Mthuli Ncube, Issa Faye, Audrey Verdier-Chouchane (Eds.) - Regional Integration and Trade in Africa-Palgrave Macmillan UK (2015)

Regional Integration and Trade in Africa
Regional Integration and

Trade in Africa
Edited by
Mthuli Ncube
Senior Research Fellow, Blavatnik School of Government, University of Oxford (on leave
from the African Development Bank)
Issa Faye
Manager of the Research Division, African Development Bank
Audrey Verdier-Chouchane
Chief Research Economist, African Development Bank
© African Development Bank 2015
Individual chapters © Respective authors 2015
Foreword © Gilbert Mbesherubusa 2015
Softcover reprint of the hardcover 1st edition 2015 978-1-137-46204-1
All rights reserved. No reproduction, copy or transmission of this
publication may be made without written permission.
No portion of this publication may be reproduced, copied or transmitted
save with written permission or in accordance with the provisions of the
Copyright, Designs and Patents Act 1988, or under the terms of any licence
permitting limited copying issued by the Copyright Licensing Agency,
Saffron House, 6–10 Kirby Street, London EC1N 8TS.
Any person who does any unauthorized act in relation to this publication
may be liable to criminal prosecution and civil claims for damages.
The authors have asserted their rights to be identified as the authors of this work
in accordance with the Copyright, Designs and Patents Act 1988.
First published 2015 by
PALGRAVE MACMILLAN
Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited,
registered in England, company number 785998, of Houndmills, Basingstoke,
Hampshire RG21 6XS.
Palgrave Macmillan in the US is a division of St Martin’s Press LLC,
175 Fifth Avenue, New York, NY 10010.
Palgrave Macmillan is the global academic imprint of the above companies
and has companies and representatives throughout the world.
Palgrave® and Macmillan® are registered trademarks in the United States,
the United Kingdom, Europe and other countries.
ISBN 978-1-349-69031-2 ISBN 978-1-137-46205-3 (eBook)
DOI 10.1057/9781137462053
This book is printed on paper suitable for recycling and made from fully
managed and sustained forest sources. Logging, pulping and manufacturing
processes are expected to conform to the environmental regulations of the
country of origin.
A catalogue record for this book is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Regional integration and trade in Africa / [edited by] Mthuli Ncube, Chief Economist
and Vice President, African Development Bank Group, South Africa, Issa Faye,
Manager of the Research Division, African Development Bank, Audrey
Verdier-Chouchane, Chief Research Economist, African Development Bank.
pages cm
Summary: “Regional integration has been adopted by many African governments
as a response to difficulties facing national markets in landlocked countries.
However, despite the aspiring targets they have experienced setbacks due to poor
implementation. This book reviews the current trends, challenges and potential
benefits of regional integration and trade in Africa, in a context of rising
interest and momentum among African leaders on the topic. This study explores the
many facets of regional trade, its implementation gaps and positive effects
through empirical studies covering various African sub-regions and using different
theoretical approaches. It analyses trade performance against full potential, its
mitigated impact on the competitiveness, industrialization and economic
diversification of African countries as well as the much debated effect of monetary
unions on regional trade in the region. It provides valuable policy recommendations
aimed at stimulating the debate among the government, private sector and
development community on the ways to promote regional trade for Africa’s
economic development. Regional Integration and Trade in Africa examines various
aspects of regional integration and trade in Africa, and will be of interest to
scholars, students, researchers, policy-makers, and professionals” —
Provided by publisher.
1. Africa—Commercial policy. 2. Africa—Commerce. 3. Africa—Economic
integration. 4. Africa—Foreign economic relations. I. Ncube, Mthuli, 1963–
II. Faye, Issa, 1972– III. Verdier-Chouchane, Audrey, 1973–
HF1611.R437 2014
337.1 6—dc23 2014028329
Contents
List of Figures vii
List of Maps viii
List of Tables ix
Foreword by Gilbert Mbesherubusa xii
Acknowledgments xiv
Notes on Contributors xv
List of Acronyms xxi
Introduction: Understanding Africa’s Regional Trade 1

Mthuli Ncube, Issa Faye and Audrey Verdier-Chouchane
Part I Intra-African Trade Performance and

Regional Integration
1 Competitiveness and Integration through Trade in
CEMAC Countries: Comparative Advantage and
Contribution to the Trade Balance 17
Joseph Parfait Owoundi
2 Economic Integration, Trade Facilitation and Agricultural

Exports Performance in ECOWAS Sub-Region 31
Wumi Olayiwola, Evans Osabuohien, Henry Okodua and
Oluyomi Ola-David
3 The Impact of Trade Liberalization on Export Growth and

Import Growth in Sub-Saharan Africa 47
Lanre Kassim
Part II Measuring Trade Potential: The Gravity

Model Approach
4 Market Integration in the ECCAS Sub-Region 71
Désiré Avom and Mouhamed Mbouandi Njikam
v
vi Contents
5 Regional Integration and Trade in Sub-Saharan Africa,

1993–2010: An Augmented Gravity Model 91
Edris Seid
Part III Industrialization Strategy and the Issue of

Deindustrialization
6 Malawi’s Trade Policies, Market Structure and
Manufacturing Performance, 1967–2002 111
Hopestone Kayiska Chavula
7 The Impact of South-South and North-South Trade on

Industrialization in Africa 125
Henri Atangana Ondoa and Henri Ngoa Tabi
Part IV Impact of Currency Union on Trade

8 Impact of Monetary Unions on Trade: The Case of
WAEMU 153
Ibrahima Camara
9 Trade Agreements and Flows in ECOWAS: Is a Single

Currency the Determining Factor? 173
Benjamin Ndong and Sokhana Diarra Mboup
10 A DSGE Model of Trade and Risk-Sharing Effects of

Currency Union on Economic Integration of the CFA Zone 197
Thierry Kame Babilla
Conclusion: Enhancing Intra-African Trade through Regional

Integration 218
Index 229
Figures
I.1 Africa’s membership of selected regional economic

communities in 2014 3
I.2 Africa’s top trade partners in 2012 (%) 4
I.3 Intra-African trade between 1995 and 2011 4
I.4 Intra-African trade by selected RECs over 2007–2011
(% total trade) 5
2.1 ECOWAS’ global exports and imports, 1999–2009 (USD
millions) 32
2.2 ECOWAS’ intra-regional exports and imports as percent
of its global trade 32
3.A.1 Definition and sources of variables 65
4.1 Total trade within communities, percentages of total
trade 75
4.2 Average trade within communities 1995–2010 (USD
millions) 76
4.3 Exports and imports within ECCAS, averages in USD
thousands 77
4.4 Total ECCAS trade with partners (USD thousands) 78
5.1 Shares of goods exports by world region, 2000–2010 93
6.1 Price–cost margins and manufacturing output growth
rate trends – 1967–2002 115
6.A.1 Clemente-Montañés-Reyes unit-root test with double
price–cost margin’s mean shifts, AO model 123
8.1 Trade balance 2000–2009 for WAEMU’s countries (CFAF
millions) 160
8.2 WAEMU’s marginal effects on exports 165
8.3 Share of exports within WAEMU and to Africa 167
9.1 Export trends by region 176
9.2 Import trends by region 176
10.1 Impulse response functions to a negative productivity
shock in CEMAC (dashed) and WAEMU (solid), years 211
10.2 Impulse response functions to a negative monetary
10.3 Impulse response functions to negative public-spending
C.1 Africa’s comparative performance 226
vii
Maps
C.1 Regional development corridor projects in Africa (2013) 222
viii
Tables
1.1 The global competitiveness of CEMAC and WAEMU

countries 25
1.2 Revealed comparative advantages (RCA) of CEMAC
countries in CEMAC and WAEMU markets 26
1.3 Contribution to trade balance indicators (CTBi) of
CEMAC countries in the CEMAC regional market 27
2.1 Effects of ECOWAS regional integration and trade
facilitation on agricultural trade, 2003–2008 39
2.A.1 Table of summary statistics 43
3.1 Average export and import growth before and after
liberalization 50
3.2 Trade liberalization and export growth (random effects) 53
3.3 Trade liberalization and export growth (GMM) 55
3.4 Timing impact of trade reforms on export growth 56
3.5 Trade liberalization and import growth (fixed effects) 58
3.6 Trade liberalization and import growth (GMM) 59
3.7 Timing impact of trade reforms on import growth 60
3.8 Comparing results 63
3.A.1 Classification of countries 64
3.A.2 Import growth regression with foreign aid 66
4.1 ECCAS exports and imports in 2009, percentages by
product 77
4.2 Gravity model estimations (1995–2010) 82
4.3 Average trade flows within ECCAS 1995–2010 (USD
millions) 84
4.4 Trade potential within ECCAS 1995–2010 (USD
millions) 86
4.A.1 List of sample countries 88
5.1 Africa’s goods exports by type, 2001–2010 (%) 93
5.2 African economic communities’ import sources,
2000–2010 average (%) 94
5.3 Intra-regional as percent of global exports by regional
economic community, in USD millions, 2001–2010 95
5.4 Variable descriptions 101
5.5 PPML estimation of augmented traditional gravity
model (N = 30,503, paired 1,793) 102
ix
x List of Tables
5.6 PPML estimation of Anderson–van Wincoop gravity

model with time-varying fixed effects for countries and
year (N = 31,477) 103
5.A.1 Countries included in the gravity model by regional
economic community 105
6.1 Malawi’s trade policies and manufacturing growth,
1964–2004 112
6.2 Effects of trade policies and market structure on
manufacturing price–cost margins, 1967–2002 118
6.A.1 Variable definitions 122
7.1 Descriptive statistics 129
7.2 Contribution of external and internal factors to
industrialization 133
7.3 Contribution of external and internal factors to
development of manufacturing industries 135
7.4 Contribution of trade with countries in transition to
development of manufacturing industries in Africa 138
7.5 Contribution of trade with countries in transition to
industrial development in Africa 140
7.A.1 Added value of manufacturing industries and trade
openness in Africa 144
7.A.2 Industrial value added and trade openness in Africa 146
8.1 Coefficients of WAEMU and WAEMU1 variables, from
cross-section analysis 166
8.A.1 Coefficients from cross-analyzes, 1990–2005 169
9.1 African exports and imports, 2011 177
9.2 Intra-African trade by economic zone in 2011,
percentages of total exports 179
9.3 Comparative trend of intra-zone trade as percent of
total exports of the economic zones concerned 180
9.4 ECOWAS’ Least Developed Countries (LDC), export
effort and ratio of exports to GDP 180
9.5 Expected signs 185
9.6 Results of estimation on cross-sectional data with OLS
on averages 186
9.7 Results of estimation using panel data 2003–2012 187
9.A.1 Trend in exports of goods by region 190
9.A.2 Trend in imports of goods by region 191
9.A.3 Intra-African trade by economic zone in 2011,
percentages of total imports 192
List of Tables xi
9.A.4 Exports and imports of ECOWAS’ Least Developed

Countries (LDC), in millions of USD 193
10.1 Calibration of household, producer and retailer
program parameters 207
10.2 Calibration of commercial bank and central bank
program parameters 208
10.3 Calibration of government program plus shock
parameters 208
Foreword
Regional integration has been at the core of Africa’s development efforts

since independence some 50 years ago. For the continent’s leaders, eco-
nomic integration was the means by which Africa would address its
economic frailty and stake claim to the global trade and investment
worthy of its size and strategic importance in the world. The Abuja
Treaty of 1991, for example, embodies such sentiments. While the ben-
efits of integration are lauded at summits, the process of integration
has been very slow and complicated by many obstacles, ranging from
language barriers to differences in political economy. In a rapidly glob-
alizing world, Africa is once again taking a close look at what can be
done to advance its integration agenda and especially to expand trade
in goods and services.
This book argues that regional integration and associated trade flows
are crucial for inclusive growth and development. But trade-boosting
economic integration will not happen by itself. It will be important
to establish flourishing economies with supportive infrastructure at
national levels and for countries to be willing to cede some of their
sovereignty on issues such as regional trade to supranational entities.
This has proved difficult to implement in the past. In other parts of the
world, the road to integration and greater trade volumes was paved with
the establishment of effective customs unions and common markets.
Free trade areas are important for attracting foreign direct investment,
promoting technology transfers and enhancing regional competition.
At the same time, they help to dismantle tariff and non-tariff barriers,
reduce distortions and raise economic welfare.
Africa’s existing regional economic communities have made some
progress in encouraging countries to allow freer movement of goods
and services and to harmonize trade policies, but the continent as a
whole still records extremely low levels of intra-regional trade. Africa
will have to do much more to harness the opportunities offered by
regional integration for trade and development.
This book is based on the presentations made at the African Economic
Conference on Regional Integration in Africa, held in Johannesburg
in November 2013, and sponsored by the African Development Bank,
Economic Commission for Africa and the United Nations Develop-
ment Programme. The book attempts to place the challenges of regional
xii
Foreword xiii
integration and trade within broader policy and practical contexts, look-
ing carefully at some of the solutions available. The chapters provide
comparative discussions of the challenges and look at the many oppor-
tunities that Africa’s abundant resources and large population provide
for trade expansion and welfare enhancement. A good measure of both
theoretical and empirical tools is used in the presentations in each
chapter.
I am confident that the book will have important impacts on debates
about economic integration and trade in Africa by offering pointers to
the solutions needed to bring countries, and the continent as a whole,
up to speed in meeting the aspirations of its leaders and people. I would,
therefore, like to recommend the book to policymakers, researchers and
practitioners in economic development and other readers interested in
development issues of the day as an additional tool in their endeavors
to ensure that Africa is not left behind in this era of rapid globalization.
Gilbert Mbesherubusa
Former Vice President, Infrastructure Private Sector and Regional
Integration African Development Bank
Acknowledgments
This book brings together a selection of articles presented at the 2013

African Economic Conference (AEC) in Johannesburg, South Africa,
under the theme ‘Regional Integration in Africa’. The articles have
been reviewed and selected by the African Economic Research Con-
sortium and enriched by comments from peer-reviewers, discussants
and rapporteurs at the conference. The book’s authors thank Steve
Kayizzi-Mugerwa, Director of the Development Research Department
and Acting Chief Economist and Vice President of the African Develop-
ment Bank (AfDB), for his comments and reviewing, as well as Charlotte
Karagueuzian for her excellent research assistance. Our appreciation also
goes to Andrew Lawday and Rick Wicks, who helped to edit the English
version of the book. We are very grateful to the AfDB Language Services
Department for assisting in the translation of articles from French to
English.
The AEC 2013 provided a unique opportunity for policymakers and
researchers, including from the diaspora, to interact and debate crit-
ical policy issues affecting Africa. This conference would not have
been possible without the full support and leadership of the African
Development Bank’s President, the Executive Secretary of the Economic
Commission for Africa and the Administrator of the United Nations
Development Programme. It also depended largely on the many staff
members of these institutions who dedicated their time and energy
to making the conference a success. Their contributions are greatly
appreciated.
Not least, we are extremely grateful to the researchers, who submitted
their papers and made presentations at the AEC 2013, for their high-
quality contributions. We sincerely hope this publication will further
stimulate debate and shape policy around regional integration and trade
in Africa.
xiv
Contributors
Désiré Avom is Professor of Economics at the University of Yaoundé

II, Cameroon. He is the Director of the Laboratoire d’Analyse et de
Recherche en Economie Appliquée (LAREA) and the Director of the
Centre d’Etudes et de Recherche en Economie et Gestion (CEREG) at the
University of Yaoundé II. He is an expert in the areas of International
Economics and Economic Integration. Désiré Avom has published sev-
eral articles in international journals such as Ecological Economics, Revue
d’Economie Politique and Revue Française d’Economie.
Thierry Kame Babilla is a researcher at the Centre d’Etudes et de

Recherche en Economie et Gestion (CEREG), University of Yaoundé II,
Cameroon. His areas of interest are Macroeconomics and Economic
Modeling. From 2014, he has been a member of the scientific commit-
tee of the Association Femmes et Développement (AFED) and a Research
Fellow at the Partnership for Economic Policy (PEP) Network and at
the African Economic Research Consortium (AERC). Babilla was laure-
ate of the International Development Research Center Doctoral Research
Awards in Economics in 2012.
Ibrahima Camara is currently working as a statistician in the depart-

ment of research, at Guinea’s central bank. He completed a training
program in statistics and Economics at the Ecole Nationale Supérieure
de Statistique et d’Economie Appliquée (ENSEA) in Abidjan in 2011. His
research interests include African regional integration, education, civil
registration and vital statistics, remittances from migrants, industrial-
ization and the ITCs (Information Technology and Communications)
sector.
Hopestone Kayiska Chavula works as an economist in the

Macroeconomic Policy Division (MPD) of the Economic Commission
for Africa (ECA) in Addis Ababa, Ethiopia. Before 2013, he worked at
the same organization’s ICT, Science and Technology Division (ISTD).
Before that, he was a lecturer in both Economics and Computer Sci-
ences at the University of Malawi. Chavula holds a PhD in Economics,
an MA in Economics, and a Bachelor’s in both Economics and Computer
Sciences.
xv
xvi Notes on Contributors
Issa Faye is Manager of the Research Division of the African Develop-

ment Bank (AfDB). Prior to joining AfDB, he worked as an economist
at the World Bank in the Research Department and the Rural Develop-
ment Network of the Africa Region. Prior to this, he was a lecturer at the
University of Auvergne/CERDI, France, from where he received a PhD
in Economics, and a researcher at the Centre National de la Recherche
Scientifique (CNRS).
Lanre Kassim is a research student at the University of Kent, United

Kingdom, working on the impact of trade liberalization in Sub-Saharan
Africa. Specifically, he analyzes how the adoption of freer trade has
affected export growth, import growth, balance of payments and tax
revenue in the region. His primary research interests are the eco-
nomics of development, growth models and international trade. His
secondary research interests are applied Micro-econometrics and Pub-
lic Economics. Kassim obtained his Master’s in international finance
and economic development with a distinction from the University of
Kent and his Bachelor’s in Economics from the University of Lagos
Nigeria.
Gilbert Mbesherubusa was Vice President for Infrastructure, Private Sec-

tor and Regional Integration of the African Development Bank between
2012 and 2014. He has held, throughout his career at the AfDB, several
key positions. Prior to joining the Bank in 1982 as a civil and transport
engineer, Gilbert concurrently worked as an Engineer/Director and as a
part-time lecturer at the University of Burundi between 1978 and 1982.
He holds a certificate in transport economics from Université d’Eté de
Lyon (France) and is a graduate of civil engineering from the Catholic
University of Louvain (Belgium).
Sokhna Diarra Mboup is a PhD student in Economics and a member of

the research team in Economics at Gaston Berger University, Saint-Louis,
Senegal. Her thesis is on the impact of the Economic Community of
West African States (ECOWAS)’s trade reforms on economic growth,
poverty and well-being. She obtained her Master’s and Bachelor’s in
Economics from the same university. Her research interests are in
international economics, international trade and regional integration,
development economics and poverty reduction.
Mthuli Ncube is on leave from the post of Chief Economist and Vice
President of the African Development Bank. He is now Senior Research
Notes on Contributors xvii
Fellow at the Blavatnik School of Government in Oxford, United King-

dom. He holds a PhD in Mathematical Finance from Cambridge Univer-
sity, UK. Before joining the Bank, he was Dean at the University of the
Witwatersrand (Wits) and before that was Dean and Professor of Finance
at Wits Business School in South Africa. He has published widely in the
area of finance and economics. He also has extensive experience as an
investment banker.
Benjamin Ndong is Professor, Head of the Department of Economics

(2009–2013), Coordinator of the Master’s program in Rural Develop-
ment and Cooperation and Coordinator of the Master’s program in
Banking and Financial Econometrics at the University Gaston Berger of
Saint Louis, Senegal. He holds a PhD in Economics from the University
of Franche-Comté (France). He has been a visiting scholar at the Inter-
national Monetary Fund (IMF) and a member of the following research
teams and networks: the Africa Growth Institute in Cape Town, South
Africa; the Centre for the Study of African Economies at the Univer-
sity of Oxford, UK; and the African Economic Research Consortium in
Nairobi, Kenya.
Mouhamed Mbouandi Njikam is a PhD candidate in Economics at

the University of Yaoundé II in Cameroon. He is a researcher at the
Laboratoire d’Analyse et de Recherche en Economie Appliquée (LAREA).
His areas of research include international trade, economic geography,
economic integration and sustainable development. He has partici-
pated in the African Economic Conference in Johannesburg in October
2013 and the biannual workshop at the African Economic Research
Consortium (AERC) in December 2013 in Nairobi, Kenya.
Henri Ngoa Tabi is Head of the Department of International Economics

at the Faculty of Economics and Management, University of Yaoundé
II, Cameroon. He obtained his PhD at the University of Versailles in
1999. Prior to this, he headed the Centre d’Etudes et de Recherche en
Economie et Gestion (CEREG) at the University of Yaoundé II. As part
of his mission in CEREG, he led several research projects in partner-
ship with international organizations, including the United Nations
Development Programme, the Economic Commission for Africa, the
International Monetary Fund, the United Nations Conference on Trade
and Development, the African Economic Research Consortium, Trust
Africa, the International Development Research Centre, Afrobarometer,
Rio Tinto Alcan and the African Development Bank.
xviii Notes on Contributors
Henry Okodua is a lecturer in Economics at Covenant University,

Nigeria, and has conducted research on African economics and develop-
ment since 2004. Previously, he worked as a regional consultant to the
Economic Community of West African States (ECOWAS) Commission
and contributed to the production of a technical report on Invest-
ment Climate Monitoring Indicators in ECOWAS countries. Okodua
holds a PhD in Economics and is a member of the Nigerian Eco-
nomic Society (NES), African Econometric Society (AES) and United
Nations Conference on Trade & Development (UNCTAD) virtual
institute.
Oluyomi Ola-David is a PhD candidate and lecturer in the Depart-

ment of Economics and Development Studies, Covenant University,
Nigeria. Her research investigates the productivity effects of foreign
direct investment in developing countries, and her interests include
the economics of innovation, industrial learning and sustainable devel-
opment. During a recent internship at the UN-Habitat, she coordi-
nated discussions on urbanization, cities and sustainable development
toward the post-2015 development framework. She is an alumnus of
Brown University’s International Research Institute on Technology and
Entrepreneurship Management and a member of the Association for the
Advancement of African Women Economists.
Wumi Olayiwola is Principal Programme Officer of the Economic

Policy Analysis Unit, Department of Macroeconomic Policy and Eco-
nomic Research, ECOWAS Commission, Abuja, Nigeria. He is also a
faculty member of Trade Policy Training Centre in Africa (TRAPCA),
Arusha, Tanzania. Prior to this, he was a consultant to the World
Bank and a Research Fellow to the African Economic Research Consor-
tium (AERC) and Trade Policy Research Training Programme (TPRTP),
Nigeria. Olayiwola holds a PhD in Economics with specialization
in International Trade, Development Economics and Macroeconomic
Analysis.
Henri Atangana Ondoa is a lecturer at the Faculty of Economics and

Management at the University of Yaoundé II, Cameroon, and Researcher
at the Centre for Studies and Research in Economics and Management of
the same university. He obtained his PhD in Economics in 2009 at the
University of Yaoundé II in collaboration with the African Economic
Research Consortium (AERC). He has participated in several impact
Notes on Contributors xix
studies with the following agencies: United Nations Development Pro-

gramme, International Development Research Centre, Economic Com-
mission for Africa, Rio Tinto Alcan. He has realized several internships in
the following international organizations: the International Monetary
Fund, the United Nations Conference on Trade and Development and
the Center for Research in Economic Analysis. He has attended several
international conferences.
Evans Osabuohien is currently a research fellow at the German Devel-

opment Institute, Bonn, with a fellowship awarded by the Alexander
von Humboldt Foundation. He holds a PhD in Economics from
Covenant University, Nigeria. His other awards include the Swedish
Institute’s Guest Doctoral Fellowship, the Council for the Development
of Social Science Research in Africa’s (CODESRIA) grant for his PhD
thesis, First Prize for FLACSO-WTO Award and research grants from the
Centre for Economic and Policy Research (CEPR) as well as from the
African Economic Research Consortium (AERC). He is currently editing
a book provisionally titled In-Country Determinants and Implications of
Foreign Land Acquisitions.
Joseph Parfait Owoundi works at the Ministry of Economy, Plan-

ning and Regional Development in Cameroon. He holds a Master’s in
Demography and Statistics and is a PhD candidate at the Institute of
Population Research and Training (IFORD), University of Yaoundé II,
Cameroon. His areas of interest include population and development,
maternal and child health and the economics of human resources. He
has published in several journals such as Statistics and Computing and
IFORD. He has participated in international conferences, such as the
annual meeting of the Population Association of America in 2014 and
the annual African Economic Conference in 2013.
Edris H. Seid is a junior research fellow at the regional independent

think tank, the Horn Economic and Social Policy Institute, in Addis
Ababa, Ethiopia. Prior to this, he was a lecturer at Addis Ababa Uni-
versity. He has a double MSc in Economic Development and Growth
from the University of Warwick, the United Kingdom, and Lund Uni-
versity, Sweden. He has also MSc in International Economics from Addis
Ababa University. His research interests are international trade, regional
economic integration, development economics, poverty and economic
growth with particular emphasis on Africa.
xx Notes on Contributors
Audrey Verdier-Chouchane is Chief Research Economist in the Devel-

opment Research Department of the African Development Bank. Prior
to joining AfDB, Audrey taught Macroeconomics and Development Eco-
nomics at the University of Nice, France, from where she received a PhD
in Economics. She has mainly published on the analysis and measure-
ment of poverty and inequalities.
Acronyms
ACP African, Caribbean and Pacific

ADMARC Agricultural Development and Marketing Corporation
AfDB African Development Bank Group
AG agricultural exporters
AMU Arab Maghreb Union
ASEAN Association of Southeast Asian Nations
BCEAO Banque centrale des Etats de l’Afrique de l’Ouest
(Central Bank of West African States)
BEAC Banque des Etats de l’Afrique Centrale (Bank of Central
African States)
BOP balance of payments
BRIC Brazil, Russia, India and China
CEMAC Communauté économique et monétaire de l’Afrique
centrale (Economic and Monetary Community of
Central Africa)
CEN-SAD Community of Sahel-Saharan States
CET common external tariff
CFA Communauté Financière Africaine (African Financial
Community)
CFAF Franc de la Communauté Financière Africaine (African
Financial Community Franc)
CIS Commonwealth of Independent States
CMA Common Monetary Area
COMESA Common Market for Eastern and Southern Africa
CTB contribution to trade balance
DC developing countries
DfID Department for International Development
DOT Data on Trade [statistics]
DPM dynamic panel model
DRC Democratic Republic of the Congo
DSGE Dynamic Stochastic General Equilibrium
EAC East African Community
ECCAS Economic Community of Central African States
ECOWAP ECOWAS Agricultural Policy
ECOWAS Economic Community of West African States
EMU European Monetary Union
xxi
xxii List of Acronyms
EPA Economic Partnership Agreement

ETLS ECOWAS Trade Liberalization Scheme
EU European Union
FDI Foreign Direct Investment
FE fixed effects
FGLS Feasible Generalized Least Squares
FTA free trade area
GATT General Agreement on Tariffs and Trade
GCI Global Competitiveness Index
GDP gross domestic product
GFCF Gross Fixed Capital Formation
GMM generalized method of moments
H-O Heckscher-Ohlin
ICT Information and Communications Technology
IGAD Intergovernmental Authority on Development
IIAG Ibrahim Index of African Governance
IMF International Monetary Fund
ITC International Trade Centre
LDC Least Developed Countries
Mercosur Mercado Comum Sudamericano (South American
Common Market)
NAFTA North American Free Trade Agreement
NAG non-agricultural exporters
OCA optimum currency area
OECD Organization of Economic Cooperation Development
OLS ordinary least squares
PIDA Programme for Infrastructure Development in Africa
PPML Pseudo-Poisson Maximum Likelihood
RCA revealed comparative advantages
RE random effects
REC regional economic community
RI regional integration
RTA regional trade agreement
SACU South African Customs Union
SADC Southern African Development Community
SAP structural adjustment programs
SEZ Special Economic Zone
SSA Sub-Saharan Africa
TF trade facilitation
TP trade potential
TSLS two-stage least squares
List of Acronyms xxiii
UNCTAD United Nations Conference on Trade and Development

UNECA United Nations Economic Commission for Africa
WAEC West African Economic Community
WAEMU West African Economic and Monetary Union
WAMZ West African Monetary Zone
WDI [World Bank’s] World Development Indicators
WEF World Economic Forum
WGI [World Bank’s] World Governance Indicators
WTI [World Bank’s] World Trade Indicators
WTO World Trade Organization
Introduction: Understanding
Africa’s Regional Trade
Introduction
Africa has registered growth rates of more than 5% annually over the
past decade (2004–2014), with a third of African countries averaging
6.5% growth per year (African Development Bank et al., 2014). This
current growth pattern has brought Africa to a turning point, if not a tip-
ping point, in its development trajectory. A more globalized world econ-
omy has brought increasing functional and spatial specialization, as well
as growing interdependence, and vulnerability of national economies
to external shocks. Meanwhile, a process of de-industrialization in the
developed world has been paralleled by an increasing competitiveness
of developing countries in manufacturing. The changing structure of
the global economy requires Africa to integrate in order to survive.
Translating economic gains into sustainable and shared growth requires
Africa to connect its markets, deepen Regional Integration (RI) and
enhance national competitiveness (World Economic Forum et al., 2013).
Embedding RI into Africa’s development agenda, whether in the form of
preferential trading area or economic and monetary union, will move
African economies to the next stage of competitiveness and integration
in the global economy as befits the continent’s rise.
RI is a means to achieving greater global integration, especially for
countries which cannot compete on a global scale by themselves such as
small and landlocked economies (DfID, 2011). As Sub-Saharan Africa is
fragmented into 48 small states, a large number of which are landlocked
countries, RI should, in theory, be an important tool for industrializ-
ing Africa through economic diversification, export competitiveness and
facilitated access to foreign markets. According to the theory of com-
parative advantages (Ricardo, 1817), greater openness is associated on
1
2 Introduction
average with faster economic growth and increased productivity, espe-

cially through the attraction of foreign direct investments and transfer
of technology. Further, the diversification in trading partners makes
countries more able to cope with external shocks, such as demand and
price fluctuations (World Bank, 2010).
However, countries face a timeframe challenge, as most of the benefits
of RI accrue in the long term while its costs are felt in the short term,
for example, through reduced fiscal revenues from diminished trade
taxes. In recent economic literature, the benefits of RI are also shown
to depend strongly on the implementation of complementary policies
at the domestic level, in favor of macroeconomic stability, trade-related
infrastructure and good governance (OECD/WTO, 2013).
The chapters in this book intend to offer a better understanding of
Africa’s regional trade and policy options for enhancing intra-African
trade through RI. In Part I, the authors look into the issues of com-
petitiveness and trade performances within specific regional economic
communities (RECs) such as the Economic and Monetary Community
of Central Africa (CEMAC) (Chapter 1) and Economic Community of
Western African States (ECOWAS) (Chapter 2), and more broadly in
Sub-Saharan Africa (Chapter 3). Although Africa’s RECs have pushed
the continent toward further RI at the legal and institutional level, the
authors remind us that there is a lack of political commitment from
national leaders, while overlapping RECs and multiple memberships
present a challenge to implementing RI (see Figure I.1).
In Part II, the authors use gravity models to measure trade poten-
tial in the Economic Community of Central African States (ECCAS)
sub-region (Chapter 4) and in four different RECs (Chapter 5). They
confirm that African RECs have underperformed so far in trade achieve-
ments. They also demonstrate that intra-African trade would increase
with countries’ economic diversification and structural transformation.
However, such causality does not seem to hold in reverse, as RI has
not caused economic diversification or industrialization. In Part III,
the authors analyze the impact of RI on industrialization, highlight-
ing a possible de-industrialization effect. Chapter 6 considers the case
of Malawi’s manufacturing performances. Chapter 7 examines indus-
trialization scenarios for 47 African countries depending on the trade
partners. In Part IV, the authors consider the mitigated effects of
currency unions on trade in the West African Economic and Mon-
etary Union (WAEMU) (Chapter 8), in ECOWAS (Chapter 9) and in
the CFA Franc Zone (Chapter 10). The following sections offer an
outline of the main issues discussed, in line with the structure of
the book.
Mthuli Ncube et al. 3
CEN-SAD
COMESA
UMA Seychelles
Algeria Comoros Congo, Dem. Rep*.
Egypt Ethiopia Madagascar
Malawi
IGAD Mauritius
Libya Zambia
Mauritania Djibouti Zimbabwe
Morocco Eritrea
Tunisia Uganda
South Sudan Swaziland
Sudan
Kenya Burundi* SACU
Rwanda Botswana
Lesotho
Somalia Namibia
WAEMU South Africa
EAC Tanzania
Benin
Burkina Faso
Cote d’lvoire Gambia, The
Guinea-Bissau Ghana Mozambique
Mali Guinea Sao Tome Burundi*
Niger Liberia and Principe SADC
Senegal Nigeria
Togo Sierra Leone Angola
Congo,
Central African Rep. Cameroon Dem. Rep*.
Chad Congo (Brazzaville)
Equatorial Guinea
Gabon
CEMAC
Cape verde
ECOWAS ECCAS
Figure I.1 Africa’s membership of selected regional economic communities in

2014
Notes: ∗ Burundi and the Democratic Republic of Congo are the only countries which appear
twice on the Figure as they are both COMESA and ECCAS countries.
Source: Authors.
1 Intra-African trade performance and regional integration
Although intra-African trade has been on the rise, the bulk of

Africa’s trade remains with high-income countries, and particularly the
European Union (Figure I.2). As shown in Figure I.3, intra-African trade
totaled USD 130.1 billion in 2011, rising from USD 32 billion in 2000.
In nominal terms, this represents a rise by a factor of 4.1 over the period
2000–2011, but the growth has been mostly driven by price increases.
In volume, intra-African trade rose only by a factor of 1.7 (UNECA,
2013). UNECA (2013) also indicates that informal intra-African trade,
which is not reported in the official figures, is relatively large. For
instance, informal cross-border trade in the Southern African Devel-
opment Community (SADC) is estimated to reach USD 17.6 billion
per year.
Figure I.3 also demonstrates that during periods of world recession
(1998–2001 and 2009), intra-African trade declined, which implies glob-
alization has a significant impact on Africa’s total trade performance.
According to UNECA (2013), intra-African trade share is higher among
non-fuel exporters (16.3% of their total trade in 2007–2011) than
4
2.4 5.6
3.1
5.3
8.9
61.1
13.6
EU United States South Africa Other

China India Japan
Figure I.2 Africa’s top trade partners in 2012 (%)

Source: Authors based on UNCTADstat database.
140
120
100
$ billions
80
60
40
20
0
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
Intra-African trade Intra-African exports Intra-African imports
Figure I.3 Intra-African trade between 1995 and 2011

ECCAS
AMU
IGAD
COMESA
CEN-SAD
ECOWAS
EAC
SADC
0 5 10 15 20 25
With Africa With same REC
Figure I.4 Intra-African trade by selected RECs over 2007–2011 (% total trade)
among fuel exporters (5.7%), supporting the argument that economic

diversification could deepen regional trade in Africa.
With the exception of ECCAS, RECs usually trade within their bloc
rather than with other African RECs (Figure I.4). This confirms that for-
mation of RECs has facilitated trade among member countries, but this
also varies significantly between national economies. Considering the
proportion of intra-African trade to the Gross Domestic Product (GDP),
only Lesotho, Swaziland and Zimbabwe had a ratio above 50% over
2007–2011 (UNECA, 2013).
The Economic and Monetary Community of Central Africa (CEMAC)
countries are compared to WAEMU in Chapter 1. Here Joseph Parfait
Owoundi evaluates CEMAC countries’ competitiveness by analyzing
their trade structure, revealed comparative advantage and contribution
to trade balance in intra-CEMAC trade. Owoundi finds that CEMAC
countries are overall competitive in the region, especially in the export
of primary products such as oil, natural gas and wood. Thus com-
petitiveness is improving, but institutional integration lags without
political will and commitment at the national level to implement
regional arrangements and suggests weakness in long-term competi-
tiveness. Institutional integration is required for the free movement of
goods and people, which is the key to international competitiveness.
Considering agricultural products, African countries have a marginal
share of agricultural world trade because they face high external tariffs;
6 Introduction
farmers in high-income countries earn subsidies to stay competitive in

the world market and because local production is limited by insufficient
land and agricultural productivity, as reflected in the continent’s ris-
ing share of world agricultural imports. However, UNECA (2013) notes
that unexploited opportunities in intra-African trade is evident in agri-
culture. Only 14.8% of African agricultural imports took place within
the continent during 2007–2011, although agriculture is the backbone
of many African countries, including the ECOWAS economies. Indeed,
Africa’s growing population and increasing urbanization, as well as
Africa’s significant share of the labor force in agriculture, all necessitate
improvement in African economies’ agricultural performance.
In Chapter 2, Wumi Olayiwola, Evans Osabuohien, Henry Okodua
and Oluyomi Ola-David analyze how trade facilitation and economic
integration, more generally, have affected agricultural exports within
the ECOWAS sub-region. Using General Method of Moments (GMM)
with instrumental variable (IV) estimation on panel data from 15
ECOWAS members during 2003–2008, they find the level of trade facil-
itation in ECOWAS to be below world average – although new and
improved infrastructure such as the Internet had facilitated trade. Agri-
cultural production, in which there has been sustained growth, affected
agricultural exports, but countries with more bureaucratic processes had
greater costs associated with trade. More efforts are needed to improve
agricultural production and facilitate trade and economic integration
in ECOWAS. Incentives are needed for full-fledged implementation by
member states of the ECOWAS Agricultural Policy (ECOWAP) and the
ECOWAS Trade Liberalization Scheme (ETLS).
The destinations of African exports vary significantly by commodity.
For instance, more than two-thirds of Africa’s exports are still directed
toward high-income countries and mainly consist of primary commodi-
ties. In contrast, African countries predominantly import manufactured
goods from high-income countries, of which 60% is machinery and
equipment. This dependence on developed markets makes African
countries highly vulnerable to external shocks, undermines their trade
balance, hinders the creation of linkages with domestic and regional
economies and limits their prospects of diversifying production and
exporting higher value-added manufactures.
In Chapter 3, using panel data, Lanre Kassim investigates the
effects of trade liberalization on imports and exports across 28 Sub-
Saharan African countries during 1981–2010, deriving liberalization
dates from a careful examination of trade policy reviews. Consistent
with the findings of other studies on less- and least-developed countries,
liberalization was found to have increased the growth of not only

exports but also imports by about two percentage points more, con-
tributing to deteriorating trade balances in the post-liberalization era.
Liberalization raised the price elasticity (but not the income elasticity)
of demand for both imports and exports. However, the price elasticity of
demand for exports is low as exports consist mainly of agricultural com-
modities, while the price elasticity of demand for imports is inelastic as
Sub-Saharan Africa has become a food importer.
2 Measuring trade potential: The gravity model
The gravity model offers a more rigorous method of assessing intra-

African trade performance and analyzing the effect of trade policies
on trade flows. Pioneered by Tinnbergen (1962), the gravity model
approach has been a success at the empirical level. Applying the orig-
inal Newtonian gravity equation to trade flows rests on the assumption
that the bigger and the closer are two geographical entities, the stronger
their attraction. For example, authors assume that trade between two
countries is proportional to the two entities’ GDP and negatively related
to the distance between the two entities. They also add dummy variables
either facilitating or restricting the flow of trade such as population size,
openness – measured by the ratio of total exports and imports to real
GDP – bilateral exchange, price variables, monetary and fiscal variables,
shared official language or membership of regional trade agreements.
In the case of Africa, authors have confirmed that the current intra-
African trade remains low relative to its potential. Assuming that policy
can influence some of the variables and increase trade, economists have
also come up with trade policy recommendations.
In Chapter 4, Désiré Avom and Mouhamed Mbouandi Njikam use a
gravity model to estimate intra-regional trade flows among the ten coun-
tries of the Economic Community of Central African States (ECCAS)
during 1995–2010. The results are then used in a simulation to iden-
tify their potential trade, with emphasis on the removal of tariff and
non-tariff barriers. Commercially, ECCAS countries were found to be
poorly integrated, with domestic commerce and out-of-Africa trade pre-
dominating over trade between member states, while low levels of
industrialization and diversification greatly reduced potential trade.
Indeed, despite the rise in regional trade agreements in Africa, some
recent studies using the gravity model approach have come up with
the conclusion that African RECs have not performed well (Geda, 2002;
Warin et al., 2009). The continent has the lowest trading volume
8 Introduction
compared to other trading blocs worldwide, including developing areas

as East and South Asia as well as Latin America (World Bank, 2010).
In Chapter 5, Edris Seid confirms this conclusion. Despite the exis-
tence of several RECs in Africa, intra-regional trade remains very low
compared to trading blocs in Europe, Asia and Latin America; this
is partly due to inefficient and costly transport systems and compli-
cated customs procedures. To uncover the main factors behind this
low level of intra-regional trade, and the role of four RECs (Com-
mon Market for Eastern and Southern Africa [COMESA], ECOWAS, the
Inter-Governmental Authority on Development [IGAD] and SADC) in
promoting it, the intuitive theoretical gravity model of Anderson-van
Wincoop is applied to panel data from 1993 to 2010. The traditional
gravity model variables (GDP, population, distance, border, language
and colonial links) as well as bilateral real exchange rates and differences
in preferences among trading partners are found to be important factors
determining bilateral trade flows. However, the effect of the RECs on
bilateral trade is found to be mixed: SADC and ECOWAS seem to have
created trade in the Vinerian sense (that is, trade creation and diver-
sion), while IGAD had a statistically non-significant positive coefficient
and COMESA had an implausible negative coefficient.
3 Industrialization strategy and de-industrialization
A key challenge for Sub-Saharan African countries is to reverse their

strong reliance on exports from extractive resource industries, which
provides few incentives to develop forward and backward linkages with
their national and regional economies or to diversify their industrial
export base.
In addition to the deterioration of trade balance in Africa, RI raises
the question of appropriate trade and industrialization strategy at the
country level. Since the 1950s, the debate has been on the adoption of
inward versus outward development policies (Dornbusch, 1992; Rodrik,
1992). Inward policies, which predominated in the 1970s, encourage
protectionism through high tariffs and quotas on imports, on the
assumption that a country will develop its manufacturing through
indigenous ‘learning by doing’ and specialization in technologies for
which the country is endowed with resources. Ultimately, some of the
manufactured items will become competitive with world prices and the
country will be able to export them. At the end of the 1960s, Ethiopia,
Ghana, Nigeria and Zambia pursued import substitution strategies
(Kirkpatrick, 1987).
Malawi also experienced an import substitution strategy over the

period 1967–1981 as described in Chapter 6 by Hopestone Kayiska
Chavula. The author looks into the effects of Malawi’s trade policies and
market structure on manufacturing performance since independence.
Using the Kaluwa and Reid (1991) modeling framework and firm-level
panel data, Chavula finds that market concentration had a positive
effect on the price–cost margins of manufacturing regardless of trade
policy, while scarcity of factor inputs (raw materials, skilled labor and
finance capital) had negative effects. Increased trade also had negative
effects on price–cost margins. Import intensity leads to a reduction in
price–cost margins as firms reduce prices in order to remain competitive
while Malawian exports have not been competitive enough to influ-
ence international prices, especially in the post-liberalization period
(1994–2004), leading to reductions in profitability. Tariff rates had no
significant effect across trade policies.
In contrast, most of the literature on import substitution strategy
acknowledges its empirical failures (Dornbusch, 1992). Most of the
industries remained largely inefficient, costly and unable to grow and
create forward and backward linkages at the national level. As opposed
to import substitution strategies, proponents of free trade gained the
upper hand in the 1990s and advocated export promotion strategies
of industrialization on the grounds that it will increase competition,
thereby promoting innovation, efficiency (through better resource allo-
cation), economies of scale and economic growth and avoids the dis-
torting price and cost effects of protectionism (Bhagwati, 1988). Part of
the literature on the outward industrialization approach focuses on the
means to redress free market failures (e.g. lack of technology transfer).
It also recommends government intervention to influence the type of
goods exported, as the country seeks to develop its comparative advan-
tage in more sophisticated and higher-value items, requiring higher
technology and skills (Rodrik, 2007).
According to Barro and Lee (2002) and Vreeland (2003), the nar-
row export base and weakening industrial sector of Sub-Saharan African
countries are related to a radical change in African trade policies in the
1980s toward economic liberalization. In line with the structural adjust-
ment programs of the World Bank and the short-term macroeconomic
stabilization programs of the International Monetary Fund, this change
in trade policy was aimed at attracting private investment in Sub-
Saharan Africa. However, countries did not experience import substitu-
tion industrialization for a sufficiently long period and trade liberaliza-
tion prematurely exposed their infant industries to global competition.
10 Introduction
The resulting poor performance contrasts the performance of other

developing economies in Latin America or Asia, which experienced
import substitution for a longer period of time and whose manufactur-
ing sector consequently was able to grow by up to 27% of total valued
added for the period 2000–2008 (Sundaram et al., 2011).
Further, adjustment programs led Sub-Saharan countries to reduce
their government expenditure, which in turn affected public investment
in infrastructure and harmed local industries as a result of the reduction
in investment in private manufacturing (UNCTAD, 2001). These adverse
consequences of liberalization explain how the average share of man-
ufacturing fell in value added, from 12% in 1980 to 9.6% in 2011 in
African developing countries (World Bank, 2013).
More recent trade theories highlight the benefits of economies of
scale, human capital investments and externalities and favor selective
tariff protection (Burton, 1998). Although countries which experienced
the most successful industrial development have been East Asian export
promoters, it should be acknowledged that they are not pure free traders
as they simultaneously adopted an import substitution strategy for
some of their industries (Lewis and Kallab, 1986; Wade, 1990). Besides,
free trade opponents argue there is a rising protectionism in high-
income countries against agricultural and manufactured goods from the
developing world. Market failures reduce opportunities for developing
countries to export more refined and higher-valued items. They also
acknowledge that the comparative advantage of developing countries is
their primary resources; as a consequence, adopting an outward policy
would work against diversification and industrialization of the national
economy.
This advent of de-industrialization is developed in Chapter 7. Using
a dynamic panel, Henri Atangana Ondoa and Tabi Henri Ngoa study
the effect of North-South and South-South trade on industrialization
in Africa. The econometric analysis on 47 African countries for the
period 1971–2010 indicates that trade with industrialized countries has
contributed to industrial development in Africa, but trade with emerg-
ing Asian economies has been fatal, with de-industrialization reaching
alarming proportions. As a result of increased competition, intra-African
trade has also led to de-industrialization in some African countries. Tar-
iffs were found to stimulate industrialization in some countries but
de-industrialization in others. It is concluded that African countries
should maintain trade links with industrialized countries to promote
industrial development through technology transfer while promoting
RI and supporting manufacturing.
4 Impact of currency unions on regional trade
A ‘full currency union’ means a group of countries that shares a

common currency and a common central bank and monetary pol-
icy. Countries engaging in a monetary union often adopt additional
trade arrangements aimed at simplifying bilateral trade. These may
include the formation of a customs union, where regional member
countries raise common external tariffs while freeing internal trade;
a free trade area, where members raise different external tariffs while
freeing internal trade; or a common market, where a customs union
is established with the free movement of labor and capital between
member countries.
Assessing the effects of currency unions on trade in Africa has become
crucial in the context of recent currency union projects. Since its
inception, ECOWAS has committed to forming a free trade area and
a currency union; but the absence of progress led six member coun-
tries (Gambia, Ghana, Guinea, Liberia, Nigeria and Sierra Leone) to
propose the WAMZ (West African Monetary Zone). In October 2013,
these countries called for the adoption of a new shared currency, the
Eco, by 2015. The other member countries of ECOWAS currently using
the West African CFA Franc within the WAEMU are to join the WAMZ
by 2020, switching from the CFA Franc to the Eco. This agreement
does not, however, include the Central African member countries of
the CEMAC, which also have a currency union under the CFA Franc.
In November 2013, five member countries of the EAC (Uganda, Rwanda,
Burundi, Tanzania and Kenya) committed to form a currency union
in the next ten years, revitalizing the currency union based on the
East African Shilling that has been used between Kenya, Tanzania and
Uganda since their independence. Africa also counts other forms of cur-
rency cooperation, which are less integrated: The Common Monetary
Area (CMA) links the currencies of Lesotho, Namibia and Swaziland
to the South African Rand and may expand in the future to include
other SADC countries. The core objective of the SADC is enhancing
trade and strengthening structural economic policies, but increasing
consideration is given to monetary coordination among members under
the frame of the CMA. This consists of a formal exchange rate union
between countries, with separate currencies and very narrow margin
fluctuation of exchange rates. The use of another country’s currency has
also temporarily existed in Africa, as Botswana used the South African
Rand from independence until 1976, and Eritrea used the Ethiopian Birr
from independence until 1997.
12 Introduction
Theoretical and empirical evidence suggests that currency unions

expand markets for goods and services, including financial products,
and that such increased trade in turn reinforces RI. The positive impact
of currency union on trade is the potential savings on transaction
cost, which in turn depends on the volume of trade among member
countries and the elimination of exchange rate volatility. Neverthe-
less, production reallocation from high-cost to more efficient low-cost
member countries could be the result of monetary unions. Benefits
for the client country are assumed to increase with the size of the
anchor country (Alesina and Barro, 2002). However, intra-regional trade
is very low in Africa, so prospects for transaction cost savings are
limited.
In the empirical literature focusing on the impact of currency unions
on trade, the standard methodology used is the gravity equation model,
which is augmented with a dummy variable indicating whether the
countries share the same currency or not (Rose, 2000; Glick and Rose,
2002). Alesina et al. (2002) give a good overview of the findings of these
empirical studies, pointing out the positive effects of currency unions
on bilateral trade. Rose (2000) shows that bilateral trade is 200% more
important between two countries using the same currency than between
two countries using different currencies.
In Chapter 8, Ibrahima Camara analyzes cross-sectional data from the
15 ECOWAS countries during 1990–2005 to determine the effect of the
WAEMU on trade. She uses a year-by-year gravity model equation to
compare the structure of regional trade before and after the implemen-
tation of the WAEMU. Though decreasing over time, the effect appears
to have been positive within the Union. Diversion of exports to the
detriment of other ECOWAS countries was not statistically significant
for any year of the study. The WAEMU economies appear to be highly
dependent on few exports, especially in their trade with Côte d’Ivoire.
The Union would therefore benefit from diversification, which would
strengthen complementarity among members.
In Chapter 9, on the same subject, Benjamin Ndong and Sokhana
Diarra Mboup look at the extent to which regional trade agreements
combined with a monetary union have a significant positive effect
on bilateral trade in the WAEMU and the larger ECOWAS zone. Have
ECOWAS and WAEMU increased trade among their members? A grav-
ity model – controlling for determinants of trade such as geogra-
phy, language, economic structure and monetary policies – is ana-
lyzed using data for all ECOWAS members except Guinea-Bissau for
the period 2003–2012. As expected, trade increased with the size of
economies (and with a common border or language) and decreased
with distance and isolation. In general, both regional trade agreements

and common currencies increased trade. Specifically, membership of
ECOWAS increased trade and the membership of WAEMU had a positive
effect though not statistically significant.
In Chapter 10, in an effort to learn from the CFA Zone (14 West and
Central African countries) and to offer policy recommendations valid
more widely, Thierry Kame Babilla applies a Dynamic Stochastic Gen-
eral Equilibrium (DSGE) model, paying special attention to differences
in trade and risk-sharing. Integration of the CFA Zone, he finds, was
not reinforced by the use of a common currency. Savings are insuffi-
cient to intensify cross-border risk-sharing and financial asymmetries
have led to amplification of national differences, so business cycles did
not synchronize. Policymakers could accelerate real integration by pro-
moting risk-sharing institutions and mechanisms to offset the effects of
asymmetric shocks. Since savings are the main channel for coping with
shocks in the CFA zone, regional policies to increase savings should
be adopted. Current and future African currency unions should also
develop and facilitate access to regional financial markets.
References
African Development Bank [AfDB] (2014), OECD Development Centre and
United Nations Development Programme, African Economic Outlook, Paris:
OECD Publishing.
Alesina, A. and Barro, R. J. (2002), ‘Currency Unions’, Quarterly Journal of
Economics, vol. 117, no. 2, pp. 409–436.
Alesina, A., Barro, R. J. and Tenreyro, S. (2002), ‘Optimum Currency Areas’,
National Bureau of Economic Research, Working Papers, no. 9072, NBER, Inc.
Barro, R. J. and Lee, J.-W. (2002), ‘IMF Programs: Who Is Chosen and What Are
the Effects?’ Working Paper,r no. 8951, National Bureau of Economic Research.
Bhagwati, J. (1988), Protectionism, Cambridge, MA: MIT Press.
Burton, H. (1998), ‘Reconsideration of Import Substitution’, Journal of Economic
Literature, vol. 36, pp. 306–936.
Department for International Development [DfID] (2011), ‘Regional Integration
and Trade in Sub-Saharan Africa’, Trade and Investment Analytical Papers, DfID:
Department for Business Innovation and Skills.
Dornbusch, R. (1992), ‘The Case for Trade Liberalization in Developing Coun-
tries’, Journal of Economic Perspectives, vol. 6, pp. 69–85.
Geda, A. (2002), ‘Debt Issues in Africa: Thinking beyond the HIPC Initiative
to Solving Structural Problems’, Working Paper Series UNU-WIDER Research
Paper, World Institute for Development Economic Research.
Glick, R. and Rose, A. K. (2002), ‘Does a Currency Union Affect Trade?’ European
Economic Review, vol. 46, pp. 1125–1151.
Kaluwa, B. M. and Reid, G. C. (1991), ‘Profitability and Price Flexibility in Manu-
facturing for a Developing Country’, The Journal of Industrial Economics, vol. 39,
no. 6, pp. 689–700.
14 Introduction
Kirkpatrick, C. (1987), ‘Trade Policy and Industrialization in LDCs’ in Norman,

Gemmell (ed.), Surveys in Development Economics, Oxford: Blackwell, pp. 71–72.
Lewis, J. P. and Kallab, V. (1986), Development Strategies Reconsidered. New
Brunswick, NJ: Transaction.
OECD/WTO (2013), Aid for Trade at a Glance 2013: Connecting to Value Chains,
WTO/OECD Publishing.
Ricardo, D. (1817), On the Principles of Political Economy and Taxation, London:
John Murray, Albemarle-Street.
Rodrik, D. (2007), ‘Normalizing Industrial Policy’, Harvard University, August.
Available at: http://ksghome.harvard.edu/
Rodrik, D. (1992), ‘The Limits of Trade Policy Reform in Developing Countries’,
Journal of Economic Perspectives, vol. 6, pp. 87–105.
Rose, A. K. (2000), ‘One Money, One Market: The Effect of Common Currencies
on Trade’, Economic Policy, CEPR & CES & MSH, vol. 15, no. 30, pp. 7–46.
Sundaram, J. K., Schwank, O. and von Arnim, R. (2011), ‘Globalization
and Development in Sub-Saharan Africa’, DESA Working Paper, r no. 102,
ST/ESA/2011/DWP/102, February.
Tinnbergen, J. (1962), Shaping the World Economy; Suggestions for an International
Economic Policy, New York: Twentieth Century Fund.
United Nations Conference on Trade and Development [UNCTAD] (2001),
Economic Development in Africa: Performance, Prospects and Policy Issues.
UNCTAD/GDS/AFRICA/2, Geneva.
United Nations Economic Commission for Africa [UNECA] (2013), Economic
Development in Africa Report – Intra-African Trade: Unlocking Private Sector
Dynamism, Addis-Ababa: UNECA.
Vreeland, J. R. (2003), The IMF and Economic Development, t New York: Cambridge
University Press.
Wade, R. (1990), Governing the Market: Economic Theory and the Role of Government
in East Asian Industrialization, Princeton, NJ: Princeton University Press.
Warin, T., Wunnava, P. V., Tengia, O. and Wandschneider, K. (2009), ‘Southern
African Economic Integration: Evidence from an Augmented Gravity Model’,
IZA Discussion Paper,r no. 4316, July.
World Bank (2010), World Development Indicators, Washington: World Bank.
World Bank (2013), World Development Indicators, Washington: World Bank.
World Economic Forum, World Bank and AfDB (2013), The Africa Competitiveness
Report,
t Geneva: World Economic Forum.
Part I
Intra-African Trade Performance
and Regional Integration
1
Competitiveness and Integration
through Trade in CEMAC
Countries: Comparative Advantage
and Contribution to the Trade
Balance
Joseph Parfait Owoundi
Introduction
Given ideal conditions of fair competition without political pressure,

David Ricardo1 showed that all countries, even the least competitive,
have an interest in joining the game of international trade and special-
izing in types of production where their relative advantage is greatest,
or their disadvantage smallest.
International trade generally revolves around theories of compara-
tive advantage and country specialization (Smith, 1776; Ricardo, 1817).
Considering other factors, such as factor endowments, Ohlin (1933)
and Heckscher (1949) sought to explain further why states engage in
trade. Samuelson (1949) complemented their work with a mathematical
formulation, but debunked their predictions.
Over the past two decades, globalization has brought about a tech-
nological and organizational revolution that is disrupting production
and trade systems. Increased product differentiation, logistical advances,
economies of scale and new strategies are increasing national and inter-
national competition. In parallel, the opening of borders and trade
liberalization are impacting world trade (Desmas, 2005).
The recent proliferation of regional trade agreements (RTAs), and in
particular customs unions, is leading governments to consider harmo-
nizing their border measures to foster regional integration and economic
competitiveness.
17
18 Competitiveness and Integration through Trade in CEMAC
Competitiveness is of prime importance to analyzing the macro-

economic performance of countries. For a country and its trade partners,
competitiveness is used to compare characteristic factors of an economy
and for analyzing trends in international trade. Further, the develop-
ment of ‘competitiveness poles’ (regional integration vectors) is essential
for boosting productive capacity and attracting investment.
Economic and Monetary Community of Central Africa (Communauté
Economique des Etats de l’Afrique Centrale – CEMAC) member states
are gradually aligning with the Treaty of Rome, which places a pre-
mium on the ‘standstill rule’, which is designed to check the erection
of barriers to inter-state trade. These members have set a number of
objectives:
• strengthening the competitiveness of member states’ economic and

financial activities, within the framework of an open and competitive
market and a rationalized and harmonized legal environment;
• creating a common market among member states, based on the free
movement of persons, goods, services, capital; the right of estab-
lishment for self-employed or employed persons; and a common
external tariff and a common trade policy.
Given these two objectives, and current competitiveness indicators

for CEMAC members, it is important to ask: Are the countries of
the sub-region competitive compared to other countries, particularly
those of West African Economic and Monetary Union (WAEMU)? How
well do CEMAC countries perform in intra-industry and inter-industry
trade?
This study assesses the competitiveness of CEMAC countries within
sectors and between sectors in the CEMAC and WAEMU regional mar-
kets. It seeks to define the profile of competitive countries in the
sub-region and evaluate the performance of exports and imports within
CEMAC and outside it (in the WAEMU zone).
First, data and indicators from the World Economic Forum (2011,
2012 and 2013 reports) are used to build a competitiveness profile of
CEMAC states. Second, the revealed comparative advantages (RCA) of
CEMAC countries is calculated to determine competitiveness scales and
thresholds. Since RCAs focus solely on exports (unilateral vision of inter-
national trade), the contribution to trade balance indicator (CTBi) is
also calculated, in line with RCA logic to allow for analysis of import
structure. These analyses use CEMAC and WAEMU data on exports and
imports.
Joseph Parfait Owoundi 19
1 Definition of concepts
Economic competitiveness is the ability of a business, an economic sec-

tor or a territory (country, economic area, etc.) to sell and supply one
or more tradable goods or services in a given market in a situation of
competition over the long term.
According to the OECD, competitiveness is ‘the extent to which a
country can, under free and fair market conditions, produce goods and
services which meet the test of international markets while simultane-
ously maintaining and expanding the real incomes of its people over the
long term’. So competitive countries are those that sell comparatively
more than others at home and abroad (that is, exports) and increase
wealth at home over the long term.
At the microeconomic level, the competitiveness of a company is its
ability to hold a strong position in a market. At the macroeconomic
level, the competitiveness of a national economy is the ability of its
productive sector to meet domestic and foreign demand, ultimately for
the purpose of raising living standards for its citizens. A national econ-
omy’s market share can also be measured (national exports compared
to a worldwide total or to a limited group of comparator countries) and
offers an excellent indicator of competitiveness.
Revealed comparative advantages

RCAs are used to identify the main characteristics of inter-sector spe-
cialization. For a given sector, they consist of dividing its percentage of
exports in relation to total exports for the country by the percentage of
the sector’s exports in relation to total exports for a reference area, for
example the world.
Trade integration
In economics, integration refers to the strategy of bundling activities
within a company. This enables a company to master the technical,
commercial or financial expertise to boost productivity and benefit
from synergy effects. ‘Horizontal’ and ‘vertical’ integration can also be
differentiated.
Vertical integration is when such bundling concerns the different
stages of production and distribution for the same type of goods or ser-
vices. Horizontal integration (or horizontal concentration) is when a
company develops its business at the same level of the value chain as its
products. The goal is to spread costs over a larger quantity of products,
to limit contingencies through substitute products or simply to avoid

competition.
By definition, economic integration is the most developed form of
a regional trade agreement: a single market reinforced with common
economic and social policies.
2 Regional integration in the CEMAC zone
‘Where there is trade, there are customs. The purpose of trade is to

export and import goods for the State and the purpose of customs is
to collect duties on the same exports and imports, also for the State’
(Montesquieu, The Spirit of Laws, 1748).
In Africa, growth rates are impressive. Increased foreign direct invest-
ment (FDI) has fuelled a rebounding economy over the past decade.
Between 2001 and 2008, Africa’s gross domestic product (GDP) recorded
a 5.9% annual average growth. While Africa is less closely linked to the
global financial markets than other regions of the world, it has not been
spared the impact of global financial crisis. Indeed, it threatens Africa’s
progress of recent years in economic development, reform of public
policies and institutions and, in particular, poverty alleviation.
In the CEMAC zone, development of the countries’ key strengths and
prospects depend on their geographic location, which offers a compara-
tive advantage in the sub-region. Their different potentials lie in forestry,
agriculture, mining and a diverse ecosystem. However, the ability to
exploit this potential is hampered by persistent constraints, mainly
related to poor basic infrastructure and weak governance.
Further, all CEMAC countries except Central African Republic (CAR)
are producers of oil, which is a key driver of growth. A breakdown of
GDP by sector shows that extractive industries alone represented 37%
of CEMAC GDP in 2005 (UNECA, 2006). This type of specialization
can only foster vertical integration, as absorption of these products by
countries of the sub-region is nil.
The countries in the sub-region that are most highly specialized in
commodities are Cameroon and the CAR, both showing Grubel–Lloyd
indices of above 70%. Those with medium specialization are Gabon,
Congo, Equatorial Guinea and Chad, showing indices of between 50
and 70%. The first two countries are relatively diversified economies,
specialized in a few products (Cameroon in agricultural, forestry and oil
products and CAR in grain, livestock and minerals). The second group
of countries, exclusively specialized in one sector, are overspecialized
mono-product economies (Mignamissi, 2008).
According to Joseph Djaowe, FDI in Central Africa (Cameroon,

Congo, Gabon, Chad and Equatorial Guinea) is mainly oriented toward
oil and gas. Four variables are statistically significant to attract FDI to the
CEMAC zone: the real GDP growth rate; the current account balance
(economic risk variables); external debt (financial risk variable); and
political stability (governance variable). CEMAC countries should pur-
sue appropriate policies that ensure improved communications infras-
tructure, business climate and good governance, to benefit from these
financial flows which are deemed more cost-effective (Djaowe, 2009).
The CEMAC economies are outward looking. Their products derive
primarily from the agricultural or mining sectors, with undiversified
exports. Thus for Gabon, Equatorial Guinea and Congo, timber and
petroleum represent over 80% of net exports. Cameroon is in a relatively
more favorable exports situation with five leading export products; oil,
cocoa, timber, coffee and aluminum together represent 70% of its total
exports (Godonou Dossou, 2008).
In the Economic Community of West African States (ECOWAS),
regional integration during 1995–2004 was rather negative in terms
of food production. The share of intra-regional trade in overall food
product imports continued to decline, from 11 to 7%. The weakness
of intra-ECOWAS regional food trade is attributable to an insufficiently
protective external tariff; but it is also the result of inadequate free trade
within the zone (Gallezot, 2006). Bilateral food trade within ECOWAS is
mostly complementary trade (inter-sector), and the Grubel–Lloyd index
is relatively low at 0.12 (Gallezot, 2006).
Countries that use the Central African Franc (Franc des Communautés
Financières d’Afrique – FCFA) face macroeconomic imbalances, stem-
ming from non-diversification in export supply structure, absence of
market power and impact of external effects on commodity mar-
kets. These imbalances will also worsen with the gradual depletion
of non-renewable resources, on the one hand, and the establishment
of Economic Partnership Agreements (EPAs), on the other (Moussone,
2010).
The production of raw materials by all these countries does not fos-
ter trade among them, as is the case for manufactured goods (trade in
similar goods). Even the 1994 CFA Franc devaluation did not benefit the
sub-region in terms of creation or diversion of trade flows – as is often
the case in a monetary adjustment within an economic and monetary
union (Moussone, 2010).
Moreover, African states are engaged in a reform process that
projects a vision of economic emergence by 2050. In this context,
competitiveness is crucial for upgrading infrastructure, creating jobs and

improving the business climate. It may only be possible and profitable
if there is understanding and harmonization of procedures for the free
movement of goods and people.
3 Methodology and data
The data used are derived from the trade statistics of the Bank of Central
African States (Banque des Etats de l’Afrique Centrale – BEAC) and Cen-
tral Bank of West African States (Banque Centrale des Etats de l’Afrique
de l’Ouest – BCEAO) member countries, and downloadable from the
websites of these two institutions. The years 2002 and 2008 are cho-
sen for the harmonization of indicators and consideration of the effects
of the global economic and financial crises (2008), respectively. In the
absence of a comprehensive database on sectors, the available data show
exports and imports of goods and non-factor services and petroleum
products.
This study relies on calculation of comparative advantages in terms
of trade in CEMAC countries, which are considered as determining for
international trade. This is based on the difference between comparative
costs in a number of countries, as each country finds it advantageous to
specialize in and export goods for which it has the greatest comparative
advantage and import other goods from its partners (Lassudrie-Duchêne
and Ünal-Kezenci, 2001). However, it is difficult to measure the com-
parative advantage directly, so a large number of empirical studies
use the RCA indicator to measure the international specialization of
countries.
The RCA concept, introduced by Balassa in 1965, holds that inter-
national trade in goods reflects cost differentials between countries,
and therefore shows their comparative advantages. So observations of
business performance can enable measurement of revealed comparative
advantage. The better a country’s relative performance in trading a spe-
cific good, the greater its comparative advantage in producing that good
(Balassa, 1965).
The following formula is used to calculate RCA, which serves as a basis
for comparing the export structure of each country to that of a reference
area (CEMAC and WAEMU).

Xk(i)
Tx(i)
RCAk (i, w) = (1)
Xk(w)
Tx(w)
Here (i, w) stands for a country’s revealed comparative advantage com-

pared to the rest of the world for the product k. Xk(i) is the value of
exports of the good k by the country i to the world w. Xk(w) is the value
of exports of the good k to the world. TX(i) is the total value of exports
of the country i to the world. And TX(w) is the total value of world
exports.
An RCA, calculated for a given country and commodity, and which
is greater than 1 would indicate a comparative advantage of that coun-
try in trade in the commodity concerned. This indicates the share of
exports for such a commodity in total exports of the country considered
is higher than the world average.
Moreover, because RCAs focus solely on exports (unilateral vision of
international trade), another indicator is calculated, the contribution to
trade balance (CTB),2 which is consistent with RCA logic and is used
to analyze imports structure. To this end, if Xr and Mr are exports and
imports of a category r, and X and M, the total exports and imports, the
relative balance of category r is:
( − Mr)
(Xr
Zr = 100 (2)
((X + M)/2
The overall relative balance is calculated as follows:
( − M)
(X
Z = 100 (3)
((X + M)/2
The contribution to trade balance indicator may also be obtained using

the following formula:

((Xr + Mr) ( − Mr)
(Xr ( − M)
(X ((Xr + Mr)
CTBi = Zr − Z = 100 − 100 =
( + M)
(X ( + M)/2
(X ( + M)/2
(X ( + M)
(X
(M. Xr − X. Mr) (M. Xr − X. Mr)
100 ==> CTBi = 100 (4)
( + M)/2)2
((X ( + M)/2)2
((X
The summation of categories corresponding to overall balance can be

done at different levels: total manufactured goods, total goods, all goods
and nonfactor services or all current transactions (in the last case, the
concepts of debits and credits replace those of exports and imports).
If there is overall trade imbalance, the country studied has a compar-
ative advantage in a sector i if the indicator is positive and otherwise,
a disadvantage. But if its overall trade balance is negative, a negative
trade balance of industry i does not necessarily mean a disadvantage.
The country will be considered to have an advantage (disadvantage) in

the industry if the deficit of the industry represents a contribution to
total trade deficit lower (higher) than its contribution to the country’s
overall trade.
In light of the above, three assumptions are tested in the analysis.
First, the greater a country’s RCA in the production of a particular good,
the greater the country’s relative performance in trade in this good. Sec-
ond, the greater a country’s RCA and contribution to trade balance in
the production of a particular good, the more this country is competitive
in terms of this good in intra-industry and inter-industry trade. Third,
the calculation of RCAs of CEMAC countries in relation to CEMAC and
WAEMU trade zones enable confirmation of these assumptions.
4 Results
The competitiveness of CEMAC and WAEMU countries is measured,

using results from the 2011, 2012 and 2013 reports of the World
Economic Forum (WEF).
This measurement takes into account the Global Competitiveness
Index (GCI), presented in the 2013 report of the World Economic
Forum (WEF, 2013). For this purpose, 38 African countries are ranked,
including three CEMAC countries: Gabon, Cameroon and Chad; and
six of WAEMU countries: Benin, Mali, Côte d’Ivoire, Burkina Faso and
Guinea-Bissau. Table 1.1 shows the Global Competitiveness Index (GCI)
rankings of CEMAC and WAEMU countries between 2010 and 2013.
As shown in Table 1.1, two of the three CEMAC countries rank first in
Africa in GCI. Only Chad is among the last two of the ranking. In spe-
cific rankings for economic environment, institutions, infrastructure
and business climate, Cameroon and Gabon also occupy the best posi-
tions for macroeconomic environment (sixth and ninth in the African
rankings, way ahead of WAEMU). Institutional management and busi-
ness climate are listed as the only major weaknesses for the CEMAC
countries.
The 2013 Ibrahim Index of African Governance (IIAG) confirms this
result, as Central Africa ranks fifth among Africa’s five regions in terms
of overall governance and has done so yearly since 2000. Only Gabon,
among the seven central African countries scores a rating higher than
the continental average of 51.6, with an Ibrahim Index of 52.8. None of
the countries in the region is among the continent’s top ten, and five
are in the bottom ten: Congo, Equatorial Guinea, the Central African
Republic, Chad and the Democratic Republic of the Congo (DRC).
Table 1.1 The global competitiveness of CEMAC and WAEMU countries
Country Global competitiveness indices
2010–2011 2011–2012 2012–2013
Ranking Rating Ranking Ranking Ranking Rating Trend
Gabon N/A N/A 99 N/A 99 3.8 New

Cameroon 111 3.6 112 116 112 3.7 ↑
Senegal 104 3.7 117 111 117 3.7 ↓
Benin 103 3.7 119 104 119 3.6 ↓
Mali 132 3.3 128 128 3.4 ≡
Côte d’Ivoire 129 3.3 129 131 3.4 ↓
Burkina Faso 134 3.2 136 133 3.4 ↑
Chad 139 2.7 142 139 3.1 ↑
Guinea Bissau N/A N/A N/A 141 2.9 New
Source: Culled from global competitiveness index rankings of the 2013 World Economic
Forum report.
Accordingly, major reforms are needed for the harmonization of

procedures and the development of growth poles to boost the compet-
itiveness of these countries. This is especially important for the DRC,
which holds a central strategic position in central Africa.
However, in terms of world ranking, countries in Africa in general and
CEMAC and WAEMU in particular occupy the last quarter of the overall
GCI ranking. These countries are not competitive with the rest of the
world. They are only competitive in the African continent.
Competition structure
To describe the structure of competition of CEMAC countries, com-
parative advantages and contributions to trade balance are calculated.
Table 1.2 shows the RCAs of the countries concerned between 2005
and 2008.
This table shows CEMAC countries are competitive in their own mar-
ket when it comes to exports of petroleum products and non-factor
goods and services (RCA = 1). The countries are specialized in these two
sectors, except Chad, which is not specialized in the export of petroleum
products in the CEMAC zone; its comparative disadvantage ranges from
0.07 to 0.03 between 2005 and 2008.
Their competitiveness diminishes slightly from one year to another,
due to a sluggish global economic environment following the economic
and financial crises of 2008 and 2009. Cameroon and Chad lead their
Table 1.2 Revealed comparative advantages (RCA) of CEMAC countries in

CEMAC and WAEMU markets
CEMAC Goods and Petroleum Goods Non-factor

countries non-factor products exportation services
services exportation exportation
exportation
2005 2008 2005 2008 2005 2008 2005 2008
CEMAC Market
Equatorial 1.374 0.959 1.869 1.121 1.437 0.996 0.1439 0.088
Guinea
Cameroon 1.705 1.0925 0.901 0.722 1.475 0.952 6.193 4.372
Gabon 1.434 0.969 1.693 1.090 1.465 0.9912 0.821 0.470
Congo 1.401 0.927 1.791 1.218 1.408 0.9257 1.259 0.961
Chad 1.989 1.328 0.0713 0.0310 2.0172 1.323 1.455 1.317
WAEMU Market
Equatorial 157.84 168.85 1.01 1.094
Guinea
Cameroon 195.92 192.24 0.48 0.705
Gabon 164.75 170.65 0.91 1.064
Congo 160.92 163.15 0.967 1.189
Chad 228.58 232.75 0.038 0.030
Source: Culled from BEAC and BCEAO trade statistics.
neighbors in the export of non-factor goods and services, while Equato-

rial Guinea, Congo and Gabon lead in petroleum products. These results
confirm the socio-economic context for CEMAC outlined above.
Compared to WAEMU countries, CEMAC countries remain more com-
petitive than the latter, because the RCAs become more substantial than
those obtained in the CEMAC zone. Only Equatorial Guinea, Congo
and Gabon are specialized in the export of petroleum products in the
WAEMU market. But all CEMAC countries remain specialized in the
export of non-factor goods and services in the WAEMU zone, with
Cameroon and Chad topping the table.
Contribution to trade balance

To analyze the structure of imports, the CTBi was calculated to mea-
sure the impact of intra-industry trade. Table 1.3 describes the CTBi of
CEMAC countries.
As observed earlier, CEMAC countries are competitive in terms of
imports, with fairly substantial intra-industry contributions. But exports
remain higher than imports regardless of product type, and due to
Table 1.3 Contribution to trade balance indicators (CTBi) of CEMAC countries in the CEMAC regional market
CEMAC Goods and non-factor Petroleum products Goods exportation and Non-factor services
countries services exportation and exportation and importation exportation and
importation importation importation
2005 2008 2005 2008 2005 2008 2005 2008
CEMAC Market
Equatorial 16.04 22.02 −14.931 −8.243 −32.087 −44.059 30.97 30.27
Guinea
Cameroon 8325.62 4487.70 −1148.51 59.006 −16651.24 −8975.42 9474.13 4428.70
Gabon 36.383 24.529 −32.137 −23.270 −72.76 −49.058 68.52 47.799
Congo 56.66 141.92 −46.73 −146.59 −113.33 −283.85 103.40 288.52
Chad −22.64 −203.50 −7292.36 −3738.78 45.28 407.00 7269.72 3535.27
Source: Culled from BEAC trade statistics.

27
unavailability of data on imports, this indicator for the WAEMU zone

could not be calculated.
Cameroon’s contribution to trade balance is by far the most signifi-
cant. Apart from its forestry, agricultural and mining potential as well as
ecosystem diversity, its political situation was characterized by relative
stability in the years 2004–2009, with the exception of 2008 when there
were riots against the high cost of living. By improving the quality of
the macroeconomic framework, it was able to reach the HIPCI decision
and completion point in 2000 and 2006. Due to its geographic location,
Cameroon is a transit zone, which gives it a comparative advantage
in the sub-region. Its population of 19 million represents about 50%
of CEMAC’s total 36 million, and 20% of ECCAS’ population of over
100 million. It accounts for nearly 40% of CEMAC’s GDP. Still, the
free movement of people remains incomplete within CEMAC, insofar as
Gabon and Equatorial Guinea continue to require entry visas for nation-
als of other member countries. This may further explain Cameroon’s
advantage.
In terms of competitiveness, Cameroon is facing high factors of pro-
duction costs, notably due to an energy supply that is below its domestic
needs, low road network density and inefficient port services. This
explains the negative balances obtained for the import of petroleum
products and goods. This situation can be generalized for all other
countries with negative balances.
5 Conclusion
In Africa, the importance of competitiveness is not universally under-

stood and shared. As a matter of urgency, CEMAC countries should
move away from simply supplying raw materials, such as crude oil and
timber, to seeking production system efficiency in markets for goods,
finance and labor and innovation. CEMAC countries are certainly com-
petitive at the CEMAC and WAEMU level, but cannot simply remain
in the last quartile of the world’s competitive economies with an index
score of 3.53 in overall competitiveness.
6 Recommendations
In light of the above, CEMAC countries are recommended to:
• Implement institutional reforms, including ethics and efficiency of

government action, judicial independence, regular payment and
right of ownership;
• Develop transport, electricity and communication infrastructures;

• Create an observatory to monitor the business climate and competi-
tiveness poles;
• Establish an export–import bank to promote the business climate;
and
• Harmonize business law in the CEMAC zone.
Notes
1. In Economics, comparative advantage is the main concept of the traditional
theory of international trade. It was broached by Robert Torrens in 1815 and
demonstrated for the first time by the British economist David Ricardo in
1817 in his On the Principles of Political Economy and Taxation. According to
Paul Samuelson, the 1970 Nobel Prize winner in Economics, it is the best
example of an undeniable but counter-intuitive economic principle.
2. Trade imbalances, whether due to economic activity (e.g. France in the years
1960–1970 recorded a trade deficit through capital goods imports following
growth) or exchange rate variations, can flaw the measurement of compara-
tive advantages. The French Centre for Prospective Studies and International
Information (CEPII) in the 1980s developed a contribution to trade balance
indicator that can correct that distortion (Cf. Gérard Lafay, 1984: Avantages
Comparatifs et Compétitivité. Commerce International, p. 52; Cf. Jean-Luc
tavernier, 1990, Echanges extérieur et avantages comparatifs: la spécialisation
de la France confrontée à celle de ses concurrents, Economie et prévision, N)
94/95.
References
Balassa, B. (1965), Trade Liberalization and Revealed Comparative Advantage, The
Manchester School of Economic and Social Studies, no. 33, May.
Desmas (2005), Analyse comparative de compétitivité: le cas de la filière
tomate dans le contexte euro-méditerranéen. Mémoire de Fin d’Etudes
Diplôme d’Agronomie Approfondie (D.A.A.); option: Politique Economique de
l’Agriculture et de l’Espace.
Djaowe, J. (2009), Investissements Directs Etrangers (IDE) et Gouvernance: les
pays de la CEMAC sont-ils attractifs? Revue africaine de l’Intégration, vol. 3, no. 1,
pp. 67–94.
Gallezot, J. (2006), les enjeux et les marges de manœuvre de la CEDEAO
face aux defis des négociations agricoles; Réseau des organisations paysannes
et de producteurs agricoles de l’Afrique de l’Ouest (ROPPA), Afrique
Nourricière.
Godonou Dossou, J. (2008), forces et enjeux de l’intégration sous régionale:
CEMAC/CEEAC; Intl. Rel., Fmr. Dean/Fac.
Heckscher, E. (1949), ‘The Effect of Foreign Trade on the Distribution of Income’
in Ellis, S., H., and L., M., Metzler (s. dir), (1919), Readings in the Theory of
International Trade, Philadelphia: Blakiston.
Lafay, G. (1984), Avantages comparatifs et Compétitivité. Commerce Interna-
tional, p. 52.
Lafay G., 1990, La mesure des avantages comparatifs révélés, Economie prospective
internationale, no. 41.
Lassudrie-Duchêne, B. and Ünal-Kezenci, D. (2001), L’avantage comparatif,
notion fondamentale et controversée, in L’économie Mondiale, La Découverte,
pp. 90–104.
Mignamissi, M. D. (2008), Réformes macroéconomique et intégration par le
marché dans la CEMAC; Université Yaoundé II – DEA 2008.
Moussone, E. (2010), Insertion des pays de la zone franc africaine dans le com-
merce mondial étude d’une spécialisation appauvrissante et le problème du
financement de l’économie; Documents de travail – no. 231 novembre 2010.
Ohlin, B. (1933), International and Interregional Trade, Cambridge, MA: Harvard
University Press.
Ricardo, D. (1817), ‘On the Principles of Political Economy and Taxation’, Posner
Memorial Collection, http://posner.library.cmu.edu/Posner/books/pages.cgi?
call=330_R4.
Samuelson, P. A. (1949), ‘International Factor-Price Equalization Once Again’,
Economic Journal, vol. 59, pp. 181–197.
Smith, A. (1925[1776]), An Inquiry into the Nature and Causes of the Wealth of
Nations. London: Methuen (Cannan ed.).
Tavernier, Jean-Luc (1990), Echanges extérieurs et avantages comparatifs: la spé-
cialisation de la France confrontée à celle de ses concurrents, Economie et
prévision, no. 94/95.
UNECA (2006), Economic Report on Africa 2006. Capital Flows and Development
Financing in Africa, Addis Abeba: Economic Commission for Africa.
World Economic Forum (2013), The Global Competitiveness Report 2013–2014,
Geneva: World Economic Forum.
2
Economic Integration, Trade
Facilitation and Agricultural
Exports Performance in
ECOWAS Sub-Region
Wumi Olayiwola, Evans Osabuohien, Henry Okodua and
Oluyomi Ola-David
1 ECOWAS – a net food importer – lacks food sufficiency
To reduce prices, increase specialization and improve efficiency, income

and welfare – including competitiveness in world markets (Interna-
tional Centre for Trade and Sustainable Development – ICTSD, 2011) –
regional economic integration (through such things as free trade areas
and customs unions) aims to facilitate international trade by reduc-
ing transaction costs such as tariff and non-tariff barriers (Dalimov,
2009; Cissokho et al., 2013). The Economic Community of West African
States (ECOWAS) was founded in 1975 for this purpose, among oth-
ers. ECOWAS’ global exports and imports increased during 1999–2009
(Figure 2.1). Statistics also has it that about 20% of its food are imported
(ECOWAS Commission, 2010a). With respect to its intra-regional trade
(exports and imports),it can be inferred from Figure 2.2 that intra-
regional imports are quite higher than intra-regional exports with
respective maximum values of 20% and 15% between 1999 and 2009.
Some ECOWAS countries did less than 1% of their trade with other
ECOWAS members (ECOWAS Commission, 2010b).
To be effective, regional integration requires transport and communi-
cations infrastructure as well as appropriate institutions (Essien, 2009).
In 2007 ECOWAS’ average telephone density was less than 18 per
100 inhabitants compared to the Sub-Saharan average over 23% and
world average over 83% (Olayiwola and Osabuohien, 2009), while its
31
32 Economic Integration and Agriculture in ECOWAS
120000
100000
80000
60000
40000
20000
0
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Exports Imports
Figure 2.1 ECOWAS’ global exports and imports, 1999–2009 (USD millions)
Source: Computed from ECOWAS Commission, 2010b and World Bank (2010).
25
20
15
10
0
1999 2001 2003 2005 2007 2009
Intra import Intra export
Figure 2.2 ECOWAS’ intra-regional exports and imports as percent of its global
trade
Source: Computed from ECOWAS Commission, 2010b and World Bank (2010).
intra-regional trade required 10–20% more documents and processing

time than the world average (World Bank, 2010).
As in most African countries, agriculture is fundamental to
ECOWAS economies for employment and income (ECOWAS Commis-
sion, 2010a; Efobi and Osabuohien, 2011). Agriculture employs about
60% of the labor force – more in rural areas – while producing about 35%
of GDP and supplying over 16% of exports (AfDB, 2012). Agriculture is
important for poverty alleviation and food security, but its development
is hindered by small markets, inefficient transportation and communi-
cations infrastructure and lack of irrigation, technical know-how and
financial resources.
The adoption of a liberal trade policy has long been a central
objective of ECOWAS. Hindrances to trade of agricultural products
within West Africa include inadequate transport and communications
infrastructure, with resulting high costs, as well as conflicting policies
and procedures. It is hoped that – through improved infrastructure
Wumi Olayiwola et al. 33
and harmonized regulations – regional integration will facilitate intra-

regional agricultural trade from surplus to deficit countries, reducing
price fluctuations and leading to greater production and income. The
ECOWAS Trade Liberalization Scheme (ETLS), adopted in 1990 and the
ECOWAS Agricultural Policy (ECOWAP), adopted in 2005 aim to reduce
hindrances to trade (Olayiwola et al., 2011).
The major focus of ETLS – a set of protocols to guide the movement of
goods and people throughout the region (West Africa Trade Hub Tech-
nical Report, 2010) – is the encouragement of local manufacturing and
entrepreneurial development more generally, by providing preferential
treatment for unprocessed goods, traditional handicrafts and industrial
products in competition with cheap imports, thus reducing unemploy-
ment (Central Bank of Nigeria, 2006). Delays in full implementation of
an ECOWAS customs union and common market via ETLS have been
ascribed largely to the unwillingness of many member states to elimi-
nate tariff and non-tariff barriers because of the lack of a functioning
compensation mechanism to mitigate the loss of tariff revenues.
The major focus of ECOWAP is increased productivity of agriculture –
to meet food needs, reduce poverty and inequality and raise the compet-
itiveness of exports – through regional development, including access
to larger markets allowing economies of scale and via a strengthened
bargaining position in global trade negotiations.
This empirical study was conducted to analyze how ETLS and
ECOWAP – regional integration and trade facilitation – influenced agri-
cultural productivity and exports. After reviewing relevant literature, the
econometric model and estimation techniques are explained and results
presented and discussed, followed by a summary and conclusions.
2 Theoretical and empirical studies of regional integration

and trade facilitation
Regional economic integration – reducing barriers to trade and thus

increasing market access – includes preferential trading areas, free
trade agreements, common markets and monetary, customs, economic
and fiscal unions (and combinations thereof) (Ndulu et al., 2005).
These processes are most effective when there are strong institutions
and strong political commitment to integration among the mem-
bers (McCarthy, 2002; Dalimov, 2009). But results do not always
match hopes. Regional economic communities have proliferated in Sub-
Saharan Africa since independence, but intra-regional trade is lower
than projected (United Nations Economic Commission for Africa-
UNECA, 2010), largely because of residual trade barriers and lack of
harmonization among member states (Yang and Gupta, 2007) as well

as inadequate, high-cost transportation infrastructure (UNECA, 2010;
Osabuohien and Efobi, 2011). The international business community
has increasingly pushed for greater transparency, efficiency and pro-
cedural uniformity regarding exports and imports, so – to increase
specialization and growth (Freund and Bolaky, 2004) – facilitating trade
is often seen as a crucial policy reform.
The extent of trade facilitation has been quantified by the effects
of port operations, of the customs environment or of the regulatory
environment and institutional quality more generally and of using
e-business and IT (Wilson et al., 2005; Martínez-Zarzoso and Márquez-
Ramos, 2008). However, the World Trade Organization (WTO)’s defi-
nition of trade facilitation – the simplification and harmonization of
international trade procedures, primarily at the border – has also been
used (Engman, 2005).
A gravity model has been used on data from 135 countries to analyze
how agricultural trade within regional economic communities in Africa
compared to elsewhere (Cissokho et al., 2013), finding that non-tariff
barriers had not been as much of a hindrance in ECOWAS as elsewhere.
Others have used gravity models augmented with measures of trade
facilitation (Wilson et al., 2003; Martínez-Zarzoso and Márquez-Ramos,
2008), including for a group of Asia-Pacific countries (Wilson et al.,
2005) or focused on Mexico (Soloaga et al., 2006). An augmented grav-
ity model has also been used to analyze the effects of time delays in the
exporting country (Djankov et al., 2010). Not unreasonably, a similar
earlier study found that lengthy export and import procedures reduced
the probability that firms would attempt to export time-sensitive prod-
ucts, thus reducing trade volumes (Nordas et al., 2006). Similar results
have been found for six groups of African, Caribbean and Pacific
(ACP) countries negotiating Economic Partnership Agreements with the
European Union-EU (Persson, 2007). With an augmented gravity model
based on data from 13 exporting and 167 importing countries, it was
found that lower transport costs and less administrative time required
positively affected trade flows (Martínez-Zarzoso and Márquez-Ramos,
2008).
Analyses of a cross-country manufacturing survey in Sub-Saharan
Africa found that exports were lower from countries with restrictive
regulations and poor customs administrations (Clarke, 2005; Elbadawi
et al., 2006). But in other analyses of the effects of trade facilitation –
including reformed customs regulations and procedures – it has been
found that improved port efficiency and service infrastructure (such as
electricity and Internet) had had the greatest effect on intra-African

trade (Wilson et al., 2004; Njinkeu et al., 2008; Yoshino, 2008). The
business and legal climate more generally have also been found to
have had a large effect on manufacturing exports (Balchin and Edwards,
2008). Another gravity model augmented with indicators of infrastruc-
ture as well as of regulatory quality – regulation, administration and
enforcement of trade policies – was used with data on 124 countries
to analyze the effects on manufacturing exports of trade facilitation in
Africa, finding that further reforms – as well as improved transport and
communications infrastructure – could have more effect in Africa than
anywhere else in the world (Iwanow and Kirkpatrick, 2009). To some
extent, the dismal trade performance of Sub-Saharan Africa may even be
attributable to the conflicting (as well as lengthy and non-transparent)
customs regulations and procedures of overlapping (non-harmonized)
regional economic communities (Njinkeu et al., 2007, 2008).
From above, it is evident that not much empirical studies with main
focus on ECOWAS’ agricultural exports have been carried.
3 Data and methods
Data (see Appendix 1) on trade came from ECOWAS as well as the

World Bank’s World Trade Indicators (WTI), while other data came
from the World Bank’s World Governance Indicators (WGI) and World
Development Indicators (WDI) (World Bank, 2013).
The econometric model

A baseline model for analysis of country i’s trade facilitation (TF, prox-
ied by the days required to process imports or exports) – as a function
of political institutions, infrastructure services and general economic
conditions – can be specified as
TF = daysi = f (Institutions
( , Infrastructure, Macro) (1)
Where the independent variables are:
Institutions political institutions proxied by a rule-of-law index and a

control-of-corruption index, both from WGI;
Infrastructure infrastructure services measured by numbers of internet users
and of telephone subscribers (both fixed and mobile) per 100
inhabitants, both from WDI;
Macro economic conditions measured by per capita real GDP, from
WDI.
More explicitly we can write the static equations
xdaysij,t = δ0 + δ1 institutionsij,t + δ2 infrastructureij,t + δ3 macroij,t + (2)

mdaysij,t = β0 + β1 institutionsij,t + β2 infrastructureij,t + β3 macroij,t + ε (3)
where xdaysij,t = days to process exports of country i to country j in

period t; mdaysij,t = days to process imports of country i from country
j in period t; and and ε are error terms assumed to be purely random.
It is expected that δ0 and β0 > 0; δj < 0 and βk < 0; that is, that processing
days go down with better political institutions, infrastructure services
and economic conditions.
Dynamically this becomes
xdaysij,t = δ0 + δ1 xdaysij,t−1 + δ2 institutionsij,t + δ3 infrastructureij,t

+ δ4 macroij,t + (4)
mdaysij,t = β0 + β1 mdaysij,t−1 + β2 institutionsij,t + β3 infrastructureij,t
+ β4 macroij,t + ε (5)
Where xdaysij,t−1 and mdaysij,t−1 are number of days required to process

exports and imports of country i to or from country j respectively, in
the previous period. It is expected that δ1 and β1 > 0; that is, that pro-
cessing days also go up with higher processing days in the previous
period.
A baseline model for analysis of agricultural exports (measured as a
percentage of a country’s GDP1 ) – as a function of the country’s agricul-
tural production, political institutions, infrastructure services, regional
integration and trade facilitation (xdays) – can be specified as
ag − exports = (ag − product, institutions, infrastructure, integration, TF) (6)
Where the independent variables are:
ag-product Annual agricultural production;

Institutions political institutions proxied by a regulatory-quality index;
Infrastructure as above;
Integration regional integration proxied by the country’s intra-regional
export share;
xdays as defined above.
More explicitly we can write the static and dynamic equations
ag − exportsij,t = γ0 + γ1 ag − producttij,t + γ2 institutionsij,t

+ γ3 infrastructureij,t + γ4 integrationij,t + γ5 TFij,t + ϕ (7)
ag − exportsij,t = π0 + π1 ag − exportsij,t−1 + π2 ag − product ij,t
+ π3 institutionsij,t + π4 infrastructureij,t + γ5 integrationij,t
+ π6 TF ij,t + ω (8)
for agricultural exports of country i to country j in year t y; where φ and

ω are error terms assumed to be randomly and normally distributed.
It is expected that γ0 , γ1 , γ2 , γ3 , γ4 > 0; and γ5 < 0 and π0 , π1 , π2 , π3 , π4 , π5 >
0; π6 < 0, that is, that exports go down with xdays but up with
everything else.
Estimation techniques
Endogeneity – often a problem in dynamic panel models – can be
dealt with using instrumental variable (IV) estimation, but the instru-
ments must be genuinely exogenous yet strongly correlated with the
potentially endogenous variables. It helps to over-identify the model
to allow tests for exogeneity and excludability. Besides Two-Stage Least
Squares (TSLS), the other commonly used IV estimation method – Gen-
eralized Method of Moments (GMM) – produces identical results in
just-sufficiently identified models, but more precise results with over-
identified models. GMM also uses internal instruments, whereas with
TSLS the researcher must search for suitable external instruments.
Panel data from the 15 ECOWAS members during 2003–2008 was
therefore analyzed using GMM with the dynamic equations (4, 5 and
8). The choice of this period – besides the issue of data availability for
most members of ECOWAS – includes the need to have the number
of country dimension to be relatively larger than the time dimension.
Another reason is to focus on the period that is relatively close to the
year ECOWAP was adopted. In addition, it was from 2003 that the data
for WGI became annual. For robustness – and to account for the pres-
ence of both time-variant and – invariant explanatory factors – the same
data was also analyzed using fixed-effects estimation with the static
equations (2, 3 and 7). A preliminary analysis using Hausman (1978) test
is also carried out. Hausman test is used to differentiate between fixed
effects model and random effects model in panel data analysis. In this
case, the analysis indicated that fixed effects model gave more efficient
results than random effects model due to higher efficiency, while the
alternative Fixed effects (FE) is at least consistent.
An advantage of dynamic GMM estimation is that all variables not
correlated with the error term (including lagged and differenced vari-
ables) may be used as instruments (Greene, 2008). The optimal set
of internal instruments was chosen via the ‘collapse’ option in the
GMM command in STATA software. The lagged dependent variables
(xdays and mdays) appear as predetermined and endogenous variables.
Hence, endogeneity is controlled for by using internal instruments. The
dynamic GMM estimator applied in this study uses the levels equation
in each case to obtain a system of two equations: one differenced and
one in levels. The variables in the differenced equation are instrumented
with the lagged levels of the regressors in the first equation, while the
variables in levels in the second equation are instrumented with their
own first differences.
4 Summary and conclusions
The estimated results from the fixed effect (static) models are broadly
as expected. A 1% increase in export share (as a measure of regional
integration) correlated with about 0.23% reduction in the number of
days required to process exports in the 15 ECOWAS countries (Table 2.1,
column 1). Similarly, a 1% increase in GDP per capita correlated with
about 1.05% reduction. A 1% increase in the density of Internet users
correlated with about 0.18% reduction and a 1% increase in the density
of telephone subscribers with about 0.06% reduction. All these variables
might thus be important potential channels for increasing trade facilita-
tion in ECOWAS. The constant term suggests that, in the absence of all
these variables, it would take about 98 days to process exports.
A 1% increase in GDP per capita correlated with about 1.21% reduc-
tion in the number of days required to process imports (column 2),
while a 1% increase in the density of Internet users correlated with about
0.19% reduction. The constant term suggests – again, in the absence of
all explanatory variables – that it would take about 106 days to process
imports.
Neither the rule-of-law index nor the control-of-corruption index
were statistically significant in determining either export or import days,
though both had large apparent negative effects (that is, reducing export
and import days, as desired).
As would be expected, agricultural production had a huge effect on
agricultural exports (column 3), a 1% increase in production correlated
39
Table 2.1 Effects of ECOWAS regional integration and trade facilitation on

agricultural trade, 2003–2008
Dependent Fixed effects (Static GMM (Dynamic

variable equations) equations)
X-days M-days X-agric X-days M-days X-agric

(1) (2) (3) (4) (5) (6)
∗ Measures of
trade
facilitation
xdays (lagged) 1.329∗
(6.51)
mdays (lagged) 1.233∗
(9.47)
ag – exports 0.655∗
(lagged) (3.41)
ag – product 7.220∗∗∗ 8.650∗∗∗
(1.69) (1.65)
∗ Indicatorsof
institutions
Rule of law −5.211 −8.275 −0.662 −1.274
index (−0.72) (−0.86) (−0.15) (−0.28)
Control of −3.338 −2.388 −0.447 −0.572
corruption (−1.09) (−0.56) (−0.09) (−0.11)
index
Regulatory −0.674 1.219
quality index (−0.43) (0.89)
∗ Measures of
infrastruc-
ture
Internet users −0.181∗∗ −0.193∗∗∗ −0.099∗∗ 0.094 0.052 −0.121∗∗
(−2.27) (−1.73) (−2.44) (0.89) (0.67) (−2.02)
Telephone users −0.064∗∗ −0.031 0.042∗∗ 0.049 0.075 0.046∗∗
(−2.26) (−0.79) (2.48) (0.88) (1.34) (2.30)
∗ Measureof
economic
conditions
GDP per capita −1.054∗ −1.216∗ 0.028 0.027
(−4.52) (−3.72) (0.45) (0.47)
∗ Measures of
regional
integration
Export share −0.233∗ 0.025 −0.119∗∗∗ 0.118∗∗
(−3.15) (0.79) (−1.91) (2.28)
Import share 0.031 −0.025
(0.23) (−0.24)
TF (xdays) −0.071∗ −0.086∗
(−3.02) (−2.95)
Table 2.1 (Continued)
Dependent Fixed effects (Static GMM (Dynamic

variable equations) equations)
X-days M-days X-agric X-days M-days X-agric

(1) (2) (3) (4) (5) (6)
Constant 98.262∗ 106.875∗ 5.675∗∗ −16.884 −20.209∗∗ 1.117

(6.70) (5.36) (2.65) (−1.39) (−2.41) (0.64)
F-test 9.03 4.34 5.48 25.30 28.94 8.89
(coefficients) (0.000) (0.003) (0.000) (0.000) (0.000) (0.000)
(p-value)
F-test (residuals) 12.40 21.05 25.81
(p-value) (0.000) (0.000) (0.000)
Arellano-Bond (0.812) (0.450) (0.501)
AR(2)
(p-value)
Sargan (p-value) (0.855) (0.181) (0.064)
Number of 13 13 15 12 12 15
countries
Number of 12 12 12
instruments
Rho 0.983 0.980 0.904
R2 0.366 0.126 0.003
Notes: t-statistics are in parentheses; ∗ , ∗∗ , ∗∗∗ indicate significant at 1%, 5% and 10% levels;
all dynamic models had time dummies.
with about 7.2% increase in exports. Similarly, a 1% reduction in the

number of days required to process exports (as a measure of trade facil-
itation) correlated with about 0.07% increase in agricultural exports.
A 1% increase in the density of telephone subscribers correlated with
about 0.04% increase in agricultural exports. Confusingly, a 1% increase
in the density of Internet users correlated with about 0.1% reduction in
agricultural exports. Export share (as a measure of regional integration)
had a small positive effect, though not statistically significant.
The estimated results from the GMM (dynamic) models are also
broadly as expected, though with lower statistical significance. A 1%
increase in export share (as a measure of regional integration) correlated
with about 0.12% reduction in the number of days required to process
exports (column 4). This result clearly shows that regional integration
effort is a booster of trade facilitation. As was the case with the fixed-
effects models, results were even less statistically significant for import
days (column 5). Past integration efforts also affected current level of
trade facilitation so also the current level of integration can also affect
future efforts. As noted, regional integration (export share) can affect

trade facilitation (processing days), which in turn affects trade (which
affects regional integration – a virtuous circle).
Again as would be expected, agricultural production had a huge effect
on agricultural exports (column 6), a 1% increase in production cor-
related with about 8.6% increase in exports. Similarly, a 1% reduction
in the number of days required to process exports (as a measure of trade
facilitation) correlated with about 0.08% increase in agricultural exports.
A 1% increase in the density of telephone subscribers correlated with
about 0.046% increase. Again a 1% increase in the density of Internet
users correlated with about 0.012% reduction in agricultural exports.
A 1% increase in export share (as a measure of regional integration)
correlated with about 0.118% increase.
Overall, these results differ from those of Cissokho et al. (2013) who –
also studying ECOWAS – found that non-tariff barriers were not espe-
cially harmful for agricultural trade, perhaps because using different
methods or studying a different period.
F-tests of the coefficients – the small-sample counterpart of the Wald
(Chi-squared) tests – indicate high overall statistical significance of the
estimated models. F-tests that the residuals are zero (columns 1–3) indi-
cate statistically significant country effects which would make the use
of pooled OLS inappropriate. The coefficients of determination (R2 ) are
presented in the last row of columns 1–3. These measure the overall fit
of the estimated regression model and are therefore indicative of the
explanatory power of the estimated model. The R2 suggests that about
36.6% the systematic variation in export days is jointly explained by
the independent variables in column 1 of Table 2.1. Similarly, about
12.6 and 0.3% of the variations in imports days and agricultural exports
is jointly explained by their independent variables in columns 2 and 3,
respectively. The R2 for agricultural export model is rather low and indi-
cate a very poor fit meaning that the fixed-effects static model in this
case failed to perform well.
The Arellano-Bond test (columns 4–6) was applied to the differenced
residuals to purge any unobserved but perfectly autocorrelated idiosyn-
cratic errors. The Arellano-Bond test results show no evidence of serial
correlation, so coefficient estimates can be regarded as consistent. The
Sargan test for instrument validity – which uses the minimized value
of the one-step non-robust GMM criterion to compare the number of
instruments to the related number of parameters – confirms that the
instrument set is valid. The dynamic analyses of export- and import-
days (columns 4 and 5) used 12 instruments for 8 parameters each, thus
having 4 over-identifying restrictions; similarly, agricultural exports

(column 6) used 12 instruments for 8 parameters thereby satisfying the
over-identifying restrictions for the instruments set.
5 Policy recommendations
General method of moments (GMM) was used with instrumental

variable (IV) estimation on a dynamic model with panel data from
15 ECOWAS members to analyze how trade facilitation (proxied by
required processing days) – and regional economic integration more
generally (proxied by export share) – affected intra-regional agricultural
exports during 2003–2008. To check robustness, a static model with
fixed effects was also estimated. Because of their effect on exports, what
determines processing days was also estimated similarly.
Not unexpectedly, agricultural production – in which there had been
sustained growth during the study period – greatly affected agricul-
tural exports, which were also affected by regional integration and trade
facilitation as well as by density of telephone subscribers (but not of
Internet users). Regional integration, GDP per capita and ICT infrastruc-
ture (telephone and Internet) affected trade facilitation itself – which
in turn affected agricultural exports – as did rule of law and control of
corruption, though those effects were less clear.
Not unnaturally, previous levels of trade facilitation also affected
current levels, so that current levels can likely affect future levels as
well. As noted, regional integration (export share) can affect trade facil-
itation (processing days), which in turn affects trade (which affects
regional integration – a virtuous circle). If ECOWAS is to enhance trade
facilitation (and trade), greater commitment to regional integration
would help.
That real GDP per capita affected trade facilitation (as also noted
by McCarthy, 2002) – which, via increased trade, presumably affects
income as well – is another virtuous circle.
That ICT infrastructure affected trade facilitation – for example,
that greater use of Internet could reduce the days required to process
exports – suggests that ECOWAS members could benefit from greater
ICT investment.
Rule of law and control of corruption seemed to have had large effects
on trade facilitation, though variation was high. Strengthening them
might make the effects clearer.
Similarly, the effect of regional integration (export share) on agricul-
tural exports directly was not clear in the static model (though clearer in
the dynamic one). Strengthening regional integration might make this

effect clearer.
Commitment to full implementation of the ECOWAS Agricultural
Policy, ECOWAS Trade Liberalization Scheme and ECOWAS Free Trade
Area would seem to offer large gains in intra-regional agricultural
trade.
Annexes
Table 2.A.1 Table of summary statistics

.xtsum xdays mdays ag_ex ag_prdt rol coc rq internet telephone pcgdp ex_shr im_shr
Variable Mean Std. dev. Min Max Observations
xdays overall 31.2 11.6164 14 59 N = 55

between 11.05216 18.5 59 n = 15
within 4.033043 23.7 51.7 T-bar = 3.66667
mdays overall 38.43636 14.96139 10 68 N = 55
between 15.10783 13.5 66 n = 15
within 4.632147 21.18636 54.68636 T-bar = 3.66667
ag_ex overall 39.2789 25.85181 .0199969 86.70892 N = 108
between 27.38025 .6496777 80.62115 n = 14
within 7.000247 21.39622 56.34284 T = 7.74129
ag_prdt overall 2.36e+09 5.54e+09 3.61e+07 3.49e+10 N = 210
between 5.28e+09 5.16e+07 2.11e+10 n = 15
within 2.14e+09 −6.64e+09 1.64e+10 T = 14
rol overall −7.531733 .6229233 −2.278 .835 N = 150
between .6062827 −1.6174 .4776 n = 15
within .2065369 −1.413773 −.1037733 T = 10
coc overall −.65104 .4981454 −1.757 .752 N = 150
between .4141412 −1.1896 .2609 n = 15
within .2949503 −1.32944 −.55426 T = 10
rq overall −.6036933 .5795659 −3.132 .902 N = 150
between .4751497 −1.8027 −.0549 n = 15
within .3518043 −1.932993 .5717067 T = 10
internet overall 1.388095 .2414941 0 20.6 N = 189
between 1.353256 .084 4.986667 n = 15
within 2.042689 −3.348571 17.00143 T-bar = 12.6
teleph∼e overall 8.318289 12.96859 .15 73.2 N = 187
between 6.578133 .8488889 24.19 n = 15
within 11.31382 −10.30171 65.21686 T = 12.4667
pcgdp overall 360.2132 266.5749 56 1631.62 N = 207
between 268.586 132.3077 1214.857 n = 15
within 54.10705 117.3614 776.9758 T-bar = 13.8
ex_shr overall 10.0629 7.926697 .01 30.67 N = 138
between 7.041481 .323 24.725 n = 14
within 2.986261 −6.734244 22.46828 T-bar = 9.85714
im_shr overall .0798571 .1425058 0 .59 N = 140
between .1306912 0 .4792857 n = 13
within .0272162 −.0294286 .1988571 T-bar = 10.7692
Note
1. The purpose of using this measure is to take into consideration the economic
sizes of ECOWAS members.
References
AfDB (2012), African Economic Outlook 2012: Promoting Youth Employ-
ment, OECD Publishing. doi: 10.1787/aeo-2012-en (Accessed 20 November
2013).
Balchin, N. and Edwards S. (2008), ‘Trade-related Business Climate and Manufac-
turing Export Performance in Africa: A Firm Level Analysis’, Paper presented
at the Trade Costs and Business Environment: A Focus on Africa Workshop,
31 May, Entebbe, Uganda.
Central Bank of Nigeria (2006), International Economic Cooperation, http://www.
cenbank.org/monetarypolicy/ecowas.asp (Accessed 5 May 2011).
Cissokho, L., Haughton, J., Makpayo, K. and Seck, A. (2013), ‘Why Is Agricultural
Trade within ECOWAS so High?’, Journal of African Economies, vol. 22, no.1,
pp. 22–51.
Clarke, G. (2005), ‘Beyond Tariff and Quotas: Why Don’t African Manufacturing
Enterprises Export More?’ World Bank Policy Research Working Paper, r no. 3617.
Dalimov, R. T. (2009), ‘The Dynamics of Trade Creation and Trade Diversion
Effects Under International Economic Integration’, Journal of Economic Theory,
vol. 1, no. 1, pp. 1–4.
Djankov, S., Freund, C. and Pham, C. S. (2010), ‘Trading on Time’, The Review of
Economics and Statistics, vol. 92, no. 1, pp. 166–172.
ECOWAS Commission (2010a), Regional Agricultural Policy, Abuja: ECOWAS.
ECOWAS Commission (2010b), The ECOWAS External Trade Statistics, Abuja:
ECOWAS
Efobi, U. and Osabuohien, E. S. C. (2011), ‘Promotion of Non-Oil Export in
Nigeria: Empirical Assessment of Agricultural Credit Guarantee Scheme Fund’,
Journal of Economic Theory, vol. 3, no. 1, pp. 20–28.
Elbadawi, I., Mengistae, T. and Zeufack, A. (2006), ‘Market Access, Supplier Access
and Africa’s Manufactured Exports: An Analysis of the Role of Geography and
Institutions’, WB Policy Working Paper, r no. 3942.
Engman, M. (2005), ‘The Economic Impact of Trade Facilitation’, OECD Trade
Policy Working Papers, no. 21.
Essien, A. E. (2009), ‘Introducing a Common Currency: The ECOWAS Experi-
ence’, Paper Presented at the ECOWAS-GIZ Regional Integration Academy in
Comparative Regional Integration, Centre for European Integration Studies,
Bonn, Germany, 16–28 March.
Freund, C. and Bolaky, B. (2004), ‘Trade, Regulations, and Growth’, World Bank
Policy Research Working Paper,r no. 3255.
Hausman, J. A. (1978). ‘Specification Tests in Econometrics’, Econometrica, vol. 46,
no. 6, pp. 1251–1271.
International Centre for Trade and Sustainable Development [ICTSD] (2011),
‘Agricultural Export Restrictions Speak Controversy at the WTO’, Bridges Weekly
Trade News Digest,t vol. 15, no. 2, pp. 1–15.
Iwanow, T. and Kirkpatrick, C. (2009), ‘Trade Facilitation and Manufactured

Exports: Is Africa Different?’ World Development, t vol. 37, no. 6, pp. 1039–1050.
Martínez-Zarzoso, I. and Márquez-Ramos, L. (2008), ‘The Effect of Trade Facilita-
tion on Sectoral Trade’, The B.E. Journal of Economic Analysis and Policy, vol. 8,
no. 1, pp. 1–41.
McCarthy, C. L (2002), ‘Macroeconomic Convergence in SADC – A Policy Per-
spective for the Central Banks of the Integration Arrangement’, Committee of
Central Bank Governors in SADC.
Ndulu, B., Kritzinger-van, L. and Reinikka, R. (2005), ‘Infrastructure, Regional
Integration and Growth in SSA’, in J. J. Teunissen and A. Akkerman (eds.),
Africa in the World Economy – The National, Regional and International Challenges,
pp. 101–119.
Njinkeu, D., Wilson, J. and Fosso, B. P. (2007), ‘Trade Facilitation: What Is It and
How Does It Help?’ Paper Prepared for AERC Collaborative Research Project on
Supply Response, July.
Njinkeu, D., Wilson, J. and Fosso, P. B. (2008), ‘Intra Africa Trade Constraints: The
Impact of Trade Facilitation’, Mimeo, Washington DC: World Bank.
Nordas, E. P., Pinali, E. and Grosso, N. G. (2006), ‘Logistics and Time as a Trade
Barrier’, OECD Trade Policy Working Papers, no. 35.
Olayiwola, K. W., Osabuohien, E. S. and Okodua, H. (2011), ‘Role of Eco-
nomic Integration and Trade Facilitation on Agricultural Export Perfor-
mance in ECOWAS Member Countries’, EPAU Monograph Series, no. 1, Abuja:
ECOWAS Commission.
Olayiwola, K. W. and Osabuohien, E. S. C. (2009), ‘Training Needs for
ECOWAS for Better Performance in International Trade’, Technical Report, t
TRAPCA, Arusha, Tanzania, July.
Osabuohien, E. and Efobi, U. (2011), ‘Trade Outcomes in Africa’s Regional Eco-
nomic Communities and Institutional Quality: Some Policy Prescriptions’,
Economic Insights-Trends and Challenges, vol. LXIII, no. 4, pp. 19–32.
Persson, M. (2007), ‘Trade Facilitation and the EU-ACP Economic Partnership
Agreements: Who Has the Most to Gain?’ Lund University Working Papers, no. 8.
Soloaga, I., Wilson, J. and Mejía, A. (2006), ‘Moving Forward Faster: Trade Facilita-
tion Reform and Mexican Competitiveness’, WB Policy Research Working Paper, r
no. 3953.
United Nations Economic Commission for Africa-UNECA (2010), Assessing
Regional Integration in Africa IV: Enhancing Intra-African Trade, Addis Ababa:
UNECA.
West Africa Trade Hub Technical Report (2010), ‘Gap Analysis ECOWAS Free Trade
Area’, Technical Reports, no. 37 and 38.
Wilson, J. S., Mann, C. L. and Otsuki, T. (2003), ‘Trade Facilitation and Economic
Development: A New Approach to Quantifying the Impact’, WB Economic
Review, vol. 17, no. 3, pp. 367–389.
Wilson, J. S., Mann, C. L. and Otsuki, T. (2004), ‘Assessing the Potential Benefit
of Trade Facilitation: A Global Perspective’, WB Working Paper Series, no. 3224.
Wilson, J. S., Mann, C. L. and Otsuki, T. (2005), ‘Assessing the Benefits of Trade
Facilitation: A Global Perspective’, World Economy, vol. 28, no. 6, pp. 841–871.
World Bank (2010), World Trade Indicators 2009/10 (Accessed 12 February 2010).
World Bank (2013), World Development Indicators 2013, Washington, DC: The
World Bank Group.
Yang, Y. and Gupta, S. (2007), ‘Regional Trade Arrangements in Africa: Past Per-
formance and the Way Forward’, African Development Review, vol. 19, no. 3,
pp. 399–431.
Yoshino, Y. (2008), ‘Domestic Constraints, Firm Characteristics, and Geograph-
ical Diversification of Firm-level Manufacturing Exports in Africa’, WB Policy
Research Working Paper,
r no. 4575.
3
The Impact of Trade Liberalization
on Export Growth and Import
Growth in Sub-Saharan Africa
Lanre Kassim
Introduction
In the 1960s and 1970s, Sub-Saharan African (SSA) countries adopted

interventionist policies aimed at protecting their domestic markets from
foreign competition. But the 1979 oil shock, followed by the debt crisis
and global recession of the early 1980s, left the region in economic dol-
drums: non-fuel primary commodity prices plummeted, debt to GDP
ratio rose to 70% and per capita income declined by 14% from 1980
to 1987 (UNCTAD, 2004). This prompted international financial insti-
tutions, such as the World Bank and the International Monetary Fund
(IMF), to offer financial aid to the region, but on condition that coun-
tries opened up their trade regimes. By the mid-1980s and early 1990s,
free trade policies and Structural Adjustment Programmes began dom-
inating the region. Early liberalizers included Niger and Ghana, while
countries such as Angola, Burundi and the Democratic Republic of the
Congo (DRC), only embarked on significant trade reforms in the early
twenty-first century.
The removal of trade barriers such as import tariffs, export duties and
quantitative restrictions stimulates the growth of exports and imports.
But if imports grow relatively faster than exports, then an economy
risks balance-of-payment problems that can constrain output growth
(see Thirlwall, 1979). This trade liberalization conundrum has not been
adequately researched in the literature on Sub-Saharan Africa. Most
empirical studies on trade liberalization adopt a narrow approach, ana-
lyzing its impact on output (GDP) growth without considering whether
growth is sustainable and consistent with long-run balance of payments
(BOP) equilibrium.
47
48 Trade Liberalization in Sub-Saharan Africa
Instead of examining the impact of trade liberalization on the trade-

off between output growth and BOP, this study investigates how the
adoption of outward-oriented policies has affected export growth and
import growth. The analysis looks into whether imports grew faster than
exports in the post-liberalization era and gives a preliminary indication
of the region’s trade balance position. Further, it assesses the impact of
trade liberalization on the price and income elasticity of demand for
exports and imports.1
The study focuses on a panel dataset of 28 SSA countries from 1981
to 2010. Export and import growth equations are specified to include
explanatory variables, such as domestic income growth, foreign income
growth and a measure of price competitiveness. Static and dynamic
panel models are employed, and trade liberalization is measured quan-
titatively and qualitatively. First, average duties applied to exports and
imports are used, where export duty represents export taxes as a per-
centage of total exports and import duty denotes import taxes as a
percentage of total imports. Second, a dummy variable is applied, taking
the value of one when uninterrupted trade reforms began in a SSA coun-
try and zero beforehand. The duty variable captures the direct impact of
trade tariffs and the liberalization dummy captures the effect of non-
tariff barriers. Liberalization dates are constructed from examining the
trade policy reviews of SSA countries.
Empirical studies into the impact of trade liberalization on export per-
formance have produced mixed results. World Bank studies by Michealy
et al. (1991) and Thomas et al. (1991), Joshi and Little (1996), Bleaney
(1999), Ahmed (2000), Pacheco-Lopez (2005) and Santos-Paulino (2002a
and 2007) all found a significant positive relationship between trade lib-
eralization and export performance. But Clarke and Kirkpatrick (1992),
Greenaway and Sapsford (1994), Jenkins (1996), Ratnaike (2012) and
UNCTAD studies by Agosin (1991) and Shafaeddin (1994) found little
or no evidence of any favorable impact of trade liberalization on export
performance.
Only a few studies have analyzed the relationship between trade
liberalization and imports. In Melo and Vogt (1984), two hypotheses
were proposed regarding the impact of import liberalization on the
behavior of import elasticities: that the income elasticity of demand
increases as the degree of import liberalization increases; and that as
economic development continues, the price elasticity of import demand
rises owing to progress in import substitution. Using disaggregated
annual data, their results provided support for these two hypotheses
in Venezuela. Santos-Paulino (2002b) also found empirical evidence
Lanre Kassim 49
for the hypotheses across a group of developing countries. Boylan and

Cuddy (1987), however, rejected the hypotheses in an investigation of
elasticities of import demand in Ireland. Mah (1999) found that income
elasticity of demand increased as a result of import liberalization in
Thailand, although price elasticity did not rise.
In SSA, there is a dearth of empirical studies on the relationship
between trade liberalization and export growth, and almost no stud-
ies on the impact of trade liberalization on import growth. UNCTAD
(2008) conducted a study on the post-liberalization export performance
of 34 African countries, using a liberalization dummy in accordance
with the Wacziarg and Welch (2008) classification and applying the
Generalized Methods of Moment (GMM) estimator. It found that trade
liberalization increased the ratio of exports-to-GDP by 0.09 percent-
age points. Although this impact is very minimal, it is surprising that
the exports-to-GDP ratio is used as the dependent variable in contrast
to the growth of exports. Babatunde (2009) examined the impact of
trade liberalization on export performance across 20 SSA countries from
1980 to 2005. Using fixed and random effects estimation techniques,
no significant relationship was found between trade liberalization and
export performance. This finding was to be expected as Babatunde
used average tariff rates (which may not directly affect exports) as
the indicator of trade liberalization in the export equation. In addi-
tion, Olofin and Babatunde (2009) examined the price and income
elasticities of Sub-Saharan African exports from 1980 to 2003, applying
a fixed effects estimation technique on a panel dataset of 20 coun-
tries. The calculated long-run income elasticity of demand for exports
was found to range between 0.94 and 1.33 while the estimated long-
run price elasticity of demand varies from −0.01 to −0.17. The real
income of trading partners and price competitiveness of exporting coun-
tries were also found to be significant determinants of SSA exports
(Table 3.1).
The present study departs from the aforementioned ones by con-
structing liberalization dates for sample countries instead of relying on
the Dean et al. (1994) or Wacziarg and Welch (2008) classifications.
A measure of relative prices is adopted for tradable goods only, while
specifications analyze whether the short-run impact of trade liberaliza-
tion on export and import growth is instantaneous or not. No other
study on SSA is known to have adopted this approach. Section 2 shows
simple descriptive statistics on export growth and import growth before
and after liberalization. Section 3 and 4 explain the methodology and
empirical results for the impact of trade liberalization on export growth
Table 3.1 Average export and import growth before and after liberalization
Country (28) Lib Year Average export Average import

growth growth
Before After Increase/ Before After Increase/

Lib (%) Lib (%) Decrease Lib (%) Lib (%) Decrease
Benin 1988 −1.13 2.06 Increase −6.24 3.53 Increase

Botswana 1994 8.84 2.85 Decrease 5.65 5.63 Same
Burkina Faso 1991 2.64 6.98 Increase 0.96 3.12 Increase
Burundi 2003 5.30 21.56 Increase 1.09 37.14 Increase
Cameroon 1989 7.89 1.55 Decrease 8.88 3.79 Decrease
Cote d’Ivoire 1994 1.01 4.33 Increase −2.19 6.24 Increase
DRC 2001 4.66 6.24 Increase 6.53 15.37 Increase
Ethiopia 1992 0.67 10.02 Increase 3.99 10.70 Increase
Gabon 1994 5.14 −1.39 Decrease 2.35 1.66 Decrease∗
Gambia 1986 8.24 2.08 Decrease −11.13 2.61 Increase
Ghana 1983 −46.70 10.43 Increase −4.5 11.78 Increase
Kenya 1993 4.07 4.35 Increase∗ 0.06 8.85 Increase
Lesotho 1994 6.30 9.67 Increase 4.26 7.37 Increase
Madagascar 1988 −5.44 6.30 Increase −9.91 7.34 Increase
Malawi 1988 1.55 3.69 Increase −5.57 3.67 Increase
Mali 1988 4.33 8.97 Increase 6.38 4.77 Decrease
Mauritius 1985 2.51 5.64 Increase −1.48 6.13 Increase
Namibia 1994 2.82 0.57 Decrease 1.21 4.28 Increase
Nigeria 1986 −6.72 3.45 Increase −3.02 4.72 Increase
Rwanda 1995 −4.77 20.36 Increase 10.27 1.98 Decrease
Senegal 1986 7.68 3.37 Decrease 6.84 3.70 Decrease
Sierra Leone 1989 −4.1 −0.56 Increase −8.04 2.90 Increase
South Africa 1994 2.20 3.90 Increase 1.38 6.05 Increase
Swaziland 1994 4.25 6.23 Increase 3.20 5.05 Increase
Togo 1994 −2.58 3.01 Increase −2.57 6.04 Increase
Uganda 1987 0.63 14.12 Increase 0.94 9.52 Increase
Zambia 1991 −3.39 14.99 Increase −4.14 15.47 Increase
Zimbabwe 1990 5.31 −1.76 Decrease 6.23 1.27 Decrease
Note: (∗ ) denotes a marginal increase or decrease while all values are the author’s calculations.
Source: WTO, 1995–2011. Policy reviews for various countries.
and import growth, respectively. Section 5 discusses and compares

results, while Section 6 offers a conclusion.
1 Export and import growth 1981–2010
The analysis begins by examining simple descriptive statistics for the

growth of exports and imports before and after liberalization (see
Figure 3.A.1). The aim is to collect preliminary information on export
growth and import growth in the post-liberalization era. Yet without
Lanre Kassim 51
controlling for other variables, changes in the growth of exports and

imports cannot be attributed to liberalization alone.
After liberalization, export growth increased in 21 countries, while it
decreased in seven countries. At the same time, import growth increased
in 21 countries, decreased in six countries and remained the same in
Botswana. Further, post-liberalization imports grew faster than exports
in all countries except Burkina Faso, Lesotho, Mali, Rwanda, Swaziland
and Uganda. Again, these results cannot be attributed to liberalization
alone, as simple descriptive statistics do not show a causal link between
two variables. To find such causal evidence, the following econometric
analysis was required.
2 Trade liberalization’s impact on export growth
Model specification
A conventional export demand equation relates the level of exports to
world real income and a measure of price competitiveness. Assuming
constant price and income elasticities of demand, the export equation
can be expressed as:
δ
EPex
EXPt = A ∗
Wtγ (1)
Pim t
Here EXP represents the level of exports; A is a constant; E is the nom-

inal exchange rate measured as the foreign price of domestic currency;
Pex /Pim
∗
is the ratio of domestic export prices to foreign import prices;

the real exchange rate (ReR) is measured as EPex /Pim∗
; W is the level of
world real income; and γ denotes the income elasticity of demand for
exports which is expected to be positive. A decrease in the foreign price
of domestic currency (devaluation), or a fall in export prices relative to
import prices, should reduce ReR and thereby increase export growth
so that the expected sign for the price elasticity of demand (δ) is nega-
tive. Taking logs and differentiating with respect to time, Equation (1)
becomes:
xt = a + δ(e + pex − p∗im )t + γ (wt ) (2)
Equation (2) can be transformed into a static panel specification in the
form of:
epggit = αi + β1 rerrit + β2 wgdpggit + εit (3)
Here epg is the growth of real exports, αi is the country-specific effect,
rer measures the rate of change of the real exchange rate, wgdpg is the
growth of world real income, and εit is the idiosyncratic error term. Also,
β1 and β2 denote the price and income elasticity of demand for exports
respectively. Equation (3) is augmented to include the two measures of
liberalization as follows:
epggit = αi + β1 rerrit + β2 wgdpggit + β3 libdumit + β4 epdit + εit (4)
Here libdum is the liberalization dummy which takes the value of 1 from
the year significant trade reforms began in an SSA country and zero
beforehand. Since trade liberalization reduces the degree of anti-export
bias, the variable libdum is expected to have a positive impact on real
export growth. epd is measured as the rate of change of export duties.
But what is the impact of trade liberalization on the price and income
elasticity of demand for exports? The implementation of trade reforms
should raise the income elasticity of demand for exports, implying
structural change in the form of resources transferred to sectors with
high income elasticities. Hence, two interaction variables are created
to examine whether trade liberalization has significantly increased or
decreased the responsiveness of export growth to world income and
relative price changes:
epggit = αi + β1 rerrit + β2 wgdpggit

+ β3 libdumit + β4 epdit + β5 rerlibit + β6 wlibit + εit (5)
Here rerlib represents an interaction between the rate of change of the

real exchange rate and the liberalization dummy, while wlib denotes
an interaction between world income growth and the liberalization
dummy.2 The expected signs of the coefficients of Equation (5) are:
β1 ( − ); β2 ( + ); β3 ( + ); β4 ( − ); β5 ( − ); β6 ( + ).
Random effects
Table 3.2 presents the random effects (RE) results, which the Hausman
test indicates as the appropriate estimator for Equations (4) and (5).
Liberalization significantly increases export growth by 3.32 per-
centage points, while a 10% decrease in export duties increases
export growth by 0.2%, albeit insignificant. The income elasticity of
demand for exports is 1.12 which means that a change in world
income will cause a marginally higher change in the demand for
SSA exports. In addition, the price elasticity of demand is −0. 24, imply-
ing that exports are much less responsive to changes in relative prices.
So SSA countries are still major exporters of primary commodities.
Lanre Kassim 53
Table 3.2 Trade liberalization and export growth (random effects)
Independent variables Dependent variable Export growth
RE (I) RE (II)
rer −0.24 −0.13

(4.44)∗ (1.80)∗∗∗
wgdpg 1.12 0.99
(2.40)∗∗ (1.38)
libdum 3.32 2.81
(2.33)∗∗ (0.87)
epd −0.02 −0.02
(0.05) (0.03)
rerlib −0.26
(2.41)∗∗
wlib 0.21
(0.22)
Diagnostic statistics
R2 0.069 0.083
F-stat [p-value] [0.0000] [0.0000]
Joint Sig − [0.0532]
No. of observations 399 399
Hausman [p-value] [0.4099] [0.6741]
Note: ∗ , ∗∗ and ∗∗∗ indicate that a coefficient is statistically significant at 1%, 5% and
10% significance level, respectively. Figures in parenthesis () are absolute t/z ratios
while figures in brackets [ ] are p-values. ‘Joint Sig’ is an F-test for the joint significance
of the two slope dummies.
The second RE regression includes the two interaction terms rerlib and
wlib. The price elasticity of demand for exports drops to −0.13, and there
is also evidence that trade liberalization increased the price elasticity of
demand for exports by 0.26 percentage points. There is no significant
evidence of the impact of trade liberalization on income elasticity of
demand for exports. Testing the joint significance of the two interaction
terms and the results shows they are jointly significant.
Generalized method of moments

Furthermore, a dynamic panel model (DPM) is specified to examine the
effect of trade liberalization on export growth. The generalized method
of moments (GMM) is employed for econometric analysis. This estima-
tor allows an instrumental variable (IV) estimation of parameters, but
also long-run and short-run effects of variables. An IV approach helps
control for measurement error which might be present in static panel
models. Thus, equation (4) and (5) are specified in dynamic form as
follows:
epggit = αi + β1 epggit −1 + β2 rerrit + β3 wgdpggit + β4 libdumit + β5 epdit + εit (6)

epggit = αi + β1 epggit −1 + β2 rerrit + β3 wgdpggit + β4 libdumit + β5 epdit +
β6 rerlibit + β7 wlibit + εit (7)
Here epggit −1 is the lagged export growth variable, αi is the fixed effect,
and εit is the idiosyncratic error term. The regressors are those defined
in Section 3.1. Long-run price and income elasticities of demand for
exports can be calculated as β2 /(1 − β1 ) and β3 /(1 − β1 ), respectively;
while long-run liberalization effect can be estimated as β4 /(1 − β1 ). The
GMM results are shown in Table 3.3.
The lagged export growth variable has a low coefficient, which implies
a small difference between short-run and long-run estimates. Trade lib-
eralization raised the growth of exports by 2.67 percentage points in the
short run with this figure increasing to 3.22 percentage points in the
long run. The short-run income elasticity is 1.43 while in the long run,
the income elasticity is 1.72. The short-run price elasticity of demand is
−0. 28, while the long-run figure stands at −0. 34. This figure is above the
estimated range of long-run price elasticity of demand for SSA exports by
Olofin and Babatunde (2009). The differences in estimates can plausibly
be attributed to the different measures of real exchange rate adopted.
Again, the export duty variable remains insignificant.
The second GMM regression produces similar results to the RE regres-
sion. Liberalization increases the price elasticity of demand for exports
by 0.25 percentage points while there is no significant evidence of the
effect of trade liberalization on income elasticity of demand. The corre-
lation tests show that the error term is serially uncorrelated while the
Sargan test shows that instruments are valid.
Timing impact
The timing impact of trade reforms on export growth in SSA is investi-
gated by replacing the average post-reform dummy variable (libdum) in
the base specification with libdum2 which equals one for the year of lib-
eralization only and libdum3 which captures the year after liberalization
alone. A significant coefficient on either of these variables will imply
that the liberalization had an instant short-run effect on export growth
(see Greenaway et al. 2002). The RE and GMM estimators are applied
and results can be found in Table 3.4.
Lanre Kassim 55
Table 3.3 Trade liberalization and export growth (GMM)
Independent variables Dependent variable Export growth
GMM (I) GMM (II)
l.epg 0.17 0.15

(1.13) (0.89)
rer −0.28 −0.15
(3.22)∗ (3.22)∗
wgdpg 1.43 1.55
(3.25)∗ (1.69)∗∗∗
libdum 2.67 3.65
(1.95)∗∗∗ (1.01)
epd −0.12 −0.03
(0.18) (0.04)
rerlib −0.25
(2.08)∗∗
wlib −0.25
(0.23)
LR income elasticity 1.72 1.82
LR price elasticity −0.34 −0.18
LR Lib effect 3.22 4.29
Joint Sig [0.1065]
Wald test [0.000] [0.000]
1st-Order serial correlation [0.015] [0.032]
2nd-Order serial correlation [0.133] [0.172]
Sargan test [0.477] [0.539]
10% significance level, respectively. Figures in parenthesis () are absolute t/z ratios
while figures in brackets [] are p-values.
The RE and GMM estimators produce similar results in terms of signs

and significance of variables. The short-run price elasticity of demand
lies between −0.24 and −0.28, while the short-run income elasticity of
demand ranges from 0.97 to 1.45. There is a decrease in export growth
in the liberalization year only. Although statistically insignificant, the
drop in export growth can be attributed to buyers delaying their pur-
chases of export goods in anticipation of a liberalization policy which
will lower export prices. In the year after liberalization, there is positive
export growth, but this is still statistically insignificant. Even when the
average post-reform liberalization dummy is included, the signs and sig-
nificance of independent variables remain unchanged. In other words,
Table 3.4 Timing impact of trade reforms on export growth
Independent Dependent variable Export growth GMM (I) GMM (II)

variables
Random Random
effects (I) effects (II)
l.epg – – 0.21 0.20

(1.38) (1.39)
libdum – 3.55 – 2.77
(2.45)∗∗ (1.83)∗∗∗
libdum2 −3.50 −5.21 −2.59 −3.86
(0.89) (1.32) (0.68) (0.93)
libdum3 3.24 1.58 3.36 2.14
(0.80) (0.38) (0.73) (0.45)
rer −0.24 −0.25 −0.28 −0.28
(4.42)∗ (4.48)∗ (3.37)∗ (3.42)∗
wgdpg 0.97 1.14 1.37 1.45
(0.08) (0.07) (0.06) (0.17)
Diagnostic
tests
No. of 399 399 380 380
observations
Wald 0.000 0.000 0.000 0.000
test/F-stat
[p-value]
1st-Order – – 0.015 0.015
serial
correlation
2nd-Order – – 0.133 0.129
serial
correlation
Sargan test – – 0.477 0.495
Note: ∗ , ∗∗ and ∗∗∗ indicate that a coefficient is statistically significant at 1%, 5% and 10%
significance level, respectively. Figures in parenthesis () are absolute t/z ratios while figures
in brackets [] are p-values.
the short-run impact of trade liberalization on export growth is not

instantaneous in SSA.
In view of the three econometric analyses adopted in the three sub-
sections of section 3, the appropriate specification for the impact of
trade liberalization on export growth is that which includes the average
post reform liberalization dummy (equations 4 and 6). The implica-
tion of this specification is that the response of export growth to trade
liberalization is slow.
Lanre Kassim 57
3 Trade liberalization’s impact on import growth
Model specification
To analyze the impact of trade liberalization on import growth, the
same countries, time period and methodology are used as in the export
growth estimations. The reason for this approach is to ensure a con-
sistent framework for comparing the impact of trade liberalization on
export growth and import growth. The static and dynamic equations to
be estimated are:
impggit = αi + δ1 rerrit + δ2 gdpggit + δ3 libdumit + δ4 impdit + δ5 rerlibit + δ6 ylibit + εit

(8)
impggit = αi + δ1 impggit −1 + δ2 rerrit + δ3 gdpggit + δ4 libdumit + δ5 impdit +
δ6 rerlibit + δ7 ylibit + εit (9)
Here impggit is the real import growth; rerrit represents the rate of change
of the real exchange rate; gdpggit is the domestic real income growth;
libdumit denotes the liberalization dummy; impdit is the rate of change
of import duties; αi is the country-specific effect while εit is the idiosyn-
cratic error term. rerlib is an interaction variable between the rate of
change of the real exchange rate and liberalization dummy while ylib
is an interaction term between domestic income growth and the liberal-
ization dummy. From equation (8), the expected signs of the coefficients
are: δ1 ( + ); δ2 ( + ); δ3 ( + ); δ4 ( − ); δ5 ( + ); δ6 ( + ).
Fixed effects and GMM

In accordance with the result of the Hausman test, equations (8) and
(9) are estimated using fixed effects (FE). Also, the GMM is applied in
order to estimate the short-run and long-run impact of trade liberaliza-
tion on import growth. The FE and GMM results are shown in Tables 3.5
and 3.6, respectively.
The FE results show that liberalization significantly increases the
growth of imports by 4.87 percentage points, which is higher than
export growth by 1.55 percentage points. Also, a 1% decrease in import
duties increases the growth of import by 0.66% while the income and
price elasticities of demand for imports are 1.20 and 0.30, respectively.
The second FE regression produces similar results to the first, except for
the liberalization dummy coefficient, which drops by almost one per-
centage point. Also, liberalization has a positive impact on price and
income elasticity of demand for imports, though the latter effect is
Table 3.5 Trade liberalization and import growth (fixed effects)
Independent variables Dependent variable Import growth
FE (I) FE (II)
rer 0.30 0.22

(5.99)∗ (3.31)∗
gdpg 1.20 1.15
(7.96)∗ (5.99)∗
libdum 4.87 3.96
(3.96)∗ (2.18)∗∗
impd −0.66 −0.66
(3.00)∗ (2.26)∗
rerlib 0.20
(1.96)∗∗
ylib 0.20
(0.62)
R2 0.2765 0.2853
F-stat [p-value] [0.0000] [0.0000]
Joint Sig [0.1288]
Hausman [p-value] [0.0007] [0.0031]
significance level, respectively. Figures in parenthesis ( ) are absolute t/z ratios while figures
in brackets [ ] are p-values.
insignificant. The joint significance test shows that the two interaction
variables are jointly insignificant.
The results in Table 3.6 indicate that liberalization raised import
growth by 4.33 and 4.92 percentage points in the short run and long
run respectively. This is above the long-run post-liberalization export
growth by 1.70 percentage points. A 10% decrease in import duties
increase the growth of imports by 5.6%. The short-run income elastic-
ity of demand for import is 1.05 while the long-run figure is 1.19. The
short-run price elasticity of demand is 0.30 while in the long run the
coefficient increases to 0.34. Again, we find a positive impact of trade
liberalization on price elasticity of demand while there is no significant
evidence of the effect of liberalization on income elasticity of demand
for import. Furthermore, the diagnostic tests show the error term is
serially uncorrelated and that the instruments used for the estimated
model are valid.3
Lanre Kassim 59
Table 3.6 Trade liberalization and import growth (GMM)
Independent variables Dependent variable Import growth
GMM (I) GMM (II)
l.impg 0.12 0.13

(1.78)∗∗∗ (2.09)∗∗
rer 0.30 0.21
(4.45)∗ (2.17)∗∗
gdpg 1.05 0.93
(4.02)∗ (2.52)∗∗
libdum 4.33 2.86
(3.25)∗ (1.23)
impd −0.56 −0.57
(2.12)∗∗ (2.20)∗∗
rerlib 0.20
(2.00)∗∗
ylib 0.35
(0.63)
LR income elasticity 1.19 1.07
LR price elasticity 0.34 0.24
LR lib effect 4.92 3.29
Joint Sig [0.1339]
Wald test [0.000] [0.000]
1st-Order serial correlation [0.004] [0.003]
2nd-Order serial correlation [0.224] [0.201]
Sargan test [0.490] [0.484]
10% significance level, respectively. Figures in parenthesis ( ) are absolute t/z ratios
while figures in brackets [ ] are p-values.
Timing impact
Again, the timing impact of trade reforms on the growth of imports is
tested by introducing two liberalization dummies (libdum2 and libdum3)
to the base specification. Results are shown in Table 3.7.
A decrease is observed in import growth in the liberalization year,
which can be attributed to an expectation of future lower prices follow-
ing import liberalization. In the following year, however, a positive sig-
nificant import growth is witnessed. Inclusion of the average post-reform
dummy variable in second GMM result does not change the significance
of the ‘year after’ dummy variable. This implies an immediate short-run
Table 3.7 Timing impact of trade reforms on import growth
Independent Dependent variable Import growth GMM (I) GMM (II)

variables
Fixed Fixed
Effects (I) Effects (II)
l.impg – – 0.12 0.12

(1.59) (1.78)∗∗∗
libdum – 4.90 – 4.23
(3.26)∗ (2.70)∗
libdum2 −1.84 −4.40 −1.27 −3.37
(0.51) (1.20) (0.68) (0.93)
libdum3 6.90 4.42 7.17 5.14
(1.84)∗∗∗ (1.17) (2.73)∗ (1.84)∗∗∗
rer 0.30 0.30 0.30 0.30
(5.98)∗ (5.94)∗ (3.37)∗ (4.26)∗
gdpg 1.26 1.19 1.12 1.04
(8.32)∗ (7.84)∗ (4.11)∗ (3.85)∗
impd −0.72 −0.68 −0.62 −0.57
(3.23)∗ (3.07)∗ (2.52)∗∗ (2.26)∗∗
Diagnostic
tests
No. of 363 363 321 321
observations
Wald 0.000 0.000 0.000 0.000
test/F-stat
[p-value]
1st-Order – – 0.005 0.005
serial
correlation
2nd-Order – – 0.157 0.173
serial
correlation
Sargan test – – 0.605 0.482
significance level, respectively. Figures in parenthesis () are absolute t/z ratios while figures
in brackets [ ] are p-values.
impact of liberalization on import growth. More importantly, GMM

II represents a more appropriate specification for the impact of trade lib-
eralization on import growth in SSA as it allows for a ‘jump’ in growth of
imports following liberalization. The price elasticity of demand is 0.30
across different estimation techniques, while the income elasticity of
demand ranges from 1.04 to 1.26.
Lanre Kassim 61
4 Discussions and comparisons
So far empirical results on export growth and import growth regressions

are specified in both static and dynamic forms, to ascertain whether
results are robust to different estimation techniques. Both qualitative
and quantitative measures of liberalization have been applied while
adopting a real exchange rate measure of tradable goods only. In general,
imports were found to grow faster than exports in the post-liberalization
era, giving a preliminary indication that trade balance deteriorated in
Sub-Saharan Africa.
In the export growth regression, all variables show the expected signs
and are statistically significant except for the export duty variable. The
insignificant effect of export duties may be caused by the presence of a
binding quota set by the exporting country. Reduction in export taxes
will only significantly raise exports if there is an increase in export quo-
tas. The measurement of export quotas in SSA is almost impossible,
but the liberalization dummy variable represents a reasonable proxy, as
it picks up the effect of non-tariff barriers. Indeed, this variable has a
positive significant effect on export growth.
Also, the price elasticity of demand for SSA exports is found to be
inelastic, which suggests exports from the region are mainly composed
of agricultural commodities. The income elasticity of demand is, how-
ever, fairly elastic. This may be explained by the fact that major exports
of some SSA countries consist of petroleum products and precious
stones, such as diamonds and gold, which are more sensitive to changes
in income. To investigate this hypothesis, regressions were run on
two sets of countries: agricultural exporters (AG) and non-agricultural
exporters (NAG). Results show income elasticity of export demand for
NAG is 1.97 while that of AG is 0.34, albeit insignificant. (Results avail-
able on request.) In addition, liberalization raises the price elasticity of
export demand, which suggests that resources are easily transferred to
sectors whose goods have a high price elasticity of demand. On the
other hand, liberalization does not have a significant impact on income
elasticity of demand, although the coefficient has the correct sign.
Further, the timing impact of trade reforms on export growth in inves-
tigated, and the impact of liberalization on the growth of export is
found not to be instantaneous. This is plausible when SSA exports have
a low price elasticity which means that demand for these goods is less
responsive to price changes. A drop in the price of exports as a result
of dismantling export taxes may not significantly alter demand imme-
diately. On the supply side, SSA countries are slow to increase their
productive capacity due to poor infrastructure, inadequate equipment

and social insecurity and similar factors. In order to check for robust-
ness, parameters are estimated using the random effects estimator and
generalized methods of moment (GMM). Both produce similar results
in terms of signs and significance of variables.
All variables in the import growth model show expected signs and are
statistically significant. The import duty coefficient is relatively high;
but this is not surprising as most of the countries were heavily protected
before the adoption of free trade. The income elasticity of demand for
imports is fairly elastic as countries import manufactured goods. How-
ever, the price elasticity of demand for imports is inelastic, and this can
be attributed to the increasing importation of food products in SSA. The
Food and Agricultural Organization (2011) found that since the 1980s
agricultural imports have exceeded agricultural exports in SSA due to the
rising population growth, low agricultural productivity and weak insti-
tutions. The major imports of some SSA countries include petroleum
products which are less sensitive to price changes. The short-run impact
of trade liberalization on import growth is immediate, as a positive sig-
nificant growth in imports was found in the year after trade reforms
began. This is expected as the removal of quantitative restrictions and
lowering of tariff rates tends to cause a sharp rise in demand for inputs
and equipment used to facilitate domestic production.
The results of this study are compared with Santos-Paulino
and Thirlwall (2004) and Santos-Paulino (2007). Santos-Paulino and
Thirlwall (2004) examine the impact of trade liberalization on export
and import growth across 22 developing countries (DC) from 1972
to 1998 while Santos-Paulino (2007) performs the same analysis for
17 Least Developed Countries (LDC) from 1970 to 2001. However,
in this study GMM results from base specifications (i.e., equation 6
and 7 for exports and equations 11 and 12 for imports) are used for
comparison as they are similar to those used in the aforementioned
studies (see Table 3.8).
The most consistent finding is that liberalization increased the growth
of imports above that of exports across all samples of countries. So the
trade balance of these countries worsened in the post-liberalization era,
ceteris paribus. Further, import duties have a significant impact on import
growth across the three sets of countries, and the coefficient for SSA is
the highest. This might be because the import duty variable in this study
is in growth form, while in the other two studies the variable appears in
levels. Export duties have a significant effect on the growth of exports
across all samples of countries except those in SSA.
Lanre Kassim 63
Table 3.8 Comparing results
Current study Santos-Paulino Santos-Paulino

(Sub-Saharan and Thirllwall (2007)
Africa) (2004)
Liberalization Exports 2.67∗ 1.91∗ 0.50∗

dummy Imports 4.33∗ 9.10∗ 1.06∗
Duties Exports −0.12 −0.16∗ −0.19∗
Imports −0.56∗ −0.43∗ −0.17∗
Price elasticity Exports −0.28∗ −0.14∗ −0.03∗
Imports 0.30∗ 0.16∗ 0.11∗
Income Exports 1.43∗ 1.42∗ 1.72∗
elasticity Imports 1.05∗ 2.60∗ 1.68∗
Liberalization Exports −0.25∗ −0.07 −0.02∗
on price elasticity Imports 0.20∗ −0.40∗ −0.08
Liberalization Exports −0.25 1.41∗ 0.15∗
on income elasticity Imports 0.35 0.93∗ 0.22∗
Note: ∗ denotes significant coefficient at either 10%, 5% or 1% significance level.
Also, price elasticity of demand for exports and imports is highest in

SSA compared to the other sets of countries. This can be explained by
the different measures of relative prices adopted. While Santos-Paulino
and Thirlwall (2004) and Santos-Paulino (2007) applied a real exchange
rate measure which contains non-tradable goods, this study adopted
the more appropriate exchange rate measure between exportables and
importables. Income elasticity of demand for exports is approximately
between 1.4 and 1.7 across the different sets of countries; while income
elasticity of demand for imports is lowest in Sub-Saharan Africa, as the
region is fast becoming an importer of food items. Across DC and LDC,
trade liberalization had a significant impact on the price and income
elasticity of demand; but in SSA, liberalization only significantly raised
price elasticity of demand for exports and imports.
5 Conclusions and policy implications
This study has investigated the impact of trade liberalization on export

growth and import growth in SSA. A dataset of 28 countries from 1981 to
2010 was used, and panel data and time series/cross-sectional estimators
were applied. Based on the empirical results obtained, trade liberal-
ization has significantly raised the growth of exports in Sub-Saharan
Africa; however, import growth has risen faster by approximately two
percentage points, which shows prima facie evidence that trade bal-
ance in the region deteriorated. This finding is robust to the different
estimation techniques adopted and also consistent with other studies

on developing countries (Santos-Paulino and Thirlwall, 2004) and least
developed countries (Santos-Paulino, 2007). In addition, the price elas-
ticity of demand for exports is low in SSA, suggesting that exports in the
region still consist mainly of agricultural commodities. Import duties
expectedly have a significant negative impact on import growth, but
no such significant relationship was found between export growth and
export duties.
Furthermore, the price elasticity of demand for imports in SSA is
inelastic as the region is fast becoming an importer of food items as
a consequence of rising population growth (see FAO, 2011). Evidence is
found that trade liberalization increases the price elasticity of demand
for exports and imports but does not have a significant impact on
income elasticity. Besides, the timing investigation of trade reforms
shows the short-run impact of trade liberalization on import growth is
instantaneous, while for export growth it is much slower.
The results in this study have important policy recommendations
with respect to future trade reforms in Sub-Saharan Africa. In partic-
ular, they suggest the need to improve the level of infrastructure in
the region. Infrastructural development would increase the response
of exports to trade reforms in the early stages and could potentially
curb the risk of balance of payments problems stemming from imports
growing faster than exports. Also, further trade reforms should be
accompanied by efficient export incentive schemes, perhaps in the form
of exempting exporters from transportation taxes and duties on their
inputs.
Annexes
Table 3.A.1 Classification of countries
All countries Agricultural Non-agricultural

goods exporters goods exporters
Benin Benin Botswana

Botswana Burkina Faso Cameroon
Burkina Faso Burundi DRC
Burundi Côte d’Ivoire Gabon
Cameroon Ethiopia Ghana
Côte d’Ivoire Gambia Lesotho
DRC Kenya Mauritius
Ethiopia Madagascar Namibia
Gabon Malawi Nigeria
Gambia Mali Sierra Leone
Lanre Kassim 65
Ghana Rwanda South Africa

Kenya Senegal Zambia
Lesotho Swaziland Zimbabwe
Madagascar Togo
Malawi Uganda
Mali
Mauritius
Namibia
Nigeria
Rwanda
Senegal
Sierra Leone
South Africa
Swaziland
Togo
Uganda
Zambia
Zimbabwe

Real export growth epg git : Exports of goods and services; annual percentage
growth rate (constant 2000 US dollars). Source: World Development Indicators,
2011

Real import growth impg git : Imports of goods and services; annual percentage
growth rate (constant 2000 US dollars). Source: World Development Indicators,
2011
Liberalization dummy (libdumit ): This variable takes the value of one when unin-
terrupted trade reforms began in a SSA country till the end of the sample period
and zero beforehand. Source: WTO Trade Policy Reviews, 2011

Export duties epd dit : These are all levies on goods being transported out of a
country or services being delivered to non-residents by residents. Source: IMF,
2011 and World Development Bank, 2011

Import duties impd dit : These are all levies collected on goods that are enter-
ing a country or services delivered by non-residents to residents. Source: IMF
Government Financial Statistics, 2011 and World Development Indicators, 2011

Domestic income growth gdpg git : This is the annual growth rate of domestic GDP
(constant 2000 US dollars). Source: World Development Indicators, 2011

World income growth wgdpg git : This is the annual percentage growth rate of
World GDP (constant 2000 US dollars). Source: World Development Indicators,
2011
Real exchange rate (rerrit ): This is the nominal exchange rate multiplied by the ratio
of domestic export prices to foreign import prices. Source: World Development
Indicators, 2011
Aid (aid
dit ): This is defined as the annual net official development assistance and
official aid received by a SSA country (constant 2010 US dollars). Source: World
Development Bank, 2011
Figure 3.A.1 Definition and sources of variables

Table 3.A.2 Import growth regression with foreign aid
Fixed effects GMM
l.impg – 0.12
(1.81)∗∗∗
libdum 5.27 4.57
(3.50)∗ (3.35)∗
gdpg 1.17 1.01
(7.66)∗ (3.95)∗
rer 0.31 0.30
(6.00)∗ (4.33)∗
impd −0.68 −0.57
(3.03)∗ (2.12)∗∗
aid 0.63 0.78
(0.25) (0.45)
Diagnostic tests
Wald test/F-stat [p-value] [0.0000] [0.0000]
1st-Order serial correlation – 0.005
2nd-Order serial correlation – 0.217
Sargan test – 0.420
Note: ∗ , ∗∗ and ∗∗∗ indicate that a coefficient is statistically significant at 1%,

5% and 10% significance level, respectively. The figures in parenthesis ( ) are
absolute t/z ratios while figures in brackets [ ] are p-values.
Notes
1. See Annex 1 for the list and classification of countries.
2. See Annex 2 for the definition and sources of variables used.
3. As a check for robustness, a variable is added for foreign aid in the import
growth regression. However, a statistically significant coefficient is not found.
This can be explained by the fact that donors pay directly to international
companies for imported goods rather than paying to the government of
SSA countries (see Annex 3 for results).
References
Agosin, M. R. (1991), ‘Trade Policy Reform and Economic Performance: A Review
of the Issues and Some Preliminary Evidence’, UNCTAD Discussion Papers,
no. 41.
Ahmed, N. U. (2000), ‘Export Responses to Trade Liberalisation in Bangladesh:
A Co-Integration Analysis’, Applied Economics, vol. 32, no. 8, pp. 1077–1084.
Babatunde, A. M. (2009), ‘Can Trade Liberalisation Stimulate Export Performance
in Sub-Saharan Africa?’ Journal of International and Global Studies, vol. 2, no. 1,
pp. 68–92.
Lanre Kassim 67
Bleaney, M. (1999), ‘Trade Reform, Macroeconomic Performance and Export

Growth in Ten Latin American Countries, 1979–1995’, Journal of International
Trade and Economic Development, t vol. 8, no. 1, pp. 89–105.
Boylan, T. A. and Cuddy, M. P. (1987), ‘Elasticities of Import Demand and Eco-
nomic Development: The Irish Experience’, Journal of Development Economics,
vol. 26, no. 2, pp. 301–309.
Clarke, R. and Kirkpatrick, C. (1992), ‘Trade Policy Reform and Economic
Performance in Developing Countries: Assessing the Empirical Evidence’,
in R. Adhikari, C. Kirkpatrick and J. Weiss (eds.), Industrial and Trade Pol-
icy Reform in Developing Countries, Manchester: Manchester University Press,
pp. 56–73.
Dean, J. M., Desai, S. and Riedel, J. (1994), ‘Trade Policy Reforms in Developing
Countries since 1985: A Review of the Evidence’, World Bank Discussion Papers,
no. 267.
Food and Agricultural Organization (2011), Why Has Africa Become a Net Food
Importer: Explaining Africa Agricultural and Food Trade Deficits, Rome: United
Nations Publications.
General Agreement on Trade and Tariffs (1991–1994), Trade Policy Reviews,
(Various Countries).
Greenaway, D. and Sapsford, D. (1994), ‘What Does Liberalisation Do for Exports
and Growth’, Review of World Economics (Weltwirtschaftliches Archiv), vol. 130,
no. 1, pp. 152–174.
Greenaway, D., Morgan, W. and Wright, P. (2002), ‘Trade Liberalisation and
Growth in Developing Countries’, Journal of Development Economics, vol. 67,
no. 1, pp. 229–244.
International Monetary Fund (2011). Government Financial Statistics, Washington,
DC: IMF.
Jenkins, R. (1996), ‘Trade Liberalisation and Export Performance in Bolivia’,
Development and Change, vol. 27, no. 4, pp. 693–716.
Joshi, V. and Little, I. M. D. (1996), India’s Economy Reforms 1991–2001, Oxford:
Oxford University Press.
Mah, J. S. (1999), ‘Import Demand, Liberalization, and Economic Development’,
Journal of Policy Modeling,
g vol. 21, no. 4, pp. 497–503.
Melo, O. and Vogt, M. G. (1984), ‘Determinants of the Demand for
Imports of Venezuela’, Journal of Development Economics, vol. 14, no. 3,
pp. 351–358.
Michealy, M., Papageorgiou, D. and Cholski, A. (1991), Liberalising Foreign Trade,
Vol.7: Lessons of Experience in the Developing World, Oxford: Basil Blackwell.
Olofin, S. O. and Babatunde, M. A. (2009), ‘Price and Income Elasticities of Sub-
Saharan African Exports’, African Journal of Economic Policy, vol. 16, no. 2,
December.
Pacheco-Lopez, P. (2005), ‘The Impact of Trade Liberalisation on Exports, Imports,
the Balance of Payments and Growth: The Case of Mexico’, Journal of Post
Keynesian Economics, vol. 27, no. 4, pp. 595–619.
Ratnaike, Y. C. (2012), ‘Is There an Empirical Between Trade Liberalisation and
Export Performance’, Economic Letters, vol. 117, no. 1, pp. 375–378.
Santos-Paulino, A. U. (2002a), ‘Trade Liberalisation and Export Performance in
Selected Developing Countries’, Journal of Development Studies, vol. 39, no. 1,
pp. 140–164.
Santos-Paulino, A. U. (2002b), ‘The Effect of Trade Liberalisation and Imports in

Selected Developing Countries’, World Development, t vol. 30, no. 6, pp. 959–974.
Santos-Paulino, A. U. (2007), ‘Aid and Sustainability under Liberalisation in Least
Developing Countries’, The World Economy, vol. 30, no. 6, pp. 972–998.
Santos-Paulino, A. U. and Thirlwall, A. P. (2004), ‘The Impact of Trade Liberal-
ization on Export Growth, Import Growth and the Balance of Payments of
Developing Countries’, Economic Journal, vol. 114, no. 493, pp. F50–F72.
Shafaeddin, M. (1994), ‘The Impact of Trade Liberalisation on Exports and GDP
in Least Developed Countries’, UNCTAD Discussion Papers, no. 85.
Thirlwall, A. P. (1979), ‘The Balance of Payments Constraint as an Explanation
of International Growth Rate Differences’, Banca Nazionale del Lavoro, vol. 32,
no. 128, pp. 35–53.
Thomas, V., Nash, J. and Edwards, S. (1991), Best Practises in Trade Policy Reform,
Oxford: Oxford University Press.
UNCTAD (2004), Economic Development in Africa: Debt Sustainability, Oasis or
Mirage? New York: United Nations Publications.
UNCTAD (2008), Economic Development in Africa: Export Performance Follow-
ing Trade Liberalisation: Some Patterns & Policy Perspectives, New York: United
Nations Publications.
Wacziarg, R. and Welch, K. H. (2008), ‘Trade Liberalisation and Growth: New
Evidence’, The World Bank Economic Review, vol. 22, no. 2, pp. 187–231.
World Bank (2012), World Development Indicators, Washington, DC: The
World Bank.
World Trade Organization (1995–2011), Trade Policy Reviews (various countries).
Part II
Measuring Trade Potential: The
Gravity Model Approach
4
Market Integration in the
ECCAS Sub-Region
Désiré Avom and Mouhamed Mbouandi Njikam
Introduction
The promotion of South-South trade is generally based on the under-

lying assumption that there is untapped potential in trade between
developing countries (DCs), as the former Director-General of the World
Trade Organization (WTO) Pascal Lamy (2006) has stated. Besides enter-
ing into negotiations on market access and preferential treatment in
trade with developed countries or economic groupings, developing
countries – particularly African countries – should promote regional
trade.
The issue of market access should not be limited to tariff cuts applied
by developed countries to products from developing countries. South-
South trade is also a vital component. Moreover, this area still has
untapped potential. We should not forget that 70% of customs duties
are derived from trade between developing countries (Lamy, 2006).
In this context, promoting trade between developing countries is
a promising solution in many respects. Many authors as well as
some international bodies like the International Trade Centre (ITC)
(UNCTAD/WTO) or OECD also bring the debate down to the regional
level, denouncing barriers to South-South cooperation (market integra-
tion), which are even more serious and harmful when customs duties of
such countries are very high and trade potential is untapped. In addi-
tion, cooperation can promote local development and enhance the
competitiveness of developing countries’ relatively small markets.
Concerning the need to increase trade flows, regional economic
cooperation has been regarded since the 1960s as a solution to the
development problems faced by poor countries, particularly those in
71
72 Market Integration in ECCAS
Sub-Saharan Africa (SSA). Regional economic cooperation may take sev-

eral forms, ranging from mere sector cooperation to the transfer of
sovereignty (Hugon, 2002). Thus, the African Union’s goal, as set out in
its Constitutive Act, is to ensure the continent’s economic and political
integration. The Act reaffirms a commitment to establishing an African
Economic Community through the progressive convergence of spaces
and patterns of integration of Regional Economic Communities (RECs),
with the aim of ‘building an Africa that is integrated, prosperous and at
peace, and that is a dynamic force on the world stage’. Despite progress
made in some areas of cooperation by some RECs over the past two
decades, the pace of progress is generally considered slow, in the context
of Africa’s numerous development challenges and the changing global
economic environment.
The concept of regional integration refers to economic, political
and/or social rapprochement among different partner countries. In 1961,
the Hungarian economist, Bela Balassa, proposed a five-stage regional
integration plan. A first stage is to establish a free trade area among
countries that decide to eliminate customs duties and protectionist mea-
sures relating to products within the area (free movement of goods and
services). Examples include the North American Free Trade Agreement
(NAFTA) and the Association of South East Asian Nations (ASEAN).1
A second stage is to establish a customs union characterized by the insti-
tution of a common trade policy with respect to third-party countries
(common external tariff).2 A third stage is to establish a common mar-
ket, a customs union that includes the free movement of people and cap-
ital. An example is the Mercado Común del Sur (MERCOSUR).3 A fourth
stage is an economic union, a common market characterized by the har-
monization of economic policies. An example is the European Union
(EU), which is considered to have achieved full economic integration.
A final stage provides for the institution of a common supranational
authority.
Integration can take many forms. No longer regarded as successive
and complementary different types of integration may reflect contradic-
tory processes: integration through capital flows to promote uniform
management regulations in companies as well as states; integration
through production, which requires resources invested toward specific
outcomes, usually by countries with outstanding training systems and
innovation; monetary integration through the use of a single cur-
rency by states making up a block, in order to optimize intra-regional
trade; integration through harmonization rules, minimum protection
and market integration – that is the removal of national barriers in
Désiré Avom and Mouhamed Mbouandi Njikam 73
transactions between the member states to enable the free movement

of factors of production. This study looks at market integration in Cen-
tral Africa, with a view to encouraging the removal of national barriers
to transactions between ECCAS members, and thus intra-regional trade
flows between them. It also evaluates their trade potential.
Reasons for renewed enthusiasm for regionalism in Sub-Saharan Africa
(SSA) in the 1990s are largely analyzed in existing literature. De Melo
et al. (1993) indicate three major reasons: the European integration suc-
cess story, the disappointing outcome of the Uruguay Round of the
General Agreement on Tariffs and Trade (GATT) and the United States’
eagerness to establish free trade areas (FTA) such as NAFTA. Krugman
(1991) adds the complexity and lack of transparency of GATT tariffs and
quotas. These various reasons also emerge in the analyses conducted by
economists pointing to the expected impacts of regional integration.
The fundamentals defined by Viner (1951) remain the basis of analyses.
Central Africa is unusual in that it has two regional groupings:
CEMAC, an economic and monetary union with six member countries4 ;
ECCAS, a free trade area comprising the six CEMAC member countries,
Angola, São Tomé and Príncipe, DRC and Burundi.
The Central African sub-region is one of the least efficient in terms of
trade flows between member countries. From 1995 to 2010 (UNCTAD,
2012), the average value of exports within ECCAS was the lowest of all
trade agreements in SSA as it did not exceed USD 320 million or 1%
of member countries’ total exports. Cameroon is the leading exporter,
accounting for more than half of the grouping’s total exports, fol-
lowed by Gabon (19%) and Congo (13%). Imports are more balanced
and the main importers are DRC (22%), Gabon, Congo and Chad. The
largest flows are Cameroon’s exports to DRC, Congo, Gabon and Chad.
By way of comparison, exports within the Economic Community of
West African States (ECOWAS) during the same period averaged USD
5.4 billion per year, or 9.4% of total ECOWAS exports. The leading
ECOWAS exporters are Côte d’Ivoire and Nigeria, together accounting
for about 70% of the community’s exports, followed by Senegal with
a little less than 10%. The main importers are Côte d’Ivoire, Nigeria
and Ghana. Although CEMAC countries have a huge trade potential,5
intra-regional trade between them remains limited, due to the narrow
national markets, a relatively insignificant intra-regional trade flows
compared to other developing regions, due to trade barriers. Empiri-
cal studies show that despite high economic growth forecasts – related
to petroleum development in Chad and Equatorial Guinea – CEMAC
is rather too small and weak to steer the sub-region on the path of
steady growth based on robust intra-regional trade relations, due to

lack of production diversification and infrastructural and institutional
problems (Avom, 2005). Trade cooperation in the sub-region, therefore,
depends on the dynamism of ECCAS which covers a larger geographical
space and is capable of developing a more positive trade potential than
CEMAC.
In this respect, the main objective of this study is to compute, using
a gravity model, the trade potential of ECCAS member states. Trade
potential is computed by simulation using gravity model estimates for
the 1995–2010 period in order to highlight trends in intra-regional
flows. Following this introduction, Section 1 deals with the main styl-
ized facts of intra-ECCAS trade; Section 2 looks at the gravity model,
its theoretical basis and empirical relevance; and Section 3 presents the
econometric techniques used, the data sources consulted, the analysis
carried out, as well as the findings and their interpretation. The conclu-
sion outlines lessons learned and economic policy recommendations for
strengthening intra-ECCAS trade.
1 Intra-ECCAS trade
The ECCAS member countries generally practice one-crop farming

so their production is not diversified. Their economic performance
depends on the activities of vulnerable sectors, particularly the pro-
duction of one or a few commodities. Consequently, their export
diversification indicator6 is low and below the African average (UNECA,
2008). Trade within the sub-region can be analyzed at three levels: trade,
product structure and geographical orientation.
Trade levels
Foreign trade trends in the ECCAS member countries between 1995
and 2012 suggest weak intra-community trade, compared to other RECs
worldwide. Figure 4.1 clearly shows that ECCAS foreign trade is minimal
and by far the weakest for all the sub-regions. Intra-ECCAS trade in 1995
represented 1.2% or USD 134.98 million of total trade, peaking at 1.7%
in 1998. During the period under review, intra-ECCAS trade remained
below 1% and represented 0.87% of total trade in 2012, that is about
USD 1,111.340 million in absolute value. The increase of about USD
880 million remains relatively insignificant, compared to the total vol-
ume of trade. By way of comparison, trade between ECOWAS member
countries rose during the same period from 10.34% to 17.15% of total
trade, or USD 2,294.44 million to USD 11,667.27 million – representing
80
70
60
50
40
30
20
10
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
ECCAS ECOWAS SADC
UE-27 ALENA Mercosur
Figure 4.1 Total trade within communities, percentages of total trade

Source: Drawn from UNCTAD data.
an increase of more than USD 9,373 million in terms of volume. This

trend clearly shows that the level of intra-ECCAS trade is low. Despite
an increase in value, the share of intra-community trade remains min-
imal compared to total trade. By way of further comparison, trade
within the Southern African Development Community (SADC) repre-
sented more than 20 times the value of intra-ECCAS trade, increasing
from USD 6,621.82 million to USD 27,134 million. Compared to total
trade, there was a slight decrease in intra-SADC trade from about 15% of
total trade in 1995 to about 13.95% in 2012. The trend however remains
upwards. Globally, the EU and NAFTA are by far the most prolific RECs
in terms of internal trade flows. Almost 70% of EU trade, and 50%
of NAFTA trade, is carried out between member countries. This trend
may be explained by the quality of products traded between member
countries.
Further, comparison of ECCAS trade performance with that of other
African trade groupings over the same period shows it is the weak-
est (see Figure 4.2). For example, average intra-regional trade (exports
and imports) stood at USD 515.29 million, 10 times less than that of
ECOWAS (USD 5,553.27 million) and 20 times less than that of SADC.
Even trade within the West African Economic and Monetary Union
(WAEMU) is more than twice that of ECCAS.
35000 32931.97
30000
25000
20000
15000 12988.33
10000
5523.57
5000
1638.43
515.28
0
Intra-ECCAS Intra-ECOWAS Intra-SADC
Intra-Mercosur Intra-WAEMU
Figure 4.2 Average trade within communities 1995–2010 (USD millions)

Product structure
The distinctive feature of central African economies is that they rely
heavily on commodities such as agricultural produce, foodstuff and bev-
erages, oil and gas products and base metals According to Hammouda
(2005), composition by product shows ‘the continent is trapped in
the cash crop inclusion snare’, where agricultural and mining prod-
ucts account for nearly 70% of total exports (Table 4.1), although global
demand structure is developing in the opposite direction, to the detri-
ment of African exports. At the same time, imports mainly comprise
manufactured goods.
The main trading partners of the countries of the sub-region are the
EU, China and the United States. All ECCAS member countries have
abundant natural wealth. Most export oil, minerals and agricultural
products, and thus suffer deteriorating terms of trade as victims of the
‘raw material curse’ (Table 4.1).
Figure 4.3 shows averages for each country’s trade with the com-
munity between 1995 and 2010, and fairly illustrates the sub-regional
trade imbalance. Regarding imports, only Cameroon and Gabon have
acceptable levels. Gabon imports the most from its sub-regional part-
ners, followed by DRC although this level remains low for its size. The
Central African Republic, Chad and São Tomé and Príncipe import very
little. Angola’s imports account for only 20% of intra-regional exports.
Regarding exports, only Cameroon shows an upward, constant and
Table 4.1 ECCAS exports and imports in 2009, percentages by product
Exports Imports
Agricultural products 13.9 15.9

Mineral products 54.9 10.1
Manufactured products 26.9 71.2
Source: WTO, World Trade Statistics 2010.
4000 3660
3500 3320
3000 2690
2460
2500
2000
1500 1310
1113
1000 920 860
670 600 490
500 320
140 0 0 120 190 6,7 34 76 86 80
0
la
go
on
AR
da
d
nd
ne
ha
ST
M
R
go
on
ab
n
ru
D
C
ui
wa
C
An
G
Bu
Exports Imports
Figure 4.3 Exports and imports within ECCAS, averages in USD thousands
Note: CAR: Central African Republic; CMR: Cameroon; DRC: Democratic Republic of the
Congo; STP: São Tomé & Prìncipe.
sustained trend over time. The Central African Republic, Chad and São
Tomé and Príncipe contribute almost zero to sub-regional exports.
Geographical orientation
The last aspect of ECCAS’ trade is the geographical orientation of
imports and exports. Graph 4.5 shows the trend in total ECCAS trade
with some partners. The volume of foreign trade has increased steadily
since 1995, when the WTO replaced the GATT and trade liberalization
worldwide accelerated with the EU, Africa’s historic trading partner, at
the forefront. A major event since then is China’s breakthrough and
the general vibrancy of its recent trade with Africa. During the period
under review, beside the EU and China, ECCAS’ main partner in terms
of imports was the United States, which has intensified its trade with
African countries since the African Growth Opportunity Act (AGOA) in
2000. Africa is the fourth trading partner for ECCAS, ranking last in
terms of the origin of its own imports, due to a low level of diversifi-
cation and similarity in national production structures (see Figure 4.4).
45,000,000
40,000,000
35,000,000
30,000,000
25,000,000
20,000,000
15,000,000
10,000,000
5,000,000
0
1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
China EU27 United States Africa ECCAS
Figure 4.4 Total ECCAS trade with partners (USD thousands)

Again, these statistics show a very slow market integration process in

the sub-region.
After illustrating the low level of market integration in the ECCAS sub-
region, the next section empirically identifies the determinants of trade
flows between countries, using an econometric gravity model to ensure
a more rigorous analysis of the sub-region’s trade potential.
2 The gravity model
Various techniques and methods can be used to evaluate regional trade.7

The gravity model of trade, one of such methods, is a simple tool that
often yields good bilateral trade forecasts.
Theoretical basis
The gravity model is a generic name for the family of quantitative
models developed by Astronaut Stewart in 1940, which have been very
successful since the early 1960s (Evenett and Keller, 2002). The analyt-
ical framework assumes (i) profit maximization through competition
by monopolistic enterprises; (ii) constrained utility maximization by
consumers; and (iii) specialization in the supply of goods between coun-
tries (Anderson, 1979; Bergstrand, 1989; Anderson and Van Wincoop,
2003; Helliwell and Schembri, 2005). Since Tinbergen (1962), the gravity
model has become a popular tool for empirical analysis of foreign trade.
The model, which was initially deduced to analyze bilateral trade flows
between countries, is based on the gravity principle, which holds that
the intensity of trade between two countries is proportional to their GDP
and inversely proportional to the distance separating them. The theo-
retical basis for the gravity model was developed by Anderson (1979),
Bergstrand (1985 and 1989), Deardorff (1995) and Evenett and Keller
(2002). The gravity model proposed in this study is inspired by empirical
literature, in particular by Fontagné et al. (2002).
Specification and basis of variables

The gravity model considered for this work is generally set out as
follows:
log Xijt = α0 + 1 log Yijt + 2 log Zij + 3 Vijt + 4 Wijt + εijt (1)
Here Xijt indicates the total exports of country i to country j during

year t. Yijt is the vector of independent variables for time t for the two
trading partners i and j. GDP and per capita GDP variables are part of
this vector.
Zijt is the vector of the independent variables which change according
to partner countries, but are constant over time. These variables include
the distance in kilometers between the capitals of both partner countries
and the area of partner countries in square kilometers. Distance is one of
the variables that are difficult to model. Intuitively, it can be understood
as a transport cost estimate. Thus, the longer the distance, the higher the
transport cost, the less countries trade with one another. In general, dis-
tance includes transaction and information costs. It may also explain
the structural differences existing between two countries, including cul-
tural, sociological and language differences. However, the fundamental
problem posed by the distance variable is how to measure it. In empir-
ical literature, three types8 of measurements are usually adopted: actual
mileage, adjusted actual mileage and great circle (Luo, 2001). In this
study, the great circle distance is used; considering the capitals of coun-
tries as trade poles, it calculates the distance between such poles, using
geographical coordinates (longitude and latitude). This measurement of
distance gives good results only under the condition of ceteris paribus (all
other things being equal).
Vijt indicates the qualitative or binary independent variables that
change over time, and according to the partner countries i and j. This
is the case of the ‘AIR’ variable that takes the value 1 if countries i and
j are part of the same regional integration agreement in the year t, and
the value 0 otherwise. It is also the case of the ‘MU’ variable that takes
the value 1 if countries i and j have a common currency in the year t
and the value 0 if not.
Wijt indicates the qualitative or binary variables that are specific to
partner countries i and j but are constant over time. This is the case of
the Encl variable that takes the value 0 if countries i and j both have a
sea front, the value 1 if one of the countries is landlocked and value 2
if both countries are landlocked. Two other variables are introduced to
consider the effects of geographical or linguistic closeness on bilateral
trade. The first is the frontier variable that takes the value 1 when both
countries have a common frontier and the value 0 if not. The second,
language, takes the value 1 when both countries have a common official
language and 0 if not. The parameter α0 and the model constant, 1 , 2
and 3 are vectors of coefficients associated with the different indepen-
dent variables. Lastly, εijt is the random part of the gravity model. The
log-linear form is adopted for continuous variables, to help to interpret
the coefficients directly as elasticities.
Computing trade potential

The calculation of trade potential relies on results from the grav-
ity model. The model used to estimate determinants of trade in
ECCAS includes some countries that are not ECCAS member countries.
Three approaches are used to measure trade potential.
The first and simplest approach is to generate the residual estimation
and to compute its deviation from zero. Thus, if the residue is positive,
the actual value is greater than the potential value. If it is negative, the
potential value is greater than the actual value. If it is zero, the variable
stands at its simulated level.
The second approach is to calculate the trade potential using an
observed/predicted trade ratio. However, this method shows that trade
potential is biased. This bias is justified by the fact that the simulated
model is based on a larger sample, that is, it contains other coun-
tries whose trade potential will not be calculated. This flaw is even
more significant when investment is not considered as an indepen-
dent variable (Fontagné et al., 2002). Another argument put forward
is the heterogeneity of the sample. The issue of purchasing power par-
ity also arises. The GDP of the countries in the sample is estimated
in dollars, suggesting that a certain exchange rate is used. In fact,
given the differences in the standards of living of the countries in
the sample, standardizing their GDP on the basis of an exchange rate
derived solely from the values of their currencies with respect to the
dollar could distort the results of the estimate (Avom and Mignamissi,
2013).
To correct these distortions, a third two-step approach borrowed from
Fontagné et al (2002) was adopted. The first step is to estimate a bilateral
trade equation using the gravity model and the second uses the equation
in the simulation for countries in the sample or for countries outside the
sample which are trading with countries in the sample. Thus, the first
step is to adjust the simulated trade flows as follows:

Xij
X −Xij
j ij

X∗ ij = X∗ i = X∗ ij (2)
−X
X j
j ij ij
−
Here X ij represents gross simulated bilateral trade flows, that is, those
obtained directly from the model estimate, Xij the level of bilateral trade
observed and X∗ ij adjusted simulated trade flows, with:
log X̂ijt = α̂0 + ˆ 1 log Yijt + ˆ 2 log Zij + ˆ 3 Vijt + ˆ 4 Wij (3)
The second step is to compute the trade potential (TP) for a given
period as the arithmetic average of gross simulated flows and adjusted
simulated flows:
1
∗
PCijt = (X ijt + X ijt ) (4)
2

Xij −Xij

ij PCijt = 1 (X
Xij
X∗ ij =
j
X∗ i = X∗ ij X
ijt + X∗ ijt ) (5)

Xij −Xij
j
2
j
3 Estimation and analysis of findings
Presentation of sample data

Access to UNCTAD’s database made it possible to obtain foreign trade
data for the various countries in the sample. Data on GDP was obtained
from the World Bank’s World Development Indicators 2012. The grav-
ity model estimation was carried out using a sample of 50 countries
(see Annex) for the 1995–2010 period. However, only bilateral trade
by ECCAS members with the rest of the countries is considered as a
dependent variable. This specification gives a total of 7,840 observa-
tions, relating specifically to 10 countries and 49 bilateral trade flows
by country for 16 periods.
Gravity model estimation method

Gravitational equations are generally estimated as a cross-section
(N countries, N partners, at a given date). However, it is not uncom-
mon for all the time data to be used in the same estimation. The data
for each pair of countries and each year is simply stacked and the equa-
tion is estimated using ordinary least squares (OLS). OLS also give robust
coefficients, especially when the dependent variable contains a low pro-
portion of zeros (Fontagné et al., 2002). In addition, OLS seem to give
a good estimation of the trade potential, which is the main objective of
this study. Thus, in line with Fontagné et al. (2002), regression by ordi-
nary least squares will be used, given the low percentage of zero values
in observations.
Findings and interpretation

Table 4.2 shows findings of the gravity model estimation using the
ordinary least squares method on pooled data from 1995 to 2010.
All the estimated coefficients are statistically significant and their
signs are in conformity with expectations. The coefficients of the GDP
of the exporting country (ln pibi ) and the importing country (ln pib
bj ) are
positive and significant at 10%. They have a strong positive correlation
with mean trade flows in the ECCAS sub-region, although the effect is
Table 4.2 Gravity model estimations (1995–2010)
Independent variables Coefficients Statistic
Constant 2.11358∗∗∗ 10.67

ln pibi 0.30615082∗∗∗ 3.36
ln pibbj 0.32585766∗∗∗ 5.24
ln pibtti 0.5815983∗∗∗ 6.16
ln pibttj −0.04644593 −0.50
ln disttij −35.283912∗∗∗ −10.84
front ij 2.9380299∗∗∗ 11.42
enclij −2.1494841∗∗∗ −6.44
MUUij 2.2231777∗∗∗ 7.21
lang ij 0.45709233∗ 1.89
AIRij 0.84785246∗∗∗ 4.15
Number of observations: 7,840

Pairs by country: 784
R2 : 0.5386
Note: p < 0. 01: significant coefficients at 1%; ∗∗ p < . 0. 5: significant coefficients at 5%;
∗∗∗ p < . 0. 1: significant coefficients at 5%.
ambivalent. A 10% increase in the GDP of the exporting country leads

to an increase of nearly 3.06% in the supply of export goods and vice
versa. This elasticity is more or less equal to that of the demand for
goods by the importing country (3.25%) following the same variation
of its GDP. The impact of GDP on trade is positive but marginal. One
of the explanations would be that the GDP of ECCAS member coun-
tries is very inadequately used in producing goods that can be exported
or traded between such countries. This poses the problem of product
quality and type of products traded in the ECCAS sub-region. The pri-
ority of these countries seems rather to be to maintain macroeconomic
balances.
However, when partner countries’ wealth is considered through per
capita GDP (ln pibt i ), elasticities are respectively 0.58 and −0.04.
It should be noted here that the wealth of the exporting country mea-
sured by the per capita GDP of its population has a positive impact on
the country’s supply of goods and it is significant at 10%. In fact, a 10%
increase in per capita GDP brings about a 5.8% variation in the sup-
ply of export goods. This result is similar to that obtained by Foroutan
and Pritchett (1993). Conversely, the impact of per capita GDP on the
importing country’s demand is negative and insignificant. Thus, the
higher an importing country’s per capita GDP, the lower the elasticity
of demand for foreign goods. This confirms the idea that the higher
the per capita income of exporting countries, the higher the produc-
tion capacity and the greater the volume of goods available for export.
This result also indicates the capacity of exporting countries to pro-
duce and export goods at lower costs, everything being equal. Thus,
economies of scale are the main feature of trade between ECCAS member
countries.
This confirms our assumptions about geographical or linguistic affin-
ity. As predicted, natural barriers have a negative impact on trade.
Intra-regional trade decreases when distance increases. Its coefficient is
negative and significant at 10%. Countries located far apart trade less
than those that are close to each other, and, as predicted, isolation has
a negative impact on trade between countries. This dummy variable has
a negative and significant coefficient at 10%. The sharing of a common
frontier has a positive impact on trade and its coefficient is positive and
significant at 10%. Thus, the sharing of a common border increases trade
between countries by 2.94 times.
The variable MU ij , which indicates the use of a single currency, seems
to have a positive impact on trade, confirming the hypothesis of Rose
(2000). However, this impact is marginal because it is less than 3%,
as predicted by Rose. The AIRij, variable, which is used to measure the

impact of regional integration agreements on trade, seems to have a
positive impact on intra-regional trade. Its coefficient is positive and
significant at 10%. Both trading partners’ membership of the same REC
increases trade between them by 0.85 times. The sharing of a com-
mon official language by countries also seems to largely explain the
increase in trade between them. Its coefficient is positive and significant
at 1% and trade between countries with a common national language
increases by 0.45%. Lastly, as predicted, isolation has a negative impact
on trade between countries. Its coefficient is negative and significant at
10%. These results are used in the simulation to determine the trade
potential of ECCAS member countries (Table 4.3).
Table 4.3 Average trade flows within ECCAS 1995–2010 (USD millions)
Trade Exports Imports

flows
Observed Simulated Simulated Observed Simulated Simulated
adjusted adjusted
ANG>BDI 3.4 80.99 2.45 0.037 14.54 0.25

ANG>CMR 1.03 222.37 9.1 0.054 122.35 2.81
ANG>CNG 7.39 131.56 3.52 0.8 89.13 1.69
ANG>GBN 0.67 123.634 4.43 0.99 157.16 3.54
ANG>CAR 0.019 88.51 3.1 0.012 28.026 0.51
ANG>DRC 13.87 194.59 3.76 5.96 50.34 0.23
BDI>ANG 0.009 13.18 1.39 0.09 36.6 0.33
BDI>CMR 0.015 14.8 1.61 0.06 32.66 0.3
BDI>CNG 0.05 8.75 0.84 0.04 23.79 0.2
BDI>GBN 0.083 8.23 0.78 0.00042 41.95 0.43
BDI>CAR 0.062 5.89 0.54 0.0023 7.48 0.06
BDI>DRC 5.13 12.95 0.056 0.98 13.44 0.02
CMR>ANG 6.74 5309.84 32.36 0.604 96.92 28.56
CMR>BDI 0.013 2171.93 11.87 0.00023 10.28 2.24
CMR>CNG 21.74 3528.14 16.54 21.2 63 11.52
CMR>GBN 28.93 3315.56 14.08 4.68 111.09 32.78
CMR>CAR 13.13 2373.64 11.52 44.55 19.81 1.67
CMR>DRC 37.4 5218.48 22.048 0.034 35.59 8.41
CNG>ANG 7.24 332.85 1.27 0.16 213.81 3.21
CNG>BDI 0.013 136.15 1.15 0.012 22.67 0.26
CNG>CMR 0.028 373.83 3.82 2.38 190.78 2
CNG>GBN 0.11 207.84 1.84 4.68 245.07 1.79
CNG>CAR 0.011 148.79 1.27 0.011 43.7 0.51
CNG>DRC 4.06 327.13 2.12 0.82 78.51 0.87
GBN>ANG 5.82 6546.12 15.13 0.53 1530.56 24.02
GBN>BDI 0.000304 2677.62 5.88 0.0063 162.32 1.83
GBN>CMR 6.47 7352.04 17.38 43.77 1365.69 3.87
GBN>CNG 8.44 4349.59 8.76 5.64 994.94 12.2

GBN>CAR 3.42 2926.3 6.12 0.012 312.83 3.65
GBN>DRC 11.04 6433.49 15.44 0.46 562 6.87
CAR>ANG 0.007 1262.29 1.9 0.05 452.6 6.39
CAR>BDI 3.4 516.327 0.55 4.30 48 0.42
CAR>CMR 13.5 1417.7 0.15 13.1 403.85 1.95
CAR>CNG 0.18 838.73 1.2 1.93 294.21 3.39
CAR>GBN 0.0017 788.2 1.13 1.16 518.77 7.25
CAR>RDC 0.83 6091.17 17.31 3.03 166.19 1.67
DRC>ANG 2.04 37.14 0.93 0.044 52.58 3.51
DRC>BDI 1.26 59.49 1.88 4.30 5.57 0.17
DRC>CMR 0.037 66.81 2.6 0.083 46.92 3.01
DRC>CNG 2.02 39.53 1.01 0.066 34.18 2.04
DRC>GBN 0.03 37,14 1.26 0.082 60.27 4.22
DRC>RCA 2.14 26.59 0.63 0.012 10.75 0.57
Note: This table shows exports from the first to the second country and imports from the
second to the first (Table 4.4).
Source: Calculations by the authors using estimation data.
The findings show there is strong trade potential for most countries.
Indeed, adjusted simulated exports and imports exceed observed values.
Still, trade potential is low for countries that are geographically close.
For example, during the period under review, Angola’s predicted exports
to neighboring countries such as Burundi, Congo and DRC are less than
observed exports. Conversely, for other countries such as Cameroon,
Gabon and DRC, which are less close geographically to Angola, exports
predicted by the model are much higher than exports observed. Thus,
geographical closeness is a ‘natural’ feature that has an impact on trade
between countries.
Predicted imports between Angola and Cameroon (USD 2.81 mil-
lion) are 52 times the value of imports observed (USD 0.054 million).
Predicted imports between Cameroon and DRC (USD 8.410 million)
represent more than 247 times the value of imports observed.
The main conclusion from predictions of the gravity equation is that
distance impedes trade between ECCAS member countries. The ‘frontier
effect’ plays a decisive role in trade between these countries.
Given that trade creation potential is the difference between the trade
potential and trade actually observed, for some ECCAS member coun-
tries, episodes of trade creation (positive sign) tend to dominate those
of trade disruption (negative sign). Indeed, the leading countries (high
GDP) like Cameroon and Angola, among others, have a very high trade
potential. This situation shows the existence of potential for the cre-
ation of trade in the ECCAS sub-region. Potential for the creation of
86
Table 4.4 Trade potential within ECCAS 1995–2010 (USD millions)
Trade flows Exports Imports
ANG>BDI 2.92 1.35

ANG>CMR 5.06 1.43
ANG>CNG 5.46 1.24
ANG>GBN 2.55 2.26
ANG>CAR 1.56 0.26
ANG>DRC 8.81 3.06
BDI>ANG 0.7 0.22
BDI>CMR 0.81 0.18
BDI>CNG 0.44 0.12
BDI>GBN 0.43 0.22
BDI>CAR 0.3 0.031
BDI>DRC 2.59 1
CMR>ANG 19.51 14.58
CMR>BDI 5.94 1.12
CMR>CNG 19.14 16.77
CMR>GBN 21.5 18.73
CMR>CAR 12.32 24.01
CMR>DRC 29.72 4.22
CNG>ANG 4.26 1.68
CNG>BDI 0.8 0.13
CNG>CMR 1.92 2.19
CNG>GBN 0.97 3.23
CNG>CAR 0.64 0.26
CNG>DRC 3.08 0.84
GBN>ANG 10.47 12.27
GBN>BDI 2.94 0.92
GBN>CMR 11.92 23.82
GBN>CNG 8.6 8.92
GBN>CAR 4.77 1.83
GBN>DRC 13.24 3.66
CAR>ANG 0.95 3.22
CAR>BDI 1.97 2.36
CAR>CMR 6.75 7.52
CAR>CNG 0.69 2.66
CAR>GBN 0.56 4.21
CAR>DRC 9.6 2.35
DRC>ANG 1.48 1.77
DRC>BDI 1.57 2.23
DRC>CMR 1.32 1.55
DRC>CNG 1.51 1.05
DRC>GBN 0.65 2.15
Source: Calculations by the authors using estimation data.

trade between Angola and Cameroon is USD 4.06 million, representing

3.94 times the current exports between both countries.
Overall (save for a few member countries), the potential for creation
of trade between ECCAS member countries is, on average, at least 50%
higher than the current trade for each country, which augurs huge
opportunities for these countries. By removing tariff and non-tariff bar-
riers between ECCAS member countries, applying and strengthening
preferential agreements, countries could increase trade between them
by more than 50% of the present value.
4 Conclusion
An obvious conclusion emerges. This zone has become a more inclu-

sive space for trade in Central Africa, but the free movement of goods
and people is not yet a reality. ECCAS is playing an important role
with regard to intra-community trade, according to the gravity model.
It has a positive impact on trade. Moreover, countries in this economic
space trade more with each other. The trade potential makes this space
a ‘natural’ common market for member countries. This confirms the
hypothesis about the impact of the removal of tariff and non-tariff bar-
riers on internal trade between countries. Despite some conflicts and
institutional difficulties, ECCAS is a dynamic space for trade. Simula-
tions show there are major opportunities in these countries in terms of
indicative export potential and the potential for the creation of trade
flows between countries.
However, intra-regional trade appears to be insignificant in view of
the specificities of these countries. The intra-ECCAS export potential
(USD 2,356.76 million) on the whole represents about 7.97 times cur-
rent exports between these countries (USD 295.65 million). For its
part, the intra-ECCAS import potential (USD 181.91 million) represents
1.4 times current imports between these countries (USD 129.94 mil-
lion). This raises the sensitive issue of trade barriers. To increase trade
flows between these countries, it is necessary to strengthen the Com-
munity’s trade policy by reorganizing member countries’ production
structures so as to achieve diversification in the sub-region and thereby
strengthen trade complementarity through interrelated demand. The
countries of the sub-region could develop, despite strong international
competition, non-natural resource-based comparative advantages. The
creation of a genuine industrial fabric will revitalize intra-community
trade and facilitate the creation of trade to offset episodes of trade
diversion (Viner, 1966). A common external tariff (CET) is the ultimate
tool for demarcating an economic space that is integrated or is being

established.
Ultimately, effective membership of WTO by ECCAS member coun-
tries is vital to regional integration. Compliance with WTO rules serves
as a common framework for all countries. Compliance with com-
mitments by the countries of the sub-region within the multilateral
framework could serve as a credible minimum basis for further regional
cooperation.
Annexes
Table 4.A.1 List of sample countries
ECCAS ECOWAS EU Other

African
countries
Angola Benin Austria Algeria

Burundi Burkina Faso Belgium Brazil
Cameroon Côte-d’Ivoire Cyprus China
CAR Gambia Estonia Egypt
Chad Ghana Finland India
Congo Guinea France Kenya
DRC Guinea Bissau Germany Morocco
Equatorial Guinea Liberia Greece Russia
Gabon Mali Ireland South Africa
São Tomé and Príncipe Mauritania Italy Tunisia
Niger Luxembourg
Nigeria Malta
Senegal Netherlands
Sierra Leone Portugal
Togo Slovakia
Slovenia
Spain
Source: Drawn up by the authors.
Notes
1. ASEAN is composed of Brunei, Cambodia, Indonesia, Laos, Malaysia,
Myanmar, the Philippines, Singapore, Thailand and Vietnam.
2. An example is the Andean Pact or the Cartagena Agreement concluded
between Bolivia, Colombia, Ecuador, Peru and Venezuela.
3. MERCOSUR is made up of Argentina, Brazil, Paraguay and Uruguay.
4. Cameroon, Central African Republic, Chad, Congo, Equatorial Guinea and
Gabon.
5. This is the trade estimate that takes into account the explanatory elements of
external trade, as well as the supply and demand structures of various coun-
tries to compute and predict the theoretical level of trade. These potentials are
usually estimated using gravity models.
6. Herfindahl-Hirschmann developed an index which has been standardized to
obtain values ranging between 0 (zero concentration) and 1 (maximum con-
centration). If the index is close to 0, exports are less concentrated, implying
that the country exports a range of products. If it is close to 1, the focus is on
a few export products.
7. Examples include the monitoring of macroeconomic indicators such as
growth and inflation (Winters, 1993), trade flows and revealed comparative
advantages (Yeats, 1997; Frankel, 1997), etc.
8. The other two forms are actual mileage and adjusted actual mileage. Actual
distance takes two forms, namely actual ground distance and actual ship-
ping distance. The first form is generally measured as the average of the road
distance and railway distance. The second is measured by the actual length
of the sea route between the two ports of both countries. For its part, the
adjusted actual mileage implies that at equal actual mileage, the quality of
infrastructure significantly influences transport cost. The better the quality of
infrastructure, the less the transport cost.
References
Agbodji Akoété, E. (2007), ‘Intégration et échanges commerciaux intra sous-
régionaux: le cas de l’UEMOA’, Revue africaine de l’intégration, vol. 1, no. 1,
pp. 161–188.
Anderson, J. E. and Van Wincoop, E. (2003), ‘Gravity with Gravitas: A Solution
to the Border Puzzle’, American Economic Review, vol. 93, pp. 170–192.
Anderson, J. E. (1979), ‘A Theoretical Foundation for the Gravity Equation’,
American Economic Review, pp. 106–116.
Avom, D. (2005), ‘Les déterminants des échanges dans la CEMAC: une évaluation
empirique’, Economie Appliquée, vol. LVIII, pp. 127–153.
Avom, D. and Mignamissi, D. (2013), ‘Évaluation et analyse du potentiel com-
mercial dans la CEMAC’, L’Actualité Économique, Revue d’Analyse Économique,
vol. 89, no. 2, pp. 115–145.
Ben Hammouda, H. (2005), Pourquoi l’Afrique s’est elle marginalisée dans le com-
merce international?, Document de travail de la CAPC no 20, Addis-Abeba:
Mai 2005.
Bergstrand, J. H. (1985), ‘The Gravity Equation in International Trade: Some
Microeconomic Foundations and Empirical Evidence’, The Review of Economic
and Statistics, pp. 474–481.
Bergstrand, J. H. (1989), ‘The Generalized Gravity Equation, Monopolistic Com-
petition, and Factor Proportions Theory in International Trade’, Review of
Economic and Statistics, vol. 71, pp. 143–153.
Deardorff, A. (1995), ‘Determinants of Bilateral Trade: Does Gravity Work in
a Neoclassical World?’ Discussion Paper, r no. 382, University of Michigan,
presented at a conference by NBER.
De Melo, J., Montenegro, C. and Panagariya, A. (1993), L’Intégration Régionale

Hier et Aujourd’hui, Revue d’Economie du Développement, t February, pp. 7–48.
Evenett, S. J. and Keller, W. (2002), ‘On Theories Explaining the Success of the
Gravity Equation’, Journal of Political Economy, vol. 110, pp. 281–316.
Fontagné, L., Pajot, M. and Pasteels, J. M. (2002), ‘Potentiels de commerce entre
les petites économies hétérogènes: un petit mode d’emploi des modèles de
gravité’, Economie et prévision 2002/1–2, no. 152–153, pp. 115–139.
Foroutan, F. and Pritchett, L. (1993), ‘Intra-Sub-Saharan African Trade: Is It Too
Little?’ Policy Research Working Papers, no. 1225, World Bank, Washington DC,
November 1993.
Helliwell, J. F. and Schembri, L. L. (2005), ‘Frontières, monnaies communes, com-
merce et bien-etre: que pouvons-nous déduire de l’observation des faits?’, revue
de la banque du canada.
Hugon, P. (2002), ‘L’intégration régionale dans les économies en développement
au regard des nouvelles théories de la régionalisation’, Revue Africaine des
sciences économiques et de gestion, no. 2, Juillet–Décembre 2002.
Krugman, P. R. (1991), Geography and Trade, Cambridge, MA: MIT Press.
Lamy, P. (2006), ‘Les Déséquilibres Sociaux et Economiques Doivent Etre
Corrigés’, Marchés Tropicaux et Méditerranées, no. 3136, pp. 7–8.
Luo, X. (2001), ‘La mesure de la distance dans le modèle de gravité: une appli-
cation au commerce des provinces chinoises avec le Japon’, Revue Régions et
Développement, t no. 13.
Rose, A. (2000), ‘One Money, One Market: the Effect of Common Currencies on
Trade’, Economic Policy, CEPR & CES & MSH, vol. 15, no. 30, pp. 7–46.
Tinbergen, J. (1962), Shaping the World Economy, Suggestion for an International
Economic Policy, New York: The Twentieth Century Fund.
UNECA (2008), Economic Report on Africa 2008. Africa and the Monterrey Consen-
sus: Tracking Performance and Progress, Addis Ababa: Economic Commission for
Africa.
UNCTAD (2012), World Investment Report 2012. Towards a New Generation of
Investment Policies, Geneva: United Nations.
Viner, J. (1950), The Customs Union Issue, New York: The Carnegie Endowment for
International Peace.
5
Regional Integration and Trade in
Sub-Saharan Africa, 1993–2010:
An Augmented Gravity Model
Edris Seid
1 Regional integration and trade
Sub-Saharan Africa (SSA) is fragmented politically and economically –

in 2010 there were at least 15 countries with population less than
6 million. Constraints to intra-African trade (which is only 10–12%
of its global exports) include inefficient and costly transport systems –
including informal roadblocks and checkpoints along many trade corri-
dors (69 between Lagos and Abidjan alone; UNECA, 2004) – as well as
complex customs regulations and procedures, with different standards
and required trade documentation including rules of origin. To increase
economies of scale, accelerate industrialization and promote global
exports and growth, regional integration has long been advocated.
The first regional economic community (REC) in Africa was proba-
bly the South African Customs Union (SACU), formed in 1910 among
Botswana, Lesotho, South Africa and Swaziland. In 1917 Kenya and
Uganda formed a Customs Union which Tanzania (then known as
Tanganyika) joined in 1927. Since the 1960s – when most African
countries became independent – many more RECs have been created.
In 1975 the Economic Community of West African States (ECOWAS)
was formed. In 1980 the Southern African Development Coordina-
tion Conference was formed, evolving in 1992 into the Southern
African Development Community (SADC). In 1981 the Preferential
Trade Area for Eastern and Southern Africa was formed, evolving in 1993
into the Common Market for Eastern and Southern Africa (COMESA).
In 1986 the Intergovernmental Authority on Drought and Develop-
ment was formed among states around the Horn of Africa, evolving in
91
92 Regional Integration and Trade in Sub-Saharan Africa
1995 into the Intergovernmental Authority on Development (IGAD).

(Appendix Table 5.A.1 lists the members with data used in the regres-
sions.) ECOWAS, SADC and COMESA have already formed free trade
areas, while IGAD is planning one.
These plus four more RECs – the Community of Sahel-Saharan States
(CEN-SAD), the East African Community (EAC), the Economic Commu-
nity of Central African States (ECCAS) and the Arab Maghreb Union
(AMU) – are recognized by the African Union Commission as pillars
of the African Economic Community, and there are still others. Every
African state is a member of at least one REC, 25 are members of 2, 17
are members of 3 and 6 are members of 4. Such overlapping membership
is in itself a problem hindering further integration.
Despite all these RECs (and the expressed strong commitment of
many African states to increased integration) – when compared, for
example, to intra-regional trade in ASEAN (the Association of Southeast
Asian Nations) – that in Africa remains low because of slow implemen-
tation of agreements to eliminate tariff and non-tariff barriers (UNECA,
2010). Nevertheless, while in 2000 intra-regional trade in Africa was
only 8.5% of Africa’s total exports, growth averaged 15% annually
thereafter, raising its global share to 10.8% by 2010.
Has proliferation of regional economic communities in Africa
expanded intra-regional trade? Why, despite the many RECs, does intra-
regional trade remain low? An augmented gravity model – including
a remoteness index – was estimated to analyze the determinants of
bilateral trade in Africa and the effects on it of ECOWAS, SADC,
COMESA and IGAD.
The next section reviews the composition of Africa’s trade, followed
by reviews of theoretical and empirical literature on regional eco-
nomic integration. Then the gravity model and estimation techniques
employed in the study are introduced, as well as the data. Results and
discussion are followed by summary and conclusions.
2 The composition of Africa’s trade
Africa’s share of global goods exports is quite low (3.3% in 2010),

just over half of Latin America’s 6% (Figure 5.1). Among Sub-Saharan
African regional economic communities, SADC had the most global
goods exports.
Africa’s global exports are mainly primary commodities, with fuel con-
stituting almost 59% in 2010 (Table 5.1). Manufactured exports had
fallen to only 17% and food to less than 9%.
Edris Seid 93
45
40
35
30
25
20
15
10
5
0
00
01
02
03
04
05
06
07
08
09
10
20
20
20
20
20
20
20
20
20
20
20
Africa European Union
Developing Developing
economies: America economies: Asia
Figure 5.1 Shares of goods exports by world region, 2000–2010

Source: UNCTADstat,
t 2012.
Table 5.1 Africa’s goods exports by type, 2001–2010 (%)
2002 2003 2004 2005 2006 2007 2008 2009 2010
Fuels 47.4 49.6 54.2 61.1 63.2 63.0 64.9 57.6 58.6
Manufactured 24.5 23.9 21.9 17.9 16.2 16.2 15.8 18.4 17.0
goods
Food 12.2 11.4 9.4 7.8 7.1 7.1 6.6 10.2 8.8
Agricultural 3.6 3.8 3.2 2.5 2.2 2.1 1.8 2.2 2.1
materials
Others 12.3 11.3 11.3 10.7 11.3 11.6 10.9 11.6 13.5
Source: UNCTADstat,
t 2012.
Given common structural constraints – weak regional infrastructure

links and undiversified exports mainly concentrated in primary com-
modities resulting in weak similarities between exports and imports –
African countries are trade-oriented toward the industrialized West,
including Japan but especially the EU, from which most RECs imported
more than from the rest of Africa (Table 5.2). China and the BRICs have
also become important import sources. However, IGAD’s main import
source during 2000–2010 was the Middle East. ECOWAS, SADC and
Table 5.2 African economic communities’ import sources, 2000–2010 aver-

age (%)
EU USA Japan China India Russia Brazil Middle South Rest of

East Africa Africa
ECOWAS 33.5 6.6 3.7 12.8 3.1 0.6 2.5 1.9 2.0 14.3
SADC 32.9 7.4 4.9 9.8 3.3 0.4 2.3 9.3 5.7 11.8
COMESA 25.9 5.9 2.8 8.3 4.3 1.6 1.5 12.1 5.6 11.8
IGAD 14.9 3.5 3.3 10.5 7.1 0.6 0.7 19.1 2.7 8.3
UMA 56.7 4.6 2.1 6.3 1.1 2.4 1.8 6.5 0.4 3.4
Source: IMF, DOTS, 2011.
COMESA imported more from other African countries than did IGAD
and UMA.
While COMESA’s intra-regional share of its global exports is low –
especially considering its focus on integration through trade – the share
grew during 2001–2010 (as did ECOWAS’s and SADC’s), whereas IGAD’s
intra-regional share fell (Table 5.3). Fuel and food items constituted 46%
of intra-regional exports in 2010, manufactured goods 43%.
3 Regional integration theory
Regional economic communities vary widely – with preferential trade

agreements, free trade areas, customs unions, common markets and
monetary unions – but a common objective is reducing internal trade
barriers (while applying common external standards) to increase intra-
regional trade, enlarging markets and increasing efficiency. Theoretically
they can be welfare enhancing (Kemp and Wan, 1976; Grinols, 1981;
Krishna and Panagariya, 2002), but not necessarily in practice (Feenstra,
2006). Welfare effects are ambiguous, for tariffs are only partially elimi-
nated, and there are thus both trade-creation and trade-diversion effects
(Viner, 1950).
Trade is created when member states import from lower-cost REC-
partners instead of relying on high-cost domestic industries, whereas
trade diversion occurs if – because of the difference between internal and
external tariffs – they now import from higher-cost REC-partners rather
than from lower-cost non-members. If the welfare effects of trade diver-
sion outweigh those of trade creation, overall welfare could be reduced.
Regional integration agreements are more likely to be trade-creating and
welfare-enhancing if member states initially account for large shares of
each other’s imports (Lipsey, 1957, cited in Baldwin and Taglioni, 2006).
Table 5.3 Intra-regional as percent of global exports by regional economic community, in USD millions, 2001–2010
ECOWAS SADC COMESA IGAD
Intra Global % Intra Global % Intra Global % Intra Global %
2001 2,255 27,155 8.3 3,983 44,531 8.9 1,626 27,598 5.9 828 4,635 17.9
2002 3,144 29,031 10.8 4,467 45,992 9.7 1,739 27,198 6.4 810 5,326 15.2
2003 3,298 35,928 9.2 5,663 55,649 10.2 2,004 35,254 5.7 970 6,485 15.0
2004 4,636 46,988 9.9 6,654 68,163 9.8 2,293 43,648 5.3 982 8,192 12.0
2005 5,546 58,872 9.4 7,799 83,556 9.3 2,694 58,602 4.6 1,094 10,385 10.5
2006 5,956 75,580 7.9 8,700 96,049 9.1 2,917 75,465 3.9 1,163 11,980 9.7
2007 6,806 86,504 7.9 12,051 118,670 10.2 4,021 89,557 4.5 1,319 16,391 8.0
2008 9,476 108,909 8.7 16,010 155,156 10.3 6,676 124,576 5.4 1,640 21,194 7.7
2009 7,379 73,569 10.0 12,004 106,626 11.3 6,122 85,759 7.1 1,435 15,337 9.4
2010 9,364 101,475 9.2 14,685 148,065 9.9 8,083 109,336 7.4 1,823 18,705 9.7
Source: Compiled from IMF, DOTS, 2011.

95
Such agreements affect the static allocation of resources as well as the

accumulation of technology and the location of production (Baldwin
and Venables, 1995). Changes in factor prices – including rates of return
on capital – in member as well as non-member states can lead to changes
in physical and human capital investment.
4 Empirical literature
Since Tinbergen (1962) and Poyhonen (1963) applied a gravity model

to analyze trade flows among 42 countries – finding that distance
between exporting and importing countries reduced their trade, while
their GDP raised it – many empirical studies have been carried out on
the determinants of bilateral trade and the effects of regional economic
communities, including dynamic models analyzing the effects of history
on trade. Countries with a history of trading with one another tended to
continue doing so, so omission of historical factors could overstate the
effects of regional economic agreements (Eichengreen and Irwin, 1995).
ASEAN and the Australia-New Zealand trade agreement have both
been found to have boosted trade five-fold or more (Frankel, 1997).
However, in the same study it was found that the high level of
intra-European Community trade in the 1960s and 1970s was almost
completely explained by country size (physical area), level of eco-
nomic development (GDP), proximity or distance, being adjacent or
not and having a common language (that is, the traditional gravity
model variables). Little internal trade could be attributed to the EC
itself until the 1980s. A similar development was found for Mercosur,
which nevertheless has now boosted trade among the member states
seven-fold.
The effects the EU, NAFTA, Mercosur, Australia-New Zealand and the
Israel-USA free trade agreement have been analyzed using both pooled
cross-sectional and fixed-effects specifications of gravity models (Cheng
and Wall, 2005), again with only modest effects found in Europe.
With a gravity model analyzing bilateral trade flows within
COMESA and SADC, the hypothesis (Linder, 1961) that countries with
similar preferences trade more than others was confirmed, while broad
money as a percentage of GDP was found to have increased trade
though, surprisingly, the coefficient for common language was neg-
ative and there was no discernible effect of the regional economic
communities themselves (Alemayehu and Haile, 2008).
Nevertheless, the actual trade of Sub-Saharan Africa – which (as noted)
is very low – has been found to be higher than would be predicted from
a traditional gravity model (Foroutan and Pritchett, 1993).
Edris Seid 97
In another study of Sub-Saharan Africa, distance and contigu-

ity seemed to be the most important factors determining bilateral
trade within regional economic communities, while trade between
communities – for example, between East and West Africa – was quite
small. Again it was found that – given low complementarity of countries’
tradable goods, inadequate infrastructure and trade barriers – actual
trade exceeded what would be predicted (Yeats, 1999).
Using panel data from the EU, EFTA, NAFTA and Mercosur with a
gravity model to analyze both bilateral trade and foreign direct invest-
ment, DeRosa (2008) found that all dummies indicating RECs yielded
the expected positive coefficients.
A gravity model was also used to estimate and compare potential to
actual trade within SADC. Not unexpectedly, transaction costs in the
trading partners, growth paths of member states and changes in per
capita income were at least as important in determining trade as were
trade policies (Cassim, 2001).
Analysis of intra-regional trade in maize, rice and wheat found that
SADC, COMESA and the East African Community had positive effects
(Makochekanwa, 2012).
A gravity model used with panel data to analyze the trade potential
between the EU and Mercosur found that exporting countries’ pop-
ulations had large negative effects – implying domestic absorption –
whereas, importing countries’ populations had large positive effects
(Martinez-Zarzoso and Nowak-Lehmann, 2001).
5 Foundations of the gravity model
The gravity model – which has had consistent empirical success analyz-
ing, for example, migration, commuting, tourism, foreign direct invest-
ment and bilateral trade flows – has nevertheless been criticized for lack
of theoretical foundation. In response, many trade economists have
formulated theoretical justifications (Anderson, 1979; Helpman and
Krugman, 1985; Bergstrand, 1985, 1989, 1990; Helpman, 1987; Deard-
orff, 1998; Eaton and Kortum, 2002; Anderson and van Wincoop, 2003).
Prior to the gravity model, the Heckscher-Ohlin (H-O) model – based
on the Ricardian theory of comparative advantage – was standard for
explaining international trade patterns. Ricardian theory explains pat-
terns as due to differences in technology, to which the Heckscher-Ohlin
model added differences in factor endowments. But it was not immedi-
ately apparent that gravity models took differences in technology and
factor endowments into account, nor, conversely, did H-O models use
country size (physical area) as a determinant (WTO, 2010).
A new theoretical approach used the ‘Armington assumption’ of

both constant-elasticity-of-substitution preferences and goods differ-
entiated by country of origin (Anderson, 1979), which implies that
countries consume some of every good from every country regardless of
prices, and all countries thus participate in international trade. National
income is then the sum of home and foreign demand for the commodi-
ties that each country produces, with larger countries tending to export
and import more.
The gravity model – based on monopolistic competition – followed,
with identical countries trading differentiated commodities because
consumers prefer variety (Bergstrand, 1985, 1989). The gravity model
was also derived under increasing returns to scale of production
(Helpman and Krugman, 1985), and then from Heckscher-Ohlin
assumptions (Deardorff, 1998). A Ricardian model – based on difference
in technology, but also incorporating geographic factors as well as devi-
ations from purchasing power parity – was also developed (Eaton and
Kortum, 2002).
A theoretically grounded estimable gravity model with con-
sumers’ homothetic preferences approximated by constant-elasticity-of-
substitution utility functions followed (Anderson and van Wincoop,
2003), with each firm producing a unique product under increasing
returns to scale, while consumers’ utility increases in both quantity and
variety, so that they enjoy products from many countries.
And the Anderson-van Wincoop model takes multilateral trade resis-
tance into account, with bilateral trade increasing to the extent that
other channels are blocked.
6 Model specification
The traditional gravity model assumes that bilateral imports and exports
can be determined by the exporting and importing countries’ GDPs
(indicative of supply and demand, increasing trade) as well as the
distance between them (a proxy for transport costs, reducing trade), or
β β β
Tij = αGDPi 1 GDP
Pj 2 Dij3 ηij (1)
where Dij is distance; α, β are parameters to be estimated; and ηij is an

error term assumed independent of the regressors.
Population was soon added as an explanatory variable (Linnemann,
1966), and then dummy variables for other potential determinants
such as being adjacent or landlocked, having a common language
Edris Seid 99
or former colonizer and membership in regional economic communi-

ties (Frankel, 1997; Cheng and Wall, 2005; Silva and Tenreyro, 2006).
Landlocked countries incur high transportation costs to access the ocean
via neighboring countries, which deters trade. In Chad and Central
African Republic, for example, it costs more than USD 5000 per con-
tainer to export or import, whereas in Ghana, Morocco, Tunisia or
Egypt (with access to the sea) the cost is far less (IFC and World Bank,
2010).
But even then the traditional gravity model did not fully explain
bilateral trade flows, which are also influenced by resistance to a coun-
try’s trade with all other possible trading partners (Anderson and van
Wincoop, 2003). A remoteness index

Mit =
REM wjt Dij for i = j (2)
j
measuring a country’s average distance from all other partner countries

was thus included as well, with Dij again distance and wjt the ratio of Yjt
and YG , where Yjt is again GDP and YG is global GDP. The estimated effect
of the index is expected to be positive since (for example) countries less
remote from their partners have more import sources so the share from
each will be smaller (Silva and Tenreyro, 2006).
This yields the augmented traditional gravity model
β β β β β β β β β
Tij = αY
Yi 1 Yj 2 Dij3 Ni 4 Nj 5 RECij6 Hij 7 REM Mjt 9 ηij
Mit8 REM (3)
where N denotes population; REC is a dummy indicating common

membership in a regional economic community; and H captures all
other dummy variables which could facilitate or hinder bilateral trade
such as area, landlockness, language, adjacency, common colonizer and
REC and others. As is customary, Equation (3) was log-linearized to
ln Tij = ln α + β1 ln Yi + β2 ln Yj + β3 ln Dij + β4 ln (N
Ni ) + β5 ln (N
Nj )
(4)
+β6 RECij + β7 Hij + β8 ln REM
Mi + β9 ln REM
Mj + ln ηij
and estimated by OLS.

Measuring multilateral resistance – such as changes in trade part-
ners, as well as global events such as inflation and financial crises –
by controlling for importer and exporter time-varying individual effects
yields consistent and unbiased estimators (Anderson and van Wincoop,
2003), though coefficients of GDP, population and other time-varying
country-specific variables then can’t be estimated. The Anderson and

van Wincoop model is
ln Tijt = β0 + β1 dt + β2 dit + β3 djt + β4 ln Dij + β5 tij + β6 Cijt + Uijt (5)
where dt is a year-dummy; dit and djt are time-varying exporter and

importer fixed effects; tij is bilateral trade-facilitating or -hindering dum-
mies (other than distance Dij ) that don’t vary over time (such as being
adjacent or having a common language or former colonizer) and Cijt
is time-varying bilateral dummies, such as exchange rate ratios and per
capita income differences as well as RECs (depreciation of exporter’s cur-
rency vis-à-vis the importer’s increases the exporter’s competitiveness
and is thus expected to increase exports).
However, log-linearization changes the properties of the error term,
raising serious econometric problems (Silva and Tenreyro, 2006). The
error term in Equation (4) is heteroskedastic, which violates the classical
OLS assumption that it be statistically independent of the regressors,
yielding inconsistent estimates.
Cross-section estimates using gravity models may also yield biased,
results since cross-section does not allow heterogeneity. A country might
export different amounts to two countries with identical GDPs and
equidistant from the exporter (Cheng and Wall, 2005).
There is also a problem – when log-linearizing and then estimating by
OLS – of zero trade between countries. It can be handled by dropping
the zero trade observations and truncating the sample (appropriate only
if the zeros are randomly distributed, otherwise important information
is lost); by adding a small constant to all trade values before taking logs;
or (best) by estimating in levels (Silva and Tenreyro, 2006).
Here, Pseudo-Poisson Maximum Likelihood (PPML) was used to esti-
mate both traditional and Anderson-van Wincoop gravity models (with
the dependent variable, export flows, in levels) using
Tij
exp ( − μij )μij
Tij ) =
Pr (T Tij = 0, 1, 2, . . . (6)
Tij !
where μij is a conditional mean exponentially related to the indepen-

dent variables Xij as
μij = exp (α0 + β Xij + ηi + γj ) (7)

Edris Seid 101
Poisson models were originally applied to count data, but can also
be used, as here, with non-negative continuous dependent variables
(Wooldridge, 2002).
7 The data
The dependent variable, consisting of 40,608 observations of bilateral

trade among 48 African countries during 1993–2010, came from the
IMF’s Direction of Trade Statistics, with about 12% of possible observa-
tions missing. GDP and population (and GDP per capita) came from the
World Bank’s World Development Indicators. Distance (between capital
cities as per Mayer and Zignago, 2011), area and dummy variables (adja-
cent, common language, common colonizer, landlocked) came from
CEPII (Research and Expertise on the World Economy). Remoteness was
calculated following Brun et al. (2005). Bilateral exchange rates came
from Darvas (2012) (Table 5.4).
Table 5.4 Variable descriptions
Exports by value, the dependent variable

GDP log of nominal GDP
Distance distance between capital cities
Population log of population
Adjacent dummy with value 1 if the exporter and importer
are adjacent (that is, share a border), otherwise 0
Common language dummy with value 1 if both exporter and
importer share a common language, otherwise 0
Common colonizer dummy with value 1 if both exporter and
importer had a common colonizer
Landlocked dummy with value 1 if country is landlocked,
otherwise 0
Regional economic dummy with value 1 if both exporter and importer
communities belonged to the same REC, otherwise 0
(ECOWAS, SADC,
COMESA, IGAD)
Remoteness log of index calculated following Brun et al.
(2005)
Exchange rate ratio ratio of exporter’s to importer’s real exchange
rate
Per capita GDP log of absolute value of the difference of GDP per
difference capita between exporter and importer
Area log of areas in square kilometers of both exporter
and importer
8 Results from the augmented traditional gravity model
As expected, the traditional gravity model variables were important

determinants of trade amongst the 48 African countries during the study
period (1993–2010). Both importers’ and exporters’ GDP affected trade
positively, as did the population of the exporter, being adjacent and
having had a common colonizer (Table 5.5). Adjacent countries traded
as much as four times more than others. Countries with a common lan-
guage traded about 14% more than others (calculated as (eβi − 1) × 100%,
where βi is the estimated coefficient, 0.134).
Again as expected, distance affected trade negatively, as did the area
of both importer and exporter, and if they were landlocked. The dis-
tance elasticity of trade (−1.3) indicates that a 10% increase in distance
reduced trade by about 13%. A 10% increase in exporter’s area reduced
trade by about 25%.
Table 5.5 PPML estimation of augmented traditional gravity model (N = 30,503,

paired 1,793)
Coefficients Standard errors
Ln (GDP) importer 1.807∗∗∗ 0.025

Ln (GDP) exporter 0.214∗∗∗ 0.024
Ln (distance) −1.300∗∗∗ 0.131
Ln (population) importer −0.062 0.046
Ln (population) exporter 0.986∗∗∗ 0.044
Ln (area) importer −0.253∗∗∗ 0.045
Ln (area) exporter −0.476∗∗∗ 0.041
Adjacent dummy 1.615∗∗∗ 0.252
Common language dummy 0.134 0.139
Common colonizer dummy 1.033∗∗∗ 0.139
Landlocked dummy (importer) −0.577∗∗∗ 0.153
Landlocked dummy (exporter) −1.382∗∗∗ 0.172
Ln (remoteness) importer 0.323∗∗∗ 0.012
Ln (remoteness) exporter 0.409∗∗∗ 0.012
Exchange rate ratio 0.058∗∗∗ 0.006
Ln (per capita GDP difference) −0.001 0.003
ECOWAS 2.003∗∗∗ 0.224
SADC 0.324∗∗∗ 0.030
COMESA −0.143∗∗∗ 0.017
IGAD 0.330 0.582
Constant −42.390∗∗∗ −1.265
Note: ∗∗∗ p < 0. 01.

Edris Seid 103
The remoteness of both importer and exporter affected trade

positively, supporting the claim that country pairs remote from the
rest of the world trade more with each other. The ratio of exporter’s to
importer’s exchange rate affected trade positively, with total trade value
increasing when goods became cheaper to import.
Membership in ECOWAS and SADC affected trade positively, with
ECOWAS members trading about 600% more than others and SADC
members about 38% (IGAD had a similar result, though ‘statisti-
cally non-significant’). Consistent perhaps with previous findings of
COMESA’s weak effect on trade (Cassim, 2001; Alemayehu and Haile,
2008) – possibly reflecting deficient infrastructure within this REC which
stretches from the Mediterranean to southern Africa – the measured
effect of COMESA was unexpectedly negative.
9 Results of the Anderson-van Wincoop gravity model
Again in the Anderson-van Wincoop model – with time-varying year

and country effects to properly control for multilateral trade resistance –
distance affected trade negatively (about 20% for every 10% increase in
distance), while adjacent countries traded about 600% more than oth-
ers; those with a common language, about 49% more; and those that
had had a common colonizer, 66% more (Table 5.6). The high effect of
distance on trade – even more in this model than in the previous one –
may reflect the lack of infrastructure in Africa. For example, only 30%
Table 5.6 PPML estimation of Anderson–van Wincoop gravity model with time-
varying fixed effects for countries and year (N = 31,477)
Coefficients Standard errors
Ln (distance) −2.054∗∗∗ 0.108

∗∗∗
Adjacent dummy 1.940 0.194
Common language dummy 0.400∗∗∗ 0.136
Common colonizer dummy 0.506∗∗∗ 0.151
Exchange rate ratio 0.776∗∗∗ 0.030
Ln (per capita GDP difference) −0.082∗∗∗ 0.005
ECOWAS 0.292 0.241
SADC 0.193∗∗∗ 0.053
COMESA −0.031 0.027
IGAD 1.651∗∗∗ 0.523
Constant −24.360 29.030
Note: ∗∗∗ p < 0. 01.

of the road network is paved (UNECA, 2010), so that it costs three times
as much to ship a car overland from Addis Ababa to Abidjan as it does
to ship it by sea from Japan to Abidjan.
Again, the ratio of exporter’s to importer’s exchange rate affected
trade positively, with total trade value increasing when goods became
cheaper to import. Consistent with Linder’s (1961) hypothesis, per-
capita income differences (as proxy for differences in preferences)
affected trade negatively, indicating that similar countries traded more.
Whereas in the augmented traditional model ECOWAS had a large
positive effect on trade, here the effect is much smaller (and ‘non-
significant’). SADC’s effect is still positive (and significant), though
smaller than before. COMESA again has a negative effect (though
smaller and ‘non-significant’). The real surprise is IGAD, which is now
found to have had a quite large, positive (and ‘statistically significant’)
effect. As Anderson and Van Wincoop (2003) and Silva and Tenreyro
(2006) showed this specification handles the problem of omitted vari-
able bias, country pair fixed effects and control for multilateral trade
resistance. Hence the positive coefficient on IGAD from the Anderson
and Van Wincoop model might give sense because some country pairs
in IGAD such as Kenya and Uganda are member of a more advanced
REC, that is, East African Community (EAC), and have extensive trade
relations.
Despite the many regional economic communities (RECs) in Africa,

intra-regional trade remains very low compared to trading blocs in
Europe, Asia and Latin America. To uncover the main factors behind this
low level of intra-regional trade – and the role of four RECs (ECOWAS,
SADC, COMESA and IGAD) in promoting it – both an augmented tra-
ditional gravity model and an Anderson-van Wincoop gravity model
were applied to panel data of trade among 48 African countries during
1993–2010.
The traditional gravity model variables (GDP, distance, being adja-
cent, having a common language, having had a common colonizer
and being landlocked or not) all had the expected effects on intra-
regional trade. Exchange rate ratios had a positive effect, indicating that
total export revenues increased with lower export prices. At least in the
Anderson-van Wincoop model, differences in preferences (proxied by
differences in GDP per capita) were also found to have affected trade,
with more similar countries trading more, confirming the Linder (1961)
Hypothesis.
Edris Seid 105
Results regarding the effects of ECOWAS, SADC, COMESA and IGAD

in promoting trade were mixed. In the augmented traditional model,
membership in ECOWAS and SADC were found to have affected trade
positively, with ECOWAS members trading about 600% more than oth-
ers and SADC members about 38%. IGAD had a similar result (though
‘statistically non-significant’).
In the Anderson-van Wincoop model, ECOWAS’s effect was found to
have been much smaller (and ‘non-significant’), though SADC’s was still
positive, but smaller than before. IGAD was now found to have had a
quite large positive (and ‘statistically significant’) effect.
In both models, COMESA was found to have had a negative effect,
perhaps reflecting its extreme geographical distribution and weak infras-
tructure links.
Though the results are mixed, regional integration can clearly pro-
mote trade. To optimize that possibility, African countries should invest
more on infrastructure to link neighboring countries, harmonize trade
policies and simplify customs procedures.
Annexes
Table 5.A.1 Countries included in the gravity model by regional economic

community
ECOWAS (1975) SADC (1992) COMESA (1993) IGAD (1995) Other
Benin Angola Burundi Djibouti Algeria

Burkina Faso Congo, D.R. Comoros Ethiopia Cameroon
Cape Verde Madagascar Congo, D.R. Kenya Central
African Rep.
Côte d’Ivoire Malawi Djibouti Somalia Chad
Gambia, The Mauritius Egypt Sudan∗ Congo, Rep.
Ghana Mozambique Ethiopia Uganda Equatorial
Guinea
Guinea Seychelles Kenya Zimbabwe Gabon
Guinea-Bissau South Africa Libya Tunisia
Liberia Tanzania Madagascar Morocco
Mali Zambia Malawi Mauritania
Niger Zimbabwe Mauritius São Tomé &
Príncipe
Nigeria Rwanda
Senegal Seychelles
Sierra Leone Sudan∗
Togo Uganda
Zambia
Zimbabwe
Note: ∗ Sudan includes South Sudan during this period.

References
Alemayehu, G. and Haile, K. (2008), ‘Regional Economic Integration in Africa:
A Review of Problems and Prospects with a Case Study of COMESA’, Journal of
African Economies, vol. 17, no. 3, pp. 357–394.
Anderson, J. E. (1979), ‘A Theoretical Foundation for the Gravity Equation’,
American Economic Review, vol. 69, no. 1, pp. 106–116.
Anderson, J. E. and van Wincoop, E. (2003), ‘Gravity with Gravitas: A Solution to
the Border Puzzle’, American Economic Review, vol. 93, no. 1, pp. 170–192.
Baldwin, R. E. and Venables, A. J. (1995), ‘Regional Economic Integration’, in
G. M. Grossman and K. Rogoff (eds.), Handbook of International Economics,
vol. 3, Amsterdam: North-Holland, pp. 1597–1644.
Balwin, R. and Taglioni, D. (2006), ‘Gravity for Dummies and Dummies for
Gravity Equation’, NBER Working Paper Series, no. 12516.
Bergstrand, J. H. (1985), ‘The Gravity Equation in International Trade: Some
Microeconomic Foundations and Empirical Evidence’, The Review of Economics
and Statistics, vol. 67, no. 3, pp. 474–481.
Bergstrand, J. H. (1989), ‘The Generalized Gravity Equation, Monopolistic Com-
petition, and the Factor-Proportions Theory in International Trade’, The Review
of Economics and Statistics, vol. 71, no. 1, pp. 143–153.
Bergstrand, J. H. (1990), ‘The Heckscher-Ohlin-Samuelson Model, the Linder
Hypothesis, and the Determinants of Bilateral Intra-Industry Trade’, Economic
Journal, vol. 100, pp. 1216–29.
Brun, J. F., Carrere, C., Guillaumont, P. and Melo, J. (2005), ‘Has Distance Died?
Evidence from a Panel Gravity Model’, World Bank Economic Review, vol. 19,
no. 1, pp. 99–120.
Cassim, R. (2001), ‘The Determinants of Intra-Regional Trade in Southern Africa
with Specific Reference to South Africa and the Rest of the Region’, Development
Policy Research Unit Working Papers, no. 01/51.
Cheng, I.-H. and Wall, H. J. (2005), ‘Controlling for Heterogeneity in Gravity
Models of Trade and Integration’, Federal Reserve Bank of St. Louis Review, vol. 87,
no. 1, pp. 49–63.
Darvas, Z. (2012), ‘Real Effective Exchange Rates for 178 Countries: A New
Database’, Bruegel Working Paper, r no. 2012/06.
Deardorff, A. (1998), ‘Determinants of Bilateral Trade: Does Gravity Work in a
Neoclassical World?’, in Jeffrey A. Frankel (ed.), The Regionalization of the World
Economy, vol. 1, Chicago: University of Chicago Press, pp. 7–32.
DeRosa, D. A. (2008), Prospects for Greater Global and Regional Integration in the
Maghreb, Washington, D.C.: Peterson Institute of International Economics.
Eaton, J. and Kortum, S. (2002), ‘Technology, Geography and Trade’,
Econometrica, vol. 70, no. 5, pp. 1741–1779.
Eichengreen, B. and D. A. Irwin (1995), ‘Trade Blocs, Currency Blocs and the
Reorientation of World Trade in the 1930s’, Journal of International Economics,
vol. 38, pp. 1–24.
Fenstra, R. (2006), ‘New Evidence on the Gains from Trade’, Review of World
Foroutan, F. and Pritchett, L. (1993), ‘Intra-Sub-Saharan African Trade: Is it too
little?’, Journal of African Economies, vol. 2, no. 1, pp. 74–105.
Edris Seid 107
Grinols, Earl L. (1981), ‘An Extension of the Kemp-Wan Theorem on the For-
mation of Customs Unions’, Journal of International Economics, vol. 11, no. 2,
pp. 259–266.
Helpman, E. and Krugman, P. R. (1985), Market Structure and Foreign Trade: Increas-
ing Returns, Imperfect Competition, and the International Economy, Cambridge:
MIT Press.
Helpman, E. (1987), ‘Imperfect Competition and International Trade: Evidence
from Fourteen Industrial Countries’, Journal of the Japanese and International
Economies, vol. 1, pp. 62–81.
IFC and World Bank (2010), Doing Business 2010: Reforming Through Difficult
Times, Washington, D.C.: IFC, Palgrave MacMillan and World Bank.
IMF (2011), Direction of Trade Statistics (DOTS) Database, CD-ROM.
Kemp, M. and Jr. HenryWan (1976), ‘An Elementary Proposition Concerning the
Formation of Customs Unions’, in Marry Kemp (eds.), Three Topics in the Theory
of International Trade: Distribution, Welfare and Uncertainty, Amsterdam: North
Holland, pp. 95–97.
Krishna, P. and Panagariya, A. (2002), ‘On Necessarily Welfare-Enhancing Free
Trade Areas’, Journal of International Economics, vol. 57, no. 2, pp. 353–367.
Linder, S. B. (1961), An Essay on Trade and Transformation, New York: Wiley
and Sons.
Linnemann, H. (1966), An Econometric Study of International Trade Flows,
Amsterdam: North-Holland.
Makochekanwa, A. (2012), ‘Impacts of Regional Trade Agreements on Trade in
Agrifood Products: Evidence from Eastern and Southern Africa’ Paper presented
at the African Economic Conference 2012, Kigali, Rwanda.
Martinez-Zarzoso, I. and Nowak-Lehmann, F. (2001), ‘Augmented Gravity Model:
An Empirical Application to Mercosur-European Union Trade Flows’, Journal of
Applied Economics, vol. 4, no. 2, pp. 291–316.
Mayer and Zignago (2011), ‘Notes on CEPII’s Distances Measures (GeoDist)’,
CEPII Working Paper, r 2011–2025, available from: http://www.cepii.fr/
anglaisgraph/bdd/distances.htm#sthash.qJGNVtLw.dpuf.
Poyhonen, P. (1963), ‘A Tentative Model for the Volume of Trade Between
Countries’, Weltwirtschaftliches Archiv, no. 90, pp. 93–100.
Silva, J. M. C. and Tenreyro, S. (2006), ‘The Log of Gravity’, The Review of
Economics and Statistics, vol. 88, no. 4, pp. 641–658.
Tinbergen, J. (1962), Shaping the World Economy, New York: The Twentieth
Century Fund.
UNCTADstat (2012), United Nations Conference on Trade and Development, t avail-
able from: http://unctadstat.unctad.org/ReportFolders/reportFolders.aspx?sCS_
referer=&ScsChosenLang=en.
UNECA (2004), ‘Assessing Regional Integration in Africa’, ECA Policy Research
Report,
t Addis Ababa.
UNECA (2010), ‘Assessing Regional Integration in Africa IV: Enhancing Intra-
African Trade’, ECA Policy Research Report,t Addis Ababa.
Viner, J. (1950), The Customs Union Issue, New York: The Carnegie Endowment for
International Peace.
Wooldridge, J. M. (2002), Econometric Analysis of Cross Section and Panel Data,
Cambridge, MA: MIT Press.
World Trade Organization (2010), A Practical Guide to Trade Policy Analysis,

Geneva: World Trade Organization.
Yeats, A. J. (1999), ‘What can be Expected from the African Regional Trade
Arrangements? Some Empirical Evidence’, World Bank Policy Research Working
Papers.
Part III
Industrialization Strategy and the
Issue of Deindustrialization
6
Malawi’s Trade Policies, Market
Structure and Manufacturing
Performance, 1967–2002
Hopestone Kayiska Chavula
1 Extending Kaluwa and Reid (1991) to include structural

breaks
As was the case with most other less-developed countries in Africa,

Asia and Latin America, Malawi went through periods of economic
crisis during its first 15 years after independence – when the guid-
ing policies were protectionism, import-substitution industrialization
and export-led agricultural development. Nevertheless, the Malawian
economy performed better during those years than later under struc-
tural adjustment programs (SAPs) or still later under trade liberalization
(Njolwa, 1982; World Bank, 1989; Mulaga and Weiss, 1996; Chirwa,
2003). Despite substantial consistent early growth, Malawi’s manufac-
turing sector has remained small and underdeveloped, with dismal
recent performance.
Following the general class of oligopolistic models, there is a grow-
ing body of empirical literature on structure-performance modeling
of the effects of trade policies – especially trade liberalization – in
a variety of countries (Smirlock, 1985; Lloyd-Williams, 1994), but no
study has looked at these issues in Malawi since Kaluwa and Reid
(1991), which only covered 1969–1972. Extending Kaluwa and Reid,
the effects of Malawi’s trade policies on price–cost margins were there-
fore estimated econometrically, using enterprise-level manufacturing
data during 1967–2002. (Consistent data after 2002 was unavailable.)
Market structure was included in the estimations as a control. Despite
prior knowledge of when policies were announced or implemented,
111
112 Malawi’s Trade Policies
the Clemente-Montañés-Reyes unit-root test was conducted to iden-

tify structural breaks, that is, when policies actually had effects, since
adjustment by firms takes time.
The next section reviews Malawi’s trade policies and manufactur-
ing performance since independence. Section 3 describes the empirical
methods and data, while Section 4 presents the results. Section 5
summarizes and draws conclusions.
2 Policies and performance since independence
Average annual manufacturing growth in Malawi (Table 6.1) was

much higher during import substitution industrialization (12.65%)
than during structural adjustment programs (2.87%) or trade liberaliza-
tion/export promotion (1.01%).
Import substitution industrialization

During import substitution industrialization (1964–1980), emphasis
was on profitable private competitive firms which – diversifying from
agriculture – were to be the engine of growth and sustainable employ-
ment in manufacturing (Malawi Government, 1971). Manufacturing
was protected both by natural barriers and by trade policy (Mulaga and
Weiss, 1996). Government Development Plans (1961–1964 and 1965–
1969), the Industrial Development Act (1966), the Control of Goods Act
(1968) and the first Statement of Development Policies (1971–1980) all
made this clear.
Though the Malawian economy became more restrictive in the late
1970s due to the government’s needing revenue to finance the budget,
Table 6.1 Malawi’s trade policies and manufacturing growth, 1964–2004
Trade policy Average annual

manufacturing
growth (%)
1964–1980 Import substitution industrialization 12.65

1981–1993 Structural adjustment programs 2.87
1994–2004 Trade liberalization/export promotion −1.01
Notes: Due to limitations on data availability, growth during import substitution industri-
alization was calculated over the period 1973–1981 and during trade liberalization/export
promotion over 1995–2002. Chirwa (2003) and Chirwa and Zakeyo (2003) report 6.7%
growth of manufacturing value-added during import substitution industrialization.
Source: Author’s calculations based on data from the Reserve Bank of Malawi Financial
and Economic Review (various issues).
Hopestone Kayiska Chavula 113
this period was generally characterized by good macroeconomic man-

agement and good financial discipline imposed on public sector enter-
prises, leading to economic growth (as noted, 12.65% in manufacturing)
and favorable balances of payments.
Structural adjustment programs

During 1981–1993 Malawi was subject to World Bank and IMF struc-
tural adjustment programs due to worsening balances of payments as a
result of the oil crises and acute deterioration in terms of trade plus and
increased international transport costs due to the civil war in nearby
Mozambique. These developments led to fiscal deficits which further
stimulated imports and worsened balances of payments (Njolwa, 1982;
World Bank, 1989; Mulaga and Weiss, 1996; Ahsan et al., 1999). The
adjustment programs were intended to stabilize fiscal and trade posi-
tions by increasing efficiency and diversifying the economy, ensuring
appropriate prices and incomes policies, improving the policy envi-
ronment for manufacturing and trade and restructuring government
expenditures (Chirwa and Zakeyo, 2003).
The first phase of structural adjustments (1981–1986) emphasized
restructuring major state-owned and private enterprises, increasing
interest rates and agricultural prices, liberalizing other prices and liber-
alizing entry into manufacturing (Mulaga and Weiss, 1996; Ahsan et al.
1999). However, macroeconomic instability increased toward the end of
this period due to intensification of the Mozambican civil war which
led to an influx of refugees as well as further increases in international
transport costs. The current account deficit rose from 7% of GDP in 1985
to 13% in 1986 (Chirwa, 2003). Average annual manufacturing growth
for 1981–1986 fell to 2.6% (author’s calculations).
The second phase of structural adjustments (1987–1994) emphasized
more trade liberalization – with continuous reduction in import tariffs
(Milner and Zgovu, 2003) – plus export promotion (Malawi Govern-
ment, 1988). However, poor sequencing and hesitant implementation
resulted in even higher macroeconomic instability. There was a con-
certed effort at domestic liberalization as well, especially in the financial
sector and agriculture. Interest rates were liberalized by 1988 and entry
into the financial sector in 1989.
In 1987, through the Agriculture (General Produce) Act, the market-
ing of smallholder produce was liberalized, allowing the participation of
private traders in both domestic and export markets. This was followed
by liberalization of prices for all agricultural products except maize, cot-
ton and tobacco. In 1990 agricultural inputs previously marketed by the
Agricultural Development and Marketing Corporation (ADMARC) were

deregulated, with the phased removal of fertilizer subsidies completed in
1991. Again with agriculture, there was poor sequencing, with markets
liberalized before deregulation of prices (Chirwa, 2003).
Entry into manufacturing was also liberalized in 1991 after phased
liberalization of prices (Chirwa and Zakeyo, 2003). Since most manu-
facturing is agriculture-based – and with increased access to financial
services – these changes were expected to have a large impact on growth.
However, average annual manufacturing growth grew to only 2.9%
during this second phase of structural adjustment programs.
Trade liberalization/export promotion

During 1994–2004 Malawi focused on trade liberalization – under multi-
lateral, regional and bilateral trade agreements – plus export promotion
and removal of constraints facing manufacturing especially. Industrial
licensing requirements were reduced to a short list of products, while
exclusive product rights were eliminated and the duty draw-back system
was revised. Manufacturing in special export processing zones was intro-
duced, while regional integration and trade openness were strengthened
with the African Growth and Opportunity Act and regional blocs includ-
ing especially the Southern African Development Community (SADC)
free trade area (Chirwa and Zakeyo, 2003).
Liberalization of agriculture continued with removal in 1995 of
restrictions on smallholders producing and marketing high-value crops
such as burley tobacco, followed in 1996 by liberalization of prices for
cotton and tobacco.
The base surtax was also reduced from 25% to 20%, and privatiza-
tion of state-owned enterprises continued, despite suspension of the
privatization program in 2001 due to lack of tangible benefits.
Despite liberalization of the financial sector, banks may have contin-
ued exercising monopoly power in setting less favorable interest rates
for both borrowers and depositors (Chirwa and Mlachila, 2004).
Due to frequent devaluations during the period, speculative attacks
on Malawi’s currency were predictable (and imminent), while it had
become difficult to do business with the rest of the world due to
low levels of foreign exchange reserves, damaging investor and donor
confidence (Reserve Bank of Malawi, 2006). To improve export com-
petitiveness and provide a more efficient foreign exchange allocation
mechanism, Malawi adopted a managed float (exchange rate regime) in
February 1994, which led to further devaluation of the currency.
40 0.7
0.6
30
0.5
Price−cost margin
20
Growth rate %
0.4
10
0.3
0
0.2
–10
0.1
–20 0
1970 1974 1978 1982 1986 1990 1994 1998
Manufacturing growth rate
Price−cost margin
Figure 6.1 Price–cost margins and manufacturing output growth rate trends –
1967–2002
Source: Author’s calculations based on data from the National Statistical Office and the
Reserve Bank of Malawi Financial and Economic Review (various issues).
As noted (Table 6.1), despite all these policy initiatives, manufacturing

actually declined during the period at an average annual rate of 1.01%.
Manufacturing price–cost margins had been higher during import
substitution industrialization, fell during the 1980s structural adjust-
ment programs and rose again in the 1990s even though manufacturing
output declined (Figure 6.1).
3 Methods and data
To analyze the effects of these changes in trade policy on manufacturing

performance in Malawi (as measured by the price–cost margin) – while
controlling for market structure – an oligopolistic model was estimated,
following Kaluwa and Reid (1991), using the relationship

y = f MC, φ (X
( ) , δ (B) , e, p (1)
where y is firm-level price–cost margin (Domowitz et al., 1986); MC

is measures of market concentration (Herfindahl-Hirschman Index,
import intensity and export intensity); X is availability of inputs (raw

materials, skilled labor, finance capital) determining how firms inter-
act; B is barriers to entry (capital/labor ratio and minimum efficient
scale) also influencing incumbent firms’ behavior; φ and δ are conjec-
tural elasticities with respect to incumbent firms and potential entrants,
respectively; e is firm elasticity of demand; and p is government trade
policies.
With variables defined as in Appendix Table 6.A.1, setting (from Eq. 1
above) MC = f (hhi, imp, exp ); φ(X( ) = f (rms, skill, fink); δ(B
( ) = f (k/l, mes);
e = f (dem); tariff as a policy variable; DU Umt as dummies representing
trade policies since independence, identified as structural breaks during
analysis; and adding εit as an error-term, yields the equation estimated,
yit = α0 + α1 hhiit + α2 impkt + α3 expkt + α4 rmtrsit + α5 skillit + α6 finkit

+ α7 k/lit + α8 meskt + α9 demandit + α10 tariffft + α11 DU
Umt + εit (2)
where the subscript i indicates the observation for the ith firm (i =
1, 2, . . . , N, where N is the number of firms in year t, 1967–2002), while
k identifies industry-level variables. DU Umt = 1 for t > Tbm and 0 otherwise,
for m = 1, 2 which is the number of structural breaks during the period.
Tb1 and Tb2 are break points to be located by grid search – expected to
coincide (with possible lag) the policy changes discussed in Section 2 –
to be located by grid-search following Baum (2001). This will let the data
reveal when policy changes actually affected manufacturing price–cost
margins.
Firm-level panel data from 141 firms in 14 four-digit ISIC industries
during 1967–2002 was obtained from responses to annual question-
naires returned to Malawi’s National Statistical Office as part of NSO’s
industrial production survey of medium and large for-profit firms. Gross
manufacturing output was deflated by the GDP deflator while capital
stock was estimated by the conventional perpetual inventory method,
taking into account the book value of all fixed assets (land and improve-
ments, buildings and other constructions, machinery, transport and
other equipment, including gross additions less depreciation). Capital
was deflated by the price of imports since most manufacturing capital is
imported.
Import and export intensities were measured at 3-digit industry level –
using data from the NSO’s (National Statistical Office, 2006) Annual
Statements of External Trade – because firm-level data was scanty. The
data was categorized and linked to the three-digit ISIC classification used
by NSO based on Malawi Foreign Trade Classification for 1984, which
establishes the relationships among the International Standard Indus-

trial Classification (ISIC), the Standard International Trade Classification
(SITC) and Harmonized System (HS) codes or the Customs Cooperation
Council Nomenclature (CCCN).
4 Empirical results and discussion
Following Baum (2001) – to identify actual structural breaks in

Malawian manufacturing’s price–cost margins (as opposed to the dates
when trade policy changes were merely announced or even imple-
mented) – the Clemente-Montañés-Reyes unit-root test was first con-
ducted (Appendix Figure 6.A.1). Statistically significant structural breaks
were identified in 1982 and 1991, as is also plausible on visual inspection
of the figure. The break in 1982 could relate to the structural adjust-
ment program initiated in 1981 after years of economic instability as
discussed earlier. The break in 1991 could relate to liberalization of trade,
finance, agriculture and manufacturing toward the end of the structural
adjustment programs.
Choice of models
Model 1 (Table 6.2) is based on all 1583 observations from 1967–2002,
using unbalanced panel data, Model 2 on those from 1967–1982 (that is,
until the first structural break), Model 3 on those from 1983–1991 (that
is, between the structural breaks) and Model 4 on those from 1992–2002
(after the second break). Hausman tests were used to choose fixed-effects
(Models 1, 2 and 3) or random-effects (Model 4). The Wooldridge test
revealed first-order autocorrelation (AR1) in Models 1, 3 and 4 while
the Wald test also revealed panel heteroskedasticity in Models 1, 2 and
3. Thus Feasible Generalized Least Squares (FGLS) – which produces con-
sistent and efficient estimates in the presence of AR(1) within panels
and cross-sectional correlation and heteroskedasticity across panels if
the time period (T) is greater than the cross-sectional units (N) (Beck
and Katy, 1995) – was used for estimating Models 1 and 3. Groupwise
heteroskedasticity without first-order autocorrelation in Models 2 and 4
prompted the use of statistical software STATA’s ‘Robust’ command to
obtain robust estimates.
Effects of trade policies and market structure on manufacturing’s

price–cost margins
Supporting Bain’s (1951) hypothesis, market concentration (hhi) was
found to be positively associated with Malawian manufacturing’s
118
Table 6.2 Effects of trade policies and market structure on manufacturing price–
cost margins, 1967–2002
Variables Model 1 Model 2 Model 3 Model 4

1967–2002 1967–1982 1983–1991 1992–2002
FGLS FE (Robust) FGLS RE (Robust)
hhi (market 0.128∗∗∗ 0.175∗∗∗ 0.091∗∗∗ 0.157∗∗∗

concentration) (0.000) (0.000) (0.000) (0.000)
imp (import −0.044∗∗∗ −0.330∗∗∗ 0.014 −0.086∗∗
intensity) (0.000) (0.000) (0.425) (0.021)
exp (export intensity) −0.006 −0.064∗∗∗ 0.025 −0.055∗∗
(0.528) (0.006) (0.174) (0.019)
rms (raw materials) −0.225∗∗∗ −0.488∗∗∗ −0.186∗∗∗ −0.241∗∗∗
(0.000) (0.000) (0.000) (0.000)
skill (skilled labor) −0.031∗ 0.038 0.019 −0.081∗∗
(0.065) (0.565) (0.644) (0.015)
fink (finance capital) −0.073∗∗∗ −0.096∗ −0.165∗∗∗ −0.062
(0.001) (0.106) (0.001) (0.226)
k/l (capital/labor 0.129∗∗∗ 0.081∗ 0.255∗∗∗ 0.084∗∗
ratio) (0.000) (0.100) (0.000) (0.029)
mes (minimum −0.013 0.090 0.090 0.274∗∗
efficient scale) (0.748) (0.525) (0.362) (0.017)
Demand −0.008 −0.008 −0.013 0.0001
(0.566) (0.750) (0.650) (0.997)
Tariff 0.008 0.058 −0.074 −0.198
(0.835) (0.314) (0.401) (0.144)
dummy83-91 −0.005
(0.927)
dummy92-02 0.154∗∗
(0.031)
Intercept 1.727∗∗∗ 2.958∗∗∗ 0.924 2.234∗∗
(0.000) (0.000) (0.183) (0.047)
Hausman test 47.09∗∗∗ 27.19∗∗∗ 50.19∗∗∗ 15.34
(0.0000) (0.0000) (0.0000) (0.1202)
Wooldridge test 6.729∗∗ 0.652 3.423∗ 17.230∗∗∗
(AR1) (0.0109) (0.4223) (0.0693) (0.0001)
Wald test 1.3e + 05∗∗∗ 1.6e + 05∗∗∗ 1.5e + 05∗∗∗ −
(0.000) (0.000) (0.0000)
Breusch & Pagan − − − 109.46∗∗∗
LM test (0.0000)
Observations 1583 678 442 463
Groups 132 88 88 94
Notes: The values in parentheses are t-values; ∗ denotes significant at 10% level; ∗∗ 5%; and
∗∗∗ 1%.
price–cost margins in all four models, that is, in the entire study period
as well as in each sub-period (Table 6.2). This contrasts with Kaluwa
and Reid (1991) – where the concentration variable was found to be
statistically non-significant and with the wrong sign – but their very
short study period (1969–1972) was also characterized by restrictions on
entry to manufacturing as well as by price controls on certain products –
where increases had to be approved by the Ministry of Trade and Indus-
try, which could only be justified by exogenous cost increases (Kaluwa,
1986) – hence their non-significant effect.
Consistent with Kaluwa and Reid (1991), import intensity (imp) was
found to have had negative (and highly significant) effects – indicating
that increased imports led to reduced price–cost margins – in Models 1, 2
and 4. These results are consistent with the imports-as-market-discipline
hypothesis, whereby it is expected that price–cost margins fall under
the pressure of imports (even – or especially – during import substitu-
tion industrialization), especially for firms that previously had market
power (Lopez and Lopez, 2003; Culha and Yalcin, 2005). The effect
was much less after 1982, which could imply improvement in manu-
facturing price–cost margins due to reduced costs of production after
liberalization.
Consistent with House (1973), Lopez and Lopez (2003) and Culha and
Yalcin (2005), as well as Kaluwa and Reid (1991), export intensity (exp)
was found to have had negative effects in Models 2 and 4, indicating
that – while possibly forced to reduce prices in order to export – more
export-oriented firms can also better exploit economies of scale and thus
operate on smaller price–cost margins.
Perhaps because they are mostly imported, leading to higher pro-
duction costs (Kaluwa, 1986), raw materials (rms) were found to have
had a negative effect on price–cost margins in all four models, with
the largest effect during import substitution industrialization (Model 2).
Skilled labor (skill) was also found to have had a negative effect over-
all (Model 1) and especially after liberalization (Model 4), perhaps then
reflecting improved labor mobility so that the effect of skilled wages
showed up more clearly. In contrast to Kaluwa and Reid (1991), finance
capital (fink) was also found to have had a negative effect in Mod-
els 1, 2 and (the largest effect, in Model) 3, perhaps reflecting the then
newly somewhat-liberalized financial sector increasing the availability
of finance, leading to a reduction in price–cost margins.
The capital/labor ratio (k/l) – a possible barrier to entry – was found
to have had a positive effect in all models, highest in Model 3 (dur-
ing structural adjustment programs). Perhaps liberalization, which led
to devaluation of the currency, raised the cost of imported capital

goods, leading to an increase in price–cost margins. Capital-intensive
firms embody the most advanced technology, the flow of which was
facilitated by liberalization. Due to the scarcity of skilled labor, labor
absorption might have been slower than the increase of capital, yield-
ing a higher capital/labor ratio and an increase in price–cost margins.
As an (unreported) alternative, the capital/sales ratio (used by Kaluwa
and Reid, 1991) was also tried, but with results contrary to theirs and to
Chirwa (2001).
Minimum efficient scale (mes) – another possible barrier to entry – was
positive (and statistically significant) in Model 4 (during full trade liber-
alization and export promotion), possibly indicating that economies of
scale led to higher price–cost margins.
Growth in demand did not seem to have any effect on price–cost
margins during the study period, nor did average tariff rates.
In Model 1 (covering the entire study period), the dummy for 1992–
2002 (the period of trade liberalization and export promotion) had a
positive effect on price–cost margins, whereas the dummy for 1983–
1991 had no statistically significant effect. One might have thought that
the statistically significant dummy for 1992–2002 had absorbed all the
effect of policy changes during that period, whereas the non-significant
dummy for 1983–1991 indicated that there were no such differences in
effects during that period and that the presence of both dummies would
have eliminated any differences between effects during the entire study
period and the first period (1967–1982), but one would be wrong.
Comparing variable coefficients between Model 1 (with both dum-
mies for later periods) and Model 2 (for the first period, 1967–1982),
values for market concentration, import intensity, raw materials and
finance capital were lower in Model 1, whereas capital/labor ratio was
higher. The value for skilled labor was statistically significant in Model 1
but not in Model 2.
Similarly, comparing variable coefficients between Model 1 (with the
non-significant dummy for 1983–1991) and Model 3 (covering the
same period by construction), values for market concentration and
raw materials were substantially higher in Model 1, whereas values for
finance capital and capital/labor ratio were substantially lower. Values
for import intensity and skilled labor, which were statistically significant
in Model 1, were not in Model 3.
Similarly, comparing variable coefficients between Model 1 (with sta-
tistically significant dummy for 1992–2002) and Model 4 (covering the
same period by construction), values for market concentration, import
intensity, raw materials and skilled labor were lower in Model 1, whereas
the value for capital/labor ratio was higher. Values for export intensity
and minimum efficient scale, which were not statistically significant in
Model 1, were so in Model 4, whereas the value for finance capital,
which was statistically significant in Model 1, was not in Model 4.
Comparing variable coefficients among Models 2, 3 and 4 – covering
each of the policy periods under study – the fact that values for
market concentration, import and export intensities and raw materi-
als were highest in Model 2 indicates that those variables had most
effect on price–cost margins during import substitution industrialization
(1967–1982).
The effects of trade policies and market structure on manufacturing

performance (price–cost margins) in Malawi during 1967–2002 were
assessed using firm-level panel data. Structural breaks were identified in
1982, attributed to economic crises in the late 1970s and early 1980s,
and in 1991, attributed to an accumulation of effects of structural
adjustment programs in the late 1980s and early 1990s.
As one might expect, greater market concentration (as measured by
the Herfindahl-Hirschman index and, to a lesser extent, by lower import
and export intensities) was found to positively affect price–cost mar-
gins across all periods, unequivocally demonstrating its relevance when
analyzing manufacturing in Malawi.
Greater import intensity led to lower price–cost margins due to pres-
sure from better quality and, sometimes, cheaper foreign goods, due
to liberalization, with firms responding by reducing prices to remain
competitive. Greater post-liberalization export intensity also led to
lower price–cost margins due to greater exposure to more competitive
international markets.
Increasing exports of high quality products might increase price–cost
margins. Emphasis should be on importing emerging technologies –
which could promote innovation – as well as on development of sci-
ence, technology and innovation skills in Malawi’s education system,
since the effectiveness of imported technologies will depend on capacity
to assimilate them.
Availability of inputs was also determinative of price–cost margins, as
also found earlier by Kaluwa and Reid (1991). Deliberate government
efforts to enhance the availability of raw materials, skilled labor and
finance capital might improve manufacturing performance.
As one might expect, barriers to entry (capital/labor ratio more so

than minimum efficient scale, except during the post-liberalization
period) were also found to positively affect price–cost margins. This
reinforces the notion that regional integration – allowing larger mar-
kets and greater economies of scale – might contribute to profitability
while at the same time increased competition might keep inefficien-
cies down.
Annexes
Table 6.A.1 Variable definitions
Description Definitions
Y Price–cost margin Ratio of the difference

between value-added and
payroll to total sales
Hhi Herfindahl-Hirschman Index Measure of concentration in
domestic production
measured at the four-digit
international standard
industrial classification
(ISIC) level
Imp Import intensity Ratio of value of imports to
total industry sales
Exp Export intensity Ratio of value of exports to
total industry sales
Rms Raw materials Total cost of raw materials
deflated by the GDP deflator
skill Skill level Average earnings
fink Finance capital Product of the estimated
minimum efficient scale
and the ratio of capital (net
book value of fixed assets)
to industry output
K/L Capital-labor ratio Capital divided by labor,
the number of employees
including production
workers
mes Minimum efficient scale Average employment of
firms accounting for 50% of
total industry employment,
as a percentage of that total
industry employment
demand Demand Percentage change in sales
tariff Average tariff rate Total import duties divided
by the volume of imports
Clemente-Montañés-Reyes double AO test for unit root

Test on lpcm: breaks at 1982,1991
–.5
–1
lpcm
–1.5
–2
–2.5
1970 1980 1990 2000
Year
D.lpcm
1.5
1
D.lpcm
.5
0
–.5
–1
1970 1980 1990 2000
Year
Optimal breakpoints 1982 1991

AR(1) du1 du2 (rho –1) Constant
Coefficients: –0.27724 0.40382 –1.07667 –0.85374
t-statistics: –2.198 2.968 –5.512
P-values: 0.036 0.006 –5.490 (5% critical value)
Figure 6.A.1 Clemente-Montañés-Reyes unit-root test with double price–cost

margin’s mean shifts, AO model
Note: ln(pcm) and D.lpcm is the log and differenced price–cost margin respectively.
References
Ahsan, A., Machinjili, C. and Mkwezalamba, M. M. (1999), The Impact of Adjust-
ment Policy Reforms on Malawi’s Manufacturing Sector, r Lilongwe: World Bank,
Malawi Field Office.
Bain, J. S. (1951), ‘Relation of Profit Rate to Industry Concentration: American
Manufacturing 1936–1940’, Quarterly Journal of Economics, vol. 65, no. 3,
pp. 293–324.
Baum Christopher, F. (2001), ‘Stata: The Language of Choice for Time Series
Analysis’, The Stata Journal, vol. 1, no. 1, pp. 1–16.
Beck, N. and Katy, J. N. (1995), ‘What To Do (and Not To Do) with Time Series
Cross-Section Data’, American Political Science Review, vol. 89, pp. 634–647.
Chirwa, E. W. (2001), ‘Privatization and Technical Efficiency: Evidence from
the Manufacturing Sector in Malawi’, Working Paper, r Economics Department,
University of Malawi.
Chirwa, E. W. (2003), ‘Trade Policy and Industrialisation in Malawi: The Need for
a Strategic Approach’, Working Paper, University of Malawi, Zomba.
Chirwa, E. W. and Mlachila, M. (2004), ‘Financial Reforms and Interest Rate
Spreads in the Commercial Banking System in Malawi’, IMF Staff Papers, vol. 51,
no. 1, pp. 96–122.
Chirwa, E. W. and Zakeyo, C. (2003), ‘Impact of Economic and Trade Policy

Reforms on Food Security in Malawi’, Report submitted to the Food and Agricul-
tural Organisation (FAO), Lilongwe Malawi and the African Economic Research
Consortium AERC, Nairobi Kenya.
Culha, A. and Yalcin, C. (2005), ‘The Determinants of Price-Cost Margins of
the Manufacturing Firms in Turkey’, Working Paper no. 05/15. Research and
Monetary Policy Department, The Central Bank of the Republic of Turkey.
Domowitz, I., R. G. Hubbard and B. C. Petersen (1986), ‘Business Cycles and
Oligopoly Supergames: Some Empirical Evidence on Prices and Margins’,
Working Paper no. 2057, National Bureau of Economic Research, October.
House, W. J. (1973), ‘Market Structure and Industry Performance: The Case of
Kenya’, Oxford Economic Papers, New Series, vol. 25, no. 3, pp. 405–419.
Kaluwa, B. M. (1986), Barriers to Entry, Price Controls and Monopoly Power in
Malawian Manufacturing, g Dissertation, Department of Economics, University of
Edinburgh, Department of Economics, PhD Thesis.
Kaluwa, B. M. and Reid, G. C. (1991), ‘Profitability and Price Flexibility in Man-
ufacturing for a Developing Country’, Journal of Industrial Economics vol. 39,
no. 6, pp. 689–700.
Lloyd-Williams, D. M., Molyneux, P. and Thornton, J. (1994), ‘Market Structure
and Performance in Spanish Banking’, Journal of Banking and Finance, vol. 18,
pp. 433–443.
Lopez, R. A. and Lopez, E. (2003), ‘The Impact of Imports on Price-cost Margins:
An Empirical Illustration’, Empirical Economics, vol. 28, no. 2, pp. 403–416.
Malawi Government (1971), Statement of Development Policies, 1971–1979,
Zomba: Government Printer.
Malawi Government (1988), Statement of Development Policies, 1987–1996,
Zomba: Government Printer.
Milner, C. and Zgovu, E. (2003), ‘Export Response to Trade Liberalisation the
Presence of High Trade Costs: Evidence for a Landlocked African Economy’,
Centre for Research in Economic Development and International Trade, University
of Nottingham, UK.
Mulaga, G. and Weiss, J. (1996), ‘Trade Policy and Manufacturing Sector Perfor-
mance in Malawi, 1970–1991’, World Development, t vol. 24, pp. 1267–1278.
National Statistical Office (2006), Quarterly Statistical Bulletin (various issues),
Zomba, Malawi: National Statistical Office (NSO). Njolwa, M. M. Z. (1982),
‘Industrial Development in Malawi: Problems and Prospects for the 1980s’,
Paper presented at the International Social Science conference at Chancellor
College: Zomba, Malawi.
Reserve Bank of Malawi (2006), ‘Monetary Policy, Exchange Rate Management
and Devaluation in Malawi – Past and Present’, an address by Governor Vic-
tor Mbewe, Reserve Bank of Malawi, at the Economics Association of Malawi
(ECAMA) Conference, Lilongwe.
Smirlock, M. (1985), ‘Evidence on the (non) Relationship between Concentra-
tion and Profitability in Banking’, Journal of Money, Credit and Banking,g vol. 17,
pp. 69–83.
World Bank (1989), Malawi Industrial Sector Memorandum, Washington, DC: The
World Bank.
7
The Impact of South-South and
North-South Trade on
Industrialization in Africa
Henri Atangana Ondoa and Henri Ngoa Tabi
Introduction
It is generally accepted that a country is industrialized when the rela-

tive contribution of industry to gross domestic product (GDP) increases
or when the proportion of people employed in the secondary sector
grows. Labor productivity improves in the secondary sector, the unem-
ployment rate falls from one year to another, per capita income increases
(Alderson, 1999) and the secondary sector’s contribution to economic
development is confirmed. No country can effectively promote its eco-
nomic and social development without strong and buoyant industry,
because it processes raw materials and creates opportunities to develop
services.
Despite their efforts, most African countries have faced obstacles
to industrial development. Africa is the least industrialized continent
because of an energy deficit, technological gap, bad governance and
narrow domestic markets. Structural adjustment programs, while advo-
cating trade liberalism, caused de-industrialization in some African
countries. Industry in some African countries deteriorated in the past
three decades. A study by Hossein and Weiss (1999), for instance,
showed that seven of 16 African countries in his sample1 experienced de-
industrialization over the 1975–1993 period. Meanwhile, from 1980 to
2009, the share of manufacturing value added to GDP increased slightly
in North Africa, from 12.6 to 13.6%, it decreased in the rest of Africa,
from 16.6 to 12.7%. More than half a century after independence, while
other regions increased their share of manufactured exports, the conti-
nent still depends on the exportation of raw materials to industrialized
125
126 South-South and North-South Trade
countries. Export of raw materials and mining and agricultural com-

modities are often the main sources of foreign exchange; so African raw
materials are processed and resold at relatively high prices to African
consumers (ECA, 2013).
Dependency on production and export of primary commodities
makes the continent vulnerable to external demand fluctuations, which
lead to pro-cyclical fiscal spending in many countries reliant on export
revenues.
At the same time, Africa is undergoing significant structural changes.
Several African countries are signing a growing number of commercial
contracts with Asian developing countries. Others are involved in nego-
tiations that could eventually lead to the creation of a common market
with the European Union. Finally, as the AfDB (2011) has observed,
exchanges within Africa are developing gradually and should inten-
sify in the next years. These changes will shape the future of African
industry.
This study aims to assess the impact of North-South trade and South-
South trade on industrialization in Africa. This is a major challenge for
the African industry. Trade openness may offer an important opportu-
nity for African economies, but it may also represent a risk if African
industries cannot compete with foreign products. It is appropriate
to identify the factors that make African industry cost-effective and
competitive.
The chapter is structured as follows: Section 1 presents the litera-
ture review; Section 2 defines the scope of the study; Section 3 outlines
the methodology; Section 4 presents the findings; and Section 5 offers
conclusions.
1 Literature review
The perceived impact of international trade on manufacturing in devel-

oping countries is varied. According to some (Wood, 1994; Rowthorn
and Coutts, 2004), North-South trade is beneficial for manufacturing
industries in the South. De-industrialization observed in developed
countries could be partly due to North-South trade. Since poor countries
have a comparative advantage in labor (Rowthorn and Coutts, 2004;
Sung and Hongshik, 2011), the importation of labor-intensive goods
manufactured in developing countries such as China, Mexico, India
and Brazil caused de-industrialization in developed countries. Manufac-
tured goods imported from developing countries resulted in job losses of
1.5 to 5% in the manufacturing sectors in developed countries over 40
years. Meanwhile, trade between developing countries and developed
Henri Atangana Ondoa and Henri Ngoa Tabi 127
countries created just 0.3 to 0.4% of employment in OECD countries.

So the net effect of North-South trade on employment in the manu-
facturing sector is beneficial for developing countries and negative for
developed countries (Rowthorn and Coutts, 2004). However, in coun-
tries such as France, Germany and Sweden, this theory was relativized
during the 1980s by Lawrence (1987), who found that external factors
account for only 25% of de-industrialization observed in France and
Germany.
The positive impacts of trade openness on industrialization are dis-
cussed in other studies. For instance, Barro and Sala -I- Martin (1995)
link innovation cost to an economy’s degree of openness; open coun-
tries can benefit from new production techniques discovered elsewhere
as international trade improves overall factor productivity through three
main effects (Rivera-Batiz and Romer 1991; Edwards, 1998):
(i) the allocative effect, trade openness improves the allocation of

inputs and the ability of states to better manage the impact of
learning-by-doing;
(ii) the amalgamation effect, openness ensures knowledge dissemina-
tion, and
(iii) the redundancy effect because trade openness reduces the costs
associated with the duplication of research activities.
Other studies (Coe and Helpman, 1995; Keller, 1999) have established
a positive correlation between goods importation and technology dis-
semination. However, Higino Schneider (2005) has demonstrated that
although the importation of high-technology goods can definitely
contribute to the industrial development in both developed and devel-
oping countries through innovation, this effect is observed mostly in
developed countries.
According to Frankel and Romer (1999), international trade boosts
industrial production through several channels:
(i) specialization, according to the principle of comparative advan-

tage;
(ii) economies of scale, through increased market size;
(iii) the exchange of ideas, through communication and travel; and
(iv) technology dissemination, through trade in machinery.
However, total factor productivity falls with the importation of goods

and improves with exports as shown in Pack’s study (1988). This
asymmetry is due to simultaneity bias as countries tend to export
the goods they produce at lower cost (comparative advantage) and

import the goods they produce at a high cost. It can also be
attributable the pro-cyclical nature of productivity. Generally, pro-
ductivity is high in a period of economic boom and decreases in
a period of economic downturn. Consequently, if imports increase
when domestic production decreases, productivity will decline as well
(Harrison, 1996).
According to Elbadawi and Helleiner (1998), industries on the African
continent are less efficient and cannot compete with foreign firms. This
underperformance of African industries is partly due to international
trade. Trade openness has exposed local industries in less developed
countries (LDCs) to competition for which they were unprepared. Large
segments of the manufacturing sector have disappeared over the last
20 years in Africa. This de-industrialization process has been more pro-
nounced in countries with a low level of development. To date, LDCs
have, on average, less diversified economies and more concentrated
exports. Instead of reducing their structural vulnerabilities, trade open-
ness has increased them.2 In short, this liberalization was premature in
LDCs given their level of development (UNCTAD, 2010).
Moreover, dependency theorists support protectionist policies.
According to them, peripheral economies export raw materials and
import manufactured goods. So trade openness can only lead to de-
industrialization in the South (Cardoso and Faletto, 1979), and free trade
has caused the closure of infant industries in some developing countries.
The only form of openness that can boost production in the manufac-
turing sector in developing countries is regional integration. According
to Brady et al. (2011), regional integration should boost industrial pro-
duction in developing countries because integrated economies have a
large market size and integration contributes to physical capital for-
mation within and between states. On the other hand, the effect of
integration on employment in the manufacturing sub-sector can be neg-
ative, as it exposes young hitherto protected enterprises to competition
from more competitive regional firms.
2 Scope of study
This study covers 47 African countries during the 1971–2010 period for
averages of five years. In other words, the study period is T = 40/5 = 8,
chosen because the main study variables such as trade openness and
customs duties vary little from year to year. The IMF database and
UN Comtrade base are used. Table 7.1 shows an industrialization rate
Table 7.1 Descriptive statistics
Variable Obs Average Std.dev. Minimum Maximum

differential
Industrialization rate (manufacturing) 371 20.67248 15.47541 0.507193 92.52468

Industrialization rate 371 25.54046 15.40954 0.644434 95.70562
Industrialization growth rate 358 −0.0820142 2.871481 −7.214132 8.591289
Industrialization growth rate (manufacturing) 358 −0.079582 3.122605 −8.632133 8.67233
Duration of war for 5-year period 376 0.6409 1.5540954 0
Customs duty level 376 18.22 9.41541 3.1507193 64.9768
Increase in degree of trade openness level/developed countries 358 −0.0132 1.094714 −4.439236 4.941442
Increase in degree of trade openness/Asian DvgCs 358 0.0759973 1.404259 −5.447616 4.507696
Degree of trade openness/African countries 358 0.0244569 1.536485 −7.355315 8.056836
Degree of trade openness/developed countries 369 49.92914 174.8159 0 2090.805
Degree of trade openness/Asian DvgCs 368 6.304589 16.12535 0 230.4244
Degree of trade openness/African countries 366 7.323723 8.050944 0 71.28955
Investment growth rate 358 −0.0072 0.6730703 −1.969168 1.830424
Investment rate as percentage of GDP 371 20.33631 10.3269 3.45412 113.3061
Real GDP per capita 371 1360.759 2288.257 82.33274 16273.92
Real GDP per capita growth rate 358 0.0386171 1.470073 −3.91794 4.211768
Labor force growth rate 342 −0.0005244 2.318324 −4.996977 5.24244
Increase in school enrolment rate 344 0.5346173 0.468409 −0.3665129 1.388126
Raw materials dependency ratio 376 20.42287 46.70091 0 433
Raw materials dependency ratio growth rate 348 0.009237 1.3729 −4.5108 3.8735
Industrialization rate 371 25.54046 15.40954 0.644434 95.70562
Source: Authors’ estimates based on IMF data (2012).

129
estimated at 25.54% of GDP. This means that African economies derive

a quarter of their GDPs from the industries.
In this period, trade between African countries and developing coun-
tries in Asia increased significantly, since the increase in the degree of
trade openness between these two groups of countries is estimated at
7.59%. At the same time, trade between African countries only grew
by about 2.44% while trade with developed countries fell by around
−1.13%. Despite increased trade between Africa and Asian develop-
ing countries, developed countries remain the leading trading partners
of African countries. For this period, trade openness between African
countries and developed countries is estimated at 49.92% compared to
only 7.32% and 6.3% respectively for the African and Asian developing
countries (Table 7.1).
3 Methodology
The authors Hossein and Weiss (1999) studied the industrialization pro-
cess in absolute terms and in relative terms. In absolute terms, analysis
of the industrialization process is based on the value added of the
secondary sector; while in relative terms, analysis focuses on the indus-
trialization level, that is, share of the absolute value of the secondary
sector in GDP. In this study, industrialization is attributable to internal
factors such as GDP, urbanization, natural resources and external factors,
in particular trade openness. Although interesting, this approach seems
incomplete as it overlooks the impact of factors such as physical cap-
ital formation, North-South trade and especially South-South trade on
industrialization.3 The approach adopted here aims to complement that
of Rowthorn and Coutts (2004) to estimate the impact of South-South
trade and North-South trade on industrialization in Africa. To this end,
the model is depicted in equation (1):
INDi,t = θi + φXi, t + ϕY
Yi,t + i,t (7.1)
Here, IND is the industrialization rate (relation between the value added
of the secondary sector on GDP), i stands for the country, t is time, X is a
vector which specifies the internal factors and Y is a vector that specifies
external factors, ϕ and φ are parameters to be estimated and ψ an error
term. θi stands for the country fixed effect.
The internal variables used are real GDP per capita, GDP2 to
test the relationship U between the GDP and industrialization or
de-industrialization postulated by Clark (1957). The country’s urbaniza-

tion (URB) level, is approximated by the proportion of the population
living in urban areas. Primary school enrollment rate and physical cap-
ital formation (FBCF) are also used as a percentage of GDP at constant
prices as other explanatory variables that measure the impact of moder-
nity on industrialization. Indeed, the production of manufactured goods
is usually capital-intensive and any change in the investment rate affects
demand for manufactured products (Rowthorn and Coutts, 2004). Brady
et al. (2011) note that the industrialization process requires a mini-
mum level of modernity that can be approximated by factors such as
urbanization, education and physical capital formation. In addition,
the total population variable will measure the local market’s impact on
the industrialization rate. The average level of customs duties is used as
an indicator of protectionism. Other control variables such as natural
resource dependence are approximated by the ratio of the value of com-
modity exports to GDP and the average duration of war in a five-year
period as an indicator of political instability (Cieslik and Tarsalewska,
2011) are also used.
Regarding the influence of business partners, trade openness is divided
into three groups: trade openness to African countries ‘opaf’; trade
openness to Asian developing countries ‘opas’; and trade openness to
developed countries ‘opded’. All these variables are expressed as a per-
centage of GDP. In the first set of estimates, trade with countries in
transition and PVD of Latin America is ignored to avoid a potential
problem of multi-collinearity. In a second set of estimates, the impact of
trade is analyzed between African countries and countries in transition
on industrialization in Africa. Results of this second series of estimates
are shown in Tables 7.A.1 and 7.A.2 (see appendix).
The estimation of equation (1) over the period 1971–2010 may pose
several econometric problems.4 Indeed, the variables trade openness,
capital and raw materials dependence index and other variables are
endogenous since the causality between economic growth and these
variables is effective in both directions. To overcome this difficulty, the
generalized Method of Moments (GMM) can be used for panel data.
This method was developed by Holtz-Eakin, Newey and Rosen (1988)
and Arellano and Bond (1991). Arellano and Bond (1991), for example,
propose moving from the baseline equation (1) to a first-order differ-
ence equation to eliminate the country fixed effects. However, such
differentiation poses another problem; the error term is by construction
correlated with the lagged endogenous variable. In addition, the instru-
ments are less relevant if the autoregressive process goes beyond order 1.
To this end, Blundell and Bond (1998), Arellano and Bover (1995) devel-
oped a method to use lagged but differentiated variables as instruments.
The advantage is that this estimator identifies the endogenous variables
of the model and thus the instruments, and sets limits to the lag of
the differentiated endogenous variables used as instruments. The latter
method is used in this study.
Equation (1) is estimated in first-order differences and all quantita-
tive variables in the study are in logarithmic form. This transformation
helps to reduce the magnitude of fluctuations and stabilize variances.
Variables are stationary in first-order differences, which confirms that
there is no correlation between the fixed effects and the explanatory
variables in first-order differences.5
4 Results
The estimation results obtained are shown in Tables 7.2, 7.3, 7.A.1 and
7.A.2. In Tables 7.2 and 7.3, the dependent variable is the industrial-
ization rate. In Tables 7.A.1 and 7.A.2, the dependent variable is the
growth rate of the industries’ added value. In both cases (industrial-
ization rate and value added), the results are overall significant and
pose no econometric problems. The Sargan tests performed show that
endogeneity problems have been corrected. These results are presented
in two parts. First, the impact of internal factors on industrializa-
tion in Africa are analyzed. Then, the impact of trade openness on
industrialization in Africa is studied.
Impact of internal factors on industrialization in Africa

Overall, the industrialization rate increases with real GDP per capita
and remains untouched by the evolution of the square of real GDP
per capita. Thus, real GDP per capita is a factor of industrialization
in Africa. The negative de-industrialization stated by Alderson (1999)
and by Hossein and Weiss (1999) and the negative de-industrialization
between 1975 and 1993 in countries such as Burkina Faso, Burundi,
Ghana, Rwanda, Sierra Leone and Tanzania would be caused by the fact
that the GDP per capita of these countries was very low or inferior to the
one corresponding to industrialization thresholds. However, the relation
between the real GDP per capita and the industrialization rate cannot
be depicted in a U-shaped curve. Indeed, the above mentioned relation
does not admit a minimum as stipulated by Clark (1957).
The estimation results also show that countries which industri-
alize rapidly are highly urbanized and populated. The effects of
urbanization particularly support manufacturing industries in Africa.
Table 7.2 Contribution of external and internal factors to industrialization
Variables Africa SSA DEP>127 DEP<12 Africa SSA DEP>12 DEP<12
Lagged endogenous 0.0372 0.0471 −0.163∗∗ −0.00116 0.0457 0.0599 −0.161∗∗ 0.0239
(0.0464) (0.0482) (0.0680) (0.0605) (0.0470) (0.0488) (0.0688) (0.0615)
Customs duties(CD) 0.256∗∗∗ 0.177∗∗∗ 0.395∗∗∗ 0.123∗ 0.253∗∗∗ 0.174∗∗∗ 0.396∗∗∗ 0.0929
(0.0582) (0.0664) (0.100) (0.0731) (0.0583) (0.0664) (0.101) (0.0730)
duration of war in 5-year −0.0168 −0.00633 −0.00577 0.0162 −0.0192 −0.00584 −0.00486 0.00555
(0.0232) (0.0234) (0.0337) (0.0295) (0.0234) (0.0235) (0.0341) (0.0295)
Urbanization rate 0.158∗∗∗ 0.155∗∗∗ 0.265∗∗∗ 0.112 0.167∗∗∗ 0.161∗∗∗ 0.269∗∗∗ 0.142∗∗
(0.0531) (0.0545) (0.0875) (0.0696) (0.0535) (0.0548) (0.0888) (0.0697)
Increase in degree of trade 0.225∗∗∗ 0.226∗∗∗ 0.182∗∗∗ 0.351∗∗∗ 0.227∗∗∗ 0.221∗∗∗ 0.189∗∗∗ 0.371∗∗∗
openness/DvgC (0.0397) (0.0396) (0.0551) (0.0694) (0.0402) (0.0402) (0.0567) (0.0695)
Increase in degree of trade −0.0428∗ −0.0346 −0.0372 −0.0516∗ −0.0428∗ −0.0413∗ −0.0290 −0.0505∗
openness/DvgC Asia (0.0228) (0.0235) (0.0334) (0.0281) (0.0237) (0.0245) (0.0364) (0.0288)
Increase in degree of trade −0.0470 −0.0200 −0.189∗∗∗ 0.0438 −0.0622∗ −0.0437 −0.176∗∗∗ 0.00614
openness/DvgC Africa (0.0315) (0.0323) (0.0557) (0.0399) (0.0343) (0.0352) (0.0636) (0.0448)
GDP growth rate (GR) 0.0435∗∗∗ 0.0592∗∗∗ 0.0654∗∗ 0.00950 0.0422∗∗ 0.0590∗∗∗ 0.0685∗∗ 0.00470
(0.0167) (0.0174) (0.0282) (0.0198) (0.0167) (0.0175) (0.0291) (0.0197)
GDP* GR growth rate −0.00251 −0.00305 0.000791 −0.00606∗ −0.00276 −0.00324 0.000661 −0.00630∗
(0.00267) (0.00288) (0.00509) (0.00363) (0.00268) (0.00290) (0.00515) (0.00359)
Investment growth rate 0.0248 0.0238 0.00238 −0.00121 0.0241 0.0238 0.00190 0.000562
(0.0161) (0.0171) (0.0239) (0.0185) (0.0161) (0.0170) (0.0242) (0.0183)
Labor force growth rate 0.0591∗∗∗ 0.0636∗∗∗ 0.0193 0.0828∗∗∗ 0.0610∗∗∗ 0.0673∗∗∗ 0.0160 0.0936∗∗∗
(0.0134) (0.0136) (0.0226) (0.0158) (0.0137) (0.0138) (0.0235) (0.0165)
133
134
Increase in school 0.0912∗∗ 0.0799∗∗ 0.219∗∗∗ 0.0926∗ 0.0822∗∗ 0.0646∗ 0.228∗∗∗ 0.107∗∗
enrolment rate (0.0363) (0.0359) (0.0546) (0.0517) (0.0373) (0.0371) (0.0573) (0.0521)
(Increase in degree of trade −0.00548 0.000813 −0.00594 −0.0124
openness/DvgCAsia)∗ (CD) (0.00730) (0.00772) (0.0106) (0.00913)
(Increase in degree of trade 0.00956 0.0117 −0.00466 0.0224∗∗
openness/DvgCAfr)∗ (CD) (0.00701) (0.00716) (0.0125) (0.00925)
Years (1976–1980) 0.389 0.775 2.137∗∗ 0.522 0.325 0.706 2.056∗∗ 0.359
(0.475) (0.539) (0.912) (0.531) (0.479) (0.540) (0.930) (0.537)
Years (1981–1985) 1.356∗∗∗ 1.530∗∗∗ 1.907∗∗ 1.495∗∗ 1.320∗∗ 1.554∗∗∗ 1.874∗∗ 1.505∗∗
(0.516) (0.542) (0.875) (0.636) (0.518) (0.545) (0.887) (0.633)
Years (1986–1990) 1.379∗∗∗ 1.385∗∗∗ −0.0131 1.294∗∗ 1.407∗∗∗ 1.454∗∗∗ −0.138 1.417∗∗
(0.499) (0.516) (1.037) (0.613) (0.501) (0.518) (1.065) (0.606)
Years (1991–1995) 0.604 0.635 0.847 0.498 0.581 0.663∗ 0.785 0.325
(0.395) (0.396) (0.552) (0.637) (0.397) (0.399) (0.566) (0.633)
Years (1996–2000) 0.667 0.897 1.271∗ 1.140∗∗ 0.609 0.868 1.199∗ 0.973∗
(0.603) (0.761) (0.692) (0.516) (0.604) (0.760) (0.710) (0.513)
Years (2001–2005) 1.541∗∗ 1.806∗∗ 1.849∗∗∗ 1.416∗∗ 1.557∗∗ 1.828∗∗ 1.792∗∗ 1.493∗∗∗
(0.644) (0.797) (0.688) (0.584) (0.644) (0.797) (0.700) (0.577)
Years (2006–2010) 0.772 0.743 1.041∗∗ 0.559 0.733 0.750 0.976∗∗ 0.398
(0.527) (0.536) (0.467) (0.493) (0.528) (0.538) (0.480) (0.491)
Observations 318 286 152 153 318 286 152 153
Number of 05 year periods 8 8 8 8 8 8 8 8
Sargan Test 336(0.00) 316(0.00) 280(0.010) 251(0.011) 334(0.00) 314(0.00) 136(0.013) 149(0.012)
Note: Dependent variable: industrialization rate (Added value of industries on GDP); DEP dependency ratio to raw materials, SSA: Sub-Saharan Africa.
∗∗∗ p < 0. 01, ∗∗ p < 0. 05, ∗ p < 0.
Source: Authors’ estimate.

Table 7.3 Contribution of external and internal factors to development of manufacturing industries
Lagged endogenous −0.39∗∗∗ −0.385∗∗∗ −0.0131 −0.294∗∗ −0.407∗∗∗ −0.454∗∗∗ −0.138 −0.417∗∗
(0.049) (0.016) (1.037) (0.0613) (0.02501) (0.018) (1.065) (0.0612)
Customs duties (CD) 0.349∗∗∗ 0.324∗∗∗ 0.549∗∗∗ 0.187∗∗ 0.343∗∗∗ 0.319∗∗∗ 0.552∗∗∗ 0.176∗∗
(0.0662) (0.0707) (0.117) (0.0859) (0.0667) (0.0712) (0.119) (0.0874)
Average duration of war in 0.0189 0.0373 −0.0146 0.0615∗∗ 0.0148 0.0346 −0.0170 0.0572∗
5-year periods (0.0262) (0.0268) (0.0409) (0.0306) (0.0265) (0.0273) (0.0415) (0.0312)
Urbanization rate 0.245∗∗∗ 0.270∗∗∗ 0.520∗∗∗ 0.183∗∗ 0.253∗∗∗ 0.279∗∗∗ 0.526∗∗∗ 0.195∗∗∗
(0.0588) (0.0604) (0.104) (0.0711) (0.0595) (0.0611) (0.106) (0.0727)
Increase in degree of trade 0.180∗∗∗ 0.180∗∗∗ 0.158∗∗ 0.229∗∗∗ 0.188∗∗∗ 0.184∗∗∗ 0.192∗∗∗ 0.236∗∗∗
openness DC (0.0476) (0.0492) (0.0705) (0.0734) (0.0487) (0.0507) (0.0740) (0.0756)
Increase in degree of trade −0.000396 0.0140 −0.0380 0.0285 0.00484 0.0161 −0.0118 0.0306
openness/DvgC Asia (0.0268) (0.0278) (0.0395) (0.0308) (0.0277) (0.0290) (0.0426) (0.0317)
Increase in degree of trade −0.093∗∗∗ −0.0655∗ −0.218∗∗∗ −0.0347 −0.0963∗∗ −0.0781∗∗ −0.202∗∗∗ −0.0482
openness/DvgC Africa (0.0333) (0.0355) (0.0627) (0.0404) (0.0376) (0.0395) (0.0724) (0.0480)
GDP growth rate (GR) 0.0622∗∗∗ 0.0714∗∗∗ 0.0516 0.0160 0.0600∗∗∗ 0.0681∗∗∗ 0.0574∗ 0.0132
(0.0191) (0.0205) (0.0320) (0.0210) (0.0193) (0.0208) (0.0333) (0.0213)
GDP* GR growth rate −0.00159 −0.00304 −0.000120 −0.011∗∗∗ −0.00191 −0.00345 −0.000210 −0.011∗∗∗
(0.00312) (0.00339) (0.00620) (0.00379) (0.00315) (0.00343) (0.00632) (0.00382)
Investment growth rate 0.0219 0.0298 0.0167 0.000338 0.0207 0.0288 0.0182 0.00130
(0.0189) (0.0203) (0.0280) (0.0201) (0.0190) (0.0204) (0.0285) (0.0203)
Labor force growth rate 0.0765∗∗∗ 0.0801∗∗∗ 0.0377 0.0890∗∗∗ 0.0762∗∗∗ 0.0808∗∗∗ 0.0308 0.0921∗∗∗
(0.0150) (0.0161) (0.0274) (0.0176) (0.0153) (0.0164) (0.0285) (0.0183)
135
136
Increase in school 0.0760∗ −0.0643 0.161∗∗ 0.175∗∗∗ 0.0746∗ −0.0559 0.176∗∗ 0.182∗∗∗
enrolment rate (0.0392) (0.0402) (0.0666) (0.0524) (0.0408) (0.0422) (0.0704) (0.0551)
(Increase in degree of trade −0.00916 −0.00679 −0.0229∗ −0.00529
openness/DvgC Asia)∗ (CD) (0.00869) (0.00951) (0.0130) (0.00996)
(Increase in degree of trade 0.00480 0.00875 −0.00645 0.00815
openness/DvgCAfr)∗ (CD) (0.00832) (0.00873) (0.0152) (0.0100)
Years (1976–1980) −0.735 −0.616 −0.0879 0.313 −0.753 −0.658 −0.111 0.322
(0.606) (0.712) (0.855) (0.921) (0.608) (0.717) (0.872) (0.927)
Years (1981–1985) 0.737 0.579 −2.649 1.114 0.710 0.658 −3.132∗ 1.118
(0.913) (1.065) (1.764) (0.924) (0.931) (1.081) (1.829) (0.965)
Years (1986–1990) 0.359 −0.0194 0.0659 1.214 0.268 −0.170 −0.131 1.129
(0.715) (0.935) (1.003) (1.256) (0.722) (0.948) (1.029) (1.270)
Years (1991–1995) 0.601 1.248 −0.782 2.105∗∗ 0.527 1.356 −1.095 2.197∗∗
(0.902) (1.334) (1.051) (0.901) (0.918) (1.361) (1.082) (0.929)
Years (1996–2000) 1.083 0.742 1.443 0.583 1.099 0.740 1.393 0.539
(1.157) (1.205) (1.282) (0.530) (1.161) (1.212) (1.310) (0.543)
Years (2001–2005) −0.595 −0.669 −0.824 0.0539 −0.620 −0.638 −1.077 0.0461
(0.777) (0.831) (0.654) (0.735) (0.783) (0.841) (0.680) (0.740)
Years (2006–2010) −0.558 −0.834 −0.610 0.390 −0.627 −0.947 −0.818 0.323
(0.629) (0.747) (0.519) (0.729) (0.633) (0.757) (0.541) (0.742)
Observations 318 286 152 153 318 286 152 153
Number of 05 year periods 8 8 8 8 8 8 8 8
Sargan Test 340(0.00) 305(0.00) 154(0.015) 178(0.012) 337(0.00) 302(0.00) 145(0.037) 176(0.022)
Note: Dependent variable: industrialization rate (added value to manufacturing industries on GDP); DEP dependency ratio to raw materials, SSA: Sub-
Saharan Africa. ∗∗∗ p < 0. 01, ∗∗ p < 0. 05, ∗ p < 0.
Such industries sell their products easily in big cities due to concentra-
tion of demand. This result had already been established by Hausmann
and Rodrik (2006) and Brady et al. (2011), when they demonstrated
the importance of factors such as domestic demand and urbanization
in industrialization.
Customs duties and schooling also contribute to industrial develop-
ment. However, the contribution of schooling has been more clearly
demonstrated in natural resource-rich countries. These results contra-
dict some previously conducted work. According to Wood and Mayer
(1998), human resources and natural resources constitute the key deter-
mining factors of the comparative advantage of manufacturing exports.
In this regard, Africa is unable to increase the volume of its manufactur-
ing exports because it is more endowed with natural and not human
resources. Similarly, AfDB (2011) notes that competitiveness in non-
mining sub-sectors is affected by the appreciation of the real exchange
rate against that of their competitors due to raw material exports, in
accordance with the Dutch disease mechanisms.6 Yet, the contribution
of the building and public works sub-sector to GDP is particularly strong
in oil-producing countries because authorities in these countries often
use oil revenues to finance the development of public infrastructure
such as roads, schools, hospitals and government buildings.
Impact of external factors on industrialization in Africa

Trade with countries in transition delays the industrialization process in
Africa (see Tables 7.4 and 7.5), because transition countries have a com-
parative advantage in labor costs. Their products are highly competitive
in some African countries. Thus, the authors’ conclusions (Rowthorn
and Coutts, 2004; Sung and Hongshik, 2011) can be relativized, that the
importation of labor-intensive manufactured goods in developing coun-
tries such as China or India causes de-industrialization in developed
countries and other developing countries. Such de-industrialization has
worsened with the structural adjustment programs of the 1980s. These
programs were based on the assumption that markets could effectively
regulate resource allocation, while government interventions were inef-
fective because they distort market signals. For this reason, all import
substitution strategies, including the protection of local businesses, tar-
iffs and other import restrictions, price controls and credit ceilings were
removed. These programs have helped to liberalize trade and privatize
public enterprises (ECA, 2013). UNCTAD (2010) notes that trade liber-
alization was premature because it exposed infant African industries to
competition from products from Asian developing countries.
138
Table 7.4 Contribution of trade with countries in transition to development of manufacturing industries in Africa
Lagged endogenous −0.295∗∗∗ −0.278∗∗∗ −0.439∗∗∗ −0.367∗∗∗ −0.289∗∗∗ −0.273∗∗∗ −0.423∗∗∗ −0.358∗∗∗
(0.0322) (0.0343) (0.0555) (0.0424) (0.0323) (0.0346) (0.0554) (0.0423)
Average duration of war – 0.0979 0.145 0.0374 0.288∗∗ 0.0938 0.138 0.0147 0.302∗∗∗
5 year periods (0.0857) (0.0921) (0.136) (0.115) (0.0854) (0.0923) (0.135) (0.114)
Urbanization rate 0.0263 0.280∗∗∗ 0.12∗∗∗ 0.0557 0.0221 0.085∗∗∗ 0.110 0.0333
(0.0519) (0.0548) (0.0111) (0.0774) (0.0517) (0.0549) (0.0906) (0.0776)
Customs duties (CD) 1.181∗∗∗ 1.108∗∗∗ 2.223∗∗∗ 0.866∗∗∗ 1.167∗∗∗ 1.100∗∗∗ 2.332∗∗∗ 0.866∗∗∗
(0.244) (0.264) (0.412) (0.306) (0.243) (0.264) (0.411) (0.303)
openness/DvdC (0.143) (0.149) (0.221) (0.236) (0.142) (0.149) (0.220) (0.234)
Increase in degree of trade −0.734∗∗∗ −0.687∗∗∗ −1.697∗∗∗ −0.393 −0.670∗∗∗ −0.646∗∗∗ −1.539∗∗∗ −0.231
openness/trans (0.150) (0.161) (0.233) (0.239) (0.157) (0.168) (0.237) (0.254)
GDP growth rate (GR) 0.891∗∗∗ 0.903∗∗∗ 0.449∗∗∗ 1.007∗∗∗ 0.885∗∗∗ 0.900∗∗∗ 0.416∗∗∗ 0.988∗∗∗
(0.0494) (0.0530) (0.0858) (0.0607) (0.0493) (0.0530) (0.0859) (0.0609)
GDP* GR growth rate 0.0303 0.0220 −0.00962 −0.0251 0.0244 0.0179 −0.0635 −0.0270
(0.0540) (0.0590) (0.0762) (0.0717) (0.0539) (0.0591) (0.0779) (0.0709)
Investment growth rate 0.0158∗ 0.0168 −0.0104 −0.00129 0.0155∗ 0.0166 −0.0111 −0.00424
(0.009) (0.0103) (0.0153) (0.0143) (0.0093) (0.0102) (0.0152) (0.0142)
Labor force growth rate 0.301∗∗∗ 0.274∗∗∗ 0.294∗∗∗ 0.232∗∗∗ 0.300∗∗∗ 0.273∗∗∗ 0.280∗∗∗ 0.231∗∗∗
(0.0571) (0.0621) (0.0868) (0.0752) (0.0569) (0.0620) (0.0865) (0.0743)
Increase in school 0.090∗∗∗ 0.095∗∗∗ 0.146∗∗∗ 0.157∗∗∗ 0.085∗∗∗ 0.090∗∗∗ 0.113∗∗ 0.148∗∗∗
enrolment rate (0.0309) (0.0321) (0.0506) (0.0499) (0.0310) (0.0325) (0.0514) (0.0496)
(DD)∗ (Increase in degree of −0.0312 −0.0223 −0.103∗∗∗ −0.0574∗
trade openness/trans (0.0228) (0.0263) (0.0339) (0.0320)
countries)
Years (1976–1980) −5.065∗ −9.017∗∗∗ −11.14∗ −4.945 −5.402∗∗ −9.309∗∗∗ −15.60∗∗ −5.526∗
(2.709) (3.289) (6.569) (3.101) (2.708) (3.300) (6.689) (3.084)
Years (1981–1985) −6.341∗∗ −0.364 −1.569 −6.144∗ −6.043∗ −0.0826 2.965 −5.964∗
(3.200) (4.171) (8.511) (3.207) (3.194) (4.177) (8.592) (3.173)
Years (1986–1990) −1.301 −5.079 −5.430 −3.748 −1.652 −5.371 −8.054 −3.907
(4.288) (4.645) (5.050) (6.301) (4.276) (4.649) (5.091) (6.232)
Years (1991–1995) −6.980 −7.651 −7.021 −7.533 −7.016 −7.624 −7.877 −8.483
(4.806) (5.337) (5.443) (5.532) (4.785) (5.327) (5.417) (5.496)
Years (1996–2000) −7.284∗∗ −3.271 −4.651 −2.694 −6.762∗∗ −2.947 −3.541 −2.508
(3.279) (3.864) (3.753) (3.244) (3.287) (3.876) (3.748) (3.210)
Years (2001–2005) −2.135 −9.751 −3.993∗ −4.773∗ −2.947 −10.48∗ −6.050∗∗∗ −5.509∗∗
(4.674) (6.032) (2.159) (2.497) (4.690) (6.081) (2.248) (2.504)
Years (2006–2010) −0.620 5.361 −1.398 −1.760 −0.436 5.648 −1.691 −1.904
(3.022) (4.524) (1.633) (2.114) (3.012) (4.529) (1.626) (2.092)
Observations 302 271 134 155 302 271 134 155
Number of 05-year periods 8 8 8 8 8 8 8 8
Sargan Test 317(0.00) 370(0.00) 220(0.02) 245(0.01) 318(0.00) 377(0.00) 214(0.023) 254(0.011)
Note: Dependent variable: industrialization rate (added value of manufacturing industries to the GDP); DEP dependency ratio to raw materials, SSA:
Sub-Saharan Africa. ∗∗∗ p < 0. 01, ∗∗ p < 0. 05, ∗ p < 0.
139
140
Table 7.5 Contribution of trade with countries in transition to industrial development in Africa
Africa SSA DEP>12 DEP<12 Africa SSA DEP>12 DEP<12
Lagged endogenous −0.30∗∗∗ −0.28∗∗∗ −0.447∗∗∗ −0.352∗∗∗ −0.298∗∗∗ −0.27∗∗∗ −0.431∗∗∗ −0.342∗∗∗
(0.0318) (0.0338) (0.0550) (0.0419) (0.0320) (0.0341) (0.0552) (0.0418)
Average duration war – 0.109 0.144∗ 0.0134 0.288 0.105 0.141 −0.0234 0.301∗∗∗
5year (0.0766) (0.0820) (0.117) (0.303) (0.0764) (0.0923) (0.118) (0.102)
Urbanization rate 0.0385 0.0438 0.0884 0.858∗∗∗ 0.0356 0.0427∗∗∗ 0.0850 0.0656
(0.0467) (0.0492) (0.0812) (0.0695) (0.0466) (0.0093) (0.0810) (0.0697)
Customs duties (CD) 1.179∗∗∗ 1.077∗∗∗ 2.029∗∗∗ 0.897∗∗∗ 1.170∗∗∗ 1.075∗∗∗ 2.098∗∗∗ 0.892∗∗∗
(0.216) (0.234) (0.367) (0.265) (0.215) (0.234) (0.366) (0.262)
openness/DvdC (0.127) (0.132) (0.195) (0.211) (0.126) (0.132) (0.195) (0.209)
Increase in degree of trade −0.72∗∗∗ −0.63∗∗∗ −1.590∗∗∗ −0.356∗ −0.67∗∗∗ −0.61∗∗∗ −1.447∗∗∗ −0.220
openness/trans (0.134) (0.143) (0.210) (0.216) (0.140) (0.150) (0.215) (0.228)
GDP growth rate (GR) 0.796∗∗∗ 0.811∗∗∗ 0.412∗∗∗ 0.920∗∗∗ 0.791∗∗∗ 0.809∗∗∗ 0.382∗∗∗ 0.905∗∗∗
(0.0446) (0.0475) (0.0750) (0.0555) (0.0446) (0.0476) (0.0755) (0.0556)
GDP* GR growth rate 0.0487 0.0358 0.0199 −0.00897 0.0439 0.0338 −0.0299 −0.0108
(0.0486) (0.0528) (0.0687) (0.0645) (0.0486) (0.0530) (0.0706) (0.0639)
Investment growth rate 0.274∗∗∗ 0.240∗∗∗ 0.256∗∗∗ 0.198∗∗∗ 0.273∗∗∗ 0.239∗∗∗ 0.238∗∗∗ 0.198∗∗∗
(0.0516) (0.0559) (0.0770) (0.0677) (0.0514) (0.0560) (0.0771) (0.0670)
Labor force growth rate 0.11∗∗∗ 0.11∗∗∗ 0.138∗∗∗ 0.180∗∗∗ 0.105∗∗∗ 0.11∗∗∗ 0.106∗∗ 0.172∗∗∗
(0.0278) (0.0289) (0.0455) (0.0452) (0.0279) (0.0293) (0.0466) (0.0449)
Increase in school 0.0127 0.0133 −0.0105 −0.00233 0.0125 0.0133 −0.0102 −0.00488
enrolment rate (IDT) (0.00845) (0.00920) (0.0138) (0.0127) (0.00843) (0.00920) (0.0137) (0.0127)
(DD)∗ (IDT) −0.0250 −0.0111 −0.091∗∗∗ −0.0498∗
(0.0205) (0.0235) (0.0308) (0.0289)
Years (1976–1980) −5.550∗∗ −7.63∗∗∗ −9.198∗ −4.352∗ −5.716∗∗ −7.71∗∗∗ −12.55∗∗∗ −4.762∗
(2.229) (2.488) (4.716) (2.568) (2.226) (2.492) (4.838) (2.552)
Years (1981–1985) −5.823∗∗ −2.798 −2.651 −6.966∗∗ −5.775∗∗ −2.779 0.326 −6.912∗∗
(2.499) (2.816) (5.865) (2.796) (2.492) (2.817) (5.938) (2.766)
Years (1986–1990) −2.942 −3.851 −6.254∗ −8.711∗ −3.098 −3.928 −7.636∗∗ −8.521∗
(3.508) (3.660) (3.527) (4.786) (3.500) (3.663) (3.550) (4.737)
Years (1991–1995) −4.897 −6.986 −2.575 −3.829 −4.954 −6.983 −3.412 −5.023
(4.488) (4.951) (4.635) (4.615) (4.475) (4.951) (4.633) (4.619)
Years (1996–2000) −6.116∗∗ −1.541 −5.368 −1.721 −5.773∗ −1.416 −4.304 −1.467
(2.969) (3.393) (3.332) (2.846) (2.973) (3.403) (3.343) (2.820)
Years (2001–2005) −2.684 −8.705 −3.446∗ −5.793∗∗∗ −3.361 −9.047 −5.369∗∗∗ −6.387∗∗∗
(5.601) (6.612) (1.896) (2.221) (5.610) (6.648) (1.998) (2.225)
Years (2006–2010) −1.290 2.547 −1.093 −2.456 −1.120 2.658 −1.205 −2.524
(3.394) (4.346) (1.468) (1.966) (3.386) (4.352) (1.465) (1.946)
Observations 302 271 134 155 302 271 134 155
Number of 05-year periods 8 8 8 8 8 8 8 8
Sargan Test 350(0.00) 372(0.00) 223(0.0213) 239(0.013) 348(0.00) 387(0.00) 218(0.022) 244(0.010)
Note: Dependent variable: industrialization rate (added value of industries to the GDP); DEP dependency ratio to raw materials, SSA: Sub-Saharan Africa.
∗∗∗ p < 0. 01, ∗∗ p < 0. 05, ∗ p < 0.
Source: Authors’ estimate

141
However, the impact of trade with developed countries fosters the

industrialization process in African countries (see Tables 7.2, 7.3, 7.A.1
and 7.A.2), as the former are the leading importers of African prod-
ucts. Further, according to Coe and Helpman (1995); Higino Schneider
(2005), productivity improves with the volume of imports into countries
that produce high technology goods. Goods imports are the channel
through which innovations from developed countries are introduced
into the production systems of poor countries. But, for most African
industries, inadequate technological capabilities required for innova-
tion and development of new industries are huge challenges. Often the
technologies used in the production process in Africa are inappropriate.
A technology lag has reduced the competitiveness of many industries
on the continent (UNCTAD, 2003).
South-South trade restriction through customs duties does not yield
any significant results. However, restriction of trade openness through
indiscriminate increases in customs duties fosters industrialization in
Africa. Customs duties impact industrialization through several chan-
nels; for instance, they contribute to investment financing and often
lead to the reallocation of resources to goods production which can
drive other sectors of the economy. This applies, in particular, to dams,
which are sometimes financed by customs revenue. In Africa, customs
duties were used to protect infant industries during the 1970s and
1980s (Greenaway et al., 2002; Kala Krishnaa et al., 2003; Yanikkaya,
2003). However, a reduction in customs duty can sustain the pro-
cess of industrialization through the demand of technology-intensive
goods (Jorn Ratts and Stokke, 2009). The latter observation has been
tested for the case of resource rich countries (see the last-but-one col-
umn of Tables 7.4 and 7.5). Indeed, these tables show restriction of
exchanges with transition countries through customs duty exacerbates
the negative contribution of trade with the transition countries on
industrialization.
5 Conclusion
This study assessed the impact of international trade on industrializa-

tion in Africa. Following a literature review, an econometric analysis
was conducted, based on data from the IMF and UN Comtrade database
for the 1971–2010 period. The estimates led to the following conclu-
sions: First, de-industrialization of the manufacturing sub-sector is really
beginning in Africa. According to UNCTAD data, the industrialization
rate of the manufacturing sub-sector decreased from 13.91% in 1971 to
9.59% in 2010. Overall, however, the continent is industrializing, since

the industrialization rate rose from 30.36% in 1971 to 36.97% in 2010.
This industrialization is mostly due to the extractive sector and min-
ing. The contribution of mining and extractive industries increased from
25.52% in 1971 to 31.61% in 2010. Second, trade with countries in tran-
sition causes de-industrialization while intra-African trade has no major
impact on industrialization in Africa. In addition to internal problems
such as lack of infrastructure, the energy deficit and poor governance,
de-industrialization is caused by competition from imported products
from some countries in transition. Third, trade with developed countries
fosters industrialization in Africa, because Africa’s natural resource-rich
countries sell their industrial products to developed countries. Trade
between developed countries and African countries ensures the transfer
of technologies favorable to industrial development. In these circum-
stances, African countries must strengthen trade ties with developed
countries and, above all, avoid discriminatory protectionist policies.
Annexes
Table 7.A.1 Added value of manufacturing industries and trade openness in Africa
144
v1 Afrique ASS DEP>12 DEP<12 Afrique ASS DEP>12 DEP<12
Lag of endogenous variable 0.287∗∗∗ 0.334∗∗∗ 0.149∗∗ 0.343∗∗∗ 0.290∗∗∗ 0.337∗∗∗ 0.161∗∗ 0.335∗∗∗
(0.0471) (0.0503) (0.0780) (0.0616) (0.0473) (0.0505) (0.0804) (0.0614)
1.183∗∗∗ 1.078∗∗∗ 2.068∗∗ 0.676∗∗∗ 1.151∗∗∗ 1.075∗∗∗ 2.075∗∗∗ 0.628∗∗∗
(0.146) (0.161) (0.226) (0.201) (0.147) (0.161) (0.231) (0.202)
0.0435 0.0677 0.131 0.0491 0.0244 0.0535 0.133 0.0180
(0.573) (0.0594) (0.0930) (0.0748) (0.0578) (0.0601) (0.0947) (0.0758)
Costums duties (CD) −0.00306 −0.0158 0.106 −0.0465 −0.00421 −0.0159 0.0988 −0.0385
(0.0389) (0.0403) (0.0712) (0.0557) (0.0391) (0.0405) (0.0730) (0.0556)
0.240∗∗∗ 0.279∗∗∗ 0.152 0.374∗∗ 0.292∗∗∗ 0.312∗∗∗ 0.175 0.452∗∗∗
(0.0912) (0.0949) (0.137) (0.156) (0.0940) (0.0984) (0.141) (0.160)
−0.21∗∗∗ −0.178∗∗∗ −0.301∗∗∗ −0.162∗∗ −0.180∗∗∗ −0.161∗∗ −0.271∗∗∗ −0.137∗
(0.0608) (0.0639) (0.0855) (0.0814) (0.0625) (0.0659) (0.0920) (0.0823)
War duration in 05 years −0.29∗∗∗ −0.218∗∗∗ −0.885∗∗∗ −0.107 −0.299∗∗∗ −0.25∗∗∗ −0.885∗∗∗ −0.155
(0.0750) (0.0803) (0.122) (0.105) (0.0854) (0.0906) (0.142) (0.121)
0.0541 0.0391 −0.0149 −0.00790 0.0446 0.0356 −0.0172 −0.00838
(0.0415) (0.0456) (0.0575) (0.0525) (0.0418) (0.0458) (0.0587) (0.0522)
Urbanization rate 0.509∗∗∗ 0.530∗∗∗ 0.247∗∗∗ 0.606∗∗∗ 0.506∗∗∗ 0.531∗∗∗ 0.244∗∗∗ 0.613∗∗∗
(0.0372) (0.0401) (0.0559) (0.0488) (0.0378) (0.0404) (0.0584) (0.0498)
0.393∗∗∗ 0.378∗∗∗ 0.310∗∗ 0.988∗∗∗ 0.390∗∗∗ 0.388∗∗∗ 0.324∗∗∗ 0.972∗∗∗
(0.0942) (0.0970) (0.0146) (0.0151) (0.0971) (0.0999) (0.0154) (0.0155)
−0.0467∗∗ −0.01317 −0.0293 −0.0463∗
(0.0190) (0.0211) (0.0273) (0.0257)
0.0141 0.0206 0.00364 0.0376
(0.0184) (0.0196) (0.0315) (0.0259)
0.751∗∗∗ 0.293∗∗∗ 0.410∗∗ 0.97∗∗∗ 0.369∗∗∗ 0.932∗∗∗ 0.552∗∗ 0.81∗∗∗
(0.854) (0.036) (0.032) (0.0249) (0.068) (0.054) (0.0602) (0.0046)
Growth rate of trade with 4.557∗∗ 5.353∗∗∗ 5.421∗ 0.689 4.821∗∗ 5.485∗∗∗ 5.806∗ 0.792
Asian LDCs (1.867) (1.928) (3.186) (2.591) (1.876) (1.935) (3.276) (2.592)
−0.0843 −0.237 2.681 −4.716 −0.606 −0.599 1.961 −4.618
(2.250) (2.276) (2.718) (5.079) (2.267) (2.297) (2.894) (5.055)
0.995∗∗∗ 0.091 0.521∗∗ 0.20∗∗∗ 0.707∗∗∗ 0.248 0.365∗ 0.58∗∗
(0.072) (0.163) (0.059) (0.020) (0.092) (0.185) (0.0817) (0.0216)
Growth rate of trade/ 4.169∗∗ 8.093∗∗∗ 5.098∗∗ 4.714∗∗∗ 4.353∗∗ 7.945∗∗∗ 5.147∗∗ 4.606∗∗∗
African countries (2.017) (2.266) (2.289) (1.351) (2.025) (2.276) (2.357) (1.344)
4.619 3.232 6.433∗∗∗ 2.182 4.771 3.391 6.158 2.344
(3.111) (3.177) (1.366) (1.646) (3.121) (3.186) (1.433) (1.647)
4.663∗∗ 4.229∗∗ 7.151∗∗∗ 3.073∗ 4.207∗∗ 3.939∗∗ 6.839∗∗∗ 2.926∗
(1.898) (1.957) (1.211) (1.595) (1.918) (1.987) (1.301) (1.586)
Observations 316 285 150 153 316 285 150 153
8 8 8 8 8 8 8 8
396 (0.00) 341 (0.000) 174 (0.001) 175 (0.01) 389 (0.000) 337 (0.00) 167 (0.012) 173 (0.013)
Note: Dependent variable: rate of industrialization (added value of industries); DEP, dependency ratio to primary commodities, SSA: Sub-Saharan Africa.
***p < 0. 01, ∗∗ p < 0. 05, ∗ p < 0.
145
146
Table 7.A.2 Industrial value added and trade openness in Africa
Africa ASS DEP>12 DEP<12 Africa ASS DEP>12 DEP<12
Lag of endogenous variable 0.0539 0.0676 −0.0117 −0.0355 0.0583 0.0730 −0.0106 −0.0329
(0.0465) (0.0500) (0.0697) (0.0592) (0.0470) (0.0505) (0.0710) (0.0599)
Customs duties (CD) 0.414∗∗∗ 0.414∗∗∗ 0.720∗∗∗ 0.183∗∗ 0.407∗∗∗ 0.407∗∗∗ 0.728∗∗∗ 0.176∗∗
(0.0675) (0.0701) (0.121) (0.0844) (0.0683) (0.0708) (0.124) (0.0863)
War duration in 05 years 0.0156 0.0303 −0.0161 0.0491 0.0121 0.0276 −0.0178 0.0478
(0.0278) (0.0280) (0.0448) (0.0308) (0.0282) (0.0285) (0.0456) (0.0314)
Urbanization rate −0.00513 −0.0127 0.0181 0.00147 −0.00495 −0.0120 0.0125 0.00220
(0.0171) (0.0175) (0.0373) (0.0204) (0.0172) (0.0176) (0.0384) (0.0207)
Growth rate of trade with 0.301∗∗∗ 0.317∗∗∗ 0.309∗∗∗ 0.347∗∗∗ 0.310∗∗∗ 0.321∗∗∗ 0.337∗∗∗ 0.353∗∗∗
developed countries (0.0447) (0.0455) (0.0724) (0.0621) (0.0464) (0.0475) (0.0775) (0.0658)
Growth rate of trade with −0.0212 −0.00705 −0.0975∗∗ 0.0104 −0.0174 −0.00559 −0.0797∗ 0.0114
Asian LDCs (0.0283) (0.0288) (0.0418) (0.0310) (0.0291) (0.0301) (0.0449) (0.0319)
Growth rate of trade with −0.112∗∗∗ −0.104∗∗∗ −0.363∗∗∗ −0.0733∗ −0.116∗∗∗ −0.117∗∗∗ −0.36∗∗∗ −0.0814∗
African countries (0.0349) (0.0362) (0.0636) (0.0400) (0.0396) (0.0406) (0.0737) (0.0479)
Growth rate of with 0.0249 0.0336 0.00686 0.00496 0.0237 0.0325 0.00802 0.00481
physical capital (0.0199) (0.0209) (0.0306) (0.0203) (0.0199) (0.0210) (0.0312) (0.0205)
Growth rate of labor force 0.0604∗∗∗ 0.0662∗∗∗ −0.00291 0.0865∗∗∗ 0.0606∗∗∗ 0.0675∗∗∗ −0.00661 0.0882∗∗∗
(0.0153) (0.0163) (0.0284) (0.0174) (0.0157) (0.0166) (0.0302) (0.0181)
Growth rate school −0.0420 −0.0274 −0.0549 0.187∗∗∗ −0.0398 −0.0193 −0.0586 0.189∗∗∗
enrollment (gross) (0.0415) (0.0420) (0.0665) (0.0543) (0.0432) (0.0441) (0.0711) (0.0564)
(Growth rate of trade/Asian −0.00705 −0.00552 −0.0178 −0.00225
LDCs)∗ (CD) (0.00914) (0.00983) (0.0144) (0.0101)
(Growth rate of trad /Afri 0.00409 0.00779 0.000502 0.00421
countries)∗ (DD) (0.00879) (0.00904) (0.0164) (0.0102)
Years (1976–1980) −0.726 −0.639 −0.352 0.849 −0.734 −0.665 −0.386 0.867
(0.641) (0.743) (0.943) (0.924) (0.643) (0.747) (0.963) (0.931)
Years (1981–1985) 1.324 1.076 −0.139 0.990 1.352 1.207 −0.405 1.032
(0.929) (1.080) (1.790) (0.942) (0.952) (1.102) (1.876) (0.991)
Years (1986–1990) 0.824 0.960 1.096 2.159∗ 0.758 0.828 1.003 2.130∗
(0.738) (0.938) (0.964) (1.261) (0.744) (0.952) (0.988) (1.272)
Years (1991–1995) 0.183 4.75e-05 −1.018 1.367 0.115 0.0890 −1.259 1.406
(0.951) (1.372) (1.100) (0.895) (0.968) (1.399) (1.139) (0.920)
Years (1996–2000) 1.612 1.307 2.712∗∗ 0.885∗ 1.665 1.361 2.713∗ 0.886
(1.200) (1.229) (1.382) (0.534) (1.206) (1.237) (1.412) (0.547)
Years (2001–2005) 0.184 −0.0844 0.0120 0.497 0.203 −0.00697 −0.161 0.519
(0.787) (0.836) (0.667) (0.737) (0.794) (0.847) (0.700) (0.744)
Years (2006–2010) −0.595 −0.560 0.368 0.585 −0.646 −0.654 0.238 0.566
(0.666) (0.779) (0.511) (0.740) (0.670) (0.789) (0.537) (0.750)
Observations 318 286 152 153 318 286 152 153
Number of 05 years periods 8 8 8 8 8 8 8 8
Test de Sargan 328 (0.00) 309(0.000) 154(0.011) 184(0.00) 326(0.000) 306(0.00) 147(0.02) 182(0.00)
Note: Dependent variable: industrialization rate (value added of manufacturing industries); DEP, dependency ratio to primary commodities, SSA: Sub-
Saharan Africa. ∗∗∗ p < 0. 01, ∗∗ p < 0. 05, ∗ p < 0.
147
Notes
1. These countries are Burkina Faso, Burundi, Ghana, Senegal, Rwanda, Sierra
Leone, South Africa and Tanzania.
2. Furthermore, trade openness is sometimes marked by the dismantling of
non-tariff barriers which improves economic efficiency and corrects the nat-
ural inequalities in resource endowment of states. Indeed, the availability
of imported inputs is essential to boost industrial production in developing
economies (Lee, 1993). Thus, trade restrictions should reduce industrial pro-
duction because they prevent domestic industries from importing some inputs
at lower cost (Lee, 1995). In addition, through trade openness, countries cer-
tainly import inputs such as machinery, but they mainly increase the market
size of their domestic producers as well as returns to innovation (Harrison,
1996; Madsen, 2009).
3. However, some of these variables have been used by Rowthorn and Ken Coutts
(2004), Brady et al. (2011), Rowthorn and Ramaswamy (1999) to analyze the
industrialization process in Europe and Latin America.
4. To take into account possible cyclical effects, the practice in empirical lit-
erature is to work with data calculated as averages over a five-year period.
Thus, the study period 1971–2010 (40 years) is relatively long to work with
five-year averages. Moreover, Chang, Kaltani and Loayza (2009) have intro-
duced dummy variables for every five-year period to improve the quality of
estimates. The dummy variables for each five-year period will estimate the
effect of independent variables on the dependent variable while controlling
the impact of cyclical fluctuations.
5. In UNCTAD terminology (2008), the secondary sector consists of three cat-
egories of industries, extractive industries (mining, oil and businesses that
produce essential goods such as water and electricity), manufacturing indus-
tries and companies involved in construction (buildings and public works).
We will, on the one hand, identify the factors that explain the contribution
of the secondary sector (three types of industry) to GDP, and on the other,
identify only the factors of industrialization of manufacturing firms.
6. 50% of the sample countries derive over 12% of their GDP from exports of
raw materials. We have adopted this classification to avoid creating dummy
variables for certain raw materials. This approach enables combining all the
raw materials into a single variable and reason about various commodities
and not just one. The classification of raw materials was made by UNCTAD
(2009).
7. However, in its 2011 report, the AfDB (2011) acknowledges that the African
manufacturing sector is unable to catch up with the Asian manufacturing
industry. Indeed, in several African countries, the share of manufactur-
ing in GDP is around 10% or less. It is in South Africa, Cameroon, Côte
d’Ivoire, Egypt, Lesotho, Madagascar, Morocco, Mauritius, Namibia, Tunisia
and Zimbabwe that such industries are developed (between 15 and 20%
of GDP). According to the AfDB, the under-performance of manufacturing
industries is due to the lack of infrastructure, unreliable energy supply and
bureaucracy. It also recognizes that African manufacturing firms face fierce
competition from products imported from emerging countries such as China
(AfDB, 2011).
References
AfDB (2011), African Economic Outlook, AfDB/OECD.
Alderson, A. (1999), ‘Explaining Deindustrialization: Globalization, Failure, or
Success?’, American Sociological Review, vol. 64, no. 5, pp. 701–721.
Arellano, M. and Bond, S. (1991), ‘Some Tests of Specification for Panel Data:
Monte Carlo Evidence and an Application to Employment Equations’, Review
of Economic Studies, vol. 58, pp. 277–297.
Arellano, M. and Bover, O. (1995), ‘Another Look at the Instrumental-Variable
Estimation of Error-Components Models’, Journal of Econometrics, vol. 68,
pp. 29–52.
Barro, R. and Sala -I- Martin, X. (1995), Economic Growth, New York: McGraw-Hill.
Blundell, R. and Bond, S. (1998), ‘Initial Conditions and Moment Restric-
tions in Dynamic Panel Data Models’, Journal of Econometrics, vol. 87,
pp. 115–143.
Brady, D., Kaya, Y. and Gereffi, G. (2011), ‘Stagnating Industrial Employment in
Latin America’, Work and Occupations, vol. 38, no. 2, pp. 179–220.
Cardoso, F. and Faletto, E. (1979), Dependency and Development in Latin America,
Berkeley: University of California Press.
Cieslik, A. and Tarsalewska, M. (2011), ‘External Openness and Economic Growth
in Developing Countries’, Review of Development Economic, vol. 15, no. 4,
pp. 729–744.
Coe, D. and Helpman, E. (1995), ‘International R&D Spillovers’, European Eco-
nomic Review, vol. 39, pp. 859–887.
Clark, C. (1957), The Conditions of Economic Progress, London: Macmillan.
ECA (2013), Industrialization for Africa’s Economic Emergence, Abstract, Sixth Con-
ference of African Ministers of Finance, Planning and Economic Develop-
ment of the Economic Commission for Africa and the Conference of African
Ministers of Economy and Finance of the African Union, Abidjan.
Edwards, S. (1998), ‘Openness, Productivity and Growth: What do we Really
Know?’, The Economic Journal, vol. 108, no. 447, pp. 383–398.
Elbadawi, I. and Helleiner, G. (1998), ‘African Development in the Context of
New World Trade and Financial regimes: The Role of WTO and Its Relationship
to the World Bank and the IMF’, Presented at AERC Workshop on Africa and
the New World Trade System, Mombasa, Kenya, 27–29 April.
Frankel, J. and Rome, D. (1999), ‘Does Trade Cause Growth?’, American Economic
Review, vol. 89, no. 3, pp. 379–399.
Greenaway, D., Morgan, W. and Wright, P. (2002), ‘Trade Liberalisation and
Growth in Developing Countries’, Journal of Development Economics, vol. 67,
pp. 229–244.
Harrison, A. (1996), ‘Openness and Growth: A Time Series, Cross-Country Anal-
ysis for Developing Countries’, Journal of Development Economics, vol. 48,
pp. 419–447.
Hausmann, R. and Rodrik, D. (2006), Doomed to Choose: Industrial Policy as
Predicament,t Cambridge, MA: Harvard University, John F. Kennedy School of
Government.
Higino Schneider (2005), ‘International Trade, Economic Growth and Intellectual
Property Rights: A Panel Data Study of Developed and Developing Countries’,
Journal of Development Economics, vol. 78, pp. 529–547.
Holtz-Eakin, D., Newey, W. and Rosen, H. (1988), ‘Estimating Vector

Autoregression with Panel Data’, Econometrica, vol. 56, pp. 1371–1396.
Hossein, J. and Weiss, J. (1999), ‘De-Industrialization in Sub-Saharan Africa: Myth
or Crisis?’, Journal of African Economics, vol. 9, no. 1, pp. 24–43.
IMF (2012), Direction of Trade Statistics (DOTS) Database.
Jorn Ratts and Stokke, H. E. (2009), Trade Barriers to Growth in South Africa:
Endogenous Investment-Productivity-Trade Interaction, Department of Economics,
Trondheim, Norway: Norwegian University of Science and Technology.
Kala Krishnaa, B., Ozyildirimc, A. and Swanson, N. R. (2003), ‘Trade, Investment
and Growth: Nexus, Analysis and Prognosis’, Journal of Development Economics,
vol. 70, pp. 479–499.
Keller, W. (1999), ‘How Trade Patterns and Technology Flows Affect Productivity
Growth’, NBER Working Paper, r no. 6690.
Lawrence, R. Z. (1987), ‘Trade Performance as a Constraint on European Growth’,
in R. Lawrence and C. Schultze (eds.), Barriers to European Growth: A Transat-
lantic View, Washington, DC: Brookings Institution, pp. 303–383.
Lee, J. (1993), ‘International Trade, Distortions and Long-Run Growth’, IMF Staff
Papers, vol. 40, no. 2, pp. 299–328.
Lee, J. (1995), ‘Capital Goods Imports and Long-Run Growth’, Journal of Develop-
ment Economics, vol. 48, no. 1, pp. 91–110.
Madsen, J. (2009), ‘Trade Barriers, Openness, and Economic Growth’, Southern
Economic Journal, vol. 76, no. 2, pp. 397–418.
Pack, H. (1988), ‘Industrialization and Trade’, Chapter 9 in Chenry and T.N.
Srinivasan (eds.), Handbook of Development Economics, vol. 16, pp. 177–206.
Rivera-Batiz and Romer, P. (1991), ‘Economic Integration and Endogenous
Growth’, Quarterly Journal of Economics, vol. 106, no. 2, pp. 531–555.
Rowthorn, R. and Coutts, K. (2004), ‘De-Industrialisation and the Balance of Pay-
ments in Advanced Economies’, Cambridge Journal of Economics, vol. 28, no. 5,
pp. 767–790.
Rowthorn, R. and Ramaswamy, R. (1999), ‘Growth, Trade and Deindustrializa-
tion’, IMF Staff Papers, Mars 1999, vol. 46, no. 1, pp. 18–40.
Sung and Hongshik (2011), ‘Foreign Direct Investment and De-industrialization’,
The World Economy, 10.1111/j.1467-9701.2010.01324.x.
UNCTAD (2003), 2003 Report on Least Developed Countries, Industrialization
United Nations, New York and Geneva, 2009.
UNCTAD (2008), 2008 Report on Least Developed Countries (LDC), Growth, Poverty
and Conditions of Partnership for Development, t United Nations, New York and
Geneva, 2008.
UNCTAD (2009), 2009 Report on Least Developed Countries, The State and Develop-
ment Governance, United Nations, Geneva.
UNCTAD (2010), 2009 Report on Least Developed Countries, Towards a new Interna-
tional Development Structure for LCDs, United Nations, New York and Geneva.
Wood, A. (1994), Nord-South Trade, Employment, and Inequality, Oxford, England:
Oxford University Press.
Wood, A. and Mayer, J. (1998), African Export Structure: A Comparative Perspective,
Geneva: Institute of Development Studies and UNCTAD.
Yanikkaya, H. (2003), ‘Trade Openness and Economic Growth: A Cross-Country
Empirical Investigation’, Journal of Development Economics, vol. 72, pp. 62–96.
Part IV
Impact of Currency Union on
Trade
8
Impact of Monetary Unions on
Trade: The Case of WAEMU
Ibrahima Camara
Introduction
The idea of establishing a monetary union is not new. ‘I want all

of Europe to have a single currency; that would facilitate trade’, said
Napoleon Bonaparte (1769–1821). The same idea is increasingly topical
in economic policy debates. There have been several monetary integra-
tion initiatives on the African continent, and the goal of establishing a
monetary union for all of Africa by 2021 was announced by the Associa-
tion of African Central Bank Governors in 2003. The African Union’s
strategy rests on the prior establishment of monetary unions within
existing regional economic communities.1 These regional unions would
constitute an intermediate step toward establishing a central bank and
an African currency (Masson and Patillo, 2004). The Economic Commu-
nity of West African States (ECOWAS) has been part of these initiatives.
ECOWAS, the West African intergovernmental organization estab-
lished in 1975, seeks mainly to promote cooperation and integration,
and works toward the establishment of a West African economic and
monetary union. In 1983, the heads of state and government at a con-
ference organized in Conakry, Guinea, outlined the vision of creating a
common currency, which was initially expected to enter into force in
1994.2
ECOWAS currently has 15 member states grouped into two major
monetary zones. Eight countries comprise the West African Economic
and Monetary Union (WAEMU): Benin, Burkina Faso, Côte d’Ivoire,
Guinea-Bissau, Mali, Niger, Senegal and Togo. Six countries form the
West African Economic Monetary Zone (WAEMZ): Gambia, Ghana,
Guinea, Nigeria, Liberia and Sierra Leone.
WAEMU was established by a treaty signed in Dakar on 10 January
1994 by the heads of state and governments of the seven West African
countries that use the African Financial Community Franc (CFAF), a
153
154 Impact of Monetary Unions on Trade
common currency. It was established in various stages. Prior to 1994,

two separate treaties governed monetary and economic integration
between WAEMU countries:
• the West African Monetary Union (WAMU) treaty signed in 1962,

which dealt with monetary integration issues;
• the West African Economic Community (WAEC) treaty signed in
1973, which was responsible for promoting trade and sectoral coop-
eration between member countries.
Until the 1980s, each member state of the Union was responsi-
ble for its macroeconomic policy and fiscal policy and all shared a
common monetary policy. This resulted in macroeconomic instabil-
ity, making the common monetary policy inefficient. Monetary policy
was also weakened by the economic and financial crisis experienced
by member states during the 1980s, after which it became necessary to
consolidate WAMU.
Considering the treaties had failed to achieve their objectives, given
the trend toward multiplication of regional blocs, and in a bid to revive
the integration process in West Africa through WAMU, the heads of state
of the Union in March 1990 instructed the then governor of the Central
Bank of West African States (BCEAO) to submit a succinct economic
integration plan. This was expected to return member states onto a path
of growth.
With support from the European Union (EU), France instigated the
establishment of WAEMU in June 1991. By 10 January 1994 in Dakar,
Senegal, following the devaluation of the CFAF, its establishment finally
became effective. The treaty to institute the Union entered into force on
1 August 1994 after ratification by the seven founding members: Benin,
Burkina-Faso, Côte d’Ivoire, Mali, Niger, Senegal and Togo. Three years
later, in May 1997, Guinea-Bissau joined the Union, becoming its eighth
member.
Among WAEMU’s objectives were to strengthen the economic and
financial competitiveness of member countries through an open and
competitive market and facilitate the free movement of people, goods,
services and capital and the establishment of a Common External Tariff
(CET) and a common trade policy through the creation of a common
market. Another was to harmonize the laws of member states, and
especially the tax system, to enable the common market to function
smoothly.
Reiterating their intention to create a common currency within
ECOWAS in Lomé in 1999, the leaders of member countries adopted
a declaration stipulating the establishment of a sub-regional monetary
Ibrahima Camara 155
union by 2004. Monetary union was supposed to be preceded by setting

up a second monetary zone, WAMZ. The initial date for establishing this
second zone was 2003. However, owing to non-fulfillment of the con-
vergence criteria, it was postponed until1 July 2005. On 6 May 2005
in Banjul, the heads of state and government decided to further post-
pone the creation of WAMZ until 1 December 2009. This date was again
shifted to 2015 to finally enable introduction of the ECOWAS common
currency in 2020.
To date, WAEMU remains the only effective monetary union within
ECOWAS.
In Africa, monetary unions are established to strengthen regional sol-
idarity and foster intra-regional trade (Masson and Patillo, 2004). So a
close examination of the phenomenon is pertinent. While the objective
of monetary unions seems clear, the relevant question is: What is the
impact of monetary unions on bilateral trade?
This question has inspired many empirical works on the case of
WAEMU (Agbodji, 2007; Diop, 2007). Most of these works assessed the
overall impact of the Union during their entire study period. But this
study seeks to assess the impact of the monetary union in force in
WAEMU member countries on trade flows. Specifically, it addresses the
following questions:
• What are the effects of the monetary union in force in WAEMU

member countries on intra-zone trade?
• What impact does it have on trade between member countries and
the rest of ECOWAS? In other words, are trade flows being diverted
to member countries?
This study is unique in assessing the year-on-year impact of WAEMU

on trade flows, and thus helps to monitor impact before and after the
establishment of the Union.
The study is divided into four sections: The first is devoted to litera-
ture review, which highlights theoretical aspects and empirical studies
of monetary unions. The second analyzes trade flows of WAEMU mem-
ber countries, in particular the volume of intra-regional trade. The
third presents the gravity model by describing it theoretically, and then
describes the empirical model used. The fourth deals with WAEMU’s
isolated impact on bilateral trade, including data and estimation results.
1 Literature review
In recent years, there has been growing interest in monetary unions, and
especially their impact on trade. Although currency is often considered
the symbol of a nation’s sovereignty, some nations have still joined

monetary unions. The idea of monetary unions has inspired both
theoretical and empirical studies.
Theoretical review
Several theories have been developed about monetary unions, the most
known being the Optimum Currency Area (OCA). An OCA refers to a
grouping of economic spaces that is subject to monetary sovereignty
and a single currency system, or characterized by the existence of a fixed
and irrevocable parity mechanism between their various currencies.
Exchange rates are fixed within the area, and left floating outside
the area.
The first theoretical OCA analyses were carried out by Mundell (1961)
who argues there are benefits for regions that use a common currency.
Monetary unions facilitate international trade, as a single means of
trade reduces the cost of transactions. Yet he notes that a common cur-
rency may be the source of problems in cases of asymmetric price and
salary shocks, and minimal rigidities. Mundell concludes the optimum
currency area is one within which the labor factor is mobile.
Highlighting the size of economies, McKinnon (1963) asserts that
small economies are best suited to monetary unions because they tend
to be much more open, and consequently, less subject to nominal
rigidities. When the share of international trade in economic activity
increases, the opportunity cost of exchange rate rigidity drops and its
benefit increases. He also argues that when a small country trades with
a big partner, it is unable to conduct a completely monotonic monetary
policy. So it is in the interest of small open countries to form monetary
unions which will ensure the stability of their trade without jeopar-
dizing the independence of their already limited monetary policy. For
his part, Kenen (1969) argues that broader diversification of economies
leads to less asymmetric shocks. So countries with broadly diversified
economies stand to gain nothing by conducting a national monetary
policy.
A currency facilitates trade, as a unit of account and means of trade.
Inasmuch as a monetary union may influence the reduction of exchange
rate and transaction cost uncertainties, it may reinforce trade flows
within the currency area. For many authors, the theoretical effects of
international trade are well identified and less controversial; but under-
standing the impacts of monetary unions on trade from the theoretical
viewpoint seems a complicated task. Still, this may be more feasible from
an empirical standpoint.
Ibrahima Camara 157
Empirical review
The impact of monetary unions on trade has been the topic of many
empirical studies in recent decades. It is generally admitted that a com-
mon currency strengthens trade, but the first empirical works were more
interested in the impact of exchange rate volatility on trade flows, with-
out arriving at conclusive findings (Lochard, 2005). The first and most
famous empirical study on the impact of monetary unions on trade
flows was conducted by Rose (2000).
Rose (2000) used a bilateral trade gravity model into which he intro-
duced the monetary union represented by a binary variable. The variable
was given the value 1 or 0 for two countries, depending on whether
they belonged to the same monetary union or not. By introducing
other control variables such as sharing a common border, a common
language, colonial links or belonging to the same free trade zone, he
found that two countries sharing a common currency carry out three
times more trade than similar countries with different currencies. Rose’s
(2000) paper attracted much criticism concerning the magnitude of the
impact of the common currency. Most criticism dealt with the existence
of a bias in the results, particularly a self-selection bias (Persson, 2001;
Tenreyro, 2001), an aggregation bias (Nitsch, 2004) and an omitted vari-
ables bias due to a correlation between the independent variables and
the error term.
According to Persson (2001), the results of Rose’s (2000) paper are
biased due to the existence of self-selection.3 But there could be a
correlation between the monetary union and trade determinants. For
example, small or poor countries may trade more between themselves
not because they have a common currency, but because they share
common geographical and cultural features. The existence of such a cor-
relation does not permit the comparison of countries having a common
currency with the sample countries. Therefore, it cannot be concluded
that monetary unions increase trade, everything being equal.
To remedy self-selection bias, Persson (2001) used a matching method.
First, he looked for the determinants of monetary unions and calculated
a monetary union propensity score for each pair of countries.4 Then, for
each of the pairs, he constituted a control group comprising countries
with their own currencies, but which have the features of the monetary
union pairs of countries according to the propensity score. Last, he esti-
mated the effects of the monetary union on trade using only pairs of
monetary union countries and their control groups. By so doing, he
found the impact of monetary unions on trade is low, ranging from 13%
to 65%. Tenreyro (2001) also tried to correct the self-selection bias in
the method used by Rose (2000). She first estimated the probability that
two countries may adopt a common currency and checked the impacts
of monetary unions on trade after explaining self-selection. She found
that monetary unions increased trade by 60%, though such impact was
not significantly different from zero.
According to Micco et al. (2002), one of the problems in the study
conducted by Rose (2000) is that most of the sample monetary unions
are made up of very small or poor countries sharing a common currency
(CFAF zone countries, for example) or of very small or poor countries
which have adopted the currency of big countries (like Réunion and
France). These findings can hardly be extended to bigger countries.
Lochard (2005) argues that introducing dummy variables which help to
isolate each type of monetary union may reduce the aggregation bias.
Using this method, she obtained a 238% trade increase. She concluded
the high impact obtained by Rose was due to the weights of unilateral
unions in her sample.
In short, most of the studies resulting from the findings of Rose (2000)
adopted other methodologies which revealed lesser impacts. Still, they
are unanimous on the fact that monetary unions increase trade.
Sub-Saharan Africa has been the subject of very many studies, though
most of them focusing on regional trade agreements rather than mone-
tary unions. Carrere (2004) assessed the impact of regional agreements
on trade in sub-Saharan Africa over the period 1962–1996 using a panel
gravity model. The findings showed the implementation of African
regional trade agreements led to a significant increase in trade between
members. She also found the monetary unions in force in the CFAF zone
strongly enhanced this positive impact.
Avom and Gbetnkom (2005) looked for intra-WAEMU trade deter-
minants using a gravity model and laying emphasis on economic
reforms undertaken from 1980–1990. Splitting their study period into
two, 1990–1994 and 1996–2000, they found regional integration sub-
stantially increases trade between the Union’s member states. Agbodji
(2007) attempted to isolate the impacts of monetary and economic
unions on intra-WAEMU trade. Using an augmented dynamic gravity
model, he showed that WAEMU membership and the implementation
of economic reforms produced significant effects in terms of diversion
of imports and exports; although economic distortions, considered to
encourage illegal trade, significantly reduced bilateral trade within the
Union.
By controlling the impact of geographical and structural factors, Diop
(2007) showed how the existence of a monetary union affects the level
of intra-community trade in West Africa. Using a gravity model with
Ibrahima Camara 159
data on bilateral trade between ECOWAS member states over the period
1997–2004, he found that geographical and structural factors as well
as WAEMU membership determine the intensity of bilateral trade flows
in West Africa. Further, use of a common currency leads to considerable
trade development. He concluded that, all things being equal, extending
WAEMU to the other ECOWAS countries should considerably increase
the volume of intra-regional trade. However, the study does not take
into account the period preceding the establishment of WAEMU to allow
for comparison.
2 Analysis of trade within WAEMU
WAEMU is a large regional entity. It covers a total surface area of

3,509,600 km2 , that is about 68.8% of the total surface area of ECOWAS,
and accounts for 33% of the region’s Gross Domestic Product (GDP).
In 2010, it had a combined overall nominal GDP of about USD 70 bil-
lion, that is 23% of the GDP of ECOWAS. Côte d’Ivoire and Senegal
respectively account for 32% and 18% of the Union’s overall GDP. Togo
and Guinea-Bissau were the Union’s smallest economies in 2010 with
respectively 6% and 1% of the overall nominal GDP. WAEMU’s trade
structure is highly dominated by Ivorian and Senegalese trade. Accord-
ing to BCEAO statistics, total WAEMU zone exports over the last decade
stood at CFAF 60.388 billion. Côte d’Ivoire’s exports represented 61%
of total exports. Those of Senegal and Mali represented 13% and 11%
respectively.
Intra-WAEMU exports were low. In fact, they represented 13% of the
Union’s total exports during the last decade while its exports to France
alone represented 14% of its total exports. While intra-WAEMU exports
were above exports to the rest of ECOWAS, imports from ECOWAS, for
their part, were above intra-WAEMU imports. These intra-zone imports
and imports from the rest of ECOWAS respectively represented CFAF
8.299 billion and CFAF 9.488 billion over the last decade. Nigeria was
the main country of origin of imports from the rest of ECOWAS with
88% of the Union’s imports from the zone. Guinea-Bissau was the
weakest WAEMU economy, and its foreign trade with the rest of the
WAEMU zone was also very low. In fact, its intra-zone imports (exports)
represented 1.8% (0.3%) of total intra-zone imports (exports).
Within the entire Union, all the trade balances showed a deficit
during the last decade, with the exception of that of Côte d’Ivoire.
On the whole, WAEMU’s imports stood at CFAF 76.870 billion com-
pared to CFAF 60.388 billion for exports, showing a trade deficit of CFAF
16.482 billion (Figure 8.1).
15,000,000
10,000,000
5,000,000
–5,000,000
–10,000,000
–15,000,000
–20,000,000
in
re
er
al
al
ga
s
sa
t
en
oi
ig
M
Fa
To
To
ne
is
Iv
N
B
B
na
Se
d'
a-
ki
ne
ôt
ur
ui
B
Figure 8.1 Trade balance 2000–2009 for WAEMU’s countries (CFAF millions)
Source: Central Bank of West African States (CBWAS), author’s calculations.
Gravity model
During the last decades, many authors have used the gravity model to
assess the impact of customs or monetary unions on bilateral trade.
Tinbergen (1962) used the model to explain the intensity of migration
flows based on the size of regions and distance between them. Since
Rose (2000), the gravity model has been increasingly used to assess the
impact of monetary unions on trade.
Theoretical model
Empirical studies on international trade increasingly rely on gravity
models. Inspired by the Newtonian theory, this type of model expresses
trade flows between two countries as proportional to their economic
weight and inversely proportional to the geographical distance between
them. In its simplest form, the model is presented as follows:
β β
Mij = A × (Y
COM Yi Yj )j 1 × Dij2 (1)
Here COM Mij , represents the value of trade flows between country i and
country j, Y, their national income, Dij the distance between these
countries and A, a proportionality coefficient. In addition to these two
traditional variables (GDP and distance), the model has been augmented
by other variables to take into account some specificities of bilateral
Ibrahima Camara 161
relations. They include sharing a common land border, the fact that the
country is an island (landlocked) or not, having the same former colo-
nial rulers and GDP per capita. These variables are often called control
variables.
To isolate the effects of monetary unions on trade since Rose (2000),
some authors introduce a dummy variable with the value 1 for two
member countries of a monetary union and the value 0 if not. In addi-
tion to the dummy variable, control variables are represented by a vector
Xij . So, the augmented model becomes:
β β β
Mij = A × (Y
COM Yi Yj )j 1 × Dij2 × Xij2 (2)
In its logarithmic form, the equation (3) becomes:
Mij ) = β0 + β1 × ln(Y
ln(COM Yi Yj ) + β2 × ln(D
( ij ) + β3 × ln(X
( ij ) (3)
Here β0 , β1 , β2 and β3 are the final coefficients; β1 is considered negative

while β2 is considered positive. β3 is the vector of coefficients of vector
Xij of the control variables. Equation (3) is referred to as an augmented
gravity model equation.
Empirical specification
Many empirical gravity model equations were derived from the theo-
retical model equation (3). Most of the empirical studies were inspired
by Rose (2000), who used the augmented gravity model to estimate the
effects of monetary unions and exchange rate volatility on trade. He
added many variables in this augmented model which take into account
monetary aspects, among others. The model used is presented as follows:
( ijt ) = β0 + β1 ln(Y
ln(X Yi Yj )t + β2 ln(yi yj )t + β3 ln(D
( ij ) + β4 Conttij + β5 Langgij
+ β6 FTAijt + β7 ComNattij + β8 ComColij + β9 Colonyij + γ CU
Uijt
+ δV(eij )t + εijt (4)
Here i and j represent the two countries, t expresses time and other
variables are defined as:
• Xij is the value of bilateral trade between countries i and j;

• Y is real GDP;
• y is GDP per capita;
• Dij is the distance between countries i and j;
• Conttij is a dummy variable with value 1 where i and j share a common

border and 0 if not;
• Langgij is a dummy variable with value 1 where i and j have the same
official language;
• FTAij is a dummy variable with value 1 where i and j are party to the
same regional trade agreement;
• ComNattij is a dummy variable with value 1 where i and j belong to
the same nation;
• ComColij is a dummy variable with value 1 where i and j were
colonized after 1945 by the same colonial power;
• Colonyij is a dummy variable with value 1 where i colonized j and
vice-versa;
• CUUijt is a dummy variable with value 1 where i and j share a common
currency during time t;
• V(eij )t is bilateral nominal exchange rate volatility (between i and j)
during the time prior to t.
In this study, the interest coefficients are expressed as γ , which repre-

sents the impact of monetary unions on trade flows and δ, which is a
measurement of the impact of exchange rate volatility on bilateral trade.
Masson and Patillo (2005) used a gravity model to assess the impact
of CMA on bilateral trade between South Africa and each other coun-
try in the zone. In this model, they included as independent variables
the product of real GDPs of the two countries in terms of level and per
capita, the distance between them and the product of their surface areas.
In addition, the following dummy variables were included: membership
of a free trade area or a monetary union, a common language and shar-
ing of a common land border or former colonial ruler. Consequently,
the gravity equation which they estimated is as follows:

n
( ij ) = β0 + β1 × ln(Y
ln(X Yi Yj ) + β2 × ln(yi yj ) + β3 × ln(Supi Sup
pj ) + β4+k × Dk
k=1
(5)
Here X is bilateral trade, Y real GDP, y GDP per capita, Sup surface
area and D the vector of dummy variables. For the purposes of proper
presentation, the time index t was omitted.
While most of the empirical studies that used the gravity model
focus on the overall impact of monetary or customs unions for the
entire study period, Micco et al. (2002) assessed year-on-year impact of
the European Monetary Union on trade. To that end, they conducted
a cross-analysis by estimating the year-on-year gravity equation using
ordinary least squares. The model estimated year-on-year is as follows:
Ibrahima Camara 163
( ij ) = β0 + β1 ln(Y
ln(X Yi Yj ) + β2 ln(Y
Yi Yj ) + β3 ln(D
+ β6 FTAijt + β7 LandLockij + β8 Islandij + β9 Surffij
+ β10 EMU
Uij + εij (6)
Here LandLock, Island and Surf represent respectively the number of

landlocked countries within the pairs of countries, the number of
islands in the pairs of countries and the product of surface areas. The
EMU variable has the value 1 where the pair of countries belongs to
the European Monetary Union (EMU), even for the period preceding the
establishment of EMU, and 0 if not. The other variables are described as
above. The authors sought to monitor the value of the coefficient of
this variable over time. Though EMU had an impact on bilateral trade,
its coefficient increased after its establishment.
In addition to this cross-analysis, these authors used two other panel
models with the following specifications. First of all, they estimated the
equation of a fixed-impact panel model:
( ijt ) = βij + β1 ln(Y

ln(X Yit Yjt ) + β2 ln(yit yjt ) + β3 FTAijt + β4 EU
Uijt
+ β5 FEMU
Uijt + εijt (7)
Here the variables, save for FEMU and EU, are described as above. FEMU Uijt
is a dummy variable with value 1 where countries i and j formally belong
to the European Monetary Union during year t and EU Uijt is a dummy
variable equal to 1 where the pair of countries belongs to the EU during
year t. In this equation, however, the other trade determinants – like
the colonization, border or common language variables – are not direct
interest values and do not vary, or vary much, over time. In fact, most of
these determinants have already been taken into account in the bilateral
specific impact (βij ) and might lead to a bias if they are not well specified
(Pakko and Wall, 2001).5 Lochard (2005) also used a fixed-impact panel
model which did not include these dummy variables.
To check the robustness of the findings of the estimation of equa-
tion (7), they estimated another panel model which, this time, included
control variables.
( ijt ) = β0 + β1 ln(Y
ln(X Yit Yjt ) + β2 ln(yit yjt ) + β3 ln(D
+ β6 FTAijt + β7 LandLockij + β8 Islandij + β9 Surffij + β10 FEMU
Uijt
+ β10 EU
Uijt + εijt (8)
The particularity of the models estimated by Micco et al. (2002) is that

they take into account monetary union trends and their impact on
bilateral trade over the years. Before joining a monetary union, coun-
tries first meet a number of criteria some of which help to strengthen
trade between them. Besides, countries do not all join a monetary union
at the same time.
For the purposes of this study, a year-on-year cross-analysis is con-
ducted to better capture and monitor WAEMU’s impact on trade flows
over time. To that end, the equation to be estimated will be similar
to equation (6). The choice of this model seems appropriate given the
events within the WAEMU area, which experienced several changes dur-
ing the last decades. The early 1990s was marked by the implementation
of Structural Adjustment Programmes (SAPs) by countries of the sub-
region, which probably modified the structure of their trade. Then there
was the devaluation of the CFAF in 1994, Guinea-Bissau’s late join-
ing of the Union in 1994 and the coup d’état and rebel uprising in
Côte d’Ivoire6 in 1999 and 2002, respectively. All these events probably
modified the intra-WAEMU trade structure. The model to be estimated
therefore is:
( ij ) = β0 + β1 ln(Y
ln(X Yi Yj ) + β2 ln(Dist
( tij ) + β3 Borderrij + β4 Langgij + β5 Colonyij
+ β6 LandLockij + β7 ln(Supi Sup
pj ) + β8 WAEMU
Uij
+ β9 WAEMU1ij + εij (9)
Here the variables are defined as follows: Xij , is the export flows of coun-
try i to country j; Supi and Sup pj are respectively the surface areas of
countries i and j; Colonyij , is a dummy variable equal to 1 where the two
countries had the same colonial ruler and 0 if not; Border, is equal to 1
where the two countries share a common border and 0 if not, WAEMU,
is a dummy variable equal to 1 where the two countries belong to the
WAEMU zone and 0 if not7 ; WAEMU1, is a dummy variable with value
1 where the partner country is a WAEMU member state and 0 if not.
The coefficient of the WAEMU1 variable helps to capture effects on the
diversion of trade flows while the WAEMU variable helps to identify the
trade surplus created by the establishment of the Union. The expected
signs are therefore positive for the WAEMU variable and negative for the
WAEMU1 variable.
Owing to the existence of zero export values for pairs of countries
at a given date, there is a modeling problem as exports are recorded
in logarithm. To solve this problem, we proceed like Bangake and
Eggoh (2008) did. Inspired by Rose (2000), they added the value 100
to each bilateral trade value. In this study, therefore, the export value
Ibrahima Camara 165
considered has been increased by 100. Consequently, the dependent

variable is ln(X
( ijt + 100), Xijt

being the initial value of exports taken from
the International Monetary Fund’s ‘Data on Trade Statistics (DOTS)’.
3 WAEMU’s impact on trade
This section presents data and empirical findings as well as related com-
mentaries. The study sample comprises all ECOWAS member countries
and covers the period 1990–2005. Data on exports (in 1987 constant
USD million) is taken from DOTS. GDP is provided by the World Bank’s
‘World Development Indicators & Global Development Finance, 2012’.
It is evaluated in 1987 constant USD million. Data on distance is culled
from the Centre d’Etudes Prospectives et d’Informations Internationales
(CEPII).
Empirical findings
Ordinary least squares were used for the year-on-year estimation by
equation (9). The results of the estimation are presented in Figure 8.2.
Only the coefficients of the WAEMU and WAEMU1 variables for each of
the cross-cutting regressions have been entered, and the overall findings
are presented in Table 8.1.
The GDPs of exporting and importing countries, their surface areas,
the distance between them and the number of landlocked countries in
the pairs of countries are significant and show expected signs for almost
all the years. As such, large countries with high GDPs tend to carry out
much more trade while the more they are landlocked and distant from
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%
90
91
92
93
94
95
96
97
98
99
00
01
02
03
04
05
19
19
19
19
19
19
19
19
19
19
20
20
20
20
20
20
Figure 8.2 WAEMU’s marginal effects on exports

Source: Authors’ estimates based on DOTS.
Table 8.1 Coefficients of WAEMU and WAEMU1 variables, from cross-section

analysis
Year WAEMU WAEMU1
Coef. T-stat Coef. T-stat
1990 0.2494 (2.46)∗∗ 0.0086 0.110

1991 0.2431 (2.28)∗∗ −0.0135 −0.170
1992 0.2342 (2.57)∗∗ −0.0364 −0.520
1993 0.2209 (2.44)∗∗ −0.0180 −0.260
1994 0.2066 (2.29)∗∗ −0.0072 −0.100
1995 0.2138 (2.44)∗∗ −0.0344 −0.510
1996 0.1992 (2.36)∗∗ −0.0668 −1.030
1997 0.2358 (2.87)∗∗∗ −0.0754 −1.190
1998 0.2471 (3.05)∗∗∗ −0.1030 −1.650
1999 0.1972 (2.46)∗∗ −0.0354 −0.570
2000 0.1660 (1.95)∗ −0.0729 −1.120
2001 0.2616 (2.72)∗∗∗ −0.0767 −1.050
2002 0.2225 (2.53)∗∗ −0.0372 −0.550
2003 0.1807 (2.07)∗∗ −0.0547 −0.810
2004 0.2452 (2.77)∗∗∗ −0.1067 −1.560
2005 0.1615 (1.82)∗ −0.0648 −0.940
Note: ∗∗∗ 1% significant; ∗∗ 5% significant; ∗ 10% significant.

Source: Authors’ estimates based on DOTS.
each other the less trade they tend to conduct. The WAEMU1 variable,
for its part, is not significant for each of the study years, but has the
expected negative sign (with the exception of 1990). It is therefore diffi-
cult to conclude that there is an effect of diversion in trade flows to the
Union each year.
The WAEMU variable coefficient is significant and positive for each
of the years, meaning WAEMU has a significant and positive impact on
export flows. However, there seems to be various phases in the Union’s
marginal effects trend on exports (Figure 8.2): the first phase from 1990
to 1996 was characterized by a reduction in coefficients with a peak in
1995, the second phase covering 1997 and 1998 was characterized by
growth and the third fluctuation phase extended from 1999 to 2005.
On the whole, these effects recorded a decline.
In 1990, intra-WAEMU trade was 28.3% (exponential [0.2494]−1)
higher than trade between countries with different currencies.8 This
marginal effect stood at 23% in 1994 and 22% in 1996. The decline
in this effect can be explained, among other things, by the lack of trade
and production policy harmonization and coordination. It can also be
explained by poor infrastructure and similarity in production structures,
most of which are based on agro-industry. These findings show that the
Ibrahima Camara 167
40.0%
30.0%
20.0%
10.0%
0.0%
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Toward Africa WAEMU
Figure 8.3 Share of exports within WAEMU and to Africa

Source: Central Bank of West African States (CBWAS), author’s calculations.
countries of the Union were already trading between themselves well

before its establishment.
Paradoxically, WAEMU’s impact on exports continued to drop follow-
ing its establishment in 1994. This could be due to the adoption, in
the same year, of devaluation by member countries. According to De
Monchy et al. (1995), devaluation benefited from a favorable interna-
tional environment with renewed demand by industrialized countries
and the increase in the price of many commodities. This is believed
to have led to the diversion of exports from WAEMU countries to
the rest of Africa and European Union countries. Also, according to
Tandian (1998), four Union countries, including Côte d’Ivoire, reduced
their share of exports to countries in the zone following devaluation.
In 1990, the share of intra-regional exports stood at 17.1%. This percent-
age was never attained again right up to 2009 (Figure 8.3). Moreover, the
Union had a growing impact on internal trade in 1996 and 1998, prob-
ably thanks to the progressive elimination of customs barriers initiated
within the Union in 1996.
Since 1999, WAEMU’s impact seems to be correlated to the vari-
ous crises in Côte d’Ivoire. After registering growth in 1997 and 1998,
the Union’s impact on internal trade dropped in 1999 and 2000, the
period of military transition in Côte d’Ivoire. After a sudden rise in 2001
following a fragile political stability, the Union’s impact on the zone’s
trade plunged again in 2002 when Côte d’Ivoire once more experienced
a rebel uprising.
In a nutshell, WAEMU’s impact on intra-zone trade before its creation

was on average higher than its impact after its establishment. From 1990
to 1993, its marginal effect was 26.7% on average and from 1994 to
2005, it was 23.6%.
4 Conclusion
This study seeks to determine WAEMU’s year-on-year impact on trade.

Specifically, it assesses the Union’s effects on intra-zone trade and its
effects on trade between Union member countries and with the rest of
ECOWAS. Using cross-analysis, the ordinary least squares method was
applied for each year during 1990–2005. This analysis allows us to moni-
tor the trend in WAEMU’s impact before and after its establishment. Esti-
mations show its impact on intra-zone trade is significant and positive
for all the study years, but this impact tends to decrease with time. From
23% in 1994, it dropped to 17.5% in 2005, showing a 23.7% decrease.
Regarding the diversion of trade within ECOWAS to Union mem-
ber countries, it is difficult to draw a conclusion given the insignif-
icance of the coefficients of the trade diversion variable. However,
there is diversion of trade to Union member countries to go by the
coefficient sign.
Therefore, to increase WAEMU’s impact on intra-zone trade, member
countries must:
• continue efforts at implementing the common trade policy initiated

within the Union in 1996;
• allow products to access the various markets of the zone by further
reducing tariff and non-tariff barriers through the liberalization of
intra-regional trade;
• promote the specialization of products within the Union taking
comparative advantages into account. This will create more trade
opportunities between member countries;
• diversify production and export structures by not limiting produc-
tion to traditional goods (agro-food) and promote complementarity
between the economies;
• develop an industrialization policy based on import substitution
through the local processing of raw materials;
• develop trade-related infrastructure to facilitate the movement of
goods and services within the Union; and
• create and preserve trade-friendly political stability and a climate of
good governance.
Annexes
Table 8.A.1 Coefficients from cross-analyzes, 1990–2005
Coefficient LPIB LSUP Ldistance Landlock WAEMU WAEMU1 Border Comland Comcol Constance
T-Stat
1990 0.177 −0.018 −0.153 −0.091 0.249 0.009 0.123 0.269 −0.179 −3.997
(5.82)∗∗∗ (−0.6) (−2.95)∗∗∗ (−1.27) (2.46)∗∗ (0.11) (1.36) (1.94)∗ (−1.15) (−5.2)∗∗∗
0.110 0.037 −0.078 −0.210 0.243 −0.013 0.236 0.074 0.050 −2.958
1991 (3.37)∗∗∗ (1.31) (−1.48) (−2.81)∗∗∗ (2.28)∗∗ (−0.17) (2.46)∗∗ (0.47) (0.3) (−3.4)∗∗∗
1992 0.134 0.032 −0.095 −0.156 0.234 −0.036 0.167 0.202 −0.115 −3.630
(5.28)∗∗∗ (1.28) (−1.72)∗ (−2.08)∗∗ (2.57)∗∗ (−0.52) (1.85)∗ (1.26) (−0.66) (−4.77)∗∗∗
1993 0.105 0.063 −0.075 −0.193 0.221 −0.018 0.148 0.219 −0.079 −3.293
(4.33)∗∗∗ (2.76)∗∗∗ (−1.36) (−2.62)∗∗∗ (2.44)∗∗ (−0.26) (1.63) (1.36) (−0.45) (−4.22)∗∗∗
1994 0.112 0.060 −0.115 −0.175 0.207 −0.007 0.086 0.187 −0.090 −3.113
(4.59)∗∗∗ (2.63)∗∗∗ (−2.12)∗∗ (−2.4)∗∗ (2.29)∗∗ (−0.1) (0.97) (1.17) (−0.52) (−4.04)∗∗∗
1995 0.126 0.046 −0.135 −0.126 0.214 −0.034 0.062 0.229 −0.151 −3.278
(5.55)∗∗∗ (2.1)∗∗ (−2.54)∗∗ (−1.79)∗ (2.44)∗∗ (−0.51) (0.71) (1.47) (−0.89) (−4.5)∗∗∗
1996 0.126 0.050 −0.160 −0.104 0.199 −0.067 0.084 0.243 −0.119 −3.240
(5.77)∗∗∗ (2.34)∗∗ (−3.13)∗∗∗ (−1.51) (2.36)∗∗ (−1.03) (1) (1.63) (−0.73) (−4.66)∗∗∗
1997 0.143 0.049 −0.133 −0.106 0.236 −0.075 0.051 0.179 −0.085 −4.073
(6.07)∗∗∗ (2.22)∗∗ (−2.67)∗∗∗ (−1.54) (2.87)∗∗∗ (−1.19) (0.62) (1.25) (−0.55) (−5.73)∗∗∗
1998 0.153 0.042 −0.133 −0.097 0.247 −0.103 0.061 0.151 −0.099 −4.290
(6.79)∗∗∗ (1.98)∗∗ (−2.7)∗∗∗ (−1.46) (3.05)∗∗∗ (−1.65) (0.75) (1.08) (−0.65) (−6.25)∗∗∗
1999 0.155 0.058 −0.194 −0.152 0.197 −0.035 −0.025 0.181 −0.090 −4.345
(6.89)∗∗∗ (2.72)∗∗∗ (−3.97)∗∗∗ (−2.32)∗∗ (2.46)∗∗ (−0.57) (−0.31) (1.31) (−0.61) (−6.36)∗∗∗
2000 0.127 0.069 −0.183 −0.129 0.166 −0.073 −0.026 0.074 0.045 −3.427
(5.2)∗∗∗ (2.97)∗∗∗ (−3.53)∗∗∗ (−1.84)∗ (1.95)∗ (−1.12) (−0.31) (0.51) (0.29) (−4.78)∗∗∗
2001 0.122 0.046 −0.186 −0.121 0.262 −0.077 0.099 −0.025 0.114 −2.739
(4.32)∗∗∗ (1.82)∗ (−3.64)∗∗∗ (−1.69)∗ (2.72)∗∗∗ (−1.05) (1.19) (−0.14) (0.59) (−3.36)∗∗∗
169
170
Table 8.A.1 (Continued)
Coefficient LPIB LSUP Ldistance Landlock WAEMU WAEMU1 Border Comland Comcol Constance
T-Stat
2002 0.136 0.040 −0.195 −0.147 0.222 −0.037 0.084 −0.011 0.111 −3.153
(5.54)∗∗∗ (1.65)∗ (−3.6)∗∗∗ (−2.05)∗∗ (2.53)∗∗ (−0.55) (0.96) (−0.07) (0.69) (−4.43)∗∗∗
2003 0.107 0.062 −0.153 −0.156 0.181 −0.055 0.081 0.016 0.138 −2.739
(4.37)∗∗∗ (2.56)∗∗ (−2.82)∗∗∗ (−2.18)∗∗ (2.07)∗∗ (−0.81) (0.93) (0.11) (0.84) (−3.81)∗∗∗
2004 0.098 0.079 −0.155 −0.179 0.245 −0.107 0.088 −0.031 0.133 −2.722
(4.09)∗∗∗ (3.27)∗∗∗ (−2.84)∗∗∗ (−2.48)∗∗ (2.77)∗∗∗ (−1.56) (1) (−0.2) (0.81) (−3.82)∗∗∗
2005 0.104 0.075 −0.161 −0.174 0.161 −0.065 0.098 −0.008 0.167 −2.846
(4.46)∗∗∗ (3.09)∗∗∗ (−2.94)∗∗∗ (−2.41)∗∗ (1.82)∗ (−0.94) (1.11) (−0.05) (1.02) (−4.07) ∗∗∗
Notes: This table shows the results of the estimations based on a sample of ECOWAS countries; 1990–2005 Period. Dependent Variable: Logarithm of
Exports (LExport).
T-stat are presented in brackets. ∗∗∗ 1% significant; ∗∗ 5% significant; ∗ 10% significant.
Source: Author’s estimates.
Ibrahima Camara 171
Notes
1. These regional economic communities are: the Arab Monetary Union (UMA),
the Economic Community of West African States (ECOWAS), the Common
Market for Eastern and Southern Africa (COMESA), the Economic Commu-
nity of Central African States (ECCAS) and the Southern African Development
Community (SADC).
2. This date was set in 1987 during the establishment of the Monetary Coopera-
tion Programme (MCP).
3. The probability of two countries adopting the same currency is non-random
and could depend on some independent variables.
4. This score evaluates the probability of belonging to monetary unions accord-
ing to other trade determinants.
5. Quoted by Lochard (2005).
6. The specific case of Côte d’Ivoire was considered owing to its economic and
trade impact within the Union.
7. The WAEMU variable had the value 1 for two Union member countries even
before the creation of the Union, the goal of the model being to monitor the
coefficient of this variable over time.
8. WAEMU’s effects on trade were calculated taking into account the exponential
of coefficients as expressed in logarithm in the model.
References
Agbodji, A. E. (2007), ‘Intégration et échanges commerciaux intra sous-régionaux:
Le Cas de l’UEMOA’, Revue Africaine de l’Intégration, vol. 1, no. 1, pp. 161–188.
Avom, D. and Gbetnkom, D. (2005), ‘Intégration par le marché: le cas de
l’UEMOA’, Région et Développement no. 22–2005.
Bangake, C. and Eggoh, J. (2008), ‘The Impact of Currency Unions on Trade:
Lessons from CFA Franc Zone and Implications for Proposed African Mone-
tary Unions’, Document de Recherche no. 2008–29, Laboratoire d’Economie
d’Orléans.
Carrere, C. (2004), ‘African Regional Agreements: Impact on Trade with or with-
out Currency Unions’, Journal of African Economies, vol. 13, no. 2, pp. 199–239.
De Monchy, G., Nancy, G., Aloy, M. and De Boisdeffre, L. (1995), ‘Effets et
perspectives de la dévaluation au Sénégal’, DIAL 1995-01/T.
Diop, C. A. (2007), ‘L’UEMOA et la perspective d’une zone monétaire unique de
la CEDEAO: les enseignements d’un modèle de gravité’, Document d’Etude et
de Recherche no DER/07/01 – BCEAO.
Kenen, P. (1969), ‘The Theory of Optimum Currency Areas: An Eclectic View’,
in Mundell and Swoboda (eds.), Monetary Problems of the International Economy,
Chicago: University Press, pp. 41–60.
Lochard, J. (2005), ‘Mesurer l’influence des unions monétaires sur le commerce’,
Economie internationale, vol. 2005/3, no. 103, pp. 5–24.
Masson, P. and Patillo, C. (2004), ‘A Single Currency for Africa?’, Finance and
Development,t vol. 41, no. 4, pp. 9–15.
Masson, P. and Patillo, C. (2005), The Monetary Geography of Africa, Washington,
DC: Brookings Institution Press.
McKinnon, R. I. (1963), ‘Optimum Currency Areas’, American Economic Review,

vol. 53, pp. 717–724.
Micco, A., Stein, E. and Ordoñez, G. (2002), The Currency Union Effect on
Trade: Early Evidence from the European Union, Washington, DC: Inter-American
Development Bank.
Mundell, R. A. (1961), ‘A Theory of Optimum Currency Areas’, American Economic
Review, vol. 51, pp. 657–665.
Nitsch, V. (2004) ‘Have a break, have a . . . national currency: when do monetary
unions fall apart?’, CESifo Working Paper Series 1113, Munich: CESifo Group.
Pakko, M. and Wall, H. (2001), ‘Reconsidering the Trade-creating Effects of a Cur-
rency Union’, Federal Reserve Bank of St. Louis Review, vol. 83, no. 5, pp. 37–45,
September–October.
Persson, T. (2001), ‘Currency Union and Trade: How Large Is the Treatment Effect’
Economic Policy: A European Forum, vol. 0, no. 33, pp. 433–448.
Rose, A. K. (2000) ‘One Money, One Market: Estimating the Effect of Common
Currencies on Trade’, Economic Policy, vol. 30, pp. 7–46.
Tandian, D. (1998), ‘Les suites de la dévaluation du franc CFA de janvier 1994’,
L’Actualité économique, vol. 74, no. 3, pp. 561–581.
Tenreyro, S. (2001), On the Causes and Consequences of Currency Union, Harvard
University. Photocopy.
Tinbergen, J. (1962), Shaping the World Economy: Suggestions for an International
Economic Policy, New York: The Twentieth Century Fund.
9
Trade Agreements and Flows in
ECOWAS: Is a Single Currency the
Determining Factor?
Benjamin Ndong and Sokhana Diarra Mboup
Introduction
Regional integration agreements are a main feature of world trade today.

The number of such agreements has multiplied in recent years, as the
World Trade Organization (WTO) notes:
Most WTO members are also party to one or more regional trade
agreements. Since the early 1990s, the sharp increase in the num-
ber of Regional Trade Agreements has continued. Some 250 Regional
Trade Agreements (RTAs) have been notified to the GATT/WTO up to
December 2002, of which 130 were notified after January 1995. Over
170 RTAs are currently in force; an additional 70 are estimated to be
operational although not yet notified.
(WTO, 2003)
RTAs are concluded by groups of countries within a region engaged

in a wider integration process. Besides RTAs, Africa maintains trading
relationships with the rest of the world based on bilateral and multi-
lateral agreements. Because Africa’s 52 or so countries are mostly small
economies, intra-African trade remains weak. The markets accessible to
African firms are often narrow and fragmented, because of poor commu-
nication channels and customs barriers. Yet African countries are also
trying to break with inherited trading practices that once gave prefer-
ence to trade with former colonial rulers over trade with each other.
Thus, African states have signed regional integration agreements.
Agreements, such as ECOWAS, might drive growth by fostering
increased trade agreements between member countries and with the rest
of the world.
173
174 Trade Agreements and Flows in ECOWAS
However, some studies show that RTAs in Africa have not led to
any increase in trade (Elbadawi, 1997; Yeats, 1999; Diouf, 2002). Diouf
(2002) attributes the lack of success of African Regional Agreements
to the adoption of an integration process that is market-based to the
detriment of transport and communication channels. Sylla (2003) and
others suggest the failure of integration in Africa is due to governments
that jealously protect national sovereignty even when states are barely
viable in the context of globalization. Similarly, the works of Gunning
(2001), ECA (2004), Yang and Gupta (2005) and Chacha (2008) sup-
port the assertion that RTAs in Sub-Saharan Africa have not improved
trade between member countries, because products are not comple-
mentary, insufficient infrastructure, minimal product differentiation,
high costs discouraging imports, narrow markets and a lack of political
commitment.
In contrast, Deme (1995), Cernat (2003), Carrère (2004), Coulibaly
(2007) and Afersorgbor and Bergeijk (2011) find that Sub-Saharan
African RTAs have increased trade flows between member countries.
In a recent study by Afersorgbor and Bergeijk (2011), a gravity model
was applied to 35 countries to analyze the impact of ECOWAS and The
Southern Africa Development Community (SADC)1 trade agreements
on bilateral trade. The impact is considered relatively more significant
than the impact from European Union-African Caribbean and Pacific
States agreements (EU-ACP agreements). Definitively, the contribution
of RTAs to regional and sub-regional integration in Africa remains con-
troversial. Similarly, different arguments are provided to explain these
divergent results. Thus, many ECOWAS actors believe that monetary
integration might contribute to the effectiveness of RTAs. Recently,
Interbank Electronic Banking Group of the West African Economic and
Monetary Union (GIM-WAEMU)2 type interbank payment project was
initiated within ECOWAS by the World Bank and GIM-WAEMU, who
advised that use of the CFA franc, Cedi or Naira should no longer be
an obstacle to trade development in the ECOWAS zone. Removing the
currency constraint in cross-border payments within ECOWAS would
be a key driver of integration. So do trade agreements associated with
the sharing of a single currency have a significant impact on bilat-
eral trade within ECOWAS? In the case of CEMAC,3 Avom (2005)
observes that economic and monetary union had a relatively no sig-
nificant impact on the development of internal trade. Whereas Debrun
et al. (2002), Masson and Patillo (2005) and Bénassy-Quéré and Coupet
(2005) point to the sub-optimal nature of the ECOWAS monetary union
project.
Benjamin Ndong and Sokhana Diarra Mboup 175
Monetary union is supposed to influence trade by easing uncertainty

over exchange rates (Clark, 1973) and transaction costs. It also simpli-
fies cost calculations and price fixing decisions (Kenen, 2003). Further,
a ‘border effect’4 (McCallum, 1995; Head and Mayer, 2000) may be bol-
stered by the adoption of a single currency. So, for Frankel and Rose
(2002), the impacts of a monetary union imply, beyond the growth of
bilateral trade, an increase in the overall openness rate and significant
trade creation. Tsangarides, Ewenczyk and Hulej (2006) indicate that
monetary unions induce significant trade creation and trade stability.
Carrère’s (2004) analysis of the impact of African regional agreements,
with or without monetary unions, supports this. Glick and Rose (2001)
find that sharing a common language, common borders, a regional trade
agreement and the same colonial history will foster trade.
The aim of this research is to demonstrate, by controlling for the
influence of geographical, linguistic and structural factors, the extent
to which trade agreements associated with a monetary union affect the
level of trade within ECOWAS. Based on the gravity model,5 this study
is intended to highlight the determining factors of bilateral trade.
The first section of the chapter offers a description and compara-
tive analysis of ECOWAS’ intra-community trade. The second presents
an analysis of the impact of the regional trade agreement associated
with monetary union on trade flows within ECOWAS using the gravity
model. The third discusses the results with a view to making economic
policy recommendations.
1 Comparative analysis of ECOWAS’ intra-community trade
Official statistics show that intra-regional trade is fairly limited within

ECOWAS. It represents about 6.3% of total exports by member countries
and 6.8% of total imports. Moreover, the marginalization of Africa and
the ECOWAS area in particular is increasing over time.
Global trade 2000–2011

During the 2000s, international trade grew steadily until 2008, when
economic crisis followed the sub-prime financial crisis. Global trade
slumped in 2009, both in terms of exports and imports (see Figures 9.1
and 9.2).
Figure 9.1 shows a generalized drop in exports across all continents
in 2009, declining 41.3% for the Commonwealth of Independent States
(CIS), 20% for South and Central America, 19.67% for Europe, 19.09%
for North America, 18.54% for the Middle East, 14.23% for Africa and
176
1,600
1,400
1,200
1,000
800
600
400
200
0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
World North America Central South America
Europe CEI Africa
Middle East Asia
Figure 9.1 Export trends by region

Note: NORTH AM. = North America / CEN.S.AM. = Central and South America / CIS =
Commonwealth of Independent States / M. EAST = Middle East.
Source: International Trade Statistics, WTO, 2012.
2,500
2,000
1,500
1,000
500
0
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
World North America Central South America
Europe CEI Africa
Middle East Asia
Figure 9.2 Import trends by region

Note: NORTH AM. = North America / CEN.S.AM. = Central and South America / CIS =
Commonwealth of Independent States / M. EAST = Middle East.
13.35% for Asia. By 2009, however, the impact of economic crisis on

world trade was contained, and in 2010, global exports rebounded with
a 21% increase compared to 2009 (see Table 9.A.1 in Annex).6
Besides a sharp downturn following the 2008 economic crisis, world
trade grew steadily in the 2000s. Between 2001 and 2011, Africa’s
exports rose considerably, by 170%. In comparison, North America’s
exports rose by 81%, Europe’s by 78%, Asia’s by 113%, South and Cen-
tral America’s by 189 percent, the Middle East’s by 204% and the CIS’s
by 239 percent. Yet Africa’s share of global trade remains marginal,
although this improved slightly, from 1.7% in 2001 to 2.2% in 2011
(see Table 9.A.1 and 9.A.2 in Annex).
Trade within Africa and ECOWAS

Africa conducts more trade with the rest of the world than with itself.
In 2011, 17% of Africa’s exports went to North America, 35% to Europe
and 25% to Asia; while only 13% was internal trade. The same trend
is observed for imports with, respectively 7%, 37%, 28% and 14% for
internal imports (see Table 9.1).
If Africa is not the only economic area with low intra-regional trade, it
remains that there is still a long way to go. In 2011, intra-regional trade
(exports) was estimated at 71% for Europe, 53% for Asia and 48% for
Table 9.1 African exports and imports, 2011
World North CEN. EUROPE CIS AFRICA M .EAST ASIA

AM S. AM.
2011 2011 2011 2011 2011 2011 2011 2011
AFRICA
AGRIC. PROD 59.49 3.5 2.04 24.82 1.19 12.02 4.81 10.55
F. EXTRAC I. P 382.21 86.92 14.65 127.34 0.37 26.84 3.48 115.24
MANUF, P. 110.3 10.6 2.68 48.29 0.25 28.18 5.86 13.68
TOT. EXPORT. 594.24 101.64 19.45 205.21 1.85 77.03 21.34 145.84
INTER. TRADE 1% 17% 3% 35% 0% 13% 4% 25%
AFRICA
TOT. IMPORT. 538.08 37.47 21.35 199.39 12.49 77.03 37.87 152.48
INTER. TRADE 1% 7% 4% 37% 2% 14% 7% 28%
Note: F.EXTRAC I. P. = fuel and extractive industry products/

MANUF. P = manufactured products/TOT. EXPORT. = total exports/
INTER. TRADE = interregional trade/TOT. IMPORT= total imports/
NORTH. AM. = North America/CEN.S.AM. = Central and South America/
CIS = Commonwealth of Independent States/M.EAST = Middle East/
Source: International Trade Statistics, WTO, 2012./AGRIC. PROD = agricultural products/.
North America (North American Free Trade Agreement, NAFTA). Mean-

while, in South and Central America, the share of intra-regional exports
was 27 percent, 20% in the Commonwealth of Independent states (CIS)
and, finally 9% in the Middle East.
The ECOWAS zone follows the overall African trend. Intra-
ECOWAS trade has been weak, with exports within the zone estimated
at 6.3% and imports at 6.8% in 2011 (see Table 9.2 and Table 9.A.3 in
Annex). Here ECOWAS performs better than the Arab Magrheb Union
(AMU), the Economic and Monetary Community of Central Africa
(CEMAC) and the Economic Community of Central Africa States
(ECCAS); but less well than the Common Market for Eastern and South-
ern Africa (COMESA), the Southern African Development Community
(SADC), West African Economic and Monetary Union (WAEMU) and
Africa as a whole. WAEMU7 is an economic and monetary union in the
ECOWAS area, so how does monetary union affect its difference in per-
formance (6.3% and 12.1% respectively in 2011) when intra-WAEMU
trade is more buoyant than intra-ECOWAS trade, with 12.1% of total
exports and 6.5% of total imports (Table 9.2 and Table 9.A.3 in Annex).
Interestingly, there is little trade between the WAMZ countries,8 which
share the ECOWAS space with WAEMU.
While the weakness of intra-regional trade is a limiting factor on the
potential gains of a monetary union, the latter tends to increase the
volume of trade (Frankel and Rose, 2002; Bénassy-Quéré and Coupet,
2005) (Table 9.3).
Export effort
Like most African countries, ECOWAS countries specialize in a limited
number of commodities. Nigeria’s exports mainly comprise oil products
(94%). Many countries export cotton, but most other products are spe-
cific to one or two countries (see Table 9.A.4 in Annex and Table 9.4).
The average export effort for the least developed countries (LDCS) in
ECOWAS was 25.72% in 2010, with countries such as Togo (37%) and
Guinea (32%) showing exports below the global level. The export efforts
of LDCs in ECOWAS are significant. The weakness of intra-community
trade, Africa’s low share in general, and ECOWAS’s in particular, is
undoubtedly due to specialization in a limited number of commodities
and a failure to diversify economies. Table 9.A.4 (in Annex) shows how
LDCs in ECOWAS specialize in commodities (oil, extractive and agricul-
tural products) and how their exports fluctuated on average by 19.2%
during 2005–2011, far above the global level of 12%.
Dependence on primary products with particularly unstable prices
leads to structural instability and high terms of trade volatility.
Table 9.2 Intra-African trade by economic zone in 2011, percentages of total exports
AMU CEMAC COMESA ECCAS ECOWAS FRANC ZONE SADC WAEMU AFRICA WORLD
−−−−−−−−−−−−→ Exports to
AMU 3.6 0.2 1.7 0.3 0.8 0.7 0.1 0.5 5.6 100.0
CEMAC 0.2 1.0 0.2 1.3 0.9 1.4 0.4 0.3 2.7 100.0
COMESA 2.8 0.2 8.9 2.1 0.5 0.4 7.6 0.2 16.1 100.0
ECCAS 0.1 0.4 1.8 0.6 0.4 0.6 3.0 0.1 4.1 100.0
ECOWAS 0.3 1.3 0.1 1.4 6.3 4.7 3.0 3.4 10.9 100.0
FRANC ZONE 0.4 1.3 0.3 1.6 7.5 5.1 1.6 3.8 10.9 100.0
SADC 0.3 0.1 6.5 1.7 1.1 0.4 9.9 0.3 12.5 100.0
WAEMU 0.9 2.0 0.4 2.5 23.2 14.1 4.3 12.1 30.5 100.0
AFRICA 1.4 0.5 3.5 1.3 2.3 1.6 4.6 1.1 10.1 100.0
WORLD 0.7 0.1 0.8 0.2 0.7 0.3 0.9 0.2 2.9 100.0
Note: AMU: Arab Maghreb Union/CEMAC: Central African Economic and Monetary Union/COMESA: Common Market of Eastern and Southern
Africa/ECCAS: Economic Community of Central African States/SADC: Southern Africa Development Community.
Source: African Statistical Yearbook, 2013.
179
Table 9.3 Comparative trend of intra-zone trade as percent of total exports of

the economic zones concerned
2008 2009 2010 2011 2012 Average Average

2008–2012 2003–2004
EU 65.68 65.94 64.30 63.75 62.78 64.49

ECOWAS 7.32 3.24 6.3 7.15 6.00 9.0
CEMAC 1.0 1.63
WAEMU 14.50 12.47 10.15 12.1 12.52 12.34 15.2
COMESA 8.04 8.77 8.9 7.13 8.21
MERCOSUR 14.96 15.06 16.10 15.36 15.05 15.31
Note: EU: European Union/MERCOSUR: South American Common Market.

Source: Trade Map, Diop and Séne (2013), Diop (2007) and the authors.
Table 9.4 ECOWAS’ Least Developed Countries (LDC), export effort and ratio of
exports to GDP
Value Ratio to GDP

(in million
US dollars)
GDP Goods and Goods Commercial

Commercial Services
Services
2010 2005 2010 2005 2010 2005 2010
Benin 6558 18 25 13 20 4 5
Burkina Faso 8825 10 21 9 18 1 3
The Gambia 1050 29 24 16 16 13 8
Guinea 4736 30 32 29 31 1 1
Guinea Bissau 835 16 20 16 15 1 5
Liberia 988 39 28 24 24 15 4
Mali 9422 26 26 21 22 5 4
Niger 5411 17 24 14 21 2 2
Senegal 12855 26 24 18 17 8 7
Sierra Leone 1910 21 22 15 19 6 3
Togo 3176 37 40 30 31 7 9
World 63134700 28 30 23 24 5 6
The shocks affecting different economies in the sub-region are not

necessarily correlated, so a significant increase in the price of crude oil
has a negative impact on most of the countries which are net importers;
but it favors the Nigerian economy, all other things being equal.
Several explanations are given for the low level of intra-community

trade: weak production, non-complementary and sometimes compet-
ing production between states, political instability and poor governance
within the community, a high volume of trade not included in the
official statistics because of its informal nature or smuggling. The low
level of intra-community trade may also be related to the absence of
a common currency for all ECOWAS countries, or the different official
languages in French and English speaking countries.
2 The Effect of RTA associated with a single currency

on trade flows
This study uses a gravity model9 based on Newton’s theory of grav-

ity. In the model, trade flows between two countries are expressed as
proportional to their economic sizes and inversely proportional to the
distance between them. It is used to calculate trade flows between coun-
tries belonging to the same monetary zone. Since Rose (2000), a dummy
variable equal to 1 is used where two countries share the same currency
(otherwise it is 0), to determine the impact of monetary unions on trade.
Specifications
The gravity model has many applications in analysis of international
trade. Since Tinbergen’s (1962) pioneering work, Newton’s gravitational
equation has been adapted to international trade flows. Newton’s theory
of gravity holds that the gravitational force, or attraction, between two
objects (planets for example) is proportional to their respective masses
and inversely proportional to the distance between them.
M i × Mj
Aij = (1)
Dij
Here A is the gravitational force or attraction between the two planets, M

is the mass of the planets and D is the distance between the two planets.
Drawing upon the Newtonian theory, Tinbergen (1962) stipulates
that trade is a function of the size of each economy estimated by the
Gross Domestic Product (GDP) and transport costs with, as a proxy,
the distance between the trading partner countries. Equation (1) is thus
reformulated as follows:
Yi × Yj )
(Y
Xij = α0 (2)
Dij
Here Xij is the total exports of country i to country j; α0 is a coefficient

of proportionality; Y is GDP; and Dij is distance between country i and
country j.
In its simplest form, the gravitation equation (2) may be reformulated
as follows for any coupling (i, j):
α
Y i × Y j )α 2
Xij = α0 × Dij1 (Y (3)
Here Xij represents the value of bilateral trade between country i and
country j; Y is GDP; and Dij measures the distance between country i
and country j.
Expressed in logarithmic form, the equation (3) becomes:
ln Xij = α0 + α1 ln ((Dij ) + α2 ln (Y
Yi × Yj ) (4)
Here α0 , α1 and α2 are coefficients; α1 is assumed to be negative while α2

is assumed to be positive.
Taking into account the existence of control variables (Cij ), the
equation (4) becomes:
ln Xij = α0 + α1 ln ((Dij ) + α2 ln (Y
Yi × Yj ) + α3 Cij (5)
Here α3 is the vector of the coefficients associated with the vector of the
control variables Cij . The control variables are added because bilateral
trade is not only explained by the economic size of partners and the
distance between them, but also by other variables, such as historical ties
(language, former colonizer) and geographical characteristics (common
border, landlockedness).
In this analysis, a dummy variable is introduced and expressed as
follows: trade agreements plus monetary union (RTA&MUij ) takes the
value 1 where two countries share both a trade bloc and monetary
union, otherwise it is 0. This variable allows us to take into account our
problem and provide an innovating factor in comparison with the pre-
vious studies where the two effects were separated in the same equation.
Initially, a cross-sectional analysis is selected where the gravity equa-
tion is estimated using the Ordinary Least Squares (OLS) method. The
OLS method is considered appropriate, when most of the variables are
time-bound (Micco and Ordônez, 2002; Diop, 2007).
ln ((Xij ) = α0 + α1 ln (Y
Yi × Yj ) + α2 ln (yi × yj ) + α3 ln Dij + α4 Comlanggij +
α5 FTRij + α6 LANDLOCKij + α7 ISLANDij + α8 Comcolij
+ α9 RTA&MU
Uij + εij (6)
Here Xij Indicates the total exports of country j to country i; Yi Yj is the

product of the countries’ GDP; and yi yj is the product of GDPs per capita;
Dij is the relative distance between the two trading partners. Comlangij
represents the official language dummy variable, with a value of 1 where
the partners have a common official language, otherwise it is 0. FTRij is
the common border dummy variable, which takes a value of 1 if the
partners have a common border, otherwise it is 0. LANDLOCKij is a
dummy variable which takes the value of 0 if neither of the two coun-
tries is without a seaboard, the value of 1 if one of the two countries
does not have a seaboard and 2 when the two partner countries do not
have a seaboard; ISLANDij is a dummy variable which takes the value
of 0 if neither of the countries is an island, the value of 1, when one
of the two countries is an island and 2 when both partner countries are
islands. Comcolij represents the colonial master dummy variable which
has a value of 1 if the two partners have the same former colonial ruler,
otherwise it is 0. RTA&MUij is a dummy variable, which takes a value
of 1 if both countries share both a trade bloc and a monetary union,
otherwise it is 0. εij is the error term.
As a second step, panel data is selected (Rose, 2000; Micco and
Ordônez, 2002; Masson and Patillo, 2005). In an initial specification,
this panel model only contains the ‘quantitative’ variables and our vari-
able of interest (RTA&MU). Indeed, the dummy variables do not vary, or
vary very little, over time; and their effects are not taken into account
in the specific effect (αij ). Yet to ensure a more accurate specification of
this panel model, the Hausman (1978) test is specified for the following
models of the equation, (7) and (8):
ln ((Xij )t = α0 + α1 ln (Y
Yi ∗ Yj )t + α2 ln (yi ∗ yj )t + α3 RTA&MU
Uij + αij + (εij )t (7)
Here αij is individual effect, and εij is residual effect. However, in a second
specification, the dummy variables are introduced to test the robustness
of the previous model (equation 7).
ln ((Xij )t = α0 + α1 ln (Y
Yi × Yj )t + α2 (yi × yj )t + α3 ln Dij + α4 Comlanggij
+ α5 FTRij + α6 LANDLOCKij + α7 ISLANDij + α8 Comcolij
+ α9 RTA&MU
Uij + αij + (εij )t (8)
It should be noted that the variables of the different models are classified
into two categories. The first concerns continuous variables, expressed
in logarithms, which is the case of the product of GDPs, the prod-
uct of GDPs per capita and the relative distance. This form allows
interpretation of the coefficients of these variables as elasticities. The

second comprises qualitative variables (they take the values of 0, 1 or 2),
which are called ‘multilateral resistance’ as in the case of language,
landlockedness and a common past (same colonial rulers). They are spe-
cific to each country and reflect individual characteristics or close ties
between the partners that are likely to have a positive or negative impact
on the volume of trade.
Expected signs of coefficients

The theoretical signs of the different coefficients are described and
explained below:
• Product of GDPs and GDPs per capita are indicators of the poten-
tial size of the market and of purchasing power, respectively. The
higher these products, the greater the opportunities for trade for both
partners. The sign of the associated coefficients should be positive.
• Relative distance allows transaction costs to be calculated approx-
imately, especially, the cost of transport for trade. Distance is a
negative factor on trade, as it generates transport costs that are
particularly high when distance between the two partners is greater.
• Landlockedness variable makes it possible to estimate what effect the
lack of a seaboard has on trade for the countries concerned. Transport
statistics show that 80% of trade is carried out by ship. A country
which is an island is thus considered to be handicapped in terms of
trade.
• Studies have shown that a common colonial past shared between
partners influences bilateral trade. Thus, common former colonial
ruler and common language variables should impact bilateral trade
positively.
• The RTA&MU variable makes it possible to calculate the assumed pos-
itive effect of a monetary union coupled with a trade agreement.
In theory, trade agreements and shared currencies foster bilateral
trade between member countries (Table 9.5).
Data and samples

On certain dates, export values for country couplings may be zero, pos-
ing problems when exports are expressed in logarithms. In line with
Rose (2000), the value of 100 is added each value of bilateral trade,
giving ln Xij = ln ((Xij + 100) where Xij is the initial value of exports
drawn from the database. In a concern for consistency of statistical
Table 9.5 Expected signs
Variables Expected signs
The product of GDPs +

The product of GDPs per capita +
Relative distance −
Common border +
Common official language +
Common colonial master +
Landlockedness −
Island −
RTA&MU +
information, especially that relating to bilateral trade, the Trade Map

database is used. The data concern the 2003 to 2012 period. The sam-
ple concerns ECOWAS countries, but excludes Guinea-Bissau for which
multi-year data is not available. For this sample, an OLS (equation 6), a
panel 1 (equation 7) and panel 2 (equation 8) are used to estimate the
effect of a regional trade agreement associated with monetary union.
Thus the OLS is applied to the averages for the following sub-periods:
the 2003–2007 sub-period which marks the pre-financial and economic
crises period, the 2008–2009 sub-period identified by the stylized facts
presented in the previous section as the time when the crisis negatively
impacted global trade (see Figures 9.1 and 9.2 and Tables 9.A.1 and
9.A.2 in annexes) and the 2010–2012 sub-period which marks the recov-
ery of international economic growth. The 2003–2012 study period is
also considered.
Estimation of the model

The results of the two estimations are presented in two tables, showing
both the cross-sectional and panel data (Tables 9.6 and 9.7).
3 Discussion and implications
With the exception of the dummy variables, the estimated coeffi-

cients of variables were directly interpreted as elasticities, since they
are expressed in natural logarithms. However, the elasticities of the
qualitative variables are given as the exponential of the estimated
coefficients.
Overall, the models are significant. The coefficients associated with
the traditional variables of the gravity model are also very significant
Table 9.6 Results of estimation on cross-sectional data with OLS on averages
Av.03–07 Av.08–09 Av.10–12 Av.03–12
Y*Y 0.52∗ 0.70∗ 0.39∗∗ 0.54∗

(4.96) (6.16) (2.97) (5.12)
y*y 0.74∗∗ 0.35 1.39∗ 0.86∗∗
(2.16) (1.05) (3.36) (2.53)
D −1.11∗ −0.96∗ −1.40∗ −1.15∗
(−4.49) (−3.86) (−5.51) (−5.06)
ComLang 0.47 0.31 0.50 0.55
(0.79) (0.52) (0.84) (1.01)
FTR 0.70∗∗ 0.99∗ 0.97∗∗ 0.82∗∗
(2.04) (2.85) (2.74) (2.58)
LANDLOCKL −0.57∗∗ −0.85∗ −0.38 −0.58∗∗
(−2.19) (−3.30) (−1.37) (−2.36)
Island −1.64∗∗ −1.80∗∗ −3.79∗ −2.27∗
(−2.18) (−2.31) (−4.08) (−3.02)
Comcol −0.73 −0.83 −0.32 −0.74
(−1.17) (−1.35) (−0.51) (−1.29)
RTA&MU 2.35∗ 2.35∗ 1.82∗ 2.15∗
(5.66) (5.61) (4.40) (5.63)
Constant −17.5∗ −21.7∗ −18.6∗ −19.65∗
(−5.52) (−6.84) (−5.79) (−6.73)
Adjusted R2 0.66 0.69 0.72 0.72
Obs. 180 180 180 180
Note: (∗ ) significant at 1%, (∗∗ ) significant at 5%, (∗∗∗ ) significant at 10%. Student’s t statistics
are reported in brackets.
and present the expected signs except for the common colonial ruler
variable. On average, the value of bilateral trade rises in line with the
size of the economies and falls with distance. The existence of a com-
mon border fosters trade; whereas the landlocked or insular position of
a partner country or two partner countries does not foster trade. Overall,
geographical and structural factors have a determining impact on intra-
regional trade. The GDP and GDP per capita variables have a positive
effect on bilateral trade, so their coefficients are positive. Geographical
characteristics such as relative distance, landlockedness and islandness
impede bilateral trade, although a common border, impacts positively
on bilateral trade.
For the cross-sectional and panel data, the coefficient associated with
the colonial ruler variable is negative, even though it is not statistically
significant. The cross-sectional data results are confirmed by the panel
data with the exception of the sign of the coefficient of the product
of GDP per capita variable which is not statistically significant in the
Table 9.7 Results of estimation using panel data 2003–2012
Equation (7) Equation (8)
Y*Y 0. 623∗ 0. 699∗ 0. 581∗ 0. 585∗

(21.84) (21.00) (23.55) (13.72)
y*y −0. 284∗ −0.009
(−4.43) (−0.098)
D −1. 293∗ −1. 29∗
(−15.37) (−15.35)
ComLang 0. 353∗∗∗ 0. 352∗∗∗
(1.746) (1.739)
FTR 0. 537∗ 0. 537∗
(4.585) (4.58)
LANDLOCK −0. 673∗ −0. 677∗
(−6.925) (−6.55)
Island −0. 604∗ −0. 584∗
(−3.884) (−2.27)
Comcol −0. 399∗∗∗ −0. 399∗∗∗
(−1.833) (−1.83)
RTA&MU 1. 678∗ 1. 608∗ 1. 66∗ 1. 664∗
(14.04) (13.42) (11.191) (11.12)
Constant −20. 63∗ −20. 45∗ −9. 444∗ −9. 47∗
(−16.21) (−16.15) (−7.822) (−7.60)
Adjusted R2 0.60 0.61 0.72 0.72
Prob. (F.stat.) 0.0000 0.0000 0.0000 0.0000
Obs. 1,635 1,635 1,635 1,635
Hausman (χ 2 ) 5.42 4.99 8.58 9.68
(prob. = 0.0663) (prob. = 0.172) (prob. = 0.379) (prob. = 0.376)
Note: (∗ ) significant at 1%, (∗∗ ) significant at 5%, (∗∗∗ ) significant at 10%. Student t statistics
are reported in brackets.
equation (8), which was used to control the results of equation (7).
Stability is noted among the different relationships described in their
dual time and spatial or static and dynamic dimensions.
The coefficients associated with the product of GDPs, product of GDP
per capita and relative distance variables are elasticities. When asso-
ciated with the product of GDPs variable, for example, these indicate
that, all other things being equal, a 1% rise in the product of GDPs is
reflected in a 0.54% increase in trade and a rise in the product of GDPs
per capita results in a 0.86% increase. Other things being equal, a 1%
increase in distance results in a 1.15% drop in trade over the 2003–2012
period.
Estimation of the model also showed that partners sharing a com-
mon border traded, on average, over the 2003–2012 period 2.27 times
(e0.82 ) more between themselves than with others. Also on average for
2003–2012, landlockedness and insularity respectively resulted in a drop
of 1.79 (e0−58 ) and 9.68 (e2−27 ) times more.
Concerning the research subject, the results showed that sharing trade
agreement and single currency has a real impact on trade between mem-
ber countries. This result is confirmed both by the cross-sectional and
panel data estimations. During 2003–2012, countries sharing a trade
agreement and a single currency traded with each other, on average,
8.41 (e2.15 ) times more than countries outside the same monetary and
free trade zone. In the ECOWAS area, countries sharing a trade agree-
ment and single currency belong to WAEMU and the result is consistent
with those of Carrère (2004) and Masson and Patillo (2005) that showed
WAEMU countries trade more with each other than with the other
ECOWAS countries.
In the case of ECOWAS, the border effect is thought to be more signifi-
cant than the effect related to the use of the same language. For example,
trade of Togo and Benin (French-speaking countries) with, respectively,
Ghana and Nigeria (English-speaking countries) is buoyant. This sit-
uation is undoubtedly due to the statistical non-significance of the
coefficient associated with the common official language variable, while
the negative sign of the coefficient of the common colonial ruler vari-
able may suggest excessive polarization of trade from former colonies
to the former colonizing country rather than to each other. This was
noted in the case of WAEMU countries and their relations with France:
the nature of their monetary system (pegged to the Euro, a guarantee of
convertibility) should favor greater trade openness (Guillaumont, 2002)
and may also explain this result.
The result of the panel data estimation relating to GDP per capita also
appears interesting. It would indicate that, in a dynamic perspective,
an increase in GDP per capita, that is, purchasing power, widens the
national market. The need to satisfy this national market above all is
likely to reduce foreign trade. Overall, it was noted that trade potential
is significant and community development avenues multifarious. The
potential knock-on effects of intense economic cooperation between
ECOWAS member countries are enormous due to the number and
diversity of the populations making up the community.
Economic policy recommendations

The main lesson from the estimated models’ results is the observation
that, despite the assumed weakness of trade among ECOWAS countries,
trade agreements associated with a single currency induce a significant

trade creation effect. This result likely indicates that the expansion of
ECOWAS into a common monetary zone should, other things remain-
ing equal, considerably increase the intensity of intra-regional trade. So
a trade agreement linked to a single currency is a determining factor in
the development of trade in ECOWAS.
However, referring to the theory of optimal monetary zones, more
emphasis should be placed on the implementation of structural reforms
to strengthen complementarity between the ECOWAS economies,
which must become more diversified and integrated, in terms of the
coordination of macroeconomic policies, harmonization of adminis-
trative rules and procedures and the development of communication
channels.
Trade growth is at least partly dependent on the characteristics of
trade agreements. With a view to standardizing trade agreements in
West Africa, particular attention should be paid to the credibility of
commitments made at the institutional level.
Limitations and prospects

In addition to the benefits of being party to a regional trade agreement
associated with a single currency, empirical analysis of trade among
ECOWAS countries highlights other factors likely to influence their trade
flows. A key obstacle to trade among ECOWAS countries is the inad-
equate physical and non-physical infrastructure, especially transport,
road and telecommunications infrastructure, as studied by Longo and
Sekkat (2004).
Economic and political governance, including institutional factors
such as poor management of economic policies and political tension,
might also be considered constraints on trade between ECOWAS coun-
tries (see Longo and Sekkat, 2004). By factoring into the gravity equation
a variable that captures the inducement of parallel markets to engage in
fraudulent trade, Agbodji (2007) shows the existence of these markets
significantly reduces formal trade among the countries of Sub-Saharan
Africa.
Finally, the sample configuration does not allow for a generalization
of the main result. A more representative sample of global trade could
confirm the impact on bilateral trade in general of a regional trade
agreement linked to a single currency.
190
Annexes
Table 9.A.1 Trend in exports of goods by region
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2008–09 2009–10 2001–11
WORLD 729.1 693.1 724.8 814.9 901.1 1026 1148.2 1287.4 1056 1278.3 1480.4 −18% 21% 103%
NORTH.AM 265 258.7 267.7 301 332.6 364.9 385.4 413.2 334.3 413.2 479.8 −19% 24% 81%
CEN.SOUTH AM 57,6 50.9 51.5 60.7 71.5 87.2 105.4 135 108 137.1 166.4 −20% 27% 189%
EUROPE 177.7 160.1 168.9 187 204 235.8 272.5 311.1 249.9 276.3 315.7 −20% 11% 78%
CIS 3.8 3.8 3.7 4.8 5.9 7.1 10.5 13,8 8.1 9.2 12.9 −41% 14% 239%
AFRICA 12.3 10.7 10.7 13.4 15.5 18.8 24.1 28.8 24.7 28.7 33.2 −14% 16% 170%
M. EAST 19.2 18.9 19.4 23.5 31.5 37.2 45.1 55 44.8 48.7 58.4 −19% 9% 204%
ASIA 193.1 189.7 202.7 224.1 239.5 274.2 304.4 329.4 285.4 364 412.1 −13% 28% 113%
Africa’s share of global exports
AFRICA/WORLD 0.017 0.015 0.015 0.016 0.017 0.018 0.021 0.022 0.023 0.022 0.022
Note: NORTH. AM = North America/CEN.SOUTH.AM = Central and South America/CSI = Commonwealth of Independent States/M. EAST = Middle
East/Africa/World = ratio of African imports/global imports.
Source: International Trade Statistics, WTO, 2012 and authors’ calculations.
Table 9.A.2 Trend in imports of goods by region
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2008–09 2009–10 2001–11
WORLD 1179.2 1200.2 1303.1 1732.7 1910.1 1910.1 2020.4 2169.5 1605.3 1969.2 2265.9 −26% 23% 92%
NORTH.AM 352.6 349.2 366.5 417.7 464.6 507.9 531.6 559 407.1 512.2 585.7 −27% 26% 66%
CEN.SOUTH AM 71.7 73.7 83.8 105.2 129.9 140.4 141.8 167.4 113.5 136.4 178.8 −32% 20% 149%
EUROPE 252.8 259.7 282 317.5 346.8 369.9 393.5 409.6 314.5 359.2 416.5 −23% 14% 65%
CIS 8.1 8.4 10.4 14.9 19.8 25.3 26.5 38.5 24.1 32.2 42 −37% 34% 419%
AFRICA 26.8 23.3 33.9 48,3 67.9 83.8 95.2 117.3 64.6 87.5 94.9 −45% 35% 254%
M. EAST 38.7 35.8 43.9 54.2 65.7 74.7 80.3 115.3 61 76.9 106.3 −47% 26% 175%
ASIA 428.6 450.2 482.6 567.7 638 716 751.4 762.4 620.4 764.7 841.7 −19% 23% 96%
Africa’s share of global imports
AFRICA/WORLD 0.023 0.019 0.026 0.028 0.036 0.044 0.047 0.054 0.040 0.044 0.042
Note: NORTH. AM = North America/CEN.SOUTH.AM = Central and South America/CSI = Commonwealth of Independent States/M. EAST = Middle
East/Africa/World = ratio of African imports/global imports.
Source: International Trade Statistics, WTO, 2012 and authors’ calculations.
191
192
Table 9.A.3 Intra-African trade by economic zone in 2011, percentages of total imports
AMU CEMAC COMESA ECCAS ECOWAS FRANC ZONE SADC WAEMU AFRICA WORLD
Imports to −−−−−−−−−−−−→
AMU 4.0 0.1 2.2 0.1 0.4 0.3 0.5 0.2 6.4 100.0
CEMAC 1.8 2.7 1.1 2.8 6.5 4.8 1.7 2.1 13.5 100.0
COMESA 1.7 0.1 6.1 1.3 0.1 0.1 9.4 0.1 14.8 100.0
ECCAS 0.9 1.4 4.9 1.6 3.2 2.5 8.2 1.1 16.6 100.0
ECOWAS 0.8 0.2 0.3 0.3 6.8 3.4 1.6 3.2 9.7 100.0
FRANC ZONE 1.8 1.2 0.7 1.3 11.9 6.2 1.6 5.0 16.9 100.0
SADC 0.1 0.1 3.4 2.0 2.2 0.2 11.4 0.1 14.7 100.0
WAEMU 1.8 0.4 0.4 0.5 14.8 7.0 1.4 6.5 18.7 100.0
AFRICA 1.4 0.2 2.5 0.8 2.6 1.1 4.4 0.9 10.1 100.0
WORLD 0.7 0.2 0.6 0.6 0.8 0.3 1.2 0.1 3.3 100.0
Source: African Statistical Yearbook 2013.

Table 9.A.4 Exports and imports of ECOWAS’ Least Developed Countries (LDC), in millions of USD
Exports Imports
Value Annual variation in % Value Annual variation in %
2011 2005–11 2009 2010 2011 2011 2005–11 2009 2010 2011
Exports of agricultural products

Burkina Faso 2000 34 30 43 55 2600 16 −7 10 27
Benin 1800 25 −4 13 29 2700 22 −10 5 25
Togo 1100 11 −6 0 38 1700 10 0 −1 13
Liberia 358 22 −39 49 61 1044 28 −32 29 47
Guinea Bissau 230 21 −5 −1 92 300 23 16 −4 36
Exports of extractive industry products excluding fuel
Mali 2391 17 −15 13 20 3250 16 −26 15 14
Guinea 1630 14 −22 40 11 2106 21 −22 33 50
Niger 1250 21 10 4 20 2400 21 30 5 4
Sierra Leone 350 17 7 48 2 1708 38 −3 48 122
Senegal 2542 10 −7 7 18 5909 11 −28 1 24
Oil exports (no ECOWAS LDC)
Manufacturing product exports (no ECOWAS LDC)
World 18255000 12 −22 22 20 18438000 11 −23 21 19

193
Notes
1. The Southern Africa Development Community (SADC).
2. The Interbank Electronic Banking Group (GIM) of the West African Economic
and Monetary Union (WAEMU) is an interbank payment system that provides
access to all automatic teller machines (ATMs) for the withdrawal of money in
the WAEMU zone. Each withdrawal is charged CFA F 500 (A transaction costs
CFAF 500).
3. The Central African Economic and Monetary Union (CEMAC).
4. A large body of empirical research finds that a pair if regions within a coun-
try tends to trade 10 to 20 times much than s an otherwise identical pair
of regions across countries. This result has been called the ‘border effect’.
McCallum (1995) was the first to find this effect.
5. This type of model has been used in three previous studies to underscore the
achievements of monetary union countries with regard to trade integration
(see Guillaumat, 2002), but this study addresses a more specific question using
more recent data.
6. The same behavior is noted for imports, see Figure 9.2 and Tables 9.A.1 and
9.A.2 in Annex).
7. WAEMU Trade Integration: Article 4 of the WAEMU Treaty (1994) stipulates
the creation of a common market with the free movement of persons, goods,
services and capital. The Common External Tariff (CET) has been in force since
January 2000.
8. ECOWAS comprises two separate sub-groups: First are the WAEMU countries
(Benin, Burkina Faso, Côte d’Ivoire, Guinea-Bissau, Mali, Niger, Senegal, Togo)
which form an economic and customs union with the CFAF as a common
currency. The second group is known as WAMZ, a project for a second West
African Monetary Zone with five countries: Gambia, Ghana, Guinea, Nigeria
and Sierra Leone). This group represents 75% of regional GDP and 70% of the
population. Cape Verde and Liberia are in neither group.
9. The gravity model is the name used for the family of quantitative models
developed by the astronomer J. Q. Stewart in 1940.
References
Afersorgbor, S. K. and van Bergeijk, P. A. G. (2011), ‘Multi-Membership and the
Effectiveness of Regional Trade Agreements in Western and Southern Africa. A
Comparative Study of ECOWAS and SADC’, ISS Working Paper Series, vol. 520.
Agbodji, A. E. (2007), ‘Integration and Intra-Sub-Regional Trade: Case of
WAEMU’, African Integration Review, vol. 1, no. 1, January 2007.
Avom, D. (2005), ‘The Determining Factors of Trade in CEMAC: An Empirical
Assessment’, Economie Appliquée, vol. LVIII, no. 2, pp.127–153.
Bénassy-Quéré, A. and Coupet, M. (2005), ‘On the Adequacy of Exchange
Rate Arrangements in Sub-Saharan Africa’, World Economy, vol. 28, no. 3,
pp. 349–373.
Carrère, C. (2004), ‘African Regional Agreements: Impact on Trade with or with-
Cernat, L. (2003), ‘Assessing South-South Regional Integration: Same Issues,

Many Metrics’, Policy Issues in International Trade and Commodities Study
Series no. 21, UNCTAD, Genève.
Chacha, M. (2008), ‘Regional Integration and Trade: Overlapping Memberships
and the Challenges of Regionalism in Africa’, International Studies Association
Conference, San Francisco, CA, March, pp. 26–29.
Clark, P. B. (1973), ‘Uncertainty, Exchange Risk and the Level of International
Trade’, Western Economic Journal, vol. 11, no. 3, pp. 302–313.
Coulibaly, S. (2007), ‘Evaluating the Trade Effect of Developing Regional
Trade Agreements: A Semi-Parametric Approach’, Working Paper WPS4220.
Washington, DC: World Bank.
Debrun, X., Masson, P. and Patillo, C. (2002), ‘Monetary Union in West Africa:
Who Might Gain, Who Might Lose, and Why?’, IMF Working Paper, r WP/02/226.
Deme, M. (1995), ‘The Impact of ECOWAS on Intra-Regional Trade Flows: An
Empirical Investigation’, The Review of Black Political Economy, Winter 1995,
pp. 113–129.
Diop, C. A. (2007), ‘WAEMU and the Prospect of an ECOWAS Single Monetary
Zone: Lessons from a Gravity Model’, Study and Research Paper No. DER (07)
01-BCEAO.
Diop, I. T. and Sène, B. (2013), ‘The Single Currency: A Factor of Economic Emer-
gence on the African Continent’, Seminar on the Senegalese Economy, September
2013, UCAD, Dakar, pp. 27–28.
Diouf, M. (2002), ‘Globalism and Regionalism, the New Regionalism in Africa’,
Codesria, Ilo.
ECA. (2004), Assessing Regional Integration in Africa, Addis Ababa: Economic
Commission for Africa.
Elbadawi, I. A. (1997), ‘The Impact of Regional Trade and Monetary Schemes on
Intra-Sub-Saharan Africa Trade’, in A. Oyejide, I. Elbadawi and P. Collier (eds.),
Regional Integration and Trade Liberalization in Sub-Saharan Africa, Volume I:
Framework, Issues and Methodological Perspectives, London: Macmillan.
Frankel, J. and Rose, A. (2002), ‘Estimating the Effect of Currency Unions on
Trade and Output’, Quarterly Journal of Economics, vol. 117, no. 2.
Glick, R. and Rose, A. K. (2001), ‘Does a Currency Union Affect Trade? The Time
Series Evidence’, Revised Draft: 21 November 2001 Comments Welcome.
Guillaumont, S. (2002), ‘Results of Monetary Integration in WAMU’, Proceedings
of the Symposium for the Fourth Anniversary of BCEAO Volume 1.
Gunning, J. W. (2001), ‘Trade Blocs: Relevant for Africa?’, Journal of African
Economies, vol. 10, no. 3, pp. 311–335.
Hausman, J. A. (1978), ‘Specification Tests in Econometrics’, Econometrica, vol. 46,
no. 6, pp. 1251–1271.
Head, K. and Mayer, T. (2000), ‘Non-Europe: The Magnitude and Causes of
Market Fragmentation in the EU’, Weltwirtschaftliches Archiv, vol. 136, no. 2,
pp. 285–314.
Kenen, P. (2003), ‘What We Can Learn from the Theory of Optimum Currency
Areas’, in Submissions on EMU from Leading Academics, London: HM Treasury.
Longo, R. and Sekkat, K. (2004), ‘Economic Obstacles to Expanding Intra-African
Trade’, World Development, t vol. 32, no. 8, pp. 1309–1321.
Masson, P. and Patillo, C. (2005), The Monetary Geography of Africa, Washington
DC: Brookings Institution Press.
McCallum, J. (1995), ‘National Borders Matter: Canada-US Regional Trade Pat-

terns’, American Economic Review, vol. 35, no. 6, pp. 1295–1309.
Micco, E. and Ordônez, G. (2002). The Currency Union Effect on Trade: Early Evi-
dence from the European Union, Washington DC: Inter-American Development
Bank.
Rose, A. K. (2000), ‘One Money, One Market: Estimating the Effect of Common
Stewart, J. Q. (1940), ‘Suggested Principles of Social Physics’, Science, 106.
Sylla, K. (2003), ‘Why Is the Economic Integration Process in Africa Mark-
ing Time?’, in Regional Integration, Democracy and Pan-Africanism: Old
Paradigms and New Challenges. Series of Codesria Books.
Tsangarides, C. G., Ewenczyk, P. and Hulej, M. (2006), ‘Stylized Facts on Bilat-
eral Trade and Currency Unions: Implications for Africa’, IMF Working Paper, r
WP/06/31.
Tinbergen, J. (1962), Shaping the World Economy: Suggestions for an International
Economy Policy, New York: The Twentieth Country Fund.
WTO (2013), World Trade Report, Geneva: World Trade Organization.
Yang, Y. and Gupta, S. (2005), ‘Regional Trade Arrangements in Africa: Past Per-
formance and the Way Forward’, IMF Working Paper, r WP/05/36, Washington,
DC: IMF.
Yeats, A. (1999), ‘What Can Be Expected from African Regional Trade Agree-
ments?’, Policy Research Working Papers, no. 2004, Washington DC: World
Bank.
10
A DSGE Model of Trade and
Risk-Sharing Effects of Currency
Union on Economic Integration of
the CFA Zone
Thierry Kame Babilla
1 The mitigating effects of monetary union on

heterogeneous shocks
Since its creation in 1963, the African Union has sought to foster
regional economic integration through monetary unions. But of six
major attempts at currency union, only the CFA Zone – using both
West African and Central African CFA francs – has been successful
(Carrère, 2004; Masson and Pattillo, 2004; Tsangarides et al., 2006,
2008; Tapsoba, 2009a, 2009b). Many have argued that African insti-
tutions are not sufficiently developed to deal with the heterogeneous
shocks to which African monetary unions have often been subject
(Eichengreen, 1992; Krugman, 1993; Fielding and Shields, 2001, 2005a,
2005b; Benigno, 2004; Debrun et al., 2003, 2005; Tsangarides and
Qureshi, 2006; Kabundi and Loots, 2007; Karras, 2007; Corsetti et al.,
2008; Masson, 2008).
But monetary unions create structural changes – in both trade and
risk-sharing – that mitigate the effects of heterogeneous shocks (Frankel
and Rose, 1997, 1998; Baxter and Kouparitsas, 2005; Caldéron et al.,
2007; Houssa, 2008; Inklaar et al., 2008). Thus the success of the
CFA Zone has renewed the interest of African policymakers in mon-
etary unions as a means of regional integration (Tsangarides et al.,
2005; Guillaumont-Jeanneney, 2006; De Grauwe, 2009; Tapsoba, 2009b;
Sénégas, 2010).
197
198 A DSGE Model of Trade and Risk-Sharing Effects
According to theory, monetary unions reduce transaction costs

by increasing exchange-rate stability, stimulating trade, while the
credibility of monetary policy for stable and sustained growth is
also strengthened, facilitating cross-border risk-sharing (Mundell, 1961;
McKinnon, 1963; Kenen, 1969; Rogoff, 1985; Tavlas, 1993; Rose, 2000;
Alesina and Barro, 2002; Frankel and Rose, 2002; Kauffmann, 2004; Rose
and Stanley, 2005).
To analyze the effects of monetary union in the CFA Zone on regional
integration – by increasing intra-regional trade and risk-sharing – a
dynamic stochastic general equilibrium (DSGE) model was estimated,
because it is one of the few suitable macroeconomic models (Bean, 2005;
Gali and Monacelli, 2005; Bean, 2006; Avouyi-Dovi et al., 2007; Peiris
and Saxegaard, 2007; Juillard & Villemot, 2010), although its use with
monetary unions is rare, especially in Africa, and hitherto nonexistent
in the CFA Zone.
The next section describes the DSGE model, parameters for which are
then calibrated. Results are then presented and discussed, and finally
conclusions are drawn.
2 A two-country DSGE model of the CFA ZONE
Trade and risk-sharing between the two CFA regions – the Economic and
Monetary Community of Central Africa (CEMAC) and the West African
Economic and Monetary Union (WAEMU) – were analyzed in a ‘two-
country’ model based on Badarau-Semenescu and Levieuge (2011) but
extended with key features of African economies. The model includes –
separately for each CFA region – the behavior of households in labor
supply, consumption and investment; of firms in both wholesale and
capital goods production; of retailers in price-setting; of commercial
banks in financing; and of the central bank and governments in pol-
icymaking. (Despite the existence of two central banks – BEAC and
BCEAO, one for each CFA region – they will be treated as one because
they maintain a common currency and thus a common monetary
policy.)
As is usual in less developed countries, households own the retail-
ers which – following Calvo (1983) – are assumed to exhibit nominal
price-stickiness. Producers are net-worth constrained, needing to bor-
row to finance production. Commercial banks use central bank credit
(refinancing) as well as deposits from households to finance production.
Risk-sharing is modeled via credit markets for bank loans because finan-
cial markets are not well developed in the CFA Zone. Heterogeneous
Thierry Kame Babilla 199
shocks are introduced in productivity, monetary policy and public

spending.
Householders’ program
We assume a continuum of infinitely-lived households represented by
the unit interval [0, 1], each maximizing the one-period utility function
σC −1 σh −1

∞
σC Ct C
σ
σh Ht h
σ
max Et βk
− (1)
k=0
σC − 1 σh − 1
where C is consumption and H labor hours; β is a subjective discount

factor between zero and one (0 < β < 1); σC (less than zero) is the inverse
intertemporal elasticity of substitution in consumption; and σh (greater
than zero) is the inverse elasticity of the disutility of labor. Consumption
in each region consists of both domestic goods and ‘foreign’ goods (that
is, produced in the other CFA region):
Cγ1 C1−γ
CEMAC, C= 2
(2)
γ γ (1 − γ )1−γ
∗ 1−γ ∗ γ
C1 C2
WAEMU, C∗ = (2∗ )
γ γ (1 − γ )1−γ
where CEMAC consumption (C) consists of both CEMAC goods (C1 )

and WAEMU goods (C2 ), while symmetrically WAEMU consumption
(C∗ ) consists of both WAEMU goods (C∗2 ) and CEMAC goods (C∗1 ). The
preference for consumption of domestic goods – assumed equal in each
region – is expressed by Y ∈ [0, 1].
Assuming the law of one price, prices (Pt ) are
CEMAC, P = P1γ P21−γ (3)

∗ γ
WAEMU, P =P P 2
1−γ
1 (3∗ )

Households supply labor Ht = 1 − lt to, and receive nominal wages
(W
Wt ) from, producers – plus profits (t ) from retailers – which they use
for consumption, taxes, or save, investing (in period t, with payoff in
period t + 1) either in nominal commercial bank demand deposits (A ( t)
or in nominal securities (Dt ) indexed to the real risk-free central bank
interest rate. Faced with an increase in the central bank rate, interest

on securities r D = RD − 1 – where RDt is their gross nominal expected
yield – reacts instantly. So as not to lose deposits, this also requires

banks to change their interest rate r A = RA − 1 , where RAt is their gross
nominal expected yield.
Households pay lump-sum taxes (T Tt ), yielding household budget
constraints
CEMAC, Pt Ct + Pt Dt + At ≤ Pt Wt Ht + At−1 RAt + Pt Dt−1 RDt − Tt + t (4)

∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗ ∗
WAEMU, P C + P D + A ≤ P W H + At−1 R + P D
t t t t t t t t
A∗
t t t−1 R D∗
t − T + ∗t
∗
t
(4∗ )
The first-order condition for intertemporal maximization in CEMAC –

describing optimal behavior WRT Ct , Dt , At and Ht – is
1 − σ1C
λt = C (5a)
Pt t

Pt+1
0 = λt − βRDt+1 Et [λt+1 ] Et (5b)
Pt
0 = λt − βRAt+1 Et [λt+1 ] (5c)
Ht = (λt Pt Wt )σh (5d)
Thus, optimally, households are indifferent between consumption and

savings and there are no arbitrage opportunities between alterna-
tive investments, while labor supply reflects the prevailing preference
between work and leisure; and symmetrically for WAEMU.
From Equation (5b) – and symmetric Equation 5*b – the common
interest rate set by the central bank is optimal when
∗
D Pt+1 Pt+1
Rt+1 Et = RD∗
t+1 E t (6)
Pt Pt∗
from which it follows that
λt+1 λ∗t+1
= ∗ (7)
λt λt
Substituted into Equation (5a) – and symmetric Equation (5*a) – this

yields
Ct = C∗t (t )σC (8)

where t – the terms of trade between CEMAC and WAEMU – is
Pt∗
t = (9)
Pt
Producers’ program
Producers use the Cobb–Douglas constant-returns-to-scale technology
CEMAC, Yt = at Ktα L1−α

t (10)
∗ ∗
∗ α ∗ 1−α
WAEMU, Y t = at K t Lt (10∗ )
where (Y) is total output, (K) is capital, (L) is labor, (at ) is total factor
productivity and (α) is the output elasticity of capital. An autoregressive
productivity (technological) shock (ρa ) is included when
CEMAC, ât = ρa ât−1 + εa (11)

∗ ∗ ∗ ∗
WAEMU, â = ρ â
t a t−1 +ε a (11∗ )

Where ρa is between zero and one (0 < ρa < 1) and εa,t → i. i. d. 0, σεa,t .
Following Badarau-Semenescu and Levieuge (2011) – in order to
ensure consistency in the credit market – labor input is assumed to

Ht ) and producers’ labor HtF so that
consist of both households’ labor (H
1−
Lt = Ht HtF (12)
where is households’ share of total labor. Producers’ labor is remuner-

ated at rate W F .
Investment (IIt ) in physical capital adds to previous capital less depre-
ciation (δ), or
Kt+1 = (1 − δ) Kt + It (13)
However, capital producers – who transform final investment goods into

the physical capital used by wholesale producers – introduce quadratic
capital-adjustment costs
2
φ It
(IIt , Kt ) = −δ Kt (14)
2 Kt
where (φ) measures their elasticity.

Following Badarau-Semenescu and Levieuge (2011), the objective of

wholesales producers – under the constraint of capital accumulation – is
assumed to be maximization of operating cash flows with
ω ω
P1,t P2,t P1,t
ρt = ; ρt = ; and Zt = (15)
P1,t P2,t P2,t
where Pω is producer prices. Retail prices include a mark-up (μt = 1/ρt ).

From Equations 3 and 3* it is easy to find Zt , the terms of trade PP1 =
2
Pt∗
Pt
= t .
The associated Lagrangian is
⎡ ⎤
⎢ ⎥
⎢ ⎥
⎢ 1−γ α 1−α ⎥
⎢∞ 1 ρt+k t+k − Wt +k Ht +k − WtF+k HtF+k ⎥
L = Et ⎢
⎢
at+k Kt+k

Lt+k
⎥
⎥
⎢ k=0
k − It +k + It +k , Kt+k + Qt +k It +k + (1 − δ) Kt+k − Kt+k+1
⎥
⎢ RKk+j ⎥
⎣ j=1 ⎦
k=1
(16)
where Qt +1 is the Lagrange multiplier for accumulation of capital and

Kt+k is the stock of capital. The Lagrangian for the other CFA region is
symmetric except for the terms of trade (t+1 )γ −1 .
Characterizing optimal behavior, profit maximization gives first-order
conditions
Yt
ρt (t )1−γ (1 − α) = Wt (17a)
Ht
Yt
ρt (t )1−γ (1 − α) = WtF (17b)
HtF
∂ (. )
Qt = 1 + (17c)
∂IIt
⎡ 2 ⎤
− φ2 δ 2 − Kt +1
Y I
ρt+1 (t+1 )1−γ α Kt +1 + (1 − δ) Qt +1
⎢ t+1 t+1 ⎥
Et RKt+1 = Et ⎢
⎣
⎥ (17d)
⎦
Qt
and symmetrically for the other CFA zone. The first two conditions
define labor demands, while the third gives Tobin’s Q and the last rep-
resents expected gross return on capital. At optimum, firms’ demand for
capital ensures equality between the expected marginal cost of external
financing and the expected marginal return on capital.
Retailers’ program
Retailers purchase goods from wholesale producers, repackaging and
slightly differentiating them – introducing nominal price-stickiness –
and retailing them in a monopolistically competitive market at price P1,t
which is marked up at rate μt over their wholesale purchase price μ1 P1,t
t
so that their real marginal cost is ρt = μ1 .
t
The total supply of retail goods in the economy – which includes the
value of capital goods used by producers – is the total product of the
economy.
The composite index CES production function and price index are
then

1 ε−1
ε
ε−1
Yt = Yt j ε dj (18)
0
1−ε
1
P1,t = (19)
()
1,t j 1−ε dj
where j designates individual retailers who – following Calvo (1983) –

symmetrically and rationally set nominal prices on a staggered basis.
In each period – regardless of the number of periods (k) since the pre-

vious change – retailers update prices to P̄1,t j with probability 1 − ς
while the remainder (ς ) keep prices unchanged, so that there is proba-

bility ς k ∀ k = [0 ∞] that P1,t j = P̄1,t j . Households, as owners of retail
firms, set prices – constrained by expected market demand – so as to
maximize the discounted sum of expected profits over k periods using
⎧ ! "#
⎪ ∞
⎨ Max ς k Et lt,t+k Yt +k j P j − μ 1 P1,t+k
t+k
−ε (20)
k=0
⎪
⎩ subject to Yt +k j = P1,t (j) C1,t+k + C∗
P
1,t+k 1,t+k
where lt,t+k is the real dividend rate at time (t + k) – subject to price

changes, so it varies over time, and discounted by the risk-free interest
rate – while the dividend payout ratio is 1, that is, all profits are paid out.
Consumption in CEMAC in period (t + k) – of goods produced and
sold in both CFA regions – is then
$ %−ε
P̄1,t
CEMAC, C1,t+1 j = C1,t+k (21)
P1,t+k
$ %−ε
∗
P̄1,t
WAEMU, C1,t+k j = C∗1,t+k (21∗ )
P1,t+k
and symmetrically for WAEMU.

ε
Following Gali and Monacelli (2005), firms apply mark-up ε−1
to real
marginal costs (ρt+k ), yielding first-order condition
∞

Et (βς)k lt,t+k Yt +k j μ 1 P1,t+k
= ε t+k
P1,t
k=0
∞ (22)
ε−1
Et (βς)k lt,t+k Yt +k j
k=0
Because, as discussed, retailers change prices with probability (1 − ς ) in

any period, overall prices are a combination of changed and unchanged
prices
! " 1
1−ε 1−ε
P1,t = ς P1,t−1
1−ε
+ (1 − ς ) P̄1,t (23)
Substituted into (22), this leads – after log-linearization around the

steady state κ = (1−ς )(1−ςβ)
ς
– to the New Keynesian Phillips curve
π̂1,t = βEt [π̂1,t+1 ] + κ ρ̂t (24)

where π1,t = log P1,t /P1,t−1 is the inflation rate in domestically priced
goods; and symmetrically for the other CFA region.
Commercial banks’ program

To finance wholesale production, commercial banks provide total loans
(Bt ) obtained either from the central bank credit (NBt ) or as deposits

from households (A ( t ) – with expected return RBt+1 . Thus

Bt = NBt iRt + At iBt (25)
In each period, wholesale producers use these funds – plus their net
worth (NFt ), which is assumed insufficient to cover the entire cost – to
acquire Kt+1 of physical capital (at price Qt ) to be used during the next
period. Thus
Bt = Qt Kt+1 − NFt (26)
Combining Equations 25 and 26, deposits required from households are
At = Qt Kt+1 − NFt − NBt (27)
In each period, wholesale producers are assumed to have a probability

( ) of remaining in business and a corresponding probability (1 − ) of
exiting. New wholesale producers – entering with the same probability –
are assumed to inherit the net worth of those exiting.

Optimally, producers’ expected return on capital RKt equals their
expected marginal financing cost (the real interest rate plus a credit
premium), or

Pt
Et RKt+1 = t+1 RBt+1 (Bt ) ηt+1 (28)
Pt+1
−ω
NF
where = q Kt+1 describes how the premium depends on the pro-
t t+1
ducers’ leverage ratio (Ben Bernanke et al., 1999), with (ω) being the

elasticity of the premium with respect to leverage and qt the real

price of capital qt = QPt , while (t+1 ) < 0 and (1) = 1. Thus, the credit
premium is an equilibrium inverse function of the aggregate financial
position in the economy, expressed by the leverage ratio.
Equation (28) provides the basis for the financial accelerator. If whole-
sale producers’ net worth goes up, the credit premium and the overall
cost of borrowing fall, yielding greater profits. Demand for capital equi-

librates its marginal return to its marginal productivity mpct – which
is the rental rate of capital – or
mpct + (1 − δ) qt
RKt = (29)
qt−1
where (δ) is the capital depreciation rate and (1 − δ) qt is the value of one
unit of capital. Wholesale producers’ accumulated net worth depends
on previous profits plus inheritance (t ) from exiting producers, which
evolves as
−ω
Pt NFt
NFt+1 = R K
q Kt − R (Bt ) ηt+1
t+1 −1
B
t qt−1 Kt − NFt
Pt+1 qt−1 Kt
+ (1 − ) t (30)
and symmetrically for the other CFA zone.
The central bank’s program

The central bank uses a standard monetary rule to set the short-term
nominal interest rate of the CFA zone (CU)

rr̂tCU = β0 rr̂tCU
−1 + (1 − β0 ) β1 π̂t
CU
+ β2 ŷtCU + εr,t (31)
where β0 ∈ [0 1] is a smoothing parameter; β1 is greater than one (β1 > 1)

and measures central bank response to deviations of inflation from its
steady-state value measured in logs (π̂tCU ); β2 is less than one (β2 < 1)
and measures response to similar deviations in output (ŷtCU ); and εr,t →

i. i. d. (0 σε2r ) represents random monetary shocks, with
1
π̂tCU = π̂t + π̂t∗ (32)
2
1
ŷt =
CU
ŷt + ŷt∗ (33)
2
where π̂tCU and ŷtCU designates, respectively, the log deviations of infla-
tion and output of the CFA zone to the steady-state values. π̂t and π̂t∗ are
the log deviations of inflation respectively of CEMAC and WAEMU to
the steady-state values. ŷt and ŷt∗ represents the log deviations of output
of each of the two economies of the CFA zone to the steady-state values.
where, as before, asterisk designates the other CFA region.
Governments’ program
The two ‘national’ governments – for CEMAC and WAEMU – maintain
independent exogenous autoregressive fiscal policies with public spend-
ing ((ggt ) equal to lump-sum taxes (T
Tt )
CEMAC, Tt = Pt gt (34)
WAEMU, Tt∗ = Pt∗ gt (35)
where ρg between zero and one and εg,t → i. i. d. (0 σεg,t ) are random
public-spending shocks.
Equilibrium of the DSGE model

Equilibrium is achieved when labor, goods and credit markets in both
CFA regions clear.
Goods production in each region must satisfy demand from both
regions for consumption plus domestic investment and government
spending, or
CEMAC, Yt = C1t + C∗1t + It + Gt (36)

∗ ∗ ∗ ∗
WAEMU, Y = C2t + C + I + G
t 2t t t (36∗ )
Substituting in Equations 2 (consumption) and 3 (prices) yields

1−γ ! "
Yt = t2γ −1 Ct γ + (1 − γ ) t C + It + Gt
1−σ
CEMAC, (37)
1−γ ! "
Yt∗ = t2γ −1 C∗t γ + (1 − γ ) t C + It∗ + G∗t
1−σ
WAEMU, (37∗ )
The labor market clears when first-order conditions for labor supply (5d)
and labor demand (15a) are equal. Substituting X into Y yields
σh +1
− σ1
Ht )
(H σh
= (Ct ) h ρt (t )1−γ (1 − α) Yt (38)
and symmetrically for the other CFA region.
Calibration of DSGE model parameters

Parameters needing calibration are organized by agents: households,
producers and retailers (Table 10.1); commercial banks and central bank
(Table 10.2); and governments plus productivity, monetary and public-
spending shocks (Table 10.3). Following Del Negro and Schorfheide
(2008), priors were taken from empirical studies of less-developed
countries or set to commonly used values.
In Equation 1, the subjective discount factor (β) was set identically
for both CEMAC and WAEMU, as found by Diop (2011); the inverse
inter-temporal elasticity of substitution in consumption (σC ) was set
identically as found by Adebiyi and Mordi (2010); and the inverse elas-
σh ) was set identically to the commonly
ticity of the disutility of labor (σ
Table 10.1 Calibration of household, producer and retailer program parameters
Parameter CEMAC WAEMU
Subjective discount factor β 0.98 0.98

Inverse inter-temporal σC 1.81 1.81
elasticity of substitution
in consumption
Inverse elasticity of the σh 0.99 0.99
disutility of labor
Preference for γ 0.4 0.6
consumption of
domestic goods
Probability of unchanged ς 0.31 0.31
price
Output elasticity of capital α 0.40 0.40
Household share of total 0.95 0.99
labor
Capital depreciation rate δ 0.05 0.05
Elasticity of capital φ 6.41 6.41
adjustment cost
ε
Mark-up on real marginal 0.9 0.9
ε−1
costs
208
Table 10.2 Calibration of commercial bank and central bank program

parameters
Parameter CEMAC WAEMU
Ratio of central bank credit to NB/B 0.1 0.15

total commercial bank loans
Ratio of producers’ physical K/NF 0.4 0.4
capital to their net worth
Elasticity of producer’s credit ω 0.002 0.004
premium with respect to
leverage
Producers’ expected return on Rk − RB 0.01 0.01
capital less interest rate on
securities
Producers’ risk-spread ηt+1 0.0001 0.0001
Wholesale producers’ 0.025 0.025
probability of remaining in
business
Monetary smoothing β0 0.51 0.51
Monetary response to inflation β1 1.5 1.62
deviations
Monetary response to output β2 0.5 0.3
deviations
Table 10.3 Calibration of government program plus shock parameters
Parameters CEMAC WAEMU
Consumption/total output C/Y 0.35 0.33

ratio
Public spending/total output G/Y 0.38 0.5
ratio
Investment in physical I/Y 4.3 5.4
capital/total output ratio
Autocorrelation of productivity ρa 0.6 0.5
shocks
Autocorrelation of monetary ρr 0.5 0.5
shocks
Autocorrelation of ρg 0.5 0.4
public-spending shocks
Standard deviation of σεa 0.02 0.01
productivity shocks
Standard deviation of σεr 0.2 0.03
monetary policy shocks
Standard deviation of σεg 0.09 0.005
public-spending shocks
used value. In Equation 2, the preference for consumption of domestic

goods (γ ) was set differently for the two CFA regions, as found by Diop
(2011). In Equation 20, the probability of unchanged price (ς ) – price
stickiness – was again set identically to the commonly used value.
In Equation 10, the output elasticity of capital (α) was set identically
as found by Buffie et al. (2012). In Equation (12) the very high house-
hold share of labor () was set slightly differently as found by Berg et al.
(2012), with producers (capitalists) supplying the very small remainder.
The capital depreciation rate (δ) in Equation 13 and the elasticity of
capital adjustment cost (φ) in Equation 14 were also set identically as
found by Buffie et al. (2012). In Equation 22, retailers’ mark-up on real
ε
marginal cost ε−1 was set to the commonly used value.
In Equation 25 – Table 10.2 – the ratio of central bank credit (NBt )
to total commercial bank loans (Bt ) was set differently for the two
CFA regions as found by Samake (2010). In Equation 26, the ratio of
producers’ physical capital (Kt+1 ) to their net worth (NFt ) was set to the
commonly used value. In Equation 28, the elasticity of producers’ credit
premium with respect to leverage (ω) was set differently as found by
Samake (2010). In Equation 29, the remainder of producers’ expected

return on capital less the interest rate on securities Rk − RB – i.e, their
finance premium – was set to the commonly used value. In Equation,
28, producers’ risk spread (ηt+1 ) was set to the commonly used value.
In Equation 30, the proportion of wholesale producers remaining in
business ( ) was set to the commonly used value.
In Equation 28, the central bank’s monetary smoothing parameter
(β0 ) was set to the commonly used value; its response to inflation devia-
tions (β1 ) was set differently as found by Taylor (1980) and its response
to output deviations (β2 ) was set differently as found by Taylor (1980).
The ratio of consumption (C) in Equation 1 to total output (Y) in
Equation 10 was set differently as found by Araujo et al. (2013); as
was the ratio of public spending (g (gt ) in Equation 34 to total output,
as found by Adebiyi and Mordi (2010), as well as the ratio of producers’
investment in physical capital to total output, as found by Berg et al.
(2012).
Autocorrelation of productivity shocks (ρa ) in Equation 11 was set
differently as found by Araujo et al. (2013). Autocorrelation of mone-
tary shocks (ρr ) in Equation 31 was set to the commonly used value.
Autocorrelation of public-spending shocks (ρ ρg ) in Equation 34 was set
differently as found by Berg et al. (2012).
Standard deviations of shocks were all set differently: of productivity
σεa ) in Equation 11 as found by Araujo et al. (2013); or monetary policy
(σ
σεr ) in Equation 31 and of public-spending (σ

(σ σεg ) in Equation 34 as found
by Berg et al. (2012).
3 Dynamic Models
Modeled effects of a negative productivity shock in the CFA regions

By definition, a negative productivity shock causes a decline in output
(Y) and reduced supply in turn causes price increases (inflation), which
in turn reduces consumption and sharply worsens the terms of trade.
The central bank – whose mandate is price stability – reacts to inflation
by increasing the nominal interest rate, but only slowly, so real inter-
est rates fall. Nevertheless investment declines in direct response to the
negative productivity shock, so imports decline (Figure 10.1).
The shock reduces net worth which increases credit premiums, first on
central bank credit to commercial banks, which flows through to bank
credit to firms. In WAEMU, with higher elasticity of producer’s credit
premium with respect to leverage (ω), output and investment responses
are amplified, though the increase in inflation is less because the invest-
ment decline reduces the gap between supply and demand generally.
These differences in responses reflect lack of thorough integration of
the CFA Zone.
Modeled effects of a negative monetary shock in the CFA Zone

A negative monetary policy shock increases the common nominal inter-
est rate which reduces investment and output – especially in WAEMU –
as well as consumption, causing deflation which improves CEMAC’s
terms of trade with WAEMU. But trade between the two CFA regions is
insufficient to synchronize them so as output returns to normal, terms
of trade do so too (Figure 10.2).
Both central bank and commercial bank credit premiums again rise
much more in WAEMU than in CEMAC, explaining the differences in
investment, output and consumption responses as well as in prices.
Households reduce consumption until higher real interest rates return to
normal. But again the changes in saving are insufficient to synchronize
the two CFA regions.
Modeled effects of positive public-spending shocks in the

CFA regions
Symmetrical positive public-spending shocks in CEMAC and WAEMU
increase output – again, more so in WAEMU than in CEMAC – with
corresponding rises in prices and raised nominal interest rates in
response, while the real interest rate falls like after a productivity
OUTPUT CONSUMPTIOM INVESTMENT
–0.6 –0.6 –0.6
06
–0.65 WAEMU –0.8 –0.8
–0.7 CEMAC –1 –1
–0.75 –1.2 –1.2
–0.8 –1.4 –1.4
–0.85 –1.6 –1.6
–0.9 –1.8 –1.8
–0.95 –2 –2
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
TERM OF TRADE INFLATION NOMINAL INTEREST RATE

1.4 4 0.08
1.2
1 3 0.07
0.8 2 0.06
0.6 1 0.05
0.4
0.2 0 0.04
0 –1 0.03
2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
REAL INTEREST RATE FIRM EXTERNAL FINANCE PREMIUM BANK EXTRNAL FINANCE PREMIUM
×10–3
0.05 0.025 4
0 0.02 3
–0.05
0.015 2
–0.1
–0.15 0.01 1
–0.2 0.005 0
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Figure 10.1 Impulse response functions to a negative productivity shock in CEMAC (dashed) and WAEMU (solid), years
211
OUTPUT CONSUMPTION INVESTMENT
0
212
0 0
–2
–10 –10
–4 WAEMU
–6 CEMAC –20 –20
–8
–30 –30
–10
–12 –40 –40
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10

0.018 0 0.3
0.017 0.2
0.016 –10 0.1
0.015 0
0.014 –20
–0.1
0.013
–30 –0.2
0.012
–0.3
0.011 –40
0 2 4 6 8 10 0 2 4 6 8 10 –0.4
0 2 4 6 8 10
REAL INTEREST RATE FIRM EXTERNAL FINANCE PREMIUM BANK EXTERNAL FINANCE PREMIUM
4 0.55 0.16
3 0.5 0.14
0.45 0.12
2 0.4 0.1
1 0.35 0.08
0.3 0.06
0 0.25 0.04
–1 0.2 0.02
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Figure 10.2 Impulse response functions to a negative monetary shock in CEMAC (dashed) and WAEMU (solid), years
OUTPUT CONSUMPTION INVESTMENT
0.2 –0.6 0.3
WAEMU –0.8
0.15 –1 0.2
CEMAC
–1.2
0.1 0.1
–1.4
0.05 –1.6 0
–1.8
0 –2 –0.1
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
1.4 0.3 0.018
1.2 0.017
1 0.2 0.016
0.8 0.015
0.1
0.6 0.014
0.4 0 0.013
0.2 0.012
0 –0.1 0.011
2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
REAL INTEREST RATE FIRM EXTERNAL FINANCE PREMIUM BANK EXTERNAL FINANCE PREMIUM
×10–3 ×10–4
0.01 0 –2
0 –0.5
–4
–1
–0.01 –1.5 –6
–0.02 –2 –8
–2.5
–0.03 –10
–3
–0.04 –3.5 –12
0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10
Figure 10.3 Impulse response functions to negative public-spending shock in CEMAC (dashed) and WAEMU (solid), years
213
shock, resulting in higher investment. Again CEMAC’s terms of trade

with WAEMU improve. As real interest rates return to normal, so does
investment (Figure 10.3).
Higher prices increase producers’ net worth which reduces credit
premiums, amplifying the positive effects on investment and output.
However, private consumption is crowded out, though less in CEMAC
than in WAEMU. As prices return to normal, so do terms of trade.
4 Conclusions and Policy Implications
Currency unions to promote regional integration through intensified

trade and broader risk-sharing have recently been high on the agendas
of African policymakers. To analyze the trade and risk-sharing effects
of currency union in the CFA zone – the only such union in Africa –
a ‘two-country’ Dynamic Stochastic General Equilibrium (DSGE) model
was calibrated and run with estimated or commonly assumed values for
the two CFA regions, CEMAC and WAEMU in the face of productivity,
monetary or public-spending shocks.
From the simulation, it appears that currency union in the CFA zone
has not fulfilled hopes and expectations regarding intensified trade
and broader risk-sharing and is thus failing in promoting those essen-
tial aspects of regional integration. Despite having a common cur-
rency, trade and risk-sharing between CEMAC and WAEMU remain
low, resulting in lack of synchronization of business cycles across the
regions. One reason is that savings have been insufficient to sup-
port broader risk-sharing, while financial asymmetries have amplified
bilateral differences.
To promote regional integration and overcome the destabilizing
effects of shocks, better risk-sharing institutions and mechanisms are
needed – including easily accessible regional credit markets – with
simultaneous encouragement of greater savings.
References
Adebiyi, M. A. and Mordi, C. N. (2010), ‘Estimating Small Scale
Macroeconometric Model (SSMM) for Nigeria: A Dynamic Stochastic General
Equilibrium (DSGE) Approach’, Central Bank of Nigeria.
Alesina, A. and Barro, R. J. (2002), ‘Currency Unions’, The Quarterly Journal of
Araujo, J., Bin Li, Poplawski-Ribeiro, M. and Zanna, L.-F. (2013), ‘Current Account
Norms in Natural Resource Rich and Capital Scarce Economies’. IMF Working
Paper,
r no. 13/80, Washington, DC: International Monetary Fund.
Avouyi-Dovi, S., Matheron, J. and Fève, P. (2007), ‘DSGE models and their
importance to Central Banks’, Banque de France Bulletin, vol. 161, no. 41.
Badarau-Semenescu, C. and Levieuge, G. (2011), ‘Assessing the Effects of Finan-
cial Heterogeneity in a Monetary Union: A DSGE Approach’, Economic Modeling, g
vol. 28, no. 6, pp. 2451–2461.
Barro, R. and Gordon, D. (1983), ‘A Positive Theory of Monetary Policy in a
Natural Rate Model’, Journal of Political Economy, vol. 91, pp. 589–610.
Baxter, M. and Kouparitsas, M. (2005), ‘What Determines Bilateral Trade Flows?’,
Working Paper Series, WP-05-11, Federal Reserve Bank of Chicago.
Bean, C. (2005), ‘Monetary Policy in an Uncertain World’, World Economics,
Economic & Financial Publishing,g vol. 6, no. 1, pp. 31–53.
Bean, C. (2006), ‘Impact of Globalization on Monetary Policy: Commentary’,
Proceedings Economic Policy Symposium, Jackson Hole, Federal Reserve Bank of
Kansas City, pp. 307–313.
Ben Bernanke, Gertler, S. M. and Gilchrist, S. (1999), ‘The Financial Accelerator in
a Quantitative Business Cycle Framework’, Handbook of Macroeconomics, vol. 1,
pp. 1341–1393.
Benigno, P. (2004), ‘Optimal Monetary Policy in a Currency Area’, Journal of
International Economics, vol. 63, pp. 293–320.
Berg, A., Portillo, R., Yang, S. and Zanna, L.-F. (2012), ‘Public Investment
in Resource Abundant Low-Income Countries’, IMF Working Paper/12/274,
Washington, D: International Monetary Fund.
Buffie, E. F., Berg. A., Patillo, C., Portillo, R. and L.-F. Zanna (2012), ‘Public
Investment, Growth, and Debt Sustainability: Putting Together the Pieces’, IMF
Working Paper,r no. 12/144, Washington, DC: International Monetary Fund.
Caldéron, C., Chong, A. and Stein, E. (2007), ‘Trade Intensity and Business
Cycle Synchronization: Are Developing Countries any Different?’, Journal of
International Economics, vol. 71, no. 1, pp. 2–21.
Calvo, G. A. (1983), ‘Staggered Prices in a Utility-Maximizing Framework’, Journal
of Monetary Economics, vol. 12, no. 2, pp. 383–398.
Carrère, C. (2004), ‘African Regional Agreements: Impact on Trade with or with-
Corsetti, G., Dedola, L. and Leduc, S. (2008), ‘International Risk-Sharing and the
Transmission of Productivity Shocks’, The Review of Economic Studies, Oxford
University Press.
Debrun, X., Masson, P. R. and Pattillo, C. A. (2003), ‘West African Currency
Unions: Rationale and Sustainability’, CESifo Economic Studies, vol. 49, no. 3,
pp. 381–413.
Debrun, X., Masson, P. R. and Pattillo, C. A. (2005), ‘Monetary Union in West
Africa: Who Might Gain, Who Might Lose, and Why?’, Canadian Journal of
De Grauwe, P. (2009), ‘The Fragility of the Eurozone’s Institutions’, Open
Economies Review, vol. 21, no. 1, pp. 167–174.
Del Negro, M. and Schorfheide, F. (2007), ‘Monetary Policy Analysis with
Potentially Misspecified Models’, Working Paper no. 13099, Cambridge, MA:
National Bureau of Economic Research.
Diop, M. B. (2011), ‘Un modèle de prévision de l’inflation à moyen terme dans
l’UEMOA: une approche DSGE’, Direction de la Recherche et de la Statistique de la
BCEAO.
Eichengreen, B. (1992), ‘Should the Maastricht Treaty Be Saved?’, Princeton Stud-

ies in International Finance, International Finance Section, Princeton University,
no. 74.
Fielding, D. and Shields, K. (2001), ‘Modeling Macroeconomic Shocks in the
CFA Franc Zone’, Journal of Development Economics, vol. 66, pp. 199–233.
Fielding, D. and K. Shields, K. (2005a), ‘Do Currency Unions Deliver More Eco-
nomic Integration than Fixed Exchange Rates? Evidence from the Franc Zone
and the ECCU’, Journal of Development Studies, vol. 41, no. 6, pp. 1051–1070.
Fielding, D. and K. Shields, K. (2005b), ‘The Impact of Monetary Union on
Macroeconomic Integration: Evidence from West Africa’, Economica, vol. 72,
no. 288, pp. 683–704.
Frankel, J. A. and Rose, A. K. (1997), ‘Is EMU More Justifiable Ex Post than
Ex Ante?’, The European Economic Review, vol. 41, pp. 753–760.
Frankel, J. A. and Rose, A. K. (1998), ‘The Endogeneity of the Optimum Currency
Area Criteria’, Economic Journal, vol. 108, no. 449, pp. 1009–1025.
Frankel, J. A. and Rose, A. K. (2002), ‘An Estimate of the Effect of Common
Currencies on Trade and Income’, The Quarterly Journal of Economics, vol. 117,
no. 2, pp. 437–466.
Gali, J. and Monacelli, T. (2005), ‘Monetary Policy and Exchange Rate Volatility in
a Small Open Economy’, Review of Economic Studies, vol. 72, no. 3, pp. 707–734.
Guillaumont-Jeanneney, S. (2006), ‘Independence of the Central Bank of West
African States: An Expected Reform?’, Revue d’Economie du Développement, t
vol. 20, pp. 43–73.
Houssa, R. (2008), ‘Monetary Union in West Africa and Asymmetric Shocks:
A Dynamic Structural Factor Model Approach’, Journal of Development Eco-
nomics, vol. 85, no. 1–2, pp. 319–347.
Inklaar, R., Jong-A-Pin, R. and de Haan, J. (2008), ‘Trade and Business Cycle
Synchronization in OECD Countries – A Re-Examination’, European Economic
Review, vol. 52, no. 4, pp. 646–666.
Juillard, M. and Villemot, S. (2010), ‘Multi-country Real Business Cycle Models:
Accuracy Tests and Testing Bench’, Journal of Economic Dynamics and Control,
vol. 35, no. 2, pp. 178–185.
Kabundi, A. and Elsabe, L. (2007), ‘Co-Movement Between South Africa and the
Southern African Development Community: An Empirical Analysis’, Economic
Modelling, vol. 24, no. 5, pp. 737–748.
Karras, G. (2007), ‘Is Africa an Optimum Currency Area? A Comparison of
Macroeconomic Costs and Benefits’, Journal of African Economies, vol. 16, no. 2,
pp. 234–258.
Kauffmann, P. (2004), Intégration Européenne et Régionalisme dans les Pays en
Développement, t Paris: L’Harmattan.
Kenen, P. (1969), ‘The Theory of Optimum Currency Areas: An Eclectic View’,
in R. Mundell and A. Swoboda (eds.), Monetary Problems in the International
Economy, Chicago: University of Chicago Press, pp. 41–60.
Krugman, P. (1993), ‘Lessons of Massachusetts for EMU’, in F. Giavazzi and
F. Torres, (eds.), The Transition to Economic and Monetary Union in Europe,
New York: Cambridge University Press, pp. 241–261.
Masson, P. R. (2008), ‘Currency Unions in Africa: Is the Trade Effect Substan-
tial Enough to Justify Their Formation?’, The World Economy, vol. 31, no. 4,
pp. 533–547.
Masson, P. R. and Patillo, C. A. (2004), ‘A Single Currency for Africa’, Finance and
Development, t December, Washington, DC: International Monetary Fund.
McKinnon, R. I. (1963), ‘Optimum Currency Areas’, American Economic Review,
vol. 53, no. 4, pp. 717–725.
Mundell, R. (1961), ‘A Theory of Optimum Currency Areas’, American Economic
Review, vol. 51, no. 4, pp. 657–665.
Peiris, S. and Saxegaard, M. (2007), ‘An Estimated DSGE Model for Monetary
Policy Analysis in Low-Income Countries’, IMF Working Paper, r no. 07/282,
Washington, DC: International Monetary Fund.
Rogoff, K. (1985), ‘The Optimal Degree of Commitment to an Intermediate
Monetary Target’, Quarterly Journal of Economics, vol. 100, pp. 1169–1190.
Rose, A. (2000), ‘One Money, One Market: Estimating the Effect of Common
Rose, A. K. and Stanley, T. D. (2005), ‘A Meta-Analysis of the Effect of Common
Currencies on International Trade’, Journal of Economic Surveys, vol. 19, no. 3,
pp. 347–367.
Samaké, I. (2010), ‘A Macro Model of the Credit Channel in a Currency Union
Member: The Case of Benin’, IMF Working Paper, r no. 10/191, Washington, DC:
International Monetary Fund.
Sénégas, M.-A. (2010), ‘Bilateral Trade, Colonial Heritage and Common
Currency Arrangement: An Oceanian Perspective’, Economie Internationale,
CEPII Research Center, vol. 129, pp. 63–98.
Tapsoba, J.-A. Sampawende (2009a), ‘Union Monétaire en Afrique de l‘Ouest:
Quelles Réponses à l’Hétérogénéité des Chocs?’, Etudes et Document, t CERDI.
Tapsoba, J.-A. Sampawende (2009b), ‘Trade Intensity and Business Cycle
Synchronicity in Africa’, Journal of African Economies, vol. 18, pp. 287–318.
Tavlas, G. (1993), ‘The “New” Theory of Optimum Currency Areas’, World
Economy, vol. 16, no. 6, pp. 663–685.
Taylor, J. B. (1980), ‘Output and Price Stability: An International Comparison’,
Journal of Economic Dynamics and Control, vol. 2, pp. 109–132.
Tsangarides, G., Charalambos, G. and Van den Boogaerde, P. (2005), ‘Ten Years
after the CFA Franc Devaluation: Progress Toward Regional Integration in
the WAEMU’, IMF Working Paper, r no. 05/145, Washington, DC: International
Monetary Fund.
Tsangarides, G., Charalambos, G. and Qureshi, M. (2008), ‘Monetary Union
Membership in West Africa: A Cluster Analysis’, World Development, t vol. 36,
no. 7, pp. 1261–1279.
Tsangarides, G. and Qureshi, M. S. (2006), ‘What Is Fuzzy about Clustering in
West Africa?’, IMF Working Papers, no. 06/90, Washington, DC: International
Monetary Fund.
Conclusion: Enhancing
Intra-African Trade through
Regional Integration
Introduction
The authors of this book recognize sizeable gaps between Africa’s

highest- and lowest-ranked economies in terms of both implementing
Regional Integration (RI) and reaping the benefits of RI. As Ancharaz
et al. (2011) note, efforts by countries to pursue regional integration
generally face the following three key challenges:
Excessive tariff and non-tariff barriers

Following the 1991 Abuja Treaty, members of Regional Economic Com-
munities (RECs) decided to eliminate tariffs and quotas between mem-
bers and to create a common external tariff. Many African countries
have become members of more than one REC, thereby applying dif-
ferent rules of origin, tariffs and customs procedures. This led to delays,
confusion and rising costs of trade, thereby acting as a further constraint
on RI. However, efforts have been made to establish a free trade area in
26 African countries with the creation of a tripartite agreement between
three of the largest RECs (the Common Market for Eastern and Southern
Africa [COMESA], the East African Cooperation [EAC] and the Southern
African Development Community [SADC]).
Tariffs have decreased, but they are still very high in Sub-Saharan
Africa; and other barriers such as customs procedures; driver- and
vehicle-related regulation; and delays at ports, weighbridges and road-
blocks remain serious impediments to trade. As DfID (2011) notes, trade
policies are more restrictive in Sub-Saharan Africa than anywhere else in
the world. They impede regional trade and trade with global markets.
According to the AfDB (2014b), from a value chain perspective, non-
tariff barriers are the main obstacle to the development of Africa’s
218
growing supply potential into the regional and global markets. This is
why trade facilitation and other measures to reduce transport costs and
increase supply capacity are so vital.
Heavy market and product concentration

Africa’s exports remain too heavily focused on natural resources,
while its share of world trade also remains low. A reported 75% of
exports from the Sub-Saharan region, and over 90% from Algeria,
Comoros, Mali, Nigeria, São Tomé and Principe and Sudan, com-
prise only five commodities (World Bank, 2010). African countries
lack the industrial capacity to innovate, absorb and effectively use
new technologies in order to diversify manufactured goods and move
toward higher-quality products. Intra-African trade is particularly hin-
dered by cumbersome border administration, limited use of information
and communication technologies (ICTs) and persistent infrastructure
deficit.
Weak competitiveness in exports

Africa’s trade competitiveness lags behind other emerging regions due
to a combination of factors, including protectionist import policies,
bad-quality institutions and low levels of infrastructure, education and
technology adoption. Many Sub-Saharan economies still lack adequate
telecommunication infrastructure, power supply, legal systems and skills
(World Bank, 2010). Weak transportation systems and limited power
supplies push up trading costs, exacerbating difficulties for firms to
engage in trade (DfID, 2011). Africa’s infrastructure deficit is certainly
the most serious impediment to RI, especially given the constraining
physical and economic geography of Sub-Saharan Africa – characterized
by fragmented markets, many small countries and 15 landlocked
economies.
The contributors to this book offer valuable policy recommenda-
tions aimed at enhancing regional trade and addressing the three main
constraints mentioned above. The first recommendation relates to the
role of institutions, notably in implementing regional trade agree-
ments (Section 2). Section 3 deals with transport infrastructure and
regional development corridors. Section 4 tackles Africa’s industrial-
ization and building regional value chains; according to the authors,
boosting private sector development in Africa will certainly encour-
age regional trade development. Then, Section 5 presents some pol-
icy measures that would enhance the development of the private
sector.
220 Conclusion
1 Highlighting the role of institutions
African RECs have set ambitious targets toward achieving full integra-
tion. But in practice, RECs suffer from delays in the ratification of
regional treaties and protocols by member states, gaps in the imple-
mentation of regional trade agenda and non-compliance with regional
arrangements (Erasmus, 2011).
The role of institutions is crucial in both implementing regional trade
agreements and in improving competitiveness and the cost of doing
business in Africa. For Joseph Parfait Owoundi (Chapter 1), the coun-
tries that form the Economic and Monetary Community of Central
Africa (CEMAC) would enhance their competitiveness and register better
macroeconomic performances by improving their business environ-
ment, institutional integration and regulatory frameworks. This requires
goodwill from political leaders, commitment from the nation and will-
ingness to surrender some national sovereignty to the regional level.
The absence of sanctions or an incentive mechanism is one of the most
important reasons for the lack of commitment from African govern-
ments for the regional trade integration agenda. Most countries fail to
clearly see what is in it for them to promote and implement a regional
integration agenda.
According to Wumi Olayiwola, Evans Osabuohien, Henry Okodua
and Oluyomi Ola-David (Chapter 2), further incentives are needed to
enhance the implementation of the Economic Community of Western
African States’ (ECOWAS) Agricultural Policy (ECOWAP) and Trade Lib-
eralization Scheme (ETLS), in order to enhance the agricultural exports
within the region. ECOWAS members should be committed to the eco-
nomic integration agenda and especially to reducing bureaucratic delays
associated with preparing export documents. The authors also advo-
cate strengthening the institutional framework to curb the threat of
corruption and ensure adherence to the rule of law.
Also looking at the ECOWAS zone, Benjamin Ndong and Sokhana
Diarra Mboup (Chapter 9) argue that Regional Trade Agreements (RETs)
within the frame of a monetary union significantly trigger bilateral
trade. Therefore, they recommend that ECOWAS evolves toward a wider
monetary union in order to increase regional trade. However, the cur-
rency union should be complemented simultaneously by more credible
commitments at the institutional level and structural reforms aimed at
promoting complementarity between the ECOWAS economies. Bad gov-
ernance and informal trade would also need to be tackled to increase
regional trade. For Thierry Kame Babilla (Chapter 10), currency unions
failed to sustain RI in the CFA Franc zone because of the lack of insti-
tutions and mechanisms for sharing risk within the zone. Financial
asymmetries have led to the widening of national differences among
members countries.
Edris Seid (Chapter 5) recommends that countries adopt and imple-
ment coherent and coordinated trade policies to promote intra-regional
trade. In particular, simplifying custom procedures makes the transit
system more efficient, faster and less costly for both importers and
exporters. This argument is corroborated by the World Bank (2011) for
which a well-functioning legal system is also a key requirement for con-
tract enforcement, enabling firms to expand their markets and networks.
Appropriate business environments facilitate risk and cost management.
The private sector, whether foreign or domestic, is more likely to commit
to new projects or the expansion of existing ones in a business-friendly
environment.
2 Investing in Africa’s infrastructure and regional

development corridors
Désiré Avom and Mouhamed Mbouandi Njikam (Chapter 4) as well as

Seid (Chapter 5) show there is a significant intra-regional trade potential
in Sub-Saharan Africa. Based on their studies of the Economic Commu-
nity of Central African States (ECCAS) zone and of COMESA, ECOWAS,
SADC and the Inter-Governmental Authority on Development (IGAD),
they argue this trade potential grows with geographical proximity. The
longer the distance between countries, the higher are transportation
costs incurred, and as a result, bilateral trade is reduced.
The authors compare regional trade potential to the low level of for-
mal intra-regional trade, which is largely constrained by undiversified
exports and weak regional infrastructure links. Logistics performance
and trade-related infrastructure affect trade volume, competitiveness,
capacity to attract Foreign Direct Investment (FDI) and the reliability
and predictability of the supply chain. These findings are corroborated
in studies by Arvis et al. (2010) and the World Bank (2013) which
show how poorly Sub-Saharan Africa performs in logistics performance
indices. For instance, in 2012, the region was estimated to have the high-
est average lead time to export (over 9.3 days) and to import (12.9 days)
(World Bank, 2013).
According to Lanre Kassim (Chapter 3), trade reforms should be aimed
at building efficient export incentive schemes, including the reduction
of transportation taxes and removal of duty exemptions on imports. His
222 Conclusion
view is corroborated by others authors (Seid [Chapter 5], Ndong and

Mboup [Chapter 9] and Ibrahima Camara [Chapter 8]) who recommend
improving the level of infrastructure in the region as a support to trade
reforms. Investing in transportation, ICTs and energy supply is essential
for both regional trade and access to global markets. It impacts signifi-
cantly on the cost of doing business, productivity and competitiveness
and ultimately affects trade volumes. It also helps in distributing wealth
more equitably, at both national and regional levels.
Developing corridors in the transport, energy, or ICT sectors is a
response to the fragmentation, isolation and high transit costs limit-
ing RI in Africa (see Map C.1). Corridors aim at fostering trade and
market integration by opening up economic hubs and giving access to
ports and are also a tool to reduce poverty. Meanwhile, RI could help
Map C.1 Regional development corridor projects in Africa (2013)

Source: Authors based on Programme for Infrastructure Development in Africa (PIDA).
Africa’s countries to fund trans-boundary infrastructure and support

cross-border trade (Map C.1).
3 Promoting industrial policy, economic diversification and

regional value chains
The lack of diversification is a weakness for Africa’s regional trade.

There is a need for renewed industrial policy in Africa, given its low
level of manufacturing production and the continent’s increasing con-
centration on exports of primary commodities. In so doing, several
region-specific recommendations are provided. Owoundi (Chapter 1)
recommends that CEMAC countries move away from the trade of pri-
mary products and look for more efficient and diversified production
processes and innovation to be more competitive at the international
level. Indeed, while manufacturing has forward and backward linkages
with other sectors generating positive externalities for the economy,
primary commodities are often poorly linked to the local economy,
in addition to being vulnerable to the world price volatility. Africa
should also invest in industrialization and manufacturing, includ-
ing agro-industry, because these sectors have a higher employment
elasticity.
Avom and Njikam (Chapter 4) acknowledge the need for ECCAS mem-
ber countries to strengthen the industrial policy of the Community
and to reorganize the production structures so as to achieve trade com-
plementarity. The authors recommend diversification, industrialization
and commercial complementarity in the region. They also suggest the
effective participation of ECCAS countries in the WTO, as the latter
will provide them with an initial framework for deeper regional coop-
eration. According to Camara (Chapter 8), the West African Economic
and Monetary Union (WAEMU) stands to gain from diversifying its
various economies and stepping up complementarity between them.
WAEMU countries should also develop an industrialization policy based
on import substitution through the local processing of raw materials.
According to Henri Atangana Ondoa and Tabi Henri Ngoa (Chapter 7),
African countries should consolidate their trading relationships with
the United States and the European Union rather than with other
African or transition countries because trade with developed countries
enables transfers of technology and contributes to industrial develop-
ment. In contrast, intra-African trade has no major impact on indus-
trialization in Africa and discriminatory protectionist policies cause
de-industrialization in some African countries.
224 Conclusion
The analyses also reveal the development of regional value chains is

another means to encourage industrialization and attract FDI. Africa’s
inability to create regional production chains is a major impediment
to the development of regional integration in the continent. Owoundi
(Chapter 1) and Hopestone Kayiska Chavula (Chapter 6) argue that
regional integration, which allows larger markets and greater economies
of scale, is cost-efficient and thus encourages firms to develop value
chains to produce more competitive products.
This is corroborated by the analysis in the African Development
Bank’s Regional Integration Strategy (AfDB, 2014a). Indeed, for the AfDB
(2014b), the new African industrialization agenda must focus on the
development of regional value chains. Similarly, the African Union
(2012) argues that intra-African trade would be stronger with the devel-
opment of regional enterprise and value chains. The importance of
regional value chains is also stressed by the OECD/WTO (2013), which
finds that agroindustry offers a major opportunity for Africa to build
regional value chains, especially given rising demand resulting from
urbanization and population growth.
4 Supporting the private sector
Chavula (Chapter 6) and Ondoa and Ngoa (Chapter 7) argue that

African economies need to develop a strong local industry, as it is a
means to process raw materials, depend less on the volatility of world
prices for the exploitation and export of natural resources, produce com-
modities with higher added value and develop more linkages with the
local economy and encourage services. However, they point out that
Sub-Saharan Africa experienced de-industrialization in the past three
decades, resulting from its private sector’s lack of competitiveness in the
global market.
In this context, the authors recommend the private sector focus on
regional trade to gradually develop a cost-effective and competitive
industry. African governments should involve more private enterprises
in regional integration processes, provide them with access to finance
and invest in education and skills. This would facilitate Africa’s indus-
trialization and regional trade. The view that supporting the private
sector is a good means for boosting regional trade is corroborated by
UNECA (2011).
Although the private sector is a key driver of Africa’s growth and
regional integration would significantly support business development
and FDI, the regional integration process in Africa is largely state-driven
and the private sector is mostly absent from the negotiations around
and design of regional integration arrangements (McCarthy, 2007). The

promotion and further engagement of the African private sector in the
regional integration process could be fostered through the following
channels:
Providing access to affordable finance

The private sector in Africa suffers from a lack of investment and inade-
quate access to finance, in particular risk capital. In Sub-Saharan Africa,
firms mainly finance their operations using internal sources (78.3%
of total resources).1 Another means for boosting Africa’s industry and
trade is to provide them with adequate financial resources. For Babilla
(Chapter 10), the African countries which are member of a currency
union should develop regional financial markets and facilitate access
to finance. Owoundi (Chapter 1) also argues that FDIs in Africa con-
stitute another significant type of financial flows for African countries,
but these are mostly directed toward natural resource exploitation.
FDI should rather be directed toward more productive sectors which
would generate forward or backward linkages with other sectors in
the local economy, be more labor-intensive and less vulnerable to the
volatility of world prices than primary commodity exports. The export
of more productive sectors could be supported by incentives such as tax
rebates and a better investment environment in the continent.
Investing in education and human capital

Education has a positive impact on the development of the private sec-
tor, as higher educational attainment increases productivity (Kumar,
2006). This applies not only to the manufacturing and service sectors,
but also to agriculture, thereby supporting the structural transformation
of economies (Sackey, 2011). According to Chavula (Chapter 6), empha-
sis should be placed on technological transfer to enhance innovation
and lead to the production of higher-quality products. However, how
much of a positive impact these imported technologies bring depends
on the country’s existing technological capacity to acquire, assimilate
and utilize new technologies. This calls for the improvement of the
African education system in science and technology, in order to broaden
innovation skills at the national level (Figure C.1).
The deficit in education and skills in Sub-Saharan African economies,
which results in the low productivity of firms stemming from labor
mismatch, is a major impediment to the development of the pri-
vate sector and partly accounts for the lag in industrial development
on the continent. UNECA (2011) indicates that small and medium
enterprises (SMEs) have very low levels of technology, while the few
226 Conclusion
Institutions
7
Innovation Infrastructure
6
5
Business Macroeconomic
sophistication 4 environment
3
2
Health and
Market size 1 primary
education
Technological Higher education

readiness and training
Financial market Goods market

development efficiency
Labor market efficiency
Africa S Latin America and the Caribbean
Figure C.1 Africa’s comparative performance

Source: World Economic Forum, World Bank and AfDB (2013).
large African enterprises have almost non-existent research and develop-

ment. The Africa Competitiveness Report 2013 also corroborates Chavula’s
(Chapter 6) recommendation, as it highlights the urgent need to bring
Africa up to higher levels of competitiveness, including with regard
to primary education and higher education and training, as well as
technological readiness, which refers to the flexibility to adopt exist-
ing technologies. Africa’s underperformance is particularly problematic
given the pressing need to move up the value chain from the natural
resource sector into more-advanced private manufacturing and services
sectors and have a workforce that can move beyond simple production
processes.
5 Conclusion
RI displays features and opportunities that could facilitate the

achievement of both political stability and economic development in
Africa. Yet the continent lacks the adequate institutions and regula-
tions to improve the overall business environment and facilitate trade.
Moreover, hard infrastructure, in particular transport, connectivity and
energy constitute a pervasive long-term challenge. Sub-Saharan Africa’s
deficit in infrastructure largely explains its low levels of intra-African
trade.
Furthermore, there are a myriad of challenges at the firm level which
limit the development of the private sector and the emergence of
high-quality exports. Referring to the industrialization policy, African
countries should move from an inward to an outward focus so as to
encourage competition and generate regional industrial capacity. Value
chain-related activities constitute an important driver of RI. African
countries need to develop regional markets and trade and to achieve
a relevant structural transformation of their economies through the
production of higher-value commodities. Improved access to afford-
able finance and skills development and training would greatly support
innovation and industrial technology. Mobilizing domestic resources
would be a first step for African countries to exploit their comparative
advantages and enter regional and then global value chains.
Finally, for RI to be more successful in Africa, political leadership is
needed to champion the process at the national level. African countries
should also reduce the disparity between actors with regards to regional
policy goals, the implementation process and the benefits from regional
trade. For instance, in the ECOWAS region, Nigeria, Côte d’Ivoire and
Senegal account for almost 90% of intra-regional exports and 50% of
intra-regional imports (UNCTAD, 2011).
Note
1. www.enterprisesurveys.org, accessed 21 April 2014.
References
African Development Bank [AfDB] (2014a), Regional Integration Strategy 2014–
2023, Tunis: AfDB.
African Development Bank [AfDB] (2014b), African Development Report 2013/14:
Leveraging Regional Integration for Inclusive Growth, Tunis: AfDB.
African Union (2012), Action Plan for Boosting Intra-Regional Trade. Available at
http://www.au.int/ [Accessed 23 May 2014].
Ancharaz, V., Mbekeani, K. and Brixiova, Z. (2011), ‘Impediments to Regional
Trade Integration in Africa’, Africa Economic Brief,
f vol. 2, no. 11, pp. 1–8, African
Development Bank.
228 Conclusion
Arvis, J. -F., Duval, Y., Shepherd, B. and Utoktham, C. (2010), ‘Trade Costs in
the Developing World 1995–2010’, Policy Research Working Paper, r no. 6309, The
World Bank.
Department for International Development [DfID] (2011), ‘Regional Integration
and Trade in Sub-Saharan Africa’, Trade and Investment Analytical Papers, DfID:
Department for business Innovation and Skills.
Erasmus, G. (2011), ‘Deeper Regional Integration in SADC. Will EPAs Undermine
the Process?’ Working Paper,r Trade Law Centre for Southern Africa.
Kumar, C. S. (2006), ‘Human Capital and Growth Empirics’, Journal of Developing
Areas, vol. 40, no. 1, pp. 153–179.
McCarthy, C. (2007), ‘Is African Economic Integration in Need of a Paradigm
Change? Thinking Out of the Box on African Integration’, in A. Bösl et al.
(eds.), Monitoring Regional Integration in Southern Africa, Yearbook Vol. 6–2007.
Stellenbosch: Trade Law Centre for Southern Africa.
OECD/WTO, Aid for Trade at a Glance 2013: Connecting to Value Chains.
WTO/OECD Publishing, 2013.
Programme for Infrastructure Development in Africa (PIDA) (2014), website,
section ‘PIDA sectors’. Available from http://www.au-pida.org/ (Accessed
23 May 2014).
Sackey, H. A. (2011), Human Capital Development in Africa: Agents, Drivers and
Implications for Growth and Structural Transformation, Addis-Ababa: UNECA.
United Nations Conference on Trade and Development [UNCTAD] (2011),
Strengthening Regional Economic Integration for Africa’s Development,t UNCTAD,
Division for Africa, Least Developed Countries and Special Programmes,
Geneva.
United Nations Economic Commission for Africa [UNECA] (2013), Economic
Development in Africa Report – Intra-African Trade: Unlocking Private Sector
Dynamism, Addis-Ababa: UNECA.
United Nations Economic Commission for Africa [UNECA] (2011), ‘Industrial
Policies for the Structural Transformation of African Economies: Options
and Best Practices’, Policy Research Paper, r no. 2, United Nations Economic
Commission for Africa, Addis Ababa, Ethiopia.
World Bank (2008), World Development Report 2008. Washington, DC:
World Bank.
World Bank (2010), World Development Indicators, Washington, DC: World Bank.
World Bank (2011), Doing Business 2011: Making a Difference for Entrepreneurs,
Washington, DC: World Bank Group.
World Bank (2013), World Development Indicators, Washington, DC: World Bank.
World Economic Forum, World Bank and AfDB (2013), The Africa Competitiveness
Report,
t Geneva: World Economic Forum.
Index
Note: The letter ‘n’ ‘t’ ‘f’ following locators refers to notes, tables and figures
respectively.
Abuja Treaty, 218 Armington assumption, 98

ACP, see African, Caribbean and Pacific Association of South East Asian
ADMARC, see Agricultural Nations (ASEAN), 72, 96, 88n1
Development and Marketing Australia-New Zealand trade
Corporation agreement with ASEAN, 96
Africa Competitiveness Report 2013,
The, 226 balance of payments, 47, 64, 113
African, Caribbean and Pacific, 34 BCEAO, see Central Bank of West
African Development Bank, 224 African States
African Financial Community Franc, Bonaparte, Napoleon, 153
153–4, 158–60, 164 BOP, see balance of payments
African Growth Opportunity Act
border effect, 175, 194n4
(AGOA), 77, 114
business cycles synchronization, 214
Agricultural Development and
business performance observation, 22
Marketing Corporation, 114
Agriculture, 6, 20, 31–44, 112–14,
117, 225 capital-intensive firms, 120
agricultural exports, 38, 41 CAR, see Central African Republic
agriculture employment, 32 Cartagena Agreement, 88n2
agro-industry, 166, 223 CCCN, see Customs Cooperation
baseline model for exports, 36 Council Nomenclature
employment, 32 CEN-SAD, see Community of
food security, 32–3 Sahel-Saharan States
hindrances to trade of agricultural Central African Economic and
products, 32 Monetary Union, 179, 194n3
intra-regional trade analysis in Central African Republic, 20, 24, 76–7,
maize, rice and wheat, 97 88n4, 99
intra-African import, 4 Central Bank of West African States,
labor force, 6, 32, 129, 133, 135, 22, 154
138, 140, 146 Centre for Prospective Studies and
liberalization of, 113, 114, 117, 168 International Information, 29n1
for poverty alleviation, 32 CEPII, see Centre for Prospective
AMU, see Arab Maghreb Union Studies and International
Andean Pact, 88n2 Information
see also Cartagena Agreement CFAF, see African Financial
Anderson-van Wincoop gravity Community Franc
model, 98, 100, 103–5 CFA Zone, 13, 197–8, 202, 205–6,
Arab Maghreb Union, 92, 178 210, 214
Arab Monetary Union, 171n1 CIS, see Commonwealth of
Arellano-Bond test, 41 Independent States
229
230 Index
Clemente-Montañés-Reyes unit-root control of corruption, 42

test, 112, 117, 123 control-of-corruption
CMA, see Common Monetary Area index, 38
Cobb–Douglas control variables, 131, 157, 161,
constant-returns-to-scale, 201 163, 182
COMESA, see Common Market for cost differentials, 22
Eastern and Southern Africa currency unions, 11–13
common currency, 11, 13, 80, 153–9, effect of WAEMU, 12
162, 181, 194n8, 198, 214 effects on trade, 11
common external tariff, 87, empirical evidences, 12
154, 194 impact of, 12
common market for Eastern and positive impact, 12
Southern Africa, 8, 91–2, 94, 97, theoretical evidences, 12
103–5, 171n1, 178, 218, 221 customs barriers, 167
common monetary area, 11 Customs Cooperation Council
Commonwealth of Independent Nomenclature, 117
States, 175–8 customs duty, 137, 142
Community of Sahel-Saharan customs unions, 17, 31, 94, 162
States, 92 cyclical effects, 148n4
comparative advantages, theory of, 1
competitiveness, 1–2, 5, 17–29, 31, 33, data on trade statistics, 165
45, 48–9, 51, 71, 100, 114, 137, see also International Monetary
142, 154, 219–21, 222, 224, 226 Fund
competition structure, 25–6 de-industrialization, 1–2, 8–11, 125–8,
contribution to trade balance 131–2, 137, 142–3, 223–4
indicators (CTBi), 27t advent of, 10
data and indicators, 18 competition from imported
definition, 19 products, 143
export competitiveness, 1, 114 manufacturing sub-sector, 142
financial crises, 25 negative, 132
horizontal integration, 19–20 structural adjustment programs, 137
importance of, 28 in the South, 128
indicator of, 19 Djaowe, Joseph, 21
objectives, 18 domestic income growth, 65
prime importance, 18 DOTS, see data on trade statistics
revealed comparative advantages DPM, see dynamic panel model
(RCA), 18–19, 26t DSGE model, 13, 198–210, 214
trade balance, 26–8 central bank credit
trade integration, 19–20 (refinancing), 198
weak competitiveness in central bank’s program, 205–6
exports, 219 commercial banks’ program,
vertical integration, 19 204–5
constant-elasticity-of-substitution equilibrium of, 206–7
preferences, 98 governments’ program, 206
Constitutive Act, 72 householders’ program, 199–201
contribution to trade balance (CTB), parameters calibration, 207–10;
5, 18, 23–4, 28 autocorrelation of productivity
contribution to trade balance shocks, 209; commercial bank
indicator (CTBi), 18, 27t and central bank, 208t;
Index 231
government program plus Economic Community of West

shock parameters, 208t; African States, 2, 6, 8, 11–13, 21,
household, producer and 31–5, 37–9, 41–3, 73–5, 91–4,
retailer, 207t; monetary 103–5, 153, 155, 159, 165, 168,
smoothing parameter, 209 171n1, 173–5, 177–8, 181, 185,
producers’ program, 201–2 188–9, 220–1, 227
retailers’ program, 203–4 African exports and imports, 177t
risk-sharing, 198 agricultural policy (ECOWAP), 6, 33,
dummy variable, 7, 12, 48, 54, 59, 61, 37, 220
83, 98–9, 101, 158, 161–4, average telephone density, 31
181–3, 185 bilateral food trade, 21
dynamic models, 210–14 bilateral trade, 159
negative monetary shock, border effect, 188
210, 212f common currency, 154
negative productivity shock, 210, common market, 11, 18, 33, 72, 87,
211f 91, 94, 126, 154, 171n1, 178,
positive public-spending shocks, 194n7
210–14 currency constraint in cross-border
symmetric positive public-spending payments, 174
shocks, 213f customs union, 33
dynamic panel model, 37, 48, 53 effects on regional integration,
Dynamic Stochastic General 39t–40
t t
Equilibrium (DSGE) model, see export effort, 178–81
DSGE model export trends, 176
exports and imports, 31, 32f, f 193t
East African Community (EAC), 92, exports of the economic zones, 180t
97, 104, 218 focus of, 33
ECCAS, see Economic Community of food importer, 7, 31–3
Central African States free trade area, 43
econometric model, 35–7 global trade 2000–2011, 175–7
econometric problems, 100, 131–2 impact of economic crisis, 177
Economic and Monetary Community import trends, 176f
of Central Africa, 2, 5, 178, intra-African trade by economic
198, 220 zone, 179t
Economic Community of Central intra-community trade, 175–81
African States, 2, 5, 7, 28, 71, intra-regional agricultural
73–8, 80, 83–8, 92, 171n1, 178, exports, 42
221, 223 intra-regional exports and
average trade flows, 84t–5t t imports, 32f
bilateral trade, 81 intra-regional trade, 31
conflicts, 87 liberal trade policy, 32
distance impedes trade between, 85 non-tariff barriers, 41
dynamic space for trade, 87 objective of, 32
economies of scale, 83 primary products dependency, 178
institutional difficulties, 87 single currency, 188
intra-community trade, 87 sub-groups, 194n8
trade flows, 83 trade agreement, 188
trade potential, 74, 84, 86t trade facilitation, 6, 38, 42, 174
232 Index
Economic Community of Central FGLS, see feasible generalized least

African States – continued squares
trade within Africa and ECOWAS, financial crises, 22, 25
177–8 first-order differences, 132
weakness of intra-community fiscal policy, 154
trade, 178 fixed-impact panel model, 163
weakness of intra-regional Food and Agricultural
trade, 178 Organization, 62
economic crises, 22, 25, foreign direct investment, 2, 20, 97,
177, 185 221, 224–5
economic diversification, 1–3, foreign markets access, 1
223 Franc devaluation, 21
economic downturn, 128 free and fair market conditions, 19
economic partnership agreements, free trade area, 11, 31, 72–3, 92, 94,
21, 34 114, 162, 218
economies of scale, 9–10, 17, 33, 83, FTA, see free trade area
91, 119–20, 122, 127, 224 F-tests, 41
ecosystem diversity, 28
ECOWAS, see Economic Community GATT, see General Agreement on
of West African States Tariffs and Trade
ECOWAS Trade Liberalization Scheme GCI, see global competitiveness index
(ETLS), 6, 33, 43, 220 General Agreement on Tariffs and
EMU, see European Monetary Union Trade, 73, 77, 173
energy deficit, 125, 143 generalized method of moments, 6,
EPAs, see economic partnership 37–8, 40–2, 49, 53–5, 57, 59–60,
agreements 62, 131
estimated regression model, 41 global competitiveness index, 24, 25t
estimation techniques, 37–8 global economic environment, 25, 72
fixed-effects estimation, 37 globalization, 3, 17, 174
generalized method of moments GMM (dynamic) models, 40
(GMM), 37–8 GMM, see generalized method of
Hausman test, 37 moments
instrumental variable (IV) gravity model, 2, 7–8, 12, 34–5, 74,
estimation, 37 78–82, 87, 92, 96–100, 102–5,
robustness, 37 155, 157–8, 160–2, 174–5, 181,
two-stage least squares (TSLS), 37 185
EU-ACP agreements, 174 circle distance, 79
European Monetary Union, CMA on bilateral trade, 162
162–3 COMESA and SADC bilateral
exchange rate volatility, 162 trade, 96
export duties, 62, 65 constrained utility
export growth regression, 61 maximization, 78
export promotion strategies, 9 consumers’ homothetic
exports of goods, trend in, 190t preferences, 98
exports-to-GDP ratio, 49 countries included, 105t
external tariffs, 5, 11, 94 country-specific variables, 100
cross-section estimates, 100
FDI, see foreign direct investment cross-section specifications of, 96
feasible generalized least squares, 117 dependent variable, 101
Index 233
distance measurement, 79 industrialization, 2, 7–11, 91, 111–2,

distance variable, 79 115, 119, 121, 126–8, 130, 132,
estimation method, 82 137, 142–3, 168, 219, 223, 227
fixed-effects specifications of, 96 absolute terms, 130
foundations of, 97–8 credit ceilings, 137
goods supply specialization, 78 customs duties, 137, 142
impact of ECOWAS, 174 domestic demand in, 137
information costs, 79 external factors, 133t–4
t tt, 137–42
measures of trade facilitation, 34 import substitution, 112–13
model specification, 98–101
import restrictions
monopolistic competition, 98
industrial value added, 146t–7t t
multilateral resistance, 99
internal factors, 130, 132–7
PPML estimation of, 102t
profit maximization, 78 modernity level, 131
residual estimation, 80 manufacturing industries and trade,
specification and basis of variables, 144t–5
t t
79–80 price controls
theoretical basis, 78–9 protection of local businesses
trade potential, 80–1 rate, 128, 130–2, 142–3
transaction costs, 79 reduction in customs duty, 142
two-step approach, 81 relative terms, 130
variable, 8, 101t schooling and, 137
Grubel–Lloyd index, 20–1 strategy, 8–11
substitution strategies, 137
Harmonized System (HS), 117 technology-intensive goods
Hausman test, 37, 52, 57, 117 demand, 142
Heckscher-Ohlin (H-O) model, 97 trade openness, 146t–7t t
Herfindahl-Hirschman index, 121 urbanization in, 137
heterogeneous shocks, 197–8 industrial value added and trade
openness, 146t–7
t t
Ibrahim Index of African instrumental variable (IV), 6, 37,
Governance, 24 42, 53
IGAD, see Inter-Governmental integration, 13, 17, 31, 71–2, 197
Authority on Development integration
IIAG, see Ibrahim Index of African
bid to revive, 154
Governance
capital flows, 72
IMF database, 128, 142
forms, 72
see also UN Comtrade database
IMF, see International Monetary Fund interbank payment system, 194n2
import duties, 57–8, 62, 64–5 Inter-Governmental Authority on
import growth Development, 8, 92–4, 103–5, 221
preliminary information, 50 International Monetary Fund, 9, 47,
regression with foreign aid, 66t 65, 101, 113, 128, 142, 165
imports of goods, trend in, 191t International Standard Industrial
import substitution strategy, 8–10 Classification, 117
import tariffs, 47, 113 International Trade Centre, 71
income elasticity of demand, 48–9, international trade, theory of, 29n1
51–4, 57–8, 60–3 intertemporal maximization, 200
234 Index
intra-African trade, 3–7 one-crop farming members, 74

affordable finance, 227 product structure, 76–7
agricultural imports, 6 total trade within communities, 75f
comparative performance, 226 trade levels, 74–6
competitive industry trade performance comparison, 75
development, 224 ISIC, see International Standard
competitiveness, 5 Industrial Classification
cost-effective, 224 islandness, 186
during periods of recession, 3 isolation, 13, 83–4, 222
duty exemptions removal, 221 Israel-USA free trade agreement, 96
economic diversification, 223–4 ITC, see International Trade Centre
by economic zone, 192t
export incentive schemes, 221 labor-intensive goods, importation of,
growth, 3 126, 137
heavy market, 219 labor productivity, 125
industrial policy, 223–4 Lagrange multiplier, 202
infrastructure investment, 221–3 Lamy, Pascal, 71
institutional integration, 5
landlockedness, 182, 184, 186, 188
institutions role, 220–1
logistical advances, 17
logistics performance, 221
log-linearization, 100, 204
non-fuel exporters, 3
product concentration, 219
regional development corridors, macro-economic framework, 28
221–3 macro-economic indicators, 89n7
regional value chains, 223–4 macro-economic policy, 154
skills development, 227 Malawi, 111–23
supporting the private sector, 224–6 Agriculture (General Produce)
tariff and non-tariff barriers, 218–19 Act, 113
trade between 1995 and 2011, 4f agriculture liberalization, 114
trade by selected RECs, 5f balances of payments, 113
trade partners, 4f base surtax reduction, 114
trade reforms, 221 Control of Goods Act, 112
trade related infrastructure, 221 development policies
transportation taxes reduction, 221 statement, 112
weak competitiveness in economic crisis, 111
exports, 219 exchange rate regime, 114
intra-ECCAS trade, 74–8 export intensity, 116
average trade within export-led agricultural
communities, 76t development, 111
ECCAS exports and imports, 77t export promotion, 114–15
ECCAS trade with partners, 78f fertilizer subsidies removal, 114
economic performance, 74 financial sector liberalization, 114
export potential, 87 fiscal deficits, 113
exports and imports within ECCAS, foreign trade classification, 116
77t government development
foreign trade trends, 74 plans, 112
foreign trade volume, 77 import intensity, 116, 119, 121
geographical orientation, 77–8 import substitution
import potential, 87 industrialization, 111–13, 119
Index 235
imports-as-market-discipline relation with trade

hypothesis, 119 determinants, 157
Industrial Development Act, 112 structural changes, 197
manufacturing growth, 112t sub-regional, 154
manufacturing output growth rate transaction cost uncertainties, 156
trends, 115t mono-product economies, 20
manufacturing’s price–cost margins,
117–21 NAFTA, see North American Free Trade
perpetual inventory method, 116 Agreement
policies and performance, 112–15 NAG, see non-agricultural exporters
price–cost margins, 115f, f 121 New Keynesian Phillips curve, 204
prices liberalization, 114 Newtonian theory, 160, 181
protectionism, 111 Newton’s gravitational equation, 181
structural adjustment programs, non-agricultural exporters, 61
113–14 North American Free Trade
structural breaks, 111–2, 121 Agreement, 72–3, 75, 96–7, 178
structure-performance North-South trade, 126–7, 130
modeling, 111 industrialization, 130
trade liberalization, 111, employment, 127
114–15
OCA, see optimum currency area
trade policies, 111, 112t,
t
oligopolistic model, 111, 115
117–21
OLS, see ordinary least squares
manufacturing industries, external
optimal monetary zones, theory
and internal factors, 135t–6
t t
of, 189
market access, 33, 71
optimum currency area, 156
market integration, 71–3, 78, 222
ordinary least squares, 41, 82, 99–100,
MERCOSUR, 88n3
182, 185–6
modeling framework, 9
Monetary Cooperation Programme
panel heteroskedasticity, 117
(MCP), 171n2
Poisson models, 101
monetary policy, credibility of, 198 PPML, see Pseudo-Poisson Maximum
monetary union, 12, 153, 155–8, 160, Likelihood
175, 197–8 preferential trade agreements, 94
binary variable, 157 preferential trade area, 91
determinants of, 157 price–cost margins, 9, 120–1
effects on CFA Zone, 198 import substitution
effects on trade, 161 industrialization, 121
exchange rate volatility on positive effect on, 120
trade, 161 price elasticity of demand, 7, 49, 51–5,
facilitate international trade, 156 58, 60–4
idea of, 156 pro-cyclical fiscal spending, 126
impact on trade, 156–8, 160, pro-cyclical nature of
162, 181 productivity, 128
interest in, 155 product differentiation, 17, 174
mitigating effects of, 197–8 protectionism, 8–10, 111, 131
propensity score, 157 protectionist policies, 128, 143, 223
reduction of exchange rate, 156 Pseudo-Poisson Maximum
regional trade agreements, 158 Likelihood, 100
236 Index
quadratic capital-adjustment trade facilitation, 33–5, 40–2;

costs, 201 cross-country manufacturing
qualitative variables, elasticities survey, 34; effects of trade
of, 185 facilitation, 34; harmonization
quantitative restrictions, removal among members, 34; political
of, 62 commitment, 33; trade
barriers, 33
random effects (RE), 52–3, 62 regional integration theory, 94–6
raw material curse, 76 regional trade agreements, 17, 220
RCA, see revealed comparative effect of, 181–5
advantages limitations, 189
real exchange rate, 65 proliferation of, 17
real export growth, 65 prospects, 189
real import growth, 65 risk-sharing, 214
Ricardian theory, 97 remoteness index, 92, 99
RECs, see regional economic revealed comparative advantages, 18,
communities 22–6
reform process, 21 Ricardian theory, 97
regional economic communities, 2–3, Ricardo, David, 17, 29n1
5, 7–8, 34–5, 72, 74–5, 91–3, 96–7, risk-sharing, cross-border, 13, 198
99–100, 104, 153, 171n1, 218, 220 RTAs, see regional trade agreements
African membership, 3f rule of law, 42
agriculture trade, 34 rule-of-law index, 38
bilateral trade, 97
integration patterns of, 72
intra-regional trade, 104 SACU, see South African Customs
monetary unions in, 153 Union
proliferation of, 92 SADC, see Southern African
in Sub-Saharan Africa, 33, 92 Development Community
tariffs and quotas elimination, 218 SAPs, see structural adjustment
trade performance, 2 programmes
regional integration, 1, 20–2, 91, 218 Sargan test, 41, 54, 132
Africa’s trade composition, 92–4 single currency, 72, 83, 153, 156,
African economic communities’ 174–5, 181–5, 188–9
import sources, 94t coefficients signs, 184
agreements, 84, 94, 173 common border fosters trade, 186
effect of, 42 common past, 184
export share on agricultural cross-sectional data, 186
exports, 42 economic policy recommendations,
exports by type, 93t 188–9
global exports by regional economic geographical and structural factors,
community, 95t 186
proliferation of regional economic insularity, 188
communities, 92 landlockedness, 184, 188
regional integration theory, 94–6 language, 184
shares of goods exports, 93f multilateral resistance, 184
static allocation of resources, 96 specifications, 181–4
structural constraints, 93 social rapprochement, 72
trade documentation and, 91 South African Customs Union, 91
Index 237
Southern African Development model specification, 51–2;

Community, 3, 8, 11, 75, 91–4, random effects, 52–3;
97, 103–5, 114, 171n1, 174, 178, generalized method of
194n1, 218, 221 moments, 53–4; timing impact,
South-South trade 54–6
customs duties restriction, 142 import growth, 48–60, 62; decrease
on industrialization, 130 in import duties, 58; fixed
Standard International Trade effects, 57–9; impact of
Classification, 117 liberalization, 60; income
STATA software, 38 elasticity of demand, 60;
Stewart, 78 liberalization dummy, 57;
see also gravity model model specification, 57; price
structural adjustment programmes, elasticity of demand, 60; real
47, 111, 113, 125, 164 exchange rate, 57; timing
structural vulnerabilities, 128 impact, 59–60
sub-prime financial crisis, 175 impact on the tradeoff, 48
Sub-Saharan Africa (SSA), 72–3, 91 indicator of, 49
price and income elasticity of, 52
Tanganyika, 91 price elasticity of demand, 64
tariff and non-tariff barriers, 7, 31, 33, relation with imports, 48
87, 92, 168, 218 second GMM regression, 54
technological gap, 125 short-run impact, 49
theory of optimal monetary Static and dynamic panel
zones, 189 models, 48
timing impact, 54, 59, 61 Timing impact of trade reforms, 56t
Torren, Robert, 29n1 Trade Map database, 185
trade agreements, 174–5 trade openness, 114, 126–8, 130–2,
currency constraint, 174 148n2, 188
lack of success, 174 degree of, 130
monetary integration monetary positive impacts, 127
integration, 174 restriction of, 142
single currency adoption, 175 trade reforms, 47–8, 52, 54, 56, 59–64,
trade balance, 6–8, 23, 25–6, 48, 221–2
61–3, 159 in early twenty-first century, 47
trade barriers removal, 47 implementation of, 52
trade facilitation, definition of, 34 timing impact of, 54, 56t, t
trade imbalance, 23, 76 59, 61
trade liberalism, 125 timing investigation of, 64
trade liberalization, 6–7, 9, 17, 47–9, uninterrupted, 48
52–3, 56–8, 60, 62–4, 77, 111–14, Treaty of Rome, 18
120, 137 two-stage least squares (TSLS), 37
average duties, 48
degree of anti-export bias, 52 UMA, see Arab Monetary Union
dummy variable, 48 UN Comtrade database, 128, 142
duty variable, 48 urbanization (URB) level, 131
effect of, 53–4
export growth, 48, 50–1, 55t, t 62; volume of trade, 12, 74, 178, 181, 184
conventional export demand African trend, 178
equation, 51; performance, 48; benefits for the client country, 12
238 Index
volume of trade – continued trade analysis, 159–65

ECOWAS member countries, 74 trade integration, 194n7
reasons for low level. 181 variable coefficient, 166
West African Economic Community
WAEMU, see West African Economic (WAEC) treaty, 154
and Monetary Union West African Monetary Union
WAEMU Treaty, 194n7 (WAMU) treaty, 154
Wald (Chi-squared) tests, 41 West African Monetary Zone (WAMZ),
West African Economic and Monetary 11, 153, 155, 178, 194n8
Union, 2, 18, 75, 174, 178, 194n2, Wooldridge test, 117
198, 223 World Bank, 2–3, 7, 9–10, 32, 35,
balance of trade, 160 47–8, 81, 99, 101, 111, 113, 165,
empirical specification, 161–5 174, 219, 221, 226
exports share, 167f World Development Indicators (WDI),
gravity model, 160 35, 65, 81, 101, 165
impact on exports, 167 World Economic Forum (WEF), 1, 18,
impact on trade, 164–8 24, 226
intra-WAEMU trade/exports, 159, World Governance Indicators (WGI),
166 35
intra-zone imports, 159 World Trade Indicators (WTI), 35
marginal effects on exports, 165f World Trade Organization (WTO), 2,
objectives, 154 34, 65, 71, 77, 88, 97, 173, 176,
theoretical model, 161–2 223–4

Mthuli Ncube, Issa Faye, Audrey Verdier-Chouchane (Eds.) - Regional Integration and Trade in Africa-Palgrave Macmillan UK (2015)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mthuli Ncube, Issa Faye, Audrey Verdier-Chouchane (Eds.) - Regional Integration and Trade in Africa-Palgrave Macmillan UK (2015)

Uploaded by

Copyright:

Available Formats

Regional Integration and Trade in Africa

Regional Integration and

List of Figures vii

List of Maps viii

Foreword by Gilbert Mbesherubusa xii

List of Acronyms xxi

Introduction: Understanding Africa’s Regional Trade 1

Part I Intra-African Trade Performance and

2 Economic Integration, Trade Facilitation and Agricultural

3 The Impact of Trade Liberalization on Export Growth and

Part II Measuring Trade Potential: The Gravity

5 Regional Integration and Trade in Sub-Saharan Africa,

Part III Industrialization Strategy and the Issue of

7 The Impact of South-South and North-South Trade on

Part IV Impact of Currency Union on Trade

9 Trade Agreements and Flows in ECOWAS: Is a Single

10 A DSGE Model of Trade and Risk-Sharing Effects of

Conclusion: Enhancing Intra-African Trade through Regional

I.1 Africa’s membership of selected regional economic

C.1 Regional development corridor projects in Africa (2013) 222

1.1 The global competitiveness of CEMAC and WAEMU

5.6 PPML estimation of Anderson–van Wincoop gravity

9.A.4 Exports and imports of ECOWAS’ Least Developed

Regional integration has been at the core of Africa’s development efforts

This book brings together a selection of articles presented at the 2013

Désiré Avom is Professor of Economics at the University of Yaoundé

Thierry Kame Babilla is a researcher at the Centre d’Etudes et de

Ibrahima Camara is currently working as a statistician in the depart-

Hopestone Kayiska Chavula works as an economist in the

Issa Faye is Manager of the Research Division of the African Develop-

Lanre Kassim is a research student at the University of Kent, United

Gilbert Mbesherubusa was Vice President for Infrastructure, Private Sec-

Sokhna Diarra Mboup is a PhD student in Economics and a member of

Fellow at the Blavatnik School of Government in Oxford, United King-

Benjamin Ndong is Professor, Head of the Department of Economics

Mouhamed Mbouandi Njikam is a PhD candidate in Economics at

Henri Ngoa Tabi is Head of the Department of International Economics

Henry Okodua is a lecturer in Economics at Covenant University,

Oluyomi Ola-David is a PhD candidate and lecturer in the Depart-

Wumi Olayiwola is Principal Programme Ofﬁcer of the Economic

Henri Atangana Ondoa is a lecturer at the Faculty of Economics and

studies with the following agencies: United Nations Development Pro-

Evans Osabuohien is currently a research fellow at the German Devel-

Joseph Parfait Owoundi works at the Ministry of Economy, Plan-

Edris H. Seid is a junior research fellow at the regional independent

Audrey Verdier-Chouchane is Chief Research Economist in the Devel-

ACP African, Caribbean and Paciﬁc

EPA Economic Partnership Agreement

UNCTAD United Nations Conference on Trade and Development

average with faster economic growth and increased productivity, espe-

Figure I.1 Africa’s membership of selected regional economic communities in

1 Intra-African trade performance and regional integration

Although intra-African trade has been on the rise, the bulk of

EU United States South Africa Other

Figure I.2 Africa’s top trade partners in 2012 (%)

Intra-African trade Intra-African exports Intra-African imports

Figure I.3 Intra-African trade between 1995 and 2011

among fuel exporters (5.7%), supporting the argument that economic

farmers in high-income countries earn subsidies to stay competitive in

liberalization was found to have increased the growth of not only

2 Measuring trade potential: The gravity model

The gravity model offers a more rigorous method of assessing intra-

compared to other trading blocs worldwide, including developing areas