93123
Public Disclosure Authorized
Public Disclosure Authorized
DIREC TIONS IN DE VELOPMENT
Public Disclosure Authorized
Trade
Trading Away from Conflict
Using Trade to Increase Resilience in Fragile States
Massimiliano Calì
Public Disclosure Authorized
Trading Away from Conflict
Direc tions in De velopment
Trade
Trading Away from Conflict
Using Trade to Increase Resilience in Fragile States
Massimiliano Calì
© 2015 International Bank for Reconstruction and Development / The World Bank
1818 H Street NW, Washington DC 20433
Telephone: 202-473-1000; Internet: www.worldbank.org
Some rights reserved
1 2 3 4 17 16 15 14
This work is a product of the staff of The World Bank with external contributions. The findings, interpreta-
tions, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its
Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the
accuracy of the data included in this work. The boundaries, colors, denominations, and other information
shown on any map in this work do not imply any judgment on the part of The World Bank concerning the
legal status of any territory or the endorsement or acceptance of such boundaries.
Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and
immunities of The World Bank, all of which are specifically reserved.
Rights and Permissions
This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://
creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to
copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following
conditions:
Attribution—Please cite the work as follows: Calì, Massimiliano. 2015. Trading Away from Conflict: Using
Trade to Increase Resilience in Fragile States. Directions in Development. Washington, DC: World Bank.
doi:10.1596/978-1-4648-0308-6. License: Creative Commons Attribution CC BY 3.0 IGO
Translations—If you create a translation of this work, please add the following disclaimer along with the
attribution: This translation was not created by The World Bank and should not be considered an official
World Bank translation. The World Bank shall not be liable for any content or error in this translation.
Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the
attribution: This is an adaptation of an original work by The World Bank. Responsibility for the views and
opinions expressed in the adaptation rests solely with the author or authors of the adaptation and are not
endorsed by The World Bank.
Third-party content—The World Bank does not necessarily own each component of the content con-
tained within the work. The World Bank therefore does not warrant that the use of any third-party-
owned individual component or part contained in the work will not infringe on the rights of those
third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to
re-use a component of the work, it is your responsibility to determine whether permission is needed
for that re-use and to obtain permission from the copyright owner. Examples of components can
include, but are not limited to, tables, figures, or images.
All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The
World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@
worldbank.org.
ISBN (paper): 978-1-4648-0308-6
ISBN (electronic): 978-1-4648-0309-3
DOI: 10.1596/978-1-4648-0308-6
Cover photo: © James Oatway / Panos Pictures. Used with the permission of James Oatway / Panos Pictures.
Further permission required for reuse.
Cover design: Naylor Design.
Library of Congress Cataloging-in-Publication Data
Calì, Massimiliano.
Trading away from conflict : using trade to increase resilience in fragile states / Massimiliano Calì.
pages cm. — (Directions in development)
Includes bibliographical references.
ISBN 978-1-4648-0308-6 (alk. paper) — ISBN 978-1-4648-0309-3 (ebk)
1. Developing countries—Commercial policy. 2. Developing countries—Commerce—Political aspects.
3. Insurgency—Economic aspects—Developing countries. 4. Economic development—Political aspects—
Developing countries. I. World Bank. II. Title.
HF1413.C355 2014
382’.3091724—dc23 2014030078
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Contents
Acknowledgments ix
About the Author xi
Abbreviations xiii
Overview 1
Introduction 1
Main Results 2
Policy Directions 4
Notes 6
References 7
Chapter 1 How Trade Can Affect Conflict 9
Introduction 9
Trade Flows in Fragile Countries Are Different 12
Why Changes in Trade Flows May Affect Conflict 18
Cross-Country Evidence on Trade Shocks and Conflict 25
Evidence from Nigerian States (2004–13) 34
Evidence from the Israeli-Palestinian Conflict (2000–04) 42
Notes 51
References 54
Chapter 2 Conditions That Affect the Impact of Trade Shocks
on Conflict 59
Introduction 59
Grievances 61
Institutional Capacity and Inclusiveness 63
Conditions in Neighboring Countries 65
Transmission of Prices to Domestic Markets 65
Cross-Country Empirical Tests 66
Testing for the Importance of Heterogeneity in
the Nigerian Conflict 72
Heterogeneity in the Israeli-Palestinian Conflict 73
Notes 73
References 75
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 v
vi Contents
Chapter 3 How Trade Policy Could Ease Tensions in
Fragile Countries 79
Trade Policies in Fragile Countries Must Take into
Account the Implications for Conflict 80
Manage Receipts from Commodity Exports in a
Conflict-Sensitive Way 81
Protect Producers, Consumers, and Workers from Adverse
Trade Shocks 85
Promote Trade with Neighbors 86
Support Labor-Intensive Exports 86
Build Long-Term Conflict Resilience 87
Notes 88
References 89
Appendix A Data Issues 93
Appendix B Estimation Methodology and Empirical Results 109
Boxes
1.1 Which Are the Fragile Countries? 12
1.2 Empirical Issues in the Early Literature on the Relationship
between Changes in Income and Conflict 26
1.3 The South Sudanese Civil War: Was Oil Export the Trigger? 28
1.4 Correcting for Endogeneity When Measuring the Relationship
between Conflict and Trade under RTAs 30
1.5 The Literature on the Israeli-Palestinian Conflict and
the Opportunity Cost of Violence 43
Figures
O.1 Most of the Poor Will Soon Be in Fragile Countries 2
1.1 Trade Represents the Major Source of Foreign Exchange in
Fragile States 13
1.2 Share of Largest Exports in Selected Fragile Countries and
Territories (in 2010) 14
1.3 For Many Fragile States, Exports Are Not Heavily Diversified 15
1.4 For Fragile States, Net Food Imports Constitute a Higher
Percentage of GDP 16
1.5 Fragile Countries Perform Worse Than Their Peers in Trade
Facilitation and the Gap Is Growing 17
1.6 Mapping the Linkages between Changes in Trade Flows and
Civil Conflict 19
1.7 Palestinian Exports to the World and to Israel, 1996–2000 44
1.8 Distribution of Changes in Palestinian Exports (1996–99) 45
1.9 Palestinians Killed by Israel in the West Bank and Gaza, 2000–04 46
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Contents vii
1.10 Israeli Imports from China and the West Bank and Gaza,
1995–2000 49
2.1 Marginal Effects of Px Across the Range of Interaction
Variables’ Values 70
B.1 Changes in Israeli MFN Tariffs (5-digit SITC Rev. 3), 1993–2004 127
Maps
1.1 The Geography of Conflict in Nigeria (2004–13) 36
1.2 Conflict Intensity across States in Nigeria 36
1.3 Violence Intensity across States in Nigeria 37
Tables
1.1 Fragile Countries’ Exports Are Less Diversified Than Other
Developing Countries’ Exports 14
1.2 Classification of the Export Commodities (with Example) 21
2.1 Under What Conditions Are the Marginal Effects of Trade
Shocks Not Significant? 69
2.2 Lebanon Has a Higher Risk of Conflict from Hydrocarbons
Exports than the Average Country 71
A.1 Fragile Countries and Territories and Number of Battle Deaths 93
A.2 Summary Statistics, Cross-Country Analysis 97
A.3 Interaction Variables for the Cross-Country Analysis 97
A.4 Household Per Capita Expenditure on Food and
Nonfood by Zone 101
A.5 Summary Statistics of the Regressors (2004–11) 101
A.6 Summary Statistics for Key Variables in the Israeli-Palestinian
Conflict Study 103
A.7 Description of Variables Used in the Palestinian Case Study 104
B.1 The Impact of Trade on Conflict, Cross-Country Analysis 112
B.2 Robustness with Fast-Moving, Country-Specific Time Trends,
Cross-Country Analysis 113
B.3 Robustness for Price Makers and Conflict Data Source,
Cross-Country Analysis 113
B.4 Splitting the Commodities’ Variables into Different Types,
Cross-Country Analysis 114
B.5 The Impact of Trade Variables on the Likelihood of Conflict
Coming to an End, Cross-Country Analysis 115
B.6 The Impact of Trade on Battle Deaths, Cross-Country Analysis 116
B.7 The Effect of Trading with Neighbors on Conflict, Revisited,
Cross-Country Analysis 117
B.8 Trade Variables without Lag Structure, Cross-Country Analysis 118
B.9 Summary Statistics of the Dependent Variable (2004–11),
Nigeria 119
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
viii Contents
B.10 The Impact of Price Shocks on Conflict Events in Nigeria
(2004–11) 121
B.11 The Impact of Price Shocks on Conflict in Nigeria (2004–11),
Robustness 122
B.12 The Impact of Price Shocks on Various Types of Conflict in
Nigeria (2004–11) 123
B.13 Mediating Factors Affecting the Impact of Price Shocks on
Conflict 124
B.14 The Impact of Changes in Trade Prices on the Boko Haram
Conflict (2010–13) 125
B.15 The Impact of Palestinian Exports on Conflict Intensity 129
B.16 The Impact of Palestinian Trade on Conflict Intensity 130
B.17 The Impact of Palestinian Trade on a Different Measure of
Conflict Intensity 131
B.18 The Impact of Palestinian Trade on Conflict Probability 131
B.19 The Impact of Palestinian Trade on Conflict: Tackling
Endogeneity 132
B.20 The Heterogeneity of the Impact of Export Changes on
Conflict Intensity 132
B.21 Instrumenting Palestinian Exports through Exogenous Shocks 133
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Acknowledgments
This report was prepared by Massimiliano Calì, Trade Economist (Trade and
Competitiveness Global Practice), on the basis of studies conducted together
with a team of consultants comprising Babatunde Abidoye (University of
Pretoria), Amir Fouad (Consultant, Trade and Competitiveness Global Practice),
Sami Miaari (Tel Aviv University), and Alen Mulabdic (Graduate Institute of
International Studies). William Shaw (Consultant, Trade and Competitiveness
Global Practice) helped put together the report.
The report benefited considerably from the overall guidance and comments
provided by Mona Haddad, Practice Manager (Trade and Competitiveness Global
Practice). Peer reviewers Quy-Toan Do, Senior Economist (Development
Research Group), and Raju Singh, Lead Economist (Macroeconomics and Fiscal
Management Global Practice), provided excellent comments that considerably
improved the report. Marie Yenko provided able editorial assistance.
Useful comments were also provided at various stages of the project by
Nicolas Berman, Elena Ianchovichina, Philip Keefer, Charles Kunaka, Eric Le
Borgne, Daniel Lederman, Hani Mansour, Renzo Massari, Gary Milante, Nadia
Piffaretti, Nicola Pontara, Espen Prydz, Nadia Selim, Radhika Srinivasan, and
Yongmei Zhou.
This project was supported in part by the governments of Finland, Norway,
Sweden, and the United Kingdom through the Multi-Donor Trust Fund for Trade
and Development.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 ix
About the Author
Massimiliano Calì is a trade economist at the World Bank. His current and recent
work focuses on the relation between economics and conflict, on the poverty
impact of trade, and on migration. Prior to joining the Bank in 2012, he served
as an economic advisor to the Palestinian Ministry of National Economy in
Ramallah, as a research fellow with the Overseas Development Institute, and as
an economist with the Italian Embassy to Bolivia. In these capacities he has pro-
vided economic policy advice to a number of Ministries in developing countries
as well as to international organizations and NGOs. His work has been published
in academic journals, books, and official reports. He holds a PhD in economic
geography from the London School of Economics, an MA in development eco-
nomics from the University of East Anglia, and a BA in economics from the
University of Pavia.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 xi
Abbreviations
ACLED Armed Conflict Location and Events Dataset
AGOA Africa Growth and Opportunity Act
CASA Conflict Affected States in Africa
COW Correlates of War Project
CPIA country policy and institutional assessment
DDPs direct dividend payments
EITI extractive industries transparency initiative
FCS Fragile and Conflict-Affected Situations
FDI foreign direct investment
FSI Failed States Index
GFRP Global Food Crisis Response Program
KP Kimberley Process
LPI logistics performance indicator
MFN Most Favoured Nation
NBS National Bureau of Statistics
NTMs nontariff measures
ODA official development assistance
PLFS Palestinian Labor Force Survey
PWYP publish what you pay
RTAs regional trade agreements
SWF Sovereign Wealth Fund
ZINB zero-inflated negative binomial estimator
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 xiii
Overview
Introduction
In the past 30 years the world has become much less poor everywhere except in
fragile countries. By 2015 most of the world’s poor are expected to live in fragile
countries (figure O.1).1 The majority of these countries has been, or still is,
affected by civil conflicts. In addition to exacting a huge toll on human life, civil
conflicts cause protracted, severe disruption of economic activities and infra-
structure, and are key constraints to development in many countries. Cognizant
of these challenges, the International Development Association (IDA) has
provided over $11.2 billion in post-conflict reconstruction assistance to “fragile
and conflict affected situations” since 2000.2 However, the challenges in these
countries remain daunting.
Fragile and conflict-affected countries are not only home to an increasing
share of the world’s poor; they are also at a greater risk of relapsing into conflict
than other countries. Nearly 90 percent of the conflicts between 2000 and 2010
occurred in countries that had already experienced a recent conflict; almost half
of the post-conflict countries relapse into conflict within 10 years (World Bank
2011a). Republic of South Sudan and the Central African Republic are but the
latest examples of fragile countries that fell back into conflict. The challenge is
particularly daunting in sub-Saharan Africa, where most countries at risk of con-
flict are concentrated.3
Trade and trade policy can greatly affect the risk of conflict. Trade encourages
the reallocation of resources to more efficient activities, and thus opens up
opportunities and creates jobs. However, changes in relative prices as a result of
trade can also destroy opportunities and jobs in declining sectors, and the people
affected by these losses may, under certain conditions, turn to violence as a
source of income. Changes in real incomes generated by trade are particularly
important in fragile states, where trade flows tend to be larger and more volatile
than other external flows, such as aid, remittances and foreign investment. This
volatility is partly due to these countries’ low diversification and their high
dependence on primary export commodities, which may exacerbate the effects
of abrupt changes in exports on conflict. For example, a sharp fall in interna-
tional coffee prices in Colombia during the 1990s lowered wages and increased
violence more in coffee-producing municipalities than in other municipalities
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 1
2 Overview
Figure O.1 Most of the Poor Will Soon Be in Fragile Countries
2,000
1,800
Number of poor people, millions
1,600
(< $1.25/day, 2005 PPP) 1,400
1,200
1,000
800
600
400
200
0
81
84
87
90
93
96
99
02
05
08
11
14
17
20
23
26
29
19
19
19
19
19
19
19
20
20
20
20
20
20
20
20
20
20
Total Nonfragile Fragile
Sources: Historic (1981–2011) data from PovcalNet (accessed Oct. 10, 2014). Projections based on 10-year
historic growth scenario from Lakner, Negre, and Prydz (2014).
Note: Estimates from counties on the OECD Fragile States list from 2014. Poverty estimates only from
countries where at least one houshold survey and PPP conversion factors available. For 18 countries of the
51 on the list, we do not have such data. These missing countries comprise about 10 percent of the total
population in fragile states.
(Dube and Vargas 2013). In addition, the majority of fragile countries are net
food importers, so they are particularly exposed to the recent swings in interna-
tional food prices. A number of governments in the Middle East and North
Africa responded to the social unrest at the onset of the Arab Spring by extend-
ing food and fuel subsidies (World Bank 2011b). Countries also may be exposed
to changes in the international demand for their products due to changes in their
trading partners’ incomes or changes in the access to foreign markets.
This report aims to understand how changes in imports and exports affect the
risk and intensity of conflict and to help policy makers use trade to reduce this risk.
tandard
In this way, it attempts to explicitly incorporate a fragility lens into the s
trade policy discussion in fragile countries. In doing so, the report also makes a
number of contributions to the nascent but growing empirical literature on the
relationship between changes in trade and conflict. It uses three different sets
of data to do so: the experience of conflict across countries from 1960 to 2010,
conflict across states in Nigeria from 2004 to 2013, and conflict during the Second
Intifada in the West Bank and Gaza from September 2000 to December 2004.
Main Results
The analysis considers three main mechanisms for how trade-related changes can
affect conflict. The opportunity cost effect holds that changes in real incomes, for
example driven by changes in trade prices, change incentives for participating in
conflict by changing the return on participation in violence compared with more
productive activities. The rapacity (sometimes called “state prize”) effect refers to
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Overview 3
the idea that valuable economic resources can provide an incentive to fight over
their control. And the resource effect recognizes that both government and reb-
els may fund their activities by taxing the production of commodities, so that
changes in their value affect the ability to sustain conflict.
The empirical results provide strong support for the rapacity effect. Increases
in the prices of exported oil and mineral commodities substantially raise the risk
conflict. An increase in the value of these exports of 10 percent raises the risk of
conflict by 2.2 percent on average across countries. This is due to the rapacity
effect. The higher the value of resources that can be easily appropriated through
fighting, such as minerals and oil, the greater is the incentive to fight over them.
The finding from Nigerian states is similar: a 10 percent increase in the price of
oil raises the number of conflict events by 2 percent. These results are also con-
sistent with other intra-country evidence from Colombia (Dube and Vargas
2013), the Democratic Republic of Congo (Maystadt et al. 2014), and from Sub-
Saharan Africa (Berman et al. 2014).
When the Nigerian government started using some of the oil revenues to
demobilize and reintegrate the militants in the oil-producing regions, the positive
relation between oil price and conflict intensity disappeared. That followed the
agreement in 2009, whereby the federal state granted amnesty and provided
employment to the militants in those states. This finding supports the “resource
effect,” which recognizes that the government (and sometimes also the rebels)
may fund their activities by taxing the production of commodities, so that
changes in their value affect the ability to repress or buy off the rebels, at least in
the short term.
On the other hand, the cross-country evidence provides little support for the
opportunity cost hypothesis. Changes in the prices of agricultural exports, in the
prices of imported commodities, and in export demand are not significantly
related to the probability of conflict. By contrast, the country case studies pro-
vide strong support for the opportunity cost effect. This difference with the
cross-country evidence is likely due to two reasons. First, the availability of data
within countries allows one to isolate the impact of commodity price changes on
real incomes. Second, the large heterogeneity across countries can mask effects
that may be important within individual countries. In Nigeria, conflict is signifi-
cantly related to changes in real incomes driven by commodity indexes that
reflect both production (higher prices, less conflict) and consumption (higher
prices, more conflict) by the households. The importance of changes in real
incomes in affecting conflict also applies to the Boko Haram attacks since 2010.
The opportunity cost hypothesis holds in the West Bank and Gaza, where exog-
enous sectoral increases in export revenues were associated with subsequent
lower levels of conflict during the Second Intifada in localities where private
sector employment in that sector was significant. These findings confirm the
evidence emerging from other within-country studies (e.g., Dube and Vargas
2013; Berman and Couttenier 2014).4
Intense trading with neighbors reduces the duration as well as the intensity of
conflict. This trade reduces the incentives of contiguous countries to fuel civil
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
4 Overview
conflict in their neighbors similarly to the case of inter-state wars. These incen-
tives may be particularly strong in areas, such as much of sub-Saharan Africa,
where there are strong ethnic ties across borders. Trading with neighbors is also
associated with a lower risk of conflict when such trade occurs under regional
trade agreements.
The strength of the effect of commodity exports on conflict varies across and
within countries and depends on a number of local conditions. Changes in eco-
nomic conditions have a much greater potential for generating conflict where there
are deep-seated, historical grievances among groups, where economic inequality is
high, and where government institutions are weak or corrupt. The report identifies
four groups of local conditions that may affect the relationship between changes
in trade flows and conflict: (i) grievances that foster tensions among groups, for
example generated by economic inequality, ethnic and religious differences, or past
conflict events; (ii) the state’s institutional capacity and the form of political
arrangements, for example democratic versus authoritarian rule; (iii) conditions in
neighboring countries, for example the level of violence, that might encourage or
discourage conflict in the country of interest; and (iv) policies that affect the
transmission of changes in international commodity prices to the domestic market.
The cross-country analysis suggests that grievances—stemming from ethnic
divisions, income inequality, and a past history of conflict—and the presence of
a conflict in neighboring countries have a particularly significant impact on the
relationship between changes in trade and conflict. While the quality of gover-
nance also helps reduce the effects of trade shocks on conflict, the impact of
political arrangements is more limited. And interventions that slow the transmis-
sion of changes in international commodity prices to domestic markets appear to
reduce the risk of conflict from changes in export prices, although not for
point-source commodities.
Similar, although not identical, results emerge from the country case studies.
In Nigeria, the impact of commodity price swings on conflict is greater in elec-
tion years, in states with high levels of ethnic divisions and inequality, and it is
smaller in states that are farther from Lagos (which tends to reduce the transmis-
sion of international commodity price changes to local markets). Interestingly,
past incidents of violence are not shown to increase the impact of commodity
prices on conflict. In the West Bank and Gaza, the impact of changes in exports
on conflict is increased by the existence of grievances such as the presence of
refugees, high unemployment rates, and a potential indicator of the number
of inhabitants in Israeli jails, but not the incidence of violent fatalities in the past
or the level of education.
Policy Directions
Following the analysis and the review of the evidence, the report highlights five
general policy directions to use trade to support stability, arguably the most
important direct policy objective in fragile countries. These include both
trade-related and complementary policies.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Overview 5
Limit government (and rebels’) access to and discretion over the spending of
the revenues from point-source commodities. Examples of policy options in line
with these principles include: a) improving transparency of the size and use of
these revenues, for example by centralizing government collection of the reve-
nues and by cooperating with international transparency initiatives; b) paying a
portion of the revenues directly to citizens or transferring the revenues to pro-
ducing areas (the former could enhance the oversight of the use of revenues and
create incentives to resist efforts to increase government or rebel control over
resources; the latter could reduce resentment in producing areas and compensate
for the economic disruption and environmental degradation that often accompa-
nies the exploitation of oil or minerals); c) channeling the resource revenues
through external financial vehicles, such as sovereign wealth funds (although the
record of such vehicles is mixed).
Protect the real incomes of producers, consumers, and workers from
adverse changes in trade flows. Targeted transfers, public works programs,
price subsidies, and temporary trade insulation are potential options to achieve
this objective. All of these policies have strengths and weaknesses, but some
evidence suggests that targeted transfers—albeit challenging to develop—
appear to be particularly useful in counteracting the losses by households
as a result of an adverse trade change (Anderson, Ivanic, and Martin 2013;
Attanasio et al. 2013).
Promote labor-intensive exports. This requires two main, mutually reinforc-
ing strategies. The first is to increase fragile countries’ market access in goods
and services in labor intensive sectors in the main trading partners. The second
is to enhance the relative competitiveness of fragile countries’ exports, particu-
larly in labor intensive sectors. This requires a broad set of interventions to
improve trade connectivity and firms’ productivity (Reis and Farole 2012). In
conflict affected and post-conflict environments, both areas are usually defi-
cient due to the destruction and insecurity caused by the conflict.
Strengthen trading relations with neighbors. Both trade policy and trade
facilitation can help foster trade relations among neighbors. There is abundant
evidence of the existence of high policy barriers to trade, especially between
fragile countries. Such barriers even constrain trade in basic food staples between
sub-Saharan African neighbors (World Bank 2012). While necessary, efficient
trade policy is not sufficient to stimulate trade between neighboring fragile coun-
tries, most of which—as this report shows—are marred by particularly poor
transit, logistics, and transport infrastructure systems.
Focus on the broader agenda of reducing some of the structural determinants
of conflict at the country level. That agenda is consistent with some of the prin-
ciples highlighted by the World Bank (2011a), and encompasses: tackling ethnic
divisions, reducing economic inequalities, resolving tensions from past conflicts,
strengthening accountability, and the control of corruption. Building these condi-
tions requires a longer term horizon than is usually adopted by a government
legislature. Yet investing in them is also likely to be necessary in order to perma-
nently break the conflict trap.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
6 Overview
The international community, including the World Bank Group, can help
fragile countries use trade to fight fragility by focusing on certain areas. Key areas
for international support include the provision of technical assistance to improve
trade facilitation and export competitiveness in fragile countries, and to enhance
transparency concerning the size and use of resource revenues; assistance with
arrangements to limit government discretion over resource revenues; the financ-
ing of programs to protect real incomes from adverse changes in trade flows; and
the provision of improved market access in both goods and services for fragile
countries.
This report is composed of three main chapters. Chapter 1 develops a con-
ceptual framework mapping the different channels through which trade may
affect conflict and political stability. The framework is based on simple eco-
nomic theory and the available empirical evidence on the impact of trade-
related changes on conflict and stability. It then tests this framework empirically
through the analysis of cross-country data and through case studies of Nigeria
and the Israeli-Palestinian conflict. The hope is that these types of intra-country
analyses could be replicated in other countries, since they use data that are
available in different countries, especially in sub-Saharan Africa. Chapter 2 uses
the same conceptual framework to show how differences in underlying condi-
tions affect the relationship between trade-related changes and conflict.
Following a review of the literature on the drivers of conflict, it examines the
importance of four groups of grievances: conditions in neighboring countries,
factors increasing grievance, government institutions, and policies that affect
the transmission of changes in international prices to the domestic market.
These relationships are tested using cross-country data and case studies of
Nigeria and the Israeli-Palestinian conflict. Finally, chapter 3 uses the existing
evidence, as well as evidence generated in this report, to discuss how the poli-
cies governing trade can reduce the probability and intensity of conflicts. Two
appendixes include detailed information on the modeling framework, the data
issues and the e stimation results.
Notes
1. Fragile countries in this case are defined according to the OECD list (see box 1.1 in
chapter 1).
2. The list of these fragile countries is slightly different from that maintained by the
OECD and comes from the World Bank African Development Bank and Asian
Development Bank Harmonized List of Fragile Situations, discussed in box 1.1.
3. According to the Failed States Index 2013 (Fund for Peace 2013), three quarters of
the twenty countries most at risk of conflict are in sub-Saharan Africa.
The cross-country study provides little support for the opportunity cost
4.
hypothesis. The difference with the within country evidence is likely to be due to
two reasons. First the availability of data within countries enables us to isolate the
impact of commodity price changes on real incomes. Second, the large heteroge-
neity across countries can mask effects that may be important within individual
countries.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Overview 7
References
Anderson, K., M. Ivanic, and W. Martin. 2013. “Food Price Spikes, Price Insulation and
Poverty.” World Bank, Mimeo.
Attanasio, O., V. Di Maro, V. Lechene, and D. Phillips. 2013. “Welfare Consequences of
Food Prices Increases: Evidence from Rural Mexico.” Journal of Development Economics
104: 136–51.
Berman, N., and M. Couttenier. 2014. “External Shocks, Internal Shots: The Geography of
Civil Conflicts.” CEPR Discussion Paper 9895.
Berman, N., M. Couttenier, D. Rohner, and M. Thoenig. 2014. “This Mine Is Mine! How
Minerals Fuel Conflicts in Africa.” OxCarre Research Paper 141.
Dube, O., and J. Vargas. 2013. “Commodity Price Shocks and Civil Conflict: Evidence
from Colombia.” Review of Economic Studies 80 (4): 1384–421.
Fund for Peace. 2013. “The Failed States Index 2013.” http://ffp.statesindex.org
/rankings-2013-sortable.
Lakner, C., M. Negre, and E. B. Prydz. 2014. “Twinning the Goals: How Can Shared
Prosperity Help Reduce Global Poverty?” Policy Research Working Paper, World Bank,
forthcoming.
Maystadt, J-F., G. De Luca, P. G. Sekeris, J. Ulimwengu, and R. Folledo. 2014. “Mineral
Resources and Conflicts in DRC: A Case of Ecological Fallacy?” Oxford Economic
Papers 66 (3): 721–49. http://oep.oxfordjournals.org/content/66/3/721.
Reis, J. G., and T. Farole. 2012. Trade Competitiveness Diagnostic Toolkit. Washington, DC:
World Bank.
World Bank. 2011a. World Development Report 2011: Conflict, Security and Development.
Washington, DC: World Bank.
———. 2011b. Middle East and North Africa: Facing Challenges and Opportunities.
Economic Developments and Prospects Report.
———. 2012. Africa Can Help Feed Africa. Washington, DC: World Bank.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Chapter 1
How Trade Can Affect Conflict
Introduction
On the face of it, it may be difficult to believe that the horrific experiences of
civil war that have plagued many poor countries over the past decades are influ-
enced by changes in international commodity prices or external demand for a
country’s exports. Nevertheless, there are many recent examples of civil conflicts
where economic motivations, along with others, appear to have played an impor-
tant role. A growing economic literature has elaborated theories of how changes
in external trade may drive conflict onset, intensity, or duration, and has tested
these theories. On the basis of this literature, this chapter aims to provide an
analytical framework for thinking about how changes in trade flows may affect
conflict, and to test this model against experiences across countries, and across
regions in two case studies (Nigeria and the West Bank and Gaza).
One reason that trade flows can have such an important impact on conflict in
fragile countries is that they are much larger than other external flows and can
have very large effects on real incomes. While trade can make an enormous con-
tribution to development by encouraging the reallocation of resources to more
productive activities, changes in relative prices as a result of trade may also involve
losses by workers (and their families) in declining sectors. Such losses may be
short-lived, as workers in declining sectors take up other activities that may have
benefited from the change in trade flows. However, the poor economic environ-
ment (weak rule of law, low levels of education and training, underdeveloped
financial sectors) in fragile countries often limits workers’ ability to take advan-
tage of the opportunities opened by trade. Thus the losses in real incomes as a
result of trade can be significant and long-lasting in fragile countries. Trade in
fragile countries can also be more volatile than other sources of foreign exchange,
in part because exports are highly concentrated in primary commodities, many of
which are subject to large and frequent fluctuations in price. Moreover, fragile
countries are highly dependent on food imports, where changes in prices can have
immediate—and at times dire—implications for large portions of the population.
Abrupt changes in trade can affect conflict through three distinct mechanisms.
The opportunity cost mechanism refers to the tendency for declines in real
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 9
10 How Trade Can Affect Conflict
incomes to reduce the income foregone by those choosing to engage in conflict.
Thus declines in export prices, increases in import prices, and declines in external
demand that reduce real incomes are associated with greater conflict. The rapac-
ity effect describes how increases in price can encourage violent competition for
point-source commodities, for example oil or diamonds. And the resource effect
refers to how increases in the value of goods subject to government (or rebel)
taxation can provide the means to suppress (or enhance) violence.
While these three motivations are analytically distinct, measuring them is
particularly challenging. First, changes in the value of some commodities can
have cross-cutting effects. For example, a rise in the price of an export commod-
ity that is also consumed may raise producers’ real incomes but reduce the real
incomes of consuming households. Thus, even if the opportunity cost mechanism
is important in motivating participation (or not) in conflict, it may be difficult to
identify this in the data. Second, the impact on conflict of trade changes may
differ depending on differences in local conditions. An increase in oil exports may
boost conflict in oil-producing Nigeria, but not in oil-producing Norway.
The cross-country analysis in this chapter finds the strongest evidence for the
rapacity effect. In a significant departure from some of the most recent literature,
we find that price swings of exported commodities do matter for the probability
of conflict. An exogenous increase in the value of a country’s exported commodi-
ties raises the probability of a civil conflict erupting in that country. The effect is
far from negligible: an increase of 10 percent in the value of exports raises the risk
of conflict by between 2.2 and 2.5 percentage points. This result is primarily
driven by competition for point-source commodities that experience rising prices.
By contrast, the cross-country evidence provides little support for the view
that conflict is fueled by reductions in real income due to commodity price
changes (the opportunity cost mechanism). Neither changes in the prices of
export commodities that are not the potential objects of predation, nor the prices
of imported commodities, nor changes in demand in export markets appear to
exert any influence on the probability or the duration of conflict. However, the
cross-country estimations do provide some indication that export and import
prices may affect conflict intensity in the direction anticipated by the opportu-
nity cost theory. This confirms previous evidence that triggering a new conflict is
more difficult than escalating an existing one (Bazzi and Blattman 2014).
The country case studies provide clearer evidence of the opportunity cost
effect, in part because of the availability of more detailed data than is feasible in
cross-country analysis (Blattman and Miguel 2010). In Nigeria, data on the com-
modity composition of household consumption and production allows the con-
struction of price indices that accurately reflect the impact of commodity price
changes on real incomes. The estimations find that increases in the prices of
commodities that are important in household production (consumption) are
negatively (positively) associated with conflict. Measuring how changes in com-
modity prices affect both production and consumption (which many studies fail
to do) is critical to accurately identifying the importance of the opportunity cost
effect in driving conflict.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 11
The Nigeria analysis also finds evidence of the rapacity effect through the
positive impact of rising oil prices on conflict. However, this effect disappears
when considering the period after the amnesty deal was signed between the state
and the militant groups in the Niger Delta, suggesting that the state may have
been able to use oil revenues to counteract the rapacity effect, at least in the
short run. Changes in the prices of both production and consumption items also
have had a major impact on the intensity of the Boko Haram conflict since 2010.
In the West Bank and Gaza, sharp changes in export revenues in the late
1990s were driven by the emergence of new foreign suppliers, chiefly China,
and the Israeli trade liberalization, which reduced Palestinians’ preferential
access to the Israeli market. Information on these changes in export revenues by
economic sector are linked to data on the sectoral composition of employment
in each locality. An increase of $10 million in export revenues (before the
Second Intifada) in a sector that accounts for at least 10 percent of private
employment in a locality reduces the number of conflict-related fatalities in that
(during the Second Intifada) by 2.1 percent. The finding that improved
locality
employment prospects are linked to lower conflict-related fatalities supports the
opportunity cost hypothesis. The fact that Palestinian exports do not include
point-source commodities facilitated identifying the importance of the oppor-
tunity cost effect.
Trading with neighboring countries was also found to be significantly related
to conflict. Higher levels of trade with neighbors reduce the duration, as well as
the intensity, of conflict, because such trade reduces the incentives of contiguous
countries to fuel conflict in their neighbors. Importantly, trading with neighbors
is also associated with a lower risk of conflict when such trade occurs under
regional trade agreements (RTAs), although it is hard to determine causality for
this result. Further, the incidence of conflict in neighboring countries is signifi-
cantly and positively related to conflict in the country of interest. The influence
of neighboring countries on conflict is a frequently observed characteristic of
many modern conflicts.
Of course, civil war does not occur in a vacuum, and changes in trade prices
and volumes that may spur conflict in some contexts may have no impact at all
on conflict in others. Diamonds have been a curse in Angola but a blessing in
Botswana. Chapter 2 is devoted to understanding how different factors may
affect the relationship between changes in trade flows and conflict.
The next section of the chapter argues that the size, volatility, and limited
diversification of trade flows in fragile countries may magnify their impact on civil
conflict. The section titled “Why Changes in Trade Flows May Affect Conflict”
outlines an analytical framework that details the channels through which changes
in trade flows affect decisions on whether to engage in civil violence. The section
on “Cross-Country Evidence on Trade Shocks and Conflict” tests this framework,
using a dataset on the occurrence of civil conflict across countries. The section on
“Evidence from Nigerian States” applies a similar empirical test to the incidence
of conflict across regions in Nigeria, and the section on “Evidence from the Israeli-
Palestinian Conflict” does the same for the West Bank and Gaza.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
12 How Trade Can Affect Conflict
Trade Flows in Fragile Countries Are Different
Sharp changes in trade prices and volumes can be an important trigger of instability
in fragile countries (see box 1.1 for a definition of fragile countries). Trade flows in
these countries are much larger than other sources of foreign exchange (i.e. official
development assistance—ODA, remittances, and foreign direct investment—FDI)
(figure 1.1). Trade in fragile countries is also more volatile than these other foreign
exchange flows. FDI, ODA, and remittances to fragile countries exhibited little
volatility and increased steadily from 2000 to 2010. Remittances, for example, rose
from just under $9 billion in 2000 to over $47 billion by 2010 (OECD 2013). By
contrast, trade flows have fluctuated to a larger degree, and these swings have
resulted in much larger absolute changes than for any other external flow. For
example, the global crisis reduced the trade-GDP ratio in fragile countries with
available data by almost 10 percentage points in 2009, and the dollar value of the
fall in trade was almost half the total value of other inflows to fragile countries.
Box 1.1 Which Are the Fragile Countries?
There is no consensus in the development community on what specific characteristics are nec-
essary to classify a country as fragile, which has led to varying definitions and lists of fragile
countries. However, each definition is predicated to some degree on the existence, relative
weakness, or lack of governance and institutional capacity.
For example, the joint World Bank/African Development Bank/Asian Development Bank
Harmonized List of Fragile and Conflict-Affected Situations (FCS) includes all low-income
countries and territories eligible for World Bank assistance with a score of 3.2 or lower on the
internally generated Country Policy and Institutional Assessment (CPIA),a a diagnostic tool
intended to measure the policies, institutional arrangements, and other key elements within a
country’s control that support sustainable growth, poverty reduction, and the effective use of
development assistance.b It also includes countries with the presence of a regional or UN
peacekeeping mission within the last three years.
The OECD extends this list to cover states meeting the following definition of fragility:
“a state is understood to be fragile when it is unable to meet its population’s expectations or
manage changes in expectation and capacity through the political process.”c The OECD’s most
recent list of 51 fragile countries is a compilation of two lists: the aforementioned Harmonized
List of Fragile Situations, as well as the 2011 Failed States Index (FSI), prepared by the Fund for
Peace and published by Foreign Policy. The OECD estimates that although “one-fifth (about
18.5 percent) of the world’s population lived in fragile states in 2010, these countries hosted
about one-third of the world’s poor (400 million out of 1.2 billion)” (OECD 2013).
Other, reasonable criteria for defining fragility could encompass even more countries. It is
well documented, for instance, that countries which have recently experienced conflict are
more likely to relapse into conflict (World Bank 2011). One approach to capturing this higher
risk in a definition of fragility is to take into account the number of recent battlefield deaths.
Between 2005 and 2010, Sri Lanka and Pakistan—neither of which appears on the World Bank’s
box continues next page
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 13
Box 1.1 Which Are the Fragile Countries? (continued)
list of fragile countries—experienced 8,413 and 6,688 battlefield deaths in a single year,
respectively. Both figures exceed the single-year total of 6,238 recorded in Afghanistan, which
topped the World Bank’s list of “deadliest” fragile states during that period. Table A.1 in
appendix A lists the countries considered fragile by the World Bank (in 2013), alongside a list of
other countries with at least one year of minor conflict (measured according to the Gleditsch
et al. (2002) definition as a year with at least 25 battlefield deaths) from 2005 to 2010.
a. See http://go.worldbank.org/NEK8GNPSO0.
b. http://go.worldbank.org/EEAIU81ZG0.
c. OECD glossary of International Network on Conflict and Fragility (INCAF) terms, http://www.oecd.org/document/13/0,3746,
en_2649_33693550_49377421_1_1_1_1,00.html.
Figure 1.1 Trade Represents the Major Source of Foreign Exchange in Fragile States
Percent of GDP
100
80
60
40
20
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Trade Remittances ODA FDI
Sources: World Development Indicators; Reflects 22 countries from the 2015 OECD list of fragile states for which data were available.
Note: Trade is defined as exports of goods and services plus imports of goods and services. FDI = foreign direct investment; ODA = official
development assistance.
Fragile countries may be more vulnerable to changes in trade flows than
many other developing countries, due to low levels of export diversification.
Fragile countries’ export markets and products are more concentrated than in
other developing countries (table 1.1).1 In more than half of the fragile coun-
tries with adequate data, the largest export accounts for over a quarter of mer-
chandise exports (figure 1.2). Several fragile countries’ exports are dominated
by only a few products to a much greater extent than comparable nonfragile
countries (figure 1.3). For example, in 2010, T-shirts, sweatshirts, and suits
accounted for 76 percent of Haiti’s exports; 87 percent of exports from the
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
14 How Trade Can Affect Conflict
Table 1.1 Fragile Countries’ Exports Are Less Diversified Than Other Developing Countries’
Exports
Index of concentration
Type of diversification Year Mean Median
Market (fragile states) 2012 0.26 0.21
Market (all developing countries) 2012 0.24 0.17
Product (fragile states) 2012 0.33 0.21
Product (all developing countries) 2012 0.25 0.16
Source: Authors’ elaboration based on UN Commodity Trade Statistics Database (HS 6 digit product classification).
Note: fragile states based on the OECD’s list of fragile states. The index is a flow-weighted concentration index normalized to
range between 0 and 1, with a higher level indicating higher concentration. Due to lack of some country’s export data,
“mirror data” is used (partner’s imports from that country).
Figure 1.2 Share of Largest Exports in Selected Fragile Countries and Territories (in 2010)
Yemen, Rep.
Congo, Rep.
Burundi
Central African Republic
Kiribati
Myanmar
Syrian Arab Republic
Afghanistan
Togo
Zimbabwe
West Bank and Gaza
Nepal
Bosnia and Herzegovina
0 10 20 30 40 50 60 70
Percent
Petroleum oils Natural gas Stone
Coffee Carpets of wool Electrical energy
Diamonds Cotton
Crude coconut oil Banknotes
Source: Authors’ elaboration based on UN Commodity Trade Statistics Database.
Central African Republic were wrapped up in only four product lines: dia-
monds (32 percent), raw wood (30 percent), sawn/chipped wood (15 percent),
and cotton (10 percent); and Iraq’s economy relies for all intents and purposes
solely on crude oil.2 On the other hand other developing countries (Honduras,
Moldova, and Peru) with similar population to each of these countries (but with
higher GDP per capita) have a considerably more diversified export basket
(figure 1.3).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Figure 1.3 For Many Fragile States, Exports Are Not Heavily Diversified
36.56%
29.41%
Source: MIT’s Observatory of Economic Complexity at atlas.media.mit.edu.
15
16 How Trade Can Affect Conflict
Moreover, fragile countries are highly dependent on primary commodity
exports, which are subject to considerable price volatility. Sharp changes in inter-
national commodity prices can have important implications for employment,
income, and investment in fragile countries. For example, exports from Benin,
Chad, and Mali grew by 30 percent following the increase in the world price of
cotton from 1994 to 1996 and declined by as much as 20 percent with the drop
in cotton prices between 1997 and 1999 (FAO 2002).
In addition, fragile countries’ high levels of food imports contribute to food
insecurity (Aksoy and Ng 2008). Food accounted for nearly 17 percent of all
fragile country imports in 2010, compared to less than 14 percent for other
developing countries.3 To the extent that international food price variability
can be a destabilizing factor (Arezki and Brückner 2011), this higher
dependence on food imports can increase the sensitivity of fragile countries to
trade shocks (figure 1.4).
Compounding these problems, fragile countries also perform particularly
poorly in terms of trade facilitation. A fragile country has lower scores on every
available indicator of trade facilitation relative to a country in the same region
and income group (left panel in figure 1.5). This penalty varies between
5 percent and 8 percent for the Logistics Performance Indicator (LPI) and
between 7 percent and 12 percent for the perception index of port
Figure 1.4 For Fragile States, Net Food Imports Constitute a Higher Percentage of GDP
2.5
2.0
1.5
Net food imports (% GDP)
1.0
0.5
0
–0.5
–1.0
2008 2009 2010 2011 2012
Nonfragile developing countries Fragile states (OECD 2013)
Sources: Author calculations based on UN COMTRADE data (via WITS) and WDI data.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 17
Figure 1.5 Fragile Countries Perform Worse Than Their Peers in Trade Facilitation and the Gap Is Growing
a. Trade facilitation measures b. LPI over time
30 –5.0
25 –5.5
20
–6.0
15
–6.5
10
Percent
Percent
–7.0
5
–7.5
0
–8.0
–5
–10 –8.5
–15 –9.0
LPI DB trade rank Port 2007 2010 2012 2014
infrastructure
FCS (WB list) FCS (OECD list)
Note: In the left panel the y-axis measures the average percentage difference of fragile countries relative to their income and regional peers across
various indicators of trade facilitation (World Bank 2014c), rank of World Bank Doing Business Trading Across Borders indicators (World Bank 2014a),
and World Economic Forum perception of port infrastructure quality (World Economic Forum 2014). This penalty is computed as the coefficient
of a fragile country dummy in a regression of the log of the indicator on income group, regional and year dummies as well as the fragile dummy.
Only countries for which LPI data are available in all years are included (constant sample of 130 countries). In the right panel the y-axis measures
the average percentage penalty of fragile countries (according to the World Bank list) relative to their income and regional peers (only countries
for which LPI data are available in all years are included, yielding a constant sample of 130 countries). All data on indicators come from World Bank
(2014b) except for the 2014 LPI data, which come from World Bank (2014c).
infrastructure according to the fragile classification one uses (i.e. OECD or
World Bank/regional development banks list). Similarly, being a fragile country
is associated with a 22–24 percent higher rank in the Doing Business “Trading
Across Borders” ranking (again relative to a similar countries for per capita
income and regional group). Worryingly, this penalty has been increasing over
time in terms of the LPI, which is the indicator with the greatest coverage of
countries-years (right panel in figure 1.5). This poor trade facilitation perfor-
mance makes export growth and diversification even more challenging for
fragile countries. In addition it creates a further penalty for the consumers in
these countries relying on imported goods.
The combination of heavy reliance on exports as a source of foreign exchange,
limited export diversification, reliance on volatile primary commodity exports,
and dependence on food imports means that trade can play a major role in trig-
gering conflict in fragile countries. However, even large, abrupt changes in trade
flows, by themselves, do not cause conflicts. Rather, changes in trade can interact
with existing tensions, for example ethnic rivalries or regional differences, which
may well be sufficient to incite conflict on their own. The next sections will try
to disentangle the channels through which changes in trade can affect country
stability. Chapter 2 considers the conditions which make countries more sensi-
tive to these shocks.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
18 How Trade Can Affect Conflict
Why Changes in Trade Flows May Affect Conflict
There are at least three main mechanisms through which economic shocks,
including trade-related changes, can affect political instability. We refer to these
as opportunity cost, rapacity, and resource effects.
First, changes in trade flows can change real incomes. As in Becker’s (1968)
seminal work on the economics of crime, an individual’s real income can be seen
as his opportunity cost of engaging in a rebellious activity.4 For example, a fall in
the price of a key export commodity can reduce employment in that sector, thus
reducing the income that workers in that sector must forfeit by engaging in con-
flict. Alternatively, rising prices of a commodity important for household produc-
tion will increase the household’s income and reduce their willingness to
participate in conflict. More generally, the opportunity cost mechanism describes
how changes in real incomes affect the willingness to participate in conflict
through changing the relative return on conflict activities compared to more
peaceful pursuits. This relatively dry language should not be taken to imply that
the choice to participate in conflict is free of compulsion. A father who joins a
rebel group after his livelihood is destroyed and his family begins to starve is
motivated by what we call the opportunity cost mechanism. But he may perceive
little choice in the matter. We should also note that a decline in real incomes may
also encourage violence due to the resentment and frustration from experiencing
a fall in social status or a deteriorating ability to care for one’s family, driven by
economic forces over which the individual lacks any control.
Second, civil conflicts are also fought over the control of valuable economic
resources. The rapacity effect refers to the willingness to engage in conflict to
control the production of commodities, such as oil or minerals, which do not
require massive amounts of labor, are highly valuable, are not perishable, and are
easily controlled. These point-source resources are generally traded in interna-
tional markets and subject to large swings in prices that affect their value, and
thus the willingness to fight to obtain them. In addition to purely economic
motives, individuals may turn to violence to protest the often serious social and
environmental consequences of the exploitation of oil or minerals. However, the
evidence presented in the remainder of the report is more consistent with the
rapacity effect than with these environmental and social consequences.
Third, the resource effect refers to how changes in the value of traded goods
will affect civil conflicts if the state or the rebels can rely on them to fund violent
activities. For example, the government may be able to capture substantial rev-
enues from oil rents, or rebels may be able to extract a portion of increased
agricultural prices from farmers in areas they control.
It is useful to distinguish between three types of trade-related change that
affect conflict: 1) changes in international commodity prices; 2) changes in trad-
ing conditions; and 3) changes in trade with neighboring countries. Figure 1.6
links these changes to the incentives of the actors to engage in conflict through
the three mechanisms described above. We explore the effects of each type of
change in turn.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 19
Figure 1.6 Mapping the Linkages between Changes in Trade Flows and Civil Conflict
Δ International Δ Trading conditions (e.g. Δ tariff, Δ NTMs, Δ Trade with
Type of trade shock
commodity prices Δ market access, Δ trade facilitation) neighbors
Channels
Δ Real incomes, Δ value of economic resources, Δ value of state revenues
Mediating factors
• Types of state institutions
• Grievances
• Transmission of prices
• Conditions in neighboring
countries
Extractive point resources Labor-intensive diffuse Major consumption
Type of commodity
(e.g. oil, diamonds) resources (e.g. rice, cocoa) commodities
Export sector Import sector
Δ Relative cost of labor
Type of effect Δ Rapacity effect Δ Opportunity cost of
Δ Resources effect
(i.e. the value of the “prize”) engaging in conflict
Δ Conflict/fragility
Commodity Export Prices
Changes in international commodity prices have drawn the most attention in the
literature examining the impact of trade-related changes on conflict. Although
the focus of the literature has been mainly on the export side, these price changes
are likely to have an impact via the import side as well.
Changes in the international price of an export commodity may affect
incentives to engage in conflict through the three mechanisms given above
(opportunity cost, rapacity, and resources). As in the case of other changes in real
incomes, the potential for changes in commodity prices to affect conflict has
received mixed empirical support. Bruckner and Ciccone (2010) find that a
reduction in the international price of a country’s main commodity export leads
to a higher chance of civil conflict in sub-Saharan Africa. Savun and Tirone
(2012) show similar evidence for a larger sample of countries. This relationship is
generally supported in recent within-country work, which studies the variation in
conflict and production across subnational units. Dube and Vargas (2013), for
example, find that during the 1990s a reduction in the price of coffee, the largest
labor-intensive commodity exported by Colombia, increased the intensity of
conflict by more in the municipalities specialized in coffee production than in the
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
20 How Trade Can Affect Conflict
others. Using data on small subunits across 48 sub-Saharan African countries,
Berman and Couttenier (2014) find that an increase in a region’s main agricul-
tural commodity exports (measured by data on imports from partner countries)
decreases the probability of a conflict and its intensity. Maystadt and Ecker
(2014) find that reductions in the price of livestock (driven by droughts rather
than international markets) substantially increase the incidence of conflict across
regions in Somalia.
On the other hand, evidence from other cross-country studies does not sup-
port the view that declines in real income are associated with greater willingness
to participate in civil conflicts (the opportunity cost mechanism). Besley and
Persson (2008) find that the price of exported commodities is positively associ-
ated with the incidence of civil conflict, a result that the authors interpret as
evidence of the rapacity effect. Bazzi and Blattman (2014) find no robust rela-
tionship between changes in the international prices of commodity exports and
either the beginning or ending of conflict across a large sample of developing
countries, although they do find some evidence that these price changes may
affect conflict intensity.
These contrasting cross-country findings suggest that the impact of commod-
ity prices on conflict varies substantially by the type of commodity (and accord-
ing to local conditions, an issue we address in chapter 2). Indeed, the conceptual
framework illustrated above suggests that all commodities are not created equal
when it comes to their effects on conflict: increases in the value of some com-
modities may increase conflict through the rapacity effect, while increases in the
value of others may reduce (or increase) conflict through the opportunity cost
mechanism. A rise in the price of a commodity whose control can be relatively
easily appropriated can foster conflict by increasing the potential prize of the
conflict, thus raising the incentive for fighting (the rapacity effect). This is usually
the case for so-called point-source resources such as minerals and fuels, which
are contestable, highly valuable, capital-intensive, and geographically concentrated
resources. At the other end of the spectrum, increases in the price of what we
refer to as “diffuse” commodities (often agricultural commodities) that are
important in household production and are produced over wide areas, labor
intensive, and more difficult (though not impossible) to control, may reduce
conflict by raising the opportunity cost of participating in a rebellion. Dal Bó and
Dal Bó (2011) develop a general equilibrium model to formalize this intuition.
The “rapacity effect” is part of the explanation for the eruption and/or the
escalation of violence in many modern conflicts. As in the case of the opportunity
cost mechanism, the evidence in support of the rapacity effect is stronger within
countries than across them. Maystadt et al. (2014) show that new mining conces-
sions spurred by increases in international mineral prices increase the level of
violence across districts in the Democratic Republic of Congo. Similarly, Dube
and Vargas (2013) find that increases in oil prices are associated with higher
violence across Colombian municipalities. Bellows and Miguel (2009) show that
the presence of diamonds was associated with higher violence during the civil
war in Sierra Leone.5
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 21
Other cross-country studies provide some evidence on the conflict-inducing
effects of oil resources in Africa (Buhaug and Rod 2006) and in low-income
countries (Lin and Michaels 2011).6 On the other hand, neither Cotet and Tsui
(2013) nor Bazzi and Blattman (2014) find any cross-country evidence support-
ive of the hypothesis that larger values of extractive resources are associated with
higher levels of conflict. One reason for these mixed results could be that
increases in value of disputable resources can generate higher fiscal revenues. The
state could use these revenues to strengthen its military capacity to repress rebel
groups’ activities and/or to buy off support, thus favoring political stability.7
On the other hand, prices of diffused agricultural commodities may be nega-
tively related to conflict. Their production is labor intensive and more difficult
than point-source commodities for the government to tax. Thus rising prices of
diffuse commodities should raise incomes, thus increasing the opportunity cost
of conflict. However, revenues from diffused commodities (as well as from min-
ing activities) can also be an important source of funding for rebel groups con-
trolling local areas. For example, in Myanmar the production and trade of timber
and other agricultural products, as well as of mining products in the bordering
areas with Thailand and China, was taxed by local rebel groups fighting the cen-
tral government.8 Whether on balance rising prices of diffuse agricultural com-
modities reduce conflict by increasing the opportunity cost facing would-be
rebels, or increase conflict by helping to fund rebel groups, is an empirical ques-
tion. In the case of the Colombian conflict the opportunity cost channel appears
to be dominant for legal crops such as coffee and bananas (Dube and Vargas
2013), while the rebel funding mechanism is more important for coca produc-
tion (Angrist and Kugler 2008).
Another potentially important distinction is between commodities that are
important consumption items domestically (e.g. rice and fuel) and those that are
not (e.g. diamonds and cocoa). A rise in price would benefit producers, but
would also penalize consumers, potentially sparking unrest. It is possible that the
majority of households are net consumers in countries that are net exporters of
that commodity. It is important to take into account this distributional impact of
price changes in the identification of the effects on conflict. Bellemare (2011)
provides some support for this relationship by linking monthly spikes in interna-
tional food prices with increased political unrest worldwide.
This classification, as well as the distinction between point-source and diffuse
commodities, generates a matrix of four types of export commodities (table 1.2).
The empirical analysis below tests for differences in the effects of price changes
on conflict across these groups of commodities.
Table 1.2 Classification of the Export Commodities (with Example)
Point-source Diffuse
Consumed Oil Rice
Not consumed Diamonds Cocoa
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
22 How Trade Can Affect Conflict
Prices of Imported Consumption Goods
The economic literature on the impact of commodity price changes on conflict
has focused mostly on the prices of fragile countries’ exports. Nevertheless,
changes in the international prices of imported commodities could affect con-
sumers, and thus their incentive to participate in violence, as much as prices of
exported commodities affect producers. The mechanisms at play are similar to
those for commodity exports, but with opposite “signs.” As the prices of com-
modities tend to be correlated, failure to consider imported commodities’ prices
may lead to a bias in the measurement of the effect on conflict of the prices of
exported commodities. In line with these findings, our analysis considers also the
impact of the price of imported commodities on conflict.
The size of these effects depends on the share in household consumption
and production of the commodities whose international price has changed.
Ivanic and Martin (2008) find, for a sample of low-income countries, that the
hike in international staple food prices in 2007–08 induced much more fre-
quent and larger poverty increases than poverty reductions in a sample of
low-income countries. As discussed above, fragile countries are particularly
vulnerable to such food price fluctuations, as most of them are net food
importing countries. Aksoy and Ng (2008) argue that the international com-
munity’s efforts in promoting food security should focus on conflict countries,
which exhibit the largest food deficits in the broad category of net food
importers.
Recent studies have found some empirical support for the idea that changes
in international food prices have affected conflict in fragile countries. To the best
of our knowledge there is no evidence on other imported commodities. Arezki
and Bruckner (2011) find that increases in the prices of imported food led to
higher levels of antigovernment riots and civil conflict in low-income countries
from 1970 to 2007. By contrast, changes in imported food prices had no impact
on conflict in high-income countries. Maystadt, Trinh Tan, and Breisinger (2014)
show this effect to be particularly strong for Arab countries, which are major
food importers. Bellemare (2011) uses a different strategy and shows that
monthly spikes in international food prices between January 1990 and January
2011 led to increased political unrest worldwide (as measured by
food-related riots).
Changes in Trading Conditions
The literature on the impact of changes in trade flows on conflict has focused
mainly on swings in commodity prices. However, other trade-related changes,
such as sharp changes in demand in destination markets, changes in a country’s
own trade policy and that of trading partners, and changes in the geography of
trade also are potentially relevant for conflict.
Changes in trading conditions with main partners may affect conflict in ways
analogous to changes in commodity prices. Consider, for example, a change in
a country’s access to a foreign market. A reduction in the preferential access to
a major exporting market X for a country could reduce the country’s exports
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 23
to X, thereby also reducing the incomes and employment of those involved in
the production of the goods exported to X.9 In general, any trade-related
change that has a large enough impact on domestic incomes and employment
opportunities could generate the same effects on conflict as the commodity
price changes described above.
We can identify four types of trade-related changes that may matter here:
1. changes in demand in a country’s trading partners, which change the demand
for products from that country;
2. changes in the country’s access to international markets;
3. changes in the country’s domestic level of protection of goods and services;
4. changes in trade facilitation.
Changes in (1) and (2) have analytically similar effects. Increases in the main
trading partners’ demand (e.g. as a consequence of a rise in income) would
increase the demand for the country’s goods and services, all else equal. Similarly,
an improvement in the country’s preferential access to its main markets (e.g. as
a consequence of a bilateral trade agreement) would also raise the demand for
the country’s products. In terms of the conceptual framework described above,
these increases in demand would have the same effects as a change in commodity
prices. Therefore, it matters which type of export experiences the rise in demand.
If the higher demand is for diffuse agricultural commodities, e.g. cocoa, or labor-
intensive manufacturing, e.g. textiles, the rise in real incomes could increase the
opportunity cost of conflict, thus reducing incentives to engage in conflict. If the
higher demand is for point-source capital-intensive commodities, such as oil,
then rising demand could increase conflict through the rapacity effect, by raising
the value of the prize.
The limited evidence available supports the view that higher external demand
for a country’s exports reduces conflict. Bruckner and Ciccone (2010) find that
economic downturns in the main OECD export destinations of sub-Saharan
African countries’ exports are associated with a higher probability of an outbreak
of conflict. Chaudion, Peskowitz, and Stanton (2012) find similar effects on the
onset as well as the intensity of conflict across a large sample of countries.
Berman and Couttenier (2014) find evidence that banking crises that reduce
demand in export destinations increase conflict. In the only study we have found
that focuses on market access, Berman and Couttenier (2014) show that
enhanced preferential access to the U.S. market through the Africa Growth and
Opportunity Act (AGOA) reduced conflict across eight African countries, espe-
cially in those countries with a high share of exports in products eligible under
AGOA. The finding that increases in demand are associated with a reduction in
conflict is consistent with the fact that most of the agricultural and manufactured
goods exported by fragile countries are relatively labor intensive, so that increases
in their price directly benefit workers.
Another trade-related change to consider is the impact of a country’s own
trade policies on the production and consumption of traded goods. Trade
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
24 How Trade Can Affect Conflict
restrictive measures, such as tariff increases or non-tariff measures (NTMs), may
benefit those domestic producers who compete with imports. However, that
advantage comes at the expense of higher prices facing the users of those prod-
ucts. Conversely, restrictions on exports may benefit consumers by lowering the
domestic price of the restricted good, while reducing income for producers.10
These trade restrictive measures usually have a net welfare-reducing effect.
Recent evidence from Africa shows that NTMs have increased poverty owing to
higher domestic prices (Cadot and Gourdon 2012; Kelleher and Reyes 2014;
Treichel et al. 2012).
Unlike demand shocks in trading partners and changes in market access,
referred to as (1) and (2) above, domestic trade policy is likely to have
ambiguous effects on conflict, as it affects different groups of people in
opposing ways. The impact of these policies on the probability of conflict
will depend on the relative power and voice of these groups. For example, in
the case of trade in food products in developing countries, a distinction is
typically made between urban dwellers who are net consumers of food, and
rural dwellers who are net producers of food. The urban group is usually
more able to organize and to voice its concern than the rural group.
Therefore, governments tend to implement policies, including trade policies,
which favor urban dwellers (Lipton 1977).11 There is very little systematic
evidence on the impact of such policies on conflict. One exception is
Bhavnani and Jha (2011), who examine the role of Britain’s trade policy in
the Quit India rebellion of 1942. They find that residents of districts
in British India that were negatively affected by the policy favoring British
manufactures over Indian producers were more likely to engage in violent
insurrection.
By influencing trade, trade facilitation policies, such as the strengthening of
transport infrastructure and the streamlining of border procedures, can also
potentially affect conflict via the same mechanisms described above. These poli-
cies typically reduce the cost of trading between countries, with similar effects
on trade as other policy changes. This reduction in cost could in turn increase
both imports and exports. The evidence suggests that trade facilitation, including
both hard and soft infrastructure, can have important effects on exports
(Portugal-Perez and Wilson 2012). For fragile countries there may be a particular
large scope for improvement given the large gaps in trade facilitation indicators
in these countries documented above. If the change in exports is large enough,
this could generate employment and income opportunities, which may reduce
the willingness of the population to engage in political violence. Similarly, the
increase in imports would have the same effects as an import increase spurred by
a reduction of trade protection.
Trading with Neighbors
Modern civil conflicts often involve substantial foreign participation, an aspect
that is not often highlighted in the economics literature. Gersovitz and Kriger
(2013) argue that almost all recent, major civil conflicts in Africa are more
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 25
properly viewed as part of a “regional war complex” than as purely domestic
conflicts.12 The authors provide many examples of foreign participation in civil
conflicts, such as the role of Côte d’Ivoire in Taylor’s invasion of Liberia in 1989,
which generated the Liberian civil war, and the role of South Africa and
Zimbabwe in the conflict between Renamo and Frelimo in Mozambique.
Consistent with this evidence, Gleditsch (2007) finds that the transnational
linkages between a country and regional countries strongly influence the risk of
civil conflict. Within the regional context, it is predominantly the neighboring
countries that exert an influence on domestic civil conflicts (Buhaug and
Gleditsch 2008).
It is thus likely that trade with neighboring countries can play an important
role in civil conflicts. A high volume of trade between two neighbors A and B
increases the costs to A of a conflict in B, thus reducing the likelihood that
A would intervene to foment civil conflict in B (and vice versa).13 Trade may also
raise the level of trust between groups in different countries (Rohner, Thoenig,
and Zilibotti 2013), for example because trade relations may require learning the
language or the customs of the other group, thus reducing the likelihood of inter-
ventions in support of civil conflicts.14 Gleditsch (2007) finds that greater trade
integration with a country’s neighbors substantially reduces the risk of civil war
in that country.15
This result, while preliminary, underlines the importance of trade integration
between neighbors, especially in more fragile contexts. Improved trade facilita-
tion and trade agreements with neighboring countries could help reduce the risk
of civil conflict. This is one of the rationales behind the trade integration pro-
grams funded by the World Bank in the Great Lakes region, an area ridden with
long-standing regional conflicts.
Cross-Country Evidence on Trade Shocks and Conflict
We offer two different approaches to testing some of the theoretical
hypotheses that emerge above. This section considers whether a relationship
between changes in trade and the onset, duration, or intensity of conflict can
be identified across countries. The next two sections present evidence on this
relationship from two country studies, on Nigeria and the Israeli-Palestinian
conflict in the West Bank and Gaza. Each of these analyses covers some, but
not all, of the trade-related changes that may affect conflict. The cross-
country analysis considers imported and exported commodity prices, changes
in economic conditions in major trading partners, and trading with neighbors;
the Nigerian case tests for the impact of changes in the prices of produced
commodities and changes in the prices of consumed goods; and the Palestinian
case focuses on the changes in trading conditions with the major trading
partner. Our empirical work builds on the young, but growing literature on
estimating the relationship between changes in income and civil conflict,
while attempting to address some of the methodological issues raised by these
studies (box 1.2).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
26 How Trade Can Affect Conflict
Box 1.2 Empirical Issues in the Early Literature on the Relationship between
Changes in Income and Conflict
Earlier empirical studies, such as Collier and Hoeffler (1998, 2004) and Fearon and Laitin (2003),
find support for a negative relationship between income levels and shocks on one side, and
coups, violence, and war on the other. However, the interpretation of these findings differs.
Collier and Hoeffler (2004) interpret the negative relationship as a confirmation of the
opportunity cost hypothesis, namely that the cost of recruiting rebels increases with income
growth. Fearon and Laitin (2003) argue that the result is instead driven by the strong positive
association between state capacity and income. When income is low, the state’s ability to
contain possible rebellions is limited.
While these papers have been influential, their cross-country empirical work suffers from
a number of drawbacks (Blattman and Miguel 2010). Importantly, these studies do not fully
account for how the relationship between income and conflict varies, depending on coun-
try circumstances (called heterogeneity). Nor do these studies address the likelihood that
changes in income and conflict are interdependent rather than causation running only
from income changes to conflict (referred to as endogneity), which can distort empirical
estimates.
The subsequent literature has tried to address these limitations. In an analysis of the impact
of income changes on conflict in sub-Saharan Africa, Miguel, Satyanath, and Sergenti (2004)
take into account much of the heterogeneity by controlling for differences among countries
that do not vary over time, but may be important in determining the relationship between
changes in income and conflict.a To deal with endogeneity, they isolate the portion of income
changes that is explained by rainfall variation, which is not affected by the conflict. Their analy-
sis confirms a significant negative effect of income on the incidence of conflict. Since changes
in income in Africa are mainly related to labor-intensive agriculture, this result lends support to
the opportunity cost hypothesis.
This work helped trigger an interest in the use of weather shocks as an instrument for
income changes or as a direct determinant of conflict. Studies almost invariably find that large
deviations from normal weather patterns increase the probability of conflict (Hsiang and
Burke 2013). This finding is particularly clear in sub-Saharan Africa. Using small geographic
cells as the unit of analysis, Harari and La Ferrara (2012) show that negative climate changes
affect conflict incidence in Africa only during the growing season. This is consistent with the
effect channeled via changes in income.b
a. More formally, they use a fixed effects model.
b. This finding is also shared by within-country studies on the determinants of conflict at the local level in Somalia
(Maystadt and Ecker 2014), Brazil (Hidalgo et al. 2010), and India (Gawande 2012). And it also applies to cross-country
studies using different kinds of changes that affect incomes, for example the movements in foreign interest rates in
relevant partner countries used in Hull and Imai (2013).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 27
Empirical Results on Trade and Conflict Onset
Our main empirical analysis estimates the impact of various trade-related vari-
ables on the onset of conflict. We use the dataset prepared by the Uppsala
Conflict Data Programme (referred to as the PRIO dataset), and include all
examples of conflict with battle deaths above 25 per year, not just major conflicts
(see appendix A). The onset of conflict is viewed as a function of: 1) the export
price index; 2) the import price index; 3) an indicator of changes in the demand
of major trading partners; 4) the share of trade with neighbors in a country’s total
trade; 5) a set of control variables that vary over time, including the presence of
conflict since 1946, the incidence of conflict in neighboring countries and in
some specifications a coup attempt in the year before; 6) a comprehensive set of
variables that do not vary over time and may influence the probability of conflict,
such as geography, ethnicity, religion, and colonial history; 7) a set of variables
controlling for any variation over time common to all countries; 8) countries’
time trends; and 9) an error term. Appendix B provides a more formal description
of the model and estimation techniques used, along with the tables showing the
estimation results (table B.1 presents the results for our preferred specification of
the model, discussed immediately below).
We find that the export price index is positively and significantly associated
with the onset of conflict. We test both contemporaneous and lagged increases
in prices, with the positive relationship mainly driven by the contemporaneous
variable. A one standard deviation increase in the export price index raises the
probability of conflict by 4 percent in the same year.16 The signs of the lagged
coefficients are consistent with the negative autocorrelation of commodity prices
(i.e. the coefficient on the export price index in t−1 is negative and that on t−2
is positive). Importantly, the sum of the three coefficients for the export price
index (the contemporaneous coefficient and two lags) is positive and significant.
It suggests that an increase of 10 percent in the export price index raises the risk
of conflict by 2.2 percentage points.
The positive and significant effect of the export price index on conflict con-
tradicts the finding in the similar analysis of Bazzi and Blattman (2014), where
the coefficient on the commodity price index was not significant.17 As the esti-
mation strategy and the data are comparable, this difference has to do with the
different way the price index is computed. Bazzi and Blattman (2014) use the
change in the price index, while we use the level of the price index. Indeed when
we compute the coefficient on the change in the price index, its contemporane-
ous coefficient becomes less significant and the sum of the contemporaneous and
lagged coefficients becomes not significant, as in Bazzi and Blattman (2014).
There is good reason to believe that it is the price level rather than its propor-
tionate change over the previous period that matters most in shaping the incen-
tives to engage in violence. Consider for instance a change in the price of oil for
an oil-exporting country. During periods of low international prices, even a large
percentage change in price in one period may be associated with a low price level
at the end of the period. In this case, the value of the oil vulnerable to predation
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
28 How Trade Can Affect Conflict
would still be limited, thus keeping the incentives for fighting over its control
relatively low.18 We therefore believe that our price index is more suited to cap-
ture changes in the incentives to engage in conflict due to commodity price
changes. This approach is also in line with other recent studies, for example
Nunn and Qian (2014) and Dube and Vargas (2013).
We argue that the positive relation between the export price index and con-
flict onset can also in part explain the timing of the recent civil war in Republic
of South Sudan (box 1.3).
As expected, the estimated relationship between conflict and the import price
index is positive: higher import prices reduce real incomes, thus reducing the
opportunity cost of conflict. However, the coefficient is not statistically signifi-
cant. This finding differs from the significant, positive impact found by Arezki
Box 1.3 The South Sudanese Civil War: Was Oil Export the Trigger?
Republic of South Sudan obtained its independence in 2011 after long years of fighting against
Sudan, which culminated in a UN-supervised popular referendum. The country has experi-
enced a very tormented period since independence. After a war with Sudan in 2012 in the oil
rich regions on the border, it descended into civil war in December 2013. The dispute started
following the sacking of the vice-president Riek Machar by the president Salva Kiir, both mem-
bers of the Sudan People’s Liberation Movement (S.P.L.M.) but long-standing political oppo-
nents. This move accelerated the collapse of the fragile government’s balance of power
established also along ethnic lines (Kiir belongs to the Dinka, the country’s largest ethnic group,
while Machar belongs to the second largest group, the Nuer). The fighting erupted in the capi-
tal Juba with an alleged coup attempt led by Machar but has since extended to much of the
country, especially in the oil-producing regions of the north. By early January 2014 the violence
had caused over 10,000 deaths and hundreds of thousands of internally displaced people.
The political roots of the war are clear, reflecting the unsuccessful state-building process so
far and the divisions within the S.P.L.M. However, the triggers behind the war’s outbreak are
less clear. What triggered the political crisis in July after two years since independence in which
the divisions of power within the government and the party had been fairly stable?
One possible explanation, which fits with the findings in this report, is the “rapacity effect.”
The political divisions between Kiir and Machar became salient again once oil exports to
Sudan resumed in April 2013. At that point the value of the unchecked control over the state,
whose fiscal revenues depend entirely on the oil exports, increased dramatically. Given the
absence of any transparency and accountability (there is not even a regular public disclosure
of the actual petroleum sales), this control is particularly appealing in Republic of South Sudan
with the government enjoying complete discretion over the management of the oil revenues.
Indeed, accusations of embezzlement of public funds from oil revenues have been frequent at
the highest political level.a By sacking Machar, Kiir ensured that the control of these revenues
would not have to be shared with his political opponent and his faction.
box continues next page
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 29
Box 1.3 The South Sudanese Civil War: Was Oil Export the Trigger? (continued)
A comparison with another oil dependent country in post-independence transition, Timor-
Leste, lends credit to this hypothesis. The country achieved independence in 2002. Its transi-
tion has also been marred by some violence in May–June 2006 (though not on the scale of that
in Republic of South Sudan), but this does not seem related to the swings in oil revenues. Since
Timor-Leste’s independence oil exports have continued to grow and even the discovery of the
large Bayu-Undan oil and gas field in 2004 did not disrupt the political context.
Why haven’t changes in oil exports triggered instability in Timor-Leste? One important
difference with Republic of South Sudan is the transparent way in which the oil revenues are
managed in Timor-Leste, which reduces the government discretion over the spending of the
revenues. The Revenue Watch Institute (2013) rates Timor-Leste among the top and Republic
of South Sudan among the lowest countries in terms of the quality of governance in the oil,
gas, and mining sector. This high quality (unprecedented among fragile countries) was
achieved also through the set up—with the World Bank assistance—of a sovereign wealth
fund in 2005 to manage most of the oil revenues in a way to maximize transparency and
accountability. The fund is structured through a bank account abroad, which can be accessed
only through parliament approval. To bolster transparency, Timor-Leste was also one of the
first countries to join the Extractive Industries Transparency Initiative (EITI). As argued by the
Independent Evaluation Group (2011), this regime has set new standards for developing coun-
tries in regard to transparency and accountability in the management of petroleum revenues,
and in limiting their arbitrary use.
a. The South Sudanese auditor-general noted that over $1 billion from oil revenues was already unaccounted for before
independence (in 2005–06), and in 2012 Salva Kiir accused senior officials of stealing over $4 billion in state funds (Al Jazeera
2012). At the onset of the civil conflict in December 2013, Riek Machar accused Salva Kiir and his government of embezzling
$4.5 billion (Wudu 2013).
and Bruckner (2011). On the other hand, this finding is consistent with the note
of caution highlighted above, i.e. a developing country’s commodity imports
account for only a limited share of total consumption of commodities, since a
large share of consumption, especially among poorer households, comes from
domestic production (Bazzi and Blattman 2014).
The estimated relationship between conflict and changes in the markets of
major trading partners is negative, consistent with the idea that increases in
demand from trading partners increases real incomes, thus increasing the
opportunity cost of conflict. However, this relationship is also not significantly
different from zero, according to standard statistical tests. This suggests that the
economic cycles in the export destination markets do not affect the probability
of conflict at home, unlike other economic shocks such as rainfall (Miguel,
Satyanath, and Sergenti 2004) and foreign interest rate movements (Hull and
Imai 2013).19
Trade with neighbors is not found to have a significant impact on conflict. One
reason may be that a country may attempt to foster trade with neighboring coun-
tries that are potential sources of instability. Thus estimates of the relationship
between trade with neighbors and conflict may find it difficult to distinguish
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
30 How Trade Can Affect Conflict
between the tendency for trade to reduce conflict and the choice of trading part-
ners that are inherently more likely to foment conflict.
Some evidence for this issue can be seen in the growth of regional trade agree-
ments (RTAs) during the second half of the 20th century, where the desire to
nurture peaceful relations with neighbors was an important motivation (Martin,
Mayer, and Thoenig 2012). While the authors focus on inter-state conflicts, the
same argument may also apply to domestic civil conflicts, as many such conflicts
are fuelled by foreign countries, especially in the same region (Gersovitz and
Kriger 2013). This argument suggests that a country may sign RTAs with the goal
of improving relationships with countries that otherwise could be a source of
instability. In the extreme, only those neighbors of a country X with which X has
signed an RTA may be important in destabilizing X. Some insight into this issue
can be gained by testing the relationship between the likelihood of conflict and
the share of trade with neighbors with which a country has signed an RTA.20 This
variable has a negative and significant association with the onset of conflict.21
However, the estimation of this relationship is difficult, given the likely two-way
causation between conflict and the signing of RTAs (box 1.4).
Mixed results are obtained for the time-varying control variables. Having had a
past conflict (since 1946) raises the probability of conflict by 18 percent, which
confirms the findings in World Bank (2011). On the other hand, neither the
Box 1.4 Correcting for Endogeneity When Measuring the Relationship between
Conflict and Trade under RTAs
We find that trade with neighbors with which a country has signed an RTA is negatively associ-
ated with the onset of conflict. One problem with this approach is that our trade under RTAs
variable may violate an important assumption of this kind of empirical test, namely that the
independent variables (in this case, trade under RTAs) are not caused by the dependent vari-
able (conflict). Indeed, this endogeneity problem likely exists, since part of the driver behind
RTAs could be the desire of intensifying the economic relations with neighbors that may oth-
erwise be able to destabilize the country (similarly to the Martin, Mayer, and Thoenig 2012
story for inter-state wars).
We attempt to correct for this issue by introducing a less endogenous representative of our
trade under RTAs variable, namely trade under RTAs that have been entered into by more than
two countries (so that strategic motives of reducing tensions may be less important). Using
this variable generates similar results as before, since it is highly correlated with the trade
under RTAs variable. The coefficients are only slightly more negative and significant, providing
some evidence that, if anything, the endogeneity biases the absolute size of the coefficients
downwards.
This instrument is not likely to fully address the endogeneity issue, and in the absence of a
suitable instrument, we can only interpret this result as suggestive evidence of the importance
of promoting trade via formal agreements with contiguous countries in order to prevent civil
conflict. Future research would need to test this hypothesis more thoroughly.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 31
presence of a major conflict in the neighboring countries, nor a coup attempt the
year before, significantly affects the conflict probability in the country of
interest.
Qualifications and Alternative Tests
In the above results, a conflict is defined as those with more than 25 battle deaths
per year. Another approach would be to perform the same empirical test but
consider only major conflicts, or those with more than 1,000 battle deaths per year.
However, the results are considerably weaker than when considering all conflicts.
The effect of the export price index on major conflicts is not significantly differ-
ent from zero overall, although its lagged coefficient is negative and significant
(see last four columns of table B.1). The results for the import price index,
changes in the markets of major trading partners, and trade with neighbors are
also not significant.
Taken at their face value, these results suggest that changes in exports affect
the eruption of minor conflicts but not of major civil wars. That would be the
case, for instance, if these shocks influence local conflicts, which do not eventu-
ally spill into large-scale civil wars. That would be consistent with evidence from
Colombia (Dube and Vargas 2013) and the Democratic Republic of Congo
(Maystadt et al. 2013). However, the lack of significance of the results for major
conflicts could also be a product of the “rare event” bias. This refers to the diffi-
culty in identifying a significant relationship when the dependent variable has a
large number of zeroes (King and Zeng 2001). This problem is more likely to
occur when the dependent variable includes only major conflicts, as this is a
much rarer event than minor conflicts. We therefore consider the specifications
that define conflicts as more than 25 battle deaths a year as our preferred ones.
These results are robust to a wide array of checks. Adding country-specific
time trends and adopting a different approach to calculating the price indices
generate very similar results to those described above (see table B.2).
One possible issue is that our results assume that the export price index is not
affected by developments in the country experiencing conflict (otherwise the
coefficient on the export price index will not be estimated correctly). This is
appropriate for most developing countries, which are typically price takers in
international commodity markets. However, some countries do account for a
significant share of the market for their main export product.
We use two strategies to deal with this concern. First, we exclude from coun-
try X’s export basket the commodities for which X’s share in world exports is
above a certain threshold (10 percent in column 1 and 20 percent in column 3
of table B.3), and obtain results that are similar to our baseline results. The main
difference is that the coefficient of the export price index is smaller, but
still significant at the 10 percent threshold. Second, we exclude the countries
that are large exporters of at least one commodity (in one year) according to
the 10 percent or 20 percent criterion. This approach has the advantage of not
generating artificial biases in the countries’ export baskets. However, this
approach also leads to a reduced sample that may be less representative than
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
32 How Trade Can Affect Conflict
the full sample of developing countries as a group. The results are the same as
in our baseline results.
Finally, we test whether our results are different if we use an alternative
source (prepared by the Correlates of War Project—COW) for the conflict
data.22 While we feel that the PRIO data is more reliable, it is reassuring that the
main results do not vary much across datasets. The comparison—reported in
columns 5–6 of table B.3—shows that the individual trade variables’ coefficients
are not significantly different across the datasets. The only exception is that con-
temporaneous coefficient on the import price index is positive and significant in
COW and negative but not significant in PRIO. However, the sum of the cotem-
poraneous and lagged import price index variables is not significant with either
dataset, a result in line with that obtained earlier using the same high threshold
for defining civil conflicts.
Differentiating between Commodities
As discussed above, the impact of commodity prices on conflict differ along two
dimensions: 1) whether they are point-source or diffused commodities; and
2) whether they are consumed domestically or not. Testing the first distinction
reveals that the positive impact of the export price index on conflict (shown
above) is mainly due to point-source commodities. By contrast, the effect of dif-
fused commodities on the probability of conflict is not significant (column 1 of
table B.4 in appendix B).23 These results are consistent with the rapacity effect,
while they provide no support to the opportunity cost effect.
For the second distinction, only commodities that are consumed domesti-
cally have a positive and significant impact on conflict onset (although the
coefficient on the price index of commodities that are not consumed domesti-
cally also has a positive sign). This result appears consistent with the expecta-
tion that increases in the prices of commodities that are consumed domestically
reduce the opportunity cost of engaging in conflict. However, further analysis
indicates that the result may be driven by the domestically consumed, point-
source commodities (e.g. oil and gas), and not domestically consumed, diffuse
commodities (e.g. food). Thus the estimated positive relationship between
conflict and domestically consumed commodities may reflect the rapacity
effect, or the competition for point-source commodities, rather than the
impact of rising prices on the real incomes of consumers.
Further evidence, albeit only partial, can be seen by splitting the export price
index into the four types of commodities that combine the two dimensions as
in table 1.2 (see column 3 of table B.4). While the coefficients of the sub-
indices are not significant, the magnitudes suggest that point-source, consumed
commodities exert the largest impact on conflict of all the subcategories. This
group comprises oil and gas, which represent important consumption items in
many developing countries, especially in the urban areas. In the absence of con-
sumption data by country, it is not possible to disentangle the rapacity effect
from the consumption effect in this case. Noting that our variable is constructed
on the basis of the export shares, we interpret this mainly as a rapacity effect.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 33
The coefficients of the diffuse export commodities are also not significantly
different from zero for both the consumed (which has a negative sign) and the
non-consumed (which has a positive sign).
This result confirms the lack of support for the opportunity cost effect, at least
via the export sector. Even the commodity group that should provide the clean-
est test for the opportunity cost effect—i.e. diffused, not consumed commodities—
does not have a significant relationship with conflict. On the other hand, the
weakly positive sign suggests that some form of resource effect may be at work
even with diffused commodities, for example as the revenues from these com-
modities may also be taxed by rebel groups to fund their struggle.
Impact on Conflict Ending or Intensity
So far the dependent variable has been the onset of conflict. It is also useful to
test the impact of the same trade variables on the probability of a conflict ending
and the intensity of conflict (equation 6 in appendix A). The results for conflict
ending (given in columns 1–4 in table B.5) are difficult to interpret. The relation-
ship between the export price index and conflict ending is not significant (while
it was significant when testing its impact on conflict onset), while the relation-
ship between trade with neighbors and conflict ending is positive and significant
(while it was not significant with conflict onset). In line with the previous results,
the trade variables do not have a significant impact on the probability of a con-
flict ending when considering only major conflicts (columns 5–8 in table B.5).24
These results suggest that the trade variables we have examined do not seem to
matter much in affecting the duration of an ongoing conflict.
More interesting results are obtained in testing the impact of the trade vari-
ables on conflict intensity (the dependent variable is the number of battle
deaths). Both the export price index and import price index have a positive and
significant impact on conflict intensity, which is consistent with the rapacity and
opportunity cost effects, respectively (column 1 of table B.6). However, for the
export price index the conflict-inducing effect is actually driven by diffused
rather than by point-source commodities (column 2).25 This is the opposite
result to that on conflict onset and may suggest that ongoing conflicts are intensi-
fied by increases in the price of diffused exported commodities. While this find-
ing is different from that in Bazzi and Blattman (2014), who find a weak effect
of export prices on conflict intensity, it is consistent with the idea that produc-
tion of diffused commodities may provide an important source of revenues for
rebel groups to fund their struggle, thus intensifying existing conflicts.
The extent to which rebel groups may support their activities through taxing
diffused commodities varies across contexts as well as across commodities. For
example, a rise in coca production fostered violence in Colombia by raising the
guerrilla’s revenues (Angrist and Kugler 2008). Similarly, the cross-country evi-
dence provided by Nunn and Qian (2014) is consistent with small armed groups
using U.S. food aid to fund local conflicts. On the other hand, Dube and Vargas
(2013) find that increases in the value of production of diffused commodities
reduce conflict intensity in Colombia. It is beyond the scope of this work to
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
34 How Trade Can Affect Conflict
identify the conditions under which one or the other channel may prevail, but it
is important to acknowledge that the impact of diffused commodities on conflict
may be more complex than what is suggested by the simple opportunity cost
theory.
Higher income growth in a country’s export markets reduces the intensity of
conflict. This raises two possible differences with the earlier results. First, this
variable had no significant impact on the onset of conflict, suggesting that such
shocks affect the intensity, but not the onset, of civil conflicts (as in Chaudion,
Peskowitz, and Stanton 2012). This confirms the finding in Bazzi and Blattman
(2014) that escalating an existing conflict seems easier than triggering a new one.
Second, the negative relationship between growth in a country’s export mar-
kets and conflict is consistent with the opportunity cost theory. However, our
finding that diffused commodities have no significant relationship with conflict
onset (see above) appears to contradict the opportunity cost theory. Examining
the source of this difference is beyond the scope of this analysis, but we can put
forward a possible hypothesis. The relationship between conflict and the prices
of diffused commodities is not significant because increases in these prices both
increase the opportunity cost of conflict (thus tending to reduce conflict) and in
some cases provided funding for rebel groups (thus increasing conflict). The latter
effect, however, applies almost exclusively to goods produced in rural areas, as
rebel groups’ ability to tax local economic activity is mostly limited to rural areas.
By contrast, increases in a country’s export markets likely have little effect on
commodity prices, which are set in international markets, but will affect demand
for manufactures, whose trade is more based on importer-exporter networks and
thus more affected by country-specific demand shocks.26 Moreover, we control
for commodity price shocks, reinforcing the idea that the effect of changes in
export markets operates through demand for manufactures. Since manufactures
production in developing countries is located mainly in urban and peri-urban
areas, changes in a country’s export markets are unlikely to provide rebels with
additional revenues. Therefore this “rebel funding” channel could explain the rela-
tion between diffused commodity export price shocks and conflict intensity, but
is less likely to apply in the case of demand shocks in export markets.
Finally, a higher share of trade with neighboring countries is significantly
associated with lower conflict intensity. This finding is consistent with the idea
that country X’s trade with its neighbors increases their opportunity cost of de-
stabilizing X, for example by supporting rebel groups in X.27 While the coeffi-
cient for the impact of regional trade on conflict onset was also negative, it was
not statistically significant.
Evidence from Nigerian States (2004–13)
Case studies that examine the impact of changes in trade on conflict across
regions within a single country are a necessary complement to cross-country
analysis. One sacrifices the opportunity to reach conclusions on conflict from a
global perspective, and to gain insight on how trade affects conflicts in many
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 35
different contexts. On the other hand, differences among regions, for example in
political conditions and the business environment, tend to be smaller than differ-
ences among countries. Thus within-country analysis can more easily control for
different conditions that may affect the relationship between trade flows and
conflict. In addition, data for a single country are usually richer than across coun-
tries (e.g. on consumption and on employment), thus enabling more precise tests
of effects than in cross-country analyses. This section considers the example of
how changes in trade have affected the conflict in Nigeria, while the next section
examines the Israeli-Palestinian conflict.
Nigeria’s Civil Conflict in the Past Decade
Although it is not officially considered fragile according to the World Bank and
the regional development banks, Nigeria has had a recent history of acute
conflict-related violence. According to the Armed Conflict Location and Events
Dataset (ACLED), from 2003 to 2013 Nigeria was the third most violent
African country and suffered the fourth-highest deaths from conflict. While the
country has not experienced a full-blown civil war and the state’s monopoly
of force does not appear to be challenged, local conflicts have been a major
constraint on the country’s development over the past few decades.
The form and intensity of violence has varied substantially, both across space
and over time. Conflict in Nigeria is highly regionalized. Both the dominant type
of violence (battles, protests, riots, and violence against civilians) and the underly-
ing determinants of conflict differ across regions. In the past decade, four
“geographies of conflict” can be identified: the North, the Niger Delta, the
Middle Belt, and the urban areas. These conflicts have important common under-
lying traits, rooted in dysfunctional public institutions and social and economic
marginalization (Joab-Peterside et al. 2012). However, the regional contexts have
also played a fundamental role in shaping the particular forms and dynamics of
violent conflict in each case.
In the past decade, violence in the so-called middle belt, and particularly in
Plateau State, has been mainly in the form of communal violence. While much
of the recent violence has occurred between Muslim and Christian communities
(though some violence also has occurred within Muslim communities), unequal
access to land appears to be a core driver of the conflict in the middle belt.28
In Kwara State, for instance, the conflicts in Offa/Erin Ile can be attributed to
disputes over land ownership and grazing rights.29 In other states, minor disputes
have escalated owing to improper handling. One example is the conflict in Ekiti
State over the permanent site of a social amenity within the neighboring towns
of Ise and Emure Ekiti.
Violence has increased since 2010 (map 1.1), particularly in the northeastern
parts of the country, in large part due to the activities of the Islamic militant
group Boko Haram. Indeed, the government declared a state of emergency in the
three most northeastern states of Borno, Yobe, and Adamawa in May 2013. These
areas also experienced some of the greatest intensification in conflict in the coun-
try, in terms of both the number of conflict events (map 1.2) and the number
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
36 How Trade Can Affect Conflict
Map 1.1 The Geography of Conflict in Nigeria (2004–13)
Source: ACLED.
Note: Conflict events are all events recorded by ACLED that involve any form of political violence (i.e., battles, protests/riots,
and violence against civilians).
Map 1.2 Conflict Intensity across States in Nigeria
Source: ACLED.
Note: Conflict events are all events recorded by ACLED that involve any form of political violence (i.e., battles, protests/riots, and violence against
civilians). The darker the color the higher the number of (any) conflict events in the period.
of fatalities (map 1.3). However, other parts of the country, particularly the
middle belt states of Platteau, Kanu, and Kaduna, have also recently experienced
an intensification of long-standing conflicts.
In addition, political demonstrations (particularly concerning fuel subsidies
and corruption) have increased in recent years, mainly in urban areas, and have
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 37
Map 1.3 Violence Intensity across States in Nigeria
Source: ACLED.
Note: The darker the color the higher the number of fatalities in the period.
expressed themselves in violence. In Abuja and Lagos, over 40 percent of conflict
activity is made up of rioting or protesting. Over the course of the period covered
by the dataset (1997–2013), over one-third of riot and protest events have
involved violence (ACLED 2013).
At the same time, conflict in other areas of the country has subsided. In par-
ticular, violence by the rebel groups in the Niger Delta states, which was among
the most violent parts of the country in the 2000s, was significantly reduced after
the agreement of 2009, whereby the state provided amnesty for local militants
along with a disarmament, demobilization, rehabilitation and reintegration
(DDRR) program. Under the amnesty, which ran from August to October 2009,
militants who handed in their weapons were pardoned for their crimes, trained
in nonviolence, and offered vocational training in various activities in Nigeria or
overseas. After attending nonviolence training they were paid US$410 per month
until they found work. Just over 26,000 young militants took the amnesty pack-
age (IRIN 2011b). While this agreement has been criticized for failing to treat
the root causes of conflict, and for promoting “warlordism,” it seems to have gone
a long way toward reducing conflict in the short run (Sayne 2013).
Model and Data
Our basic model measures the impact on conflict across states over the
period 2004–11 (as measured by the number of conflict episodes, the number
of violent episodes, and the number of conflict-related fatalities) of price
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
38 How Trade Can Affect Conflict
indices for goods produced by households, goods consumed by households,
and oil. We control for various other determinants of conflict at the state
level, including the size and density of the population at the beginning of the
period, past incidence of various types of conflict, the poverty gap and share
of the population under the poverty line at the beginning of the period, and
three indicators of the potential for ethnic tensions (whether the president’s
ethnicity is the same as that of the dominant group in a state, whether there
are at least two significant ethnic minority groups, and whether there is
more than one dominant ethnic group). In addition we control for all possible
time varying covariates of conflict at the regional level (for each of the six
macro-regions in Nigeria).
The data on conflict are from the Armed Conflict Location and Event Data
Project (ACLED), which records individual conflict events from media sources,
allowing for the construction of statewide measures of conflict intensity over
time. The price indices are based on urban prices reported by Nigeria’s National
Bureau of Statistics (NBS), which collects monthly data for 143 food and non-
food items by state. In the consumer price index, the state-level NBS prices are
weighted by the share of each good in household consumption. Similarly, in the
producer price index, the NBS prices are weighted by the share of each good in
household production. Data on production and consumption behavior are taken
from household surveys. The oil price index is calculated by multiplying the
value of oil production in 2003 by the international oil price. A more formal
explanation of how these indices are calculated, and an explanation of the several
choices required to deal with data issues (e.g. the use of urban and not rural data
on prices, how the NBS price data are matched to survey data on goods produced
and consumed by households, and the calculation of oil production at the state
level) is provided in appendix A.
While our goal is to estimate the impact on conflict of our two price indices
(of household production and consumption goods), it is equally true that conflict
will affect these price indices. For instance, high levels of conflict may reduce local
production, and if markets are imperfectly integrated across space, this may boost
local prices. Conflict may also reduce local demand, which would have an oppo-
site effect on prices. This endogeneity would bias the relationship between the
price indices and conflict. Typically, this endogeneity problem is corrected for by
using an independent variable (called an instrument) that is similar to the inde-
pendent variable of interest, but not affected by the dependent variable. Many
studies (e.g. Bazzi and Blattman 2014; Dube and Vargas 2013) use international
prices as instruments for domestic prices. However, this does not work very well
in our study, because international prices are available only for internationally
traded commodities, which often do not include many local products important
for consumption and production in Nigeria (e.g. yam and cassava). Moreover,
international prices do not account for the price transmission from international
to domestic markets, which is often limited. Thus international prices may not
provide an ideal representation of the size of the change in price at the local level,
and thus have only a weak relationship with conflict. However, the international
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 39
price indices are useful for checking the robustness of the results. In addition
(unlike the domestic price indices) they are available through 2013.
Our main specification uses instruments that are constructed on the basis of
domestic prices of faraway states, following the same logic of Jacoby (2013) for
changes in rice prices in Indian districts. The price data for the other Nigerian
states should reflect exogenous international price changes, their transmission to
the domestic market, and shifts in demand and supply in the large domestic
markets outside of the particular state in question. We exclude neighboring states
that may be affected by the conditions in the state in question, to ensure that
they are not affected by conflict within that state. A formal presentation of
how these various price indices are constructed and of the method of derivation
of exogenous prices indices is given in appendix A.
Results
In our preferred specification, the relationship between the producer price index
(calculated with prices of faraway states) and conflict is negative and statistically
significant.30 A 10 percent increase in the producer price index will lead to a
reduction of 2.3 conflict events (column 3 of table B.10), while holding all other
variables in the model constant. The relationship between the consumer price
index (again calculated with the prices of faraway states) and conflict is positive,
statistically significant, and slightly larger than that of the producer price index.
These results provide substantial evidence that changes in real incomes affect
conflict by changing the opportunity cost of participating in violence: as real
incomes rise (both through increases in producer prices and declines in consumer
prices), the incidence of conflict falls, and vice versa.
This result is consistent with that from an alternative specification which
excludes the consumer price index. In this case, the producer price index has
no statistically significant impact on conflict. This suggests that failure to
include this consumption impact severely biases (toward zero) the conflict-
reducing effect of increases in prices of agricultural commodities produced by
the households. The issue here is that the prices used in construction of the
producer price index involve goods that are both consumed in, and produced
by, households. Thus increases in the producer price index may raise real
incomes (and thus the opportunity cost of fighting) for households that are
predominantly affected through the goods they produce, and lower real
incomes for households that are predominantly affected through the goods
they consume. Once we control for this (positive) consumption effect
through the consumer price index, we are able to isolate the true (negative)
production effect of producer prices on conflict. This is an important finding,
as the literature has tended to focus solely on the impact of prices of pro-
duced goods on conflict, thus potentially suffering from an important omitted
variable bias. This may also help explain the lack of consensus on the effects
of agricultural commodity prices on conflict.
The oil index has a positive and significant effect on the number of conflict
events the following year, in line with the state prize hypothesis: exogenous
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
40 How Trade Can Affect Conflict
increases in the value of oil raise the incentive for fighting in the production
areas. However, this effect disappears after the amnesty agreement in 2009,
confirming that the agreement was effective in curbing violence in the Niger
Delta states. In fact, the insignificant association between the oil price index
and conflict intensity after 2009 is consistent with the idea that the oil funds
may have helped to demobilize militant groups in these areas (Sayne 2013).
In the period up to the agreement, a 10 percent increase in the oil index only
increases conflict events by 0.14. This smaller average effect (relative to the
other commodities’ effect) is partly due to the fact that it only applies to a
few states and partly to the fact that the oil price index varies more than the
producer and consumer price indices.
While we do not report the coefficients of the control variables (full results are
available upon request), an interesting result is that the ethnicity of the president
matters in determining the level of conflict in each state. When the president’s
ethnicity is the same as that of the dominant group in a state, conflict intensity
subsides, confirming the importance of ethnic allegiance in state politics. This
result is weaker for those states with more than one dominant ethnic group.
These results are robust to various tests. If we use contemporaneous (rather
than lagged) values of the price indices, the consumption and oil price indices are
still significant, but the production price index and the oil price index (after
2009) are not. The weak result for the contemporaneous producer price index
implies that the impact of price changes on conflict occurs with a lag (even the
significant result for the contemporaneous consumption price index is weaker
than with its lagged value).31 Similar results are found if we include both the
contemporaneous and lagged values of the price indices.32 The results are also
robust to the inclusion of the lagged unemployment rate as a further control,
which, however, makes the producer price coefficient less significant (column 3
of table B.11).
We also test for using international prices to calculate the producer and con-
sumer price indices, rather than the domestic prices described above. Once again,
the relationship between the consumer price index and conflict is weaker than
in the preferred specification, while the relationship between the producer price
and conflict is no longer significant. Comparable results are obtained when using
an alternative econometric technique (the poisson estimator, see appendix B).
The results for the use of contemporaneous variables and international price
indices confirm that consumption, production, and oil prices have a significant
impact on conflict events, although production effects are somewhat less robust
than the others.
Comparable results also are obtained using alternative measures of conflict
(events that result in fatalities, battle events, protests and riots, and violence
against civilians) as the dependent variable (columns 1–12 in table B.12). The
consumption and oil price effect, but not the production price effect, remain
significant when using the number of fatalities as the dependent variable (col-
umn 13 of table B.12).33 This analysis also shows an important dichotomy in the
effect of the price changes. While the effect of consumer and producer prices is
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 41
particularly large for protests/riots, changes in oil prices have no effect on this
type of conflict event (and that is the only type of conflict event that oil does not
affect). This result is consistent with the view that the violence surrounding oil
extraction was mainly organized around militant groups, and was unrelated to
popular protests.
The Boko Haram Conflict
As noted above, the most devastating Nigerian conflict in recent years has been
associated with Boko Haram. According to the International Crisis Group (2014)
Boko Haram (usually translated loosely as “Western education is forbidden”)
emerged in the early 2000s as an Islamic movement in northern Nigeria led by
the charismatic cleric Mohammed Yusuf. Its aim is to establish an Islamic state in
the north with strict adherence to Sharia law as it “believes that corrupt, false
Muslims control northern Nigeria” (p. i).
The clashes between the group—which wanted to change the political and
religious order of the region—and the police started in 2009 and quickly esca-
lated into an armed insurrection, which was crushed by the state forces.
Hundreds of Boko Haram’s members were killed and the group’s principal
mosque was destroyed. Yusuf was captured by the army, handed over to the
police, and shortly thereafter extra-judicially executed in public (International
Crisis Group 2014; Nossiter and Kirkpatrick 2014). This spurred the retaliation
of Boko Haram, which went underground and a year later launched attacks on
police stations and military barracks, explicitly in revenge for the killings of Yusuf
and his comrades (International Crisis Group 2014).
However, the attacks continued to escalate, including attacks against civilians.
As a response the Nigerian government assembled a joint task force of military
and police units to battle Boko Haram and declared a “state of emergency” in
three northeastern states—Borno, Yobe, and Adamawa—in May 2013. However,
that has not seemed to reduce the violence, which included the murder of sixty-
five students at the agricultural college in Yobe State in September 2013, chain-
saw beheadings of truck drivers, and the killing of hundreds on the roads of
northern Nigeria (Council on Foreign Relations 2014). Most recently the abduc-
tion of 200 schoolgirls in Borno State spurred a worldwide wave of condemna-
tion of the group.
While the conflict has a clear religious dimension, analysts suggest that griev-
ances, including those motivated by poverty and unemployment, are fundamen-
tal drivers of the increased militant activities of Boko Haram (International Crisis
Group 2014; IRIN 2011a).34 This type of economic grievances may have facili-
tated the recruitment of Boko Haram, which incentivized poor youth to join
their ranks by offering food, shelter, and other forms of assistance not provided
by the government (Copeland 2013).
We use the same approach as above to determine to what extent trade-related
changes in income explain the surge in the Boko Haram conflict in the past few
years. As we don’t have state-level domestic prices after 2010, the changes in
consumption and production measures are based on international prices.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
42 How Trade Can Affect Conflict
The results of the analysis suggest that income shocks via both consumption
and production price changes exert an important influence on the intensity of
the Boko Haram conflict. The direction of the effects is consistent with that of
the previous analysis, although this time the impact of production shock on
conflict events is larger than consumption. On average a 10 percent reduction
in the value of the production index in a year leads to 2.38 additional conflict
events involving Boko Haram in the following year, while the same reduction in
consumption prices leads to 0.95 fewer events (column 1 in table B.14). The
effects are even larger (3.7 and 2.5 respectively for production and consumption
prices) when considering only the three northern regions where Boko Haram
activities are concentrated: production (column 2 in table B.14). These results
are even more striking, considering that the data on international prices may not
reflect accurately local market conditions as argued above. On the other hand,
the oil index does not bear any significant relation with Boko Haram conflict
activities.
Interestingly, neither consumption nor production price indices appear to
be associated with non Boko Haram conflict events either in the full sample or
in the northern regions (columns 3 and 4 in table B.14). This suggests that
income shocks, at least driven by consumption and production price changes,
appear to matter for current conflicts in Nigeria mainly through the Boko
Haram insurgency. On the other hand, the oil index is associated with a reduc-
tion in non Boko Haram conflict, confirming our previous results for the post-
2009 period.
The conflict inducing effects of increases in the consumption basket price
carry through also for the number of violent conflict events and of fatalities while
that is not the case for the prices of the production basket, which becomes insig-
nificant. The effect of consumption prices on fatalities is substantial: a 10 percent
increase in the price of the consumption basket in a state generates an additional
19 fatalities in that state.
These results confirm the view that the Boko Haram conflict has an impor-
tant economic dimension, which affects the intensity of the militants’ activities.
They are consistent with the view that changes in incomes affect the propensity
of the local populations to support the Boko Haram insurgency and/or the ability
of Boko Haram to recruit fighters. This opportunity cost story seems to be even
more relevant than for the other current conflicts in Nigeria.
Evidence from the Israeli-Palestinian Conflict (2000–04)
A case study of the decades-long Israeli-Palestinian conflict has an important
advantage from the perspective of our model of how changes in trade affect
conflict. Unlike the cross-country analysis or the Nigerian study, Palestinian
exports do not include point-source commodities such as oil or minerals that
might be the target of appropriation through violence. Moreover, in the cross-
country analysis the possibility that rebels might use diffuse commodities to fund
their activities complicated the interpretation of the estimated relationship
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 43
between changes in the prices of these commodities and conflict. By contrast, the
Palestinian Authority has only a limited ability to tax private earnings, thus mini-
mizing the potential to use increased exports to curb (or to increase) violence
against Israel. Eliminating the rapacity effect and the resources effect as possible
interpretations of our results allows us to focus on testing the opportunity cost
hypothesis. So far, tests of the opportunity cost hypothesis in studies of the
Israeli-Palestinian conflict have shown mixed results (box 1.5).
Palestinian Trade Prior to the Second Intifada
Understanding the impact of changes in trade flows on the Israeli-Palestinian
conflict requires understanding the degree to which Palestinian tradable produc-
tion is dependent on Israel. Almost 90 percent of Palestinian merchandise
exports is destined for Israel.35 Further, Palestinian external trade is de facto regu-
lated by Israel. Following the Oslo accords in 1993, the West Bank and Gaza and
Israel have become part of a de facto custom union with a common external
tariff decided by Israel, which during the 1990s controlled all the borders of the
custom union, with no tariffs or quotas imposed between Israel and the West
Bank and Gaza.36 While the latter can de jure have its own trade policy, e.g. it
can sign trade agreements with third parties, in reality any imports destined for
the West Bank or Gaza have to enter the union via an international border con-
trolled by Israel, which automatically charges the Israeli import tariff for goods
from the specific country of origin.
Box 1.5 The Literature on the Israeli-Palestinian Conflict and the Opportunity Cost
of Violence
Studies of the Israeli-Palestinian conflict have provided evidence both for and against the
opportunity cost mechanism. Berrebi (2007) shows that an individual with a higher education
and standard of living is more likely to become a suicide bomber (which is not consistent with
the opportunity cost hypothesis), while Sayre (2009) and Saleh (2009) using district-level data
find the opposite relationship. In addition, Cali, Miaari, and Fallah (2014) find no support for the
opportunity cost mechanism by relating districts’ public sector employment and Palestinian
fatalities in the West Bank and Gaza during and after the Second Intifada. One interpretation of
public
these latter findings is that engaging in political violence has little opportunity cost for
sector employees, as they do not face a high cost from shirking.
In a related study, Miaari, Zussman, and Zussman (2014) find that localities which were rela-
tively more dependent on employment in Israel experienced relatively more fatalities after
Israel’s abrupt imposition of severe restrictions on the employment of Palestinians within its
borders at the outbreak of the Second Intifada. This assumes that the large variation in the pre-
Intifada employment rates in Israel across West Bank localities was unrelated to prior levels of
involvement in the conflict. These results may provide some prima facie support for the oppor-
tunity cost mechanism for private employees. Our analysis will take this channel into account
by including the localities’ share of employment in Israel.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
44 How Trade Can Affect Conflict
Moreover, Israeli control of the international borders implies that Palestinian
exporters and importers have a strong incentive to use Israeli intermediaries to
clear their goods. Israeli intermediaries reduce the cost and time of trading rela-
tive to those faced by Palestinian traders. Palestinian imports and exports are
subject to twice the costs of Israeli imports and exports using the same port facili-
ties in Israel (World Bank 2010). Importing procedures take on average as much
as four times longer for Palestinians than for Israelis (40 days vs. 10 days). It is
estimated that 58 percent of the Palestinian imports from Israel in 2008 were
through trading companies, most of which was for reexport (Bank of Israel 2010).
The opening up of the Israeli import regime in the 1990s eroded the
preferential access of Palestinian goods in their dominant export market. As
a result, imports from the rest of the world have progressively replaced
those from the West Bank and Gaza, especially in the main labor-intensive sec-
tors. Partly as a consequence of this shift, manufacturing production in the West
Bank and Gaza declined in real terms by almost 20 percent between 1994 and
2009.37 Palestinian merchandise exports slowed in nominal terms prior to the
Second Intifada (figure 1.7), and declined as a share of GDP from over 10 per-
cent in 1996 to less than 9 percent in 1999.38 Palestinian exports also declined
slightly in constant prices during this period. Palestinian exports generally per-
formed worse than Israeli imports.
The limited changes in Palestinian trade as a result of the liberalization of the
Israeli import regime mask a large variation across sectors (figure 1.8).39 For
example, exports of cucumbers, and of marble and alabaster, rose by more than
$6 million each, while exports of building stone dropped by $8 million. All in all,
the shape of the distribution of changes in figure 1.8 suggests that more sectors
Figure 1.7 Palestinian Exports to the World and to Israel, 1996–2000
Thousand US dollars
450,000
400,000
350,000
300,000
250,000
200,000
1996 1997 1998 1999 2000
Palestinian exports to the world Palestinian exports to Israel
Source: Elaboration on data from the Palestinian Central Bureau of Statistics.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 45
Figure 1.8 Distribution of Changes in Palestinian Exports (1996–99)
Thousand US dollars
0.10
0.08
k density di _exp
0.06
0.04
0.02
0
–10 –5 0 5 10
Change in Palestinian exports, 1996–99
Source: Palestinian Central Bureau of Statistics.
Note: Sectors are recorded at the 5-digit SITC level.
had negative than positive changes, and the large variation in changes likely
resulted in very different effects across local areas, depending on where these
goods are produced and their degree of labor intensity. This large variation across
sectors, and thus across local areas, provides a useful dataset for testing the
impact of changes in trade on the Second Intifada.
As in previous studies (Calì and Miaari 2013; Miaari, Zussman, and Zussman
2014), the number of Palestinian fatalities killed by Israeli forces is the main
measure of conflict intensity across the West Bank and Gaza. This is a suitable
measure as most of these fatalities were the result of political demonstrations
suppressed by the Israeli army or direct confrontation between the Israeli army
and Palestinian armed factions. The evolution of Palestinian fatalities, depicted in
figure 1.9, shows that violence in the West Bank peaked in 2002 and declined
through 2003 and 2004, when the Second Intifada finally drew to a close. This
period was followed by periods of relatively low-intensity conflict. In Gaza, after
the drop in 2003, violence picked up again in 2004 and lasted until the first half
of 2005. During the respective sample periods, 1,278 Palestinian fatalities were
recorded in the West Bank. In Gaza, 1,702 Palestinian fatalities were recorded.
Empirical Results—Exports
We find that changes in Palestinian exports had a significant impact on conflict
intensity during the Second Intifada. We model conflict intensity (the number of
conflict-related fatalities in a locality) as a function of changes in overall sectoral
exports weighted by the share of that sector in each locality’s private employ-
ment, along with indicators of sociodemographic conditions, the quality of
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
46 How Trade Can Affect Conflict
Figure 1.9 Palestinians Killed by Israel in the West Bank and Gaza, 2000–04
Number of deaths
700
600
500
400
300
200
100
0
2000 2001 2002 2003 2004
West Bank Gaza
Source: B’tselem.
nfrastructure, and the region (West Bank, Jerusalem, or Gaza).40 An increase
i
of $10 million in the export revenue variable reduces conflict-related fatalities
in that locality by 2.1 percent, significant at the 1 percent level (column 1 of
table B.15).41 Given the weighting scheme for the export variable, this effect
reflects the conflict-reducing impact of exports via employment. Adding eco-
nomic controls, including permits to work in Israel and the unemployment rate,
as well as the Palestinian fatalities prior to the Second Intifada (1995–2000),
slightly raises the absolute size of the export coefficient, which remains signifi-
cant at 1 percent (column 2 of table B.15).
These results support the opportunity cost hypothesis: better employment
opportunities should raise the opportunity cost of involvement in the conflict,
thus lowering its intensity. This result is also in line with experts’ opinion on the
determinants of violence in the West Bank in the current phase (Xinhua 2014).42
We next add other labor market indicators from the Palestinian Labor Force
Survey (PLFS), including the share of private and public employment in the
locality, the share of the locality’s workers who are employed in Israel, and
the average wage in 1999 (column 3 of table B.15). Adding these potentially
important controls, which capture the pre-Intifada reliance of the localities on
different sources of employment, increases the estimated impact of changes in
export revenues on the conflict. Now a rise in $10 million in Palestinian
exports of a sector covering 10 percent of the locality’s private employment
reduces conflict-related fatalities in that locality by 2.8 percent, significant at
the 1 percent level. This larger absolute value of the coefficient is entirely due
to the reduction in sample size (to 199 localities) caused by the inclusion of
the additional labor market controls.43 This smaller sample size provides a
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 47
robustness check in its own right, as it restricts the analysis to comparatively
larger localities.
Thus far, we have constrained the effect of exports on conflict to be similar
across localities in the West Bank and the Gaza Strip. Testing our preferred speci-
fication for the West Bank and the Gaza Strip separately confirms the finding
that increases in Palestinian exports are negatively associated with the number
of fatalities during the Second Intifada (columns 5 and 6 of table B.15).44 The
results for Gaza should be interpreted with caution since they only rely on a
population of 37 localities.
We also estimate the impact on conflict of other control variables used in
the export equation. Consistent with the opportunity cost mechanism, a
higher level of education and higher private sector wages are associated with
lower conflict intensity. Some other control variables have a relatively
straightforward interpretation. Higher shares of large households and of
married individuals are associated with lower conflict intensity. More permits
to work in Israel in 1999 (and higher wages to Palestinian employees in Israel)
are associated with a higher number of fatalities during the Second Intifada,
probably because Israel imposed employment restrictions at the outset of the
Intifada (Miaari, Zussman, and Zussman 2014). More populated localities
have a higher number of fatalities, although this association is not robust across
specifications.
Empirical Results—Other Trade Variables
Changes in imports are found to have no significant impact on conflict intensity
in the next period (column 1 of table B.15). This suggests that the eventual dis-
placement effect of increased imports on domestic producers is not substantial.
Alternatively, higher imports may reflect higher demand, and thus better eco-
nomic conditions. This addition does not affect the coefficient on change in
Palestinian exports, which remains significant and of similar magnitude as before.
One possible issue with our results is that Israel might have changed its trade
policies in anticipation of the Second Intifada. If so, then our estimate of the
impact of exports on conflict would be biased. However, this is unlikely. To show
this, we construct a measure of how Israeli imports from the rest of the world
may have affected individual Palestinian localities. Similar to our export index,
this measure is, for each locality, the weighted average of changes in sectoral
imports by Israel from the rest of the world, where the weights are the share of
local private employment in that sector. This measurement of the impact of
Israeli imports has no significant impact on conflict.45 This result is consistent
with the view that Israel did not change its trade policies in connection to the
expected surge in violence during the 1990s. It also suggests that once we control
for changes in Palestinian exports, any residual effect of Israeli imports on
Palestinian employment is marginal.
A further check on our results involves splitting the export variable into
changes in exports to Israel and changes in exports to the rest of the world, again
weighted by each locality’s employment by sector. While both of these variables
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
48 How Trade Can Affect Conflict
have a negative relationship with conflict, only the variable reflecting exports to
Israel is statistically significant.46 It thus appears that exports to Israel were the
main channel through which exports affected violence during the Second
Intifada, which is not surprising as Israel accounted for over 90 percent of
Palestinian exports in the 1990s. However, the large absolute magnitude of the
coefficient on exports to the rest of the world suggests potentially relevant effect
on conflict from those exports as well.
Empirical Results—Alternative Measures of Conflict
So far we have used the total number of Palestinians killed by Israeli forces as the
measure of conflict intensity. However, some of these fatalities were not related
to Palestinian participation in politically motivated violence against Israeli forces.
The B’tselem dataset does identify various instances of Palestinians who were
not taking part in the hostilities, but were killed by Israeli forces. Those fatalities
should not be considered when testing for the opportunity cost motive to engage
in violence. To address this issue we use as the dependent variable only the num-
ber of Palestinian fatalities as a result of participation in violence against Israeli
forces or political demonstrations. We do not use this dependent variable for our
main specification, because information on participation in violence is not pro-
vided for every recorded fatality. The estimate for the impact of exports on
conflict remains significant with this alternative dependent variable. In fact, the
estimated impact is larger than in our main specification, and (contrary to the
results given above) with this dependent variable the coefficient on exports to
the rest of the world becomes significant at the 10 percent level (see column 3
of table B.17). The import variables continue to be irrelevant in explaining con-
flict intensity.
Changes in Palestinian exports also affect the probability of conflict. Here
we define a dependent variable equal to 1 if a locality experiences any fatalities
throughout the Second Intifada and zero otherwise. The coefficient on the
export variable indicates that an increase in Palestinian exports by $10 million in
a sector employing 10 percent of private employees in a locality reduces the
probability of conflict in that locality by between 5.3 percent and 5.5 percent
(columns 1–2 in table B.18).47 This effect increases to between 6.3 percent and
6.9 percent in the case of exports to Israel, while it is not significant for the
exports to the rest of the world (columns 3–4). Again, the effect of changes in
imports is not significant (columns 2 and 4).
Empirical Results—The Issue of Endogeneity
One assumption of the econometric techniques used in these estimations is that
our trade measures are exogenous, that is, they are not affected by the dependent
variable (conflict intensity or probability of conflict), nor are they related to
other unobserved factors affecting also local conflict. This assumption appears to
be plausible for a number of reasons. First, as the distribution of employment
across sectors in each locality is measured at the beginning of the period, it
should not be affected by the eruption of the local-level conflict after three years.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 49
Second, controlling for a large number of local-level factors should help address
the concern that employment shares may reflect local characteristics such as skill
intensity and labor productivity that may also drive conflict locally. Third, given
the large number of localities, each of them should not exert an important influ-
ence over the export variable, which is the aggregation over all Palestinian locali-
ties. Finally, these changes in trade, and exports in particular, are mainly driven by
two factors exogenous to the Palestinian economy. The first is the emergence of
new global suppliers, chiefly China, competing in similar sectors (and markets)
where Palestinian exports are concentrated. The second is the reduction in Israeli
import tariffs during the 1990s, which eroded the preferential access of
Palestinian exporters to their most important external market.
It is nevertheless useful to check whether our trade values are indeed exog-
enous. As with the Nigeria case study, we select two alternative variables
(called instruments) that are related to our export variable but are demonstra-
bly not influenced by conditions in Palestinian localities. The first instrument
is the change in Chinese export supply over the same period as our export
variable (1996–99). This is arguably an important source of competition for
Palestinian exports, especially in Israel. Indeed, unlike imports from the West
Bank and Gaza, Israeli imports from China dramatically increased between
1995 and 2000 (figure 1.10). In order to ensure that our measure of Chinese
exports is not affected by conditions in Israel (which might possibly be related
to the prospects for conflict), we take the changes in Chinese sectoral exports
to the world, excluding Israel. The second instrument is the 1990s decline in
Israeli import duties in most sectors, which reduced Palestinian exporters’
preferential access to the Israeli market, thus reducing the demand for
Palestinian goods.
Figure 1.10 Israeli Imports from China and the West Bank and Gaza, 1995–2000
Million US dollars
600
500
400
300
200
100
0
1995 1996 1997 1998 1999 2000
West Bank and Gaza China
Sources: UN Commodity Trade Statistics Database and Palestinian Central Bureau of Statistics.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
50 How Trade Can Affect Conflict
Both of these instruments are effective predictors of Palestinian exports. And
these predictions of Palestinian exports are significantly related to conflict inten-
sity during the Second Intifada. We can therefore conclude that our export
variable is not driven by conditions in the West Bank and Gaza, so that our esti-
mation of the impact of Palestinian exports on conflict intensity is not affected
by endogeneity. A more formal description of the procedure used to test for
endogeneity is given in appendix B.
Empirical Results—Conclusions
Overall, the results suggest that changes in Palestinian exports (to Israel and to a
more limited extent to the rest of the world) during the period before the
Second Intifada reduce the intensity (and the probability) of conflict during the
Second intifada. An increase of $10 million in Palestinian exports of a sector
employing 10 percent of a locality’s private employment reduces conflict-related
fatalities in that locality by between 2.1 percent and 2.8 percent. This increase
also reduces the probability of the eruption of conflict in the locality by between
5.3 percent and 5.5 percent.
Each of the trade variables is the weighted average of changes in exports
(or imports) by sector, where the weights reflect the sectoral distribution of pri-
vate employment in each locality. Therefore, these results can be interpreted as
the outcome of variation in employment opportunities induced by export
(and import) changes. These findings support the opportunity cost hypothesis,
that participation in violence increases with declines in real incomes, since the
cost in terms of foregone income declines.
The Israeli-Palestinian conflict provides an unusually straightforward test of
the opportunity cost hypothesis. That is, some studies find that export prices are
positively related to conflict, because competition for valuable, point-source
commodities intensifies as their price rises. This channel is unlikely to operate in
the Israeli-Palestinian context, as the Palestinian export basket does not include
such goods. Similarly, some studies find a positive relationship between agricul-
tural commodity prices and conflict, because increases in the revenues from
agriculture are used to fund rebel activities. This channel is also unlikely to oper-
ate in the Palestinian context, because of the relative inability of the Palestinian
authority to tax earnings or trade.
Implications for Trade Policies Affecting the West Bank and Gaza
Our findings suggest that dealing with adverse employment shocks is critical to
avert conflict, or reduce its intensity, in fragile environments. The promotion of
labor-intensive export sectors appears to be a useful strategy in this context. The
evidence in the paper suggests that the policies of fragile country’s trading part-
ners can effectively encourage exports, for example by increasing preferential
market access.
Facilitating Palestinian trade can be an important strategy to reduce the risk of
conflict. This entails better control by Palestinians of their own border, an
improved system for tax collection, and renewed trade agreements with Israel on
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 51
a more equal footing. Improving trade will also require the development of a
strong Palestinian private sector. A large amount of evidence suggests that this
development cannot be achieved without the removal of the many Israeli mea-
sures which restrict the activity of the Palestinian private sector, such as move-
ment and access restrictions, the West Bank wall, the blockade in Gaza, and
impediments to access to natural resources (Calì and Miaari 2013; UNCTAD
2011; World Bank 2007; Niksic et al. 2014).
In addition, our results point to the crucial importance of Palestinian sales to
the Israeli market (rather than to any other market) as a way to reduce the prob-
ability of intense conflict. On the one hand, this suggests the importance of
facilitating the purchase of Palestinian goods and services by Israelis. That is also
in line with the finding by Miaari, Zussman, and Zussman (2014) on Palestinian
workers in Israel. On the other hand, this dependence on Israel implies a high
exposure of the Palestinian private sector to changes in the Israeli market, such
as the Israeli trade liberalization of the 1990s. As we have shown, such economic
shocks can have important consequences for conflict. Therefore, Palestinian
exporters need to be able to diversify their markets. Intensifying trade links with
the Arab economies would be a natural starting point. That would require
Palestinian sovereignty over its own trade policy, as well as the development of
trade facilitation measures linking the West Bank and Gaza with neighboring
Arab countries.
Finally, our findings support the hypothesis that economic opportunities
affect the dynamics of the Israeli-Palestinian conflict. How to guarantee such
opportunities and how changes in trade policies might interact with local politi-
cal institutions operating within the Palestinian-Israeli conflict are fruitful ave-
nues for future research.
Notes
1. The measurement of export concentration used is the herfindahl index, where higher
values indicate greater concentration.
2. Figures obtained via MIT’s Observatory of Economic Complexity at atlas.media.mit.edu
(see Hausmann et al. 2011; Simoes and Hidalgo 2011).
3. Based on available WDI data for 22 fragile countries from the OECD list. The differ-
ence was statistically significant at the 0.05 level.
4. When incomes and employment are higher, so is the expected income foregone due
to allocating time to violence rather than to an economically productive activity.
5. This rapacity channel is also supported by an emerging literature which finds that aid
may increase violence in conflict contexts by raising the value of contestable resources
that groups fight over (Nunn and Qian 2013).
6. This mechanism can plausibly account for some of the “resource curse” hypothesis
(van der Ploeg 2011), whereby resource rich countries tend to have worse economic
performance than other countries.
7. That stability can also be (and in fact it is often) associated with authoritarian, non-
democratic political regimes.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
52 How Trade Can Affect Conflict
8. This information is based on the report by Kesan (2012) as well as personal commu-
nications with the main author of the report.
9. This effect is likely to be smaller the larger is the mobility of labor across industries,
which would allow workers to adjust to negative industry-specific shocks.
10. In fact, if the reduction in the incentive to produce is large enough to substantially
reduce domestic production, the net effect of an export ban could even be negative
for consumers.
11. Carter and Bates (2012) argue that that is the case only in countries where there is little
political competition (i.e. authoritarian governments) while democracies have to please
the majority of the voters, who reside in rural areas in a typical low-income country.
12. The authors define a “regional war complex” as a conflict which has high foreign par-
ticipation, and domestic participation inside at least one of the countries involved in
the violent conflict that is high enough to challenge the government’s monopoly of
force in that country.
13. The same principle explains why countries which have had past bilateral wars are more
likely to sign trade agreements with each other (Martin, Mayer, and Thoenig 2012).
14. Gleditsch (2007) also notes that countries support conflicts in neighboring countries
to a large extent on the basis of their affinity or antipathies to existing regimes. Trade
is one indicator of compatibility between states.
15. Other studies have focused on the impact of trade, including with neighbors, on inter-
state wars (e.g. Martin, Mayer, and Thoenig 2008).
16. This estimate is conditional on the changes in the export price index in the preceding
two periods.
17. Our result is also different from the negative coefficient of Bruckner and Ciccone
(2010). However, our specification is not comparable to theirs, as we run the regres-
sions for onset and ending separately on split samples as in Bazzi and Blattman (2014).
In addition we also use the log of the price level instead of price change.
18. In the model of Dal Bó and Dal Bó (2011) this would lead to a relatively small size
of the appropriative conflict sector, i.e., a low probability of conflict.
19. This result appears to contradict those of two other recent studies, Berman and
Couttenier (2014) and Chaudion et al. (2012). However, the former considers only
large economic downturns in trading partners (i.e. banking crises) and focuses on
conflict eruption at the subnational level. The latter uses the indicator of market
demand as an instrument for economic growth rather than as an independent regres-
sor, and no lagged structure is used.
20. In order to test this hypothesis, we replace the share of trade with neighbors with the
same variable interacted with an RTA dummy which takes the value of 1 for each
country-pair that had an RTA between them by the year before the observation. In
this way the variable becomes the share of trade with neighbors under RTAs in a
country’s total trade.
21. As showed in table B.1, this variable has a negative and significant association with any
conflict onset (columns 1–4), while it has no significant association with major conflict
onset (columns 5–8).
22. In order to make the regressions comparable, we rerun the regression with PRIO data
using the same battle deaths threshold of major conflict as in COW (i.e. at least 1,000
battle deaths over the entire course of the conflict).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 53
23. To save clutter, for each price variable in table B.4 we only report the value and signifi-
cance of the sum of the three coefficients (the contemporaneous and the two lagged
variables). Unless otherwise stated, we also do not report the coefficients of the other
trade variables as they are little affected by the splitting of the price indices (results
available upon request).
24. The only exception is the first lag of the export price index, which has a negative and
significant effect on a conflict ending. However, this effect is offset by that of the
contemporaneous and second lag terms, and the sum of the export price index coef-
ficients is not significant.
25. The result is not significant when using fast moving weights (column 5).
26. A drop in demand from a country’s main market for an agricultural commodity typi-
cally will not greatly affect export volumes, given the opportunities for switching to
other markets.
27. Trade may affect the incentive for the neighbors to intervene in a country’s conflict
also for other reasons than the opportunity cost channel. For example trade may raise
the trust between the peoples of the neighboring countries (Rohner, Thoenig, and
Zilibotti 2013) thus reducing the propensity of countries to fuel conflict in their
neighbors.
28. The land rights related to indigenous people are of particular concern for Fulani pas-
toralist in Plateau (and other states), as pastoralists by definition do not own the land
their herds graze upon when they are on the move. Expanding cities and agriculture,
in addition to the outermost northern pastoralist routes becoming irregularly dry, has
led the Fulani pastoralists to clash with farmers, many of whom are indigenous. This
is not exclusive to Plateau State, as seen in the 2013 small outbreaks of violence in
Benue State (Human Rights Watch 2013).
29. There was tension in the state in October 2013 following bloody clashes between
Fulani herdsmen and Yoruba inhabitants at Alapa/Onire in Asa Local Government
Area of the state.
30. Tables showing the estimated coefficients for each of the independent variables, under
various specifications of the model, are presented in appendix B.
31. The lack of significance of the oil price index after 2009 occurs because we can
include only one year after the 2009 agreement, as opposed to two years with the use
of the lag.
32. The purpose is to control for potential negative autocorrelation of prices over time
periods (Bazzi and Blattman 2014).
33. Note we are unable to run the model with fatalities using indices based on broad
matching and on international prices due to lack of convergence.
34. In an interview with IRIN Human Rights Watch researcher Eric Guttschuss noted
that “Boko Haram is essentially the fallout of frustration with corruption and the
attendant social malaise of poverty and unemployment.” Similarly Paul Lubeck, a
professor at the University of California, Santa Cruz, who studies the group, argues in
an interview that Boko Haram tapped into growing anger among northern Nigerians
at their poverty and lack of opportunity (Nossiter and Kirkpatrick 2014).
35. Part of this export in the 1990s was due to Israeli firms using Palestinian firms as
subcontractors in a number of sectors, such as textile, garments, and furniture.
36. Since the withdrawal of Israel from Gaza in 2005, the border between Gaza and the
Arab Republic of Egypt is no longer controlled by Israel although the blockade of
Gaza effectively implies no formal trade between Gaza and Egypt.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
54 How Trade Can Affect Conflict
37. Authors’ calculations based on PCBS’ National Accounts.
38. Authors’ calculations based on PCBS’ trade data and National Accounts.
39. The distribution is obtained through a kernel density function (with 1.5 bandwidth).
40. The sociodemographic variables refer to 1997 and include the total population, share
of males in the population, share of the population aged 15–40, share of population
with elementary education or below, share of households with more than 8 members,
and the share of married individuals. We also control for factors that may foster
Palestinian grievances, including the availability of public utilities such as water, elec-
tricity, sewage, and landline telephones; the unemployment rate in 1997 (computed
from the Census); and the number of permits to work in Israel in 1999, which Miaari,
Zussman, and Zussman (2014) show to be associated with the subsequent violence.
Finally, to control for the cyclicality of the conflict, we include the number of
Palestinian fatalities in each locality between January 1995 and August 2000.
41. Marginal effects are calculated as incidence rate ratios and are equal to 1-e.
42. The agency reports that Yoram Cohen, the chief of Israel’s Shin Bet security agency,
in a meeting of the Knesset Foreign and Security Affairs Committee in February 2014
argued that “the uptick in Palestinian militant attacks against Israelis in the past year
can be traced back to the economic grievances Palestinians are suffering from in the
West Bank” (Xinhua 2014).
43. In fact this inclusion reduces the absolute size of the export coefficient (from −0.159
in column 4 of table B.16 to −0.137 in column 3). Column 4 uses the same specifica-
tion as in column 2 but run over this restricted sample.
44. The coefficient is more precisely estimated in the West Bank than in Gaza, probably
due to the small sample size in the latter. However it is larger in magnitude in Gaza.
45. The period covered is 1996–1999. We subtract Palestinian imports from Israel from
the Israeli imports from the rest of the world, as a large part of Palestinian imports are
processed through Israeli firms and are recorded as part of Israel’s total imports. The
result is also robust to including Palestinian imports from Israel (results available from
the authors upon request).
46. Although even the exports to Israel variable is significant only at the 15 percent level
over the Palestinian Labor Force Survey sample, with import variables included in the
estimation (see column 7 of table B.17).
47. This estimation uses a probit model. The impact of exports on the probability of
conflict increases to between 6.3 percent and 6.9 percent in the case of exports to
Israel, while is it not significant for the exports to the rest of the world (columns 3 and
4 of table B.20). Again, the effect of changes in imports is not significant (columns 2
and 4).
References
ACLED. 2013. Country Report: Nigeria. April.
Aksoy, A., and F. Ng. 2008. “Who Are the Net Food Importing Countries?” Policy Research
Working Paper 4457, World Bank.
Angrist, J., and A. Kugler. 2008. “Rural Windfall or a New Resource Curse? Coca, Income,
and Civil Conflict in Colombia.” The Review of Economics and Statistics 90 (2):
191–215.
Arezki, R., and M. Brückner. 2011. “Food Prices, Conflict, and Democratic Change.” Mimeo.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 55
Bank of Israel. 2010. “Recent Economic Developments.” Tel Aviv, May–August, 128.
Bazzi, S., and C. Blattman. 2014. “Economic Shocks and Conflict: Evidence from
Commodity Prices.” American Economic Journal: Macroeconomics.
Becker, G. 1968. “Crime and Punishment: An Economic Approach.” Journal of Political
Economy 76: 169–217.
Bellemare, M. F. 2011. “Rising Food Prices, Food Price Volatility, and Political Unrest.”
Duke University, Mimeo.
Bellows, J., and E. Miguel. 2009. “War and Local Collective Action in Sierra Leone.”
Journal of Public Economics 93 (11–12): 1144–57.
Berman, N., and M. Couttenier. 2014. “External Shocks, Internal Shots: The Geography of
Civil Conflicts.” CEPR Discussion Paper 9895.
Berrebi, C. 2007. “Evidence about the Link between Education, Poverty and Terrorism
among Palestinians.” Peace Economics, Peace Science and Public Policy 13 (1): Article 2.
Besley, T., and T. Persson. 2008. “The Incidence of Civil War: Theory and Evidence.”
Mimeo.
Bhavnani, R. R., and S. Jha. 2011. “Trade Shocks, Mass Mobilization and Decolonization:
Evidence from India’s Independence Struggle.” Mimeo.
Blattman, C., and E. Miguel. 2010. “Civil War.” Journal of Economic Literature 48: 3–57.
Bruckner, M., and A. Ciccone. 2010. “International Commodity Prices, Growth and the
Outbreak of Civil War in Sub-Saharan Africa.” The Economic Journal 120 (May):
519–34.
Buhaug, H., and K. S. Gleditsch. 2008. “Contagion or Confusion? Why Conflicts Cluster
in Space.” International Studies Quarterly 52 (2): 215–33.
Buhaug, H., and J. K. Rod. 2006. “Local Determinants of African Civil Wars, 1970–2001.”
Political Geography 25: 315–35.
Cadot, O., and J. Gourdon. 2012. “Assessing the Price-Raising Effect of Non-Tariff
Measures in Africa.” CEPII Working Paper 2012–16.
Calì, M., and S. Miaari. 2013. “The Labor Market Impact of Mobility Restrictions:
Evidence from the West Bank.” Policy Research Working Paper 6457, World Bank.
Calì, M., S. Miaari, and B. Fallah. 2014. “More Jobs, Less Fighting? Evidence from the
Second Intifada.” Mimeo.
Chaudion, S., Z. Peskowitz, and C. Stanton. 2012. “Beyond Zeroes and Ones: The Effect
of Income on the Severity and Evolution of Civil Conflict.” Mimeo.
Collier, P., and A. Hoffler. 1998. “On Economic Causes of Civil War.” Oxford Economic
Papers 50 (4): 563–73.
Collier, P., and A. Hoffler. 2004. “Greed and Grievance in Civil War.” Oxford Economic
Papers 56 (4): 563–96.
Copeland, F. 2013. “The Boko Haram Insurgency in Nigeria, Comprehensive Information
on Complex Issues.” https://www.cimicweb.org/cmo/medbasin/Holder/Documents
/r028%20CFC%20Monthly%20Thematic%20Report%20(21-FEB-13).pdf.
Cotet, A. M., and K. K. Tsui. 2013. “Oil and Conflict: What Does the Cross Country
Evidence Really Show?” American Economic Journal: Macroeconomics 5 (1): 49–80.
Council on Foreign Relations. 2014. “Boko Haram.” Background Paper.
Dal Bó, E., and P. Dal Bó. 2011. “Workers, Warriors and Criminals: Social Conflict in
General Equilibrium.” Journal of the European Economic Association 9 (4).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
56 How Trade Can Affect Conflict
Dube, O., and J. Vargas. 2013. “Commodity Price Shocks and Civil Conflict: Evidence
from Colombia.” Review of Economic Studies 80 (4): 1384–1421.
FAO (Food and Agriculture Organization of the United Nations). 2002. “FAO Papers on
Selected Issues Relating to the WTO Negotiations on Agriculture.” Commodities and
Trade Division, Rome. ftp://ftp.fao.org/docrep/fao/004/Y3733E/Y3733E00.pdf.
Fearon, J., and D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American Political
Science Review 97 (1): 75–90.
Gersovitz, M., and N. Kriger. 2013. “What Is a Civil War? A Critical Review of Its
Definition and (Econometric) Consequences.” World Bank Research Observer 28 (2):
159–90.
Gleditsch, K. S. 2007. “Transnational Dimensions of Civil War.” Journal of Peace Research
44: 293–309.
Gleditsch, N. P., P. Wallensteen, M. Eriksson, M. Sollenberg, and H. Strand. 2002. “Armed
Conflict 1946–2001: A New Dataset.” Journal of Peace Research 39 (5): 615–37.
Harari, M., and E. La Ferrara. 2012. “Conflict, Climate and Cells: A Disaggregated
Analysis.” IGIER Working Paper 461, Bocconi University.
Hsiang, S. M., and M. Burke. 2014. “Climate, Conflict, and Social Stability: What Does the
Evidence Say?” Climatic Change 123: 39–55.
Hull, P., and M. Imai. 2013. “Economic Shocks and Civil Conflict: Evidence from Foreign
Interest Rate Movements.” Journal of Development Economics 103: 77–89.
Independent Evaluation Group. 2011. “Timor-Leste Country Program Evaluation,
2000–2010.” Independent Evaluation Group, World Bank, Washington, DC.
International Crisis Group. 2014. Curbing Violence in Nigeria (II): The Boko Haram
Insurgency, Africa Report 216.
IRIN. 2011a. “Understanding Nigeria’s Boko Haram Radicals.” July.
———. 2011b. “Analysis: Niger Delta Still Unstable despite Amnesty.” November.
Ivanic, M., and W. Martin. 2008. “Implications of Higher Global Food Prices for Poverty
in Low-Income Countries.” Agricultural Economics 39 supplement: 405–16.
Jacoby, H. 2013. “Food Prices, Wages, and Welfare in Rural India.” Policy Research Working
Paper 6412, World Bank.
Kelleher, S., and J. D. Reyes. 2014. “Technical Measures to Trade in Central
America: Incidence, Price Effect, and Consumer Welfare.” World Bank Working
Paper 6857.
Kesan. 2012. “Economic Development and Peace in Karen State: Scenarios, Principles &
Policies.” Unpublished report.
King, G., and L. Zeng. 2001. “Logistic Regression in Rare Events Data.” Political Analysis
9: 137–63.
Lin, Y-H., and G. Michaels. 2011. “Do Giant Oilfield Discoveries Fuel Internal Armed
Conflicts?” CEPR Discussion Paper 8620.
Lipton, M. 1977. Why Poor People Stay Poor: Urban Bias in World Development. Cambridge,
MA: Harvard University Press.
Martin, P., T. Mayer, and M. Thoenig. 2012. “The Geography of Conflicts and Regional
Trade Agreements.” AEJ: Macroeconomics 4 (4): 1–35.
Maystadt, J-F., G. De Luca, P. G. Sekeris, J. Ulimwengu, and R. Folledo. 2014. “Mineral
Resources and Conflicts in DRC: A Case of Ecological Fallacy?” Oxford Economic
Papers 66 (3): 721–49. http://oep.oxfordjournals.org/content/66/3/721.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Can Affect Conflict 57
Maystadt, J.-F., and O. Ecker. 2014 “Extreme Weather and Civil War: Does Drought Fuel
Conflict in Somalia through Livestock Price Shocks?” American Journal of Agricultural
Economics.
Maystadt, J.-F., J.-F. Trinh Tan, and C. Breisinger. 2014. “Does Food Security Matter for
Transition in Arab Countries?” Food Policy 46: 106–15.
Miaari, S., A. Zussman, and N. Zussman. 2014. “Employment Restrictions and Political
Violence in the Israeli-Palestinian Conflict.” Journal of Economic Behavior and
Organization 101 (May): 24–44.
Miguel, E., S. Satyanath, and E. Sergenti. 2004. “Economic Shocks and Civil Conflict:
An Instrumental Variables Approach.” Journal of Political Economy 112 (4): 725–53.
Niksic, O., N. Nasser Eddin, and M. Calì. 2014. Area C and the Future of the Palestinian
Economy. Washington, DC: World Bank.
Nossiter, A., and D. D. Kirkpatrick. 2014. “Abduction of Girls an Act Not Even Al Qaeda
Can Condone.” The New York Times, May 7.
Nunn, N., and N. Qian. 2014. “U.S. Food Aid and Civil Conflict.” American Economic
Review 104 (6): 1630–66.
OECD. 2013. Fragile States 2013: Resource Flows and Trends in a Shifting World. Paris:
OECD.
Portugal-Perez, A., and J. S. Wilson. 2012. “Export Performance and Trade Facilitation
Reform: Hard and Soft Infrastructure.” World Development 40 (7): 1295–307.
Revenue Watch Institute. 2013. “The 2013 Resource Governance Index.” Washington,
DC: Revenue Watch Institute.
Rohner, D., M. Thoenig, and F. Zilibotti. 2013. “War Signals: A Theory of Trade, Trust and
Conflict.” Review of Economic Studies 80 (3): 1114–47.
Saleh, B. A. 2009. “An Econometric Analysis of Palestinian Attacks: An Examination of
Deprivation Theory and Choice of Attacks.” European Journal of Social Sciences 7 (4):
17–29.
Savun, B., and D. C. Tirone. 2012. “Exogenous Shocks, Foreign Aid, and Civil War.”
International Organization 66: 363–93.
Sayne, A. 2013. What’s Next for Security in the Niger Delta? U.S. Institute for Peace Special
Report 333.
Sayre, E. A. 2009. “Labor Market Conditions, Political Events, and Palestinian Suicide
Bombings.” Peace Economics, Peace Science and Public Policy 15 (1): 1–26.
Simoes, A. J. G., and C. A. Hidalgo. 2011. “The Economic Complexity Observatory: An
Analytical Tool for Understanding the Dynamics of Economic Development.”
Workshops at the Twenty-Fifth AAAI Conference on Artificial Intelligence.
Treichel, V., M. Hoppe, O. Cadot, and J. Gourdon. 2012. “Import Bans in Nigeria Increase
Poverty.” Africa Trade Policy Note 28, World Bank.
UNCTAD. 2011. Developments in the Economy of the Occupied Palestinian Territory.
Geneva: UNCTAD Secretariat.
van der Ploeg, F. 2011. “Natural Resources: Curse or Blessing?” Journal of Economic
Literature 49 (2): 366–420.
World Bank. 2007. Movement and Access Restrictions in the West Bank: Uncertainty and
Inefficiency in the Palestinian Economy. Washington, DC: World Bank.
———. 2010. “Doing Business Indicators, Trading across Borders.” Online data, http://
www.doingbusiness.org/data/exploretopics/trading-across-borders.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
58 How Trade Can Affect Conflict
———. 2011. World Development Report 2011: Conflict, Security and Development,
Washington, DC: World Bank.
———. 2014a. “Doing Business Indicators, Trading across Borders.” Online data. http://
www.doingbusiness.org/data/exploretopics/trading-across-borders.
———. 2014b. “World Development Indicators.” Online data. http://databank
.worldbank.org/data/views/variableSelection/selectvariables.aspx?source=world
-development-indicators.
———. 2014c. “Logistics Performance Index.” Online data. http://lpi.worldbank.org/.
World Economic Forum. 2014. “Quality of Port Infrastructure.” Online data. http://data
.worldbank.org/indicator/IQ.WEF.PORT.XQ.
Wudu, W.S. 2013. “Machar Accuses Government of Embezzling $ 4.5 Billion.” December 7.
http://www.gurtong.net/ECM/Editorial/tabid/124/ctl/ArticleView/mid/519
/articleId/14062/Machar-Accuses-Government-Of-Embezzling-45-Billion.aspx.
Xinhua. 2014. “Palestinian Militancy Stems from Economic Hardship: Israeli Security
Chief.” February 2.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Chapter 2
Conditions That Affect the Impact
of Trade Shocks on Conflict
Introduction
Changes in economic activity, including those originating in external trade, do
not affect peace and stability in a vacuum. Deep-seated factors, often dating back
in history, need to be in place in order to create the conditions for conflict and
political instability. Norway, Canada, the Republic of South Sudan, and Nigeria
are all oil rich economies, but swings in oil prices have little potential to create
instability in the first two countries and a great deal of potential in the last two.
Changes that affect trade in goods have different effects across countries even in the
same region. Increased diamond production and trade, for example, has been asso-
ciated with higher levels of conflict in the Central African Republic (International
Crisis Group 2010) and in Sierra Leone (Bellows and Miguel 2009), but not in
Botswana, where the rents from diamonds were shared across all groups in the
society. The widespread perception that this sharing agreement was fair helped
ensure stability (Robinson, Acemoglu, and Johnson 2003).
These examples point to a basic but often forgotten principle: the expected
effects of changes in trade on instability are likely to differ across countries, and
across regions within countries. Various factors are likely to facilitate or hinder
the extent to which these changes can create instability. Improving our under-
standing of how the relationship between conflict and changes in trade differs
across contexts is important for at least three reasons. First, it would help identify
some of the conditions under which changes in trade are more harmful. This
information would be important in developing proper monitoring frameworks
for future sharp changes in trade volumes or prices. Second, it would allow policy
makers to identify, and possibly address, the conditions that make countries
vulnerable to changes in trade flows. And third, it would improve our analysis of
the extent to which trade-related changes affect conflict. That is, analysis which
bundles together countries with very different conditions may conclude that
trade-related changes have no impact on conflict, while actually they may be
important for conflict, but only in countries that meet certain conditions.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 59
60 Conditions That Affect the Impact of Trade Shocks on Conflict
Unfortunately, our understanding of the conditions under which changes in trade
trigger conflict remains limited. Scholars have acknowledged the importance of this
heterogeneity. Blattman and Miguel (2010, p. 31) note that “there is good reason to
believe that the relationships between civil conflict and income shocks … should
be conditional ones, evident primarily when interacted with other contextual
variables.” However, macroeconomic studies of the relationship between trade-
related changes and conflict have rarely tested systematically for these condition-
ing factors, and no consensus has emerged on their effects.
The local conditions that might affect the relationship between changes in trade
and conflict include: (a) grievances that foster tensions among groups, for example
generated by economic inequality, ethnic and religious differences, and past conflict
events; (b) the state’s institutional capacity and political structure, particularly
whether government policies foster inclusiveness; (c) conditions in neighboring
countries that might encourage or discourage conflict; and (d) policies that affect the
transmission of changes in international commodity prices to the domestic market.
In general, our cross-country analysis confirms the expectation that local con-
ditions, such as more peaceful neighborhoods, more limited ethnic divisions,
lower inequality, and better governance, tend to reduce the impact of changes in
trade on conflict. Variables related to grievances and to conditions in neighboring
countries appear to be particularly important in this context. While the quality
of governance (for example the degree of accountability, level of corruption, and
quality of the bureaucracy) has a significant impact on the relationship between
export price changes and conflict, the impact of political arrangements (for
example the degree of democracy) is more limited. Finally, interventions that
slow the transmission of changes in international commodity prices to domestic
markets appear to reduce the risk of conflict from changes in export prices,
except in the case of point-source commodities.
Similar, although not identical, results emerge from the country case studies.
In Nigeria, political factors play an important role in mediating conflict, as the
impact of commodity price increases in conflict rises in election years. Delays in
the transmission of international prices to local markets reduce the impact of
price changes in conflict. Ethnic divisions and economic inequality, but not unem-
ployment or poverty rates, magnify the impact of higher oil prices on conflict.
Interestingly, past incidents of violence are not shown to increase the impact of
commodity prices on conflict. In the West Bank and Gaza, the impact of changes
in exports on conflict is increased by the existence of grievances such as the pres-
ence of refugees and high unemployment rates, but not the incidence of violent
fatalities in the past. Export changes have a greater impact on conflict in localities
with a larger share of males (perhaps reflecting a higher share of local residents in
Israeli jails), but not in localities with lower levels of education (despite the likeli-
hood that more educated individuals face a higher opportunity cost of conflict).
The chapter begins with a brief review of the possible factors mediating the
impact of trade-related changes on conflict. The evidence on the interaction
between these factors and trade changes in determining conflict is relatively thin,
and much of the economic literature on the causes of conflict has focused on
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 61
changes in income from sources other than external trade. The section titled
“Grievances” discusses various kinds of grievances that may affect the relationship
between changes in trade and conflict.The section titled “Institutional Capacity and
Inclusiveness” does the same for the institutional and political context. The section
titled “Conditions in Neighboring Countries” discusses the role of conditions in
neighboring countries, and the section titled “Transmission of Prices to Domestic
Markets” the transmission of changes in international prices to the domestic market.
We then test some of the tentative hypotheses based on this review. The section
titled “Cross-Country Empirical Tests” expands on the models discussed in
chapter 1 to provide an empirical test of the importance of these factors across
countries. The next two sections do the same for our two case studies, Nigeria
and the West Bank and Gaza.
Grievances
The literature suggests—albeit with some relevant exceptions—that grievances are
a major source of conflict and political unrest. Grievances may arise from a variety of
factors, such as the exclusion of certain groups from access to resources and/or
political participation, inequality in ownership of assets and in opportunities, or
past injustices that have not been redressed. All of these factors can be direct
determinants of conflict, but can also create conditions that result in trade-
related changes having an impact on conflict. In order to illustrate the r elationship
between grievances and the impact of trade on conflict, we focus on three impor-
tant determinants of grievances: ethnic divisions, economic inequality, and past
incidence of violence. Unfortunately, the evidence on the importance of these
factors in determining the outcome of trade-related changes on conflict is
extremely limited. Instead, we rely on evidence on the interaction between
general changes in real incomes and g
rievances in determining conflict.
Ethnic/Religious Differences
Divisions along ethnic lines have been associated with adverse economic and
political outcomes. Easterly and Levine (1997) and Alesina et al. (2003) argue
that African countries that are more divided along ethnic and linguistic lines are
more susceptible to competitive rent-seeking across different groups. One conse-
quence is that these countries are less likely to develop the public goods of infra-
structure, education, and strong political institutions necessary to sustain
economic growth. Another consequence of these kinds of division is a greater
tendency toward conflict (Buhaug, Cederman, and Gleditsch 2011; Cederman,
Girardin, and Gleditsch 2009), although this finding is controversial (Collier and
Hoeffler 2004; Fearon and Laitin 2003). More recent evidence supports the idea
that ethnic and religious diversity plays a key role in conflict across communities
within a country. For example, Blair, Blattman, and Hartman (2012) find that
ethnic and religious diversity is one the most powerful predictors of the onset
of conflict in 247 communities in Liberia. The risk of conflict appears to be
lower in communities populated predominately or exclusively by a single tribe.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
62 Conditions That Affect the Impact of Trade Shocks on Conflict
Other studies provide evidence that the role of ethnic divisions is important in
understanding the impact of changes in income on civil conflict (Hull and Imai
2013) and political risk (Bruekner and Gradstein 2014).1
The direct effect of ethnic divisions on conflict suggests that political stabil-
ity may be more vulnerable to economic changes in ethnically polarized countries
than in more ethnically homogenous countries. Blimes (2006) finds that ethnic
cleavages have an important role in conditioning the effects of various determi-
nants of conflict. In his cross-country analysis, he finds that these effects are under-
estimated for countries with higher levels of ethnic cleavages and overestimated
for those with lower levels. Bruekner and Gradstein (2014) also find that the
higher the degree of a country’s ethnic polarization, the more adverse is the impact
of income growth on political risk.2 By contrast, Hull and Imai (2013) find that the
higher the degree of ethnic fractionalization, the more that income growth reduces
conflict (and the more that recessions increase conflict).3 The diverging results
found in Bruekner and Gradstein (2014), who measure polarization, and Hull and
Imai (2013), who measure fractionalization, are consistent with other studies sug-
gesting that these two concepts may have opposite effects on conflict (Montalvo
and Reynal-Querol 2005). Alternatively, these different results may reflect meth-
odological differences between the two studies, such as the dependent variables
used and the source of the change in real incomes.4 In any event, this difference in
the empirical findings is striking and calls for investigation into how ethnic diver-
sity affects the relationship between trade-related changes and conflict.
Economic Inequality
Grievances may be generated by economic inequality. However, the extent to
which this type of grievance generates conflict is disputed. Collier and Hoeffler
(2004) and Fearon and Laitin (2003) find no support for this hypothesis across
countries. Acemoglu and Robinson (2006) note that with high asset inequality,
the wealthy may be more willing to invest in repression in order to deter revolu-
tion. However, income inequality has been found to increase conflict where
conflicts have a strong component of political ideology. For example, income and
asset inequality have been shown to be key determinants of Nepal’s Maoist insur-
gency (Nepal, Bohara, and Gawande 2011), the Maoist/Naxalite conflict in India
(Gomes 2011), and the conflict in South Mexico (Maystadt 2008).
Inequality in land ownership is particularly important in developing countries,
because land often represents a major household asset, offers key income opportu-
nities, and serves as the main collateral for rural dwellers’ access to credit. Inequality
in land ownership often is manifested in a high percentage of landless households.
In this case, a reduction in real incomes may increase the incentive of the population
to rebel. For example, Hidalgo et al. (2010) find that reductions in real incomes
induced by adverse weather caused the rural poor to invade large landholdings, and
that this effect was twice as large in municipalities with high land inequality than
in municipalities with low land inequality. Hidalgo et al. (2010) is one of the rare
empirical studies on the mediating power of inequality in the relation between
economic shocks and conflict.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 63
Incidence of Past Conflict
One of the reasons why countries tend to fall back into conflict is that their citizens
carry the burden of past violence and injustice. This burden may increase the sensi-
tivity of these countries to changes in economic conditions. Blair, Blattman, and
Hartman (2012) identify the presence of ex-combatants and the exposure to
wartime violence as one of the four factors predicting intra-communal violence
in Liberia. This finding is consistent with the idea that the experiences of war-
time violence may continue to foment tensions even in peacetime. Past victim-
ization, participation in war, and loss of land during the war are all associated
with the future likelihood of violence in Liberia (Blair, Blattman, and Hartman
2012). Evidence from the Israeli-Palestinian conflict also supports the notion that
past violence is related to successive violence (Miaari, Zussman, and Zussman
2014). However, we are not aware of any studies that relate the sensitivity of
trade-related changes to grievances from past violence.
Institutional Capacity and Inclusiveness
The effectiveness of government institutions, and the degree to which government
strives to take the concerns of minorities in consideration when formulating poli-
cies, are likely to be important in determining whether trade-related changes
have a significant impact on the onset or intensity of conflict.
Institutional Capacity
One of the mechanisms through which changes in trade may affect conflict is by
raising the value of the “prize” of the conflict. However, the extent to which such
a “prize” can be appropriated through fighting also depends on the capacity of
the state to maintain control over the resources. The higher such capacity, the
higher the probability that increases in the value of disputable wealth may be
appropriated by the state, e.g. through higher fiscal revenues, rather than used by
insurgents. For example, Angrist and Kugler (2008) argue that the weakness of
the Colombian government in the countryside allowed guerrillas and paramili-
revenues from increased coca prices at the end of the 1990s. That
taries to raise
helps explain why the rural areas which expanded coca production subsequently
became considerably more violent, while urban areas were virtually unaffected.
The issue of state capacity may be particularly important for oil exporters. Fearon
(2005) argues that these countries are particularly prone to civil war because of their
relatively weak state institutions, at least compared to other countries with similar
per capita income levels. States with high oil revenues tend to have less incentive to
develop administrative capabilities and control of their territory (Fearon and
Laitin 2003). This relatively low state capacity may also make oil wealth a par-
attractive prize for potential insurgents.
ticularly
Type of Political Regime
While it seems reasonable that states with more effective institutions are better at
mitigating the potential impact of trade change on conflict, it is more difficult to
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
64 Conditions That Affect the Impact of Trade Shocks on Conflict
evaluate whether more inclusive governments, or alternatively more democratic
governments, are more effective in this context. One view is that countries with
more inclusive state institutions, or stronger democratic institutions (Miguel,
Satyanath, and Sergenti 2004), may be better able to negotiate compromises among
social groups to avoid unrest in the face of an adverse change in trade. On the other
hand, strong, authoritarian states that are not viewed as inclusive may use increases
in the prices of their exports (or reductions in the prices of their imports) to ramp up
military repression, rather than expanding social expenditures to pacify the popula-
tion. It will be difficult to distinguish in empirical analysis between a reduction in
conflict owing to greater repression versus a reduction owing to a more inclusive
approach to government, as in either case the probability of conflict onset (or its
intensity) would decrease.
The empirical evidence is mixed. Besley and Persson (2008) find that com-
modity prices affect conflict only in countries with weak constraints on executive
power. Their empirical results confirm diametrically different effects of commod-
ity price changes between parliamentary and nonparliamentary democracies.5 It is
possible that ethnic groups are more likely to have legitimate, effective channels
to seek redress of their grievances when checks and balances are strong, than when
there is little constraint on executive power. However, this result may just reflect
other structural differences between countries with strong constraints, mostly
high-income Western democracies, and countries with weak constraints, mostly
developing countries. Caselli and Tesei (2013) find that positive commodity price
shocks have no effect on political stability when they occur in democracies.
On the other hand, Bazzi and Blattman (2014) find that commodity price
changes have no significant impact on conflict when considering only high-risk
countries, defined by the type of political system (and, for that matter, the level
of economic inequality and the degree of ethnic polarization).6 Similarly, Miguel,
Satyanath, and Sergenti (2004) do not find any difference in the impact of
weather shocks on civil war between countries differing in terms of democratic
rule, ethno-linguistic fractionalization, type of terrain, income per capita, or oil-
exporting status.
How could studies using similar data, variables, and approaches find such dif-
ferent results? Note that the findings of Bazzi and Blattman (2014) and Miguel,
Satyanath, and Sergenti (2004) also contradict some of the studies cited above, for
example Hull and Imai (2013) and Bruekner and Gradstein (2014) on the importance
of ethnic divisions on conflict. One reason is that the studies differ in significant
details. The analysis in Miguel, Satyanath, and Sergenti (2004) is not strictly compa-
rable with the others, as they restrict the sample to sub-Saharan African countries.
Importantly, Besley and Persson (2008) and Hull and Imai (2013) use a different
way of coding civil wars than Bazzi and Blattman (2014) (the latter’s approach
is explained in appendix A) and Bruekner and Gradstein (2014) examine the
impact on political stability rather than on conflict per se. In addition, Bazzi and
Blattman (2014) use various conflict datasets and find a significant effect of the
conditioning factors in some specifications, although it is not clear in which ones
as they do not report the individual interaction terms.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 65
Conditions in Neighboring Countries
External actors—and neighboring countries in particular—can play a key role in
triggering and developing a domestic civil conflict. Therefore, a country’s politi-
cal stability is likely affected by the characteristics of countries in the region.
Countries in regions with more democracies are less subject to the risk of con-
flict, and the presence of a conflict in a neighboring country increases the risk of
a civil war by almost two-thirds. Gleditsch (2007) finds that the effects of “bad
neighborhoods” can be as important, if not more, in determining the risk of con-
flict than the profile of individual states themselves. In particular, the presence of
transboundary ethnic groups increases a country’s risk of conflict, particularly if
these ties are to groups in a neighboring conflict (Buhaug and Gleditsch 2008).
These results—which are consistent with a broad strand of literature emphasiz-
ing the ethnically driven motives of third party intervention (alluded to above)—
suggest that transnational ethnic linkages constitute a central mechanism for
transmitting conflict between countries.
More importantly for our immediate purpose, ties to neighboring countries
may also change the likelihood that a trade-related change may trigger a conflict.
A conflict in a neighboring country may make it easier for potential rebels to
access weapons to stage a rebellion in their own country, taking advantage of
political disaffection due to an adverse change in trade. Alternatively, countries
that are surrounded by democracies without conflicts and without transnational
ethnic links may be less sensitive to the effects of trade changes on conflict.
Transmission of Prices to Domestic Markets
The extent to which changes in international prices affect civil conflict also
depends on the transmission of prices from international to domestic markets. In
many countries, food price subsidies provide domestic commodity markets some
insulation from changes in international prices. For example, the pass-through
from international to domestic food prices in the Middle East and North Africa
region varies substantially across countries, reflecting different use of consump-
tion subsidies, although some pass through occurs in all of the countries exam-
ined (Ianchovichina, Loening, and Wood 2012). The degree of price transmission
may also vary by commodity market, as Minot (2011) shows in his analysis of
food prices across African countries in 2007–08.
Trade policies, such as price regulation and export bans, can dampen the effects
of changes in international prices on the domestic prices of exported commodities.
For example, until 1999 the largest exports in Côte d’Ivoire—coffee and cocoa—
were regulated by a state-owned marketing board (the Caisse de stabilisation),
which fixed producer prices at a guaranteed level. After it was dismantled, changes
in international prices were fully transmitted to exporters. Losch (2002) argues that
the fall in international cocoa prices in the subsequent years was one of the
causes of the ensuing civil and political unrest in the country.7
Domestic policies that influence commodity price transmission are often
motivated by the need to maintain political order. Carter and Bates (2012) find
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
66 Conditions That Affect the Impact of Trade Shocks on Conflict
olitical competition matters in determining the government’s response to
that p
agricultural price shocks. In particular, authoritarian governments tend to enact
policies that insulate consumers from price hikes, while more democratic ones
tend to focus on protecting rural producers. Their findings confirm that failure to
account for policies that affect the transmission of international prices to domes-
tic markets can bias analysis of the effects of commodity price shocks on conflict.
Similar attention to policies that affect the transmission of prices should be made
in analysis within countries, as the extent of price transmission can vary across
regions. Berman and Couttenier (2014) find that commodity price shocks have
much weaker effects on civil conflict in locations distant from the main seaports
through which imports and exports must transit.
Cross-Country Empirical Tests
These four groups of factors are used to estimate the extent to which differences
among countries affect the impact of trade on conflict: (a) factors affecting griev-
ances, including income inequality, ethno-linguistic and religious fractionalization
and polarization, and the incidence of past conflicts; (b) institutional capacity and
inclusiveness, including the type of political system (e.g. parliamentary versus
nonparliamentary democracies; federal versus unitary government), proxies for
the quality of governance (e.g. bureaucratic quality and degree of government
accountability) and proxies for the quality of parties in office (age of party in
office and the extent to which parties take an ideological approach to economic
issues); (c) conditions in neighboring countries that may affect their general pro-
pensity to destabilize their neighbor, including the presence of a conflict in neigh-
boring countries and the country’s trade ties with its neighbors; and (d) policies
that affect the transmission of international prices to the domestic economy,
as measured by the nominal rate of assistance to agricultural markets (Anderson
et al. 2008). The full list of conditioning factors is reported in table A.3.
We use the model developed in chapter 1 to evaluate how these c onditioning
factors affect the impact of trade-related changes on conflict across countries. We
multiply each of three trade variables (export price index, import price index, and
changes in a country’s principal trade markets) by each of the conditioning vari-
ables (the result is referred to as an interaction term), using our preferred specifica-
tion for the relationship between trade changes and conflict (see column 2 of table
B.1).8 Running these regressions separately maximizes the number of observations,
as most of the conditioning variables are not available for the entirety of our base-
line sample. This strategy also has the advantage of maximizing the degrees of
freedom, compared to including all the conditioning variables in one regression.
With 25 conditioning variables and three types of trade variables, we run a
total of 75 regressions. We also add a set of regressions using the price index for
point-source commodities alone, as this is the main driver of the relationship
between the export price index and conflict. To keep the number of coefficients
to interpret manageable, we only include the contemporaneous trade variable
(and its interaction with the relevant conditioning variable) in the regression,
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 67
without its lagged terms. We do this for two reasons. First, the coefficient of the
contemporaneous variable incorporates to some extent the coefficients of the
lagged trade variables, especially for the export price index.9 Second, the con-
temporaneous term is of particular interest in and of itself, as in our analysis
(and in others) it tends to be the most important one to explain the impact of
trade changes on conflict.
We also include dummy variables to control for conditions in countries
that do not vary over time. This specification effectively restricts our analysis
to those countries where a conflict has begun during our period of analysis
(since 1960).
Results
As expected, the impact on conflict of the export price index, and of the export
price index for point-source commodities, depends significantly on many of the
conditioning variables. Changes in export prices have lower effects on conflict in
countries that are located in peaceful neighborhoods, that have lower ethnic divi-
sions and economic inequality, that have better governance, and that have poli-
cies that reduce the transmission of international prices to domestic markets.
Variables related to grievances and to conditions in neighboring countries appear
to be important in determining the effect of export prices on conflict. Changes in the
overall export price index, and the point-source export price index, have no signifi-
cant effect on the probability of conflict in countries with sufficiently low levels of
economic inequality and of ethnic divisions (whether measured through ethnic
polarization or fractionalization), in countries that didn’t experience any conflict in
the previous 10 years, and in countries whose neighbors are not in conflict, or have
sufficiently intense trade with their neighbors.
The quality of state institutions, as measured by various dimensions of gover-
nance, has a significant impact on the relationship between changes in the export
price index and conflict. In particular, changes in export prices have no effect on
conflict in countries with a high degree of government accountability, a low degree
of corruption, and high bureaucratic quality, although the latter two results do not
carry over to prices on point-source commodities. Other potentially relevant mea-
sures of governance, such as the rule of law and the presence of the military in the
government, do not seem to be important in this context.
Having a federal government appears to consistently reduce the probability that
a change in export prices triggers conflict. That may be because regions tend to have
greater autonomy under federal systems than under unitary government systems,
which may reduce the incentive for challenging the central state at the local
level.10 The other political variables, including the share of programmatic parties
(those with an ideological orientation with respect to economic policy) in power,
the age of the party in office, whether elections were held the previous year, and
the degree of democracy (including also parliamentary democracy) do not yield
robust results for either export prices or point-source commodities prices.
Democratic countries (i.e. a score higher than 5 – out of 10 – in the polity index)
are less subject to the conflict-inducing impact of changes in the prices of
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
68 Conditions That Affect the Impact of Trade Shocks on Conflict
point-source commodities, a result broadly in line with Besley and Persson
(2008).11 These results suggest that the type of state institutions, at least as
defined through commonly available measures, seem to have a limited impact in
determining the effects of export prices on conflict.
Finally, our results support the hypothesis that substantial nominal rates of
assistance to agricultural commodities (whether positive or negative) reduce the
risk of conflict due to changes in export prices. This finding is consistent with the
hypothesis that international prices are transmitted more rapidly to domestic
prices in countries with low distortions to agricultural markets (Anderson et al.
2008). However, this result does not hold for changes in the prices of point-
source commodities. This finding comes with two notes of caution. First, these
distortions also cause reductions in welfare relative to more direct policy instru-
ments aimed to achieve similar domestic policy objectives (Bhagwati 1978;
Corden 1971). Second, the effect we are capturing is inherently short term, as
prices are eventually transmitted to domestic prices in the longer run (Ivanic and
Martin 2013).
On the other hand, the impact on conflict of the price of imports and of changes
in demand in trade partners do not appear to be affected by the conditioning factors
we analyze. These trade variables have generally no significant impact on conflict
across the entire range of the conditioning factors’ values.This result underscores the
weakness of these variables in explaining conflict onset in our model.
Table 2.1 may be helpful in providing a snapshot of these results, although it
requires some explanation. The table reports the range of values of each condition-
ing variable for which the marginal effect of the export price index (and the point-
source price index) on the onset of conflict is not significantly different from zero (at
the 10 percent level). We report this range only when in the rest of the values’ range
the marginal effect is significant. If the marginal effect is always or never statisti-
cally different from zero across the entire values’ range, then we report “No diff.”
For example, ethnic polarization (the fourth variable in table 2.1) has values that
range between zero and one. For all values of less than 0.37, indicating a low
degree of ethnic polarization, the variable has no significant effect on the rela-
tionship between trade changes and conflict. However, where ethnic polarization
is high, it does have a significant impact. This presentation facilitates identifying
those conditions that make certain countries particularly vulnerable or resilient
to changes in the export price index.
Some further insight into a subset of our results can be seen in figure 2.1,
which shows how two different conditioning factors affect the impact of export
prices on conflict. The upper panel presents the curve describing the marginal
effect of the price index, along with its 90 percent confidence interval, across the
range of values of the accountability measure (from the International Country
Risk Guide database). The downward slope indicates that the effect of the price
index becomes less significant as one moves from low to high accountability
observations (i.e. from left to right).12 Thus, when accountability is low, the
estimated impact of the price index on conflict is almost twice as great as in obser-
vations with average accountability. The effect becomes not significantly different
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 69
Table 2.1 Under What Conditions Are the Marginal Effects of Trade Shocks Not Significant?
Type Range Px Point source
Grievance
Economic inequality Contin. [22;65] <26 or >58 <46 or >52
gini_net Contin. [15;75] <32 <48
gini_market Contin. [17;80] <37 <52
Ethnic polarization Contin. [0;1] <0.37 <0.42
Ethnic fractionalization Contin. [0;1] <0.28 <0.30
Religious fractionalization Contin. [0;0.7] No diff. <0.28
Religious polarization Contin. [0;0.96] No diff. <0.49
Any conflict in last 10 yrs Dummy [0;1] 0 0
Political institutions
Elections (t-1) Dummy [0;1] No diff. Elections
Federal Government Dummy [0;1] 1 Federal
Polity2 Contin. [−10;10] No diff. >5
Parliamentary democracy Dummy [0;1] No diff. No diff.
Programmatic party 3 groups [0;3] < or > 2nd tercile No diff.
Age of party in office 3 groups [1;191] No diff. No diff.
Governance
Law Contin. [0;6] No diff. No diff.
Military Contin. [0;6] No diff. No diff.
Accountability Contin. [0;6] >3 >2.2
Corruption Contin. [0;6] >2.5 No diff.
Bureaucratic Contin. [0;4] >1.5 No diff.
Composite index Contin. [12;84.7] No diff. No diff.
Neighbors
Neighbors’ conflict (any) Dummy [0;1] 0 0
Share trade neighbors RTA Contin. [0;0.8] >0.35 >0.08
Share trade neighbors Contin. [0;0.9] No diff. >0.24
Price transmission
NRA (output + input) 3 groups [−0.9;1.4] < or > 2nd tercile No diff.
NRA (output) 3 groups [−0.9;3.5] < or > 2nd tercile No diff.
Note: The column range indicates the range of values taken by each variable in our sample; the column type groups the
variables into three types, i.e. continuous, dummy (0 or 1), and 3 groups (the variable is split into three mutually exclusive
continuous groups of values). The latter category is included for those variables for which the marginal effect of Px on conflict
appeared to be non linear. The other columns report the values of the interaction for which the marginal effect of the
relevant trade variable becomes not significantly different from zero (at the 10 percent level); “No diff.” indicates either that the
marginal effect is always or never statistically different from zero across the distribution of the interactions’ values. The
variables are defined in table A.3.
from zero once the accountability index rises above 3. Similarly, the upward slop-
ing curve in the lower panel indicates that the effect of the price index on conflict
becomes larger for observations characterized by at least one conflict in contigu-
ous countries. In fact, the effect of the price index (while positive) is not signifi-
cantly different from zero for countries in years that they have peaceful neighbors.
These results can also be used to determine the relative importance of the condi-
tioning factors in explaining the resilience to trade-related changes.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
70 Conditions That Affect the Impact of Trade Shocks on Conflict
Figure 2.1 Marginal Effects of Px Across the Range of Interaction Variables’ Values
a. Degree of accountability (1–6)
0.10
Marginal effect of price shock on conflict
0.05
0
−0.05
0 1 2 3 4 5 6
b. Conflict in contiguous countries
0.06
Marginal effect of price shock on conflict
0.04
0.02
0
t
ct
lic
li
nf
nf
Co
co
No
Note: The bands indicate the confidence around the estimate line and points represent the 90 percent confidence interval.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 71
The importance of this cross-country heterogeneity is also highlighted by
the different association between swings in oil exports and conflict in the Republic
of South Sudan and in Timor-Leste, two countries with different degrees of
accountability and transparency in oil revenues management (box 1.3 in chapter 1).
These marginal effects of the mediating factors could be used also to predict
the risk of conflict from an increase in the value or a new discovery of point-
source commodities in a specific country. To illustrate it let us take the example
of Lebanon, where recent 3D seismologic mapping of the country’s offshore
have revealed the likely presence of sizeable hydrocarbon resources. According
to some estimates this can be worth several times Lebanon’s annual GDP (World
Bank 2014). We can apply the estimated coefficients of the interaction terms
computed above to the values of the relevant variables for Lebanon to calculate
the additional risk of conflict that this discovery would entail.
The country has a relatively high level of economic inequality and of religious
fractionalization, and it is experiencing a low-intensity civil conflict.13 These
characteristics increase the risk of conflict induced by hydrocarbons discovery in
Lebanon relative to the average country. In table 2.2 we quantify this increased
risk due to all the factors that have a statistically significant marginal effect as
computed in the regressions underlying table 2.1. On the basis of these marginal
effects we know, for example, that for a country with the average level of reli-
gious fractionalization (i.e., 0.35), a standard deviation increase in point-source
commodity exports raises the risk of conflict by 2.5 percent. At the level of
fractionalization of Lebanon (i.e., 0.79) the increased risk is more than double at
5.7 percent. A similar story applies to the two other main factors—economic
inequality and the presence of a recent conflict. On the other hand, other factors,
such as the level of democracy, the level of accountability, the presence of con-
flict in neighboring countries, and the share of trade with neighbors, yield smaller
differences between Lebanon and the average country.
Table 2.2 Lebanon Has a Higher Risk of Conflict from Hydrocarbons Exports than the
Average Country
Mean Lebanon Average effect LBN effect Significance
Inequality (Gini) 45.44 55.33 2.1% 4.2% Yes
Religious fractionalization 0.35 0.79 2.5% 5.7% Yes
Recent conflict 0.28 1 2.2% 4.1% Yes
Level of democracy −1.21 7.00 1.9% 1.9% No
Accountability 3.24 5.00 2.4% 0.9% No
Neighbors in conflict 0.52 Yes 2.0% 2.3% No
Share trade neighbors 0.12 0.03 2.0% 2.1% No
Source: Computations in the table are based on the following sources for the Mean and Lebanon values shown: Standardized
World Income Inequality Database for inequality data; Alesina et al. (2003) for religious fractionalization; UCDP/PRIO (for
mean) and ICRG for Lebanon (see endnote 13 for more details) for recent conflict; polity 2 data for level of democracy; ICRG
for accountability; UCDP/Prio (for mean) and ICRG for neighboring conflict; COMTRADE for share of trade with neighbors.
Note: Mean is the mean value of the variable across the entire sample over which the marginal effects of the export price
index are computed in the regressions underlying table 2.1; Lebanon is the value of the variable for Lebanon (latest available);
Average effect is the increased probability of conflict due to a standard deviation increase in the point-source commodity
export index measured for the mean value of the variable; LBN effect is the same increased probability measured at the value
of the variable for Lebanon; and Significance indicates whether the marginal effect of the variable is significant.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
72 Conditions That Affect the Impact of Trade Shocks on Conflict
Testing for the Importance of Heterogeneity in the Nigerian Conflict
Chapter 1 found that producer prices, consumer prices, and oil prices appear to
be important determinants of conflict across Nigerian states, on average.
However, states’ vulnerability to a change in a price index may vary. In particular,
the presence of deep-seated roots of conflict is usually a necessary condition for
any change in real incomes, including that generated by changes in trade prices,
to have an impact on violence. A better understanding of what conditions matter
in this respect could help direct policy interventions to address the vulnerability
to increases in conflict from trade changes.
We consider how political issues, the conditions affecting the transmission of
prices, and grievances may affect the impact of changes in trade on conflict across
Nigerian states. The procedure is to multiply each factor times the price indices
for production, consumption, and (usually) oil. The role of politics is represented
by a dummy variable for an election year. As in other African countries (Ksoll,
Macchiavello, and Morjaria 2010), elections are perilous times in Nigeria.
The degree of transmission of international prices to the domestic markets is
measured by the distance to Lagos for each state. Internal trade costs are high in
Nigeria (Atkin and Donaldson 2014). As Lagos is the largest market and the
international gateway for the country’s trade, distance to Lagos could affect the
extent to which changes in international prices translate at the local level. We do
not follow this procedure for the oil index, as price transmission should not be
an issue for oil.
We use variables representing ethnic divisions, economic inequality, and the
level of past conflict intensity to capture the extent to which grievances affect
the impact of price changes on conflict. Among the factors affecting grievances,
ethnic divisions feature prominently in African conflicts, and in Nigeria in par-
ticular (NNoli 2003). Ethnic divisions are represented by the three dummy
variables used in chapter 1 as controls: whether the president has the same
ethnicity as the dominant group, whether there are more than two significant
ethnic minorities, and whether there are multiple dominant ethnic groups in
the state.
We use poverty measures, including the poverty gap, the poverty headcount,
the gini index of inequality, and the unemployment rate (all computed at the
beginning of the period in 2003–04) to reflect the potential for economic
conditions contributing to grievances.
Finally, we assume that the level of past conflict will affect the extent to which
price indices affect current conflicts. The level of past conflict is an important
predictor of future violence by generating grievances (World Bank 2011), a
finding that has been confirmed in this analysis as well.
We conduct several repetitions of our preferred specification for the Nigeria
model. In each repetition, we add three (only two for the price transmission
variable) new terms, which are each of the three price indices multiplied by one
of the mediating factors listed above. Further, we repeat this exercise with the
different measures of conflict as dependent variables. The results of this analysis
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 73
are presented in table B.13, which reports only the coefficient of the additions to
the separate regressions, along with their degree of significance.
Various findings emerge. First the magnitude of the price effects on conflict is
amplified in election years, especially for consumption and oil. Second, the
effects of the production price indices on conflict events and on battle events are
reduced the farther one moves away from Lagos. Third, various factors related to
grievances significantly magnify the conflict-inducing effect of a rise in the price
of oil. That is particularly the case for our measurements of ethnic factors and
economic inequality.
On the other hand, and more surprisingly, most past conflict events do not
magnify the effects of prices on conflict. Past protests are an exception (i.e. the
impact of prices on the incidence of protests is higher in states with a past history
of protests). In addition, neither unemployment nor poverty appears to affect the
impact of prices on conflict.
Heterogeneity in the Israeli-Palestinian Conflict
In chapter 1 we show that changes in exports exert a sizable effect on conflict
across Palestinian localities, on average. However, these effects may differ across
localities, depending on conditioning factors that could shape the importance of
the opportunity cost channel of conflict. The conditions we consider include the
share of refugees, the unemployment rate, the share of males, past fatalities, the
level of education, and the share of young adult males in the population. As
above, we multiply each conditioning factor times the export variable, and
include this interaction term in the regression equation, along with the export
variable and the controls mentioned in chapter 1.
Changes in exports appear to affect conflict more in localities with a higher
share of refugees and a higher unemployment rate, pointing to the importance of
grievances in exacerbating the impact of adverse trade changes on conflict.14 In
particular, refugee status is connected with the displacement by Palestinians as a
result of the creation of the state of Israel in 1948, which is often associated with
Palestinian grievances vis-à-vis Israel. On the other hand, past fatalities, presum-
ably an important indicator of grievances, does not significantly affect the impact
of export changes on conflict.15 A lower share of males (but not of young adult
males) in the population is associated with a greater impact of export changes on
conflict. A lower share of males in the local population may mean that a higher
share of males related to local residents are in Israeli jails, so that this finding also
reflects the impact of grievances. Perhaps surprisingly, the level of education is
not associated with significantly different effects of changes in exports on conflict
intensity.
Notes
1. The former study captures ethnic divisions via the ethno-linguistic fractionalization
index, while the latter focuses on ethnic polarization from Montalvo and Reynal-
Querol (2005).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
74 Conditions That Affect the Impact of Trade Shocks on Conflict
2. Ethnic polarization refers to the extent to which most of the population in a country
or region is divided between two ethnic or linguistic groups. Ethnic fractionalization
refers to the probability that two randomly selected individuals will belong to differ-
ent ethnic or linguistic groups.
3. These results support the argument by Fearon (2005) that slowed economic devel-
opment may widen preexisting ethnic rifts in countries with highly fractionalized
societies.
4. In particular, Hull and Imai (2013) use a measure of instability (conflict onset) more
relevant to the arguments discussed here.
5. This distinction is defined according to Persson and Tabellini (2003).
In particular, these subsets of countries include nondemocracies, autocracies,
6.
regimes with low executive constraints, low-income countries, highly unequal coun-
tries, countries with high ethnic polarization, and countries in sub-Saharan Africa.
7. These price stabilization mechanisms may point toward a second order effect of inter-
national price changes on conflict, i.e. via price variability (higher price variability may
lead to greater unrest due to the higher variability of incomes). There has been no
empirical evidence so far on the extent to which such variability may affect
violence.
8. The dependent variable includes all conflicts that result in at least 25 deaths per year.
We are mainly interested in modeling the cross-country heterogeneity in the effect
of trade-related changes on conflict. Therefore, we do not interact the conditioning
variables with the share of trade with neighbors’ variables (results for this separate
set of regressions are available upon request).
9. Compare the significance of the coefficients in table B.8 vis-à-vis the joint coefficients
in table B.1, column 2 and table B.2, column 1.
10. For example, regions within federal countries tend to have more influence over the
allocation of natural resource revenues in their territory (Brosio and Singh 2014), which
our empirical results suggest is a potentially contentious issue in fragile countries.
11. However, unlike Besley and Persson (2008), these changes in prices do not exert any
differential impact on conflict in parliamentary democracies versus other forms of
governments.
12. Each observation is one country example for one year.
13. This conflict categorization follows from the International Country Risk Guide (ICGR)
data as the UDCP/PRIO armed conflict database. In the first seven months of 2014,
ICRG assigns an average value of 2 to Lebanon in the “civil war” category (on a scale
of 1 to 4), placing the country in the top decile of that category. The current low-
intensity conflict in Lebanon would also be in line with the definition of UCDP/PRIO
of “a contested incompatibility that concerns government and/or territory where the
use of armed force between two parties, of which at least one is the government of a
state, results in at least 25 battle-related deaths.”
14. The results are presented in table B.22, where we include one interaction term at a
time along with all the usual controls (columns 1–7) and then all of them together
(column 8).
15. Note that chapter 1 shows that past violence did have a significant relationship with
violence during the Second Intifada. Here we are reporting only the impact of past
violence on the relationship between trade changes and conflict, not the direct impact
on conflict.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 75
References
Acemoglu, D., and J. Robinson. 2006. The Economic Origins of Dictatorship and Democracy.
Cambridge, UK: Cambridge University Press.
Alesina, A., A. Dzevleeschauwer, R. Wacziarg, S. Kurlat, and W. Easterly. 2003.
“Fractionalization.” Journal of Economic Growth 8 (2): 155–94.
Anderson, K., M. Kurzweil, W. Martin, D. Sandri, and E. Valenzuela. 2008. “Measuring
Distortions to Agricultural Incentives, Revisited.” Policy Research Working Paper
4612, World Bank.
Angrist, J., and A. D. Kugler. 2008. “Rural Windfall or a New Resource Curse? Coca,
Income, and Civil Conflict in Colombia.” The Review of Economics and Statistics 90 (2):
191–215.
Atkin, D., and D. Donaldson. 2014. “Who’s Getting Globalized? The Size and Implications
of Intranational Trade Costs.” Mimeo.
Bazzi, S., and C. Blattman. 2014. “Economic Shocks and Conflict: Evidence from
Commodity Prices.” American Economic Journal: Macroeconomics.
Bellows, J., and E. Miguel. 2009. “War and Local Collective Action in Sierra Leone.”
Journal of Public Economics 93 (11–12): 1144–57.
Berman, N., and M. Couttenier. 2014. “External Shocks, Internal Shots: The Geography of
Civil Conflicts.” CEPR Discussion Paper 9895.
Besley, T., and T. Persson. 2008. “The Incidence of Civil War: Theory and Evidence.”
Mimeo.
Bhagwati, J. N. 1978. Foreign Trade Regimes and Economic Development: Anatomy and
Consequences of Exchange Control Regimes. Cambridge, MA: Ballinger.
Blair, R., C. Blattman, and A. Hartman. 2012. “Patterns of Conflict and Cooperation in
Liberia (Part 2): Prospects for Conflict Forecasting and Early Warning.” Yale University
and IPA, Mimeo.
Blattman, C., and E. Miguel. 2010. “Civil War.” Journal of Economic Literature 48: 3–57.
Blimes, R. J. 2006. “The Indirect Effect of Ethnic Heterogeneity on the Likelihood of Civil
War Onset.” Journal of Conflict Resolution 50, 4.
Brosio, G., and R. Singh. 2014. “Revenue Sharing of Natural Resources: A Review of
International Practices.” World Bank, Mimeo.
Bruekner and Gradstein. 2014. “Income Growth, Ethnic Polarization, and Political
Risk.” Journal of Comparative Economics, forthcoming.
Buhaug, H., L.-E. Cederman, and K. S. Gleditsch. 2011. “Square Pegs in Round Holes:
Inequalities, Grievances, and Civil War.” Mimeo.
Buhaug, H., and K. S. Gleditsch. 2008. “Contagion or Confusion? Why Conflicts Cluster
in Space.” International Studies Quarterly 52 (2): 215–33.
Carter, B. L., and R. Bates. 2012. “Public Policy, Price Shocks, and Civil War in Developing
Countries.” Georgetown University, Mimeo.
Caselli, F., and A. Tesei. 2013. “Resource Windfalls, Political Regimes, and Political
Stability.” Mimeo.
Cederman, L., L. Girardin, and K. S. Gleditsch. 2009. “Ethnonationalist Triads: Assessing
the Influence of Kin Groups on Civil Wars.” World Politics 61 (3): 403–37.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
76 Conditions That Affect the Impact of Trade Shocks on Conflict
Collier, P., and A. Hoffler. 2004. “Greed and Grievance in Civil War.” Oxford Economic
Papers 56 (4): 563–96.
Corden, W. M. 1971. The Theory of Protection. Oxford: Clarendon Press.
Easterly, W., and R. Levine. 1997. “Africa’s Growth Tragedy: Policies and Ethnic Divisions.”
Quarterly Journal of Economics 112 (4): 1203–50.
Fearon, J. 2005. “Primary Commodity Exports and Civil War.” Journal of Conflict Resolution
49 (4): 483–507.
Fearon, J., and D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American Political
Science Review 97 (1): 75–90.
Gleditsch, K. S. 2007. “Transnational Dimensions of Civil War.” Journal of Peace Research
44: 293–309.
Gomes, J. F. 2011. “The Political Economy of the Maoist Conflict in India: An Empirical
Analysis.” Mimeo.
Hidalgo, F. D., S. Naidu, S. Nichter, and N. Richardson. 2010. “Economic Determinants of
Land Invasions.” The Review of Economics and Statistics 92 (3): 505–23.
Hull, P., and M. Imai. 2013. “Economic Shocks and Civil Conflict: Evidence from Foreign
Interest Rate Movements.” Journal of Development Economics 103: 77–89.
Ianchovichina, E., J. Loening, and C. Wood. 2012. “How Vulnerable Are Arab Countries
to Global Food Price Shocks?” Policy Research Working Paper 6018, World Bank.
International Crisis Group. 2010. “Dangerous Little Stones: Diamonds in the Central
African Republic.” Africa Report 167.
Ivanic, M., and W. Martin. 2013. “World Food Price Rises and the Poor 2006–12: A Slow
Food Price Crisis?” World Bank, Mimeo.
Ksoll, C., R. Macchiavello, and A. Morjaria. 2010. “The Effect of Ethnic Violence of an
Export-Oriented Industry.” Mimeo.
Losch, B. 2002. “Global Restructuring and Liberalization: Côte d’Ivoire and the End of the
International Cocoa Market.” Journal of Agrarian Change 2 (2): 206–27.
Maystadt, J.-F. 2008. “Does Inequality Make Us Rebel? A Revisited Theoretical Model
Applied to South Mexico.” Households in Conflict Network Working Paper 41.
Miaari, S., A. Zussman, and N. Zussman. 2014. “Employment Restrictions and Political
Violence in the Israeli-Palestinian Conflict.” Journal of Economic Behavior and
Organization 101 (May): 24–44.
Miguel, E., S. Satyanath, and E. Sergenti. 2004. “Economic Shocks and Civil Conflict: An
Instrumental Variables Approach.” Journal of Political Economy 112 (4): 725–53.
Minot, N. 2011. “Transmission of World Food Price Changes to Markets in Sub-Saharan
Africa.” IFPRI Discussion Paper 1059.
Montalvo, J. G., and M. Reynal-Querol. 2005. “Ethnic Polarization, Potential Conflict, and
Civil Wars.” American Economic Review 95 (3).
Nepal, M., B. K. Bohara, and K. Gawande. 2011. “More Inequality, More Killings: The
Maoist Insurgency in Nepal.” American Journal of Political Science 55: 886–906.
NNoli, O. 2003. “Ethnic Violence in Nigeria: An Historical Perspective.” Indiana University,
Bloomington. http://www.indiana.edu/~workshop/papers/nnoli_021003.pdf.
Persson, T., and G. Tabellini. 2003. The Economic Effects of Constitutions. Boston, MA:
MIT Press.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Conditions That Affect the Impact of Trade Shocks on Conflict 77
Robinson, J. A., D. Acemoglu, and S. Johnson. 2003. “An African Success Story: Botswana.”
In In Search of Prosperity: Analytic Narratives on Economic Growth, edited by D. Rodrik,
80–119. Princeton, NJ: Princeton University Press.
World Bank. 2011. World Development Report 2011: Conflict, Security and Development.
Washington, DC: World Bank.
———. 2014. “Lebanon Economic Monitor, Spring 2014.” World Bank, Washington, DC
and Beirut.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Chapter 3
How Trade Policy Could Ease
Tensions in Fragile Countries
The first two chapters showed that international trade matters for civil conflict.1
This general finding confirms growing evidence gathered over the past decade that
changes in prices and incomes are important determinants of civil conflict
(Blattman and Miguel 2010; Miguel, Satyanath, and Sergenti 2004). It also
complements the evidence on the importance of trade for inter-state conflicts
(Martin, Mayer, and Thoenig 2012).
Part of the reason why trade is important for civil conflicts is that it is the
largest external flow in fragile countries, which are the ones most exposed to the
risk of conflict. But this report shows that its importance extends beyond that.
For example, trade enables countries to monetize their natural resources, which
in fragile countries often represent the bulk of economic resources. This affects
the potential for conflict by raising the economic value involved in gaining hold
of these resources through violence, and also by providing funds to the govern-
ment which can be used to repress or buy off rebel groups.
Trade also affects real incomes and thus individuals’ opportunity costs of
engaging in violence. In many fragile countries, open trade regimes have provided
essential consumption goods to households, especially food. Moreover, open
trade regimes increase real incomes by improving efficiency and providing
opportunities for exports. On the other hand, an open trade regime can also
quickly transmit the effects of international price swings onto households. In the
context of fragile countries, sharp, adverse changes in the prices households face
on the goods they consume or produce can increase the potential for conflict.
The natural question that follows is how to use these findings to inform the
policy debate, and trade policy in particular. This chapter attempts to do this by
focusing on a narrow but arguably fundamental policy objective in fragile coun-
tries, i.e. preventing civil conflict or quelling existing ones. As argued by scholars
such as Collier (2008) and del Castillo (2011), this should be the guiding prin-
ciple of engagement in conflict affected and post-conflict countries. This is not
the first time that the World Bank has explored how policies affect conflict.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 79
80 How Trade Policy Could Ease Tensions in Fragile Countries
Illustrious antecedents include Collier et al. (2003) and more recently the 2011
World Development Report on conflict, security, and development (World Bank
2011). However, this is the first policy analysis that explicitly focuses on trade.
The policy ideas discussed are not only directed to domestic policy makers.
Whenever possible, we consider how the international community, including the
World Bank, can assist countries to formulate policies that take into account the
risk of conflict linked to trade. This is important also in light of IEG (2013b),
which stresses the need for the World Bank to better tailor its assistance to
fragility and conflict contexts.2
As documented in this report, the drivers of conflict differ across countries.
Moreover, governments, particularly in fragile countries, have different degrees of
capacity. In this sense, the options discussed below have to be interpreted more
as general directions that need to be tailored to the local situation, rather than as
policy prescriptions. Similarly, this policy discussion does not attempt to evaluate
the feasibility of the various policy options, which would require a country-
specific political economy analysis. The absence of such an analysis does not,
however, detract from the value of the general directions d iscussed below.
Trade Policies in Fragile Countries Must Take into Account the
Implications for Conflict
Trade policies in fragile countries need to be compatible with the objective of
supporting political stability. This requires an understanding of the relationship
between trade and conflict, which our analysis suggests can be adequately devel-
oped only at the level of the country or below. That is, the differences in this
relationship among countries are so great that cross-country analysis is an unreli-
able guide to country-specific advice. For example, our analysis suggests that
point-source commodities, like oil or gas, are the exports that most frequently
become drivers of conflict. However, this is true to a different extent in different
contexts, and even across time. In Nigeria, the estimated relationship between oil
production and conflict was positive for most of the period studied, but then
turned not significant for the period following the amnesty agreement with rebel
groups. By contrast, in the West Bank and Gaza changes in export revenues were
an important driver of conflict, despite the absence of point-source commodity
exports.
A growing literature, including this report, shows the data required to analyze
the relationship between trade and conflict at the country level. These data are
essential to develop a framework to monitor the implications for the risk of
conflict of both trade policies and trade-related changes. Such a framework
should fulfill at least two main objectives:
a. It should identify the changes in trade that would matter most for the coun-
try’s stability. For example, it would identify the subset of traded goods
(and services) that are most relevant for the economy and the types of
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Policy Could Ease Tensions in Fragile Countries 81
economic changes (e.g. international price swings, changes in trade policies in
country concerned or its trading partners) that affect the domestic prices
the
of these goods (and services).
b. The framework should also help assess the likely distribution of gains and
losses across different groups within the country as a result of the change in
trade. This assessment, along with a political economy analysis, could help
policy makers understand to what extent the losers may be willing and capable
of destabilizing the country (or areas of it) following an adverse change in
trade flows.
A monitoring framework of this type would be particularly important for
those countries most exposed to the risk of conflict due to changes in trade
flows. These countries are not necessarily limited to those in the World Bank’s
FCS list; nor is the inclusion in the FCS list sufficient to be considered exposed
to this risk. As the IEG (2013b) argues, the World Bank’s definition of fragility
does not always adequately take into account indicators of conflict, violence, and
political risk.
In this respect, an empirical analysis like in chapter 2 can identify the
conditions that increase the conflict risk connected to trade changes. Important
conditions that affect the relationship between trade changes and conflict include
the presence of conflict in neighboring countries, ethnic divisions, a recent history
of past conflicts, and weak and/or corrupt government institutions. As many of
the countries fitting this profile lack adequate human and financial resources,
technical assistance from organizations like the World Bank may be important to
putting such a monitoring framework in place.
Manage Receipts from Commodity Exports in a Conflict-Sensitive Way
Previous evidence across and within countries (Bellows and Miguel 2009; Dube
and Vargas 2013; Lin and Michaels 2013; Maystadt et al. 2014) as well as the
analysis in the previous chapters, suggest that exports of point-source commodi-
ties substantially raise the risk of conflict. In many countries, these exports rep-
resent the bulk of the state’s revenues or of the local area’s resources. Thus their
value is often the prize which the different parties fight over.
The way in which point-source commodity export revenues are managed is
thus a key factor in determining their effect on conflict risk.3 According to the
evidence in chapter 2, that is especially the case in countries located in unstable
regions, with a recent history of conflict, and with weak governance. In particular,
the evidence suggests two general principles that should reduce the conflict
inducing effect of point-source exports.
First, the discretion of the central or local governments in managing these
resource revenues should not be absolute. Effective limitations on the spending
of revenues by government, and procedures to ensure the transparency of such
expenditures, are essential for two reasons. First, they can reduce the potential
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
82 How Trade Policy Could Ease Tensions in Fragile Countries
for governments to discriminate against some groups, and thus limit the resent-
ments and disputes that can result in civil conflict. In addition, if it is possible,
for example through international arrangements, to reduce government access to
resource revenues, then the incentive to rebel in order to control these revenues
is reduced. This perspective is consistent with the analysis in chapter 2, which
shows that increases in the prices of point-source commodity exports are
associated with a larger increase in conflict where government accountability is
low. Often windfall increases in oil or minerals prices are captured by the govern-
ment, state-owned entities, or national-resource companies, which have total
discretion on how to spend such resources. This appears to have been the case in
the Republic of South Sudan, where the renewal of oil exports to Sudan in April
2013 was followed by increased political violence, which eventually triggered the
current civil war. On the other hand, the development of a transparent and
accountable oil revenue management system was associated with a peaceful
post-independence political transition in oil dependent Timor-Leste (see
box 1.3).
Second, transferring part of the revenues from point-source commodities to
the producing areas can reduce the risk of conflict. This can be deduced from the
results of recent studies. First, federal systems, which are typically more inclined
to reward natural resources–producing areas (Brosio and Singh 2014), are less
subject to conflict risk from trade-related changes. Second, conflicts over extrac-
tive resources occur overwhelmingly in producing areas, as also confirmed by
Dube and Vargas (2013), Berman et al. (2014), and Maystadt et al. (2014). Third,
transferring resources to local militant groups helped reduce the intensity of the
conflict in the oil-producing Niger Delta, at least in the short run. Producing
areas can be rewarded by channeling resources through individuals, organiza-
tions, or subnational governments. Besides conflict prevention, this transfer is also
justified from an economic standpoint as compensation for the environmental
degradation and socioeconomic evils generated locally by natural resource
extraction (Brosio and Singh 2014).
These principles may not apply when nonstate actors (e.g. local rebel groups)
rather than the state control the extraction and the sale of the natural resources.
In those instances, the revenues are usually employed to fund the fighting. Thus,
other strategies that are not completely dependent on the government may have
to be implemented to break the link between increased revenues and conflict
(see below).
Using these principles in formulating policies to manage point-source
commodity revenues requires adapting them to the local context. Examples of
policy options in line with these principles include the following:
a. Increase the transparency of the flow of revenues from extractive commodities.
Domestic policies could help achieve that objective, for example by centraliz-
ing the collection of the revenues into a single account under the authority of a
Ministry (typically the Ministry of Finance) as suggested by Haysome and Kane
(2009). In addition, a number of international initiatives help governments to
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Policy Could Ease Tensions in Fragile Countries 83
enhance transparency. The Extractive Industries Transfer Initiative (EITI),
a global coalition of governments, companies, and civil society helping to ensure
the disclosure of taxes and other payments made by oil, gas, and mining com-
panies to governments, is perhaps best known in this respect. Recent years have
seen a flurry of other global initiatives with similar aims, such as the Publish
What You Pay (PWYP) network, the Kimberley Process (KP) on the diamond
sector, and the Conflict Free Gold Standard on the gold sector. By enhancing
the transparency around the flow of these revenues, these initiatives can also
make it more difficult for nonstate actors to use the sale of extractive com-
modities to fund their own fighting.
b. Transfer a percentage of extractive commodities’ rents to citizens. This option—
perhaps best known as Direct Dividend Payments (DDPs)—has been gaining
momentum in recent years among scholars (see Gillies 2010; Moss 2011;
Devarajan and Giugale 2012; and Devarajan et al. 2013). It has also been
implemented especially at the subnational states, notably in the states of Alaska
(USA) and Alberta (Canada). It has not yet gained much traction in develop-
ing countries but its proponents argue that DDPs can maximize the direct
effects of resource revenues on poverty reduction and improve the account-
ability and the efficiency of public spending through enhanced scrutiny by
citizens. For example, Devarajan et al. (2013) estimate that in Equatorial
Guinea the per person distribution of just 10 percent of resource revenues
through a DDP would be one and a half times larger than the money the aver-
age poor person needs to exit poverty. According to their proponents, DDPs
would also help establish a stronger fiscal contract between the citizens and
their government, by increasing citizens’ incentives to oversee how the govern-
ment manages natural resources’ revenues, and possibly by taxing back some
of the DDPs.
While not initially proposed with conflict in mind, DDPs have two poten-
tially important benefits in the quest for conflict prevention in resource rich
countries. First, if citizens benefit directly from the revenues, their incentive to
oppose any attempt by government to regain full control over resource reve-
nues (e.g. by eliminating the DDPs) would be higher. Similarly, citizens bene-
fiting from DDPs may oppose attempts by nonstate actors to gain control of
extractive resources. Second, as argued by Devarajan and Giugale (2013),
greater citizens’ involvement could strengthen oversight of the spending of
resource revenues in general. That could limit the state’s ability to capture the
natural resource rents and thus the incentives for fighting to appropriate the
resources.
This option could also be combined with the principle of rewarding areas
where resources are extracted. In that case the amount of the direct transfer
to the citizens may be differentiated on the basis of their location of resi-
dence, with citizens from producing areas receiving a higher amount than the
others.4
c. Create a financial vehicle outside the direct control of the government to channel
part of the resource revenues. This is another way to limit the ability of the
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
84 How Trade Policy Could Ease Tensions in Fragile Countries
government to capture point-source commodity rents. The most common of
such vehicles is the Sovereign Wealth Fund (SWF), which have been success-
fully established by various resource rich countries, such as Norway,
Timor-Leste, Brazil, Mexico, Chile, and Colombia. Some of these funds have
helped improve the long-term returns of extractive revenues, smooth the busi-
ness cycle, and enhance transparency in the spending of revenues from natural
resources. The high quality of Timor-Leste’s SWF has also been associated
with a peaceful post-independence transition (box 1.3).
However, the record of SWFs in furthering development is mixed. Several
SWFs operate with no disclosure, limiting their accountability and increasing
the risk of corruption (Revenue Watch Institute 2013). In addition, saving
resource revenues through an SWF may not be the optimal strategy in many
developing countries, especially those with high levels of poverty where these
revenues may be more profitably channeled toward investments and con-
sumption. IEG (2011) points out that in Timor-Leste itself the management of
oil revenues—in line with the World Bank’s advice—favored high levels of
savings and placed too little emphasis on short-term interventions that would
have yielded immediate benefits to the population.
d. Use resource revenues to placate rebel groups in the producing areas. The evidence
in chapter 1 suggests that the use of oil revenues for the DDRR program for the
militant groups appears to have been effective in reducing violence in the Niger
Delta, at least in the short run. The evidence in favor of this strategy remains
thin, and even in Nigeria it has been criticized for failing to treat the root causes
of conflict, and for providing incentives toward “warlordism” (Sayne 2013).
Indeed, without parallel policies addressing the deep determinants of conflict
(according to Joab-Peterside et al. [2012], youth unemployment, ineffective
and corrupt public institutions, human and ecological insecurity), civil unrest is
likely to resume and the program risks ending up promoting “warlordism.”
However, the evidence supporting the success of the alternative—usually the
preferred option by autocracies (Cotet and Tsui 2013)—of using extractive
commodity revenues to increase armed repression is also pretty limited. For
example, in Nigeria heightened military repression in the oil-producing areas
before the 2009 agreement was not associated with a reduction in violence.
In fact, the military repression may have even contributed to the escalation of
violence (Asuni 2009; Rosenau et al. 2009). Similarly, increased foreign military
assistance in Colombia was ineffective in reducing guerrilla attacks and even
increased paramilitary attacks (Dube and Naidu 2013). The decision as to
whether to attempt to reach agreement with, or to fight, militant groups is
ultimately a political one. Each of these strategies needs to be accompanied by
attempts to address the root causes of the conflicts for a sustainable peace.
Again, the international community could be instrumental in helping
ountries in the identification and the implementation of the most effective
c
policy options in some of these areas. First, it could provide support to the gov-
ernment in developing cost-benefit analyses of the alternative options. Second, it
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Policy Could Ease Tensions in Fragile Countries 85
could offer technical assistance in the actual implementation of the policies, as
the World Bank did in the establishment of Timor-Leste’s petroleum fund. Third,
it could exert pressure on governments when the most effective option for con-
flict prevention is not in line with the policy makers’ interests.
Protect Producers, Consumers, and Workers from Adverse Trade Shocks
The evidence from Nigeria and the West Bank and Gaza presented in chapter 1
confirms that adverse changes in trade flows can increase conflict intensity. By
reducing real incomes of consumers, producers, and workers, such changes
reduced the opportunity cost of fighting. That is also the evidence emerging from
other within-country studies (e.g. Berman and Couttenier 2014; Dube and Vargas
2013; Maystadt and Ecker 2014). On the other hand, the cross-country evidence
is inconclusive in this respect (see for example the evidence in chapter 1 and in
Bazzi and Blattman 2014).
Where adverse trade changes increase conflict via the opportunity cost
channel, protecting the real incomes of consumers, producers, and workers
should be a priority. A variety of policy options exist to do this, for example
targeted transfers, public works programs, price subsidies, and temporary trade
insulation.
All of these policies have strengths and weaknesses, and the identification of
the right policy tool in each context is beyond the scope of this discussion.
However, the evidence suggests that targeted transfers appear to be particularly
useful in counteracting the losses by households as a result of an adverse trade
change (Anderson, Ivanic, and Martin 2013; Attanasio et al. 2013). Furthermore,
it is important that the policies adopted avoid, to the extent possible, impeding
adjustment to the changes in relative prices caused by changes in trade. From this
perspective, targeted transfers (as well as public works programs) may be prefer-
able to price subsidies and temporary trade insulation.5 In general, however,
effective safety nets are difficult to develop. Providing protection to all possible
losers is costly, and proper targeting is inherently difficult, especially in fragile
countries.
The international community could help fund programs that fragile countries
use to protect their citizens from adverse trade changes. One relevant antecedent
in this respect is the Global Food Crisis Response Program (GFRP), which was
set up by the World Bank to help countries address the immediate needs arising
from the international food price hike in 2007–08. The GFRP supported
35 countries, with Sub-Saharan Africa accounting for about 60 percent of fund-
ing. Most of the short-term assistance was to the agricultural sector (via input
subsidy and distribution operations to increase food supply) and to social safety
nets (in-kind transfers and public works programs). A recent independent evalu-
ation considered the program effective in helping poor countries deal with the
immediate consequence of the price crisis (IEG 2013a).6 It may be useful to
think about a similar international shock absorption fund to help fragile countries
deal with adverse trade shocks, including international price swings.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
86 How Trade Policy Could Ease Tensions in Fragile Countries
Promote Trade with Neighbors
The previous analysis suggests that promoting trade with a country’s neighbors
reduces the risk of conflict, or at least its intensity and duration. This trade is
particularly effective in preventing conflict when it occurs under an RTA. While
this result is novel in the empirical literature, it is consistent with the idea that a
high volume of trade between two neighbors A and B increases the costs to A of
a conflict in B, thus reducing the likelihood that A would intervene to foment
civil conflict in B (and vice versa). It is also in line with the idea that trade may
raise the level of trust between the peoples of neighboring countries (Rohner,
Thoenig, and Zilibotti 2013).
Trade policy as well as trade facilitation can help foster these trade relations
among neighbors. Reducing tariff and nontariff barriers is a necessary step in
this direction. There is abundant evidence of the existence of high policy barri-
ers to trade, especially between fragile countries. Such barriers even constrain
trade in basic food staples between sub-Saharan African neighbors (World Bank
2012). Similarly, policy barriers to trade between neighbors in other major
conflict-affected regions, the Middle East and North Africa and South Asia, are
particularly high.
While necessary, efficient trade policy is not sufficient to stimulate trade
between neighboring fragile countries, most of which are marred by particularly
poor transit, logistics, and transport infrastructure systems (see the analysis in
chapter 1). Improving the whole trade facilitation system is therefore crucial to
increase trade between neighbors in conflict prone areas. That is part of the
objective, for example, of the large program of assistance of the World Bank to
the Great Lakes region, where increasing cross-border trade is considered as con-
tributing to regional stability. Similarly, the African Union (AU) established the
African Union Border Program in 2007 with a view to preventing conflict by
promoting cross-border cooperation and trade (World Bank 2011). Unfortunately,
no evidence is available so far on the effectiveness of this initiative for conflict
prevention.
Support Labor-Intensive Exports
The evidence from the Israeli-Palestinian conflict presented in this report sug-
gests that the main channel through which exports (of nonpoint-source com-
modities) affect conflict is through employment. This is consistent with the
finding that changes in employment in Palestine had an important effect on
conflict intensity (Miaari, Zussman, and Zussman 2014) as well as with recent
theoretical contributions (Dal Bó and Dal Bó 2011, 2012).7
This finding suggests that promoting labor-intensive export sectors in fragile
countries may help reduce conflict intensity and risk. There are two main
(mutually reinforcing) ways to promote labor-intensive exports. First, a country
can improve its market access in labor-intensive sectors in its main trading partners.
This access is essentially dependent on the trade policy of a country’s trading part-
ners. As far as tariffs are concerned, the data of Carpenter and Lendle (2011) show
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Policy Could Ease Tensions in Fragile Countries 87
that there is still much room to improve fragile countries’ preferential access to
the main markets.8 Among the 25 developing countries with the largest prefer-
ence margins for their exports in the main importers, only three are FCS (Haiti,
Afghanistan, and Nepal). Reducing nontariff barriers is likely to p rovide even
further mileage in increasing this access (Hoekman and Nicita 2011).
Market access for developing countries is even more restricted in trade in
services, especially via the temporary movement of natural persons (Saez 2013).
This type of service export (mode 4 services trade in WTO terminology) has a
direct impact on employment. Expanding market access via mode 4 in labor-
importing countries could yield direct employment benefits for fragile
countries.
The second way to enhance the employment effect of exports is by increasing
the relative competitiveness of fragile countries’ exports, particularly in labor-
intensive sectors. This requires a broad set of interventions to improve trade con-
nectivity and firms’ productivity (Reis and Farole 2012). In conflict-affected and
post-conflict environments, both areas are usually particularly deficient (see the
evidence in chapter 1 on the poor performance of fragile countries in trade facili-
tation), mainly due to the destruction and insecurity caused by the conflict.
The international community has tried to address these constraints. One
example is the Conflict Affected States in Africa (CASA) initiative by the
International Finance Corporation, which specifically focuses on post-conflict
reconstruction.9 By providing long-term private sector development support, the
initiative helps countries address constraints that stifle economic recovery, such
as poor government institutional capacity, weak business associations, weak basic
infrastructure, and the deterioration of financial services. One approach to
improving the efficiency of export production in fragile countries is to develop
special economic zones. The idea behind it is that in conflict-affected countries,
it should be easier and more effective to achieve international competitiveness
for firms in a concentrated environment than in the whole country.10 While
this approach could be effective in promoting competitiveness for a subset of
producers, the evidence suggests that the implementation challenges are often
daunting (Farole and Akinci 2011).
Build Long-Term Conflict Resilience
The policies that we have so far focused on can be implemented over a relatively
short time horizon. This perspective is intrinsic to the report’s focus on how to
use trade, which is by its nature fast changing, to prevent conflict. However, the
analysis in the report also points to the need to modify a country’s structural
characteristics to build resilience to changes in trade flows. In chapter 2 we point
to several long-term conditions that make conflict more sensitive to trade-related
changes. The level of grievances and institutional capacity appear particularly
important.
Countries may need to focus on a number of areas to reduce grievances. It is
essential to achieve some resolution of the tensions arising from past conflicts,
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
88 How Trade Policy Could Ease Tensions in Fragile Countries
both because countries with a recent past of civil conflict are more likely to fall
back into conflict in general (World Bank 2011), and because a history of conflict
exacerbates the tendency to react to adverse trade changes through violence.
Reducing interpersonal economic inequality and tackling ethnic divisions are also
priorities to build conflict resilience to trade-related changes.
The analysis also underscores the importance of accountable and honest
government institutions to build resilience to changes in trade flows. In addition,
it suggests that federal political systems are more resilient to such changes than
central systems, probably because the former are better equipped than the latter
to respond to the needs of the regions within a country. Clearly, the governments’
responsiveness to tensions at the subnational level is likely to be important for
stability.
Building these conditions requires a longer term horizon than is usually
adopted by a government legislature. Yet investing in them is also likely to be
necessary to permanently break the conflict trap.
Notes
1. The concept of conflict in this chapter refers to both conflict risk (for fragile countries
not currently in conflict) and conflict intensity (for those where one or more civil
conflicts are ongoing).
2. In particular, the IEG evaluation points out that the World Bank is most responsive to
FCS in the immediate aftermath of conflict, while the Bank’s effectiveness in the
medium term has been lower because of a lack of proper understanding of the
countries’ drivers of conflict.
3. Note that we do not discuss the optimal fiscal instruments to extract revenues from
natural resources exports, nor the issue of assigning the ownership of natural
resources and the power to regulate their exploitation. These issues are important in
determining the size of revenues extractable from the natural resources endowment
as well as the political economy of the distribution of the resources. However, the
management of the revenues appears to be the really important issue shaping the way
resource revenues may affect conflict risk (see Brosio and Singh 2014 for an extensive
treatment of the optimal fiscal instruments and Haysome and Kane 2009 for a
discussion of the other issues).
4. That may also require a first allocation of the resources across subnational levels of
government. See Brosio and Singh (2014) for a discussion on the options for such an
allocation.
5. See Do, Levchenko, and Ravallion (2013) for an alternative view suggesting trade
insulation, such as export restrictions, as potentially preferable options to social
protection.
6. The evaluation highlights five lessons that may be important for any future such
initiative (IEG 2013a). First, a detailed strategic framework for crisis response is
necessary but not sufficient for the effectiveness of interventions. Second, the expan-
sion in the scale of operations requires commensurate enhancement of administrative
budgets. Third, owing to the small amount of additional funding made available, many
countries received only modest assistance that could not have had significant crisis-
mitigating impact. Fourth, the effectiveness of the assistance depends critically on
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Policy Could Ease Tensions in Fragile Countries 89
adequate analytical work and staff resources. Finally, for short-term responses to any
types of crises, having social safety net systems in place before a crisis hits is key to
protecting vulnerable households and individuals.
7. This evidence notwithstanding, a note of caution is in order. The empirical basis on
the effect of employment generation on conflict is still very thin. A recent systematic
review identifies just seven relevant studies, the majority of which either present
anecdotal evidence or are based on secondary literature reviews (Holmes et al. 2013).
8. The exporter-level data do not distinguish between sectors in terms of their labor
intensity.
9. See http://www.ifc.org/wps/wcm/connect/region__ext_content/regions/sub-saharan
+africa/advisory+services/strategicinitiatives/casa.
10. Among other countries, this approach has been tried most recently in Haiti, where a
SEZ was started in 2012 mainly to process exports for the U.S. market.
References
Anderson, K., M. Ivanic, and W. Martin. 2013. “Food Price Spikes, Price Insulation and
Poverty.” World Bank, Mimeo.
Asuni, J. B. 2009. “Blood Oil in the Niger Delta.” United States Institute for Peace Special
Report 229.
Attanasio, O., V. Di Maro, V. Lechene, and D. Phillips. 2013. “Welfare Consequences of
Food Prices Increases: Evidence from Rural Mexico.” Journal of Development Economics
104: 136–51.
Bazzi, S., and C. Blattman. 2014. “Economic Shocks and Conflict: Evidence from
Commodity Prices.” American Economic Journal: Macroeconomics.
Bellows, J., and E. Miguel. 2009. “War and Local Collective Action in Sierra Leone.”
Journal of Public Economics 93 (11–12): 1144–57.
Berman, N., and M. Couttenier. 2014. “External Shocks, Internal Shots: The Geography of
Civil Conflicts.” CEPR Discussion Paper 9895.
Berman, N., M. Couttenier, D. Rohner, and M. Thoenig. 2014. “This Mine Is Mine! How
Minerals Fuel Conflicts in Africa.” OxCarre Research Paper 141.
Blattman, C., and E. Miguel. 2010. “Civil War.” Journal of Economic Literature 48: 3–57.
Brosio, G., and R. Singh. 2014. “Revenue Sharing of Natural Resources: A Review of
International Practices.” World Bank, Mimeo.
Carpenter, T., and A. Lendle. 2011. “How Preferential Is World Trade?” CTEI Working
Paper, The Graduate Institute, Geneva.
Collier, P. 2008. “Post-Conflict Economic Policy.” In Building States to Build Peace, edited
by C. T. Call. Boulder, CO: Lynne Rienner.
Collier, P., V. L. Elliott, H. Hegre, A. Hoeffler, M. Reynal-Querol, and N. Sambanis. 2003.
Breaking the Conflict Trap: Civil War and Development Policy. Washington, DC: World
Bank and Oxford University Press.
Cotet, A. M., and K. K. Tsui. 2013. “Oil and Conflict: What Does the Cross Country
Evidence Really Show?” American Economic Journal: Macroeconomics 5 (1): 49–80.
Dal Bó, E., and P. Dal Bó. 2011. “Workers, Warriors and Criminals: Social Conflict in
General Equilibrium.” Journal of the European Economic Association 9 (4).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
90 How Trade Policy Could Ease Tensions in Fragile Countries
———. 2012. “Conflict and Policy in General Equilibrium: Insights from a Standard Trade
Model.” In M. R. Garfinkel and S. Skaperdas (eds.), The Oxford Handbook of the
Economics of Peace and Conflict. New York: Oxford University Press.
del Castillo, G. 2011. The Economics of Peace: Five Rules for Effective Reconstruction. Special
Report 286, U.S. Institute of Peace.
Devarajan, S., and M. Giugale. 2012. “The Case for Direct Transfers of Resource Rents in
Africa.” World Bank, Mimeo.
Devarajan, S., M. Giugale, H. Ehrhart, T. M. Le, and H. M. Nguyen. 2013. “The Case for
Direct Transfers of Resource Revenues in Africa.” CGD Working Paper 333.
Do, Q. T., A. Levchenko, and M. Ravallion. 2013. “Trade Insulation as Social Protection.”
Policy Research Working Paper 6448, World Bank.
Dube, O., and S. Naidu. 2013. “Bases, Bullets and Ballots: The Effect of U.S. Military Aid
on Political Conflict in Colombia.” Mimeo.
Dube, O., and J. Vargas. 2013. “Commodity Price Shocks and Civil Conflict: Evidence
from Colombia.” Review of Economic Studies 80 (4): 1384–21.
Farole, T., and G. Akinci. 2011. Special Economic Zones: Progress, Emerging Challenges, and
Future Directions. Directions in Development, World Bank.
Gillies, A. 2010. “Giving Money Away? The Politics of Direct Distribution in Resource-
Rich States.” Center for Global Development Working Paper 231.
Haysome, N., and S. Kane. 2009. “Negotiating Natural Resources for Peace: Ownership,
Control and Wealth-Sharing.” Briefing Paper, Center for Humanitarian Dialogue.
Hoekman, B., and A. Nicita. 2011. “Trade Policy, Trade Costs, and Developing Country
Trade.” World Development 39 (12): 2069–79.
Holmes, R., A. McCord, and J. Hagen-Zanker, with G. Bergh and F. Zanker. 2013. What Is
the Evidence on the Impact of Employment Creation on Stability and Poverty Reduction in
Fragile States: A Systematic Review. London: ODI.
IEG. 2011. Timor-Leste Country Program Evaluation, 2000–2010. Washington, DC:
Independent Evaluation Group.
———. 2013a. The World Bank Group and the Global Food Crisis: An Evaluation of the
World Bank Group Response. Washington, DC: Independent Evaluation Group.
———. 2013b. World Bank Group Assistance to Low-Income Fragile and Conflict-Affected
States: An Independent Evaluation. Independent Evaluation Group, Washington, DC.
Joab-Peterside, S., D. Porter, and M. Watts. 2012. “Rethinking Conflict in the Niger Delta:
Understanding Conflict Dynamics, Justice and Security.” Working Paper 26, Berkeley:
University of California.
Lin, Y-H., and G. Michaels. 2013. “Do Giant Oilfield Discoveries Fuel Internal Armed
Conflicts?” LSE, Mimeo.
Martin, P., T. Mayer, and M. Thoenig. 2012. “The Geography of Conflicts and Regional
Trade Agreements.” American Economic Journal: Macroeconomics 4 (4): 1–35.
Maystadt, J-F., G. De Luca, P. G. Sekeris, J. Ulimwengu, and R. Folledo. 2014. “Mineral
Resources and Conflicts in the Democratic Republic of the Congo.” Oxford Economic
Papers 66 (3): 721–49. http://oep.oxfordjournals.org/content/66/3/721.
Maystadt, J.-F., and O. Ecker. 2014. “Extreme Weather and Civil War: Does Drought Fuel
Conflict in Somalia through Livestock Price Shocks?” American Journal of Agricultural
Economics.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
How Trade Policy Could Ease Tensions in Fragile Countries 91
Miaari, S., A. Zussman, and N. Zussman. 2014. “Employment Restrictions and Political
Violence in the Israeli-Palestinian Conflict.” Journal of Economic Behavior and
Organization 101 (May): 24–44.
Miguel, E., S. Satyanath, and E. Sergenti. 2004. “Economic Shocks and Civil Conflict: An
Instrumental Variables Approach.” Journal of Political Economy 112 (4): 725–53.
Moss, T. 2011. “Oil to Cash: Fighting the Resource Curse through Cash Transfers.” CGD
Working Paper 237.
Reis, J. G., and T. Farole. 2012. Trade Competitiveness Diagnostic Toolkit. Washington,
DC: World Bank.
Revenue Watch Institute. 2013. The 2013 Resource Governance Index. Washington, DC:
Revenue Watch Institute.
Rohner, D., M. Thoenig, and F. Zilibotti. 2013. “War Signals: A Theory of Trade, Trust and
Conflict.” Review of Economic Studies 80 (3): 1114–47.
Rosenau, W., P. Chalk, R. McPherson, M. Parker, and A. Long. 2009. Corporations and
Counterinsurgency. Rand Corporation.
Saez, S. 2013. Let Workers Move: Using Bilateral Labor Agreements to Increase Trade in
Services. Washington, DC: World Bank.
Sayne, A. 2013. What’s Next for Security in the Niger Delta? U.S. Institute for Peace Special
Report 333.
World Bank. 2011. World Development Report 2011: Conflict, Security and Development.
Washington, DC: World Bank.
———. 2012. Africa Can Help Feed Africa. Washington, DC: World Bank.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Appendix A
Data Issues
Table A.1 Fragile Countries and Territories and Number of Battle Deaths
Avg. deaths Max. death WB non fragile Avg. deaths Max. death
WB fragile 2005–10 2005–10 (with high deaths) 2005–10 2005–10
Afghanistan 4,354 6,238 Sri Lanka 3,420 8,413
Angola 8 25 Pakistan 2,768 6,688
Bosnia and Herzegovina 0 0 India 1,148 1,336
Burundi 104 285 Colombia 538 1,389
Central African Republic 30 86 Russian Federation 437 696
Chad 431 1,250 Philippines 436 692
Comoros 0 0 Algeria 342 499
Congo, Dem. Rep. 585 1,978 Turkey 296 453
Congo, Rep. 0 0 Thailand 159 214
Côte d’Ivoire 0 0 Uganda 157 655
Eritrea 0 0 Ethiopia 150 515
Guinea-Bissau 0 0 Lebanon 129 774
Haiti 0 0 Georgia 104 621
Iraq 2,055 3,658 Iran, Islamic Rep. 78 133
Kiribati 0 0 Nigeria 68 405
Kosovo 0 0 Mali 51 126
Liberia 0 0 Indonesia 36 213
Libya 0 0 Peru 28 56
Marshall Islands 0 0 Niger 23 81
Micronesia, Fed. Sts. 0 0 Tajikistan 16 98
Myanmar 159 232 Ecuador 8 45
Nepal 260 1,104 Saudi Arabia 7 31
Sierra Leone 0 0 Djibouti 6 35
Solomon Islands 0 0 Azerbaijan 4 26
Somalia 1,209 2,158
Republic of South Sudan — —
Sudan 563 1,029
Syrian Arab Republic 0 0
Timor-Leste 0 0
Togo 0 0
Tuvalu 0 0
West Bank and Gaza NA NA
Yemen, Rep. 43 175
Zimbabwe 0 0
Sources: World Development Indicators and World Bank 2013.
Note: — = not available. Battle-related deaths are deaths in battle-related conflicts between warring parties in the conflict dyad (two conflict units
that are parties to a conflict). The left-hand side of the table comprises the FY13 list of “fragile countries and situations” as classified by the World
Bank as: a) having a harmonized average CPIA country rating of 3.2 or less, or b) the presence of a UN and/or regional peace-keeping or peace-
building mission during the past 3 years. The right-hand side of the table comprises the list of other countries which have had at least a minor civil
conflict (i.e. with at least 25 battle deaths in a year) between 2005 and 2010.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 93
94 Data Issues
Cross-Country Analysis
We collect data from different sources. For the conflict data we rely on the
Uppsala Conflict Data Programme, Peace Research Institute, Oslo (PRIO).1
Although there are other sources for conflict data (e.g. Fearon and Laitin 2003;
Sambanis 2004 and the Correlates of War (COW), Sarkees and Wayman
2010), PRIO has become the key dataset for cross-country analyses of conflict
determinants.
While all sources define a civil war as an internal conflict with at least 1,000
battle deaths, there are significant differences in the data concerning the number
of fatalities and the criteria used to code the onset of wars, what counts as a war,
and how to treat breaks in violence. These differences lead to dramatically differ-
ent civil war variables.2 We are reluctant to test the relationship between trade
shocks and conflict using all sources of conflict data, because the differences
among them would likely affect the analysis, and could lead to very different
results.
Instead, we follow the same practice as Bazzi and Blattman (2014), who
base their analysis on the more episodic PRIO and COW measures of conflict.
Bazzi and Blattman (2014, p. 13) assert that these databases are most relevant
to measuring the impact of commodity price changes on conflict, “as they
capture the ebb and flow of incentives for war as incomes rise or fall.” We use
the PRIO dataset as our main source of conflict data, and only use the COW
data for robustness, for three main reasons. First, the PRIO dataset has effec-
tively become the standard reference for cross-country studies on the determi-
nants of conflict. The majority of the recent studies we have reviewed use only
this dataset (e.g. Bruckner and Ciccone 2010; Lin and Michaels 2011; Hull and
Imai 2013; Nunn and Qian 2014), while a few use it along with other data
sources (e.g., Besley and Persson 2008; Bazzi and Blattman 2014). Second, we
have performed independent checks on both datasets and found PRIO to be
generally better at identifying civil conflicts than COW.3 Finally, along with
major civil conflicts, PRIO also codes minor conflicts as those above 25 battle
deaths per year. We believe these smaller conflicts provide a relevant comple-
ment to the more episodic, full-blown civil conflicts. In fact, the problem with
estimating the determinants of a rare event as conflict onset strengthens the
importance of adding these smaller conflicts to the database (more on this
below).
One issue with the major conflict data is that temporary reductions in the
yearly number of casualties (i.e. below 1,000) are automatically counted as peace
years. For instance, in the Angolan Civil War (1975–2002) the number of fatali-
ties was below 1,000 in 1991 and 1995, even though the war was continuing
with a large number of yearly casualties. To check the sensitivity of the results to
this coding procedure, we also construct an additional conflict variable, which
defines peace years in between war years as those with a number of battle-related
deaths that fall below 300. The results below for major conflict onset are based on
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Data Issues 95
this variable although the results are very similar when using the major conflict
onset from PRIO (results available upon request).
We use various trade variables in the empirical analysis. Two of them (i.e. export
and import price indices) rely on a combination of international commodity
prices and country-specific trade shares. The former come from data coded by
Bazzi and Blattman (2014), who combine price data for 65 commodities from
various sources, including the IMF International Financial Statistics (IFS), the U.S.
Bureau of Labor Statistics (BLS), Global Financial Data (GFD), and others.4 We
complement data from Bazzi and Blattman (2014) with commodity price data
from the World Bank to obtain prices for 73 internationally traded commodities.5
Using this data we are able to construct the export price index:
1
Σ j expij ,t − k
Px it = ln
∏
j
exp
Pjt ij ,t − k
(A.1)
where Px is defined as the log of the geometric average of the international com-
modity prices (Pjt).6 Each commodity is weighted by its country i-specific aver-
age value of exports between t−k and t−2, where in our case k takes the value of
12 or of 4 (see below). The lagged structure of the export variable ensures the
exogeneity of the weights.7 In addition, the moving average weights can capture
changes in the export structure such as large discoveries of oil or gas, which
static weights cannot do (Deaton 1995). To compute the weights we construct
exports and imports series for the 1962–2010 period at 4-digit (SITC Rev. 2)
level by combining the NBER-United Nations Trade data with data from UN
COMTRADE. This standard formulation allows the effects of commodity
prices on countries to vary according to the commodity’s share in a country’s
export bundle.
In the baseline specification we use “slow moving weights,” defined as the
average exports of each commodity from (t−2) to (t−12). Lagged averages of
10 years of data ensure a balance between the time invariant weights and “fast
moving weights” (t−2; t−4) used by Bazzi and Blattman (2014). This weighting
scheme also allows us to control for changes in the export structure without
giving excessive weight to changes of the export structure due to temporary
shocks.8
Two other differences in our index compared to what Bazzi and Blattman
(2014) use are important. First, their index is based on percentage changes in the
commodities’ price. As it turns out, this difference is important for the results.
Second, our baseline index is not scaled by size of the exported commodities
compared to gross domestic product (GDP). This is, in principle, a desirable
property of such an index (Bazzi and Blattman 2014 reflect the size of exports
relative to GDP by multiplying the index by the commodity exports-GDP
ratio at the mid-point of the period). Using this scaling does not affect our
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
96 Data Issues
results in any meaningful way, but it does slightly reduce the number of obser-
vations, as GDP data are not available for a few countries in some years (results
available upon request).9 Therefore we do not incorporate this scaling in the
index.
In addition to the export price indices, we develop a similar country-specific
variable for the import sector:
1
Σ j impij ,t − k
PM it = ln
∏
j
impij , t − k
Pjt
(A.2)
The variable is constructed by using import rather than export data to calcu-
late a weighted average of international commodity prices.
Changes in trade flows arise from other sources than simply changes in inter-
national commodity prices. Changes in demand in partner countries may be as
important as changes in international prices especially for countries that are not
diversified in terms of destination markets. To capture this potential effect, we
define a demand shock variable—constructed as a market potential
measure—as:
MPit = ln
∑ w i , j ,t − kGDPjt (A.3)
j
where we use nominal GDP (normalized to 100 in 2000) from World Bank
(2014) to build an index that is a weighted average of trade partners’ GDPs, with
each weight w being the lagged average share of country j in total exports of
country i over the period defined by t−k.
The last trade variable captures the trade relation with neighboring countries
and it is defined as:
Σ j borderij ∗ ( import ijt + export ijt )
Trade Neighbors it = (A.4)
Σ j ( import ijt + export ijt )
We use the GeoDist dataset from the Centre d’Études Prospectives et
d’Informations Internationales (CEPII) to define contiguous country pairs.
The variable Trade Neighbors (TN) proxies for the level of integration of coun-
try i with its neighbors. We expect countries that trade less with neighbors to
be at greater risk because the neighboring countries’ cost of fueling conflict is
lower.
Table A.2 presents the summary statistics of the main variables used in the
analysis, while table A.3 lists the interaction terms used in the analysis of how
local conditions affect the relationship between changes in trade and
conflict.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Data Issues 97
Table A.2 Summary Statistics, Cross-Country Analysis
Type Variable N Mean SD Min Max
Onset Civil conflict UCDP/PRIO 3,465 0.044 0.205 0 1
Major conflict UCDP/PRIO 4,053 0.021 0.144 0 1
Major conflict (modified) UCDP/PRIO 3,861 0.015 0.123 0 1
Major conflict COW 3,830 0.027 0.161 0 1
Ending Civil conflict UCDP/PRIO 1,135 0.1304 0.33689 0 1
Major conflict UCDP/PRIO 420 0.20714 0.40574 0 1
Major conflict phasing UCDP/PRIO 559 0.10912 0.31207 0 1
Battle-related Constructed PRIO 4,300 930.60 4,908 0 150,000
deaths High PRIO 4,300 1474.4 7,472 0 250,000
Low PRIO 4,300 306.2 1,802 0 50,000
Trade variables Px slow moving 4,300 0.068 0.947 −4.332 2.629
Pm slow moving 4,300 0.072 0.940 −3.308 2.001
MP slow moving 4,300 0.041 1.002 −1.120 7.404
Px fast moving 4,300 0.067 0.949 −4.280 2.550
Pm fast moving 4,298 0.072 0.941 −3.373 1.937
MP fast moving 4,300 0.040 1.003 −1.103 7.157
Controls Any conflict since 1946 4,300 0.646 0.478 0 1
Share trade with border 4,238 0.116 0.156 0 0.891
Neighbors’ conflict (any) 4,300 0.243 0.429 0 1
Sources: Authors’ calculations based on UCDP/PRIO and COW data.
Note: UCDP/PRIO refers to Uppsala Conflict Data Program/Peace Research Institute Oslo Conflict Data Set. COW refers to the
Correlates of War Project.
Table A.3 Interaction Variables for the Cross-Country Analysis
Variable Description Source
Neighbors
Neighbors’ conflict (any) Dummy equal to one if there is a conflict in a PRIO
neighboring country.
Share trade neighbors RTA Share of trade with neighbors with an RTA. COMTRADE + WTO
Share trade neighbors Share of trade with neighbors. COMTRADE
Grievances
Economic inequality Dataset derived from the econometric relationship EHII University of Texas
between UTIP-UNIDO, other conditioning variables,
and the World Bank’s Deininger & Squire dataset.
gini_net Estimate of Gini index of inequality in equivalized SWIID
(square root scale) household disposable (post-tax,
post-transfer) income, using Luxembourg Income
Study data as the standard.
gini_market Estimate of Gini index of inequality in equivalized SWIID
(square root scale) household market (pre-tax, pre-
transfer) income, using Luxembourg Income Study
data as the standard.
Ethnic fractionalization The probability that two randomly selected individuals Montalvo and Reynal-Querol 2005
in a country will belong to different ethno-linguistic
groups.
table continues next page
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
98 Data Issues
Table A.3 Interaction Variables for the Cross-Country Analysis (continued)
Variable Description Source
Ethnic polarization How far the distribution of the ethnic groups is from Montalvo and Reynal-Querol 2005
the bipolar distribution (i.e. 1/2, 0, 0, ... 0, 1/2)
Religious fractionalization The probability that two randomly selected individuals Montalvo and Reynal-Querol 2005
in a country will belong to different religious groups.
Religious polarization How far the distribution of the religious groups is from Montalvo and Reynal-Querol 2005
the bipolar distribution (i.e. 1/2, 0, 0, ... 0, 1/2)
Any conflict in last 10 yrs Dummy equal to one if there was a conflict in the last PRIO
10 years.
Political system
Elections Dummy for election in that year. NELDA
Federal Govt. Dummy for a government with a federal system. Institutions and Elections Project
Polity2 Captures the political regime characteristics. Polity
Parliamentary democracy Dummy for being a parliamentary (from DPI) DPI + Polity
democracy (Polity).
Programmatic party Share of the major four political parties with an DPI
ideological orientation with respect to economic
policy, weighted by number of votes.
Age of party in office DPI
Governance
Law Strength and impartiality of the legal system. ICRG
Military Indicates the degree of military participation in politics. ICRG
Accountability A measure of how responsive a government is to its ICRG
people.
Corruption A measure of the level of corruption. ICRG
Bureaucratic Indicates the “strength and quality of the bureaucracy.” ICRG
Composite index The mean value of the ICRG variables “Corruption,” ICRG
“Law and Order,” and “Bureaucracy Quality.”
Price transmission
nra_covt Value of production-weighted average of covered Anderson et al. 2008
products.
nra_cov_o Value of production-weighted average of covered Anderson et al. 2008
products.
Note: PRIO is the Peace Research Institute, Oslo. ICRG is the International Country Risk Guide. COMTRADE refers to the Commodity Trade Statistics
Database. WTO is the World Trade Organization. EHII is the Estimated Household Income Inequality dataset. DPI is the World Bank’s Database of
Political Institutions. SWIID is the Standardized World Income Inequality Database. NELDA is the National Elections Across Democracy and
Autocracy database. Sources are given in the table.
Nigeria
The data on conflict we use in this study is the Version 4 (1997–2013) of ACLED.
This version of the data covers all countries on the African continent from 1997.
ACLED definitions mainly concern actors and events. ACLED c ollects and codes
reports from the developing world on civil and communal conflicts, militia inter-
actions, violence against civilians, rioting, and protesting. ACLED covers activity
that occurs both within and outside the context of a civil war.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Data Issues 99
The calculation of consumption and production price indices is essential to
the model estimation. While there are a number of surveys in Nigeria, we use the
Nigeria Living Standards Survey (NLSS) 2003/04. This is the first survey of the
income and expenditure patterns of Nigerian households with sufficient data to
analyze conflict over time.10 Before describing the survey itself, we summarize
the methodology used in calculating the price indices.
The consumption price index CI for state s at time t is constructed as a geo-
metric average of prices weighted by the budget shares (computed from the
2003/04 NLSS):
N 2003
Σ j =1Exptot sj
j =1 ( p jst )
2003
Expshrsj
CI st = Π N × (A.5)
TotExps 2003
where pjst is the price of good j in state s at time t and Expshr 2003 is the share of
j in total expenditures in 2003/04 across households in s on all the N items for
which price data are available. In this way the sum of the shares always equal to
1. As we can only match a subset of consumed items with prices (the list of items
matched is available on request), we scale this index by the importance of those
expenditure items in total household expenditures in the state TotExp (the latter
term in equation A.5).11
The main advantage of the geometric over the arithmetic average is that it
allows the index to incorporate some substitution effect across commodities as
relative prices change. This type of formulation is common in the literature on
commodity prices and conflict (e.g. Arezki and Brückner 2011; Bazzi and
Blattman 2014; Calì and Mulabdic 2014).
The domestic price data come from Nigeria’s National Bureau of Statistics
(NBS), which collects monthly data for 143 food and nonfood items by state in
both rural and urban areas. The price data we use covers 2000–10.12 Our analysis
relies on the urban data, assuming that rural prices will be a markup/discounted
value of the urban prices. The rural data are not used because the Nigerian clas-
sification of the areas into urban and rural has not been updated since 1991, and
thus they are not representative of the current division into urban and rural. We
use two approaches to determining which price index from the NBS data is
matched to which production or consumption item from the household survey.
The first is a narrow price match, where the good is matched to price data with
exactly the same name. However, the limited number of items in the price data
means that relying on a narrow price match alone could exclude potentially
important consumption items that have no exact match in the price data. The
second approach is a broad price match, where the price of a food crop is also
applied to products which are complements of, or derived from, that food crop
(e.g. the price of cassava is used for its extract gavi).13 The value of the scaling
factor for both consumption and production indices by state, for the narrow and
broad match, are available on request.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
100 Data Issues
We construct the production price index in a similar fashion:
K 2003
Σ j =1Prodtot sj
j =1 ( p jst )
2003
Prodshrsj
PI st = Π K × (A.6)
2003
TotInc s
Where Prodshr are the shares of j in all K products for which price data are avail-
able (thus the shares sum to 1) and TotInc is the total household income from all
sources in the state.
Because each commodity price may refer to a different unit of measurement,
we normalize the price of every commodity to 100 in 2003 and then construct
the price index on the basis of the normalized series.
The oil price index is constructed by interacting the oil production value in
oil
2003 with the international oil price (Pst = oils × oilprt). We use oil production
data published in the Nigerian National Petroleum Corporation (NNPC) Annual
Statistical Bulletin. However, because these data are only reported at the oil well
level and not at the state level, we had to manually map the oil wells to a state.
To do that we use a combination of online google search and geo-mapping using
longitudes and latitudes of the oil well mapped to the state.
The oil price index variable should be exogenous to conflict. First, Nigeria is
a price taker in the international oil market, as it is a small producer (Nigeria
produced approximately 2.8 percent of world oil production in 2012).14 In
addition, oil production at the beginning of the period should not be influenced
by subsequent conflict, especially as we control for the level of past conflict (in
case there is persistence over time). Given the lack of GDP data by state, we
normalize the production by state-wise receipts of Value Added Tax (VAT) in
2003 (Nigeria Bureau of Statistics 2010). The VAT is a tax levied on products
and services, based on the contribution to output at each stage of production.
Thus low levels of VAT receipts indicate low levels of economic activity, and
vice versa. Table A.5 presents the summary statistics for these main regressors
along with the other control variables used in the analysis (the statistics for the
dependent variable are presented and discussed in appendix B).
The NLSS was designed to collect household characteristics, such as demo-
graphic, education, health, and migration, for the purpose of poverty analysis.
The survey covered the urban and rural areas of all the 36 states of the Federation
and the Federal Capital Territory. Ten Enumeration Areas (EAs) were studied in
each of the states every month, while 5 EAs were covered in Abuja. Information
on food expenditure and production by 18,770 households was considered.
Part B of the questionnaire asked respondents questions on household’s
consumption, including both expenditures and agricultural activities at the
household level. Household expenditure is categorized into nonfood and food
expenses.15 The former is, in turn, divided into frequently and less frequently
purchased items. Table A.4 shows that the mean per capita food expenditure
is highest in the South South and South East regions, which house the major
oil-producing wells. The South East region had mean total per capita expen-
diture of N 45,216, which is well above the national average. However, the
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Data Issues 101
Table A.4 Household Per Capita Expenditure on Food and Nonfood by Zone
Per capita food Per capita nonfood Total per capita
expenditure expenditure expenditure
South South 17,287 19,199 36,486
South East 22,314 22,902 45,216
South West 16,533 26,696 43,229
North Central 14,740 15,067 29,806
North East 15,364 12,171 27,535
North West 16,907 11,176 28,083
Total 17,094 18,506 35,600
Source: Elaboration on Nigeria Living Standards Survey 2003/04.
Table A.5 Summary Statistics of the Regressors (2004–11)
Obs. Mean SD Min Max
CI 296 77.40 23.98 21.92 137.72
PI 296 66.28 26.67 21.19 138.08
oth
CINAR 296 87.44 24.46 45.27 143.69
oth
PINAR 296 75.76 25.62 28.62 127.31
oth
CIBR 296 70.55 27.18 23.42 142.07
oth
PI BR 296 83.22 25.41 27.30 149.43
Intl
CINAR 296 69.43 21.34 30.18 132.61
Intl
PINAR 296 56.33 13.00 31.43 84.29
Oil index 296 60.27 155.41 0 925.30
President 296 0.28 0.45 0 1
Pop (2003) ln 296 8.15 0.40 7.25 9.15
Pop dens. (2003) ln 296 5.26 0.89 3.93 7.90
Past conflict event 296 29.86 49.31 2 264
Past fatalities 296 202.46 344.81 1 1,892
Past event with fat. 296 17.19 29.59 1 156
Past battle events 296 12.76 21.52 1 99
Past protest 296 5.70 10.20 0 60
Past civil. violence 296 10.84 18.35 0 103
Poverty gap 2003 296 18.46 11.07 5 54
Headcount poverty 2003 296 48.78 17.90 21 87
Multiple dominant groups dummy 296 0.32 0.47 0 1
Ethnic minorities > 2 296 0.68 0.47 0 1
more urban South West region had the highest levels of per capita nonfood
expenditures.
The agricultural production section of the survey collects information on agri-
culture income and assets; land, livestock, and equipment; harvest and disposal of
crops; seasonality of sales and purchases (key staples only); and other agricultural
income (both in cash and kind). Information on the production of agricultural
food is collected at a different frequency. Information on household produce
sales during the last 12 months is collected for certain items, such as staple
grains, field crops, and cash crops, including the value of sales from hunting,
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
102 Data Issues
honey, fruit/berries, milk, other dairy products, eggs, hides, wool and skin, and
mushrooms output. On the other hand, for roots, fruits, vegetables, and other crops
harvested piecemeal, respondents are asked how much the household sold in the
last two weeks. We converted these two week estimates to a yearly value of sales.16
The West Bank and Gaza
The data in this study were taken from various Palestinian and Israeli sources that
include information on the Palestinian labor market, on economic and socio-
demographic characteristics of Palestinian localities, on Palestinians fatalities
from the conflict, and on Palestinian and Israeli trade. This information was
aggregated to the level of the locality, which serves as the unit of analysis and
represents the smallest spatial unit for which economic data is available in the
West Bank and Gaza. Our sample consists of 532 localities in the West Bank and
37 localities in the Gaza Strip.
Data on the number of Palestinians fatalities from politically motivated vio-
lence (Palestinians killed by Israelis) during the Second Intifada (September
2000–December 2004) in each locality are taken from B’Tselem—the Israeli
Information Center for Human Rights in the Occupied Territories.17 B’Tselem
publishes detailed data records which include every Israeli and Palestinian fatality
during the Second Intifada.
We are also able to identify those Palestinian fatalities which occurred as a
result of political demonstrations or confrontation with the Israeli army by exam-
ining the description of each fatality provided by B’tselem. As this description is
not available for all fatalities, this subset of fatalities is incomplete; so in the
subsequent analysis we use it only for robustness purposes.
Locality-level data on Palestinian sectoral employment as well as on socio-
demographic characteristics come from the 1997 Palestinian population census
carried out by the Palestinian Central Bureau of Statistics (PCBS). The
Palestinian census data includes information about various localities’ character-
istics, such as total population, share of males in the population, share of popula-
tion between the ages 15–40, share of population with up to elementary
education, share of refugees in the population, share of households with more
than 8 persons, share of people married in the population, as well as availability
of public utilities. Importantly for our purposes, the census also records data on
the employment of the population for all private establishments in the PT at the
2-digit ISIC level.
We match this distribution of employment at the locality level with Palestinian
and Israeli trade data at the 5-digit SITC level (revision 3) in order to compute
the measure of exposure to the trade shocks. Palestinian annual import and export
data are taken from the Palestinian Central Bureau of Statistics. Israeli import and
export data (at the 5-digit SITC level) come from the COMTRADE dataset in
WITS.18 We also use world and Chinese exports data from the same source.
Other labor market variables were constructed from two sources.
Administrative data on all Palestinians employed in Israel with a permit in
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Data Issues 103
1999 come from the Israeli Ministry of Industry Trade and Labor, which is in
charge of issuing the permits. In addition we gathered further labor data,
including information on Palestinian private and public employment in Israel
from the Palestinian Labor Force Survey (PLFS). This survey has been admin-
istered every quarter since 1995 to a nationally representative sample of house-
holds. We restrict the sample from the PLFS to individuals in the labor force
between the ages of 15 and above and surveyed during at least one of the four
quarters in 1999. As the survey is not stratified at the locality level, we exclude
localities in which less than 30 individuals were interviewed in each round in
1999.19
Summary statistics for the key variables at the district level are provided in
table A.6.
Table A.6 Summary Statistics for Key Variables in the Israeli-Palestinian Conflict Study
Obs. Mean SD Min Max
Palestinian fatalities 569 4.33 21.84 0 331
Localities with at least one fatality 569 0.35 0.48 0 1
Δ Palestinian exports96–99 569 3.53 4.36 −6.21 13.48
Δ Palestinian exports to Israel96–99 569 4.53 5.48 −5.2 17.16
Δ Israeli imports from RoW96–99 569 −11.73 32.78 −144 120.12
Δ Palestinian imports96–99 569 25.33 17.61 −4.87 95.18
Δ Palestinian exports to RoW96–99 569 −1.00 1.31 −5.29 2.42
Socioeconomic Total population 569 4,542 18,499 4 353,113
characteristics in Share of males in the population 569 0.51 0.03 0.385 1
1997 in locality Share of population between the ages 15–40 569 0.39 0.04 0.273 0.727
Share of population with up to elementary education 569 0.61 0.12 0.25 0.973
Share of refugees in the population 569 0.28 0.32 0 1
Share of households with more than 8 persons 569 0.34 0.11 0 1
Share of married in the population 569 0.33 0.03 0.174 0.438
Localities in Gaza Strip 569 0.07 0.25 0 1
Pre-Intifada fatalities 569 0.25 1.44 0 20
Localities in Jerusalem 569 0.05 0.22 0 1
Availability of public Water 569 0.69 0.46 0 1
utilities in 1997 in Electricity 569 0.84 0.37 0 1
locality Sewage 569 0.11 0.31 0 1
Telephone (landline) 569 0.63 0.48 0 1
Share of employment in Israel in 1999 222 10.63 6.91 0 36.186
Locality type Urban 222 0.23 0.42 0 1
Refugee camp 222 0.10 0.30 0 1
Average wage of employees in Israel in 1999 212 100.96 16.08 52.03 150.160
Share of public sector employment in 1999 222 5.65 3.87 0 18.947
Average wage of public sector employees in 1999 209 59.57 35.03 30.66 538.280
Share of private sector employment in 1999 222 9.75 6.81 0.53 41.523
Average wage of private sector employees in 1999 221 61.92 17.35 16.62 150.135
Sources: Authors’ elaboration using different datasets; see text for details.
Note: See table A.7 for variables’ description.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
104 Data Issues
Table A.7 Description of Variables Used in the Palestinian Case Study
Variable Description
Palestinian Fatalities Average number of fatalities from politically motivated violence
(Palestinians killed by Israel) from the outbreak of the Second
Intifada (September 28, 2000) until December 2004 in the locality.
For Palestinian fatalities, the locality is the locality where the fatal
wounding occurred. There are a handful of cases in which the fatal
wounding occurred inside Israel. In those cases, we considered the
locality of residence, or the closest geographical locality.
Localities with at least one fatality Dummy variable which takes 1 if the locality has more than zero
fatalities and 0 if it has 0 fatalities.
Change in the Palestinian exports to Israel Sum of the change in the Palestinian exports to Israel in each
between the years 1999 and 1996 sector between the years 1999 and 1996, weighted by the 1997
employment share in that sector in each locality of the total
employees in the same locality.
Change in the Israeli imports from the rest Sum of the change in the Israeli imports from the rest of the world
of the world between the years 1999 (except the West Bank and Gaza) in each sector between the years
and 1996 1996 and 1999, weighted by the 1997 employment share in that
sector in each locality of the total employees in the same locality.
Change in the total Palestinian exports Sum of the change in the total Palestinian exports in each sector
between the years 1996 and 1999 between the years 1996 and 1999, weighted by the 1997
employment share in that sector in each locality of the total
employees in the same locality.
Change in the total Palestinian imports Sum of the changes in the total Palestinian imports in each sector
between the years 1996 and 1999 between the years 1996 and 1999, weighted by the 1997
employment share in that sector in each locality of the total
employees in the same locality.
Change in the total Palestinian exports to Sum of the change in the total Palestinian exports from the rest of the
the rest of the world between the years world (except Israel) in each sector between the years 1996 and
1996 and 1999 1999, weighted by the 1997 employment share in that sector in each
locality from the total employees in the same locality.
Share of employment in Israel in 1999 Share of Palestinian workers employed in Israel out of total working age
in 1999 in the locality.
Average wage of employees in Israel in 1999 Average daily wage of employees in Israel in 1999 in the locality.
Share of public sector employment in 1999 Share of public sector employees out of total working age in 1999 in
the locality.
Average wage of public sector employees Average daily wage of public sector employees in 1999 in the locality.
in 1999
Share of private sector employment in 1999 Share of private sector employees out of total working age in 1999 in
the locality.
Average wage of private sector employees Average daily wage of private sector employees in 1999 in the locality.
in 1999
Fatalities between January 1995 and Average number of fatalities from politically motivated violence
August 2000 (Palestinians killed by Israelis) from 1995 until the outbreak of the
Second Intifada (September 28, 2000) in the locality.
Notes
These datasets are available at http://www.prio.no/Data/Armed-Conflict
1.
/ UCDP-PRIO/.
2. For example Bazzi and Blattman (2014) report that in the PRIO dataset major civil
wars are coded in 7 percent of the country-years, compared to 20 percent of the
country-years in the Fearon and Laitin (2003) data.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Data Issues 105
3. We first checked the conflict instances that were missing in PRIO according to
Gersovitz and Kriger (2013) (i.e. Cameroon, the Central African Republic, the
Republic of Congo, Côte d’Ivoire, Kenya–Shifta War [1963–1967], Mali, Niger,
and Senegal). All of these conflicts, with the exception of the one in Kenya, have been
included in the latest versions of the PRIO dataset while they were still missing in
COW. As an additional check, we identified those conflict incidences recorded in
COW but missing in the PRIO data (e.g. Liberia in 1996). We find that in various
instances, these conflict years had been there in earlier versions of the PRIO data but
had been removed in the more recent versions, suggesting a frequent process of updat-
ing of the conflict episodes in the PRIO dataset.
4. See the web appendix to Bazzi and Blattman (2014) for the complete list of
sources.
5. The list of these commodities is available in a separate web appendix.
6. To avoid the effects of different units of measure across commodities, all international
commodity prices are normalized to 100 in 2000.
7. For example, commodity exports could react in anticipation of a conflict or as a
reaction to changes in international commodity prices themselves, thus making the
contemporaneous export shares invalid as weights.
8. One important concern comes from the fact that some countries are large exporters
of these commodities and their internal conditions may influence international prices.
We address this issue in the next sections.
9. The index scaled by export-GDP ratio is defined as:
1
exp Σ j expij,mid point Σ j expij, t − k
Px it = Π j Pjt ij, t − k ×
GDPmid point
10. The Living Standard Measurement Survey for Nigeria 2010/11 is not appropriate for
the purpose of our analysis for two reasons: the survey is only representative at the
geopolitical zone and not at the state level, and the period 2010/11 covers periods
after the Niger Delta conflict but before the core of the Boko Haram crisis.
11. Available domestic price data are matched with food and nonfood items in the survey
in order to estimate the indices. Items not matched are not used in the indices, but
contribute to the weights as described.
12. Though another batch of data is available for 2010–13, there are a number of
inconsistencies in the data that make it difficult to use at this point. The NBS
changed the methodology of data collection for the prices in those periods and
some of the prices were totally different when compared to the 2000–10 dataset.
Also the items in the 2010–13 datasets were different with more items included
and disaggregated.
13. The broad matching procedure relies on subjective judgments, based on our under-
standing of the country and the consumption items.
14. http://www.eia.gov/countries/country-data.cfm?fips=NI. Accessed April 29, 2014.
15. The expenditure on food by household is a sum of expenditure on each individual
food item over 6 visits. That is, aggregation of the response to the question, “How
much was spent on … since my last visit?”
16. One way of converting this is to multiply the two week estimate by 26 to get a total
of 52 weeks’ value of sale. However, inconsistency in the values reported for cassava,
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
106 Data Issues
yam, and plantain, which include data on both two weeks and annual sales, shows that
multiplication of the two weeks value by 26 is not a consistent estimate of the yearly
value. We therefore elected to predict the yearly value produced by each household,
by applying an average of the relationships between the yearly value and the two
weeks value reported for cassava, yam, and plantain to the other items.
17. Available at: http://www.btselem.org.
18. This is available at http://wits.worldbank.org/wits/.
19. Given this narrow geographical definition, many localities do not meet this criterion,
leaving us with 241 localities. We drop 42 additional localities for which key variables
in the analysis are missing.
References
Anderson, K., M. Kurzweil, W. Martin, D. Sandri, and E. Valenzuela. 2008. “Measuring
Distortions to Agricultural Incentives, Revisited.” Policy Research Working Paper
4612, World Bank.
Arezki, R., and M. Brückner. 2011. “Food Prices, Conflict, and Democratic Change.”
Mimeo.
Bazzi, S., and C. Blattman. 2014. “Economic Shocks and Conflict: Evidence from
Commodity Prices.” American Economic Journal: Macroeconomics.
Bellows, J., and E. Miguel. 2009. “War and Local Collective Action in Sierra Leone.”
Journal of Public Economics 93 (11–12): 1144–57.
Besley, T., and T. Persson. 2008. “The Incidence of Civil War: Theory and Evidence.”
Mimeo.
Bruckner, M., and A. Ciccone. 2010. “International Commodity Prices, Growth and the
Outbreak of Civil War in Sub-Saharan Africa.” The Economic Journal 120 (May):
519–34.
Calì, M., and A. Mulabdic. 2014. “Trade and Civil Conflict: Revisiting the Cross-Country
Evidence.” Policy Research Working Paper, forthcoming.
Deaton, A. 1995. “International Commodity Prices, Macroeconomic Performance, and
Politics in Sub-Saharan Africa.” Princeton Studies in International Finance, 79,
Princeton University.
Fearon, J., and D. Laitin. 2003. “Ethnicity, Insurgency, and Civil War.” American Political
Science Review 97 (1): 75–90.
Gersovitz, M., and N. Kriger. 2013. “What Is a Civil War? A Critical Review of Its
Definition and (Econometric) Consequences.” World Bank Research Observer 28 (2):
159–90.
Hull, P., and M. Imai. 2013. “Economic Shocks and Civil Conflict: Evidence from Foreign
Interest Rate Movements.” Journal of Development Economics 103: 77–89.
Lin, Y-H., and G. Michaels. 2011. “Do Giant Oilfield Discoveries Fuel Internal Armed
Conflicts?” CEPR Discussion Paper 8620.
Montalvo, J. G., and M. Reynal-Querol. 2005. “Ethnic Polarization, Potential Conflict, and
Civil Wars.” American Economic Review 95 (3): 796–816.
Nigeria Bureau of Statistics. 2010. Annual Abstract of Statistics.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Data Issues 107
Nunn, N., and N. Qian. 2014. “U.S. Food Aid and Civil Conflict.” American Economic
Review 104 (6): 1630–66.
Sambanis, N. 2004. “What Is Civil War? Conceptual and Empirical Complexities of an
Operational Definition.” Journal of Conflict Resolution 48: 814–58.
Sarkees, M. R., and F. Wayman. 2010. Resort to War: 1816–2007. Washington, DC:
CQ Press.
World Bank. 2013. “Harmonized List of Fragile Situations, FY 2014.” World Bank,
Washington, DC.
———. 2014. World Development Indicators, online dataset.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Appendix B
Estimation Methodology and
Empirical Results
Cross-Country Analysis
The main empirical analysis estimates the impact of various trade-related vari-
ables on the onset of conflict. We follow Bazzi and Blattman (2014) and model
conflict onset and ending separately, using split samples.1 This strategy allows
incorporating the dynamic properties of conflicts (Beck and Katz 2011). These
are highly persistent, so past years of conflicts affect current conflict. As the for-
mer are in turn affected by past shocks, not modeling the dynamics introduces a
bias in the estimation. The basic specification reads as follows:
j = 0 β j Px it − j + ∑ j = 0 γ j Pmit − j
COit = α i + ∑ n n
(B.1)
j = 0 θ j MPit − j + ϑ TN it −1 + ΓX it + ρt + ε it
+ ∑n
j = 0 β j Px it − j + ∑ j = 0 γ j Pmit − j
CEit = α i + ∑ n n
(B.2)
j = 0 θ j MPit − j + ϑ TN it −1 + ΓX it + ρt + ε it
+ ∑n
where CO and CE are conflict onset and ending (as defined above) for country
I at time t, Px is the (country-specific) export price index, Pm is the import price
index; MP is a country-specific market potential variable; TN is the share of i’s
trade with its neighbors (i.e. countries with which it shares a border) in total
trade; X is a vector of time varying controls, α are country fixed effects, r are
time effects, and ε is the i.i.d. error term. The errors are corrected for clustering
at the country level. The use of country fixed effects controls for any time invari-
ant factors that may influence the probability of conflict, such as geography,
ethnicity, religion, and colonial history. In addition, the different lags in the price
regressor account for the time-dependence of these shocks, which are usually
negatively autocorrelated and can take many periods to affect earnings.2
Unlike the other trade variables, TN does not have a lag structure, since it is
highly persistent and one single lag appears appropriate to capture its effect on
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6 109
110 Estimation Methodology and Empirical Results
conflict while reducing the potential endogeneity concerns. We argue that the
other three trade variables—Px, Pm, and MP—are also exogenous to the indi-
vidual countries’ conditions that are associated with the probability of conflict.
These variables capture the variation in demand and supply in international
goods markets and in a country’s trading partners. These factors should not be
influenced significantly by conditions in individual countries.
An exception to this rule is when the countries are large enough to influence
the international supply and/or demand in certain markets. If world prices rise in
anticipation of conflict, this may lead to a spurious positive correlation between
conflict and the lagged price index. In our sample there are 16 cases of a country
producing on average more than 20 percent of global exports of a particular
commodity. One way to deal with this issue is to exclude the commodity in
question from the country’s price index, which is what Bazzi and Blattman
(2014) do. However, that may introduce another source of bias. Consider for
instance coffee production in Colombia. The country is responsible for 14 per-
cent of global coffee exports. Dube and Vargas (2013) show that the sharp fall
in international coffee prices in the 1990s substantially increased conflict inten-
sity in Colombia. Thus excluding coffee from Colombia’s price index would bias
the Px coefficient upward. Because of this reason we do not exclude any com-
modities from the price variables in the baseline specifications. Instead, we check
the robustness of the results to the exclusion of relevant commodities (according
to different thresholds) as well as of some countries that are price makers in at
least one commodity (average share in global exports above the threshold).
The vector X contains controls that are likely to have an independent effect
on conflict. First, we include a dummy for whether the country has had any
conflict since World War II, which captures the higher likelihood of starting a
conflict for countries which already experienced one recently (World Bank
2011). Second, in order to control for the spillover effects of conflict, we include
a dummy for whether any of the neighboring countries has a civil conflict. The
variable is lagged one year to reduce the endogeneity concerns. In some specifica-
tions we also include the incidence of coup attempts (whether successful or not)
in the two years prior to t, based on Powell and Thyne (2011). This can be inter-
preted as an indicator of a weak state (Kuhn and Weidmann 2013), but it could
also act as a trigger for a civil conflict itself.
To differentiate the effects across classes of commodities (see chapter 1), we
split the Px variables into different groups of commodities. In particular, we dis-
tinguish between point-source and diffused commodities as well as between
commodities that are consumed in the country and those that are not. This yields
four different export price variables, each with a lag structure.
We use the linear probability model as in Bazzi and Blattman (2014) to esti-
mate equations (B.1) and (B.2), although the results are very similar when using
the conditional fixed effect logit and the probit estimator as well (results avail-
able upon request).
One problem with estimating equations (B.1) and (B.2) is that the dependent
variables have a very large number of zeros relative to the number of ones, i.e. the
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 111
models try to explain rare events.3 That is inherent in the nature of this type of
empirical analysis, although it is rarely acknowledged in other studies. Importantly,
this problem can lead to biased coefficients’ estimates in finite samples (King and
Zeng 2001). In addition, the bias in the standard errors tends to go in the same
direction as the bias in the coefficients. Thus the rare event bias may cause the
underestimation of event probabilities. King and Zeng (2001) suggest that one
way to correct the problem is to decrease the rareness of the event. That could be
done, for example, by lowering the threshold of what constitutes an event or by
expanding the data selection period. In our case this would strengthen the case for
using any civil conflict as the dependent variable. For this variable the number of
events is 25 times smaller than the number of zeros while it is 50 times smaller
for major conflict onset. We explore in more detail in the main text the implications
of this possible bias in interpreting the empirical results.4
While the main focus of the analysis is to identify the effects of the trade
variables on conflict onset (and ending), we also examine their impact on the
intensity of conflicts. This is an important outcome in its own right and may not
necessarily follow the same dynamics of conflict onset. To that end we also run
the following specification:
j = 0 β j Px it − j + ∑ j = 0 γ j Pmit − j
BDit = α i + ∑ n n
(B.3)
+∑ n
j =0 θ j MPit − j + ϑ TN it −1 + ΓXit + t + ε it
where BD is the number of battle-related deaths in country i at time t. The vector
X includes also a dummy for the first year of conflict to capture the impact of
conflict onset on the intensity of conflict (relative to both the years without
conflict and the years with conflict after the first). The vector X also includes a
variable measuring the number of years of conflict since the onset.
As we estimate the model over the entire sample (including nonconflict years)
this specification combines the effects on the extensive margin (the probability
of conflict) with those on the intensive margin (conflict intensity once the con-
flict has started). This specification is different to that used in Bazzi and Blattman
(2014), who run it only on conflict years. We argue that it is important to keep
the nonconflict years as well, to capture the effect of the trade variables on a
continuum of conflict intensity which varies from none to severe conflict. This
specification is similar to that used in most micro studies on the determinants of
conflict (e.g. Dube and Vargas 2013). This specification also allows us to avoid
the contentious issue of defining the correct threshold for a civil conflict
(Chaudion, Peskowitz, and Stanton 2012).
A number of options are available to estimate equation (B.3) in a way that
accounts for the fact that the dependent variable is a count variable with a large
proportion of zeros, i.e. the poisson, negative binomial or zero-inflated negative
binomial estimator (ZINB). In our setup the ZINB estimator is ruled out due to
a lack of convergence. As in our dependent variables the unconditional variance
is larger than the mean, the negative binomial is to be preferred to the poisson
estimator (Long and Freese 2006; Cameron and Trivedi 2013).
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Table B.1 The Impact of Trade on Conflict, Cross-Country Analysis
112
(1) (2) (3) (4) (5) (6) (7) (8)
Any onset Any onset Any onset Any onset Major onset Major onset Major onset Major onset
Px (t) 0.040** 0.039** 0.039** 0.010* 0.012 0.008 0.008 0.003
(0.016) (0.016) (0.016) (0.005) (0.011) (0.011) (0.011) (0.004)
Px (t−1) −0.014 −0.014 −0.015 0.003 −0.020** −0.016* −0.016* −0.003
(0.020) (0.020) (0.020) (0.006) (0.009) (0.009) (0.009) (0.003)
Px (t−2) 0.000 0.001 0.001 −0.005 0.007 0.005 0.005 −0.003
(0.017) (0.017) (0.017) (0.005) (0.009) (0.008) (0.008) (0.003)
Pm (t) −0.001 −0.001 0.002 −0.022 −0.022 −0.005
(0.024) (0.024) (0.006) (0.014) (0.014) (0.004)
Pm (t−1) −0.002 −0.003 0.003 0.017 0.017 0.000
(0.026) (0.026) (0.007) (0.016) (0.016) (0.004)
Pm (t−2) −0.009 −0.009 −0.001 −0.001 −0.001 0.002
(0.023) (0.022) (0.006) (0.014) (0.014) (0.003)
−0.006 −0.006 −0.001 0.005 0.005 0.000
MP (t)
(0.023) (0.023) (0.004) (0.010) (0.010) (0.002)
−0.041 −0.042 −0.008 −0.038 −0.038 −0.005
MP (t−1)
(0.054) (0.054) (0.010) (0.036) (0.036) (0.007)
0.061 0.063 0.007 0.052 0.052 0.006
MP (t−2)
(0.066) (0.065) (0.012) (0.047) (0.047) (0.011)
−0.033 −0.031 −0.037 −0.018 −0.018 −0.018
Trade with neigh. (t−1)
(0.053) (0.054) (0.057) (0.032) (0.032) (0.035)
0.179*** 0.183*** 0.185*** 0.177*** 0.046*** 0.045*** 0.045*** 0.047***
Any conflict since 1946
(0.021) (0.021) (0.021) (0.021) (0.012) (0.012) (0.012) (0.012)
−0.006 −0.007 −0.007 0.002 0.002 0.003
War border (t−1)
(0.012) (0.012) (0.012) (0.008) (0.008) (0.008)
−0.018 −0.015 −0.001 −0.001
Coup
(0.016) (0.018) (0.008) (0.008)
Observations 3,465 3,428 3,428 3,327 3,861 3,812 3,812 3,704
R-sq. (within) 0.069 0.072 0.072 0.068 0.025 0.026 0.026 0.026
Countries 114 114 114 114 115 115 115 115
Shocks logP logP logP ∆logp logP logP logP ∆logp
Weight [t−12; t−2] [t−12; t−2] [t−12; t−2] [t−12; t−2] [t−12; t−2] [t−12; t−2] [t−12; t−2] [t−12; t−2]
Time trends NO NO NO NO NO NO NO NO
Sum Px 0.026** 0.026** 0.025** 0.008 −0.001 −0.002 −0.002 −0.002
Sum Pm −0.013 −0.013 0.004 −0.006 −0.006 −0.003
Sum MP 0.015 0.016 −0.002 0.019 0.019 0.001
Note: Robust standard errors (clustered at the country level) in parentheses.
***p < 0.01, **p < 0.05, *p < 0.1.
Estimation Methodology and Empirical Results 113
Table B.2 Robustness with Fast-Moving, Country-Specific Time Trends, Cross-Country Analysis
(1) (2) (3) (4) (5) (6)
Any onset Any onset Any onset Major onset Major onset Major onset
Px (t) 0.037** 0.035** 0.033** 0.010 0.006 0.007
(0.017) (0.015) (0.016) (0.011) (0.010) (0.010)
Px (t−1) −0.015 −0.016 −0.015 −0.015* −0.015* −0.014*
(0.020) (0.020) (0.020) (0.009) (0.008) (0.008)
Px (t−2) 0.008 0.003 0.008 0.008 0.006 0.009
(0.019) (0.017) (0.018) (0.009) (0.007) (0.008)
Pm (t) 0.007 0.011 0.017 −0.021 −0.005 −0.003
(0.026) (0.023) (0.025) (0.015) (0.011) (0.012)
Pm (t−1) −0.007 −0.021 −0.023 0.015 0.003 0.004
(0.026) (0.022) (0.022) (0.016) (0.014) (0.014)
Pm (t−2) 0.001 0.001 0.010 0.004 0.001 0.004
(0.025) (0.017) (0.020) (0.016) (0.010) (0.011)
MP (t) −0.015 −0.016 −0.027 −0.007 −0.009 −0.024**
(0.023) (0.022) (0.024) (0.011) (0.011) (0.012)
MP (t−1) −0.043 −0.031 −0.052 −0.038 −0.026 −0.031
(0.055) (0.048) (0.050) (0.038) (0.038) (0.042)
MP (t−2) 0.068 0.057 0.093 0.059 0.053 0.068
(0.070) (0.057) (0.066) (0.051) (0.054) (0.060)
Trade with neigh. −0.084 −0.030 −0.080 −0.038 −0.015 −0.036
(t−1) (0.059) (0.054) (0.057) (0.041) (0.033) (0.041)
Controls YES YES YES YES YES YES
Observations 3,428 3,425 3,425 3,812 3,809 3,809
R−sq. (within) 0.156 0.072 0.157 0.064 0.027 0.066
Countries 114 114 114 115 115 115
Shocks logP logP logP logP logP logP
Weight [t−12; t−2] [t−4; t−2] [t−4; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2]
Time trends YES NO YES YES NO YES
Sum Px 0.030* 0.022** 0.025 0.003 −0.003 0.002
Sum Pm −0.00003 −0.009 0.003 −0.002 −0.002 0.005
Sum MP 0.010 0.010 0.014 0.013 0.018 0.013
Note: Robust standard errors (clustered at the country level) in parentheses; controls include any conflict since 1946 and war border (t−1).
***p < 0.01, **p < 0.05, *p < 0.1.
Table B.3 Robustness for Price Makers and Conflict Data Source, Cross-Country Analysis
(1) (2) (3) (4) (5) (6)
Conflict data PRIO PRIO PRIO PRIO PRIO COW
Dep. Var. Any onset Any onset Any onset Any onset Major (1,000) Major (1,000)
Px (t) 0.028* 0.040** 0.040** 0.043** 0.006 0.018**
(0.016) (0.018) (0.016) (0.016) (0.010) (0.009)
Px (t−1) −0.012 −0.023 −0.022 −0.025 −0.019* −0.025**
(0.020) (0.022) (0.020) (0.021) (0.011) (0.011)
Px (t−2) 0.001 0.007 0.005 0.008 0.004 0.011
(0.017) (0.019) (0.017) (0.017) (0.008) (0.009)
Pm (t) −0.006 0.002 −0.001 0.009 −0.007 0.037**
(0.024) (0.027) (0.024) (0.025) (0.017) (0.017)
Pm (t−1) −0.001 −0.006 −0.004 −0.009 0.005 −0.017
(0.026) (0.031) (0.026) (0.028) (0.017) (0.021)
table continues next page
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
114 Estimation Methodology and Empirical Results
Table B.3 Robustness for Price Makers and Conflict Data Source, Cross-Country Analysis (continued)
(1) (2) (3) (4) (5) (6)
Conflict data PRIO PRIO PRIO PRIO PRIO COW
Dep. Var. Any onset Any onset Any onset Any onset Major (1,000) Major (1,000)
Pm (t−2) −0.008 −0.017 −0.008 −0.013 −0.008 −0.007
(0.022) (0.027) (0.022) (0.024) (0.015) (0.014)
MP (t) −0.004 0.001 −0.007 −0.001 0.005 0.001
(0.023) (0.024) (0.023) (0.023) (0.012) (0.009)
MP (t−1) −0.042 −0.062 −0.038 −0.035 −0.032 −0.030
(0.054) (0.059) (0.054) (0.056) (0.046) (0.043)
MP (t−2) 0.061 0.081 0.059 0.049 0.043 0.057
(0.065) (0.070) (0.066) (0.067) (0.057) (0.061)
Trade with neigh. −0.034 −0.053 −0.033 −0.025 −0.037 0.003
(t−1) (0.053) (0.055) (0.053) (0.053) (0.034) (0.036)
Observations 3,428 2,491 3,428 2,992 4,001 3,834
R-sq. (within) 0.070 0.084 0.071 0.079 0.024 0.036
Countries 114 85 114 98 115 115
Threshold 10% 10% 20% 20% None None
Exclude Commod. Countries Commod. Countries None None
Sum Px 0.017* 0.024** 0.023** 0.025** −0.009 0.004
Sum Pm −0.015 −0.022 −0.014 −0.013 −0.010 0.013
Sum MP 0.015 0.020 0.015 0.013 0.016 0.028
Note: Robust standard errors (clustered at the country level) in parentheses; all regressions include country fixed effects, year effects and controls (any
conflict since 1946 and war border (t−1)); trade shock variables are weighted using the slow-moving averages (without country-specific time trends).
***p < 0.01, **p < 0.05, *p < 0.1.
Table B.4 Splitting the Commodities’ Variables into Different Types, Cross-Country Analysis
(1) (2) (3)
onset_any onset_any onset_any
Sum Px point-source 0.018*
(0.01)
Sum Px diffused 0.013
(0.013)
Sum Px consumed 0.017*
(0.009)
Sum Px not consumed 0.022
(0.017)
Sum Px point-source consumed 0.032
(0.033)
Sum Px point-source not consumed 0.013
(0.015)
Sum Px diffused consumed −0.019
(0.015)
Sum Px diffused not consumed 0.021
(0.015)
Other trade variables YES YES YES
Controls YES YES YES
Weight [t−12; t−2] [t−12; t−2] [t−12; t−2]
Observations 3,403 3,426 3,104
R-sq. (within) 0.070 0.072 0.072
Countries 114 114 114
Note: Robust standard errors (clustered at the country level) in parentheses; all regressions include country fixed effects, year
effects and controls (any conflict since 1946 and war border (t−1)); other trade variables include MP and Pm with their three
lags and trade with neighbors (t−1).
***p < 0.01, **p < 0.05, *p < 0.1.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Table B.5 The Impact of Trade Variables on the Likelihood of Conflict Coming to an End, Cross-Country Analysis
(1) (2) (3) (4) (5) (6) (7) (8)
Any ending Any ending Any ending Any ending Major ending Major ending Major ending Major ending
Px (t) −0.047 −0.053 −0.037 −0.038 0.019 0.020 0.038 0.061
(0.036) (0.034) (0.033) (0.030) (0.059) (0.077) (0.041) (0.057)
Px (t−1) 0.040 0.033 0.036 0.028 −0.004 −0.013 −0.045 −0.056
(0.054) (0.055) (0.052) (0.052) (0.065) (0.059) (0.061) (0.055)
Px (t−2) −0.002 −0.024 −0.011 −0.043 0.009 0.034 0.014 0.030
(0.036) (0.036) (0.037) (0.036) (0.053) (0.058) (0.050) (0.057)
Pm (t) −0.044 −0.046 −0.033 −0.042 0.242* 0.237* 0.094 0.179*
(0.070) (0.080) (0.059) (0.063) (0.127) (0.140) (0.088) (0.105)
Pm (t−1) 0.129 0.098 0.101 0.077 −0.142 −0.143 −0.131 −0.142
(0.102) (0.101) (0.083) (0.079) (0.169) (0.159) (0.127) (0.119)
Pm (t−2) 0.020 0.007 0.021 0.002 0.075 0.082 0.101 0.044
(0.080) (0.089) (0.071) (0.078) (0.134) (0.167) (0.098) (0.098)
MP (t) −0.093 −0.078 −0.050 −0.030 0.123 0.181 0.357 0.399
(0.180) (0.191) (0.148) (0.172) (0.347) (0.343) (0.252) (0.305)
MP (t−1) 0.060 −0.064 0.187 0.094 0.114 −0.084 −0.379 −0.571
(0.332) (0.312) (0.268) (0.258) (0.479) (0.432) (0.409) (0.394)
MP (t−2) −0.079 −0.018 −0.278 −0.249 −0.320 0.032 −0.038 0.229
(0.293) (0.334) (0.234) (0.260) (0.426) (0.604) (0.331) (0.412)
Trade with neigh. (t−1) 0.249** 0.199* 0.221** 0.155 0.402 0.463 0.359 0.375
(0.095) (0.116) (0.094) (0.108) (0.276) (0.436) (0.268) (0.415)
Observations 1,104 1,104 1,101 1,101 540 540 537 537
R-sq. (within) 0.058 0.196 0.056 0.196 0.168 0.297 0.168 0.303
Countries 80 80 80 80 41 41 41 41
Shocks logP logP logP logP logP logP logP logP
Weight [t−12; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2] [t−12; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2]
Time trends NO YES NO YES NO YES NO YES
Sum Px −0.008 −0.045 −0.011 −0.053 0.024 0.041 0.007 0.035
Sum Pm 0.105*** 0.058 0.089** 0.037 0.174** 0.177 0.064 0.081
Sum MP −0.113 −0.16 −0.141 −0.185 −0.082 0.129 −0.059 0.056
Note: Robust standard errors (clustered at the country level) in parentheses; all regressions include country fixed effects, year effects and controls (any conflict since 1946 and war border (t−1)).
115
***p < 0.01, **p < 0.05, *p < 0.1.
116
Table B.6 The Impact of Trade on Battle Deaths, Cross-Country Analysis
(1) (2) (3) (4)
Battle deaths Battle deaths Battle deaths Battle deaths
Sum Px 0.202*** 0.169***
(0.053) (0.052)
Sum Pm 0.217*** 0.278*** 0.276*** 0.326***
(0.073) (0.088) (0.074) (0.088)
Sum MP −0.55*** −0.534*** −0.486*** −0.543***
(0.168) (0.172) (0.162) (0.169)
Trade with neigh. (t−1) −2.119*** −1.972*** −2.097*** −1.939***
(0.284) (0.289) (0.285) (0.295)
Sum Px point-source −0.038 0.011 (0.069)
(0.066)
Sum Px diffused 0.130** 0.062 (0.057)
(0.057)
Sum Px consumed
Sum Px not consumed
First year 1.704*** 1.690*** 1.699*** 1.679***
(0.088) (0.087) (0.088) (0.089)
Duration 0.106*** 0.106*** 0.106*** 0.105***
(0.004) (0.004) (0.004) (0.005)
Controls YES YES YES YES
Weights [t−12; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2]
Observations 3,125 3,110 3,121 3,051
Countries 76 76 76 76
Note: Robust standard errors (clustered at the country level) in parentheses; models are estimated through the negative binomial estimator; all regressions include country fixed effects, a time trend and other
controls (any conflict since 1946 and war border (t−1)); trade shock variables are weighted using the slow-moving averages (no country-specific time trends).
***p < 0.01, **p < 0.05, *p < 0.1.
Table B.7 The Effect of Trading with Neighbors on Conflict, Revisited, Cross-Country Analysis
(1) (2) (3) (4) (5) (6) (7) (8)
Any onset Any onset Any onset Any onset Major onset Major onset Major onset Major onset
Trade with neighbors(RTA) (t−1) −0.077* −0.106 −0.084** −0.113* 0.000 −0.006 −0.005 −0.008
(0.040) (0.065) (0.040) (0.064) (0.021) (0.025) (0.022) (0.026)
Instrumented NO NO NO NO YES YES YES YES
Time trends NO YES NO YES NO YES NO YES
Weights [t−12; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2] [t−12; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2]
Observations 3,464 3,464 3,461 3,461 3,461 3,461 3,461 3,461
R-squared 0.071 0.158 0.071 0.158 0.071 0.158 0.071 0.158
Nr. of countries 114 114 114 114 114 114 114 114
First stage Kleibergen-Paap
(9) (10) (11) (12) (13) (14) (15) (16)
Trade with neighbors(RTA) (t−1) −0.077** −0.108* −0.077** −0.108* −0.007 −0.015 −0.007 −0.015
(0.038) (0.055) (0.038) (0.055) (0.018) (0.021) (0.018) (0.021)
Instrumented NO NO NO NO YES YES YES YES
Time trends NO YES NO YES NO YES NO YES
Weights [t−12; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2] [t−12; t−2] [t−12; t−2] [t−4; t−2] [t−4; t−2]
Observations 3,860 3,860 3,857 3,857 3,857 3,857 3,857 3,857
R-squared 0.026 0.067 0.026 0.065 0.026 0.065 0.026 0.065
Number of countries 115 115 115 115 115 115 115 115
First stage Kleibergen-Paap 22,743 6,948 2,2743 6,948 27,094 8,503 27,094 8,503
Note: Robust standard errors (clustered at the country level) in parentheses; all regressions include country fixed effects, year effects and the trade shocks variables as in the previous tables and other controls (any
conflict since 1946 and war border (t−1)); trade shock variables are weighted using the slow-moving averages (no country-specific time trends).
***p < 0.01, **p < 0.05, *p < 0.1.
117
118 Estimation Methodology and Empirical Results
Table B.8 Trade Variables without Lag Structure, Cross-Country Analysis
(1) (2) (3) (4)
Any onset Any onset Major onset Major onset
Px (t) 0.029*** 0.028* −0.000 0.005
(0.010) (0.015) (0.005) (0.009)
Pm (t) −0.005 0.003 −0.007 −0.003
(0.016) (0.022) (0.008) (0.010)
MP (t) −0.001 −0.021 0.003 −0.003
(0.012) (0.018) (0.004) (0.008)
Trade with neigh. (t − 1) −0.017 −0.060 −0.014 −0.030
(0.048) (0.056) (0.029) (0.036)
Controls YES YES YES YES
Observations 3,628 3,628 4,017 4,017
R-sq. (within) 0.074 0.152 0.025 0.059
Countries 115 115 115 115
Shocks YES YES YES YES
Time trends logP logP logP logP
Note: Robust standard errors (clustered at the country level) in parentheses; all regressions include country
fixed effects, year effects and controls (any conflict since 1946 and war border (t-1)); trade shock variables are
weighted using the slow-moving averages.
***p < 0.01, **p < 0.05, *p < 0.1.
Nigeria
We use the indices developed in appendix A in the regression framework to
measuring the impact of price shocks on conflict. The basic specification reads as
follows:
Csrt = α rt + β1CI st −1 + β2 PI st −1 + β3Pst
oil
−1
(B.4)
+β P oil
4 st −1 × d2009 + AZ s + BX st + ε st
where C is a measure of conflict (e.g. number of conflict episodes, number
of violent episodes, number of conflict-related fatalities), Poil is the oil price
index, d2009 is a post-2009 dummy which captures the period after the
amnesty deal between the state and the militant groups in the Niger Delta;
Z and X are vectors of time invariant and time varying state-level covariates
of conflict, respectively, α are region-time fixed effects, which capture any
time varying effect at the regional level. The standard errors are clustered at
the state level.
The count nature of the data on conflict makes applying an ordinary linear
regression model problematic. The Poisson and Negative Binomial models are
the two commonly used models for this kind of data characteristics (count data),
because they ensure a positive conditional mean of the conflict variables. The
Poisson model has the advantage that it does not require that the model be
Poisson distributed to use it—that is, the model requires a weaker distributional
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 119
Table B.9 Summary Statistics of the Dependent Variable (2004–11), Nigeria
Variable Obs Mean Std. Dev. Min Max % of Zeroes
Nr. of fatalities from conflict episodes 296 18.4 85.5 0 1,001 41.6
Nr. of conflict events in a year 296 6.9 12.7 0 118 21.3
Number of conflict events with fatalities 296 2.2 6.0 0 79 41.6
Source: ACLED.
assumption than the negative binomial model. However, the negative binomial
model is designed to handle over-dispersion in the data and will lead to higher
efficiency in estimation.5 The summary statistics of the main dependent variables
(reported in table B.9) confirm that the data is over-dispersed with standard
deviations much larger than the mean. Thus we opt to use the negative binomial
over the poisson estimator.
Among the controls, Z includes a number of important variables measured
at the beginning of the period of analysis (from the NSS), i.e. population,
population density, and measures of poverty and inequality, including the
headcount poverty rate, poverty gap, and the Gini index of inequality. Z also
includes the cost of travelling to Lagos (Nigeria’s main trading center), the
number of conflict events between 1997 and 2003, and ethnic variables.
Ideally we would use m easures of ethnic divisions traditionally used in the
conflict literature, such as ethnic fractionalization (Alesina et al. 2003) and
polarization (Montalvo and Reynal-Querol 2005). However, in the absence of
state-level data on the ethnic composition of the population, the next best
variable we can construct is a dummy for whether there are more than two
ethnic minorities in the state.6
We also construct two time-varying ethnic measures of the relation between
the state’s dominant ethnic group(s) and the ethnic group holding the presi-
dency, which are included in X. The first (president) equals 1 if the ethnicity of
the nation’s president is the same as that of one of the state’s dominant groups.
This variable captures the idea that federal policies toward the states may be
driven, in part, by ethnic allegiance. The second is a dummy variable for those
states in which the president variable equals 1 and which have only one domi-
nant ethnic group. This allows us to differentiate the president’s effect between
these two types of states.
This wide range of state-level covariates, along with region effects, should
compensate for the absence of state fixed effects in the regressions.
We propose four sets of indices—two for the consumption and two for the
production indices—as instruments to check the endogeneity of price indices to
conflict. These indices are constructed in the same way as CI and PI (see appen-
dix A), but using prices which are arguably exogenous to the conflict at the state
level. The first set of price indices is the standard one based on international
prices that the literature usually employs as a direct regressor in the absence of
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
120 Estimation Methodology and Empirical Results
domestic price data (e.g. Bazzi and Blattman 2014; Dube and Vargas 2013).
The instrument is constructed as follows:
I 2003
∑ j =1 Exptot sj (B.5)
Intl
)
2003
Expshrsj
Intl
Cst = ∏ Ij =1 p jt ×
2003
TotExps
where PjtIntl is the international price of good j at time t. This approach has several
difficulties. Replacing domestic with international prices requires changing the
set of goods included in the index. The range of goods for which international
prices are available (from 1 to I) is more limited than the N or K goods included
in equations (A.5) and (A.6) of the section on Nigeria of appendix A.
International prices are available only for internationally traded commodities,
which often do not include many local products important for consumption and
production in Nigeria (e.g. yam and cassava). The amount of survey data matched
to international prices is more limited than the amount matched with domestic
prices, i.e. the scaling term for these instruments is smaller than in the CI and PI
in equations (A.5) and (A.6) of appendix A. The same applies to the PI instru-
ment as well. Moreover, international prices do not account for the price trans-
mission from international to domestic markets, which is often limited. Thus
international prices may not provide an ideal representation of the size of the
price shock at the local level.
We propose another set of instruments to address both issues. The instruments
are constructed using domestic prices of faraway states, following the same logic
of Jacoby (2013) for changes in rice prices in Indian districts. The price data for
the other Nigerian states should reflect exogenous international price changes,
their transmission to the domestic market, and shifts in demand and supply
in the large domestic market outside of the particular state. We exclude neigh-
boring states that may be affected by the conditions in the state in question, to
ensure the exogeneity of the instruments.7
For each state s, we compute the weighted average of prices of states located
beyond a certain travel distance (D) to the capital of state s—weighted by the
inverse of D:
other 1
p jst = ΣN S
m =1 × p jmt (B.6)
Dm
where pjmt is the price of j in state m at time t for all the Ns states whose capital
is located beyond 11 hours travel distance. The eleven hours threshold is based
on both the mean and median bilateral distance between the state capitals.
We argue that this threshold excludes all the states that are close to the state’s
geopolitical zone of influence. On average, about 10 states are included on
the basis of this threshold. Differently from Jacoby (2013), we penalize far-away
state’s prices, conditional on being more than 11 hours away, by applying the
inverse distance weight. This ensures that within the set of states beyond 11
hours, those relatively closer to the state in question have a greater weight.8
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 121
other
We then replace the pjst in equation (A.5) of appendix A with p jst to obtain
oth
the instrument CI st . We also do the same for PI. Note that the rest of equation
(B.5) is unchanged, as the goods j are the same in equations (A.5) and (A.6) in
appendix A, since the price data come from the same source (Nigeria Bureau of
Statistics). That is, of course, the case for both the narrow and the broad match-
ing of goods between the price and the survey data. It is also the case for the
production indices defined in equation (A.6) of appendix A.
Following Cameron and Trivedi (2013), we use these instruments to extract
the endogenous component of CI and PI through the first stage regressions.
We use the instruments in separate regressions:
CI srt = α rt + δ1othCI st
oth
+ δ2
oth oth
PI st + δ 3Pst
oil
+ δ 4 Pst
oil
× d2009 + Z s + X st + µst (B.7)
PI srt = α rt + δ1intlCI st
oth
+ δ2
intl oth
PI st + δ 3Pst
oil
+ δ 4 Pst
oil
× d2009 + Z s + X st + v st (B.8)
CI srt = α rt + δ1CI st
intl
+ δ 2 PI st
intl
+ δ 3Pst
oil
+ δ 4 Pst
oil
× d2009 + Z s + X st + µst (B.7’)
PI srt = α rt + δ1CI st
intl
+ δ 2 PI st
intl
+ δ 3Pst
oil
+ δ 4 Pst
oil
× d2009 + Z s + X st + v st (B.8’)
Table B.10 The Impact of Price Shocks on Conflict Events in Nigeria (2004–11)
(1) (2) (3) (4) (5) (6)
Dep. Variable Any event Any event Any event Any event Any event Any event
oth oth oth oth oth oth Intl Intl Intl Intl
Instruments PINAR PINAR CINAR , PINAF CIBR , PIBR CINAR , PINAR CIBR , PIBR
CINAR (t−1) 0.052*** 0.053*
(0.013) (0.030)
PINAR (t−1) 0.054 0.024 −0.047*** −0.001
(0.053) (0.052) (0.012) (0.025)
Oil ind. (t−1) 0.002** 0.003*** 0.003*** 0.003*** 0.003***
(0.001) (0.001) (0.001) (0.001) (0.001)
Oil ind. (t−1) x post-09 −0.003*** −0.003*** −0.003*** −0.003*** −0.003***
(0.001) (0.001) (0.001) (0.001) (0.001)
CIBR (t−1) 0.038*** 0.078*
(0.011) (0.043)
PIBR (t−1) −0.036*** −0.008
(0.01) (0.026)
Marginal effects: change in number of conflict events caused by a 10% increase in price index
CI 2.97 2.31 2.83 4.51
PI − −2.30 −1.86 −0.05 −0.39
Oil ind. 0.01 0.14 0.15 0.12 0.14
Oil ind. x post-09 −0.02 −0.02 −0.02 −0.02 −0.02
Note: Dependent variable is the number of any conflict events in the state in year t; all regressions include residuals from the
first stage regressions of the endogenous price index on the excluded instruments as a control with appropriate standard
errors (clustered at the state level) calculated; *, **, *** indicate statistical significance at the 10, 5, and 1 percent level,
respectively. Data are for 37 states for 8 years (2004–11). All regressions include year and region-year effects and various
controls (the log of population in 2003, the log of population density in 2003, the number of conflict events in 1997–2003,
the poverty gap and headcount poverty in 2003, a dummy for multiple dominant ethnic groups, a dummy for more than 2
ethnic minorities, a dummy for whether the federal president is of the same ethnicity as the dominant group in the state, and
the interaction between this variable and the multiple dominant groups dummy). The models are estimated through the
population-averaged negative binomial estimator for panel data.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
122 Estimation Methodology and Empirical Results
Table B.11 The Impact of Price Shocks on Conflict in Nigeria (2004–11), Robustness
(1) (2) (3) (4) (5)
Method Nbreg Nbreg Nbreg Nbreg Nbreg
Period 2004–10 2004–10 2004–11 2004–11 2004–11
oth oth oth oth oth oth
Instruments CINAR , PINAR CINAR , PINAR CINAR , PINAR
CINAR (t) 0.024* −0.021
(0.012) (0.034)
PINAR (t) −0.010 0.035
(0.011) (0.043)
CINAR (t−1) 0.079 0.066**
(0.053) (0.031)
PINAR (t−1) −0.049 −0.032
(0.070) (0.054)
oth
CINAR (t−1) 0.051***
(0.009)
oth
PINAR (t−1) 0.015
(0.012)
Intl
CINAR (t−1) 0.013*
(0.008)
Intl
PINAR (t−1) −0.001
(0.024)
Oil ind. (t) 0.003*** 0.001
(0.001) (0.001)
Oil ind. (t−1) 0.002*** 0.003*** 0.002*** 0.002***
(0.001) (0.001) (0.001) (0.001)
Oil ind. (t) x post-09 −0.001
(0.001)
Oil ind. (t−1) x post-09 −0.003*** −0.004*** −0.003***
(0.001) (0.001) (0.001)
Controls YES YES YES YES YES
Observations 259 259 296 296 296
CI + CI (t−1) 0.058**
PI + PI (t−1) −0.014
Oil + Oil (t−1) 0.003***
Note: Dependent variable is the number of any conflict events in the state in year t; all regressions include residuals from the first stage regressions
of the endogenous price index on the excluded instruments as a control with appropriate standard errors (clustered at the state level) calculated;
*, **, *** indicate statistical significance at the 10, 5, and 1 percent level, respectively. Data are for 37 states for 8 years (2004–11). All regressions
include year and region-year effects. The models are estimated through the population-averaged negative binomial estimator for panel data.
and retrieve the respective estimated residual components of the price indices
oth
µsrt
oth
and v srt or µ srt
intl intl
and v srt , which should contain the endogenous component
of CIsrt (Cameron and Trivedi 2013) We then add these endogenous components
of CI and PI to equation (B.4), which becomes (when using the domestic price
indices of faraway states as instruments):
Csrt = α rt + β1CI st −1 + β2 PI st −1 + β 3Pst −1 + β 4 Pst −1 × d2009 + AZ s
oil oil
(B.4’)
+ BX st + µ srt
oth oth
+ v srt + ε st
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Table B.12 The Impact of Price Shocks on Various Types of Conflict in Nigeria (2004–11)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
oth oth Intl oth oth Intl oth oth Intl oth oth Intl oth
CINAR CI BR CI NAR CINAR CI BR CI NAR CI NAR CI BR CI NAR CI NAR CI BR CI NAR CINAR
oth oth Intl oth oth Intl oth oth Intl oth oth Intl oth
PI NAR PI BR PI NAR PI NAR PI BR PI NAR PI NAR PI BR PI NAR PI NAR PI BR PINAR PINAR
Events with fatalities Battle events Protests/riots Violence against civilians Fatalities
CINAR (t−1) 0.059*** 0.063* 0.054*** 0.017 0.080 −0.002 0.038** 0.033 0.068***
(0.016) (0.033) (0.015) (0.034) (0.058) (0.041) (0.016) (0.032) (0.013)
PINAR (t−1) −0.027** 0.008 −0.043*** 0.012 −0.025 0.029 −0.034** −0.020 −0.021**
(0.011) (0.023) (0.014) (0.028) (0.107) (0.026) (0.014) (0.024) (0.009)
CIBR (t−1) 0.042*** 0.044*** 0.040*** 0.029**
(0.014) (0.012) (0.014) (0.014)
PIBR (t−1) −0.014 −0.036** −0.032** −0.027**
(0.012) (0.012) (0.014) (0.012)
Oil ind. (t−1) 0.003*** 0.003*** 0.003*** 0.004*** 0.003* 0.003*** 0.000 0.000 −0.001* 0.003*** 0.003*** 0.003*** 0.001*
(0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001)
Oil ind. (t) x post-09 −0.004*** −0.004*** −0.004*** −0.004** −0.004** −0.004*** −0.003 −0.003 −0.003* −0.002** −0.002** −0.002*** −0.006***
(0.001) (0.001) (0.001) (0.002) (0.002) (0.002) (0.002) (0.002) (0.002) (0.001) (0.001) (0.001) (0.001)
Note: All regressions include residuals from the first stage regressions of the endogenous price index on the excluded instruments as a control with appropriate standard errors (clustered at the state level)
calculated; *, **, *** indicate statistical significance at the 10, 5, and 1 percent level, respectively. Data are for 37 states for 8 years (2004–11). All regressions include year and region-year effects, and a full set of controls
as in tables B.10 and B.11. The models are estimated through the population-averaged negative binomial estimator for panel data.
123
124 Estimation Methodology and Empirical Results
Table B.13 Mediating Factors Affecting the Impact of Price Shocks on Conflict
Any events Events with fat. Battles Protests and riots civ. violence
election x CI (t−1) 0.034** 0.034** 0.030* −0.003 0.030**
election x PI (t−1) −0.023 −0.036** 0.010 −0.024 −0.049***
election x Oil Ind (t−1) 0.002*** 0.002 0.001 0.002 0.002**
cost_lagos x CI (t−1) −0.000 0.000 −0.001 −0.001** −0.000
cost_lagos x PI (t−1) 0.001** 0.000 0.001** −0.001** 0.001
president x CI (t−1) −0.001 −0.019 −0.018 0.006 0.020
president x PI (t−1) −0.002 0.008 0.009 −0.015 −0.013
president x Oil Ind (t−1) −0.002 −0.000 −0.003*** 0.002* −0.001
unem03 x CI (t−1) −0.000 −0.001 −0.001 −0.000 −0.000
unem03 x PI (t−1) 0.001 0.001 0.001 0.001 0.001
unem03 x Oil Ind (t−1) −0.000 −0.000 −0.000 0.000** −0.000
mult_domin x CI (t−1) −0.016 0.022 0.003 −0.002 −0.006
mult_domin x PI (t−1) −0.008 −0.035** −0.039 −0.015 −0.015
mult_domin x Oil Ind (t−1) 0.003*** 0.002** 0.004*** 0.001 0.003***
many_minor x CI (t−1) −0.011 0.029 −0.002 −0.045 0.009
many_minor x PI (t−1) 0.007 −0.027 −0.000 0.022 −0.000
many_minor x Oil Ind (t−1) 0.004*** 0.002*** 0.005*** 0.001 0.004***
pov03 x CI (t−1) 0.000 −0.000 −0.000 0.000 0.000
pov03 x PI (t−1) −0.000 0.000 −0.000 −0.000 −0.000
pov03 x Oil Ind (t−1) −0.000 −0.000 −0.000 −0.000 −0.000
gini03 x CI (t−1) −0.270** −0.223 −0.187 −0.072 −0.708***
gini03 x PI (t−1) 0.256* 0.270 0.184 0.070 0.729***
gini03 x Oil Ind (t−1) 0.038*** 0.046*** 0.063*** 0.002 0.030***
Past conflict x CI (t−1) −0.000 −0.000 −0.001** 0.002* −0.001
Past conflict x PI (t−1) −0.000 −0.000 0.001* −0.003** 0.001
Past conflict x Oil ind (t−1) −0.000 −0.000 −0.000 −0.000*** 0.000
Note: The table reports the coefficients of the interaction terms between the price indices and various conditioning factors
obtained from different regressions; all regressions include residuals from the first stage regressions of the endogenous price
index on the excluded instruments as a control with appropriate standard errors (clustered at the state level) calculated;
*, **, *** indicate statistical significance at the 10, 5, and 1 percent level, respectively (based on adjusted standard errors).
Data are for 37 states for 8 years (2004–11). All regressions include year and region-year effects and a full set of controls as in
tables B.10 and B.11. The models are estimated through the population-averaged negative binomial estimator for panel data.
The coefficients of CI and PI should not suffer from endogeneity bias as the
residuals from the first stage should purge the endogenous component of the
price indices. This formulation ensures the computation of consistent standard
errors (Cameron and Trivedi 2013). The signs of the residuals (not reported in
the tables but available upon request) confirm the direction of the endogeneity
bias, that is negative for CI and positive for PI. In other words, by reducing the
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 125
Table B.14 The Impact of Changes in Trade Prices on the Boko Haram Conflict (2010–13)
(1) (2) (3) (4) (5) (6) (7) (8)
Region All North All North North North North North
Dep. variable BH event BH event Non BH Non BH BH fat. Non BH fat. BH fatalities Non BH
event event event event fatalities
Intl
CINAR (t−1) 0.046** 0.067*** −0.013 0.014 0.116*** 0.010 0.078*** −0.025
(0.020) (0.018) (0.011) (0.013) (0.018) (0.016) (0.024) (0.022)
Intl
PINAR (t−1) −0.105** −0.093** −0.003 0.008 −0.024 −0.000 0.020 0.050
(0.043) (0.043) (0.018) (0.023) (0.026) (0.023) (0.038) (0.034)
Oil ind. (t−1) 0.001 −0.001***
(0.002) (0.000)
Observations 148 80 148 80 80 80 80 80
Nr. of states 37 20 37 20 20 20 20 20
Note: Robust standard errors in parentheses; *, **, *** indicate statistical significance at the 10, 5, and 1 percent level, respectively. All regressions
include year and region-year effects and a full set of controls as in in tables B.10 and B.11. The models are estimated through the population-
averaged negative binomial estimator for panel data.
demand for products, conflict reduces the prices of consumed goods and thus it
generates a spurious negative correlation between CI and conflict. The opposite
is true for PI. Therefore in both cases the endogeneity biases the CI and PI
coefficient toward zero.
As it turns out, the international price indices have a weak predictive power
for CI and PI because of the different composition of the items’ basket and the
limited transmission of international commodity prices to the Nigerian market.
However, the international price indices are useful, since (unlike the domestic
price indices) they are available through 2013.
These results are robust to using the instrumented PI and CI constructed
through the broad matching of the items (column 4). However, the results for
production are not robust to using the PI instrumented through the international
prices whether through the narrow (column 5) or the broad matching
(column 6). On the other hand, the result for CI holds although the coefficient
is less significant. These weaker results suggest that the two problems described
above (limited transmission of international to domestic prices and limited
number of items matched) in using international prices to capture price shocks
at the local level may be important in the case of Nigeria.9 The oil index
coefficients are unaffected by the use of international prices for CI and PI.
Israeli-Palestinian conflict
We estimate the following equation using a Negative Binomial regression, which
is an appropriate method for analyzing count data characterized by many zeros
and by over-dispersion (Long and Freese 2006):
Flr = gr + b1∆EXPl + BXl + el,(B.9)
where F is the number of fatalities in each locality l in region r during the period
of September 2000 and December 2004. The change in Palestinian exports in
each locality, ∆EXPl, is calculated as the sum of the change in exports in each
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
126 Estimation Methodology and Empirical Results
sector s during the 1996–99 period weighted by the share of employment in that
sector in the locality’s total private employment in 199710:
N
emp1997
∆EXPl = ∑s =1
∆exp s × ls
1997
empl
(B.10)
This measure of exposure to the export changes is close in spirit to Topalova
(2010) and should capture the effect of exports on conflict via employment.
Thus it should provide for a direct test of the opportunity cost hypothesis. An
appealing feature of this measure for our identification is that it should be exog-
enous to changes in local conditions related to conflict. We provide below
evidence in support of this exogeneity hypothesis by showing that the ∆EXP
coefficient is not affected when instrumenting DEXP through measures using
plausibly exogenous sources of export changes.
The specification (B.9) includes also region effects (where the regions are
Gaza Strip, Jerusalem, and the West Bank) as well as the vector X, which con-
tains a host of sociodemographic variables at the locality level that might affect
the distribution of employment across sectors and the level of conflict. These
variables refer to 1997 and include the total population, share of males in the
population, share of the population aged 15–40, share of population with ele-
mentary education or below, share of households with more than 8 members,
and the share of married individuals. We also control for other factors that may
foster Palestinian grievances. First, we include controls for the availability of
public utilities such as water, electricity, sewage, and landline telephones.
Second, we control for the unemployment rate in 1997 (computed from the
census). Third, we also control for the number of permits to work in Israel in
1999, which Miaari, Zussman, and Zussman (2014) show to be associated with
the subsequent violence. Finally, to control for the cyclicality of the conflict, the
vector X includes also the number of Palestinian fatalities in each locality
between January 1995 and August 2000. The error term is ε and standard errors
are robust.
In some specifications, we also account for other local labor market charac-
teristics that might be correlated with changes in exports as well as violence.
These controls include the locality-wise distribution of Palestinian workers
across the main types of employment (i.e. private and public sector and Israel),
their relative wages in 1999 (i.e. the first year these data are available from the
PLFS), and the distribution across location types, i.e. urban, rural, or refugee
camp.
In some specifications, we split the change in total Palestinian exports into
exports to Israel, the dominant export market, and exports to the rest of the
world, both weighted according to equation (B.10). We also test for the effects
of changes in Palestinian imports between 1996 and 1999, DIMP, weighted by
the employment shares as described in equation (B.10):
Flr = gr + b1∆EXPl + b2∆IMPl + BXl + el (B.11)
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 127
We also estimate an alternative model to investigate whether changes in
Palestinian exports affect the probability of violence in different localities. In
order to do this, instead of using the number of fatalities in each locality we
construct an indicator variable that equals 1 if the locality experienced a positive
number of fatalities during the Second Intifada and zero otherwise.11 We esti-
mate this specification using a probit model.
In order to substantiate the argument that the trade measures are exogenous
to local level conditions related to the conflict, we also instrument the export (and
other import) measures. In order to do so we use instruments based on plausibly
exogenous sources of Palestinian export changes. In particular we use two such
sources. The first is the changes in Chinese export supply over the same period
(1996–99). This is arguably an important source of competition of Palestinian
exports especially in Israel. Indeed, unlike imports from the West Bank and Gaza,
Israeli imports from China dramatically increased between 1995 and 2000
(figure 1.13). In order to ensure the exogeneity of Chinese exports changes to
Israeli conditions (which may be related to the conflict prospects), we take the
changes in Chinese sectoral exports to the world but Israel. We replace the
Palestinian exports with this variable in equation (B.10) to generate the first
instrument (∆ChnExp). A rise in Chinese sectoral exports may also be related to
a general increase in the world demand in those sectors. To control for this effect
we also add an instrument, which is constructed again as in equation (B.10) but
using world sectoral exports instead of Palestinian exports (∆WldExp).
The second source of exogenous changes in Palestinian exports is Israeli
trade policy. In particular Israel, as much of the rest of the world, undertook a
substantial process of (unilateral) tariff liberalization during the 1990s, which
reduced import duties in most sectors. The scale of this reduction can be
gauged in figure B.1, which reports the distribution of the nonzero Most
Figure B.1 Changes in Israeli MFN Tariffs (5-digit SITC Rev. 3), 1993–2004
0.08
0.06
Density
0.04
0.02
0
–50 0 50 100 150 200
Di erence MFN 1993–2004
Source: WITS.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
128 Estimation Methodology and Empirical Results
Favoured Nation (MFN) tariff differences between 1993 and 2004 (the only
years during that period for which data are available in WITS). This reduction
in Israeli tariffs caused a reduction in the preferential access of Palestinian
exports to their major destination market, thus effectively acting as a reduc-
tion in demand for Palestinian goods from Israel.12 Again we use the change in
Israeli tariffs to replace the changes in Palestinian exports in equation (B.10) to
generate the instrument for ∆EXP (i.e. ∆IsrMFN).13 The main problem with
this instrument is that it covers a larger period of time than do the changes in
pre-Intifada trade that we are considering. Because of that we also check the
robustness of the results to excluding this instrument from the first stage
estimation.
We instrument the ∆EXP with this series of variables in the following first
stage specification:
∆EXPlr = gr + a1∆ChnExpl + a2∆WldExpl + a 3∆IsrMFNl + BXl + ml (B.12)
The results of this specification—presented in table B.21—suggest that these
instruments are good predictors of changes in Palestinian exports over 1996–99.
In particular ∆ChnExp is negatively associated with Palestinian export changes in
line with the idea that Chinese exports may displace Palestinian exports.
Conversely the coefficient of ∆WldExp is positive and significant as world
demand positively affects Palestinian exports as well. Similarly, ∆IsrMFN has a
positive and significant coefficient consistent with the hypothesis that a reduc-
tion in Israeli import tariffs to the rest of the world induces also a reduction in
Palestinian exports (as competition in the Israeli market increases). These instru-
ments appear to be relevant also in explaining Palestinian exports to Israel and to
the rest of the world. It is noticeable that the effect of Israeli tariffs is of opposite
sign in the case of Palestinian exports to the rest of the world than in the case of
Palestinian exports to Israel (column 5). This suggests a reorientation of
Palestinian exports away from the Israeli market as a result of a reduction in
preferential access to Israel.
We follow a similar procedure to that of the Nigerian analysis to correct for
the endogeneity bias on the basis of this first stage. Armed with the estimated
endogenous component of ∆EXP computed from equation (B.12), we add that
in equation (B.9), which becomes:
Flt = γ r + b1∆EXP + BXl + µl + ε l (B.9’)
Table B.19 presents the results of this equation, which confirm the robust-
ness of the effects of changes in Palestinian exports on subsequent conflict
intensity. The b1 coefficient is slightly larger in absolute term than the compa-
rable coefficient in table B.15. That is especially when excluding ∆IsrMFN from
the instrument set (column 2). However, the magnitude of the coefficient is
not statistically different from that in table B.15, thus confirming that the
endogeneity bias is limited in this context. The results also hold for Palestinian
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 129
Table B.15 The Impact of Palestinian Exports on Conflict Intensity
(1) (2) (3) (4) (5) (6)
Region All All All All West Bank Gaza
Sample All All LFS LFS All All
∆Palestinian exports96–99 −0.125*** −0.128*** −0.137*** −0.159*** −0.086*** −0.528*
(0.029) (0.030) (0.051) (0.050) (0.028) (0.283)
Population 0.000* 0.000 0.000 0.000 0.000* 0.000
Male share −5.701 −5.959 −10.663 −4.296 −5.143 −84.722
Age 15–40 −0.739 −0.123 9.395 14.063* −0.034 −85.317*
Education −1.917* −1.865* −1.491 −3.292* −2.047* −16.864*
Share refugees 0.209 0.276 −0.215 0.501 0.110 2.173*
Large households (%) −3.845*** −4.017*** −3.301* −3.489* −3.352*** 7.556
Married (%) −12.299*** −12.242*** −22.521*** −22.195*** −12.432*** −2.373
Gaza 0.272 0.100 −0.332 −0.059
Jerusalem 0.165 0.140 −0.473 −0.956** 0.330
Public water 0.167 0.162 −0.269 −0.183 −0.049 −0.603
Public electricity 0.113 0.095 −0.300 −0.215 0.428 −6.386***
Public sewage 1.329*** 1.339*** 0.326 1.049*** 1.146*** 0.836
Telephone lines 1.290*** 1.285*** 1.323*** 1.501*** 1.084*** 4.026*
Work permits to Israel 0.001 0.002*** 0.002*** 0.002 0.001**
Past Pal. Fatalities −0.021 0.076** 0.059 −0.198*** −0.001
Unemployment 0.275 −0.823 −1.356 0.694 −1.645
Empl. Israel (%) −0.015
Avg. wage emp. Israel 0.018**
Empl. Public (%) 0.065
Avg. wage emp. Public −0.009
Private empl. (%) 0.038
Avg. wage emp. Private −0.013**
Urban (%) 1.179***
Refugee camp (%) 1.535***
Observations 569 569 199 199 532 37
Marginal effect ∆EXP −0.091 −0.092 −0.312 −0.430 −0.049 −0.691
Note: The dependent variable is the number of fatalities from politically motivated violence (Palestinians killed by Israelis) from
the outbreak of the Second Intifada (September 28, 2000) until December 2004 in locality. See table A.7 for the definitions of the
independent variables. The regressions are estimated using a Negative Binomial model. Robust standard errors (clustered at the
locality level) are reported in parentheses. The symbols *, **, *** represent statistical significance at the 10, 5, and 1 percent levels.
exports to Israel (columns 3–4), while they are weaker for Palestinian exports
to the rest of the world, probably due to the relatively weak power of the
instruments in predicting this variable.
The results are also robust to including the predicted import variables using
the same instrument set (columns 5–7). These variables continue to be not
significant even in these specifications.14 The b1 coefficient is also robust to
using the restricted LFS sample (column 8) and it is also robust when consider-
ing the 532 localities in the West Bank (columns 9–10). On the other hand, it
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
130
Table B.16 The Impact of Palestinian Trade on Conflict Intensity
(1) (2) (3) (4) (5) (6) (7)
Sample All All All LFS All All LFS
∆Palestinian exports96–99 −0.121*** −0.127*** −0.122*** −0.146***
(0.029) (0.030) (0.029) (0.052)
∆Palestinian exports to Israel96–99 −0.147*** −0.133*** −0.120
(0.041) (0.042) (0.075)
∆Palestinian exports to RoW96–99 −0.231 −0.179 −0.028
(0.162) (0.175) (0.302)
∆Palestinian imports96–99 −0.006 −0.004 −0.016 −0.004 −0.017
(0.005) (0.008) (0.012) (0.008) (0.012)
∆Israeli imports from RoW96–99 0.003 0.001 −0.005 0.001 −0.005
(0.003) (0.004) (0.004) (0.005) (0.004)
Other controls YES YES YES YES YES YES YES
Observations 569 569 569 199 569 569 199
Notes: The dependent variable is the number of fatalities from politically motivated violence (Palestinians killed by Israelis) from the outbreak of the Second Intifada (September 28, 2000) until December 2004 in
locality. See table A.7 for the definitions of the independent variables. Other controls include all the controls in table B.15 column 2. The regressions are estimated using a Negative Binomial model. Robust standard
errors (clustered at the locality level) are reported in parentheses. The symbols *, **, *** represent statistical significance at the 10, 5, and 1 percent levels.
Estimation Methodology and Empirical Results 131
Table B.17 The Impact of Palestinian Trade on a Different Measure of Conflict Intensity
(1) (2) (3) (4)
Dep. variable Palestinian fatalities during hostilities and demonstrations
∆Palestinian exports96–99 −0.150*** −0.137***
(0.041) (0.042)
∆Palestinian exports to −0.187*** −0.175***
Israel96–99 (0.050) (0.051)
∆Palestinian exports to −0.355* −0.340
RoW96–99 (0.213) (0.218)
∆Palestinian imports96–99 −0.012 −0.011
(0.009) (0.009)
∆Israeli imports from −0.003 −0.004
RoW96–99 (0.005) (0.005)
Other controls YES YES YES YES
Observations 569 569 569 569
Notes: The dependent variable is the number of fatalities from politically motivated violence during demonstration
(Palestinians killed by Israelis) from the outbreak of the Second Intifada (September 28, 2000) until December 2004 in locality.
See table A.7 for the definitions of the independent variables. Other controls include all the controls in table B.17, column 2.
The regressions are estimated using a Negative Binomial model. Robust standard errors (clustered at the locality level) are
reported in parentheses. The symbols *, **, *** represent statistical significance at the 10, 5, and 1 percent levels.
Table B.18 The Impact of Palestinian Trade on Conflict Probability
(1) (2) (3) (4)
Dep. variable Dummy at least 1 fatality
∆Palestinian exports96–99 −0.053*** −0.055***
(0.017) (0.018)
∆Palestinian exports to −0.063** −0.069***
Israel96–99 (0.025) (0.026)
∆Palestinian exports to −0.107 −0.131
RoW96–99 (0.103) (0.105)
∆Palestinian imports96–99 0.002 0.002
(0.006) (0.006)
∆Israeli imports from −0.000 −0.001
RoW96–99 (0.003) (0.003)
Other controls YES YES YES YES
Observations 569 569 569 569
Notes: The dependent variable is a dummy for whether the locality experienced at least one fatality from politically motivated
violence (Palestinians killed by Israelis) from the outbreak of the Second Intifada (September 28, 2000) until December 2004.
See table A.7 for the definitions of the independent variables. Other controls include all the controls in table B.15, column 2.
The regressions are estimated using a probit model. Robust standard errors (clustered at the locality level) are reported in
parentheses. The symbols *, **, *** represent statistical significance at the 10, 5, and 1 percent levels.
is not robust for the 37 localities in Gaza (columns 11–12). Finally, the results
also hold for fatalities as a result of hostilities or demonstrations (column 13)
as well as for conflict probability (column 14). The instrumentation again
slightly raises the effect of changes in exports on the latter: a USD 10 million
increase in export in a sector which employs 10 percent of private employees
in a locality is associated with a reduction in the probability of conflict in that
locality by 7.8 percent.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
132 Estimation Methodology and Empirical Results
Table B.19 The Impact of Palestinian Trade on Conflict: Tackling Endogeneity
(1) (2) (3) (4) (5) (6) (7)
Dep. Variable Palestinian fatalities during the 2nd Intifada
∆Palestinian −0.161*** −0.174*** −0.144*** −0.169*** −0.164***
exp96–99 (0.036) (0.037) (0.044) (0.036) (0.051)
∆Palestinian exp. to −0.161*** −0.136**
Israel96–99 (0.053) (0.057)
∆Palestinian exp. to −0.175 0.053
RoW96–99 (0.216) (0.271)
∆Palestinian 0.024 0.000
imports96–99 (0.034) (0.041)
∆Israeli imp. from 0.008 0.006
RoW96–99 (0.007) (0.009)
Instruments All ∆ChnExp All ∆ChnExp All All All
∆WldExp ∆WldExp
Observations 569 569 569 569 569 569 569
(8) (9) (10) (11) (12) (13) (14)
Sample LFS All All All All All All
Region All West Bank West Bank Gaza Gaza All All
Dep. Variable Palestinian fatalities during the 2nd Intifada Fat demo Dummy
∆Palestinian −0.190** −0.124*** −0.169 −0.139*** −0.082***
exp.96–99 (0.071) (0.034) (0.356) (0.052) (0.024)
∆Palestinian exp. to −0.131*** 1.969
Israel96–99 (0.046) (1.281)
∆Palestinian exp. to −0.173 8.067
RoW96–99 (0.195) (4.995)
Instruments All All All All All All All
Observations 199 532 532 37 37 569 569
Note: The regressions are estimated using a Negative Binomial model except column 14 which is estimated through a probit model. See table A.7
for the definitions of the independent variables. Other controls include all the controls in table B.15, column 2. All regressions include the
estimated residual term of the first stage (table B.23). Robust standard errors (clustered at the locality level) are reported in parentheses. The
symbols *, **, *** represent statistical significance at the 10, 5, and 1 percent levels.
Table B.20 The Heterogeneity of the Impact of Export Changes on Conflict Intensity
(1) (2) (3) (4) (5) (6) (7) (8)
Dep. variable Palestinian fatalities during the 2nd Intifada
Pre-Intifada Fatal. x ∆PalExp 0.003 0.004
(0.032) (0.034)
Refugee (%) x ∆PalExp −0.204** −0.214**
(0.087) (0.094)
Male (%) x ∆PalExp 1.696*** 1.681***
(0.589) (0.603)
Large HH (%) x ∆PalExp −0.162 −0.166
(0.192) (0.222)
Unemployment (%) x ∆PalExp −0.426* −0.158
(0.245) (0.273)
table continues next page
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 133
Table B.20 The Heterogeneity of the Impact of Export Changes on Conflict Intensity (continued)
(1) (2) (3) (4) (5) (6) (7) (8)
Dep. variable Palestinian fatalities during the 2nd Intifada
Age 15–40 (%) x ∆PalExp 0.359 0.623
(0.487) (0.634)
Educated (%) x ∆PalExp 0.119 0.251
(0.164) (0.176)
Other controls YES YES YES YES YES YES YES YES
Observations 569 569 569 569 569 569 569 569
Notes: The dependent variable is the number of fatalities from politically motivated violence during demonstration (Palestinians killed by Israelis)
from the outbreak of the Second Intifada (September 28, 2000) until December 2004 in locality. See table A.7 for the definitions of the
independent variables. Other controls include ∆Palestinian exports1996–99 and all the controls in table B.15, column 2. The regressions are estimated
using a Negative Binomial model. Robust standard errors (clustered at the locality level) are reported in parentheses. The symbols *, **, ***
represent statistical significance at the 10, 5, and 1 percent levels.
Table B.21 Instrumenting Palestinian Exports through Exogenous Shocks
(1) (2) (3) (4) (5) (6)
Dep. variable ∆Palestinian exports96–99 ∆Palestinian exports to Israel96–99 ∆Palestinian exports to RoW96–99
∆Chinese exports to −0.002*** −0.006*** −0.002*** −0.007*** −0.000** 0.002***
the world but Israel (0.001) (0.000) (0.001) (0.001) (0.000) (0.000)
∆World exports 0.001*** 0.001*** 0.001*** 0.001*** −0.000*** 0.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
∆MFN tariff 0.801*** 1.303*** −0.502***
(0.130) (0.155) (0.034)
Other controls YES YES YES YES YES YES
Observations 569 569 569 569 569 569
R-squared 0.557 0.501 0.585 0.492 0.687 0.445
Note: The regressions are estimated using an OLS model. Robust standard errors (clustered at the locality level) are reported in parentheses. The
symbols *, **, *** represent statistical significance at the 10, 5, and 1 percent levels. Other controls include all the controls in table B.15, column 2.
Notes
1. This means that conflict onset is coded as zero in the nonconflict years, 1 in the first
year of conflict and missing in the following years of conflict. This is a departure from
much of the previous literature, which uses a dummy for the first year of conflict (and
zero otherwise) to measure conflict onset. Ending would take the value of zero during
the conflict; the value of 1 in the first year of peace and missing in the other years.
2. In particular Bazzi and Blattman (2014) and Bruckner and Ciccone (2010) use
j ∈[0; 2].
3. King and Zeng (2001) define “rare events” dependent variables as those for which the
number of zeros are larger than that of ones by at least an order of magnitude in the
dozens.
4. King and Zeng (2001) also propose a rare logit estimator to correct the rare event bias.
Unfortunately the lack of convergence in the estimation of our model does not allow
us to implement that estimator.
5. Within the negative binomial model option in STATA, we use the population-
averaged (PA) option that relaxes the assumption of independence of Csrt to allow
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
134 Estimation Methodology and Empirical Results
for different correlations over time of the conflict. The relaxation of this assump-
tion is useful for the purpose of our analysis, given that conflict in 2001 may not
be correlated with conflict in 2003 and 2010 the same way. The other two
options, random effect (RE) and fixed effect (FE), are attractive but do not ade-
quately capture our data and impose additional structure on the data that cannot
be validated. Also, they attempt to model over-dispersion rather than capturing
fixed effects at the state level and may have convergence issues given the size of
our data.
6. We also tried a dummy for whether the state has more than one dominant ethnic
group, but its effect was never significant in explaining conflict.
7. In addition, an inspection of the consumption and production data from the house-
hold survey suggests that none of the states is large enough to substantially influence
the prices of the top-produced and -consumed commodities across the entire country.
This adds confidence to the claim of exogeneity of this instrument.
8. The results do not change without weights. We also experiment with different dis-
tance thresholds, i.e. 3, 4, 6, and 7 hours obtaining similar results (results available
upon request).
9. Indeed, in parallel preliminary work we document the limited pass-through from
international to domestic prices in Nigeria for various agricultural items.
10. Note that all the results are robust to computing the changes over the 1997–1999
period or the 1996–2000 period (results available upon request).
11. This variable is also identical to an indicator variable that equals 1 if the number of
fatalities in the locality is above or equal to the median number of fatalities across
localities and zero otherwise.
12. As a de facto customs union, Palestinian and Israeli goods do not pay any duties to
access the reciprocal markets.
13. Unlike trade data, we cannot sum tariff rates over the sectors to match the 5-digit
SITC level data of MFN tariff with the 2-digit ISIC employment data. Instead we
take the average of tariff rates across 5-digit SITC sectors weighted by Palestinian
total exports in each sector. In this way we ensure that the changes in Israeli tariffs
are weighted by the importance they may have for Palestinian sectoral exports.
14. The results are also robust to including the non instrumented import variables (results
available upon request).
References
Alesina, A., A. Devleeschauwer, R. Wacziarg, S. Kurlat, and W. Easterly. 2003.
“Fractionalization.” Journal of Economic Growth 8 (2): 155–94.
Bazzi, S., and C. Blattman. 2014. “Economic Shocks and Conflict: Evidence from
Commodity Prices.” American Economic Journal: Macroeconomics.
Beck, N., and J. Katz. 2011. “Modeling Dynamics in Time-Series-Cross-Section Political
Economy Data.” Annual Review of Political Science 14: 331–52.
Bruckner, M., and A. Ciccone. 2010. “International Commodity Prices, Growth and the
Outbreak of Civil War in Sub-Saharan Africa.” The Economic Journal 120 (May):
519–34.
Cameron, C., and P. K. Trivedi. 2013. Regression Analysis of Count Data. Cambridge, UK:
Cambridge University Press.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Estimation Methodology and Empirical Results 135
Chaudion, S., Z. Peskowitz, and C. Stanton. 2012. “Beyond Zeroes and Ones: The Effect
of Income on the Severity and Evolution of Civil Conflict.” Mimeo.
Dube, O., and J. Vargas. 2013. “Commodity Price Shocks and Civil Conflict: Evidence
from Colombia.” Review of Economic Studies 80 (4): 1384–421.
Jacoby, H. 2013. “Food Prices, Wages, and Welfare in Rural India.” Policy Research Working
Paper 6412, World Bank.
King, G., and L. Zeng. 2001. “Logistic Regression in Rare Events Data.” Political Analysis
9: 137–63.
Kuhn, P., and N. Weidmann. 2013. “Unequal We Fight: Between- and Within-Group
Inequality and Ethnic Civil War.” Princeton University, Mimeo.
Long, J. S., and J. Freese. 2006. Regression Models for Categorical Dependent Variables Using
Stata. 2nd ed. College Station, TX: Stata Press.
Miaari, S., A. Zussman, and N. Zussman. 2014. “Employment Restrictions and Political
Violence in the Israeli-Palestinian Conflict.” Journal of Economic Behavior and
Organization 101 (May): 24–44.
Montalvo, J. G., and M. Reynal-Querol. 2005. “Ethnic Polarization, Potential Conflict, and
Civil Wars.” American Economic Review 95 (3).
Powell, J. M., and C. L. Thyne. 2011. “Global Instances of Coups from 1950 to 2010:
A New Dataset.” Journal of Peace Research 48 (2): 249–59.
Topalova, P. 2010. “Factor Immobility and Regional Impacts of Trade Liberalization:
Evidence on Poverty from India.” American Economic Journal: Applied Economics
2: 1–41.
World Bank. 2011. World Development Report 2011: Conflict, Security and Development.
Washington, DC: World Bank.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
Environmental Benefits Statement
The World Bank Group is committed to reducing its environmental footprint.
In support of this commitment, the Publishing and Knowledge Division lever-
ages electronic publishing options and print-on-demand technology, which is
located in regional hubs worldwide. Together, these initiatives enable print runs
to be lowered and shipping distances decreased, resulting in reduced paper
consumption, chemical use, greenhouse gas emissions, and waste.
The Publishing and Knowledge Division follows the recommended standards
for paper use set by the Green Press Initiative. Whenever possible, books are
printed on 50 percent to 100 percent postconsumer recycled paper, and at least
50 percent of the fiber in our book paper is either unbleached or bleached using
Totally Chlorine Free (TCF), Processed Chlorine Free (PCF), or Enhanced
Elemental Chlorine Free (EECF) processes.
More information about the Bank’s environmental philosophy can be found
at http://crinfo.worldbank.org/wbcrinfo/node/4.
Trading Away from Conflict • http://dx.doi.org/10.1596/978-1-4648-0308-6
The world is getting richer, but fragile countries and territories are not. By 2015, most of the poor will live in
these economies, most of which have been or still are affected by civil conflicts. Trading Away from Conflict
finds evidence of an important factor affecting conflict: trade, and in particular, trade policy.
On the one hand, trade brings resources and creates jobs, and in turn, those add to stability. On the other
hand, trade itself is also volatile and can increase the value of predating exportable commodities. This book
tries to help policy makers in fragile countries to use trade to reduce the risk of conflict. To do so, it lays out
a framework for studying the relationship between trade and conflict, reviews previous studies, revisits and
expands the cross-country evidence, and analyzes new country case studies. On the basis of the analysis,
it offers trade-related policy directions to reduce this risk in fragile economies.
The results provide convincing evidence that trade and trade policy have a large impact on the risk and
intensity of conflict. In particular, higher minerals and oil exports raise the risk of conflict because there is a
high incentive to fight over their control. This is seen in a number of countries—and was the case in Nigeria
until the government reached an agreement that directed some resource revenues to militants.
The results in the book also show that trade-related increases in real incomes and employment reduce the
risk and intensity of conflict. A high volume of trade with neighbors is likely to shorten conflicts and make
them less severe. These effects are particularly strong in more ethnically divided and unequal settings, in
countries and territories with a recent history of or ongoing conflict, such as the West Bank and Gaza,
as well as in those with low levels of accountability and transparency.
Trading Away from Conflict will help readers understand part of the dynamics of some current conflicts such
as those in Nigeria and the Republic of South Sudan, quantify the effect of various trade shocks on conflict
across and within economies, identify some conditions when this effect is larger, and evaluate measures to
use trade to increase conflict resilience in fragile settings.
ISBN 978-1-4648-0308-6
SKU 210308