and economic institutions. Insofar as

102

domestic institutions inﬂuence a state’s foreign policy preferences, then
measuring the similarity of domestic institutions provides a broader measure of

satisfaction.

Political Institutions

“Political rules [institutions] broadly deﬁne the hierarchical structure of the
polity, its basic decision structure, and the explicit characteristics of agenda
control (North, 1990: 47).” Phrased differently, political institutions focus on two
basic questions. First, what is the power relationship between the government
and the people? In other words, what means, short of violence, do the citizens of
a state have for removing a political leader? Second, what is the relationship
between the different branches of government? Does one branch or person
dominate the political process? Are there checks and balances?

I use the Polity IV dataset to measure the nature of the political institutions
in a state (Marshall and Jaggers, 2000). The Polity project produces an index of
democracy for each state in the international system. This index is largely a
function of four different authority dimensions: competitiveness of executive
recruitment, openness of executive recruitment, executive constraints, and
competitiveness of political participation. Although the Polity data provides a
summary index measure of democracy in each state, there are different ways to
attain the same index score (Gleditsch and Ward, 1997). For this reason, I follow
Werner (2000) and disaggregate the democracy index into the above four

dimensions, and create a measure, dyadic political similarity, of political similarity

103

.-
, .-
.

based on the distance between states’ scores on these dimensions. Like
Werner, I standardize each dimension by dividing the distance by the maximum
possible value for that dimension. I also multiply it by negative one so that the
measure reﬂects political similarity instead of political dissimilarity.

Next, I created a second variable, systemic political similarity, measuring a
dyad’s satisfaction with the systemic status quo. First, I calculate the distance
between state “A’s” political institutions and the political institutions in the
dominant state. Since this research only covers the post-World War II period,
the dominant state is always the United States. Second, I calculate the distance
between state “B’s” political institutions and the political institutions in the
dominant state. Then, I add one to each of the above scores and multiply them
together. I add one to prevent multiplying by zero. Finally, I take the square root
of this value and divide by the highest possible value, 21, so that the ﬁnal value
ranges between 0 and 1. Again, I multiply by negative one so that higher scores

represent more similarity with the system leader.

Economic Institutions

Where the focus of political institutions is participation in the political
process and protection of civil rights, the focus of economic institutions is the
protection of private property rights. “Economic rules [institutions] deﬁne property
rights, that is the bundle of rights over the use and the income to be derived from
property and the ability to alienate an asset or a resource (North, 1990: 47).

Institutions enhance property rights, and in turn, economic activity in two ways.

104

First, institutions reduce transaction costs. Transaction costs include the costs
involved in negotiating, implementing, and enforcing a transaction. Transaction
costs limit economic activity; they make for imperfect property rights, and
therefore transaction costs are a central reason property rights need protection
(North, 1990: 8). “The costs of transacting arise because information is costly
and asymmetrically held by the parties to exchange (North, 1990: 108).”
Institutions counter transaction costs by providing information to actors. Second,
institutions enhance property rights by reducing enforcement problems. All
transactions depend on contracts and the enforcement of contracts. When
enforcement is uncertain, haphazard, or costly property rights are less secure.
Commenting on the importance of contract enforcement, Douglass North writes
that “the inability of societies to develop effective, low-cost enforcement of
contracts is the most important source of both historical stagnation and
contemporary underdevelopment in the Third World (1990: 54).” In short,
institutions protect property rights by reducing the uncertainty involved In
economic transactions.

The central dimension along which to measure economic institutions is a
state’s protection of property rights. I use data from the International Country
Risk Guide (ICRG) to measure the extent that the economic institutions in a state
protect private property rights. The ICRG is produced by an international risk
services ﬁrm. It is designed to provide potential foreign investors information on

the protection of private property rights. ICRG data covers about 100 countries

105

over the period 1982 to present (see Appendix A for a list of states in the ICRG
dataset).

Following Sobel (1999) and the Keefer and Knack research group
(Clague, Keefer, Knack, and Olson, 1996; Knack and Keefer, 1995), I use ﬁve
ICRG variables to create a measure of private property protection in a state. The
variables comprising this “regulatory index” are the risk of expropriation, rule of
law, repudiation of contracts, corruption in government, and bureaucratic quality.
As this dataset is unfamiliar to most political scientists, I brieﬂy describe each
variable in the index.

Expropriation risk measures the probability of “outright conﬁscation” or
“forced nationalization.” As the probability that the government will conﬁscate an
investment increases, economic actors are less likely to invest. This measure of
private property protection varies from 0 to 10, with lower scores indicating higher
risk.

The rule of law variable “reﬂects the degree to which the citizens of a
country are willing to accept the established institutions to make and implement
laws and adjudicate disputes” (Knack and Keefer, 1995: 225). Higher scores
indicate greater conﬁdence in one’s government for impartial adjudication of
disputes. This variable ranges from 0 to 6.

Assessing the probability that contracts are repudiated by the government
taps into the amount of conﬁdence private actors can have in the government.
“In the absence of impartial state enforcement, the only impersonal exchanges

taking place between private economic actors will be those that are self-

106

n)

enforcing” (Knack and Keefer, 1995: 211). This variable ranges from 0 to 10,
with lower scores indicating higher risk of contract repudiation.

The corruption in government variable measures the extent to which “high
government ofﬁcials are likely to demand special payments” and “illegal
payments are generally expected throughout lower levels of government” (Knack
and Keefer, 1995: 225). This variable ranges from 0 to 6, with lower scores
indicating more corruption in government.

The bureaucratic quality variable measures the bureaucracy’s “autonomy
from political pressure” (Knack and Keefer, 1995: 225). States with low scores
on this variable have bureaucracies more concerned with political pressures than
efﬁciency in making decisions. This variable ranges from 0 to 6.

I use these variables to create a regulatory index score for each state.
The regulatory index is the sum of the ﬁve variables, where the bureaucratic
quality, corruption in government, and rule of law variables are ﬁrst transformed
into 10 point scales to ensure equal weight for each component of the index.
With the regulatory indices, I create a measure of the similarity of economic
institutions between two states in a dyad. This variable, dyadic economic
similarity, is the absolute value of the difference between the regulatory score of
state “a” and the regulatory score of state “”.b I, then, divide by 50 to make the
variable range between 0 and 1, with higher values indicating more similar
economic institutions in the dyad.

I also created a variable, systemic economic similarity, measuring the

similarity of the economic institutions in a dyad with the system leader. The

107

construction of this variable is similar to the construction of the systemic political
similarity variable. First, I found the difference between the ICRG scores of
states a and b and the ICRG score of the United States. Again, I added one to
each of these scores to prevent multiplication by zero. Then, I multiplied them
together and took the square root. Higher values of systemic economic similarity

represent less similarity with the system leader.

Empirical Model

The above arguments lead me to specify the following empirical model of
interstate conﬂict.
(EQ 1) Conﬂict = 80 + 81*Political Similarity + 82*Economic Similarity 4-
83*Joint Democracy + [34*Trade Interdependence + [35*Foreign Direct
Investment + 86*Power Parity + 87*Allies + 88*Distance + 89*Contiguity
+ [310*Major Power Dyad + e

I operationalize the dependent variable as whether or not a dyad
experienced the onset of a new militarized interstate dispute (MID) in a particular
year. Accordingly, the unit of analysis is the dyad year. “The term “militarized
interstate dispute” refers to united historical cases in which the threat, display or
use of military force short of war by one member state is explicitly directed
towards the government, ofﬁcial representatives, ofﬁcial forces, property, or
territory of another state” (Jones, Bremer, and Stuckey, 1996: 168). I also
include interstate wars in the analysis; the Correlates of War project deﬁnes wars

as disputes that exceed 1000 battle casualties.

108

Data on economic institutions are only available as far back as 1982.
While the economic institutions data extend to the present, data on the
dependent variable only extend to 1992. Thus, the temporal domain of this
analysis covers the period 1982 to 1992. In cross-sectional terms, the analysis
covers over 100 states, including developed and underdeveloped states in all
regions of the world. See Appendix A for a list of countries included in the
analysis.

To estimate the above model, I employ a general estimating equation
(GEE), specifying a logistic link function and a ﬁrst order autoregressive
correlation structure. I use a logistic link function to account for the unique
properties of a dichotomous dependent variable. Further, Beck, Katz, and
Tucker (1998) argue that it is important to control for temporal dependence in
time series, cross-section data. They do so by including temporal dummy
variables or a spline function of these variables. Oneal and Russett (1999) and
Zorn (2000) note that a better, and more theoretical, way to address issues of
temporal dependence is to model them directly. They advocate the use of GEE
models. Given the nature of the data generating process, it is likely that values
of the independent variables in the preceding period strongly inﬂuence values in
the current period, so I specify an AR (1) correlation structure.35 I also lag all of
the independent variables by one year to ensure that the independent variables

are producing changes in the dependent variable and not the reverse. GEE

 

35 While estimating an empirical model similar to one here, Oneal and Russett (1999) also specify an AR

(1) correlation structure. This model produces similar results to one that uses the Beck, Katz, and Tucker
(1998) spline technique. Zorn (2000) also speciﬁes an AR (1) correlation structure for a similar empirical
model.

109

models are advantageous for another reason. They produce population
averaged results, which is more useful for making comparisons across groups.
Since the focus here is on particular types, or groups of states, the GEE
estimator is most appropriate.

I run two different versions of the model speciﬁed above, one with dyadic
political and economic similarity and one with systemic political and economic
similarity. These variables are measured as described in section three. I use
data from the Polity project to measure joint democracy (Marshall and Jaggers,
2000). Speciﬁcally, I use the smaller of two dyadic democracy index scores in a
dyad. This variable ranges from 0 to 10.

I measure trade interdependence as exports plus imports divided by a
state’s gross domestic product. As with the democracy variable, I employ the
weak link operationalization and only include the less dependent state’s trade
interdependence in the empirical model. Data on trade interdependence comes
from the International Monetary Fund.

Unlike most research analyzing economic interdependence, in addition to
trade interdependence I also include a measure of foreign direct investment.
Like trade interdependence, the expectation is that as a dyad’s level of foreign
direct investment increases, militarized conﬂict becomes less likely.
Unfortunately, dyadic foreign direct investment data is not available for a large
number of dyads. Instead I employ a monadic measure of foreign investment.
The expectation is the same. The more an individual state is open to foreign

investors, the less likely it should be to enter into conﬂict with other states. Data

110

on foreign direct investment comes from the World Bank’s World Development
Indicators (1997). I also use the weak link operationalization of this concept.

Power preponderance is the ratio of the larger state’s Correlates of War
Composite Capabilities (COWCAP) index over the smaller state’s COWCAP
index. Higher values on this measure represent less power parity.36 Given the
arguments of power transition theory and previous empirical ﬁndings, I expect
this variable to be negative.

I also use the Correlates of War data to measure the existence of an
alliance in a dyad. The COW project classiﬁes alliances into four types: no
alliance, entente, non-aggression pact, and mutual defense pact. l dichotomize
this scheme, coding ententes, non-aggression and mutual defense pacts as
alliances. If none of these types of alliances exists in a dyad, the variable takes
on a value of 0.

Contiguity is a dichotomous variable, taking a value of 1 when two states
share a land border or are separated by less than 150 miles of water. I measure
distance as the great circle distance between each state’s capital.

Since I analyze all dyads, as opposed to only including politically relevant
dyads, I include a variable for major power dyads. Owing to greater logistic
capabilities and international commitments, major powers are more likely than
other states to become involved in militarized disputes. This variable takes on a
value of 1 when at least one state in a dyad is a major power, 0 othenNise. The

COW project classifies the following states as major powers for this period: the

 

3" I use Bennett and Stam’s (2000) EUGene program (version 2.101) to generate the COWCAP, alliance,
contiguity, distance, and major power status data.

111

United States, the Soviet Union/Russia, France, Britain, China, Germany after

1990, and Japan after 1990.

Empirical Results

First, I analyze the inﬂuence of dyadic institutional similarity on conﬂict
onset. The empirical results in Table 1 support hypotheses 1a and 1b (see
Model 1a in Table 1). Consistent with Werner (2000), I ﬁnd that higher levels of
dyadic political similarity reduce the likelihood of a militarized interstate dispute
(MID) occurring. Unlike previous research, however, I argue that the similarity of
economic institutions in a dyad also inﬂuences the onset of MIDs. The results of
Model 1a support this expectation. The greater the similarity of economic
institutions in a dyad, the less likely that dyad is to experience a MID. Moreover,
increasing the similarity of economic institutions in a dyad by one standard
deviation reduces the likelihood of a militarized dispute by 57%, while a similar
increase in the similarity of political institutions only reduces the likelihood of a
dispute by 23% (see Table 2). What is perhaps most important is that the
independent effects of both political similarity and economic similarity are
statistically signiﬁcant. Models omitting economic similarity, then, may have
biased results due to model misspeciﬁcation.

The other variables in Model 1a generally perform as expected. Power
preponderance reduces the likelihood of conﬂict onset. This is especially
important as it is one of the two central concepts identiﬁed in the power transition
research program. Contiguous dyads are more likely to experience conﬂict than

non-contiguous dyads. Similarly, the greater the distance between two states,

112

the less likely they are to experience militarized conﬂict. As expected, major
power dyads are also more likely than other dyads to experience militarized
conﬂict. Foreign direct investment is signiﬁcant and negative, yet trade
interdependence is not statistically signiﬁcant. These results suggest that a
liberal peace may be more a function of institutions and foreign direct investment
than trade interdependence. Finally, the allies variable is not statistically
signiﬁcant.

An alternative explanation of the above results is that political and
economic similarity are signiﬁcant due to the omission of a joint democracy
variable. In other words, only the portion of the political and economic similarity
variables that includes joint democracies may be signiﬁcantly inﬂuencing the
onset of disputes. I examine this argument by including a control variable for
joint democracy (see Model 1b). Including the joint democracy variable does
wash out the effect of the political similarity variable. Jointly democratic dyads
are less likely than other dyads to experience a militarized dispute, but the
inﬂuence of political similarity does not extend to other regime types. Economic
similarity, however, is statistically signiﬁcant and negative. In other words, the
greater the similarity between two states economic institutions, the less likely
they are to experience militarized conﬂict. As indicated in Table 2, a one
standard deviation increase in economic similarity reduces the likelihood of a
dispute by 60%.

Next, I analyze the inﬂuence of joint systemic satisfaction on interstate

conﬂict. The empirical results in Table 3 support hypothesis 2b, but not

113

hypothesis 2a. As anticipated, dyads in which both states have economic
institutions similar to the economic institutions of the system leader are less likely
to experience a militarized dispute than dyads where the economic institutions
are more different from the institutions in the dominant state. The similarity of
political institutions with the dominant state is also statistically signiﬁcant and
negative. As before, jointly democratic dyads are less likely than other types of

dyads to experience a militarized dispute. Still, the effects of economic and

political institutional similarity hold even in the presence of joint democracy.

4'

Further, increasing the similarity of the economic institutions in a dyad with
the system leader’s institutions leads to a signiﬁcant reduction, 55%, in the
likelihood of conﬂict. This effect is greater than a similar change in the similarity
of political institutions, power preponderance, orjoint democracy. In brief,
economic institutions matter.

While economic institutions matter, the effect of the economic
interdependence is less clear. Trade interdependence is not signiﬁcant.
However, foreign direct investment is signiﬁcant and negative. Greater levels of
foreign investment reduce the likelihood of conﬂict onset. Moreover, the effect of
foreign direct investment is substantive. A one standard deviation increase in
foreign direct investment reduces the likelihood of conﬂict by about 20%. The
other control variables in Model 2 perform similarly to the control variables in
Model 1. Contiguity, distance, major power status, and allies are all statistically
signiﬁcant, and exert the expected effect. Power parity, contiguity, smaller

distances, and major power involvement all increase the probability of a dispute

114

 

occurring. Again, the allies variable is not statistically signiﬁcant. As before, this

may suggest a temporal dimension to the inﬂuence of this variable.

Dyadic Satisfaction, Systemic Satisfaction, and Types of Conﬂict

Outside of research on territorial disputes, few studies of interstate conﬂict
consider the issues at stake. In this section, I explore the relationship between
institutional similarity and types of militarized conﬂict. In particular, I investigate
three questions. First, do political and economic institutional similarity reduce the
likelihood of similar types of conﬂict? Second, does dyadic similarity affect one
type of conﬂict while systemic similarity affects another type of conﬂict? Third,
are some types of militarized disputes less likely with high political or economic
similarity than other types of militarized disputes?

A primary reason most research on interstate conﬂict does not analyze the
issues under dispute is that we lack a comprehensive and sound dataset of the
reasons for militarized disputes. The Correlates of War project includes a
variable, revision type, for the issue at stake in a dispute. Unfortunately, this
variable is coded into four very general categories: territorial, policy, regime
change, and other. For present purposes, this coding scheme does not permit a
clear identiﬁcation of disputes for political versus economic reasons. For
instance, a dispute classiﬁed as territorial may be motivated to free members of a
similar ethnic group from oppression or a territorial dispute may be motivated to
acquire natural resources. The ﬁrst is an example of a politically motivated

dispute, while the second is an example of an economically motivated dispute.

115

 

 

Nonetheless, an analysis of institutional similarity and types of dispute may
provide some preliminary answers to the questions above.

To simplify the analysis, I ﬁrst dichotomized the institutional similarity
variables into a high similarity category and a low similarity category. A dyad is
coded as highly similar if the value on the institutional similarity variable is greater
than or equal to one standard deviation above the mean, othenlvise it falls into
the low similarity category . This results in four new dummy variables: high
dyadic political similarity, high dyadic economic similarity, high systemic political
similarity, and high systemic economic similarity.

When comparing political and economic institutional similarity, I ﬁnd that
dyads with highly similar economic institutions experience fewer territorial
militarized disputes than dyads with similar political institutions (see Table 4).
This holds at both levels of analysis. Dyadic economic institutional similarity
results in fewer Mle than dyadic political similarity, and systemic economic
similarity results in fewer MIDs than systemic political similarity. In an analysis of
territorial disputes, Huth found that “economically valuable bordering territory . .
.was associated with a higher probability of a territorial dispute” (1996: 75). It is
possible that more economically similar dyads are satisﬁed with each other
economically, and thus have fewer incentives for a territorial dispute. Similarly,
systemic economic similarity implies greater satisfaction with the economic status
quo; thus explaining why these dyads are less likely to experience territorial

disputes. To verify these suppositions, one would have to examine the

116

 

 

relationship between institutional similarity and disputes over economically
valuable territory.

The difference between economic and political institutions with respect to
disputes over policy issues is less clear. As displayed in Table 4, economic
similarity at the dyadic level results in slightly fewer disputes, 16, than political
similarity, 19. However, economic similarity at the systemic level is associated
with more disputes, 38, than political similarity at the systemic level, 13.

Table 4 also reveals that fewer territorial disputes occur under high
systemic institutional similarity than under high dyadic institutional similarity. At
the systemic level, there are 10 territorial disputes in dyads with high political
similarity, but at the dyadic level there are 13 territorial disputes under high
political similarity. The results are similar for economic institutions. Systemic
economic similarity is associated with only 2 disputes, while dyadic economic
similarity is associated with 10 disputes. This is consistent with power transition
theory. Territorial disputes are clear attempts to change the status quo. Insofar
as systemic similarity indicates greater satisfaction with the status quo, then
systemic similarity should be associated with fewer territorial disputes.

Finally, Table 4 affords a preliminary answer to whether or not some types
of militarized disputes less likely with high political or economic similarity than
other types of militarized disputes? Where the ﬁrst two questions call for a
comparison of the rows in Table 4, this question requires a comparison of the
columns within a single row in Table 4. It is evident in Table 4 that the most

common type of militarized dispute associated with institutionally similar dyads is

117

 

 

a policy issue. This applies for both economic and political similarity, as well as
at the systemic and dyadic levels of analysis.

Finally, militarized disputes where a regime change is the dominant issue
are highly unlikely to occur under conditions of high institutional similarity. At a
minimum, this offers face validity to the institutional measures, for a regime
change motivation should be unlikely when the institutions are similar. Still, one
may ask why there are any cases with institutional similarity and a regime
change was the central issue? Further analysis shows that a regime change
motivation is only present when either the political or economic institutions are
not highly similar. In other words, there are no cases in which a regime change
militarized dispute occurs when both the economic and political institutions are

dyadically similar.

Conclusion

Where most research on interstate conﬂict focuses on the opportunity
dimension of conﬂict, this paper analyzes the willingness component. Contrary
to realist theories, which imply that the anarchical structure of the international
system creates an ever present willingness for conﬂict and the primary constraint
on interstate conﬂict is the balance of power, I argue that the motivation for
conﬂict varies across dyads. Satisﬁed dyads are less likely to engage in
militarized disputes than dissatisﬁed dyads. This argument not withstanding, the
primary contributions of this paper lie in addressing three more speciﬁc research

questions. First, what factors contribute to satisfaction? Second, does it matter if

118

 

 

we focus on satisfaction with a dyadic status quo or satisfaction with the systemic
status quo? Third, does satisfaction reduce the likelihood of all types of
militarized disputes or are some types of disputes more likely with satisfaction
than other types of disputes?

With respect to the ﬁrst question, I argue that satisfaction is primarily

about preference similarity. In turn, preference similarity is largely a function of

the similarity of domestic institutions. Unlike previous research, I argue that there

are two relevant dimensions to satisfaction, a political component and an
economic component. As such, I create measures of satisfaction based on both
political and economic institutions. Given the theoretical uncertainty regarding
the appropriate level at which to measure satisfaction with the status quo, l
construct measures for both the dyadic and systemic levels.

An empirical analysis of the onset of militarized disputes over the period
1982 to 1992 supports the theoretical argument. Institutional similarity, at either
the dyadic or systemic levels, reduces the likelihood of militarized interstate
disputes. Perhaps most importantly, economic institutional similarity increases
the likelihood of peace in a dyad more than political similarity.

On the third question, I ﬁnd that the strongest effects of institutional
similarity are to reduce the likelihood of militarized disputes over regime changes.
I also ﬁnd that economic similarity tends to reduce territorial disputes more than
political similarity.

The results of this research are consistent with Fukuyama’s argument on

the end of history. Political and economic institutions are a reﬂection of a state’s

119

1

FA

 

ideology, and ideologically similar dyads are less likely to experience conﬂict than
non-ideological similar dyads. Nonetheless, the results of this research should
be viewed as preliminary. First, the empirical domain of the analysis is limited to
ten years. It is important to test the sensitivity of these results over additional
time periods. Second, the analysis of dispute types employs an unreﬁned
measure of the issue at stake in a dispute. As further issue data sets become

available, it will be important to re-examine the ﬁndings of this research.

120

Chapter 5

Conclusion

In this concluding section, I brieﬂy review the contributions of this research
to our understanding of world politics. I also apply insights from this research to
the topic of terrorism. Too often academic work does not shed insights on real
world events, and no topic is more relevant at this time than terrorism. While this
dissertation is not on terrorism, it is about interstate conﬂict. As such, it should
provide some understanding of this issue. Finally, I conclude with some
observations on how to extend this research in the future.

The aim of this dissertation is to improve our understanding of the effects
of globalization on international relations. The phenomenon of globalization has
two dimensions: economic and political/cultural. The economic effect of
globalization is to increase the economic connectedness between nations. “Over
the past decade, trade has been growing twice as fast as output and foreign
direct investment three times as fast” (Economist, September 28, 1996). The
political/cultural effect of globalization is to increase the number of states with
liberal political and economic systems. The globalization process, or using
Richard Rosecrance’s term virtualization process, “requires an effective
regulatory state . . . This means that essentially Western commercial codes, legal
SYStems, and relatively incorruptible political practices should be emulated in

Other regions of the world” (Rosecrance, 1999: 90). This dissertation explores

121

 

the inﬂuence of both of these factors, economic connectedness and institutional
similarity, on interstate conﬂict.
In summary form, the major arguments and ﬁndings of this dissertation are

as follows. I discuss each in turn.

1. Economic interdependence includes both trade and foreign direct investment.
2. To analyze competing arguments concerning economic interdependence, it is
necessary to disaggregate trade into strategic and non-strategic commodities.

3. Non-strategic trade interdependence reduces dyadic conﬂict.

4. The effect of strategic trade interdependence is conditional on the political
relationship in the dyad. Dyads lacking political afﬁnity with high levels of
strategic trade interdependence are more likely than other dyads to experience
militarized disputes.

5. Foreign direct investment reduces the likelihood of dyadic conﬂict.

6. States dependent on importing both strategic and non-strategic commodities
are less likely than other states to initiate militarized disputes.

7. Highly militarized states dependent on importing strategic commodities are
more likely than other states to initiate militarized disputes.

8. Democratic institutions reduce the likelihood of dispute initiation.

9. Weak, that is economically under-developed, states dependent on exporting
strategic commodities are more likely than other states to be targets of militarized

disputes.

122

 

10. Higher levels of foreign direct investment reduce both the likelihood of a
state initiating a dispute and the likelihood that a state will be the target of a
militarized dispute.

11. A state’s ideology and foreign policy preferences are a reﬂection of the
political and economic institutions in a state.

12. States with similar political and economic institutions are less likely to
experience militarized conﬂict.

13. The strongest effect of institutional similarity is to reduce disputes over
regime changes.

14. States with similar economic institutions are less likely to experience

disputes over territory than states with similar political institutions.

In chapter two, I analyze the relationship between economic
interdependence and militarized interstate conﬂict. I argue that previous studies
examining the relationship between economic interdependence and international
conﬂict are ﬂawed because they fail to disaggregate trade into strategic and non-
strategic commodities and neglect the role of foreign direct investment. The
incorporation of these additional variables also permits a more rigorous research
design in that I am able to address competing arguments.

I argue that non-strategic trade interdependence reduces the likelihood of
conﬂict between states because it increases the opportunity costs of conﬂict. In
one sense, the inﬂuence of non-strategic trade interdependence is similar to the

inﬂuence of military power on interstate dynamics. Both increase the costs of

123

 

conﬂict, thereby constraining the actions of state leaders. But, trade does not
always promote peace. I also posit that strategic trade interdependence
increases the likelihood of militarized conﬂict between states. This effect,
however, is conditional on the political afﬁnity of the states in the dyad. The
reason for the difference between non-strategic and strategic trade is that
strategic trade increases a state’s vulnerability and may be capturable; thus,
strategic trade provides incentives for conﬂict that outweigh the disruption that
conﬂict creates. Put differently, political leaders aim to retain their position in
ofﬁce. To do so, they must protect their state's security and enhance its welfare.
States dependent on strategic commodities face both a security and welfare
vulnerability. If the importing state feels the commitment to provide the goods is
weak, and it will have this perception when there are different foreign policy
preferences, then it has an incentive to enter into conﬂict.

In addition to trade interdependence, I argue that economic
interdependence also involves foreign direct investment. Like trade
interdependence, foreign direct investment also exercises a peaceful inﬂuence
on interstate relations. Since conﬂict not only increases the costs of trade
transactions, but also the risk, and in turn cost, of investment in another state,
higher levels of foreign investment increase the opportunity costs of conﬂict.
Foreign investment also contributes to peace between nations because it
decreases the potential beneﬁts of conquest. Under foreign direct investment,
the investing nation is acquiring access to goods in the recipient nation so it has

less incentive to pursue a policy of conquest to attain these same goods. In

124

 

short, foreign direct investment differs from strategic trade interdependence in
that it reduces the incentives for conﬂict.

An empirical analysis of the onset of militarized interstate disputes for all
dyads over the period 1970 to 1992 supports these arguments. I ﬁnd that non-
strategic trade interdependence does reduce the likelihood of conﬂict onset,
though its effect is not statistically signiﬁcant. However, the effect of strategic

trade interdependence is signiﬁcant and positive. Dyads with strategic trade and

 

lacking in political afﬁnity are more likely than other dyads to experience
militarized disputes. I also ﬁnd support for the paciﬁc beneﬁts of foreign
investment. The more foreign direct investment in a state, the less likely the
state is to experience a militarized dispute.

Nonetheless, these ﬁndings should be viewed as tentative. First, the
empirical domain is limited. To the extent systemic forces inﬂuence the onset of
militarized disputes and to the extent cycles of conﬂict characterize world politics,
ﬁndings based on a single twelve year period can only provide limited support.
Second, the arguments concerning foreign direct investment require dyadic data.
Dyadic investment data, however, does not exist for a large number of countries.
As such, this analysis uses state-level investment data. The ﬁnding that states
with higher levels of foreign direct investment reduce the likelihood of dyadic
conﬂict is consistent with, even if it is not a direct test of, the arguments set forth
here.

lmportantly, the statistical signiﬁcance of these ﬁndings are present even

when other relevant control variables are included. Moreover, the empirical

125

results are robust to alternative operationalizations of all the central theoretical
variables.

In chapter three, I apply the arguments concerning the differences
between strategic and non-strategic goods to state level behavior. This is
important for two reasons. First, the arguments of theoreticians like Paine and
Rosecrance emphasize the actions of the state. To evaluate those arguments,
then, it is necessary to conduct a monadic analysis. Second, scientiﬁc ﬁndings
always contain a certain amount of uncertainty. Yet, if a theoretical argument
contains implications for multiple levels of analysis and we ﬁnd empirical support
for all of the implications, then we can express more conﬁdence in the theory.

Because conﬂict requires at least two actors, state level arguments are
difﬁcult to assess. In particular, one cannot simply examine conﬂict participation,
for we expect a state to respond if attacked. When discussing state level
arguments, it is more appropriate to distinguish between the initiation of and
participation in conﬂict. I also distinguish between initiating a dispute and being
the target of a militarized dispute.

In chapter three, I argue that non-strategic and strategic import
dependence, in general, reduce the likelihood that a state will initiate a militarized
dispute. The reason is straightforward. States dependent on trade have more to
lose if they enter into conﬂict with other states, even if the other disputant is not a
trading partner. Conﬂict increases the costs of conducting business, so states
with the most commerce have the most to lose. Nonetheless, states aim to

alleviate dependence on strategic commodities. Since political leaders aim to

126

 

retain ofﬁce and it is politically beneﬁcial to reduce one’s dependence on key
commodities, militarized states dependent on importing strategic commodities
are more likely than other states to initiate militarized disputes.

An empirical analysis of militarized interstate disputes over the period
1970 to 1992 supports the arguments concerning the differential effects of non-
strategic and strategic trade interdependence. In addition, I ﬁnd that foreign
direct investment reduces the likelihood of a state initiating a dispute.

In chapter three, I also examine which states are more likely to be targets
of militarized disputes. The conditional economic dependence argument laid out
above suggests speciﬁc expectations regarding the targets of disputes. In
particular, I expect that weak states exporting strategic commodities are more
likely than other states to be the targets of aggressive foreign policies by other
states. These states have something others want, and if they are not able to
protect their resources, someone else is likely to attack them. On the other hand,
foreign direct investment reduces the likelihood that a state will become the
target of a militarized dispute. Foreign direct investment reduces the likelihood of
being a target for two reasons. First, states investing in them have less incentive
to attack since they are already extracting wealth. Second, states attracting
investment have benefactors who do not want their investments jeopardized,
thus deterring other, non-investing, states from attacking them. The empirical
ﬁndings support these arguments.

In chapter four, I focus on the political/cultural dimension of globalization.

Globalization is not only increasing economic connectedness between nations, it

127

 

 

is also spreading liberal ideology. Whereas previous research only emphasizes
the relationship between a state’s political institutions and its preferences, I argue
that a state’s ideology is a reﬂection of its political and economic institutions.
States with similar domestic institutions will have similar foreign policy
preferences, be more satisﬁed with the status quo, and less likely to enter into
military conﬂict. Accordingly, I construct two new measures of satisfaction, one
based on the similarity of political institutions across states and the other based
on the similarity of economic institutions. I also distinguish between dyadic and
systemic satisfaction.

After examining the onset of militarized disputes over the period 1982 to
1992, I conclude that institutional similarity at either the dyadic or systemic levels
reduces the likelihood of a dispute occurring. Perhaps most importantly,
economic institutional similarity is a stronger paciﬁc inﬂuence than political
similarity.

In chapter four, I also examine the relationship between institutional
similarity and different types of militarized disputes. I ﬁnd that the strongest
effects of institutional similarity, whether political or economic, are to reduce the
likelihood of militarized disputes over regime changes. I also ﬁnd that economic

similarity tends to reduce territorial disputes more than political similarity.

Economic Dependence, Institutional Similarity, and State Sponsored
Terrorism

128

 

 

The events of September 11, 2001 have brought the issue of terrorism,
and state sponsors of terrorism, to the forefront of international relations. This
dissertation is not about terrorism, but it is about interstate conﬂict. As such, this
research should provide some insights on the issue of state-supported
terrorism.37 A central argument of this research is that economic dependence

encourages more peaceful foreign policies. Clearly, state sponsors of terrorism

_“‘I’
3..
I

are not pursuing peaceful foreign policies. This dissertation, then, expects that

states sponsoring terrorism should exhibit low levels of economic dependence.

i'Vd ~

All the states identiﬁed by the United States Department of State as sponsors of
terrorism exhibit trade dependence below the mean value of trade dependence
for the international community as a whole.38 In four of the states, their economic
dependence is more than one standard deviation below the mean for the rest of
the world. Each of these states is also far below the world mean in terms of

foreign direct investment.39

The lack of trade and investment dependence in
these states is consistent with the argument of this dissertation.
In chapter four, I argue that institutional similarity taps into satisfaction with

the status quo and a willingness for conﬂict. Insofar as this is accurate, we

should expect the states labeled as “sponsors" of terrorism to be dissatisﬁed,

 

37 The United States Department of State identiﬁes the following as states sponsoring terrorism: Cuba,
Iran, Iraq, Libya, North Korea, Sudan, and Syria. This list has not changed since 1993. In the next version
of this report, Afghanistan is sure to be listed as well.

38 Iran and Sudan have trade dependence scores of 22.92 and 21.22. Iraq’s trade dependence is 50.85.
Syria’s trade dependence is 53.28, and assuming minimal trade for states not reporting to international
agencies gives Cuba, Libya, North Korea, and Afghanistan trade dependence scores near zero. The mean
of trade dependence is 61.55, and the standard deviation is 41.28.

39 The World Bank reports that Cuba, Iraq, and Sudan had no foreign direct investment. The foreign direct
investment as a percentage of a state’s gross domestic problem in Iran and Libya was negative, -.3006 and -
.306 respectively. Syria’s foreign investment was .5768. There is no foreign direct investment information
available for North Korea and Afghanistan. The world mean for foreign investment is 1.20 with a standard
deviation of 2.99.

129

 

meaning their institutions are signiﬁcantly different than the institutions of the
United States. The following are the political and economic similarity scores
between the United States and each state sponsor of terrorism: USA and Cuba,
.705 and -.400; USA and Libya, .705 and -.573; USA and Sudan, .705 and -.693;
USA and Iran, .955 and -.585; USA and Iraq, .842 and —.684; USA and Syria,
.842 and -.587; USA and North Korea, .842 and -.453; USA and Afghanistan,
.878. The mean political similarity score for all dyads is 1.128 with a standard
deviation of .333, while the mean economic similarity score is -.239 with a
standard deviation of .182. All states sponsoring terrorism are below the mean
value on both the political and economic measures, most by more than one
standard deviation. This suggests these states are not satisﬁed with the status

quo.

Future Research

While this dissertation improves our understanding of the effects of
globalization on international relations, more research is still necessary. In
particular, the discussion in chapters two and three emphasizes an opportunity
cost argument in explaining the paciﬁc beneﬁts of economic interdependence.
Others have placed more emphasis on the signaling aspects of economic
interdependence. In future research, it would be useful to distinguish between

the constraint and signaling explanations.

130

 

 

 

The discussion in chapter three on measuring the militarization of a state
suggests an additional research enterprise. What are the determinants of major
power status? Presently, we lack a deﬁnition of major powers. The low
correlation between militarization and major powers suggests that major power
status includes more than military power.

Finally, the discussion in chapter four underscores the importance of
ideology and the similarity of ideology between states for understanding
motivations for militarized conﬂict. However, the present indicators of political,
and to a lesser extent economic, similarity are crude. Distinguishing between
autocracies may be especially helpful. Presently, monarchies and military
dictatorship are treated similarly, yet monarchies may be more peaceful toward
other monarchies than they are toward military dictatorships. Addressing this
issue may shed light on the causes of the democratic peace. Some scholars
suggest that the democratic peace is a result of shared preferences among
democracies. If this is true, then we are likely to observe peace in autocratic
dyads of similar regime type. Others argue that democratic institutions facilitate
the sending and receiving of costly signals, permitting democracies to resolve
disputes at lower levels of hostility. If this is the case, then we are not likely to

observe a separate peace in autocratic dyads with similar political institutions.

131

APPENDICES

132

APPENDIX A

Table 2.1: NBER SITC Trade Categories

 

SITC one-digit commodity group

Description

 

—_1

Food and live animals

 

 

 

 

 

 

 

 

 

 

80

S 1 Beverages and tobacco

$2 * Crude materials except fuels

S3 * Mineral fuels, lubricants

84 Animal, vegetable oil, fat

S5 Chemical related products

86 * Basic manufactures

S7 Machines, transport equipment

88 * Miscellaneous manufactured goods
39 Goods not classiﬁed

 

 

* Indicates Strategic Commodities

133

 

Table 2.2: Total Dyadic Trade Interdependence and Militarized Conﬂict,
1970-1992

 

 

 

 

 

 

 

 

 

 

 

 

Variable [3 Change in Predicted

s.e. Probability of Event
Occurrence

Total Dyadic Trade - 48.454 -15.67 %
32.362

Democracy -.045 ** -22.15 %
.020

Power Parity -.168 ** -27.76 °/o
.070

Allies .407
.298

Contiguity 3.665 ***
.261

Ln Distance -.261 ***
.105

Dyad with a Major Power 2.000 ***
.257

Constant -3.010 ***
1.081

N 103386

Wald Chi-Square 779.17 ***

 

 

 

 

*** p < .01, ** p < .05 All signiﬁcance tests are one-tail.

Note: The ﬁrst differences in column 3 reﬂect a one standard deviation increase
in each variable, while all other variables are held at their mean or modal values.

134

 

Table 2.3: Disaggregated Trade, Foreign Direct Investment and Militarized

Conﬂict, 1970-1992

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable [3 Change in Predicted

s.e. Probability of Event
Occurrence

Non-Strategic Trade -955.574 -15.80 %
82.766

Strategic Trade 5.015 ** 14.95 %
2.300

FDI -.063 *** -9.39 %
.023

Afﬁnity -2.112 ** -26.17 %
.954

Democracy -.036 -18.39 °/o
.023

Power Parity -.214 ** -33.92 %
.088

Allies .378
.350

Contiguity 3.522 ***
.353

Ln Distance -.344 **
.153

Dyad with a Major Power 2.202 ***
.356

Constant -2.363
1.448

N 64249

Wald Chi-Square 573.73 ***

 

 

 

*** p < .01, ** p < .05 All signiﬁcance tests are one-tail.

Note: The ﬁrst differences in column 3 reﬂect a one standard deviation increase
in each variable, while all other variables are held at their mean or modal values.

135

 

 

 

Table 2.4: Conditional Trade Interdependence Model, 1970-1992

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable [3 Change in Predicted

s.e. Probability of Event
Occurrence

Non-Strategic Trade -70.316 -11.88 %
79.246

Strategic Trade 13.709 *** 46.33%
3.240

FDI -.062 *** -9.29 °/o
.023

Afﬁnity -1.820 ** -23.00 %
.965

Strategic Trade*Afﬁnity -20.219 ** -14.73 "/0
8.929

Democracy -.034 -17.32 %
.023

Power Parity -.190 ** -30.74 °/o
.090

Allies .292
.347

Contiguity 3.522 ***
.351

Ln Distance -.370 ***
.152

Dyad with a Major Power 2.186 ***
.356

Constant -2.385
1.450

N 64249

Wald Chi-Square 678.31 ***

 

 

 

*** p < _01, ** p < .05 All signiﬁcance tests are one-tail.

Note: The ﬁrst differences in column 3 reﬂect a one standard deviation increase
in each variable, while all other variables are held at their mean or modal values.

136

 

 

 

Table 2.5:
Disputes, 1970-1992

Conditional Trade Interdependence and Fatal Militarized

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable [3 Change in Predicted

s.e. Probability of Event
Occurrence

Non-Strategic Trade - 156.756 -37.22 %
141.667

Strategic Trade (SITC 3) 41.773 ** 126 %
24.928

Afﬁnity -.261 -3.68 %
1.297

Strategic Trade*Afﬁnity - 187.668 * -83.61 %
1 17.369

Democracy -.040 -19.82 %
.042

Power Parity -.170 -28.02 %
.123

Allies .406
.521

Contiguity 3.674 ***
.437

Ln Distance -.264 *
.171

Dyad with a Major Power 1.618 ***
.457

Constant - 6.270 ***
1.764

N 103360

Wald Chi-Square 309.03 ***

 

 

 

*** p < .01, ** p < .05 All signiﬁcance tests are one-tail.

Note: The ﬁrst differences in column 3 reﬂect an increase of one standard

deviation for each variable.

137

 

 

Table 2.6: Conditional Trade Interdependence, FDI and Fatal Militarized

Disputes, 1970-1992

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable [3 Change in Predicted

s.e. Probability of Event
Occurrence

Non-Strategic Trade - 83.388 -21.94 °/o
104.631

Strategic Trade (SITC 3) 42.181 129 %
24.481

FDI -.092 ** -13.49 %
.050

Affinity .978 15.09 %
1.976

Strategic Trade*Afﬁnity - 179.501 -77.09 %
137.823

Democracy -.015 -8.24 %
.042

Power Parity -.031 -5.87 %
.155

Allies -.369
.856

Contiguity 3.709 ***
.735

Ln Distance -.601 **
.343

Dyad with a Major Power 1.635 ***
.621

Constant - 5.119 **
2.963

N 64232

Wald Chi-Square 232.68 ***

 

 

 

*** p < .01, ** p < .05 All signiﬁcance tests are one-tail.

Note: The ﬁrst differences in column 3 reﬂect an increase of one standard

deviation for each variable.

138

 

 

 

Table 3.1: Economic Dependence and the Initiation of Militarized Interstate

Disputes: An Event Count Analysis

 

 

 

 

 

 

 

 

 

 

 

 

MID Initiation Model Changes in Relative
Variable 1a, 1970-1992 Risk of MID Initiation
B
s.e.
Non-Strategic - 2983.408 *** - 65 %
Import Dependence 409.288
Strategic Import - 10192.830 *** - 65 %
Dependence 1854.765
Strategic Import 400734 *** 29 %
Dependence * 95830.230
Militarization
Foreign Direct -.223 *** - 48 %
Investment .023
Militarization 29.771 *** 28 %
2.438
Democracy -0.072 *** -26 %
0.012
Economic 0.370 * 98 %
Development 0.032
Major Power 0.364 ***
0.125
Constant -.814 ***
0.081
N 2080

 

 

 

 

 

Note: *** p < .001: ** p < .05; * p < .10. All signiﬁcance tests are one-tailed.
Changes in relative risk reﬂect a one standard deviation increase for the interval

level variables from a baseline model where all interval level covariates are set at
their mean value and categorical variables are set at zero.

139

 

Table 3.2:

Developed Democracies, Economic Dependence, and the
Initiation of Militarized Interstate Disputes: An Event Count Analysis

 

 

 

 

 

 

 

 

 

 

 

 

 

MID Initiation Model Changes in Relative
Variable 1b, 1970-1992 Risk of MID Initiation
l3
s.e.
Non-Strategic - 1928.983 *** -49 %
Import Dependence 205.138
Strategic Import -17857.94O *** -84 %
Dependence 3329.631
Strategic Import 564752 *** 42 %
Dependence * 124477
Militarization
Foreign Direct - .208 *** ~45 %
Investment .020
Militarization 30.455 *** 29 %
3.022
Democracy -.373 *** ~79 %
0.122
Economic 0.502 *** 152 %
Development 0.049
Development*Demo -0.577 *** -40 %
cracy 0.099
Major Power 1.095 ***
0.109
Constant -O.900 ***
0.091
N 2080

 

 

 

 

 

Note: *** p < .001: ** p < .05; * p < .10. All signiﬁcance tests are one-tailed.
Changes in relative risk reflect a one standard deviation increase for the interval

level variables from a baseline model where all interval level covariates are set at
their mean value and categorical variables are set at zero.

140

 

Table 3.3: Economic Dependence and Targets of Militarized Interstate
Disputes: An Event Count Analysis

 

MID Targeting Changes in

 

Variable Model 2, 1970- Relative Risk of
1992 MID Initiation
B
s.e.
Non-Strategic Export 657.684 13 %
Dependence 459.200
Strategic Export 1407.171 *** 26 %
Dependence 176.244
Strategic Export -789.585 *** - 20 %
Dependence * 152.941
Development
Foreign Direct Investment -.084 *** -22 %
.007
Militarization 17.858 *** 16 %
5.484
Democracy -.030 -12 %
.027
Economic Development - .048 -8 %
.055
Major Power .115
.164
Constant -1.716 ***
_ , , 7 2-, f .189? i g? __g g
N 2016

 

Note: *** p < .001: ** p < .05; * p < .10. All signiﬁcance tests are one-tailed.
Changes in relative risk reﬂect a one standard deviation increase for the interval

level variables from a baseline model where all interval level covariates are set at
their mean value and categorical variables are set at zero.

141

b;

 

Table 3.4: Initiation of Militarized Disputes With An Alternative Measure of

Militarization

 

 

 

 

 

 

 

 

 

 

 

 

Variable B
s.e.
Non-Strategic Export Dependence -1598.84 ***
247.76
Strategic Export Dependence -15803.73 ***
4297.10
Strategic Export Dependence * 52181 ***
Development 10085
Foreign Direct Investment -.206 ***
.020
Militarization 2.68 ***
.282
Democracy -.101 ***
.01 1
Economic Development .348 ***
.035
Major Power -.282 ***
.078
Constant _.915 H.
.112
N 2030

 

 

Note: *** p < .001: ** p < .05; * p < .10

. All signiﬁcance tests are one-tailed.

142

 

 

 

Table 4.1: Dyadic Institutional Similarity and Militarized Interstate Conﬂict,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1982-1992
Variable Model 1a Model 1b
l3 l3
s.e. s.e.
Dyadic Political Similarity -.786 * -.101
.452 .445
Dyadic Economic Similarity -2.495 ** -2.594 **
.822 .860
Joint Democracy -.109 *
.060
Trade Interdependence -20.447 1 1.900
32.820 26.083
Foreign Direct Investment -.1953 -.180 **
.035 .035
Power Preponderance -.246 * -.242 *
.133 .130
Allies -.115 -.038
.352 .358
Contiguity 3.855 ** 3.776 **
.438 .443
Ln Distance -.385 ** -.383 **
.174 .176
Major Power Dyad 2.000 ** 2.076 **
.453 .420
Constant -3.075 ** -3.700 **
.1.455 1.491
N 26712 26627
Model chi-square 342.05 ** 347.52 **

 

 

 

** p < .01, * p < .05 one tail signiﬁcance test.

143

 

 

Table 4.2: Substantive Impact of Dyadic Satisfaction on Militarized Dispute

Onset

 

First Difference

Model 1a

Model 1 b

Model 2

 

Dyadic Political Similarity
Increased 1 standard
deviation

-23.17 %

- 3.31 °/o

 

Economic Similarity
Increased 1 standard
deviation

- 57.58 %

- 60.44 %

 

Joint Democracy
Increased 1 standard
deviation

- 26.55 %

- 32.50 %

 

Systemic Political Similarity
Increased 1 standard
deviation

-29.10%

 

Systemic Economic Similarity
Increased 1 standard
deviation

- 55.45 %

 

Power Preponderance
Increased 1 standard
deviation

- 19.94 %

- 37.24 %

- 38.25%

 

 

Foreign Direct Investment
Increased 1 standard
deviation

 

- 26.43 %

 

- 24.68 %

 

- 20.92 %

 

Note: Changes in predicted probabilities are based on a baseline model where
interval variables are set at their mean value and categorical variables are set at

the modal category.

144

 

 

 

Table 4.3: Systemic
Conﬂict, 1982-1992

Institutional Similarity and Militarized Interstate

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Model 2
B
s.e.
Systemic Political -1.086 **
Similarity .616
Systemic Economic -2.731 **
Similarity .987
Joint Democracy -.139 **
.059
Trade Interdependence 1 .101
20.652
Foreign Direct -.149 **
Investment .040
Power Preponderance -.250 **
.133
Allies -.425
.394
Contiguity 3.116 **
.456
Ln Distance -.406 **
.174
Major Power Dyad 2.569 **
.420
Constant -3.477 **
1.467
N 55726
Model Chi Square 580.00 **

 

 

145

 

 

 

Table 4.4: Institutional Similarity and Type of Militarized Interstate Dispute,

 

 

 

 

 

 

1 982-1 992

Institutional Number of Number of Number of

Similarity Territorial Policy Change Regime Change
MIDs Mle Mle

Dyadic Political 13 19 1

Similarity

Dyadic Economic 10 16 2

Similarity

Systemic Political 10 13 1

Similarity

Systemic 2 38 2

Economic

Similarity

 

 

 

 

 

146

5L

Appendix B

Model Derivations

A state’s Political Welfare Function = W(C, Z) where C represents total goods
and services consumption and 2 represents conﬂict toward a particular country.

More speciﬁcally, C approximates the national income of a state; thus
C=q_la'Xs—Xn+Md+Me (Eq-1)

Where q represents domestic consumption, Ia denotes foreign direct investment
in another nation, X8 and Xn signify strategic exports and non-strategic exports,
and Md and Me stand for strategic and non-strategic imports.

The portion of a state’s economy dependent on trade and international
investment can be represented as:

Rla + PXn + Asz- PMd — APMe (Eq. 2)

Where R.a signiﬁes the rate of return on an investment, PX", APXS, Pm, and APMe
signify the prices of non-strategic and strategic exports and imports. Notice that
the price of strategic exports and imports are modiﬁed by the political afﬁnity of
the trading states. In other words, the utility of strategic commodities is different
in a dyad with close political afﬁnity than it is in a dyad lacking political afﬁnity. In
politically dissimilar dyads, strategic goods are costly in both economic and
political terms, but in politically similar dyads, strategic goods still have an
economic cost but not a political cost. Assuming no balance of payments
problem, then equation 2 equals zero.

Insofar as conﬂict affects the cost of conducting economic transactions, then la,
X8, Xn, Md, and Me, are functions of conﬂict.

R|a = f(Z) (Eqs. 3.1-3.5)
PX" = 11(2)

APXs = “2)

PMd = HZ)

APMe = “2)

Further, since conﬂict increases transactions costs, it decreases the gains from

trade and investment. Thus, conﬂict raises import prices, lowers export prices,
and lowers the return on investments.

147

 

Riva = f' < 0 (Eqs. 4.1-4.5)
PX'n = h, < O

APX.s = j' < o
PM'd = kI > 0
APMve = I' > 0

An actor chooses a particular level of conﬂict, 2, that maximizes W (2; C). Stated
formally

Max W(Z, EC)
Z

Maximizing this function is subject to a balance of payments constraint.
Therefore, an actor maximizes the following Lagrangian (L):

(Eq. 5)

Differentiating L with respect to conﬂict, Z, yields a set of ﬁrst order conditions
(FOC) for optimal conﬂict.

FOCs:

aL/az = W’(Z; q — Ia — Xs — Xn + Md + Me) + k[la'(z) + X'nPn'(z) + X'sAPs' (z)
- M'de' (z)- M’eAPe' (2)] (Eq. 6)

due). = lar (z) + XnPn (z) + XsAPs (z) - Mde (z) — MeAPe (2) (Eq. 7)

To determine how conﬂict responds to changes in trade or investment, I derive
comparative static equilibria. To determine comparative static equilibria on a
system of equations, one has to take the total differential of the FOCs (equations
6 & 7) yielding a set of second order conditions (SOCs).

SOCs:

aL2/az2 =

w" (z; q — la — Xs — Xn + Md + Me) + Ml 6R2/622 (Z) + x5 aAPs2/822 (2) + Xn
aPnzlaZZ (Z) — Md an2/az2 (Z) — Me aAPeZ/az2 (2)] dz {Part A}

+ [awzlazaq (2; q — la — XS — Xn + Md + Me)] dq {Part B}

+ [- awzlazal (2; q — la — Xs — Xn + Md + Me) + A mm (2)] dl {Part C}

148

 

+ [- 6W2/aZc‘9Xs (Z; q — la — Xs — Xn + Md + Me) + A aAPs/aZ (Z)] dXs {Part D}
+ [- awzlazaXn (Z; q — la — Xs — Xn + Md + Me) + A aPn/aZ (Z)] an {Part E}
+ [aWZIaZBMd (Z; q - la — Xs — Xn + Md + Me) + 7t 6AMd/aZ (Z)] de {Part F}
+ [awzlazaMe (z; q — Ia — x5 — Xn + Md + Me) + 2. dMe/dz (2)] dMe {Part G}

+ [la + Xs + Xn - Md - Me] dx {Part H} (Eq. 8)

aLZIaitaz =

[I (BR/62 (q — Ia-Xs-Xn + Md + Me) +Xs aAPs/aZ (q— la—Xs—Xn + Md + Me)
+Xn aPn/aZ (q-la-Xs-Xn + Md + Me)-Md aMd/aZ (q- Ia—Xs-Xn + Md +
Me) — Me 6AMe/62 (q — Ia - Xs - Xn + Md + Me)] dZ {Part I}

+ R(Z) dl {Part J}

+ APs(Z) dXs {Part K}

+ Pn(Z) an {Part L}

- APd(Z) de {Part M}

- Pe(Z) dMe {Part N} (Eq. 9)

Next, I arrange the above equations into a linear system, isolating the two
variables, Z and it.

PartA PartH dZ _ qu — C d1 — DdXs - Ean + F de + GdMe
Part] 0 d}. 0 + JdI + KdXs + Lan — Mde — NdMe

To ﬁnd the effects of any particular parameter on Z, use Cramer’s Rule: Di] / D,
where D signiﬁes determinant.

Proposition 1a: In a dyad with negative political affinity, the greater the
amount of strategic exports, the more likely conﬂict is to occur.

Proposition 1b: In a dyad with positive political afﬁnity, the greater the amount of
strategic exports, the less likely conﬂict is to occur.

149

 

 

 

 

 

 

dZ/dXs = Par”) PM” PartA PartH
— PartK 0 Part] 0
= KH / -(|H)
= - (K / I)

= - [APs (Z)] I [I aR/az (q — la — Xs - Xn + Md + Me) + Xs aAPs/aZ (q — Ia
—Xs-Xn + Md + Me) +Xn aPn/az(q—la-Xs—Xn + Md + Me)-Md
aMd/aZ (q — Ia — Xs - Xn + Md + Me) - Me 6AMe/aZ (q — Ia — Xs - Xn +
Md + Me)]

Prices, e.g. Ps (Z), are always positive. However, A may be either negative or
positive depending on the political afﬁnity in the dyad. If the dyad has close
political afﬁnity, A is positive and the numerator is negative. If the dyad lacks
political afﬁnity, A is negative and the numerator is positive. Assuming a well-
behaved utility function, then the Hessian must be negative deﬁnite; thus, the
principle minors alternate sign making the above second order conditions
produce a positive denominator. Thus,

dZ/dXs < 0 ifA> 0
and
dZ/dXs > 0 ifA< 0

Proposition 1: QED.

Proposition 2: As the amount of non-strategic exports increase, conﬂict
decreases.

 

 

 

 

d 2! d Xn ___ PartE PartH PartA PartH
PartL 0 Part] 0
= LH / - IH
= - (K/l)

=-[Pn (Z)]/[I aRIaZ(q-la—Xs—Xn + Md + Me) +Xs aPsIaZ(q— Ia—
Xs-Xn+Md+Me)+Xn aPnlaZ(q-la—Xs-Xn+Md+Me)—Md
aMd/aZ(q—la-Xs—Xn + Md + Me)—Me aMe/aZ(q-la-Xs-Xn + Md
+Me)]

Prices, e.g. Ps (Z), are always positive, but multiplying this by a negative makes

the numerator negative. Assuming a well-behaved utility function, then the
Hessian must be negative deﬁnite; thus, the principle minors alternate sign

150

 

 

making the above second order conditions produce a positive denominator. The
whole term is then negative. Thus,

dZ/an < 0
Proposition 2: QED.

Proposition 3a: In a dyad with negative political afﬁnity, the greater the
amount of strategic imports, the more likely conﬂict is to occur.

Proposition 3b: In a dyad with positive political afﬁnity, the greater the amount of

 

 

 

 

strategic imports, the less likely conﬂict is to occur. I»
d 2] d M d = - PartF PartH PartA PartH in
PartM 0 Part] 0
= - MH / - lH
= M l l

= - APd (Z) / [I aR/az (q — la -Xs - Xn + Md + Me) + Xs aAPs/aZ (q — Ia —
Xs—Xn + Md + Me) +Xn aPn/aZ(q—la-Xs-Xn + Md + Me)—Md
aMd/aZ (q — Ia — Xs - Xn + Md + Me) - Me 6AMe/aZ (q — la — Xs — Xn +
Md + Me)]dZ

Prices, e.g. Ps (Z), are always positive. However, A may be either negative or
positive depending on the political afﬁnity in the dyad. If the dyad has close
political afﬁnity, A is positive and the numerator is negative. If the dyad lacks
political afﬁnity, A is negative and the numerator is positive. Assuming a well-
behaved utility function, then the Hessian must be negative deﬁnite; thus, the
principle minors alternate sign making the above second order conditions
produce a positive denominator. Thus,

dZ/de < 0 ifA > 0
and

dZ/de > 0 ifA< 0
Proposition 3: QED.

Proposition 4: As the amount of non-strategic imports increase, conﬂict
decreases.

151

 

 

 

 

 

 

d Z/ dM d = — PartG PartH PartA PartH
PartN 0 Part] 0
= - NH / - IH
= N / l

=-Pe(Z)/[laR/62(q—Ia—Xs-Xn+Md+Me)+XsaPs/az(q—la—Xs
—Xn + Md + Me) + Xn aPn/BZ (q — la—Xs-Xn + Md + Me)— Md aMd/az
(q-la-Xs—Xn+Md+Me)-MeaMe/aZ(q—la-Xs-Xn+Md+Me)]
dZ

The numerator is again negative, and making the same assumption as before
that the utility function is well behaved yields a positive denominator. Thus,

dZ/dMe < 0
Proposition 4: QED.

Proposition 5: As the amount of foreign direct investment increases,
conﬂict decreases.

PartC PartH
PartJ 0

/ PartA PartH
Part] 0

dZ/dl =

 

 

 

 

=JHl-IH
=-(J/l)

=-[R(Z)]/[IaR/aZ(q-la—Xs-Xn+Md+Me)+XsaPs/aZ(q—Ia—Xs
-Xn + Md + Me) +Xn aPnlaZ (q — la—Xs—Xn + Md + Me)- Md aMd/aZ
(q-Ia-Xs-Xn+Md+Me)-MeaMe/aZ(q—Ia—Xs-Xn+Md+Me)]

Assuming the return on investment is positive and multiplying it by a negative
makes the numerator negative. Assuming a well-behaved utility function, then
the Hessian must be negative deﬁnite; thus, the principle minors alternate sign
making the above second order conditions produce a positive denominator. The
whole term is then negative. Thus,

dZ/dl < 0

Proposition 5: QED.

152

 

 

Because an increase in either exports or imports decreases conﬂict, one can
combine Propositions 1 and 3 and Propositions 2 and 4. This produces three
testable hypotheses.

 

153

Appendix C

States in the ICRG dataset and year for which data is ﬁrst available

Albania
Algeria
Angola
Argentina
Australia
Austria
Bahamas
Bahrain
Bangladesh
Belgium
Bolivia
Botswana
Brazil
Brunei
Bulgaria
Burkina Faso
Cameroon
Canada
Chile
Colombia
Congo
Costa Rica
Cote d'lvoire
Cuba
Cyprus
Czechoslovakia
Denmark
Dominican
Republic
East Germany
Ecuador
Egypt

El Salvador
Ethiopia
Finland
France
Gabon
Gambia

1984
1982
1984
1982
1982
1982
1984
1984
1982
1982
1982
1984
1982
1984
1984
1985
1982
1982
1982
1982
1985
1982
1982
1984
1984
1984
1982

1982
1984
1982
1982
1982
1984
1982
1982
1982
1985

Germany, FR
Ghana
Greece
Guatemala
Guinea
Guinea-Bissau
Guyana
Haiti
Honduras
Hong Kong
Hungary
Iceland

India
Indonesia
Iran

Iraq

Ireland
Israel

Italy
Jamaica
Japan
Jordan
Kenya
Korea, Republic
Kuwait
Lebanon
Liberia

Libya
Luxembourg
Madagascar
Malawi
Malaysia
Mali

Malta
Mexico
Mongolia
Morocco
Mozambique

154

1982
1982
1982
1982
1984
1985
1982
1982
1982
1982
1984
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1982
1984
1984
1982
1982
1984
1986
1982
1986
1982
1984

Myanmar
Namibia
Netherlands
New Zealand
Nicaragua
Niger
Nigeria
North Korea
NonNay
Oman
Pakistan
Panama
Papua New
Guinea
Paraguay
Peru
Philippines
Poland
Ponugal
Romania
Saudi Arabia
Senegal
Sierra Leone
Sierra Leone
Singapore
Somalia
South Africa
Spain

Sri Lanka
Sudan
Suriname
Sweden
Switzerland
Syria
Taiwan
Tanzania
Thailand
Togo

1982
1990
1982
1982
1982
1985
1982
1985
1982
1984
1982
1982

1984
1982
1982
1982
1984
1982
1984
1982
1982
1984
1985
1982
1984
1982
1982
1982
1982
1985
1982
1982
1982
1982
1982
1982
1982

 

 

Trinidad &
Tobago
Tunisia
Turkey
United Arab
Emirates
Uganda
United Kingdom
United States
Uruguay
USSR
Venezuela
Vietnam
West Germany
Yemen, Arab
Republic
Yemen, PDR
Yugoslavia
Zaire

Zambia
Zimbabwe

1982
1982
1982

1982
1982
1982
1982
1982
1984
1982
1984
1982

1984
1985
1982
1982
1982
1982

155

 

 

BIBLIOGRAPHY

156

. ‘1.

 

BIBLIOGRAPHY

Anderton, Charles H., Roxane A. Anderton, and John R. Carter. 1999. Economic
Activity in the Shadow of Conﬂict. Economic Inquiry 37 (1): 166-179.

Angell, Norman. 1912. The Great Illusion: A Study of the Relation of Military
Power in Nations to their Economic and Social Advantages. London:
Heinemann.

Baldwin, David. 1985. Economic Statecra_ft. Princeton, NJ: Princeton
University Press.

Barbieri, Katherine. 1996. Economic Interdependence: A Path to Peace or a
Source of Interstate Conﬂict? Journal of Peace Research 33 (1): 29-49.

Barbieri, Katherine and Jack S. Levy. 1999. Sleeping With The Enemy: The
Impact of War on Trade. Journal of Peace Research 36 (4): 463-479.

Barbieri, Katherine and Gerald Schneider. 1999. Globalization and Peace:
Assessing New Directions in the Study of Trade and Conﬂict. Journal of
Peace Research 36 (4): 387-404.

Beck, Nathaniel, Jonathan Katz and Richard Tucker. 1998. Taking Time
Seriously: Time Series—Cross Section Analysis with a Binary Dependent
Variable. American Journal of Political Science 42: 1260-1288.

Bennett, D. Scott, and Allan Stam. 2000. EUGene: A Conceptual Manual.

International Interactions 26:179-204. Website: http:l/eugenesoftware.org.

Benoit, Kenneth. 1996. Democracies Really Are More Paciﬁc (in General).
Journal of Conﬂict Resolution 40: 636-657.

Brawley, Mark R. 1993. Liberal Leadership: Great Powers and Their
Challemers in Peace gnd Wg. Ithaca, NY: Cornell University Press.

Brooks, Stephen G. 1999. The Globalization of Production and the Changing
Beneﬁts of Conquest. Journal of Conﬂict Resolution 43 (5): 646-670.

Bueno de Mesquita, Bruce. 1990. Pride of Place: The Origins of German
Hegemony. World Politics 43: 28-52.

Bueno de Mesquita, Bruce. 1981. The War Trag. New Haven, CT: Yale
University Press.

157

Emga
_. l _ .

‘1 .

4 a

 

Bueno de Mesquita, Bruce and David Lalman. 1992. War and Reason: Domestic
and International Imperatives. New Haven, CT: Yale University Press.

Bueno de Mesquita, Bruce, James Morrow, Randolph Siverson, and Alastair
Smith. 1999. “An Institutional Explanation of the Democratic Peace.”
American Political Science Review 93 (4): 791-807.

Bueno de Mesquita, Bruce and Randolph Siverson. 1995. “War and the Survival
of Political Leaders: A Comparative Study of Regime Types and Political
Accountability.” American Political Science Review 89: 841-855.

Burton, John Hill. 1846. Life and Correspondence of David Hume. Edinburgh,
UK: W. Tait.

Buzan, Barry, Charles Jones and Richard Little. 1993. The Logic of Anarchy:
From Neorealism to Structural Realism. New York, NY: Columbia
University Press.

 

Chan, Steve. 1997. In Search of Democratic Peace: Problems and Promise.
Mershon International Studies Review 41: 59—91.

Clague, Christopher, Philip Keefer, Stephen Knack, and Mancur Olson. 1996.
Property and Contract Rights in Autocracies and Democracies. Journal of
Economic Growth 1: 243-276.

Deutsch, Karl W, et al. 1957. Politigal Community and the North Atlantic Area:
lntemational Organization in the Light of Historical Egerience. Princeton,
NJ: Princeton University Press.

Dicken, Peter. 1998. Global Shift: Transforming the World Economy. London:
Paul Chapman.

Domke, William K. 1988. War and the Changi_ng Global System. New Haven,
CT: Yale University Press.

Dorussen, Hans. 1999. "Balance of Power Revisited: A Multi-Country Model of
Trade and Conﬂict." Journal of Peace Research 36(4): 443-462.

The Economist. February 24, 2001. “The Cutting Edge.”

The Economist. July 29, 1999. The Thing that Won't Go Away: Why the State
Stays the Basic Unit.

 

The Economist. June 18, 1998. "The World as a Single Machine."

The Economist. November 8, 1997. Trade Winds.

158

Fearon, James. 1994. Domestic Political Audiences and the Escalation of
International Disputes. American Political Science Review 88: 577-592.

Foreign Policy. January 2001. Measuring Globalization.

Forland, Tor Egil. 1991. “Economic Warfare and Strategic Goods: A Conceptual
Framework for Analyzing COCOM.” Journal of Peace Research 28 (2):
191 -204.

Friedman, Thomas L. 1999. Dueling Globalizations: A Debate Between Thomas
L. Friedman and Ignacio Ramonet. Foreign Policy Fall: 110-127.

Fukuyama, Francis. 1992. The End of History and the Last Man. New York, NY:
Avon Books.

Fukuyama, Francis. 1989. The End of History? The National Interest 16: 3-18.

Gartzke, Erik. 2000. Preferences and the Democratic Peace. International
Studies Quarterly 44: 191-212.

Gartzke, Erik. 1998. Kant We All Just Get Along? Opportunity, Willingness, and
the Origins of the Democratic Peace. American Jogrnal of Political
Science 42 (1): 1-27.

Gartzke, Erik, Quan Li, and Charles Boehmer. 2001. Investing in Peace:
Economic Interdependence and International Conﬂict. Internationa_l
Organization 55 (2): 391-438.

Gasiorowski, Mark J. 1986. “Economic Interdependence and International
Conflict: Some Cross-national Evidence.” International Studies Quarterly
30: 23-38.

Gilpin, Robert. 1981. War and Charge in World Politics. New York, NY:
Cambridge University Press.

Gleditsch, Kristian. 2000. Expanded Trade and GDP Data, Version 1. httpzl/k-
gleditschsocscigla.ac.uk/proiectslhtml.

Gleditsch, Kristian and Michael Ward. 1997. Double Take: A Reexamination of
Democracy and Autocracy in Modern Polities. Journal of Conﬂict
Resolution 41: 361-383.

Gowa, Joanne. 1989. “Bipolarity, Multipolarity, and Free Trade.” American
Political Science Review 83 (4) 1245-1256.

159

 

Gowa, Joanne and Henry Farber. 1997. Common Interests or Common Polities?
Reinterpreting the Democratic Peace. Journal of Politics 59: 393-417.

Hegre, Havard. 2000. Development and the Liberal Peace: What Does It Take
to be a Trading State? Journal of Peace Research 37 (1): 5-30.

Helpman, Elhanan and Paul R. Krugman. 1985. Market Structure and Foreign
Trade: Increasing Returns, Imperfect Competition, andrthe International
Economy. Cambridge, MA: MIT press.

Holsti, Kalevi. 1991. Pgace and War: Armed Conﬂict_s and International Order.
1648-1989. New York, NY: Cambridge University Press.

Huth, Paul K. 1996. Standing Your Ground: Territorial Disputes and International
Conﬂict. Ann Arbor, MI: University of Michigan Press.

Jones, Daniel M., Stuart Bremer and J. David Singer. 1996. “Militarized Interstate
Disputes, 1816-1992: Rationale, Coding Rules, and Empirical Patterns.”
Conﬂict Management and Peace Science 15: 163-213.

Kant, Immanuel. [1795] 1957. Perpetual Peace. Edited by Lewis White Beck.
Indianapolis: Bobbs-Merrill.

Kenen, Peter B. 2000. The Internationalgconomy. New York, NY: Cambridge
University Press.

Keohane, Robert. 1984. After Hegemony: Cooperation and__Discorg in the
World Political Economy. Princeton, NJ: Princeton University Press.

Keohane, Robert O. and Joseph S. Nye. 1989. Power and Interdependence.
Glenview, IL: Scott, Foresman.

King, Gary. 1989. Unifyigg Political Methodology: The l__ikelihood Theory of
Statistical Inference. NY: Cambridge University Press.

King, Gary and Langche Zeng. 2001. Explaining Rare Events in lntemational
Relations. International Organization 55 (3): 693-715.

Knack, Stephen and Philip Keefer. 1995. Institutions and Economic Performance:
Cross Country Tests Using Alternative Institutional Measures. Economics
and Politics 7: 207-227.

Krasner, Stephen. 1978. Defending the National Interest: Raw Materials
Investments and US. Foreign Policy. Princeton, NJ: Princeton University
Press.

160

Kugler, Jacek and Douglas Lemke. 1996. Parity and War. Ann Arbor, MI:
University of Michigan press.

Kugler, Jacek and Douglas Lemke. 2000. The Power Transition Research
Program: Assessing Theoretical and Empirical Advances. In Handgiok of
War Studies 11. Manus Midlarsky, editor. Ann Arbor, MI: University of
Michigan press.

Lake, Anthony. 1993. From Containment to Enlargement. Speech given at
Johns Hopkins University September 21.
httpzllwww.mtholyokeedulacad/intreI/lakedoc.html

Landes, David S. 1998. The Wealth and Poverty of Nations. New York, NY: W.
W. Norton.

Lemke, Douglas. 1995. Tyranny of Distance: Redeﬁning Relevant Dyads.
International Interactions 21: 23-38.

Lemke, Douglas and William Reed. 1996. Regime Types and Status quo
Evaluations: Power Transition Theory and the Democratic Peace.
International Interactions 22 (2): 143-164.

Lemke, Douglass and William Reed. 2001. The Relevance of Politically Relevant
Dyads. Mrnal of Conﬂict Res_olt_1tion 45 (1): 126-143.

Lewis-Beck, Michael. 1988. Economics and Elections: The Maior Western
Democracies. Ann Arbor, MI: University of Michigan Press.

Liang, K. and S. Zeger. 1986. “Longitudinal Data Analysis Using Generalized
Linear Models.” Biometrika 73: 13-22.

Liberman, Peter. 1996. Does Congtigst Pay? The Igploitation of Occupied
lndgstrial Societies. Princeton, NJ: Princeton University Press.

Liberman, Peter. 1996. Trading with the Enemy: Security and Relative
Economic Gains. International Securiy 21 (1): 147-175.

Liberman, Peter. 1993. The Spoils of Conquest. International Seam 18 (2):
125 -153.

Mansﬁeld, Edward D. 1994. Power, Trade. and War. Princeton, NJ: Princeton
University Press.

Maoz, Zeev and Bruce Russett. 1993. Normative and Structural Causes of the
Demoratic Peace. American Political Science Review 87: 624-638.

161

Maoz, Zeev and Nasrin Abdolali. 1989. Regime Types and lntemational Conﬂict,
1816—1976. Journal of Conﬂict Resolution 33: 3-35.

Marshall, Monty and Keith Jaggers. 2000. Polity lV Project: Political Regime
Characteristics and Transitions, 1800-1999.
http://www.bsos.umd.edu/cidcm/polity/

Mayhew, David R. 1974. Congess: The Electoral Connection. New Haven CT:
Yale University Press.

McMillan, Susan M. 1997. Interdependence and Conﬂict. Mershon International
Studies Review 41: 33-58.

Mill, John Stuart. 1909 (1848). Principles of Political Economy. New York, NY:
Longmans Green, and company.

de Montesquieu, Baron. 1900. The Spirit of Laws. Translated by Thomas Nugent.
New York, NY: Collier Press.

Morrow, James. 1999. How Could Trade Affect Conﬂict? Journal of Peace
Research 36: 481-489.

Morrow, James, Randolph Siverson, and Tressa Tabares. 1998. The Political
Determinants of International Trade: The Major Powers 1907-90.
Amerign Political Science Review 92: 649-661.

Most, Benjamin and Harvey Starr. 1989. Inquiry, Logic, aﬂl International Politics.
Columbia, SC: University of South Carolina Press.

Mousseau, Michael. 2000. Market Prosperity, Democratic Consolidation, and
Democratic Peace. Journal of Conﬂict Resolption 44(4): 472-507.

North, Douglass C. 1990. lnstitptions, Institutional Change and Economic
Performance. New York, NY: Cambridge University Press.

North, Douglass. 1981. Structure ancﬁhagqe iLEwnomic History. New York,
NY: W. W. Norton.

Olson, Mancur. 1996. Big Bills Left on the Sidewalk: Why Some Nations are
Rich, Others Poor. Journal of Economic Perspectives 10 (2): 3-24.

Oneal, John R and Bruce M. Russett. 1997. The Classical Liberals Were Right:

Democracy, Interdependence, and Conﬂict, 1950-1985. International
Studies Quarterly 41: 267-294.

162

 

Oneal, John R. and Bruce Russett. 19993. Assessing the Liberal Peace with
Alternative Speciﬁcations: Trade Still Reduces Conﬂict. Journal of Peace
Research 36 (4): 423-442.

Oneal, John R. and Bruce Russett. 1999b. The Kantian Peace: The Paciﬁc
Beneﬁts of Democracy, Interdependence, and International Organizations.
World Politics 52: 1.

Organski, A. F. and Jacek Kugler. 1980. The War Ledger. Chicago, IL:
University of Chicago press.

Partell, Peter and Glenn Palmer. 1999. Audience Costs and Interstate Crises:
An Empirical Assessment of Fearon’s Model of Dispute Outcomes.
International Stggies @arterly 43 (2): 389-405.

Pipes, Richard. 1999. Property arﬁFreedpm. New York, NY: Vintage Books.

Polachek, Solomon W. 1980. “Conﬂict and Trade.” Journal of Conﬂict Resolution.
24: 55-78.

Polachek, Solomon W., John Robst, and Yuan-Ching Chang. 1999. Liberalism
and Interdependence: Extending the Trade-Conﬂict Model. Journal of
Peace Research 36 (4): 405-422.

Prestowitz, Clyde V. 1988. Trading Places: How We Allowed Japan To TLKe the
Lead. New York, NY: Basic Books.

Prins, Brandon. 2001. Domestic Politics and Interstate Disputes: Examining U.S.
MID Involvement and Reciprocation,1870-1992. Forthcoming in
International Interactions.

Reich, Robert B. 1991. The Work of Nations: Pregaring Ourselves for 21St
Centum Capitalism. New York, NY: A. A. Knopf.

Reuvany, Rafael. 1999. The Trade and Conﬂict Debate: A Survey of Theory,
Evidence and Future Research. Peace Economics. Paape Science, and
Public Policy 6: 23-49.

Reuvany, Rafael and Heejoon Kang. 1998. Bilateral Trade and Political
Conﬂict/Cooperation: Do Goods Matter? Journal of Peace Research 35:
581-602.

Rosecrance, Richard. 1999. The Rise of tha Virtual State. New York: Basic
Books.

163

 

Rosecrance, Richard. 1986. The Rise of the Trading State. New York: Basic
Books.

Rotte, Ralph. 1997. Economics and peace-theory on the eve of World War I. In
Economics of Conﬂict and Peace. Eds. Jurgen Brauer and William G.
Gissy. Brookﬁeld, VT: Ashgate Publishing Company.

Rousseau, David L., Christopher Gelpi, Dan Reiter, and Paul K. Huth. 1996.
Assessing the Dyadic Nature of the Democratic Peace, 1918-1988.
American Political Science Review 90: 512-33.

Russett, Bruce and Harvey Starr. 2000. From Democratic Peace to Kantian
Peace: Democracy and Conﬂict in the International System. In Handbook
of War Studies II. Manus Midlarsky, editor. Ann Arbor, MI: University of
Michigan Press.

Russett, Bruce, John Oneal, and David Davis. 1998. The Third Leg of the
Kantian Tripod for Peace: International Organizations and Militarized
Disputes, 1950-1985. International Organization 52(3): 441-67.

Schultz, Kenneth A. 1999. Do Democratic Institutions Constrain or Inform?
Contrasting Two Institutional Perspectives on Democracy and War.
International Organization 53 (2): 233-266.

Signorino, Curtis and Jeffrey Ritter. 1999. Tau-b or Not Tau-b. International
Studies Quarterly 43: 115: 144.

Sobel, Andrew C. 1999. State Institutions. Private Incentives, and Global
Capital. Ann Arbor, MI: University of Michigan Press.

Summers, Robert and Alan Heston. 1991. The Penn World Table (Mark 5): An
Expanded Set of International Comparisons, 1950-1988. The guartemr
Journal of Economics. 106: 327-368.

Uchitel, Anne. 1993. Interdependence and Instability. In Coping with Complexity
in the International System. Editors Jack Snyder and Robert Jervis.
Boulder, CO: Westview Press.

Waltz, Kenneth N. 1979. Theory of International Politics. Reading, MA: Addison
Wesley.

Werner, Suzanne. 2000. “The Effects of Political Similarity on the Onset of
Militarized Disputes, 1816-1985." Political Research Quarterly 53 (2):

Werner, Suzanne and Douglas Lemke. 1997. Opposites Do Not Attract: the

164

 

Impact of Domestic Institutions, Power, and Prior Commitments on
Alignment Choices. International Studies CMrterly 41: 529-546.

Werner, Suzanne and Douglas Lemke. 1996. Power Transitions and Military
Buildups. In Parity and War, editors Jacek Kugler and Douglas Lemke.
Ann Arbor, MI: University of Michigan Press.

World Bank. 1998. The 1998 World Development Indicators CD-ROM.
Washington, DC: World Bank.

Zorn, Christopher J. 2001. General Estimating Equation Models for Correlated
Data: A Review With Applications. American Journal of Political Science
45 (2): 470-490.

165

Ill