AFRICA’S DEVELOPMENT IN HISTORICAL PERSPECTIVE

his edited volume addresses the root causes of Africa’s persistent poverty through an investigation
of Africa’s longue durée history. It interrogates the African past through disease and demography,
institutions and governance, African economies and the impact of the export slave trade, colo-
nialism, Africa in the world economy, and culture’s inluence on accumulation and investment.
Several of the chapters take a comparative perspective, placing Africa’s developments alongside
other global patterns. he readership for this book spans from the informed lay reader with an
interest in Africa, to academics and undergraduate and graduate students, policy makers, and those
in the development world.

Emmanuel Akyeampong is a professor of history and of African and African American Studies
at Harvard University. He is the former editor of the Journal of African History and of African
Diaspora and the author or editor of several books, including Drink, Power and Cultural Change: A
Social History of Alcohol in Ghana (1996); Between the Sea and the Lagoon: An Eco-Social History of
the Anglo of Southeastern Ghana (2001); hemes in West Africa’s History (2006); and Dictionary of
African Biography (six volumes) (2013).

Robert H. Bates is Eaton Professor of the Science of Government at Harvard University. His research
focuses on the political economy of development, particularly in Africa, and on violence and state
failure. Professor Bates has conducted ieldwork in Zambia, Sudan, Uganda, Kenya, Colombia,
and Brazil. He currently serves as a researcher and resource person with the Africa Economic
Research Consortium, Nairobi. Among his most recent books are Analytic Narratives with Avner
Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry Weingast (1998); When hings Fell Apart
(Cambridge University Press, 2007); and Prosperity and Violence (2009).

Nathan Nunn is a professor of economics at Harvard University. His primary research interests
are in economic history, economic development, political economy, and international trade. He is
an NBER faculty research Fellow, a research Fellow at BREAD, and a faculty associate at Harvard’s
Weatherhead Center for International Afairs. He is also currently coeditor of the Journal of
Development Economics.

James A. Robinson is the David Florence Professor of Government at Harvard University. His main
research interests are in political economy, comparative economic development, and economic
history with a focus on sub-Saharan Africa and Latin America. He is the coauthor, with Daron
Acemoglu, of Economic Origins of Dictatorship and Democracy, (Cambridge University Press,
2006), which was awarded the 2007 Woodrow Wilson Foundation Award by the American Political
Science Association for “the best book published in the United States during the prior year on gov-
ernment, politics, or international afairs.” He edited the book Natural Experiments in History with
geographer and ecologist Jared Diamond in 2010. His most recent book, also written with Daron
Acemoglu, is entitled Why Nations Fail and was declared one of the ten best books of 2012 by the
Washington Post.
Africa’s Development in Historical
Perspective

Edited by

EMMANUEL AKYEAMPONG
Harvard University

ROBERT H. BATES
Harvard University

NATHAN NUNN
Harvard University

JAMES A. ROBINSON
Harvard University
 32 Avenue of the Americas, New York, NY 10013-2473, USA

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781107691209
© Emmanuel Akyeampong, Robert H. Bates, Nathan Nunn, and James A. Robinson 2014
his publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2014
Printed in the United States of America
A catalog record for this publication is available from the British Library.
Library of Congress Cataloging in Publication data
Africa's development in historical perspective / editors: Emmanuel Akyeampong,
Harvard University, Robert H. Bates, Harvard University, Nathan Nunn,
Harvard University, James A. Robinson, Harvard University.
pages cm
Includes bibliographical references and index.
ISBN 978-1-107-04115-8 (hardback) – ISBN 978-1-107-69120-9 (pbk.)
1. Economic development – Africa – History. 2. Africa – Economic conditions – History.
3. Africa – Politics and government. I. Akyeampong, Emmanuel Kwaku, editor of
compilation. II. Bates, Robert H., editor of compilation. III. Nunn, Nathan, editor
of compilation. IV. Robinson, James A., 1960– editor of compilation.
HC800.A57255 2014
338.96009–dc23 2014001822
ISBN 978-1-107-04115-8 Hardback
ISBN 978-1-107-69120-9 Paperback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party Internet Web sites referred to in this publication
and does not guarantee that any content on such Web sites is, or will remain,
accurate or appropriate.
 Contents

Author Biographies page vii

Acknowledgments xiii

Introduction: Africa – he Historical Roots of Its

Underdevelopment 1
Emmanuel Akyeampong, Robert H. Bates, Nathan Nunn, and
James A. Robinson

PART I. LA LONGUE DUR É E

1. Africa in World History before ca. 1440 33

Christopher Ehret
2. Reversal of Fortune and Socioeconomic Development in the
Atlantic World: A Comparative Examination of West Africa and
the Americas, 1400–1850 56
Joseph E. Inikori
3. he Impact of Malaria on African Development over the Longue
Durée 89
David N. Weil
4. African Population, 1650–2000: Comparisons and Implications
of New Estimates 131
Patrick Manning

PART II. CULTURE, ENTREPRENEURSHIP, AND DEVELOPMENT

5. Redistributive Pressures in Sub-Saharan Africa: Causes,

Consequences, and Coping Strategies 153
Jean-Philippe Platteau

v
vi Contents

6. Accumulation and Conspicuous Consumption: he Poverty of

Entrepreneurship in Western Nigeria, ca. 1850–1930 208
Ayodeji Olukoju
7. Commerce, Credit, and Mobility in Late Nineteenth-Century
Gold Coast: Changing Dynamics in Euro-African Trade 231
Emmanuel Akyeampong
8. he Textile Industry of Eastern Africa in the Longue Durée 264
William G. Clarence-Smith
9. Explaining and Evaluating the Cash Crop Revolution in the
“Peasant” Colonies of Tropical Africa, ca. 1890–ca. 1930:
Beyond “Vent for Surplus” 295
Gareth Austin
10. Reinventing the Wheel: he Economic Beneits of Wheeled
Transportation in Early Colonial British West Africa 321
Isaías Chaves, Stanley L. Engerman, and James A. Robinson
11. Mbanza Kongo/São Salvador: Culture and the Transformation
of an African City, 1491 to 1670s 366
Linda Heywood

PART III. INSTITUTIONS

12. he Fragile Revolution: Rethinking War and Development in

Africa’s Violent Nineteenth Century 393
Richard Reid
13. he Imperial Peace 424
Robert H. Bates

PART IV. EXTERNAL FORCES

14. Dahomey in the World: Dahomean Rulers and European

Demands, 1726–1894 447
John hornton
15. he Transatlantic Slave Trade and the Evolution of Political
Authority in West Africa 460
Warren C. Whatley
16. Gender and Missionary Inluence in Colonial Africa 489
Nathan Nunn

Index 513
 Author Biographies

Emmanuel Akyeampong is a professor of history at Harvard University. He

has written on social history and on the history of disease and the environ-
ment in West Africa. His more recent work has looked at trade and political
economy, interests strengthened by his participation in the Harvard work-
ing group on “Understanding African Poverty over the Longue Durée.”
His recent articles in this area include: “Africa, the Arabian Gulf and Asia:
Changing Dynamics in Contemporary West Africa’s Political Economy,”
Journal of African Development, 13: 1 (Spring 2011), 73–105; and Emmanuel
Akyeampong and Hippolyte Fofack, “he Contribution of African Women
to Economic Growth and Development: Historical Perspectives and Policy
Implications – Part I: he Pre-colonial and Colonial Periods,” World Bank,
Policy Research Working Paper, WPS 6051, April 2012.
Gareth Austin is a professor of international history at the Graduate Institute
of International and Development Studies, Geneva. Previous employers
include the University of Ghana and the London School of Economics. His
research and teaching interests are in African, comparative, and global eco-
nomic history. His primary research has focused on West Africa, especially
Ghana and the precolonial kingdom of Asante. His publications include
Labour, Land and Capital in Ghana: From Slavery to Free Labour in Asante
(2005); “Resources, Techniques and Strategies South of the Sahara: Revising
the Factor Endowments Approach to African Economic History,” Economic
History Review, 61: 3 (2008), 587–624; and (edited with Kaoru Sugihara)
Labour-Intensive Industrialization in Global History (2013).
Robert H. Bates is Eaton Professor of the Science of Government at
Harvard University. During the past decade, he has focused on conlict and
state failure and, working with the Africa Economic Research Consortium,
on the political economy of the “lost decades” following independence. In

vii
viii Author Biographies

recent years, he has returned to the study of agricultural policy, focusing on

political responses to international price shocks and policy responses to the
enfranchisement of rural voters. He spends his summers in Ghana trying to
igure out when the absence of violence transforms into the rule of law and
peace becomes an institutionalized feature of political order.
Isaías Chaves is a PhD candidate in political science at Stanford University.
His past research projects have examined the determinants of electoral
fraud in 1920s Colombia and the divergent political consequences of civil
war in Sierra Leone and Colombia. His current work in historical political
economy studies the domestic political coalitions that drove British impe-
rial expansion ater the Great Reform Act. At the same time, he is pursuing
projects in American politics that adapt techniques from empirical indus-
trial organization to better understand the strategic behavior of legislators
and campaign donors.
William G. Clarence-Smith is Professor of the Economic History of Asia
and Africa in the School of Oriental and African Studies at the University of
London. His major current research project is on Middle Eastern migrants
in the colonial Philippines, who came mainly from “Greater Syria.” his has
taken him of late to Manila, Beirut, Madrid, and Washington, DC, and has
involved probing the history of the global trade in embroidery and lace. He
also works on tropical commodities and the manufacturing related to them,
not only cotton, but also rubber, palm oil, hot beverages, masticatories, and
large livestock. A third part of his research concerns the interplay between
Islamic law and social history, notably in regard to slavery, sexuality, and
technological change.
Christopher Ehret is Distinguished Research Professor in the Department
of History at the University of California at Los Angeles. His research over
the past four decades has dealt with a wide variety of African historical
and cultural topics and a wide range of time periods. A major method-
ological focus in this work has been to develop the tools and techniques
for reconstructing history, especially early African history, from the evi-
dence of language and for correlating the evidence of language with other
types of historical evidence. One long-running theme in his research has
been the origins and early history of agriculture in Africa. A second major
focus has been on understanding the ways cross-cultural interactions have
shaped the courses of long-term social, cultural, and economic change
across the continent. Recently he has embarked on a new course of research
into the longue durée histories of African kinship systems, and he has also
 Author Biographies ix

undertaken collaborative work with geneticists and biological anthropolo-

gists on the possibilities of correlating their indings with those of linguis-
tics and archaeology.
Stanley L. Engerman is John H. Munro Professor of Economics at the
University of Rochester. He is best known for his quantitative historical
work along with Nobel Prize–winning economist Robert Fogel. His coau-
thored books include Time on the Cross: he Economics of Slavery (1974)
with Robert Fogel and Economic Development in the Americas since 1500:
Endowments and Institutions (2012) with Kenneth L. Sokolof.
Linda Heywood is a professor and the director of African American
Studies at Boston University. In November 2011, Heywood published an
article entitled “Angola, the Violence Years: Civilian Casualties,” and she has
just completed her book Queen Njinga of Angola/Matamba: A Biography.
During the summer of 2011, she spent three weeks in Angola working in
the archives, where she collected materials for her Njinga project. She also
visited the locations in Malange, Angola, where the people called Njingas
live. here she conducted video interviews with the “king of the Njingas”
and other oicials who all claim to be the direct descendants of the Njinga
dynasty and her followers. Heywood was allowed to visit the place where
Njinga and other members of the Ndongo ruling elite are supposed to be
buried. She was told that she was only the third outsider accorded this
honor. Heywood has now turned her attention to the second part of the
Njinga project, entitled “he Remaking of Njinga: Angola, Portugal and the
African Diaspora.”
Joseph E. Inikori is a professor of history at the University of Rochester.
He was previously chairman of the Department of History, Ahmadu Bello
University, Zaria, Nigeria. One of the pioneers of the study of Atlantic
World history, he has published extensively on Atlantic World economic
history. His most recent book in the ield, Africans and the Industrial
Revolution in England: A Study in International Trade and Economic
Development (Cambridge University Press, 2002), won the 2003 American
Historical Association’s Leo Gershoy Award for “the most outstanding
work in English on any aspect of the ield of 17th- and 18th-century west-
ern European history,” and also the 2003 African Studies Association’s
Herskovits Award. he book has been entered in the American Council
of Learned Societies Ebook Project “dedicated to selecting and creat-
ing an electronic collection of important scholarly monographs that are
expected to have continuing relevance in the ield of history.” His lifetime
x Author Biographies

achievement honors include he Distinguished Africana Award from the

African New World Studies Program, Florida International University
(2007) and the Distinguished Africanist Award from the New York State
African Studies Association (2008).
Patrick Manning is Andrew W. Mellon Professor of World History and
director of the World History Center at the University of Pittsburgh. For
many years, he has analyzed historical African populations, especially as
they were inluenced by enslavement, and he is near completion of a col-
laborative work with Scott Nickleach on African population from 1650 to
1950. He has written widely on migration in world history, most recently in
he African Diaspora: A History through Culture (2009). His current project
is a collaborative project, with historical linguist Christopher Ehret, on the
history of migration and social change in early human communities (ca.
100,000–5,000 years ago) as seen through historians’ reading of the evi-
dence on language, genetics, archaeology, climate, plants, and animals.
Nathan Nunn is a professor of economics at Harvard University. Professor
Nunn’s research has focused on empirically examining the determinants of
the long-term economic development of societies both within Africa and
across the globe. His research traces the impacts of large historical events
like Africa’s slave trades, missionary activities in Africa, the transfer of food
crops during the Columbian Exchange, plantation slavery in the Americas,
historical state formation, historical conlict, and technology adoption. His
ongoing research seeks to better understand exactly why many historical
events have impacts that are persistent and continue to inluence economic
development today.
Ayodeji Olukoju is a professor of history at the University of Lagos and
vice-chancellor, Caleb University, Lagos, Nigeria. He is the author of he
“Liverpool” of West Africa: he Dynamics and Impact of Maritime Trade in
Lagos (2004) and Culture and Customs of Liberia (2006). His other pub-
lications on maritime, social, economic, urban, and African history have
appeared as authored and edited books, chapters in books, and articles in
high-impact specialist and African journals. He serves on the boards of
the Journal of African History and History in Africa: A Journal of Method.
His latest contributions include “Food and Food Production” in homas
Spear (ed.) Oxford Bibliographies in African Studies (2013) and a chapter on
Lagos in Miguel Suarez Bosa (ed.), Atlantic Ports and the First Globalisation,
c. 1850–1930 (Cambridge Imperial and Post-Colonial Studies Series,
Macmillan-Palgrave, 2014, chapter 6).
 Author Biographies xi

Jean-Philippe Platteau is a professor of economics at the University of

Namur. Most of his work has examined the understanding of the role of
institutions in economic development, and the processes of institutional
change, especially under the joint impact of population growth and mar-
ket penetration. he inluence of noneconomic factors and various frontier
issues at the interface between economics and sociology are a central focus
of his research projects, hence his continuous interest in other social sci-
ences than economics and his continuous emphasis on the potential contri-
butions of sociology to the ield of economics in general, and to economic
development in particular. Recently, he has embarked upon two ambitious
research projects. he irst one deals with the transformation of farm-cum-
family structures when land becomes more scarce and market integration
increases. Field data of the panel type are collected in central Mali for the
purpose of testing theoretical predictions. he second project concerns the
interaction between the statutory law and the custom when the lawmaker
wants to ight oppressive social norms and both informal and formal judges
act strategically. Fieldwork in relation to the latter project is being carried
out in Senegal, Guinea Bissau, and Guinea Conakry.

Richard Reid is a professor of African history at the School of Oriental and

African Studies, University of London. He has recently completed a project
dealing with long-term conlict in the Horn of Africa, which resulted in
Frontiers of Violence in Northeast Africa (Oxford University Press, 2011).
His thinking on African warfare more broadly has now been distilled into
Warfare in African History (Cambridge University Press, 2012). While he
continues to ponder the links between war and development in the making
of modern Africa, he has also begun to travel in a slightly diferent direc-
tion, namely research on the role of historians and historical consciousness
in African state and society during the nineteenth and twentieth centuries,
using the case study of Uganda.

James A. Robinson is David Florence Professor of Government at Harvard

University. He is currently on sabbatical in Colombia researching the ori-
gins of paramilitarism and how and why some paramilitary groups created
state-like structures while others did not. He is also collecting data on elite
networks in Haiti to investigate how elite control of the economy inluences
the development of the society. He has just inished an empirical project on
the consequences of indirect rule for development in rural Sierra Leone and
has been in the DRC, Uganda, and Rwanda researching the roots of delayed
political centralization in Africa.
xii Author Biographies

John hornton is a professor of history and African American Studies at

Boston University. He is primarily an Africanist, with a specialty in the his-
tory of West Central Africa before 1800. His work has also carried him
into the study of the African Diaspora, and from there to the history of the
Atlantic Basin as a whole, also in the period before the early nineteenth
century.
David N. Weil is the James and Merryl Tisch Professor of Economics at
Brown University. Most of Professor Weil’s research examines long-run
economic growth. Among the topics he has studied are technology transfer,
population growth, accumulation of physical and human capital, and the
transition from the Malthusian equilibrium to modern economic growth.
Much of his recent work has looked at how population health afects the
level of income in a country and how demographic changes such as declin-
ing fertility and population aging impact growth. In his latest completed
research, he and his coauthors explore techniques for measuring economic
growth using satellite observations of light that is visible at night. his
method can be applied both in countries where standard national income
accounts data are of low quality and to subnational or supranational regions
for which no other data are available.
Warren C. Whatley is a professor of economics at the University of Michigan.
His research interests include the economic history of Africans and African
Americans – primarily African American workers in the twentieth century
north and south of the United States, and the impact of the transatlantic
slave trade and colonialism on African political and economic develop-
ment. His articles have appeared in the Quarterly Journal of Economics, the
Journal of Economic History, and the Journal of Labor Economics, among
others. He is currently writing a book on the transatlantic slave trade and
the political and economic development of West Africa.
 Acknowledgments

his edited volume, African’s Development in Historical Perspective comes

out of the initiative of a working group on “Understanding African Poverty
over the Longue Duree,” funded by the Weatherhead Center for International
Afairs at Harvard University. he working group organized a conference
in Ghana in the summer of 2010, co-hosted by the International Institute
for the Advanced Study of Cultures, Institutions and Economic Enterprise
(IIAS), under the then director Professor Irene Odotei. We acknowledge
the tireless assistance of Jaronica Fuller at Harvard and Sharon Okantey,
Ebi Kanga Landry, and others from IIAS (Ghana), who made the Ghana
conference a great success. We also express our deepest appreciation to all
the scholars who participated in the Accra conference, and whose presence
made this one of the most stimulating intellectual events ever.

xiii
 Introduction

Africa – he Historical Roots of Its

Underdevelopment

Emmanuel Akyeampong, Robert H. Bates, Nathan Nunn,

and James A. Robinson

Africa poses the development challenge of our time. Once devastated by

famine and mired in poverty, India and much of Asia are now growing
economically. While Africa’s economies too are now growing, in many
African countries, people are just now returning to the levels of per capita
income they enjoyed more than a half century ago.1 he current poverty
and the widespread economic decline following independence in Africa
have led to a great deal of research by development economists and politi-
cal scientists (Bates 1981; Easterly and Levine 1997; Ndulu et al. 2007; Sachs
and Warner 1997). Yet this work is oten very policy focused and does not
place Africa within the larger processes that created the Great Divergence
in the early modern and modern worlds (Pomeranz 2000).
his is somewhat odd, because the adoption of a much more historical
approach has revolutionized research on comparative economic develop-
ment in the past iteen years. For example, the economic divergence that
took place in the Americas during the nineteenth and twentieth centuries
is now seen as deeply rooted in the diferent institutional structures created
during the colonial period (Bruhn and Gallego 2012; Dell 2010; Engerman
and Sokolof 1997, 2011). It is commonplace, though not uncontroversial,
to blame Africa’s economic problems on its economic, political, and per-
haps social institutions. It is less common to ind studies that situate these

1
In its focus on the historical roots of Africa’s underdevelopment, this volume does not
include contributions on more contemporary developments, such as the World Bank’s
structural adjustment programs or the current commodity boom, driven by rising Asian
demand, among other factors (Mkandawire and Soludo 2003; World Bank 1994). On the
current growth cycle in Africa, see IMF (2009) and Kasekende, Brixova, and Ndikumana
(2010). See also Radelet (2010).

1
2 Akyeampong, Bates, Nunn, and Robinson

issues in a historical context and consider why they diverged from the rest
of the world in the irst place (see Austin 2008; Nunn 2008a; and Acemoglu,
Johnson, and Robinson 2001, 2002; Acemoglu and Robinson, 2010, for
exceptions).
he aim of this book is to trigger a research agenda of the type that has
reoriented our understanding of the Americas (see Hopkins, 2009, for sim-
ilar suggestions). Although scholars have published a rich literature on the
economic history of Africa, contemporary approaches are yet to be inte-
grated into the broader literature on its development, and for Africa we
have nothing like the type of synthesis or consensus that has emerged for
the Americas.
Attempts to provide grand narratives of African economic history are not
new, of course, though they took some time to emerge. he economic prob-
lems of Africa were an intense subject of interest for European colonial oi-
cials. Colonial powers oten justiied their colonization of Africa on the basis
of the continent’s poverty, poor institutions, and backward technology. hey
put forth various explanations for African poverty. An example is the notion
that the incidence of the tsetse ly limited the use of draught animals that
impeded the adoption of the plow and wheeled transportation (see McPhee
1926). Yet colonial oicials did not create a coherent view of the economic
development of Africa and were typically content to argue that colonialism
would lead to a straightforward dissemination of modern technology and
institutions with immediate positive efects on African living standards.
As early as 1958, however, Suret-Canal provided a famous Marxist inter-
pretation of African poverty by applying the notion of the “Asiatic mode
of production.” his approach took it for granted that Africa was relatively
poor and explained this by the fact that Africa was trapped with precapi-
talist institutions. he research he pioneered long lourished in African
studies (see Birnberg and Resnick 1975; Law 1978; Sender and Smith 1986)
and triggered a fascinating debate over the role of feudalism in Africa (for
example, Crummey 1980; Goody 1971). Another grand narrative emanated
from anthropology, when the “substantivist school,” also accepting Africa’s
historical poverty as given, argued that African societies were not based
on market principles (though they did have marketplaces), and therefore
did not respond to the economic incentives created since the Industrial
Revolution in ways that generated prosperity (Dalton 1976; Polanyi 1966;
Polanyi, Arensberg, and Pearson eds. 1957). he 1960s and 1970s also
saw a great deal of work by dependency and world systems theorists who
attempted to explain African underdevelopment not as a consequence of
primarily internal dynamics, as in the Marxist and substantivist approaches,
 Introduction 3

but as a consequence of its integration on unfavorable terms with the world

economy (e.g., Amin 1974, 1976; Wallerstein 1974–2011). his literature has
many variants; for example, Rodney (1972) developed an inluential syn-
thesis of Marxist and dependency theory approaches with a heavy emphasis
on the impact of formal colonization. Other scholars pointed to speciic
features that were quantitatively so important as to become the dominant
force shaping African economic history. For Goody (1971), this was the
relative abundance of land and scarcity of labor. For Inikori (1992), it was
the Atlantic slave trade. For Douglas (1962, 1963) and Vansina (1978), it
was the historical dynamics of state formation in Africa. For Law (1981)
and Miller (2009), it was the absence of the horse. For others, it was the
impact of colonization (Palmer and Parsons 1977).
Common to this research was the acceptance of Africa’s poverty relative
to the rest of the world in the early modern period. In a seminal symposium
based on the paper by hornton (1992), those unconvinced by the histor-
ical evidence for Africa’s poverty challenged this view. Other scholars also
critiqued these early approaches for their overly static view of the African
past (see Jerven 2010).
Alongside such grand narratives a large monographic literature devel-
oped focusing on speciic themes. One focused on precolonial trade
(Birmingham and Gray 1970; Bohannon and Dalton 1962; Meillassoux
1971). A second investigated the impact of the slave trade (Curtin 1975;
Manning 1990; Miller 1996). hese literatures have been surveyed by many
(e.g., Austen 1987; Hopkins 1973; Wickens 1980, 1986).2
Yet while many of these early narratives linger (for example, the relation-
ship between the substantivist school and Hydén’s, 1980, 1983, “economy
of afection”) and have sparked productive and lively debates (for example,
that between Hopkins, 1973, and Dalton, 1976, on the empirical applica-
bility of substantivist claims and the book by Hill, 1963, arguing for the
sophisticated economic rationality of Africans and the adaptability of tra-
ditional economic institutions), their conceptual frameworks have fallen
out of favor. Political scientists focused much more on the politics of con-
temporary Africa. Anthropologists moved away from research on political
and economic institutions. Economists studied contemporary development
problems as if they were timeless, and conducted no research within the
subield of economic history in economics departments. he consequence
is a relative void of convincing narratives of African economic history even
in its own terms, let alone in a comparative context. For example, the big

2
Austin (2008) provides a recent synthesis of much of this research.
4 Akyeampong, Bates, Nunn, and Robinson

debates in comparative economic history over the past twenty years have
focused squarely on questions such as the divergence between Europe and
Asia (Findlay and O’Rourke 2007; Jones 1981; Morris 2011; Pomeranz 2000;
Rosenthal and Wong 2011; Van Zanden 2009; Wong 1997) or on why Britain
diverged from Western Europe (Allen 2009; Mokyr 2009). In no study is
there any reference to Africa or to its divergence from the other economies
of the world. Why this occurred has never been one of the big questions in
either economic history or comparative economic development.
Our book is a relection of the fact that this situation of neglect is now
changing. Allen (2011) has made a pioneering attempt to integrate Africa
into his theory of the great divergence, and we hope that the chapters in this
volume will provide stimulus to make further progress toward this goal.
At the same time as this nascent interest by comparative economic histo-
rians, economists who saw the economic problems of Africa as essentially
those of any poor economy began to recognize the problematic nature of
this view. As Africa emerged from colonialism in the 1960s, the irst wave
of development economists proposed simpliied models of African pov-
erty that either wrapped countries into general models of the problems of
“backward economies,” implying few diferences between Africa and Asia,
for instance, or proposed ahistorical models of the sources of poverty (see
Killick, 1978, for a good overview). he starting point for this is easily seen
by reading the characterization of African development problems by late
colonial scholars such as Stamp (1953) and Batten (1954). Only with the
successive failure of the models of development economics (see Easterly
2001) did a perception emerge that it is critical to understand the speciici-
ties of African society and institutions and where these came from in order
to unlock the vast economic potential of the continent. here certainly were
voices in the wilderness urging this (Hopkins 1986), but the larger intel-
lectual reorientation over the past decade in comparative economic devel-
opment toward a focus on historical long-run forces shaping institutional
development has emphasized the need for completely rethinking the nature
of African poverty.
he attempt to develop a convincing comparative account of African
development is necessarily interdisciplinary. It involves absorbing not just
the recent research in economic history and comparative development, but
also the rich literature on the economic and political history and anthro-
pology of Africa. It fact, it was historians, such as McGreevey (1971) and
Coatsworth (1978), as much as anyone who initially framed the new ques-
tions that scholars began to ask about the comparative economic history
of the Americas. he task is neither one of embedding Africa within some
 Introduction 5

preexisting paradigm nor of developing some explanation for Africa’s

uniqueness. Rather, it is to reine and reevaluate the types of explanations
that now reign in comparative economic history and development. he his-
torical evolution and development of African political systems, economic
institutions such as product and factor markets, or the nature of landown-
ership provide an unrivaled chance to learn about economic history.
We do not pretend that this book can provide the type of synthesis to
which we aspire. We can only hope that it marks the start. Its origins lie in
a conference we organized in Accra, Ghana in July 2010, with the type of
interdisciplinary group of scholars that seemed necessary for this venture.
At the conference, we focused on several main questions: When did Africa
become poor? Why did Africa become that way? What mechanisms made
it so? And why did it stay that way?
While each of the chapters in this volume addresses the roots of Africa’s
poverty, they fall into four distinct categories. Contributors to the irst
category view Africa from a global perspective, asking “When did Africa
become poor?” and comparing its development to that in other regions.
In response to our next question, “Why did Africa become that way?,” a
second set of chapters focuses on institutions, both those that antedate the
colonial encounter and the slave trade and those that were shaped by them.
he third cluster of chapters highlights the values and practices of some of
Africa’s cultures and assesses the possibilities they open and the limitations
they impose for better understanding long-term economic development.
he last category of chapters addresses the impact of external forces, most
notably the slave trade, colonialism, and global markets, in shaping African
development.

AFRICA IN HISTORICAL PERSPECTIVE

he chapter by Christopher Ehret reaches far back in history – in some
instances to 10,000 BCE – to argue that the level of economic develop-
ment on the African continent was on par with that in the rest of the world.
Agriculture, he contends, emerged sometime between 9000 and 6000 BCE
in the southern part of the Eastern Sahara, which is comparatively early by
global standards. Some of the earliest developments in ceramics took place
in Africa around 9000 BCE, as did the domestication of cattle between 8500
and 7200 BCE by the Northern Sudanese in the Eastern Sahara. His read-
ing of the evidence suggests that the people of this region also numbered
among the irst to cultivate crops. Far from lagging behind the rest of the
world, African societies of this period were positioned at the technological
6 Akyeampong, Bates, Nunn, and Robinson

frontier. Like all issues in early African economic history, the speciic dat-
ing of developments is debated. Africa is far less studied archaeologically
than the rest of the world, and the absence of written records and material
remains means that a great deal of creativity has to be used, including such
controversial techniques as glottochronology.3 Yet it was just such innova-
tions that put precolonial African history on the map in the 1960s.4
Ehret’s chapter suggests that it is most likely that the economic diver-
gence of Africa from Eurasia is a feature of the past millennium. Much other
work points in this direction. Phillipson (1998), for instance, illustrated the
compelling comparisons between the Kingdom of Axum and the Eastern
Roman Empire at the start of the Dark Ages, and even pointed out how
theories such as Piernne’s “Mohammed and Charlemagne” applied equally
to Ethiopia. Haour (2000) showed great continuities during the medieval
period between the central Sahel and northwestern Europe.
Further evidence on these issues comes from David Weil’s chapter, which
manages to narrow down the timing of Africa’s divergence to an even
greater degree. He presents evidence on some proxies for Africa’s pros-
perity relative to that of other regions, using such commonly employed
measures as population density, city size, and technology.5 His data sug-
gest that at 1500 BCE, rather than being the poorest region of the world,
Africa appears quite average. Population density, conditional on land qual-
ity, was higher in Africa than in the Americas and Oceania, but lower than
in Europe and Asia. A similar picture emerges from the data on urbaniza-
tion and technological choice: Africa was on par with the Americas, but
signiicantly less urbanized and less technologically advanced than Europe
and Asia. Weil also examines state development, measured from 1 to 1500
CE, using the State Antiquity Index developed by Chanda and Putterman
(2007). Again, taken as a whole, Africa attained levels of state development
that were lower than Asia and Europe, but higher than the Americas or
Oceania.
As with Ehret’s data, these estimates are conjectural. Population density,
for example, has to be reconstructed by backward projection from early
colonial censuses, and estimates of urban population may be even more
speculative. he fact that diferent authors highlighted in this book marshal
quite diferent numbers highlights this diiculty and the need for further

3
See Barker (2006) for a diferent assessment of what the evidence says about Africa, and
see Vansina (2004) for a critique of the use of glottochronology.
4
See, for instance, Vansina (1985).
5
See Prados de la Escosura (2011) for a complementary exercise.
 Introduction 7

research. However, a great deal of qualitative evidence suggests that the

end judgment is broadly accurate and unlikely to be overturned by further
reinements in the data.
As Weil stresses, the average diference between Africa and the rest of
the world masks signiicant heterogeneity within Africa, with the most
developed parts being North Africa, West Africa, and Ethiopia. he mea-
sures Weil examined suggest that these parts of Africa had attained levels of
development that were just below or even sometimes on par with Eurasia.
Elsewhere, acephalous societies were common, for example, in the Horn of
Africa or Southern Nigeria, and these places were probably quite a bit poorer
and less technologically advanced. Moreover, even in the more advanced
places, we know that there were important elements of technology that
lagged behind technology in other parts of the world. While Ethiopia and
North Africa had writing, the wheel, and the plow, West Africa did not.
To examine the trajectory of African development between 1650 and
2000 CE, Patrick Manning traces its demographic history. Total population
both provides a proxy for economic development and is of interest for its
own sake, particularly given the export of human beings during the slave
trade. In many ways an update and extension of Manning (1990), the chap-
ter argues that from the seventeenth century to the twentieth century, the
continent’s population was much larger in size yet growing at a slower rate
than previously thought. he implication is clear: by 1650 Africa entered a
period of sustained stagnation, one that lasted for two and a half centuries.
Such a fact, if it holds up to further empirical scrutiny, is undoubtedly a big
part of the story behind Africa’s divergence since 1500.
Joseph Inikori takes a comparative perspective in his chapter and probes
the relative development of West Africa and the Americas between 1400
and 1850 CE. Inikori examines estimates of city sizes and population den-
sities, as well as the history of market development, trade, and specializa-
tion, and argues that, prior to 1500, the level of economic development was
higher in West Africa than in the Americas, but that the economic develop-
ment of the Americas thereater increased dramatically while that in West
Africa did not.
When Inikori turns to the causes of this reversal, he points to the trans-
atlantic slave trade, which shipped between 11 and 12.5 million Africans to
the Americas. he use of enslaved Africans enabled the growth of highly
specialized plantation agriculture and mining: an advantage that accrued
to the Americas rather than to Africa. In West Africa, participation in
the transatlantic slave trade diverted efort from and actively hindered
productive activity, thus contributing to Africa’s stagnation and relative
8 Akyeampong, Bates, Nunn, and Robinson

underdevelopment, he argues. See Inikori (1992) for a development of

this view.
he chapters thus suggest that Africa’s underdevelopment is not a histor-
ical constant or something inevitable. Regions within Africa – namely West
Africa and Ethiopia – were once more highly developed than the Americas
and Oceania and comparable to Europe and Asia. Not until ater 1500 did
Africa stagnate and lag behind, as measured by population, city growth,
state development, or technological progress.
hese chapters thus set the stage for our attempt to explain Africa’s cur-
rent underdevelopment. Because Africa’s economic fortunes have varied,
time invariant factors, such as natural endowments, are unable to explain
its present underdevelopment. Rather, the authors turn to more malleable
factors: institutions, culture, and relations with the outside world. It is easy
to see that these things are likely interconnected. For example, one potential
explanation for why Africa exported slaves given that it was labor scarce is
that economic institutions within Africa were so poor that while physical
marginal productivity was potentially very high, the appropriable product
was low, thus making it relatively attractive to sell people into slavery rather
than to exploit them on the continent (Robinson 2012).

INSTITUTIONS
Among the institutions examined in this book’s chapters, the political
receive the most attention. And among the most important institutions are
the state and the regimes that governed it. Also critical, though less dis-
cussed in this book, are economic institutions.
Early accounts of Africa’s interior oten stressed the level of conlict
and war making, which, some researchers claimed, helped to account for
the continent’s relative lack of development (a lucid nineteenth-century
account with this lavor is Speke, 1863). Inspired by subsequent work in
comparative history, some of the authors in this volume mount a coun-
terargument. Early modern Europe, which experienced dramatic changes
in political and economic institutions, was violent, Reid and Bates note in
their chapters. But according to some (e.g., Bates 2001; Brewer 1988; Herbst
2000; Roberts 1956; Tilly 1975), the result was the emergence of the “iscal-
military” state: one that mobilized political power in support of economic
development. Be it ofensive or defensive, warfare is expensive – it must
be inanced – and those seeking to enhance the military capabilities of the
state may therefore use their power to promote economic growth. In pur-
suit of this theme, Bates joins Reid in noting the protection states in Africa
 Introduction 9

ofered to trade routes and commercial centers, as well as states’ provision

of public goods: roads, pontoons, and peace in the market. Using data from
the Standard Cross-Cultural Sample, Osafo-Kwaako and Robinson (2013)
show that greater precolonial political centralization is indeed correlated
with improved transportation infrastructure, greater occupational speciali-
zation, and the use of more advanced technologies. Heywood and hornton,
moreover, examine in their chapters attempts by embattled African rulers
to promote technological change by securing trained artisans from abroad.
hornton’s chapter applies the paradigm of the iscal-military state to
Dahomey, which sent its armies out virtually every year. But, hornton
argues, these armies “may not have simply been economic instruments
destined to shape the development of the state.” He draws on the corre-
spondence of Dahomean monarchs in the eighteenth and early nineteenth
centuries (Agaja in the eighteenth century, Adandozan between 1810 and
1812, and others). hornton underscores how in the letters the Dahomean
monarchs refute the allegation that they went to war to capture slaves.
Agongolo noted how, unlike England that was protected by the sea and had
commerce everywhere, Dahomey was hedged in by other nations similarly
armed, which was the cause of incessant conlict. We certainly can read the
Atlantic trade as the backdrop to this militarization and constant wars, as
the slave trade both militarized societies and absorbed the captives of wars.
But for hornton, the slave trade is not suicient explanation for Africa’s
lack of industrialization. hornton opines that the Industrial Revolution
was not state run in Europe, but a private enterprise, oten in spite of state
interference and monopoly concessions. And it is this little understood
dimension of precolonial African economies – the private sector – that may
shed light on Africa’s failure to industrialize.6
hese works resonate with a long history in African studies that addresses
the processes of state formation.7 One tradition, recently surveyed by
McIntosh (1999), argues that African states have developed along historical
paths that difer from those traced by early states in Eurasian; by implica-
tion, it would reject the arguments of Bates and Reid.8 he historical nature
and dynamics of the state in Africa and its comparison with Eurasian states
remains an important area for future research, particularly because recent
empirical studies suggest that precolonial political institutions have let a

6
See Law (1977) and Manning (2004) for discussion in the case of Dahomey.
7
Going as far back as Evans-Pritchard and Fortes (1940).
8
McIntosh (1988) for the Niger basin in West Africa and Sherif (2010) for the East African
coast stress the feature of heterarchy in inter-polity relations in early Africa, not hierarchy
based on aggressive expansion.
10 Akyeampong, Bates, Nunn, and Robinson

long shadow in Africa (Gennaioli and Rainer 2007; Michalopoulos and

Papaioannou 2013).
Early accounts by explorers, adventurers, and missionaries tended to
stress the malevolence of Africa’s rulers: their violent temperaments, their
unbridled appetites, and their sadistic dispositions. But, Bates notes from
the evidence we possess on Africa’s traditional political institutions, while
there certainly were despotisms, for example in the Kongo or Buganda, in
many places political institutions existed that constrained such behavior (see
also Beattie 1959). Even had these leaders been as benighted as European
travelers held them to be, they would have found their actions checked by
nonroyals who held high oice or by independent councils. While travel-
ers’ accounts may stress the fearsome character of Africa’s monarchs, then,
the empirical record suggests a more nuanced portrayal: Africa’s preco-
lonial institutions were oten imbricated with agencies of constraint. An
important example of this theme is the literature that attempts to explain
Botswana’s extraordinary economic success since independence and that
stresses the importance of precolonial institutions of accountability in the
Tswana states in helping to create good governance (Acemoglu, Johnson,
and Robinson 2003; Leith 2005).
In Africa as in Europe, there was warfare, of course; less obviously per-
haps, Reid and Bates conclude, military insecurity appears to have fortiied
the hand of political leaders who sought to protect and promote economic
activity. As Reid notes, in West Africa, most states possessed an urban core;
in support of merchants and artisans, rulers employed the forces of the
state to protect markets and to promote trade. In East Africa, he writes,
such eforts were devoted to long-distance trade, which promoted regional
specialization and “Smithian” growth, that is, increases in welfare resulting
from specialization and exchange. By incorporating those who could pillage
and destroy into administrative agencies subject to command and control,
the formation of states demobilized banditry and transformed warfare into
an instrument of public policy – sometimes for the worst, but sometimes
too for the better.
he arguments of Bates and Reid suggest that rather than Africa diverg-
ing from Eurasia around 1500, the similarities were much closer until the
past two hundred years. Indeed, both chapters suggest that African states
were in many places during the early modern period evolving along similar
lines to those in Eurasia, and only colonialism and the postcolonial inter-
national system blocked this path of institutional change. hey both agree
with hornton (1992) that African economic divergence must have come
much more recently.
 Introduction 11

he consequences of the strength and nature of the state and institutions

of governance are crucial as well for understanding the nature of economic
institutions in Africa. For instance, Douglas (1962, 1963) and Colson (1969)
suggest that the absence of state authority in parts of precolonial Africa
created disorder and insecurity of property rights. Yet just as some of the
scholars featured in this book ind parallels between African and Eurasian
political institutions, others, for example, Hopkins (1973), found similarities
between African economic institutions during the early modern and mod-
ern periods.9 his thesis too is controversial.10 It is clear, for example, that
slavery was endemic in Africa until well into the twentieth century (Miers
and Roberts 1988), and a potent source of economic ineiciency. Finley
(1965) argues that slavery was the fundamental reason for the lack of tech-
nological change in the classical world, and the data surveyed by Lovejoy
(2000) suggest that slavery in Africa was far more extensive as a percentage
of the population than it was in ancient Greece or Rome. Finley’s thesis
might help us to comprehend the technological lag that Africa experienced
(documented by Austen and Headrick 1983; Goody 1971; and Law 1980)
and to resolve a long-standing puzzle: Why did African societies not adopt
modern technologies of which they were clearly aware?11 Recent research
by Bezemer, Bolt, and Lensink (2012) has indeed found that the historical
incidence of slavery in Africa is strongly associated with poor economic
performance today.

CULTURE
Another set of chapters examines the importance of societies’ values and
cultural practices, something that resonates with the substantivist approach
to African development. While aware of the existence of cultural diversity
within Africa, Platteau identiies the existence of common cultural traits,
while Olukoju, focusing on the culture of one society, the Yoruba, addresses
themes of general signiicance. Each contributor stresses the signiicance
of beliefs about the importance of equality and of resistance to hierarchy.
Each points as well to the quest for social recognition. In so doing, they take

9
Hopkins (1973) is the most famous statement.
10
See Dalton (1976) for a critique.
11
Other economic institutions of crucial importance being the way that land is owned,
transferred, and allocated and the available evidence (Goldstein and Udry 2008) suggests
very large economic ineiciencies in the way the system is currently organized, which
itself relects deep historical continuities in Africa (Colson 1971).
12 Akyeampong, Bates, Nunn, and Robinson

common aim at the ability of Africa’s peoples to accumulate, to invest, and

so to forge the means of production necessary for an industrial economy.
Reaching deep into his knowledge of Yoruba history, Olukoju marshals
an array of instructive examples, some from the wars of the nineteenth cen-
tury, others from the time of the foreign occupation of Lagos, and still oth-
ers from the time of the creation of the cash crop economy. Olukoju’s vivid
portrayals conirm the entrepreneurial impulse that permeates Yoruba cul-
ture and the value placed on wealth. But he also points to the challenges
facing those who might wish to conserve it or to invest. hose who are rich,
he argues, are urged to celebrate their wealth, to display it, and to share it.
Being wealthy is a social, not an individual, act. It has more – much more –
to do with consumption than with investment. In Yoruba-land, one rarely
encounters ascetic capitalists; nor does one oten encounter capitalist insti-
tutions – irms or foundations designed to remain productive long ater
their founders have died.
One way of reading Olukoju’s chapter is through the lens of sociolo-
gist Max Weber, who, when discussing social stratiication, distinguished
between class, status, and power. Viewed in this way, Olukoju is suggesting
that Yoruba values lead people to trade wealth for status.
Platteau also underscores the importance of status and redistribution in
Africa’s cultures, but in so doing, illustrates the wide repertoire of defenses
that the rich can deploy (see Baland, Guirkinger, and Mali, 2011, for an
empirical study). Platteau illustrates how they can withdraw from the local
cultures, forming communities of their own: thus his discussion of the indus-
trial organization of commercial diaspora in West Africa. He also notes the
role of conversion, leading to his discussion of Long’s (1978) account of
the use of religion as a defense against the “beer culture” of the Copperbelt
in Central Africa and his discussion of Islam in Western Africa.12 While
ofering an implicit critique of the arguments of Olukoju, in other respects,
Platteau ampliies them, as in his discussion of witchcrat.
Nunn’s chapter turns to the impact of early twentieth-century Christian
missions on contemporary values. Controlling for possible alternative inlu-
ences – some historical and others contemporary – Nunn inds that Africans
from ethnic groups exposed to the inluence of foreign missions achieve
higher levels of educational attainment today. Interestingly, when Nunn
examines the education efect along gender lines, he inds very diferent
impacts for Catholic and Protestant missions. Although both had a similar

12
Another classic discussion of the use of conversion to Islam to avoid traditional social
exactions is Parkin (1972).
 Introduction 13

positive impact on long-run educational attainment, for Catholic missions

the increased education was amongst males only, while for Protestant mis-
sions the impact was across both genders, and actually greater for females.
As with the chapters on the state, these chapters can be read as attempts
to address the internal roots of Africa’s underdevelopment. But clearly, cul-
tural practices in other societies – particularly in societies that achieved
growth – contributed to Africa’s decline. One such force was Western sci-
ence. he absence of writing throughout most of Africa may have played an
important role in limiting the growth of science and therefore the growth
of Africa’s economies.
While the Enlightenment gave rise to a culture of skepticism and reason
in Europe, it was the print revolution that underpinned organized science
(Mokyr 2008). Printing made it possible for investigators to document the
priority of “discoveries,” while also disseminating them. hrough corre-
spondence and debate, ideas were shared and their merits subject to scru-
tiny. Reputations were thus made or lost; and with the invention of patents,
fortunes were as well. Writing, record keeping, and communication enabled
private eforts at innovation to yield inancial reward, thereby strengthen-
ing the incentives to achieve scientiic progress and promoting technical
change (see also Goody 1986).
Africa too developed advanced technologies: its smiths forged iron of
the highest quality, and its weavers, as recorded by Clarence-Smith in his
chapter, produced superb textiles (see also Herbert 1993; Kriger 2006; and
hornton 1992). In East Africa, some societies also kept records and cor-
responded in writing, although oten using Islamic scholars or Christian
missionaries to do so. But without widespread literacy and a print indus-
try to underpin it, Africa failed to generate widespread technical progress.
It failed to produce an industrial revolution. Energetic, risk-taking, and
appropriately celebratory of wealth, Africa nonetheless did not transform
its economic foundations. Export agriculture, mining, services, transport:
these industries existed in Africa. But while Europe and North America
beneited from the harnessing of coal, petroleum, and other sources of
energy for the production of goods, Africa did not.
Yet even an example as apparently simple as literacy is more complex
than meets the eye. Chaves, Engerman, and Robinson show in their chapter
that African societies knew about technologies such as the wheel but did
not adopt them. Such was the case with writing. While we do not under-
stand why Africa failed to widely adopt such technologies, the literature
ofers hints. Consider the work of Ewald (1988), for example, who argues
that the use of writing was a political act, stripping power away from those
14 Akyeampong, Bates, Nunn, and Robinson

who previously controlled oral knowledge and tradition.13 Perhaps then the
lack of literacy in Africa is but another product of its political economy.

EXTERNAL IMPACT
As noted at the outset, the preponderance of chapters suggests that Africa
became relatively poor ater the sixteenth century; until then, it occupied a
middle position in the global rankings, as measured by technology, popu-
lation, or power. What could explain the subsequent divergence? he argu-
ments thus far stress the importance of institutions and cultures, but they
are inconclusive. While some like Dalton (1976) or Acemoglu and Robinson
(2010, 2012) argue that economic institutions in precolonial Africa were not
conducive to development, others like Hopkins (1973), hornton (1992), or
Jerven (2010) dispute this. While Reid and Bates may argue that African
states were developmental, this view is also hotly contested (Acemoglu
and Robinson 2010). he inconclusive nature of these debates whets our
appetite for an alternative explanation. For most, following the scholars of
dependency theory, though emphasizing diferent mechanisms, the lead-
ing candidate is Africa’s encounter with Europe. he date at which Africa’s
fortunes began to diverge, ater all, marks as well the time of Europe’s escape
from the conines of the Mediterranean and the North Sea and its entry
onto the global stage. he possibility therefore arises that Africa, like other
portions of the globe, was itself subject to the great divergence.

Commerce and Agriculture

In some respects its encounter with Europe enriched Africa, promoting
exchange, trade, and the growth of institutions that underpinned markets.
And, as Akyeampong notes, its merchants, initially at least, grew prosper-
ous from their participation in foreign trade during the nineteenth century.
Drawing on the private papers of Gold Coast trading families such as the
Ocanseys and the Caesars, examining the career of the young Christian
Jacobson, and focusing on the major palm oil-exporting port of Ada,
Akyeampong explores the impact of declining transport costs and rising
foreign demand on the fortunes of merchants in West Africa. He studies
as well the impact of global inancial crises and the manner in which Gold
Coasters interpreted them. Compiling detailed case studies, he explores

13
Predating arguments more recently proposed by Feldman (2008) and Scott (2009).
 Introduction 15

the economic practices of families and irms – subjects that others, such as
Platteau, theorize.
Focusing on agriculture rather than commerce, Austin’s chapter also
explores the impact of global markets on West Africa during the late nine-
teenth and early twentieth centuries. While doing so, he compares and
critiques two major approaches: the “vent for surplus” school and the
“dependency” school. he latter, he indicates, stresses the importance of
coercion and views Africa’s integration into global markets as implying a
loss of welfare. While conceding that this interpretation might well it the
spread of cotton production in West Africa, Austin rejects it when applied
to the spread of other cash crops: palm oil, ground nuts, or cocoa, for exam-
ple. Nor does he fully embrace “vent for surplus” as a valid model. For while
properly emphasizing, in Austin’s view, the gains in income that resulted
from the production of cash crops, the “vent for surplus” theory treated
these gains as essentially free; they resulted, it was held, from the mobiliza-
tion of slack resources. But, Austin notes, labor in fact had alternative uses.
Put another way, the cash crop revolution in West Africa did not represent
the birth but rather the restructuring of West Africa’s economy.
Advancing this theme, Austin highlights the importance of that crucial
period when slave labor became free, with the redistribution of income and
the restructuring of contracts that must have accompanied that transition.
His chapter thus bridges the precolonial and imperial eras in Africa, and
our discussion of its economies during and ater the slave trades.

he Slave Trades
For many, the answer to the question “Why did Africa become poor?” is
obvious: the slave trades. Africa’s slave trades did not merely subtract from
the stock of labor, they also altered the way power and production were
organized and the nature of African institutions (see Inikori 1992, 1994;
Lovejoy 1989, 2000; Nunn 2008a). In short, the slave trades encouraged the
creation of institutions that favored the extraction rather than the creation
of wealth.
Were the standard models of international trade applied to Africa dur-
ing the slave trade, they would have predicted economic gains on the part
of exporters and importers alike. he evidence presented in this volume
suggests that the standard models would have been wrong and that with
the commencement of the slave trade, Africa’s relative fortunes began to
decline. What is suggested, then, is an approach that looks beyond the gains
from trade to institutions and cultures. his, like the colonial period, may
16 Akyeampong, Bates, Nunn, and Robinson

have shaped contemporary Africa, inluencing the use of power and atti-
tudes toward the production and accumulation of wealth. he analysis of
the slave trade that emerges its better with approaches advanced by, say,
Brenner (1976), who stresses the role of power relations, rather than by,
say, Postan (1937) or others who focus on the relative abundance of factors,
prices, and the functioning of markets (see Acemoglu and Robinson, 2010,
2012, for a development of this approach).
Reid and Bates, echoing Fage (1969), note that the rise of some states
in Africa was associated with the slave trade, though Nunn (2008a) shows
empirically that the average efect of the slave trade on political centrali-
zation was negative. But while some of those who seized slaves may have
forged states that encouraged trade, supplied public goods, and promoted
the welfare of those they ruled, they did so at the expense of those they
attacked and plundered. In addition, a major portion of the slaves withheld
from export were assigned to plantations, where they grew food for mem-
bers of the court, civil servants, and the military or for export abroad. As
Inikori noted, plantation agriculture based on slave labor became a major
form of production on both sides of the Atlantic: one that was privately
proitable for slave owners, but that failed – in either region – to induce
long-term growth (Engerman and Sokolof 1997; Nunn 2008b).
he slave trade can thus be said to have launched Africa along a trajec-
tory marked by slow growth and relative decline. It may also have impeded
subsequent development by promoting ethnic fractionalization (Easterly
and Levine 1997; Nunn 2008a). As Miguel and Gugerty (2005) show in
their study of school committees in western Kenya and Habyarimana and
colleagues (2007) reveal in their experiments on public good provision in
Uganda, ethnic diversity makes it more diicult to share in the costs of pro-
viding public goods. It therefore leads to the under-provision of education,
energy, roads and transport, and so forth that make economic development
possible.
Of equal importance, although more diicult to measure, is the impact
of slavery on economic institutions. Slavery involved extraction rather than
the creation of value.14 he slave trade may have led to the deterioration
of economic institutions in Africa, a process that continued ater the slave
trade was abolished (see the essays in Law, 2002). Even more provocative
is the inding of Nunn and Wantchekon (2011) of a relationship between
slaving and modern levels of interpersonal and political distrust. Because

14
As Lovejoy (2000) stressed, the evidence suggests that the rise of slave exports led to a
large increase in the extent of slavery within Africa.
 Introduction 17

it is premised on labor repression, slavery presumes the existence of polit-

ical forces capable of defending the claims of the slave-owning minority
against those of the working majority. Its very existence therefore requires
that power be employed to repress. his creates a rather diferent picture
of Africa’s historical states relative to the “developmental” one implicit in
the iscal-military hypothesis discussed by Bates and Reid. he nature and
dynamics of precolonial African states is undoubtedly a crucially important
area for future research.
he chapter by Warren Whatley explicitly tests for the impact of the
transatlantic slave trade on political structure, measured by the form of
succession of local headman. He does this by combining information from
Murdock’s Ethnographic Atlas, reporting ethnicity characteristics, with he
Trans-Atlantic Slave Trade Database, reporting the total number of slaves
shipped from each African port. His analysis shows that ethnic groups
located close to ports that exported a large number of slaves were more
likely to have hereditary succession; less likely to have succession based on
inluence, wealth, or status; and less likely to have succession determined
by election or other forms of consensus. He concludes that the slave trade
had a causal impact on succession, generating political structures that were
more absolutist.
Whatley also provides evidence that the transatlantic slave trade not only
impacted precolonial institutions, but also colonial institutions. Using data
from Nunn (2008a), he documents a positive relationship, among British
African colonies, between the intensity of the slave trade and British use
of indirect rule (measured using data from Lange, 2004, 2009). Whatley
argues that greater precolonial absolutism would have increased the ei-
ciency of indirect rule for the British, making its adoption more likely.

he Colonial Period
When the external world intruded upon Africa, it took the form of the slave
trade on the one hand. On the other, it assumed the guise of colonialism.
he academic literature contains many channels via which colonialism
could have created poverty in Africa (see Heldring and Robinson, 2013, for
a survey), a claim made more plausible if one accepts the idea that Africa
was not far behind the rest of the world during the early modern period.
Most who address the impact of colonialism focus not on the colo-
nial period itself, but rather on the consequences of colonialism for the
postcolonial period. Claims include those that the authoritarian nature
of colonialism created despotism ater independence (Mamdani 1996;
18 Akyeampong, Bates, Nunn, and Robinson

Young 1994); or let distorted “gate-keeper states” unable to promote

development (Cooper 2002); or let arbitrary “nation-states” with no
workable social contract, leading to low public good provision and to
conlict (Engelbert 2000).
While the studies in this volume do not focus directly on these issues,
Chaves, Engerman, and Robinson come close to doing so. In their chap-
ter, they examine the economics of transportation, particularly those of
the railways during the colonial period. In contrast to earlier revisionists
who argue that the wheel was an “inappropriate technology” for Africa
(Hopkins 1973), Chaves and colleagues argue that the introduction and use
of wheeled transportation was in fact highly beneicial. Indeed, they show
that the social rate of return on capital invested in the railways was huge, as
much as 100 percent per year. If Africa was not isolated and therefore knew
about these technologies (the Chaves, Engerman, and Robinson chapter
contends that they did), and the technologies were economically eicient,
why then were so few rail lines constructed in colonial Africa?
he authors provide evidence that the failure to adopt railways (if not
other forms of wheeled transportation) was the result of eforts of the colo-
nial powers to block their dissemination in some instances, and of the fears
of African leaders in others. Illustrative is the fate of European investment
in Ashanti in the 1890s. In this case, the Asante dispatched a mission to
England in 1895 in search of the inances that would enable them to develop
their economy through investments in mechanized gold mining and rail-
ways in partnership with European irms. But the project was blocked by
colonial secretary Joseph Chamberlain, who instead ordered a military
invasion (Wilks 1989). Rather than railways, Britain bound Ashanti with
the coast through the use of force, invading Kumasi even while its delega-
tion was in London. As Chaves and colleagues noted, in other instances –
as in Ethiopia – local elites resisted the construction of railways, viewing
them as a harbinger of colonial occupation. he latter response highlights
an informative contradiction. In the precolonial era, those who ruled oten
sought to enhance political security by accelerating economic development.
However, in the era of rapid colonization, rulers oten sought to enhance it
by retarding it. hus did political calculations shape the pace and nature of
Africa’s development.
his chapter complements recent research by economic historians on
colonial Africa. hey have shown that in West African nations, where there
were few European settlers, colonialism was consistent with rising living
standards for Africans (Frankema and van Waijenburg 2012) and improving
stature (Cogneau and Rouanet 2011; Moradi 2008, 2009). But where settler
 Introduction 19

populations were large and intent upon mobilizing African labor, colo-
nialism led to immiserization (the calculations in Heldring and Robinson
2013; Mosley 1983; Wilson 1972). While these chapters help to account for
the level of development the colonial occupation bequeathed to indepen-
dent Africa, they fail to inform us about the manner in which colonialism
shaped institutions and values, and thus the future development of the con-
tinent. Important research is, however, beginning to look more precisely at
these issues, for example, Frankema’s (2011) work on colonial educational
systems, and Gardner’s (2012) work on systems of taxation. For additional
evidence, we can turn to the works of scholars such as Anderson, Von der
Mehden, and Young (1967), Young (1994), Firmin-Sellers (1995, 1996), and
Boone (2003), who provide qualitative accounts and in-depth studies of the
use of power in the colonial period. Important too is the recent research of
Zaaruka and Fedderke (2010, 2011). Heldring and Robinson (2013) discuss
the problems involved in assessing the impact of colonialism on African
development (particularly the issues of the absence of clear counterfactu-
als), while nonetheless concluding that it is diicult to believe that colo-
nialism stimulated development. hough Botswana has been an economic
success since independence, it would surely have been a lot richer today
had it not been colonized by the British (unless of course the counterfactual
is that it would have instead been taken over by the British South Africa
Company).
Colonialism did not just impede development, it also distorted it. he
occupation of Africa resulted in a shit in control over war making, justice,
taxation, and expenditures – that is, the core issues of politics – from the
hands of the chiefs and headmen to those of the foreign occupier. As Bates
and Reid argue, it also inserted colonial oicials into the set of principals
whom the local power holders needed to please to remain in oice (see
also Gluckman, Mitchell, and Barnes 1949). he irst change weakened the
incentives to invest efort in costly projects: traditional leaders now lacked
the resources to mount them. he second expanded the opportunities for
oiceholders to advance their private agendas. When negotiating with one
principal – say, the colonial oice – a chief could plead pressure from the
royal family. When arguing with indigenous power holders, in turn, he
could cite directives from the occupier. he possibilities for self-serving
maneuvering thus increased, enabling indigenous oiceholders to gain pri-
vately even though their status and power had declined. In short, colonial-
ism appears to have divorced political ambition from public service.
Pondering this argument, we can return to one of Bates’s sources, the
Standard Cross-Cultural Sample, which lists what it calls the “focal year”
20 Akyeampong, Bates, Nunn, and Robinson

of the sources coded. For a given society, the “focal year” constitutes the
central tendency of the annual distribution of the accounts from which the
data were taken. Drawing from the Africa portion of this sample, we can
compare data for societies that were observed during the nineteenth cen-
tury with those observed during the twentieth: periods that roughly corre-
spond to before and ater the consolidation of imperial rule. When doing
so, we focus on societies that possessed executives or centralized political
institutions. We then ind that during the nineteenth century, that is, before
colonial rule, in 44 percent of societies with executives, the executives had
to rule through councils. However, during the twentieth century, that is,
during the colonial period, in only 25 percent of such societies were the
executives similarly constrained. During the nineteenth century, moreover,
89 percent of the centralized societies possessed a judiciary. During the
twentieth century, that is, ater the introduction of colonialism, the per-
centage fell to 40 percent. hese – and other – measures thus suggest that
while Africa’s political elites lost power while under colonial occupation,
they were also freed of traditional impediments to the self-serving use of
what powers they retained.
It is important to emphasize the scarcity of the data. he sample contains
fewer than thirty societies in all. But the direction of the indings corre-
sponds with the position taken in the qualitative literature: Africa’s political
institutions changed during the colonial period, and in a way that facilitated
the self-interested use of power. Independence may therefore have found
Africa governed by political institutions that had been rendered more
authoritarian by colonial rule and people more inclined to view political
elites as self-seeking (Cooper 2002).

CONCLUSIONS
he chapters in this volume ofer rich insight into the nature and trajectory
of Africa’s development. If one thing unites them, it is the belief that an
ahistorical understanding of Africa’s current economic problems is of little
value. Understanding Africa’s relative standing in the global distribution
of wealth entails understanding its development path and the legacies that
shaped it. he absence of such an understanding undermines the eforts of
practitioners to “save Africa,” a point Hopkins (1986) lucidly made some
time ago. And its presence will surely inspire new scholarship. By heighten-
ing interest in Africa among those with an interest in comparative historical
development, we surely will induce creative insights, as those who study
other regions encounter a history so diferent from their own.
 Introduction 21

Many important areas of African economic history have been well

researched, and do not appear in this volume given our focus. One is the
extent to which economic factors inluenced the creation of colonialism and
subsequently decolonization (Hopkins 1973; Austen, 1987, and Fieldhouse,
1986, 1994, 1999, review many of the theories, and see Milburn, 1977, and
Marseille, 1984, as well as Cooper, 1996, 2005, for important monographs).
his introduction and the chapters of this book also neglect the impact of
Africa’s economic development on Europe and the Industrial Revolution, an
important theme since the work of Williams (1944). (Also see Acemoglu,
Johnson, and Robinson 2005; Inikori 2002; Lynn 1997.) Labor issues, partic-
ularly the transition from slavery to free labor and the forms of labor coer-
cion used during the colonial period, surface only transitorily (see Austin
2005; Cooper 1980, 1987). Important and vivid as well is the regional eco-
nomic history of southern Africa (see Houghton, 1976, and Feinstein, 2005,
for overviews), which includes important studies of mining (Worger 1987),
business (Mager 2010), and labor (Bundy 1979; Van der Horst 1942; and
Van Onselen 1982).
Our intention in putting together the conference and in editing this vol-
ume was not to provide a comprehensive guide to the enormous amount
of high-quality research on many aspects of African economic history and
development. It was rather to focus on the key themes we have stressed
throughout and to inspire students and scholars to take this rich base of
evidence and ideas and build it more efectively into a convincing compar-
ative narrative.

References
Acemoglu, Daron, Simon Johnson, and James A. Robinson. (2001). “he Colonial
Origins of Comparative Development: An Empirical Investigation.” American
Economic Review, 91(5): 1369–1401.
(2002). “Reversal of Fortune: Geography, Institutions in the Making of the Modern
World Income Distribution.” Quarterly Journal of Economics, 117(4): 1231–94.
(2003). “An African Success Story: Botswana.” In In Search of Prosperity: Analytic
Narratives on Economic Growth, edited by Dani Rodrik, 80–122. Princeton, NJ:
Princeton University Press.
(2005). “Rise of Europe: Atlantic Trade, Institutional Change and Economic Growth.”
American Economic Review, 95: 546–79.
Acemoglu, Daron and James A. Robinson. (2010). “Why is Africa Poor?” Economic
History of Developing Regions, 25: 21–50.
(2012). Why Nations Fail. New York: Crown.
Allen, Robert C. (2009). he British Industrial Revolution in Global Perspective. New
York: Cambridge University Press.
22 Akyeampong, Bates, Nunn, and Robinson

(2011). Global Economic History: A Very Short Introduction. Oxford: Oxford University
Press.
Amin, Samir. (1974). Accumulation on a World Scale, 2 vols. New York: Monthly Review
Press.
(1976). Unequal Development. New York: Monthly Review Press.
Anderson, Charles W., Fred R. von der Mehden, and Crawford Young. (1967). Issues of
Political Development. Englewood Clifs NJ: Prentice-Hall.
Austen, Ralph A. (1987). African Economic History: Internal Development and External
Dependency. London: Heinemann.
Austen, Ralph A., and Daniel Headrick. (1983). “he Role of Technology in the African
Past.” African Studies Review, 26: 163–84.
Austin, Gareth. (2005). “Labour, Land, and Capital in Ghana: From Slavery to Free
Labour in Asante 1807–1956.” Rochester: University of Rochester Press.
Austin, Gareth. (2008). “Resources, Techniques, and Strategies South of the Sahara:
Revising the Factor Endowments Perspective on African Economic Development,
1500–2000.” Economic History Review, 61: 587–624.
Baland, Jean-Marie, Catherine Guirkinger, and Charlotte Mali. (2011). “Pretending
to Be Poor: Borrowing to Escape Forced Solidarity in Cameroon.” Economic
Development and Cultural Change, 60(1): 1–16.
Barker, Graeme. (2006). he Agricultural Revolution in Prehistory: Why Did Foragers
Become Farmers? New York: Oxford University Press.
Bates, Robert H. (1981). Markets and States in Tropical Africa. Berkeley : University of
California Press.
(2001). Violence and Prosperity. New York: W. W. Norton.
Batten, T. R. (1954). Problems of African Development. London: Oxford University Press.
Beattie, J. H. M. (1959). “Checks on the Abuse of Political Power in some African States.”
Sociologus, 9, 97–115.
Bezemer, Dirk J., Jutta Bolt, and Robert W. Lensink (2012). “Slavery, Statehood and
Economic Development in Sub-Saharan Africa.” African Economic History
Network Working Paper no. 6.
Birmingham, David and Richard Gray, eds. (1970). Pre-Colonial Trade: Essays on Trade
in Central and Eastern Africa before 1900. London: Oxford University Press.
Birnberg, homas B. and Stephen A. Resnick. (1975). Colonial Development: An
Econometric Study, New Haven, CT: Yale University Press.
Bohannan, Paul and George Dalton, eds. (1962). Markets in Africa. Evanston, IL:
Northwestern University Press.
Boone, Catherine. (2003). Political Topographies of the African State: Rural Authority
and Institutional Choice. Cambridge: Cambridge University Press.
Brenner, Robert. (1976). “Agrarian Class Structure and Economic Development in Pre-
Industrial Europe.” Past and Present, 70: 30–75.
Brewer, John. (1988). he Sinews of Power: War, Money and the English State, 1688–1783.
New York: Knopf.
Bruhn, Miriam and Francisco Gallego. (2012). “Good, Bad, and Ugly Colonial Activities:
Do hey Matter for Economic Development?” Review of Economics and Statistics,
94(2): 433–61.
Bundy, Colin. (1979). he Rise and Fall of the South African Peasantry. Berkeley :
University of California Press.
 Introduction 23

Chanda, Areendam and Louis Putterman. (2007). “Early Starts, Reversals and Catch-Up
in the Process of Economic Development.” Scandinavian Journal of Economics, 109:
387–413.
Coatsworth, John H. (1978). “Obstacles to Economic Growth in Nineteenth-Century
Mexico.” American Historical Review, 83(1): 80–100.
Cogneau, Denis and Léa Rouanet. (2011). “Living Conditions in Côte d’Ivoire and
Ghana 1925–1985: What Do Survey Data on Height Stature Tell Us.” http://www
.parisschoolofeconomics.eu/fr/cogneau-denis/.
Colson, Elizabeth L. (1969). “African Society at the Time of the Scramble.” In Colonialism
in Africa, 1870–1960. Vol. 1: he History and Politics of Colonialism, edited by Lewis
H. Gann and Peter Duignan, 27–65. Cambridge: Cambridge University Press.
(1971). “he Impact of the Colonial Period on the Deinition of Land Rights.” In
Colonialism in Africa, 1870–1960. Vol. III: Proiles of Change: African Society and
Colonial Rule, edited by Victor Turner, 193–215. Cambridge: Cambridge University
Press.
Cooper, Frederick. (1980). From Slaves to Squatters: Plantation Labor and Agriculture in
Zanzibar and Coastal Kenya, 1890–1925. New Haven, CT: Yale University Press.
(1987). On the African Waterfront: Urban Disorder and the Transformation of Work in
Colonial Mombasa. New Haven, CT: Yale University Press.
(1996). Decolonization and African Society: he Labor Question in French Africa. New
York: Cambridge University Press.
(2002). Africa since 1940: he Past of the Present. New York: Cambridge University
Press.
Cooper, Frederick. (2005). Colonialism in Question. Berkeley : University of California
Press.
Crummey, Donald. (1980). “Abyssinian Feudalism.” Past and Present, 89(1): 115–38.
Curtin, Philip D. (1969). he Atlantic Slave Trade: A Census. Madison: University of
Wisconsin Press.
(1975). Economic Change in Precolonial Africa: Senegambia in the Era of the Slave
Trade. Madison: University of Wisconsin Press.
Dalton, George H. (1976). “Review of A. G. Hopkins, Economic History of West Africa.”
African Economic History, 1: 51–101.
Dell, Melissa. (2010). “he Persistent Efects of Peru’s Mining Mita.” Econometrica,
78(6): 1863–1903.
Douglas, Mary. (1962). “Lele Economy Compared to the Bushong.” In Markets in Africa,
edited by Paul Bohannan and George Dalton, 211–33. Evanston, IL: Northwestern
University Press.
(1963). he Lele of the Kasai. Oxford: Oxford University Press.
Easterly, William. (2001). he Elusive Quest for Growth. Cambridge: MIT Press.
Easterly, William and Ross Levine. (1997). “Africa’s Growth Tragedy: Policies and Ethnic
Divisions.” Quarterly Journal of Economics, 112(4): 1203–50.
Englebert, Pierre. (2000). State Legitimacy and Development in Africa. Boulder, CO:
Lynne Rienner Publishers.
Engerman, Stanley L. and Kenneth L. Sokolof. (1997). “Factor Endowments, Institutions,
and Diferential Paths of Growth among New World Economies.” In How Latin
America Fell Behind, edited by Stephen Haber, 260–306. Stanford, CA: Stanford
University Press.
24 Akyeampong, Bates, Nunn, and Robinson

(2011). Economic Development in the Americas since 1500: Endowments and

Institutions. New York: Cambridge University Press.
Engelbert, Pierre. (2000). State Legitimacy and Development in Africa. Boulder, CO:
Lynne Rienner Publishers.
Evans-Pritchard, E. E. and Meyer Fortes. (1940). African Political Systems. Oxford
University Press.
Ewald, Janet. (1988). “Speaking, Writing and Authority: Explorations in and from the
Kingdom of Taqali.” Comparative Studies in History and Society, 30(2): 199–224.
Fage, John. (1969). “Slaving and the Slave Trade in the Context of West African History.”
Journal of African History, 10: 393–404.
Feinstein, Charles H. (2005). An Economic History of South Africa. New York: Cambridge
University Press.
Feldman, Noah. (2008). he Fall and Rise of the Islamic State. Princeton, NJ: Princeton
University Press.
Fieldhouse, David K. (1986). Black Africa, 1945–1980: Economic Decolonization and
Arrested Development. New York: Routledge.
(1994). Merchant Capital and Economic Decolonization: he United Africa Company
1929–1987. Oxford: Clarendon Press.
(1999). he West and the hird World: Trade, Colonialism, Dependence and
Development. New York: Wiley.
Findlay, R. and K. H. O’Rourke. (2007). Power and Plenty: Trade, War and the World
Economy in the Second Millennium. Princeton, NJ: Princeton University Press.
Finley, Moses. (1965). “Technical Innovation and Economic Progress in the Ancient
World.” Economic History Review, 18: 29–45.
Firmin-Sellers, Katherine. (1995). “he Politics of Property Rights.” American Political
Science Review, 89(4): 867–82.
(1996). he Transformation of Property Rights in the Gold Coast. Cambridge:
Cambridge University Press.
Frankema, Ewout. (2011). “Colonial Taxation and Government Spending in British
Africa, 1880–1940: Maximizing Revenue or Minimizing Efort?,” Explorations in
Economic History, 48(1): 136–49.
(2012). “he Origins of Formal Education in Sub-Saharan Africa: Was British Rule
more Benign?” European Review of Economic History, 16(4): 335–55.
Frankema, Ewout and Marlous van Waijenburg. (2012). “Structural Impediments to
African Growth? New Evidence from British African Real Wages, 1880–1965,”
Journal of Economic History, 72(4): 895–926.
Gardner, Leigh. (2012). Taxing Colonial Africa: he Political Economy of British
Imperialism. Oxford: Oxford University Press.
Gennaioli, Nicola with Ilia Rainer. (2007). “he Modern Impact of Pre-Colonial
Centralization in Africa.” Journal of Economic Growth, 12(3): 185–234.
Gluckman, Max, J. C. Mitchell, and J. A. Barnes. (1949). “he Village Headman in British
Central Africa.” Africa, 19(2): 89–106.
Goldstein, Markus and Christopher Udry. (2008). “he Proits of Power: Land Rights and
Agricultural Investment in Ghana.” Journal of Political Economy, 116: 981–1022.
Goody, Jack. (1971). Technology, Tradition and the State in Africa. New York: Cambridge
University Press.
 Introduction 25

(1986). he Logic of Writing and the Organization of Society. New York: Cambridge
University Press.
Habyarimana, J., M. Humphreys, Daniel Posner, and Jeremy Weinstein. (2007). “Why
Does Ethnic Diversity Undermine Public Goods Provision?” American Political
Science Review, 101(4): 709–25.
Haour, Anne. (2000). Rulers, Warriors, Traders, Clerics: he Central Sahel and the North
Sea, 800–1500. Oxford: Oxford University Press.
Heldring, Leander and James A. Robinson. (2013). “Colonialism and Development.” In
Handbook on the Politics of Development, edited by Carol Lancaster and Nicolas
van de Walle. New York: Oxford University Press.
Herbert, Eugenia W. (1993). Iron, Gender, and Power: Rituals of Transformation in
African Societies. Bloomington: Indiana University Press.
Herbst, Jefrey I. (2000). States and Power in Africa. Princeton, NJ: Princeton University
Press.
Hill, Polly. (1963). he Migrant Cocoa Farmers of Southern Ghana: A Study in Rural
Capitalism. New York: Cambridge University Press.
Hopkins, Anthony G. (1973). An Economic History of West Africa. New York: Addison
Wesley Longman.
(1986). “he World Bank in Africa: Historical Relections on the African Present.”
World Development, 14(12): 1473–87.
(2009). “he New Economic History of Africa.” Journal of African History, 50:
155–77.
Horst van der, Sheila. (1942). Native Labour in South Africa. London: Oxford University
Press.
Houghton, D. Hobart. (1976). he South African Economy. 4th ed. Cape Town: Oxford
University Press.
Hydén, Göran. (1980). Beyond Ujamaa in Tanzania: Underdevelopment and an
Uncaptured Peasantry. Berkeley : University of California Press and Heinemann.
(1983). No Shortcuts to Progress. Berkeley : University of California Press and
Heinemann.
IMF. (2009). “World Economic Outlook: Crisis and Recovery.” In World Economic and
Financial Survey, he International Monetary Fund.
Inikori, Joseph E. (1992). he Chaining of a Continent: he Export Demand for Captives
and the History of Africa South of the Sahara, 1450–1870. Mona: University of the
West Indies.
(1994). “Ideology versus the Tyranny of Paradigm: Historians and the Impact of the
Atlantic Slave Trade on African Societies.” African Economic History, 22: 37–58.
(2002). Africans and the Industrial Revolution in England. New York: Cambridge
University Press.
Jerven, Morten. (2010). “African Growth Recurring: An Economic History Perspective
on African Growth Episodes, 1690–2010.” Economic History of Developing Regions,
25(2): 127–54.
Jones, Eric. (1981). he European Miracle. New York: Cambridge University Press.
Kasekende, Louis, Zuzana Brixova, and Leonce Ndikumana. (2010). “Africa: Africa’s
Counter-Cyclical Policy Responses to the Crisis.” Journal of Globalization and
Development, 1(1): Article 16.
26 Akyeampong, Bates, Nunn, and Robinson

Killick, Tony. (1978). Development Economics in Action. London: Heinemann.

Kriger, Colleen E. (2006). Cloth in West Africa. Lanham, MD: AltaMira Press.
Lange, Matthew. (2004). “British Colonial Legacy and Political Development.” World
Development, 32(6): 905–22.
(2009). Lineages of Despotism and Development: British Colonialism and State Power.
Chicago: University of Chicago Press.
Law, Robin C. (1977). “Royal Monopoly and Private Enterprise in the Atlantic Trade:
he Case of Dahomey.” Journal of African History, 18: 555–77.
(1978). “In Search of a Marxist Perspective on Pre-Colonial Tropical Africa.” Journal
of African History, 19, 441–52.
(1980). “Wheeled Transportation in Pre-Colonial West Africa.” Africa, 50: 249–62.
(1981). he Horse in West African History. Oxford: Oxford University Press.
Law, Robin C., ed. (2002). From Slave Trade To “Legitimate” Commerce: he Commercial
Transition in Nineteenth-Century West Africa. New York: Cambridge University
Press.
Leith, Clark J. (2005). Why Botswana Prospered. Montreal: McGill University Press.
Lovejoy, Paul E. (1989). “he Impact of the Atlantic Slave Trade on Africa: A Review of
the Literature.” Journal of African History, 30: 365–94.
(2000). Transformations in Slavery: A History of Slavery in Africa. 2nd ed. New York:
Cambridge University Press.
Long, Norman. (1978). Social Change and the Individual: A Study of the Social and
Religious Responses to Innovation in a Zambian Rural Community. Manchester:
Manchester University Press.
Lynn, Martin. (1997). Commerce and Economic Change in West Africa. New York:
Cambridge University Press.
Mager, Anne. (2010). Beer, Sociability and Masculinity in South Africa. Bloomington:
Indiana University Press.
Mamdani, Mahmood. (1996). Citizen and Subject: Contemporary Africa and the Legacy
of Late Colonialism. Princeton, NJ: Princeton University Press.
Manning, Patrick. (1990). Slavery and African Life. Cambridge: Cambridge University
Press.
(2004). Slavery, Colonialism and Economic Growth in Dahomey, 1640–1960. New
York: Cambridge University Press.
McGreevey, William P. (1971). An Economic History of Colombia, 1845–1930, New York:
Cambridge University Press.
McIntosh, Roderick J. (1988). he Peoples of the Middle Niger. Malden: Blackwell
Publishers.
McIntosh, Susan K. (1999). “Pathways to Complexity: An African Perspective.” In
Beyond Chiefdoms: Pathways to Complexity in Africa, edited by S. K. McIntosh,
1–30. Cambridge University Press.
McPhee, Allan. (1926). he Economic Revolution in British West Africa. London: George
Routledge & Sons.
Marseille, Jacques. (1984). Empire Colonial et Capitalisme Français. Paris: A. Michel.
Metcalfe, G. E. (1964). Great Britain and Ghana: Documents of Ghana History, 1807–
1957. London: T. Nelson for the University of Ghana.
Meillassoux, C., ed. (1971). he Development of Indigenous Trade and Markets. London:
International African Institute.
 Introduction 27

Michalopoulos, Stelios and Elias Papaioannou. (2013). “Pre-Colonial Ethnic Institutions

and Contemporary African Development.” Econometrica, 81(1): 113–52.
Miers, Suzanne and Richard L. Roberts, eds. (1988). he End of Slavery in Africa.
Madison: University of Wisconsin Press.
Miguel, Edward and Mary K. Gugerty. (2005). “Ethnic Diversity, Social Sanctions, and
Public Goods in Kenya.” Journal of Public Economics, 89: 2325–68.
Milburn, Josephine. (1977). British Business and Ghanaian Independence. London:
Hurst.
Miller, Joseph C. (1996). Way of Death: Merchant Capitalism and the Angolan Slave
Trade, 1730–1830. Madison: University of Wisconsin Press.
(2009). “Africa in World History: Lecture 1 History without Horses.” W.E.B. du Bois
Lectures Presented at Harvard University.
Mkandawire, handika and Charles Soludo, eds. (2003). African Voices on Structural
Adjustment. Dakar: Codesria.
Mokyr, Joel. (2008). “he Institutional Origins of the Industrial Revolution.” In
Institutions and Economic Performance, edited by E. Helpman, 64–119. Cambridge,
MA: Harvard University Press.
Mokyr, Joel. (2009). he Enlightened Economy. New Haven, CT: Yale University Press.
Monga, Celestin. (2006). “Commodities, Mercedes Benz and Structural Adjustment:
An Episode in West African History.” In hemes in West Africa’s History, edited by
Emmanuel Akyeampong, 227–64. Athens: University of Ohio Press.
Moradi, Alexander. (2008). “Confronting Colonial Legacies – Lessons from Human
Development in Ghana and Kenya.” Journal of International Development, 20:
1107–21.
(2009). “Towards an Objective Account of Nutrition and Health in Colonial Kenya:
A Study of Stature in African Army Recruits and Civilians, 1880–1980.” Journal of
Economic History, 96(3): 720–55.
Morris, Ian. (2011). Why the West Rules – For Now. New York: Picador.
Mosley, Paul. (1983). he Settler Economies: Studies in the Economic History of Kenya
and Southern Rhodesia 1900–1963. New York: Cambridge University Press.
Ndulu, Benno, Stephen A. O’Connell, Robert H. Bates, Paul Collier, and Chukwuma
C. Soludo, eds. (2007). he Political Economy of Economic Growth in Africa, 1960–
2000, 2 vols. New York: Cambridge University Press.
Nunn, Nathan. (2008a). “he Long-Term Efect of Africa’s Slave Trades.” Quarterly
Journal of Economics, 23(1): 139–76.
(2008b). “Slavery, Inequality and Economic Development in the Americas: An
Examination of the Engerman-Sokolof Hypothesis.” In Institutions and Economic
Performance, edited by Elhanan Helpman, 148–80. Cambridge, MA: Harvard
University Press.
Nunn, Nathan and Leonard Wantchekon. (2011). “he Slave Trade and the Origins of
Mistrust in Africa.” American Economic Review, 101(7): 3221–52.
Osafo-Kwaako, Philip and James A. Robinson. (2013). “Political Centralization in Pre-
Colonial Africa.” Journal of Comparative Economics, 41(1): 6–21.
Palmer, Robin and Q. Neil Parsons, eds. (1977). he Roots of Rural Poverty in Central
and Southern Africa. Berkeley : University of California Press.
Parkin, David. (1972). Palms, Wine, and Witnesses: Public Spirit and Private Gain in an
African Farming Community. San Francisco: Chandler Publishing Co.
28 Akyeampong, Bates, Nunn, and Robinson

Phillipson, David W. (1998). Ancient Ethiopia. London: British Museum Press.

Pirenne, Henri. (2001). Mohammed and Charlamagne. New York: Dover.
Polanyi, Karl. (1966). Dahomey and the Slave Trade: An Analysis of an Archaic Economy.
Seattle: University of Washington Press.
Polanyi, Karl, Conrad M. Arensberg, and Harry W. Pearson, eds. (1957). Trade and
Market in the Early Empires: Economies in History and heory. New York: he Free
Press.
Pomeranz, Kenneth. (2000). he Great Divergence: China, Europe and the Making of the
Modern World. Princeton, NJ: Princeton University Press.
Postan, M. M. (1937). “he Chronology of Labour Services.” Transactions of the Royal
Historical Society, 70: 169–93.
Prados de la Escosura, Leandero. (2011). “Human Development in Africa: A Long-
Run Perspective.” University Carlos III, Working Papers in Economic History, WP
11–09.
Radelet, Steven. (2010). Emerging Africa: How 17 Countries Are Leading the Way.
Washington, DC: Center for Global Development.
Roberts, Michael. (1956). he Military Revolution 1560–1660: An Inaugural Lecture
Delivered before the Queen’s University of Belfast. Belfast: M. Boyd.
Robinson, James A. (2012). “Why Did Africa Export Scarce Labor?” Unpublished.
Rodney, Walter. (1972). How Europe Underdeveloped Africa. London: Bogle-L’Ouverture
Publications.
Rosenthal, Jean-Laurant and R. Bin Wong. (2011). Before and beyond Divergence:
he Politics of Economic Change in China and Europe. Cambridge, MA: Harvard
University Press.
Sachs, Jefrey and Andrew Warner. (1997). “Sources of Slow Growth in African
Economies.” Journal of African Economies, 6: 335–76.
Scott, James C. (2009). he Art of Not Being Governed: An Anarchist History of Upland
Southeast Asia. New Haven, CT: Yale University Press.
Sender, John and Sheila Smith. (1986). he Development of Capitalism in Africa. London:
Methuen.
Sherif, Abdul. (2010). Dhow Cultures and the Indian Ocean: Cosmopolitanism,
Commerce, and Islam. Amsterdam: KIT Publishers.
Speke, John H. (1863). he Discovery of the Source of the Nile. London: William
Blackwood and Sons.
Stamp, L. Dudley. (1953). Africa: A Study in Tropical Development. London: John Wiley
& Sons.
Suret-Canale, Jean. (1958). Afrique Noire, Occidentale et Centrale: Geographie,
Civilisations. Histoire, Paris: Éditions Sociales.
hornton, John K. (1992). “Pre-Colonial African Industry and the Atlantic Trade, 1500–
1800.” African Economic History Review, 19: 1–19.
(1998). Africa and Africans in the Making of the Atlantic World, 1400–1800. Cambridge:
Cambridge University Press.
Tilly, Charles, ed. (1975). Relections on the History of State Making. he Formation of
Nation States in Western Europe. Princeton, NJ: Princeton University Press.
Van de Walle, Nicolas. (2001). African Economies and the Politics of Permanent Crisis,
1979–1999. New York: Cambridge University Press.
 Introduction 29

Van Onselen, Charles. (1982). Studies in the Social and Economic History of the
Witwatersrand, 1886–1914, 2 vols. New York: Longman.
Van Zanden, Jan Luiten. (2009). he Long Road to the Industrial Revolution: he European
Economy in a Global Perspective, 1000–1800. Leiden: Brill Publishers.
Vansina, Jan. (1978). he Children of Woot: A History of the Kuba Peoples. Madison:
University of Wisconsin Press.
(1985). Oral Tradition as History. Madison: University of Wisconsin Press.
(2004). How Societies Are Born: Governance in West Central Africa before 1600.
Charlottesville: University of Virginian Pres.
Wallerstein, Immanuel. (1974–2011). he Modern World System, 4 vols. New York:
Academic Press.
Wickens, Peter L. (1980). An Economic History of Africa from the Earliest Times to
Partition. Cape Town: Oxford University Press.
(1986). Africa, 1880–1980: An Economic History. Cape Town: Oxford University
Press.
Williams, Eric. (1944). Capitalism and Slavery. University of North Carolina Press.
Wilks, Ivor. (1989). Asante in the Nineteenth Century: he Structure and Evolution of a
Political Order. Cambridge: Cambridge University Press.
Wilson, Francis. (1972). Labour in the South African Gold Mines, 1911–1969. Cambridge:
Cambridge University Press.
Wong, R. Bin. (1997). China Transformed: Historical Change and the Limits of European
Experience. Ithaca, NY: Cornell University Press.
Worger, William. (1987). South Africa’s City of Diamonds. New Haven, CT: Yale
University Press.
World Bank. (1994). Adjustment in Africa: Reform, Results, and the Road Ahead. New
York: Oxford University Press.
Young, Crawford. (1994). he African Colonial State in Comparative Perspective. New
Haven, CT: Yale University Press.
Zaaruka, Bernie P. and Johannes W. Fedderke. (2010). “Indicators of Economic and
Political Institutions in Tanzania: 1884–2008.” ERSA Working Paper 231.
(2011). “Measuring Institutions: Indicators of Political and Economic Institutions in
Namibia: 1884–2008.” University of Witswatersrand, Department of Economics,
Working Paper 236.
 PART I

LA LONGUE DUR É E
 1

Africa in World History before ca. 1440

Christopher Ehret

1.1 THE AFRICAN ORIGINS OF HISTORY

Barely more than ity thousand years ago, the ancestors of every single
human being alive today lived in Africa. World history to that point was
African history. hat is now becoming accepted knowledge. Less generally
understood is that, just because a few Africans let the continent around
ity thousand years ago and began to expand across the rest of the globe,
history did not come to a halt in Africa. he Africans who stayed behind in
our ancestral continent did not fall out of time into some kind of ahistorical
stasis. hey passed through the same great transitions of human history –
from foraging to agricultural ways of life; from small-scale, local communi-
ties to societies of towns and states; and from localized, irregular exchanges
of surplus to complex systems of formal, long-distance commerce – and
they passed through those transitions in the same broad periods of time as
did people elsewhere in the world.

1.2 THE FIRST GREAT TRANSITION

For thirty-ive thousand years, human beings, wherever they spread in the
world, continued to be gatherers and hunters of wild food. hen, separately
in diferent parts of the world, the climatic shits at the end of the latest ice
age set of a long, episodic “First Great Transition” of human history, from
foraging to food production – from the gathering and hunting of wild food
to the deliberate tending and protection of animals and plants.

he author thanks Oxford University Press for permission to reuse several previously pub-
lished passages in this chapter.

33
34 Ehret

he initial warming of climate during the Bølling-Allerød interstadial,

12,700–10,900 BCE, brought increased rainfall and warmer conditions in
many African regions. hree sets of peoples, speaking languages of the three
language families that predominate across the continent today, probably
began their early expansions in this period. Nilo-Saharan peoples spread
out in the areas around and east of the middle Nile River in what are today
the countries of Sudan and South Sudan. Peoples of a second family, Niger-
Kordofanian, spread across an emerging east-west belt of savanna vegeta-
tion from the eastern Sudan to the western Atlantic coast of Africa. During
the same era, communities speaking languages of the Erythraic branch of
the Afrasian (Afroasiatic) family expanded beyond their origin areas in the
Horn of Africa, northward to modern-day Egypt.1
his last point needs special emphasis, considering how widespread the
notion is that the Afrasian family somehow originated in the Middle East.
he linguistic, genetic, and archaeological evidence combines in locating
the origins of this family far south in Africa, in Eritrea or Ethiopia, and not
at all in Asia.2 A complex array of lexical evidence conirms that the proto-
Afrasian society belonged to the preagricultural eras of human history.3
None of these episodes of expansion proceeded into entirely empty
lands. Other communities already lived in most areas into which the speak-
ers of these three families of languages spread. An issue deserving of future
historical attention is the question of what kinds of advantages each set of
societies might have had that allowed their expansions into the lands of
others. What little is known of the archaeology of that era in the middle
Nile River regions suggests either or both of two developments may have
contributed to the earliest spread of Nilo-Saharans. One would have been
the expansion of grasslands northward toward the Sahara, opening up large
new areas to the hunting of large herbivores. he other would have been the
increased low of the Nile, which would have enhanced the attractions of a
more ancient regional economy based on the exploitation of aquatic food
resources. In the case of the Afrasians, the reconstructed ancient Afrasian
lexicon of subsistence indicates that they collected wild grains. he spread
of grassland areas in the Red Sea hills and Egypt away from the Nile during
the thirteenth millennium BCE would have opened up new areas to their
kind of subsistence.4 What subsistence advantages might have favored the

1
Ehret 2006, 2008, 2011, 2013.
2
Ehret and colleagues (2004) concisely summarize the convergent indings of these three
ields.
3
Ehret (2006) provides a detailed exposition.
4
Ehret (2006); Ehret and colleagues (2004).
 Africa in World History before ca. 1440 35

spread of the Niger-Kordofanians is a question the historical literature has

yet to consider.
he return of colder, drier conditions in the Younger Dryas, 10,900–9500
BCE, would have brought these expansions to a halt and in some cases
reversed them. At the same time, however, the period of climatic setback
encouraged a new round of subsistence innovation before fully postglacial
conditions took hold in the tenth millennium BCE. In a few areas of the
world, people began the irst protecting of plants or animals, in this fashion
laying the earliest foundations for agriculture. From 9500 BCE onward, sep-
arately and independently in diferent parts of the world, the stage-by-stage
developments toward agriculture quantitatively transformed the directions
of human history. he deliberate tending of plants and animals multiplies
by magnitudes the amounts of food potentially obtainable from the same
amount of land. he growth of human populations from a few hundred
thousand to billions, social stratiication, urban life, states, and the develop-
ment of complex technologies all rest on the creation of agriculture.
Africans actively contributed to these new directions in world history.
In Africa, peoples of two distant parts of the continent’s middle belt inde-
pendently set in motion developments leading to agriculture, and they did
so during the same broad era as did peoples in the Middle East, India, East
Asia, New Guinea, and the Americas.
No later than the middle of the tenth millennium in the savannas of mod-
ern-day Mali, communities speaking the early daughter languages of proto-
Niger-Congo, itself an ofshoot of the Niger-Kordofanian family, began to
intensively collect wild grains, among them probably fonio. heir Ounjougou
culture is the earliest identiied facies of the West African Microlithic,5 the
archaeological complex associated with the early Niger-Congo peoples.6
Integral to their new subsistence system was their invention of one of the
two earliest ceramic technologies in world history sometime before 9400
BCE. Rather than grinding whole grains into lour, the Ounjougou people
made the grains edible by cooking them whole in pots.
When did the shit from gathering to the cultivation of grains begin
among Niger-Congo peoples? he archaeobotanical evidence is as yet
uncertain or unknown for the crucial periods. Provisional reconstruc-
tions of several early Niger-Congo verbs speciically connoting cultiva-
tion suggest, however, that the transition from collecting to cultivating
grains in the grassland savannas of West Africa took place broadly in the

5
Huysecom and colleagues (2009).
6
Ehret 2002, 44.
36 Ehret

period 9000–6000 BCE.7 Black-eyed peas and African groundnuts (Vigna

subterranea), along with grains, may have been the earliest protected plants
of this incipient agriculture. Niger-Congo peoples domesticated both
black-eyed peas and groundnuts by or before the fourth millennium BCE,
earlier than can yet be demonstrated directly from the archaeology. he
linguistic testimony shows that they were among the food crops the earli-
est Bantu took with them into the equatorial rainforests ater 3000 BCE,8
and both crops had spread all the way to East Africa equally early; we can
reconstruct words for both in the proto-Southern Cushitic language of
northern Kenya, spoken around the late fourth millennium BCE.
West African agricultural history entered a new stage around the sixth mil-
lennium BCE, with the cultivation of two new crops, Guinea yams and oil
palms. he technological signature of this development was the adding of
polished stone axes to the West African Microlithic toolkit.9 he new crops
and tools opened the way for communities of the Benue-Kwa branch of
Niger-Congo to spread between 5000 and 3000 BCE into the rainforest zones
of West Africa, from modern-day Côte d’Ivoire to Cameroon. With polished
stone axes, they could clear forest for raising yams and oil palms, both of
which require direct sunlight.10 An additional technological innovation prob-
ably dating to this period was the Benue-Kwa invention of broadlooms for
weaving raia-cloth. Ater 3000 BCE, one ofshoot of the Benue-Kwa group,
the Bantu, carried the yam-based variety of West African agriculture farther
southward and eastward through the equatorial rainforests of central Africa.
Far to the east, the Northern Sudanians, a Nilo-Saharan people of the
southern eastern Sahara, took a very diferent irst step toward agricul-
ture. In the mid-tenth millennium BCE, a belated shit to wetter conditions
spread Mediterranean climate, with cool-season rains and Mediterranean
wild animals, most notably the cow, south to the middle of the Sahara.
Contemporaneously, tropical grassland and steppe environments advanced
north to the middle of the Sahara. he Northern Sudanians, following the
climatic shit northward, encountered cattle at the interface of the two
climatic regimes and, between 8500 and 7200 BCE, initiated the earliest
herding of cattle in world history.11 Like the Ounjougou people twenty-ive
hundred kilometers to the west, they collected wild grains, but, diferently,
they ground their grain into lour.

7
Ehret 2010.
8
Ehret 1998, 104–5.
9
Shaw 1980.
10
Ehret 2002, 84.
11
Wendorf and Schild 1998.
 Africa in World History before ca. 1440 37

Around 7200 BCE, a new development appeared in the eastern Saharan

archaeology: neighborhoods of substantial homesteads, with thornbush
cattle pens, round houses, and grain storage pits, and with sorghum as the
notable grain.12 he linguistic evidence in this case strongly backs these
indirect archaeological indicators that the Northern Sudanians of this era
had begun to cultivate. hey had important contacts, too, with the con-
temporary Afrasian communities immediately east of them in the Red Sea
Hills region. hese communities spoke early daughter dialects of the proto-
Cushitic language. In the second half of the seventh millennium, the north-
ernmost Cushites, ancestral to the modern-day Bedauye (the Medjay of the
ancient Egyptians), were the intermediaries in the difusion of sheep from
the Middle East to their Northern Sudanian neighbors. Even earlier, the
Cushites began, like the Northern Sudanians, to raise cattle, and they either
collected or cultivated sorghum.13
he Northern Sudanians of the ninth millennium, along with a closely
related set of Nilo-Saharan peoples whom archaeologist J. E. G. Sutton
called the “Aquatic Civilization of Middle Africa,”14 participated in a second
African invention of ceramic technology. he Aquatic societies responded
in a diferent fashion to the mid-ninth-millennium climatic amelioration.
hey became specialist ishing and hippopotamus-hunting peoples along
the new rivers and lakes of the Sahara, and in the later ninth millennium
they spread this economy westward across the southern Sahara as far as the
Hoggar mountains and the great bend of the Niger River.
Drier climates in the mid-Holocene then shrank many Saharan streams
and lakes, shiting the balance of advantage away from the Aquatic com-
munities. As a result, during the sixth millennium the descendants of the
Northern Sudanians spread their agripastoral economy across the southern
Sahara, displacing or assimilating many of the Aquatic communities. Where
perennial water resources existed, such as along the Nile, the Aquatic liveli-
hood persisted, but combined now with herding and probably cultivation.
he inhabitants of the Khartoum Neolithic site of 5000 BCE along the Nile
participated in a particularly notable invention, of cotton textile technol-
ogy, attested by their possession of spindlewhorls.15

12
Ibid.
13
Ehret 2011.
14
Sutton 1974.
15
For published plates showing the spindlewhorls, see Arkell (1949). I am indebted to
Merrick Posnansky for bringing these materials to my attention and explaining their
signiicance.
38 Ehret

As with ceramic technology, African societies were leaders in innovation

in the early agricultural eras. he history of cotton provides a lesson as well,
of wider relevance in history – that peoples with no knowledge whatsoever
of each other can and do arrive at parallel inventions. he domestication of
cotton as a iber plant for textile production took place separately in three
distant parts of the globe: the eastern Sudan of Africa, India, and the New
World. In each region, the inventors of cotton weaving domesticated their
own indigenous species of cotton. he evidence from Khartoum places this
development as early in Africa as in India.

1.3 THE ERA OF AGRICULTURAL ELABORATION,

6500–3500 BCE
he long period from 6500–3500 BCE typically marks a second stage in the
emergence of agricultural economies in world history. As in other world
regions, in Africa these centuries were a time of growth in the variety of
crops and animals that farmers raised, and a time of growth in the propor-
tional contribution of agriculture to the diet. he women farmers of Niger-
Congo civilization brought the two savanna legumes, the African groundnut
(Vigna subterranea) and the black-eyed pea (V. unguiculata), into cultivation
during this time. he Sudanic agripastoralists of the southern Sahara and the
Sahel added, it appears, melons and gourds of several kinds to their original
emphasis on sorghum, and they also began raising castor beans as a source
of a nonedible oil. Niger-Congo farmers and Sudanic agripastoralists appear
to have separately domesticated an additional major grain crop, pearl mil-
let (Pennisetum glaucum). In the Ethiopian highlands, reconstructed early
farming lexicon reveals that the Cushites during this time began to supple-
ment their stock raising with two highland African grain crops, inger millet
and t’ef, and probably a variety of gourds and calabashes.16
he seventh millennium to the fourth millennium BCE was also a period
of world history during which crops and animals domesticated in one or
another seminal region of agriculture irst spread to other such world regions.
In Africa between 6500 and 4000 BCE, Cushitic peoples domesticated the
donkey, native to the Red Sea Hills and the arid foothills of the northern
Ethiopian highlands.17 Donkeys then spread via Egypt to the Middle East,
where they became the earliest important beasts of burden in all of world
history. Sheep and goats, of course, difused in the opposite direction even

16
Ehret 2011.
17
Ibid.
 Africa in World History before ca. 1440 39

earlier, during the second half of the seventh millennium, and rapidly
became important animals in the Sudanic and Cushitic agripastoral tradi-
tions. From the Sudanic herders both goats and cattle spread west to the
Niger-Congo societies of West Africa, again at a still uncertain period, but
certainly well before 3000 BCE. Among the important early crop spreads,
melons, gourds, and castor, brought under cultivation by Nilo-Saharan
peoples of the Sudan belt and the southern Sahara, spread northward to the
ancient Egyptians probably no later than the fourth millennium.
An especially interesting historical problem far from solved is the ques-
tion of how three important grain crops domesticated in Africa, sorghum,
pearl millet, and inger millet, reached India between 3000 and 1000 BCE18
without passing through the Middle East irst. Might seagoing trade have
already connected northeastern Africa and India by that time? Sorghum
soon spread even farther eastward, becoming a crop of northern Chinese
agriculture before 1000 BCE.
What may surprise is that Egypt was not an initiating region of these sem-
inal developments. he indigenous Afrasian communities of the Egyptian
Nile in the seventh millennium still consisted of hunter-gathers. hey grad-
ually transformed their subsistence economy by adopting two staple crops,
barley and wheat, along with sheep and goats, which difused to them from
the ancient Middle Eastern center of domestication. Melons, gourds, and
donkeys reached them from the Sudanic agripastoralists to the south; sur-
prisingly, cotton did not. Word borrowings in ancient Egyptian, along with
recent archaeological discoveries, conirm that Sudanic herders also signif-
icantly inluenced Egyptian beliefs and practices relating to cattle.19

1.4 THE SECOND GREAT TRANSITION: THE AFRICAN

BEGINNINGS OF URBAN CENTERS AND STATES
By the ith millennium BCE, the growing variety and productivity of agri-
culture brought about a growth in the size and density of human popula-
tions, such that a “Second Great Transition,” from villages and tiny local
political units to towns and states, began to take place in several world
regions. Historians have long identiied Egypt as an early locus of this tran-
sition in the African continent. But because of the dominant Western idea
of Egyptian exceptionalism, historians have generally failed to note that the
formative area of ancient Egyptian culture, southern Upper Egypt, was the

18
Fuller 2006.
19
Ehret 2001.
40 Ehret

northern outlier of a wider nexus of emerging complexity in the fourth mil-

lennium. Predynastic Upper Egypt of the fourth millennium BCE did have
encounters at a distance with a separate nexus of town and state develop-
ment centered on the Fertile Crescent. But as recent archaeological discov-
eries reveal, the cultural and political world we identify as ancient Egypt
grew out of institutions and ideas already present among its Nilo-Saharan
neighbors in the ith and early fourth millennia BCE. he southern eastern
Sahara and eastern Sudan, though greatly neglected relative to Egypt by
scholars, remained the locus of towns and cities and powerful states from
the fourth millennium BCE until recent centuries.
he irst evidence of emerging complexity in the ith millennium
appeared not along the Nile itself, but in the then steppe country west of
northern Lower Nubia. Two hundred kilometers from the river, the inhab-
itants of Nabta Playa erected an extensive megalithic archaeoastronomical
array. he associated burials, of both cattle and people, reveal a wealthy
pastoral society, with a complex ritual basis, in existence centuries before
similar complexity and before the same speciic cultural features, such as
cattle burial, appeared in Upper Egypt.20
A further progression toward social and political complexity followed in
the fourth millennium BCE, this time along the Nile itself, with states and
the irst towns appearing between the Nile-Abbay conluence in the south
and southern Upper Egypt in the north. Town life along the river grew in
importance, even as the drying of the Sahara during the fourth millennium
brought the Nabta Playa culture to an end. Because of the relative archae-
ological neglect of Nubia, just two excavated sites, Shaheinab and Qustul,
provide most of our knowledge of this era south of Egypt. he two towns
lay respectively at the far northern and far southern ends of a thousand-
kilometer stretch of cultural commonality along the Nile. On sites of rit-
ual importance, the people of this Middle Nile culture built large conical
earthen mounds, reshaped since then by rain and wind into more formless-
seeming tumuli. Ritual sites of this type represent a very long-lived cultural
and political tradition, lasting in some cases until recent centuries.
Qustul was the capital of wealthy kings from the mid-fourth millennium
BCE almost until to the uniication of Egypt late in the millennium. Like
the earlier Nabta Playa pastoralist sites, the Qustul sites include numerous
cattle burials. Pictorial documents in the royal graves explicitly depict the
kings of the Qustul state as having conquered Upper Egypt. here is no a

20
Kobusiewicz and Schild (2005); Malville and colleagues (2008).
 Africa in World History before ca. 1440 41

priori reason to reject these claims. If one sets aside the received notion of
Egyptian exceptionalism, it is quite evident, as archaeologist Bruce Williams
argues, that here was a kingdom every bit as signiicant as its late predynas-
tic contemporaries in Upper Egypt.21
Behind the rise of the highly centralized kingship of dynastic Egypt may
have been an additional factor, the adoption in late predynastic Upper
Egypt of elements of the rituals and royal ideology of the Qustul kingdom.
Early Egyptian royal tombs, before the shit to pyramid building in stone,
were covered with a conical mound of earth, mimicking the practice known
as early as the fourth millennium in Nubia and still prevalent two thou-
sand years later in the kingdoms to the south. hese outward resemblances
accompany resemblances in ideology as well, from the special ritual signif-
icance accorded cattle to the claims of both Sudanic and Egyptian kings to
a degree of personal sacredness unparalleled in the Middle East. Did Upper
Egyptian rulers build their power in the later fourth millennium BCE by
adopting legitimizing ideas from Nabta Playa and Qustul? he outward
signs, at least, favor that proposal.
Two notable kingdoms persisted in Nubia through the Old Kingdom
period. he more powerful state, Kerma, ruled the Dongola Reach in
Upper Nubia and probably other lands farther south. he great fortiica-
tions at Buhen in Lower Nubia, built by the rulers of the Middle Kingdom
(2040–1700 BCE) ater their conquest of the northern Sai kingdom, sug-
gest an Egyptian concern with the potential threat from Kerma farther
south. he placement of Kerma’s capital at the northern end of its territo-
ries, closest to Egypt, may imply a similar concern in Kerma about Egypt.
Alternatively, it may indicate that the kings of Kerma wished to situate
their court so as to better oversee and control trade with Egypt. he mas-
sive royal funerary sites at Kerma City give a sense of the power of this
kingdom at its height. But as almost the sole major excavations relating
to the Kerma state, they leave us little idea of urban life more generally
in Kerma, and no knowledge of how much farther south Kerma’s power
might have extended.
In the late 1500s, hutmose I accomplished something new, a conquest
that extended Egyptian power into the Dongola Reach between the third
and fourth cataracts and imposed a thoroughgoing colonial rule over the
region. A common historical presumption is that this conquest destroyed
the Kerma kingdom. But was that the case? With the decline of Egyptian

21
Williams and Seele 1986.
42 Ehret

power in Dongola Reach during the twelth and eleventh centuries

BCE, many features of material culture reminiscent of Kerma’s high era
reemerged in the archaeological record. Somewhere to the south, beyond
the reach of Egyptian rule, the political and cultural traditions of Kerma
apparently persisted.
In the ninth century BCE, there arose south of Egypt a new major king-
dom, called Kush by Egyptians. Historians today give this kingdom two
names, Napata during the period until the sixth century when its capital
was at Napata City on the Dongola Reach, and Meroe ater the sixth cen-
tury, when the capital shited to the southern city of Meroe. Around 750
during the reign of Piye, Napata conquered Egypt, and Piye’s successors
ruled large parts of Egypt for much of the next century. he future capital,
Meroe City, already existed in the seventh century and surely lay within its
southern territories.
he Western scholarly tradition of Egyptian exceptionalism obscures a
startling fact. Piye’s immediate successors ruled over an empire probably
larger in territory than native Egyptian kings ever did, even at the height
of the New Kingdom. In Egypt, they tailored their religious relations and
political propaganda to appeal to their Egyptian subjects, and for their
eforts got recognition, from the parochial perspective of the Egyptians,
as Egypt’s twenty-ith dynasty. From the record they let behind in Egypt,
it is nevertheless evident that the Napata-Meroe rulers understood them-
selves as ethnically diferent and were not hesitant to represent themselves
as such in royal art. Nor were they hesitant to make changes in the relations
of political to religious power, with lasting efects even ater Egypt regained
its independence in the seventh century.22
he Napatan rule in Egypt from the mid-eighth to the mid-seventh cen-
tury was a conquest by a foreign power as much as were the subsequent
Assyrian and Persian conquests. he foreign power did not fade away just
because it eventually lost the conquered Egyptian lands. he lands from
the irst cataract to the Nile-Abbay conluence remained the territory of
one Napata-Meroe empire from the eighth century BCE until the third or
fourth century CE, a period of more than one thousand years. he empire
became a manufacturing center of cotton textiles in its several cities along
the Nile; Meroe City itself became a major iron-producing center by the
fourth or third century BCE. he kings built dams and encouraged new
irrigation techniques to enhance animal husbandry and cultivation, and the
state began to keep written records in its own Meroitic language.

22
Gordon 2009.
 Africa in World History before ca. 1440 43

1.4.1 Economy and Society in West Africa

A “hird Great Transition” in world history was the emergence of merchant-
managed commerce, an economic changeover beginning in the later sec-
ond millennium BCE that transformed the nature of economic transactions
and created the earliest forms of the relations of capital and production that
characterize the world today. More signiicant, these new kinds of relations
began the long, drawn-out process of building global connections among
human societies around the world. Over the later second millennium and
the irst millennium BCE, merchant enterprise of the First Commercial
Revolution eventually brought into being a series of long-distance links
extending from the Pillars of Hercules to the Indian Ocean, Southeast Asia,
and China.23
What has not been understood until recently is that a separate West
African commercial revolution was under way during almost precisely the
same era of time. It began with the towns and manufacturing villages of
Tichit during the middle second millennium BCE and grew into an exten-
sive network of trade routes and commercial towns and cities all across
the eastern and central Sudan over the course of the irst millennium BCE.
Sporadic contacts between the two commercial revolutions in the irst mil-
lennium BCE changed into regular trans-Saharan trade by the early irst
millennium CE, bringing West Africans in as full participants and contrib-
utors to the networks of contact and movement of things and ideas from
one end of the Eastern Hemisphere to the other.
In Tichit, located in modern-day southern Mauretania along a low
escarpment with reliable water sources, a skein of large villages and at
least one town lourished during the middle and second half of the second
millennium BCE. Diferent settlements seem to have specialized in a par-
ticular product for trade: one produced grindstones, another arrowheads,
still another beads, and so forth. At the middle of the skein lay one town
larger than all the rest. Its location and greater size mark it out as the prob-
able capital town of an early polity ruling over that skein of settlements.24
An additional region where the development of more complex polities
appears to have been under way during the mid-second millennium was
the Aïr Mountains of modern-day Niger. Archaeologist Augustin Holl
argues for an independent invention of copper metallurgy in Aïr in the

23
he characteristics and consequences of this “First Commercial Revolution” are described
in Ehret 1998, 16–20.
24
Augustin F. C. Holl, Holocene Saharans (London, New York: Continuum, 2004), pp. 122–
137, 164–184.
44 Ehret

period 2500–1500 BCE. Five large megalithic elite burial sites existed in the
mid-second millennium in diferent parts of the region, indicative of the
existence of ive chiefdoms or small kingdoms, each associated with copper
production and strongly pastoral in other aspects of economy.25 Possibly
these statelets formed provinces of an overarching larger polity.
Most intriguing, sub-Saharan Africans may have separately invented
ironworking. Iron smelting dates to the eleventh and tenth centuries
BCE in sites as far apart as Rwanda and Lake Chad, too early and too far
south to be reasonably explained as having difused from an origin just
ive hundred years earlier, three thousand kilometers away in Anatolia –
especially as ironworking did not reach the intervening lands, such as
Carthage, until ater the tenth century BCE, and Egypt and Meroe until
still later than that. Even earlier dates for African ironworking, extend-
ing well back into the second millennium BCE, come from new research
sites in the present-day Central African Republic, which lies between
two other early ironworking regions, the Lake Chad Basin and the Great
Lakes.26
As early as the eleventh century, the centers of lasting urban development
and commerce in West Africa shited south to the better-watered Sahel belt.
Central in the new developments was the Inland Delta of the Niger River
in modern-day Mali. Well before 1000 BCE, peoples of this region special-
ized in diferent kinds of food production for trade. Farmers among the
bayous of the Delta domesticated African rice (Oryza glaberrima) probably
as early as the fourth millennium, while other communities became ishing
specialists. Savanna farmers outside the Delta supplied sorghum and other
savanna crops, along with domestic animals, to the Delta communities. By
the early irst millennium BCE, the growth of manufacturing turned these
long-existing trade relations into an emerging commercial revolution, with
merchants, regular market centers, and long-distance transport of goods by
both boat and donkey.
Urban life in the western Sudan and Sahel evolved in a unique fashion.
he towns and cities developed as market centers for earlier village clusters,
in which each village had engaged in a diferent kind of production – cotton
textile weaving in one village, potting in another, and leather working in
still another. A fourth manufacturing specialization, ironworking, further
diversiied production during the irst millennium BCE, while the impor-
tation of copper from Aïr and from new mines in the far western Sahara

25
Holl 2004, 122–37, 164–84.
26
Holl (2009); Zangato and Holl (2010).
 Africa in World History before ca. 1440 45

greatly expanded the long-distance sector.27 Another valued metal, gold,

coming from upper Niger and Senegal River goldields, further enhanced
these trends by the late irst millennium.
In the north-central parts of modern-day Nigeria, the new directions of
economic change eventuated, between 900 and 400 BCE, in the emergence
of the earliest signiicant state as yet known from West Africa, associated
with the Nok culture. Around the central areas of this culture, excavators
have unearthed many huge terracotta sculptures, broken and buried in
the graves of high-ranking persons. From the distribution of its cultural
remains, the Nok state appears to have been as large as any kingdom of
more recent centuries in the region. Iron was a major product, and Nok was
probably a center for tin mining as well.28
It is oten assumed that the rise of the Garamantes in the Fezzan oases of the
north-central Sahara ater 900 BCE owed to their position as intermediaries
between the First Commercial Revolution of world history and the West
African commercial revolution.29 he evidence for a direct Garamantes role
is still sparse, but at least occasional contacts of some kind did exist between
the Carthaginian and Inland Niger Delta commercial spheres: the peoples
around the Delta acquired the Punic name for the horse, and thus presum-
ably the animal, sometime during the irst millennium BCE.
Regular trans-Saharan trading networks likely did not develop, however,
until the establishment of camels as beasts of burden and primary food ani-
mals among the inhabitants of the northern Sahara. Timothy Garrard has
proposed that the opening of the Roman mint at Carthage in the late third
century CE relects the arrival of West African gold from across the Sahara.
Most telling, he shows that the system of gold weight measures West African
merchants used until the nineteenth century preserved the particular sys-
tem put into efect at the Carthage mint.30

1.5 EARLY TOWNS AND STATES IN THE HORN OF AFRICA

Urban centers also emerged in the Horn of Africa in the irst millennium
BCE. Diferently from the endogenous rise of towns and cities in the west-
ern Sahel and Sudan, the founding of urban centers in the Horn owed in
large part to the First Commercial Revolution. South Arabians from the

27
McIntosh 1998.
28
A planned twelve-year archaeological investigation of Nok, led by Peter Breunig of the
Goethe Universität, Frankfurt, is currently under way.
29
E.g., Kea 2004.
30
Garrard 1982.
46 Ehret

opposite shore of the Red Sea, like the Phoenicians who founded Carthage
and the Greeks who founded Cyrene, came to Africa seeking new com-
modities and new sources for old commodities – initially frankincense and
myrrh, but subsequently tortoise shell and ivory – and their settlements
took the form, like Carthage and Cyrene, of city-states, planted amidst
the indigenous Cushitic pastoral and farming populations of the northern
Ethiopian Highlands.31
At irst, the routes tying the Horn of Africa to the First Commercial
Revolution of world history passed overland through South Arabia to the
Levant. Ater 300 BCE, the Red Sea itself became the central conduit of
trade between the Mediterranean and the expanding commercial networks
of the Indian Ocean. Sea routes passed from the Gulf of Aden across the
Arabian Sea to India and from India to Indonesia, and south down the East
African coast at least as far as modern-day Dar-es-Salaam in Tanzania. he
terminus of this latter route at the beginning of the irst millennium CE was
Rhapta, the earliest known East African town.
During the early irst millennium CE, the kings of one far northern
Ethiopian city-state, Aksum, brought all the towns and the countryside of
the northern Horn of Africa under their rule. By controlling, protecting,
and taxing commercial enterprise in the Horn and in the southern Red Sea,
the Aksumite kingdom grew into a major regional power, with its hege-
mony periodically extending to South Arabia as well. A notable outcome of
Aksum’s dominant position along the main route linking the Mediterranean
to the Indian Ocean was the spread of Christianity to Aksum, with King
Ezana adopting it as the oicial religion about thirty years ater Constantine
had taken the same step for Rome.
he Sassanian conquest of South Arabia in the 570s undermined Aksum’s
predominance in the Red Sea trade, and the rise of the irst Islamic empire
in the 640s to 750s completed Aksum’s isolation from the main lines of
commerce. In establishing Damascus as their capital, the Umayyad caliphs
shited the pivotal commercial sea link between east and west to the Persian
Gulf. For a century the Red Sea became a commercial backwater.
From a comparative world history perspective, the signiicant conse-
quence was that the Aksumite kings built a new material basis for their
state, feudal in character. he transformation of Aksum from the later sev-
enth to the ninth century strikingly parallels the course of change in con-
temporary, early medieval Western Europe. Urban life collapsed, with even
the city of Aksum shrinking to an episcopal center of perhaps a thousand

31
Ehret (1988); Schmidt and colleagues (2008).
 Africa in World History before ca. 1440 47

people. he kings created a horse-mounted military class by granting iefs

to their soldiers, with the peasants of each ief owing a portion of their prod-
uct to their lord. he titles of provincial oicials in late prefeudal Aksum
became the titles of the higher nobility. Monasteries became the principal
centers of education and literacy. Kings granted iefs of land to the monas-
teries to support their activities as religious centers, and the monks oten
acted as missionaries in spreading Christianity into the outlying areas of
the kingdom.
One notable diference distinguishes the feudalisms of Europe and the
Ethiopian highlands. Rights to land in prefeudal Aksum were vested in
the local peasantry, rather than in great landed magnates, as in the west-
ern Roman Empire. As a consequence, a ief in feudal Aksum – and in its
successor states, the Zagwe kingdom of the twelth and thirteenth centuries
and the Solomonic kingdom from 1270 onward – gave the lord a right to a
portion of the peasants’ production and certain other manorial privileges,
but let the local farmers not as serfs, but as free people, able to bequeath the
land they worked to their relatives and descendants.

1.6 AN ERA OF EMPIRES, AN ERA OF CITIES

AND COMMERCE
Even as the northern Horn of Africa was entering a long period of feudal
governance during the later irst millennium CE, in the savannas of west-
ern Africa an age of empires was beginning. Wagadu (Ghana), the earliest
known large empire, rose to prominence before mid-millennium. Stretching
from the Inland Delta of the Niger to Senegal, Wagadu lay athwart the key
trade routes linking the goldields far to the south to the merchant net-
works of the Sahara. In an age when cities as such did not exist north of the
Pyrenees, urban life lourished not just in Wagadu, but all across the west-
ern and central Sudan belt.
A series of empires succeeded Wagadu during the centuries ater its
decline in the twelth century: Susu from the mid-twelth to the early thir-
teenth century, with its power resting on control of the actual goldields32;
Mali from the 1240s to the mid-iteenth century, controlling access to both
the gold sources and the northern outlets of the trade; and Songay from the
mid-1400s to the late 1500s, commanding the major Sahel trading cities
and the salt trade of the Sahara. In the Chad Basin, the Kanem empire built
its wealth and power, from the ninth to the iteenth century, on a similar

32
Bühnen 1994.
48 Ehret

control over the access of neighboring states to the main trade routes of the
central Sahara.
he commercial interests of these empires gave them strong ties to the
Muslim world of those times. Islam had become established initially across
North Africa following the early Muslim conquests between 642 and 710
CE. During the next several centuries, it became the religion of the trans-
Saharan trading networks. In the Wagadu and Mali empires as well, it
became the religious allegiance of the merchants and the commercial cen-
ters. During the eleventh century, the rulers of the Takrur kingdom of the
Senegal Valley and the Kanem Empire of Lake Chad converted to Islam.
he rulers of the later Mali and Songay empires also professed Islam, but
the rural majority population in all those areas continued to follow their
older religions. Islam also spread with commercial relations along the East
African coast, becoming integral to urban identity in the Swahili city-states
by the twelth century CE. In the Horn of Africa, the spread of Islam, again
among merchants, but also among the Cushitic pastoralist populations
of the eastern Horn, provided religious backing for the military jihad of
Ahmad Gurey (1527–43) against the Christian Solomonic kingdom of the
Ethiopian highlands.
In these various fashions, Islam linked large areas of Africa to major
currents of world history between the seventh and iteenth centuries.
Timbuktu in the thirteenth century, for example, was not only a trade cen-
ter intimately connected to the Mediterranean and Middle Eastern worlds,
but a university town in the early sense of that term, with the university as
a place where noted scholars gathered to write and teach.
West of the lower Niger River in today’s Nigeria, a rich city-state-based
urban life developed during the second half of the irst millennium CE.
Ife, an early Yoruba city-state, grew into a major commercial entrepôt,
manufacturing glass beads and dealing in goods from the rainforest and
from the savannas to the north. Home to a splendid sculptural tradition of
brass casting, using the lost wax method, Ife became, as well, the leading
religious and ritual center of the Yoruba. An equally notable contemporary
of Ile-Ife was the Igbo city of Igbo-Ukwu, also an artistic center for brass
sculpture and the capital of a state, whose highly ritualized kings ruled the
lands across the lower Niger, east of the Yoruba.

1.7 FROM FARMING TO COMMERCE AND STATES:

THE SOUTHERN HALF OF AFRICA
In the southern half of Africa, where agriculture did not arrive until
ive thousand years ago, the irst appearance of towns and states, not
 Africa in World History before ca. 1440 49

surprisingly, lagged behind areas farther north. In the equatorial rainforests

of west-central Africa, long-distance commerce on the rivers of the Congo
Basin developed during the last millennium BCE out of an earlier trade
in ish, farm products, products of the hunt, and stone tools among the
Bantu societies, who had spread agriculture across the region from 3000–
1000 BCE, and the ancient foraging peoples of equatorial Africa, the Batwa
(“Pygmies”). By the mid-irst millennium BCE, the spread of iron across
the Congo Basin introduced a new manufacturing component to this trade.
Other industries, notably raia textile weaving and boat building, further
fueled trade expansion. he Batwa carved out their own niche in the new
economy by becoming specialist providers of honey, wax, skins, ivory, and
other forest products.33
In the Great Lakes region of East Africa, iron-using Mashariki Bantu set-
tlers set of a diferent chain of developments during the early irst millen-
nium BCE. Encountering Sudanic and Cushitic agripastoralist communities,
the Mashariki added sorghum and pearl millet from the Sudanic tradition
and inger millet from the Cushites to their previously yam-based farming.
he new crops, which required less rainfall than yams, allowed the Mashariki
peoples to scatter between 300 BCE and 300 CE across most of eastern and
southern Africa. Iron technology spread with them, and the demand for iron
helped stimulate new kinds of regional trade wherever they settled.34
hose Mashariki communities who arrived at the East African coast
toward the close of the irst millennium BCE soon came into contact with the
Indian Ocean developments of the First Commercial Revolution. he most
salient and lasting efect of this encounter came not from the merchants who
frequented East Africa’s earliest town, Rhapta, but from Indonesian immi-
grants, who followed the Indian Ocean trade routes to East Africa and set-
tled for a time at the coast before moving on to Madagascar around 300 CE.
hese ancestral Malagasy brought along several Southeast Asian crops, most
important bananas, well suited to the wetter African tropical environments.35
Banana cultivation, which spread rapidly west to the Great Lakes and into
the Congo Basin, was far less labor-intensive as well as more productive than
yam raising. Historians Jan Vansina and Kairn Klieman have argued that the
arrival of bananas in the Congo Basin fostered a major leap in commercial
activity in those areas, because reliance on the new crop freed up time for
people to engage in trade and in the production of trade commodities.36

33
Klieman (2003) explores this history in considerable detail.
34
Ehret 1998.
35
Ehret 1998, chapter 6; Ehret 2010; Gonzales 2009.
36
Klieman 2003; Vansina 1990.
50 Ehret

In the Congo Basin, political growth followed on the heels of these

developments, with chiefdoms emerging between 500 and 1100 and, ater
1100, kingdoms. he two earliest known kingdoms of the deep interior,
the Songye and Upemba states of the middle Lualaba River region, date
to roughly 1100–1400.37 In the lower Congo areas near the Atlantic coast,
Kongo with its large capital city, Mbanzakongo, and several smaller pro-
vincial capitals lourished from around 1300 until 1665. Both regions lay
along major routes of long-distance trade and close to prime copper- and
iron-producing areas.
In the African Great Lakes region, the earliest large states also date to
1100–1400. Supported by great wealth in cattle rather than trade, the rulers
built extensive earthworks in their capitals. Several thousand people lived
in these capitals, with their residential areas scattered over several square
kilometers, interspersed with ields and pasture. hese dispersed towns
undoubtedly attracted trade in salt, iron, and foodstufs, but their primary
function was as political and ritual centers.38
In contrast, Southern Africa’s irst town, Mapungubwe, which lourished
in the Limpopo Valley during the eleventh and twelth centuries CE, was
both a royal capital, with large stone structures, and the central entrepôt
connecting the source areas of ivory and gold in the interior to the sea
routes of the Indian Ocean.39 he establishment of the Zimbabwe Empire
in the thirteenth century shited the heartland of urban development north
to modern-day Zimbabwe. he capital city of Great Zimbabwe, famous for
its great stone buildings, had iteen thousand to eighteen thousand inhabit-
ants during the fourteenth century. Several provincial capitals in the empire,
though smaller, appear also to have deserved the appellation of town.

1.8 AFRICA IN A GLOBAL HISTORICAL PERSPECTIVE

To view Africa over the very long term is to discover that the notable devel-
opments of Africa’s past followed similar pathways and proceeded at similar
paces as comparable changes elsewhere in the world. hese developments
airm something historians more widely in the world, including unfortu-
nately only too many historians of the African continent itself, have not
yet assimilated – namely, the comparability in the timing and content of

37
Vansina (1990) shows that the Songye, although an oligarchic republic in recent centuries,
most probably evolved out of an earlier monarchy. See also Ehret 2002, chapter 6.
38
Sutton 1998.
39
homas N. Hufman, Mapungubwe (Johannesburg: Witerwatersrand Press, 2005).
 Africa in World History before ca. 1440 51

historical change in Africa to the timing and content of historical change

elsewhere. Two great transitions of human history during the Holocene –
from foraging to farming and, several thousand years later, from villages
and informal governance to towns and states – not only were not late in
emerging in Africa, but Africa was a continent of primary invention in
those times.
Cultivation of crops and herding of animals began in Africa as early as
anywhere else except the Middle East, and only slightly later than there. he
irst domestication of cattle in world history took place in the southern half
of the eastern Sahara one thousand to iteen hundred years earlier than
the separate domestication of cattle around the eastern Mediterranean.
West Africans living south of the Sahara were among the irst peoples in
the world to invent ceramic technology, before eleven thousand ive hun-
dred years ago. A probable second African invention of ceramics took place
almost as early in the eastern Sahara. he independent inventions, by difer-
ent African populations in diferent regions south of the Sahara, of cotton
textile weaving seven thousand or more years ago, of raia-cloth weaving
and polished stone tools almost equally early, and probably of copper met-
allurgy in the Sahara by 2000 BCE and ironworking in north-central Africa
before 1000 BCE, reairm something that historians have long under-
stood – that particular advances in human technological capacities oten
arise more than once and in disparate parts of the world.
Within the overall progression of human history since the end of the
latest ice age, the lag time between the earliest agriculture and the earli-
est towns and states in Africa accords with wider world history patterns.
Typically, whether in the Middle East, Middle America, China, or Africa,
the earliest urban centers and states came into being around four thou-
sand to ive thousand years ater the irst deliberate plant or animal tend-
ing. In the eastern Sahara, the initial stage of the First Great Transition,
from foraging to cultivation and herding, began around the mid-ninth
millennium BCE. he Second Great Transition in those areas, from local-
ized political relations to states and from villages to towns, began as early
as anywhere in the world – during the ith millennium at Nabta Playa
and during the fourth millennium along the Nile itself, at such places as
Shaheinab and Qustul in Nubia and Naqada in southern Upper Egypt. In
West Africa, the earliest towns and larger polities date to the second mil-
lennium. In those regions, the span between irst farming and irst towns
may have been as much as six thousand or as little as four thousand years,
depending on how early the shit from gathering to cultivating wild grains
came about.
52 Ehret

African history cautions, as well, against the mistake of attributing all

invention to the early regions of cities and states – to the cultural complexes
historians have traditionally called “civilizations.” Ironworking, for instance,
rapidly established itself as a key productive sector in the commercial towns
of West Africa during the irst millennium BCE. But the men who inno-
vated this technology lived elsewhere, in regions where village-scale res-
idence patterns long prevailed; and ironworking spread equally rapidly
across the nonurbanized, nonstate parts of the continent. Civilization is not
a thing; it is an evaluation, and only too oten an evaluation that nonhisto-
rians turn into a judgment on human worth. A historian’s job is to seek to
give historical ordering to and make historical sense of the changes socie-
ties and people have undergone, to the extent possible from the evidence
available. When historians fall into the trap of deining some cultures as
civilizations and others as not, they make a value judgment even if they
mean not to. More to the point, they inevitably narrow the scope of their
inquiry and, as a result, very oten fail to recognize the more encompass-
ing history that makes the culture they single out in this fashion worthy of
attention in the irst place.
he gloriication of ancient Egypt is a prime illustration of the need to
discard the fallacious Western idea that there really is something called
“civilization.” Putting Egypt back into Africa, where it has always properly
belonged, is fundamentally important, but it is a bootless exercise if we
perpetuate the equally fallacious Western attribution of everything nota-
ble in Africa to the ancient Egyptians. As important as Egypt was in the
history of early state formation, it was not an initiating region in the cru-
cial transitions of Holocene human history to food production, but rather
a crossroads in the subsequent difusion of crops, animals, and technol-
ogy. he crucial early innovative areas in the continent lay in the Sudan
belt and in West Africa. Even within the gyre of political history, Egypt’s
irst large state, the Old Kingdom, took shape not at the center, but at one
side of a wider nexus of early complexity that included Saharan pastoral-
ists, Nubian town dwellers, and the Qustul state and probably other, as yet
undiscovered small states farther south in Nubia. he major technological
turning points of early African history, from the invention of ceramics to
the invention of ironworking, did not originate in Egypt, but spread there
from other areas.
he events of the hird Great Transition of the last millennium BCE,
the First Commercial Revolution, and the developments of the irst if-
teen hundred years CE deepened and extended the ways Africans from
many parts of the continent participated in the wider compass of global
 Africa in World History before ca. 1440 53

history. he Sudan belt and northern Africa had numerous urban centers
already during the irst millennium BCE, when the western and northern
Europeans of the same period had none at all. During much of the irst
thousand years CE, the northern three-iths of Africa, as well as the eastern
coast and its hinterlands, were intertwined economically with other world
regions to an extent that Europe north of the Pyrenees and Alps began to
match only between 1000 and 1400 CE. he irst millennium and a half CE,
it can be argued, were times of economic development and advance overall
for Africa, in which large parts of the continent contributed to the far-lung
currents of historical change across the Eastern Hemisphere.
hrough all the eras before the mid-iteenth century, Africa did not fol-
low behind or lie outside the main trends and pathways of human history.
he great question that we as historians of Africa must grapple with is how
and why the developments of the past one thousand years, and more par-
ticularly the past ive hundred years, in the end so greatly redirected history
across large parts of the continent. A fuller understanding of African his-
tory over the very long term casts in sharpest relief the salience and com-
plexity of that problem for historians.

References
Arkell, Anthony John. (1949). Early Khartoum: An Account of the Excavation of an Early
Occupation Site Carried Out by the Sudan Government Antiquities Service in 1944–
5. London, New York: Oxford University Press.
Bühnen, Stephan. (1994). “In Quest of Susu.” History in Africa, 21: 1–47.
Ehret, Christopher. (1988). “Social Transformation in the Early History of the Horn of
Africa.” In Proceedings of the Eighth International Conference of Ethiopian Studies,
edited by Taddese Bayene. Addis Ababa: Institute of Ethiopian Studies 1: 639–51.
(1998). An African Classical Age. Charlottesville: University of Virginia Press.
(2001). “he African Sources of Egyptian Culture and Language.” In África Antigua:
El Antiguo egipto, una civilizatión Africana, edited by Josep Cervelló, 121–8.
Barcelona: Centre D’estudis Africans.
(2002). he Civilizations of Africa. Charlottesville: University of Virginia Press.
(2006). “Linguistic Stratigraphies and Holocene History in Northeastern Africa.” In
Archaeology of Early Northeastern Africa, edited by Marek Chlodnicki and Karla
Kroeper, 1019–55. Posnan: Posnan Archaeological Museum.
(2008). “Reconstructing Ancient Kinship in Africa.” In Early Human Kinship: From
Sex to Social Reproduction, edited by Nicholas J. Allen, Hilary Callan, Robin
Dunbar, and Wendy James. Oxford: Blackwell.
(2010). “Linguistic Testimony and Migration Histories.” In Migration in World History,
edited by J. Lucassen, L. Lucassen, and P. Manning. Leiden, Boston, MA: Brill.
(2011). “A Linguistic History of Cultivation and Herding in Northeastern Africa.” In
Windows on the African Past: Current Approaches to African Archaeobotany, Vol. 3,
54 Ehret

edited by Ahmed G. Fahmy, S. Kahlheber, and A. C. D’Andrea. Frankfurt am Main:

Africa Magna Verlag.
(2013). “Holocene Migrations as Determined from Linguistics in Sub-Saharan
Africa.” In he Encyclopedia of Global Human Migration, Vol. 1, edited by I Ness
and P. Bellwood, Chapter 13. Malden, MA, and Oxford: Wiley-Blackwell.
Ehret, Christopher, S. O. Y. Keita, and Paul Newman. (2004). “he Origins of Afroasiatic.”
Science, 306: 1680–1.
Fuller, Dorian. (2006). “Agricultural Origins and Frontiers in South Asia: A Working
Synthesis.” Journal of World Prehistory, 20(1): 1–86.
Garrard, Timothy F. (1982). “Myth and Metrology: he Early Trans-Saharan Gold
Trade.” Journal of African History, 23(4): 443–61.
Gonzales, Rhonda M. (2009). Societies, Religion, and History: Central-East Tanzanians
and the World hey Created, c. 200 BCE to 1800 CE. New York: Columbia University
Press.
Gordon, William. (2009). “Cultural Identity of the 25th Dynasty Rulers of Ancient
Egypt in Context: Formulation, Negotiation and Expression.” PhD dissertation,
University of California at Los Angeles.
Holl, Augustin F. C. (2004). Holocene Saharans. London, New York: Continuum.
(2009). “Early West African Metallurgies: New Data and Old Orthodoxy.” Journal of
World Prehistory, 22: 415–38.
Huysecom, E., M. Rasse, L. Lespez, K. Neumann, A. Fahmy, A. Ballouche, S. Ozainne, M.
Maggetti, Ch. Tribolo, and S. Soriano. (2009). “he Emergence of Pottery in Africa
during the Tenth Millennium cal BC: New Evidence from Ounjougou (Mali).”
Kea, Ray. (2004). “Expansions and Contractions: World-Historical Change and the
Western Sudan World-System (1200/1000 B.C.–1200/1250 A.D.).” Journal of
World-Systems Research, 10: 723–816.
Klieman, Kairn A. (2003). he Pygmies Were Our Compass: Bantu and Batwa in the
History of West Central Africa, Early Times to c. 1900 C.E. Portsmouth, NH:
Heinemann.
Kobusiewicz, Michal and Romuald Schild. (2005). “Prehistoric Herdsmen.” Academia,
3(7): 20–4.
Malville, J. M., R. Schild, F. Wendorf, and R. Brenner. (2008). “Astronomy of Nabta Playa.”
In African Cultural Astronomy: Current Archaeoastronomy and Ethnoastronomy
Research in Africa, edited by Jarita C. Holbrook, Rodney T. Medupe, and Johnson
O. Urama. Dordrecht, New York: Springer.
McIntosh, Roderick J. (1998). he Peoples of the Middle Niger. Malden, MA: Blackwell
Publishers.
Schmidt, Peter R., Matthew C. Curtis, and Zelalem Teka, eds. (2008). he Archaeology of
Ancient Eritrea. Trenton, NJ: Red Sea Press.
Shaw, T. (1980). “Holocene Adaptations in West Africa: he Late Stone Age.” Early Man
News, 3/4: 51–82.
Sutton, J. E. G. (1974). “he Aquatic Civilization of Middle Africa.” Journal of African
History, 15(4): 527–46.
(1998). “Ntusi and Bigo: Farmers, Cattle-Herders and Rulers in Western Uganda, AD
1000–1500.” Azania, 33: 39–72.
Vansina, Jan. (1990). Paths in the Rainforests. Madison: University of Wisconsin Press.
 Africa in World History before ca. 1440 55

Wendorf, Fred and Romuald Schild. (1998). “Nabta Playa and Its Role in the Northeastern
African History.” Anthropological Archaeology, 20: 97–123.
Williams, Bruce B. and Keith C. Seele. (1986). he A-Group Royal Cemetery and Qustul.
Chicago: University of Chicago Oriental Institute.
Zangato, Étienne and Augustin F. C. Holl. (2010). “On the Iron Front: New Evidence
from North-Central Africa.” Journal of African Archaeology, 8(1): 7–23.
 2

Reversal of Fortune and Socioeconomic

Development in the Atlantic World

A Comparative Examination of West Africa

and the Americas, 1400–1850

Joseph E. Inikori

he comparatively poor performance of the economies of sub-Saharan

Africa in the current global economy has been a challenge to scholars and
policy makers alike. Economic historians, economists, and political sci-
entists have struggled to understand the puzzle of sub-Saharan African
economic development with little success. Without helpful guidance from
economic historians and economists, policy makers have oten applied
ill-informed policy prescriptions with disastrous outcomes. For a while,
economic historians seem to have abandoned sub-Saharan Africa out of
frustration (Hopkins 2009). Happily, there appears to be a turnaround; eco-
nomic historians are returning to sub-Saharan Africa as the growing popu-
larity of globalization has helped to focus attention, once again, on long-run
historical processes as a source of knowledge for the understanding of con-
temporary situations (Inikori 2007a). We have begun to see imaginatively
articulated articles seeking to explain long-run development processes in
sub-Saharan Africa.1
his welcome development, however, appears unconnected, or at least not
seriously connected, to a major historiographical development with much
potential for a clearer understanding of the long-run development process
in western Africa. As part of the growing globalization discourse, Atlantic
world history is now one of the fastest growing ields of historical scholar-
ship (Bailyn 2005; Canizares-Esguerra and Seeman 2007; Coclanis 2005;
Eltis, Lewis, and Sokolof 2004; Inikori 1979, 2002a; Mancke and Shammas

1
Acemoglu, Johnson, and Robinson (2002, 2005a); Austin (2008a, 2008b); Inikori (2002b,
79; 2009a); Nunn (2007, 2008); Nunn and Wantchekon (2011).

56
 Reversal of Fortune and Socioeconomic Development 57

2005; McCusker and Morgan 2000; O’Rourke and Williamson 1999; Solow
1991). Studying the Atlantic basin as a unit of historical analysis enables
us to overcome the myopic view of long-run development processes in the
major regions of the Atlantic basin as self-contained and isolated phenom-
ena (Davis 1973, xi). he relational framework of Atlantic world history
gives us a clear view of the interconnections between the long-run develop-
ment processes in the individual subregions of the Atlantic basin. Recent
research seeking to explain long-run economic development processes
in western Africa has yet to fully incorporate this relational framework
of Atlantic world history. Like mainstream European economic historio-
graphy (Inikori 2009b), the highly imaginative recent works generally lose
sight of the fact that long-run economic development processes in western
Africa during the formative centuries, 1450–1850, were essentially part of
an integrated historical process in the Atlantic basin. If we are not fully con-
scious of this fact, we are unlikely to accurately identify the causal factors
in the process, properly understand how they operated and interacted with
other factors, and assign to them their correct causal weights.
his chapter attempts to show what we can learn when the economic his-
tory of western Africa during the critical period, 1450–1850, is studied with
the perspective of Atlantic world history. It examines West Africa’s long-run
economic development process, 1400–1850, in the light of the clear view
ofered by the Atlantic world history framework. A discussion of the com-
parative levels of economic development in West Africa and the Americas
in two “long years” – ca. 1400 and ca. 1850 – with emphasis on markets and
the development of the market economy, is presented in the irst part of the
chapter.2 his part of the chapter aims to establish a clear fact of reversal
of fortune among the major regions examined. Emphasis on markets and
the market economy arises from the fact that eicient market allocation
of resources is critical to sustained growth of per capita income and the
further development of resources. he degree of market development is,
therefore, a good barometer for measuring the level of economic develop-
ment at a given moment. he second part of the chapter presents currently
available evidence to demonstrate the extent to which the reversal can be
explained in terms of the transatlantic slave trade and the employment of
enslaved Africans in the Americas to produce commodities for Atlantic

2
he choice of West Africa as the appropriate region in Africa for this comparative study
arises from the fact that it was the region irmly linked to the Americas in a single inter-
connected long-run development process that produced, over the period 1450–1850,
an integrated Atlantic economy that was the nucleus of the current global economy. See
Inikori (2007a).
58 Inikori

commerce from 1500 to 1850. Because the end result of the developments
in the Atlantic world during this formative period constituted the nucleus
of the current global economy (Inikori 2007a), this chapter’s indings have
a signiicant bearing on our understanding of the historical origins of the
relative position of West African economies in the current global economic
order.

2.1 COMPARATIVE LEVELS OF ECONOMIC DEVELOPMENT

IN WEST AFRICA AND THE AMERICAS
Global comparison of regional development levels in the distant past, when
data for measuring Gross Domestic Product (GDP) are nonexistent, relies
heavily on demography – its absolute size, density, and urbanization – as a
proxy (Acemoglu et al. 2005b; Allen 2009, 16–21). For West Africa in the
“long 1400,” a combination of Arabic sources, archaeological evidence, and
oral sources forms the basis for researchers’ demographic estimates. he
more heavily documented studies focus largely on micro-regions, particu-
larly the Niger Bend – the inland delta of the Niger River from the modern
city of Segu in the west to where the river enters modern Nigeria. Specialists
generally agree that the Niger Bend subregion of West Africa was heavily
populated, with very large cities, from the early centuries of the irst mil-
lennium CE to the irst half of the second millennium CE (Cissoko 1984;
Hunwick 1999; McIntosh 1998, 2005, 2008, 31–43; Moraes Farias 2003;
Niane 1984a, 1984b, 1984c). Niani, Jenne, Timbuktu, and Gao are among
the best known cities in the region. Niane puts the population of Niani
during the fourteenth century at one hundred thousand; based on a late
sixteenth-century census, Cissoko estimates the population of Gao at one
hundred thousand, that of Timbuktu at eighty thousand, and that of Jenne
at thirty thousand to forty thousand; Hunwick’s estimate for Gao, based
on the same data as Cissoko’s, ranges between thirty-eight thousand and
seventy-six thousand (Cissoko 1984, 206; Hunwick 1999, xlix; Niane 1984a,
156). he Niger Bend had many other towns during the late fourteenth- to
the late sixteenth-century period. In fact, Mahmud Kati, an Arab visitor to
the region during the period, reported that there were four hundred towns
in the Mali Empire (Niane 1984a, 156). he research of Roderick McIntosh
and that of P. F. de Moraes Farias, among others, strongly support earlier
arguments that the Niger Bend was a highly urbanized region in West
Africa, with relatively large overall populations, from the last centuries
of the irst millennium CE to the late sixteenth century (McIntosh 1998,
 Reversal of Fortune and Socioeconomic Development 59

2005, 2008; Moraes Farias 2003). To the east of the Niger Bend, in what
is generally known as West Africa’s Central Sudan, the Hausa city-states,
such as Katsina and Kano, and the Kanem-Borno Empire in the Lake Chad
region also had many large towns during the period. So, too, did the West
African forest belt, with large urban centers such as Ile-Ife and Benin City.
Among modern researchers competent in combining Arabic sources,
archaeological evidence, and oral sources, only D. T. Niane has ventured
to estimate the total population of the African continent between the four-
teenth century and the sixteenth century. Because his estimate has been
generally overlooked in recent discussions of the continent’s population, it
is pertinent to present the basis of his informed guesstimate to give readers
a proper context for a fair comparative assessment. Ater detailing the gen-
eral sociopolitical and economic conditions in Africa between the twelth
and sixteenth centuries to support his contention that “Africa was not an
under-populated continent” during the period, he ofers regional speciics
to reach his overall numerical statement:
We have some indications of the population of certain regions. According to
Mahmud Kati, Mali had 400 towns or large population centres; the farming villages
stretched in a continuous line along the waterways. Agricultural production was
very extensive. Professor Cissoko . . . has drawn attention to the extent of rice pro-
duction; for example, among the Songhay in the iteenth and sixteenth centuries, a
single fanfa or overseer, directing the work of tributaries, could supply the king with
more than 1000 sounu (large leather sacks which held about 70 kg). he royal stocks
of food were immense; we may gain some idea of them from the fact that the king
of Gao had a regular army of 100,000 men, garrisons near the main commercial
towns and a very large court, and that all these people were fed and maintained by
the king almost entirely from the agricultural revenues. It is diicult to estimate the
size of the population, but the large number of well-populated trading towns and
constructions on the scale of the Great Zimbabwe monuments would suggest that
the population was dense. Over the whole continent, at this period of commercial
expansion, the towns perhaps accounted for 10 percent of the population. . . . At this
time Africa must have sufered epidemics, periods of drought and major loods, but
the documents available make little mention of famine. Arab travelers – for exam-
ple, Ibn Battuta, the fourteenth-century globe-trotter – noticed the abundance of
food on the east coast and in the Sudan. For the whole of the continent, the popula-
tion in Ibn Battuta’s lifetime [the fourteenth century] may be estimated at 200 mil-
lion. (Niane 1984c, 683–4)

Based on the writings of contemporary Timbuktu historians and informa-

tion from Portuguese explorers who made contact with the emperor of Mali
in the mid-iteenth century, Niane puts the empire’s population in the mid-
iteenth century at 40–50 million (Niane 1984a, 156).
60 Inikori

Like all demographic estimates for these early periods across the globe,
we can all agree, Niane’s estimates must have a large margin of error. In
some important sense, it may be best to treat them as orders of magnitude
instead of exact igures. But, comparatively speaking, it is diicult to under-
stand why scholars in recent times appear to prefer estimates, including
those by Angus Maddison, that are not better documented than Niane’s
(Benjamin 2009; Ilife 1995; Maddison 2001, 2003; Manning 2006). Based
on Carr-Saunders’s 1934 work, Patrick Manning employs in his analysis a
total population of 100 million for the African continent in 1500, one-half
of Niane’s igure, and 50–60 million for West and Central Africa (Manning
2006, 53). From what we know about West Africa and Central Africa, it
seems reasonable to suppose that at least two-thirds, that is, 40 million, of
Manning’s upper igure can be assigned to West Africa. With an area of
2.4 million square miles, this gives a density of 16.7 persons per square mile
for West Africa in 1500. Density in the Niger Bend could not have been less
than 25 per square mile during the iteenth and sixteenth centuries. Again,
readers should treat these igures as orders of magnitude and base their rea-
soned judgment on evidence concerning sociopolitical organization, agri-
cultural production, and land use.
Coming to the demography of the Americas in the “long 1400,” the sub-
ject has been a matter of dispute since the irst reports by explorers dur-
ing the sixteenth century (Diie 1945; excerpts in Hanke 1967, 193). More
recent discussions center on the works of what has become known as the
Berkeley school (Borah and Cook 1960, 1963; Cook and Simpson 1948).
he Berkeley school has been severely criticized. Francis Brooks’s criticism
probably represents the most severe:
Embedded in the received story of the “conquest of Mexico” are two questionable
assertions. he irst is that the population of Central Mexico in 1519 was approx-
imately 25 million people. he second is that these people were struck in 1520 by
smallpox and, possibly, by other diseases; and that as a result perhaps one-third
of them died. . . . Every element of this received account is false, epidemiologically
improbable, historiographically suspect, or logically dubious. he population of
Mexico in 1519 was almost certainly nowhere near as large as 25 million. (Brooks
1993, 1–2)3

he controversy continues, but igures derived from the synthesis by William

Denevan, 53,904,000 for all the Americas in 1492, are probably the best

3
Brooks adds that “the conclusions of the physical anthropologists, derived from the carry-
ing capacity of the available agricultural technology,” suggest a total population of 5 mil-
lion to 10 million for Mexico in 1500 (Brooks 1993, 4–5). he literature cited includes
Newson (1985) and Denevan (1976).
 Reversal of Fortune and Socioeconomic Development 61

currently available (Denevan 1992, xxix; Newson 2006, 148, table 5.1).4 he
bulk of the population was concentrated in two regions: Mexico, 17,174,000
(31.9 percent); the Andes (including the Andean highlands and coastal
Venezuela, Colombia, Ecuador, Peru, and Bolivia), 15,696,000 (29.1 per-
cent). Two other regions had relatively large populations in proportion to
their geographical area: Central America, 5,625,000 (10.4 percent); and the
Caribbean, 3,000,000 (5.6 percent). he huge geographical area of North
America (what later became the United States of America and Canada),
7 million square miles, had only 3,790,000 (7 percent); Paraguay, Uruguay,
and southern Brazil had 1,055,000 (2 percent).
In terms of density, extremely low population densities characterized
the Americas in 1492. he regional distribution of Denevan’s total igure
of 53,904,000 implies an average density of approximately 6.3 persons
per square mile for Latin America and the Caribbean, and 0.5 for North
America. he large concentration in Mexico and the Andes suggests that
density in most areas of Latin America and the Caribbean was less than
one person per square mile in 1492. As would be expected, the few urban
centers in the Americas were located in Mexico and the Andes (Prados de
la Escosura 2006, 476–7, table 13.3).
he uncertainty concerning the population estimates for West Africa
and the Americas between the late fourteenth and late iteenth century
(the “long 1400”) notwithstanding, it is clear enough from the evidence
that West Africa was by far more densely populated than the Americas. As
we now propose to show, market exchange and the market economy also
developed to a much greater extent in West Africa than in the Americas
during the irst half of the second millennium CE.
he subregions of West Africa had their own internal factors (in partic-
ular, population growth and natural resource endowments) that stimulated
the growth of trade and the evolution of the market economy in the regions.
But the growth of trade in the economies of the Niger Bend operated as a
powerful growth pole for all the subregions. For more than one thousand
years, from the mid-irst millennium CE to the mid-second millennium CE,
the Niger Bend enjoyed considerable geographical location advantages –
the agricultural resources provided by the lood plains of the inland Niger
delta; the long stretch of river transportation along the River Niger; the
ease of movement in the open savanna (including drat animals for trans-
portation); the trade links to the copper- and salt-producing economies of
4
While some scholars do not accept the Denevan synthesis igures, they are the most widely
and frequently cited igures in the current literature, both in scholarly works and in major
textbooks.
62 Inikori

the southern Sahara and to the commercial centers of the Mediterranean

world – all this made the region the center of trade and manufacturing in
West Africa until the middle centuries of the second millennium.
An important internal factor in the region was the growth of popula-
tion and the emergence of large urban centers. he demographic evidence
discussed earlier suggests strongly that the bulk of West Africa’s popu-
lation at this time was located in the region. Urbanization in the region
arose from the combined efects of population growth and the expansion
of commerce. As McIntosh has shown, the growth of urban populations
in the Niger Bend was not associated initially with administration in cen-
tralized states or religious worship as was the case with early urbanism in
other parts of the world. Urbanism emerged spontaneously in the Niger
Bend and the urban centers were self-organizing before they became part of
centralized, bureaucratic, and military states (McIntosh 1998, 2005, 2008,
32). his must mean that urbanism in the region was largely the product of
demography and commerce, and the urban centers functioned principally
as commercial centers. he subsequent rise of centralized state systems in
the region simply provided further impetus for the growth and geographi-
cal extension of the commercial business of the cities’ merchants.
hus, the Niger Bend became the center of a merchant diaspora that
connected all the subregions of West Africa in an integrated network of
long-distance interregional trade (Inikori 2010). Niger Bend merchants dis-
tributed the region’s manufactures and other products to the subregional
trading centers in West Africa in exchange for the commodities of the subre-
gions. In addition, the Niger Bend traders conducted a large entrepôt trade,
re-exporting southern Saharan products (largely salt and copper products)
and products of the Mediterranean trading world (especially cowries from
Asia and European textiles) to the forest and savanna subregions of West
Africa. Gold and kola nuts were the most important products imported
from the subregions into the Niger Bend. Much of the gold was re-exported
to the Mediterranean trading world to feed the commercializing process in
the economies of the Mediterranean and Europe. he rising Middle Eastern
and European demand for gold beginning during the fourteenth century
expanded the gold trade greatly, West Africa and its famed “Island of Gold”
being the source of about two-thirds of global gold supply at this time
(Hunwick 1999, liii; McIntosh 1998, 269, 273–4). Niger Bend merchant dias-
pora colonies spread out across West Africa, linking the subregional trading
centers to each other and to the Niger Bend commercial cities (in particular,
Jenne, Timbuktu, and Gao) in an embryonic West African regional global-
ization process centered in the Niger Bend (Inikori 2010).
 Reversal of Fortune and Socioeconomic Development 63

As a result of this embryonic regionally based globalization, a gradual

process of interregional specialization and division of labor was in progress
in West Africa by the sixteenth century. A major feature of the process was
the general tendency of the coastal regions (from the Upper Guinea coast
to southeastern Nigeria) and their immediate hinterlands to export largely
primary products (gold, kola nuts, aquatic products, etc.) to the interior
savanna in exchange for mainly manufactures from the latter (Inikori 2002b,
2009a). he combined operation of local, intraregional, and long-distance
interregional trade stimulated sustained commercial expansion and wide-
spread evolution of the market economy all over West Africa during the
iteenth and sixteenth centuries.
he extent of market development and the geographical spread of the
market economy in West Africa during the iteenth and sixteenth centuries
is indicated by the scale and widespread use of well-established exchange
currencies in the region. Both in magnitude and geographical spread,
the cowrie currency was probably the most important (Hogendorn 2006;
Hogendorn and Johnson 1986; Johnson 1970). he cowrie shells were gath-
ered from the Indian Ocean Maldive Islands. Early imports into West Africa
came through the Middle East. As would be expected, evidence shows the
upper and middle Niger (Niger Bend) was where they were irst used as
early as the eleventh century CE, and remained in regular use in the com-
mercial cities of the region (Jenne, Timbuktu, Gao) thereater. hey spread
to ancient Ghana by the thirteenth century; and through the gold and kola
nut trade, the cowrie currency reached the Gold Coast (southern modern
Ghana), from where it expanded to Whydah, Ardra, and southwest Nigeria
several decades before the arrival of Portuguese explorers in those regions
(Johnson 1970, 32, 35–6, 332–3). he other major currency in use in West
Africa before the arrival of European explorers and traders was copper cur-
rency, imported from the southern Sahara in the form of small copper rods
weighing about one pound each. he main area of the copper currency was
southeastern Nigeria, where a local metal currency in the form of small iron
hoes (called anyu) also developed (Inikori 2002b, 64). he quantity and rel-
ative value of currency in the imports European traders brought during the
early decades of their trade in West Africa provide unambiguous evidence
of the extent and continuing expansion of the market economy in West
Africa (Inikori 2007b). hus, taking all the evidence together, it is clear that
the development of markets and the evolution of the market economy in
West Africa that had been on since medieval times had advanced greatly by
the late iteenth century. he predominance of products derived from West
African labor and natural resources in the irst one hundred ity years of
64 Inikori

European trade in the region further intensiied this ongoing development

and geographical spread of the market economy.
he story in the Americas is radically diferent. here was market exchange
of goods in all regions of the Americas during the iteenth century. But in
all the sparsely populated regions, with one person per square mile or less,
subsistence production (production for direct consumption by producers
without market exchange) was overwhelmingly dominant. his was the
case for all of North America and the larger part of South America where
modern Brazil, Uruguay, Paraguay, and Argentina are located (Bakewell
and Holler 2010; Carmack, Gasco, and Gossen 1996; Hanke 1967; Katz
1972; Lockhart and Schwartz 1983; Trigger and Washburn 1996). As Storey
and Widmer (2006, 77) put it:

In pre-Columbus Latin America, there were far more autonomous societies and
cultures than in Europe at the time. Most of these societies had economies char-
acterized by a “domestic mode of production.” Economies on this scale were self-
suicient. Each family constructed, produced, and gathered all that it needed to
survive, and each family was economically redundant with the next – it engaged in
the same economic activities:

In the more densely populated regions with complex sociopolitical orga-

nizations – the Inca and Aztec Empires – market development was lim-
ited partly by physical geography and partly by political institutions. In the
expansive Inca Empire, the high mountains running north-south impeded
east-west movement; in the Aztec Empire, physical geography also made
movement between the plains and the highlands, which separated the west-
ern from the eastern plains, extremely diicult.
Possibly in adaptation to the physical environment, the quasi-socialist orga-
nization of the Inca Empire (Baudin 1967; Katz 1972, 282–94; Metraux 1967)
imposed the greatest limit on the development of markets and the market
economy in the polity. he state parceled out land among the families in the
communes to meet their subsistence needs. he members of the communes
paid labor dues employed on state lands to produce for the state and the
priestly nobility, as well as surpluses kept for emergency purposes. he state
also employed the labor dues in its massive construction works – the exten-
sive state roads, bridges, irrigations, temples, and so forth. Cratsmen worked
only for the state and religious institutions. Members of the communes were
allowed to exchange goods in their local communities. But only the state could
distribute products from one province to another, from one ecological region
to another. he famous Inca roads, running north-south for a total of thirty
thousand to ity thousand kilometers, were not built for commerce; only those
 Reversal of Fortune and Socioeconomic Development 65

traveling on state business were allowed to use them (Bakewell and Holler
2010, 58; Katz 1972, 287). As Friedrich Kantz wrote of the Inca state: “there
is no other state in pre-Spanish America, in the ancient East or in European
antiquity in which trade was so uniquely controlled by the state that not a
single general article of barter served as currency” (Katz 1972, 293).
here was much greater private trading organized by a merchant class in
the Aztec Empire. Commerce, therefore, developed to a higher level than
in the Inca Empire. Even so, the development and geographical spread of
the market economy was very limited. he vast majority of the population
had little or no connection with the market. Somewhat similar to the Inca
economy, production was organized around the needs of the state and its
supporting nobility in a tributary system. Trading itself was largely depen-
dent on the tribute system. At the time of the Spanish conquest, 371 towns
spread across the thirty-eight provinces of the empire paid tribute (Storey
and Widmer 2006, 100; Von Hagen 1961, 170). A large bureaucracy of trib-
ute collectors connected the provinces to the capital city of Tenochtitlan
(Katz 1972, 198–200). he amount of tribute goods brought to the capital
yearly was immense. For example, the province of Cuauhnahuac sent annu-
ally twelve thousand eight hundred cloaks, sixteen hundred loin cloths,
sixteen hundred women’s tunics, eight warriors’ costumes, thirty-two
thousand bundles of paper, eight thousand bowls, and four bins of maize
and beans. he province of Toctepec on the Gulf Coast sent, among other
things, ninety-six hundred decorated cloaks, sixteen hundred women’s
tunics, sixteen thousand rubber balls, eighty handfuls of quetzal feathers,
twenty-four thousand little bunches of feathers, one hundred pots of liq-
uid amber, and three hundred loads of cacao (Storey and Widmer 2006,
100–1). he state redistributed these tribute goods among the nobility and
state functionaries. Tribute recipients sold to the merchants what they did
not need and bought from them what they needed. he merchants thus
provided a reshuling service that enabled recipients to secure the basket
of goods they needed. Crat production was also linked largely to the trib-
ute system. By law, the most important manufactures the cratsmen pro-
duced – feather ornaments, gold and silver, cotton clothing, and precious
stones – could only be used by the tribute-receiving nobility; and they alone
were allowed to live in two-story houses and palaces, the building of which
needed the services of experienced cratsmen. Merchants and cratsmen
were thus dependent largely on the state: “he majority of cratsmen by the
end of the iteenth century were either employed in the palaces and worked
for the state or were maintained on the tributes exacted from various parts
of the [capital] city” (Katz 1972, 218).
66 Inikori

It is clear from the operation of the tribute system that the vast majority of
producers in the Aztec economy were hardly touched by market exchange.
As Katz wrote:
he Aztec peasant did not in fact require any cratsmen to satisfy his needs in hous-
ing, clothing, and agricultural implements. he houses constructed of dried adobe
bricks were built by communal labour by members of the calpulli, articles of cloth-
ing woven out of maguey ibre were mostly made by the women and the agricul-
tural tool – the digging stick, called coa – was so simple to make that nobody was
required for this purpose. As yet no real division of labour between town and coun-
try existed in Mesoamerica. he cratsmen manufactured essentially luxury goods
for the nobility and helped to build great state buildings. (Katz 1972, 218)

Some writers have attempted to show that Mesoamerica constituted “a world

system” at the time of Spanish contact (Carmack 1996, 83–6). In some sense,
the case can be made that a regionally based globalization process, centered
on the Aztec empire, was in progress at the time. But the integration pro-
cess was more political and cultural than commercial; the economic links
were based largely on tributes and much less on market-based division of
labor. It should be stressed at this point that, unlike the cities of West Africa’s
Niger Bend, those of the Inca and Aztec kingdoms originated as religious
and administrative centers. hus, while the state systems of the West African
Middle Niger – Ancient Ghana, Mali, Songhay – were merchant empires by
deinition, those of the Americas of the same period were not. It is signii-
cant that there is no evidence of commercial intercourse between the two
great empires of the Americas, the Inca and Aztec (Bakewell and Holler 2010,
249; Hagen 1961, 580).5 he chronicles of the Aztec merchants’ proclamation
that “commerce began in Mexico-Tenochtitlan in the year 1504”6 should be
understood in this context to mean the beginning of commercial expansion,
not the actual beginning of commerce. here can be no doubt, therefore, that
the market economy had developed to a much greater extent in West Africa
than in any part of the Americas at the close of the iteenth century. Taken
along with evidence on agricultural technology and pastoral farming, it can
also be guessed that per capita incomes in West Africa should be higher.
By the “long 1850,” the situation had changed completely. he predomi-
nantly subsistence (nonmarket-oriented) economies of the “long 1400”
had become fully commercialized and the market economy had taken hold

5
As Bakewell and Holler put it, “Peru had never before [Spanish contact] dealt directly
with Central America and Mexico; nor the Andean province of Charcas with the Rio de la
Plata; nor the Greater Antilles with Panama” (Bakewell and Holler 2010, 249).
6
Quoted by Hagen (1961, 174).
 Reversal of Fortune and Socioeconomic Development 67

Table 2.1. Commodity production for export and population in Latin America
and the United States of America, 1850–1912

Latin America United States

Exports Population Exports Exports Population Exports
(3-yr. average per head (3-yr average per head
p.a. US$) (US$) p.a. US$) (US$)
1850 159,484,000 30,381,000 5.25 162,000,000 23,192,000 6.99
1870 344,123,000 38,628,000 8.91 400,000,000 39,818,000 10.05
1890 602,147,000 51,662,000 11.66 859,667,000 62,948,000 13.66
1912 1,580,534,000 77,456,000 20.41 2,307,000,000 94,569,000 24.39
Source: Computed from Victor Bulmer-homas, he Economic History of Latin America since
Independence. 2nd ed. (Cambridge: Cambridge University Press, 2003), appendix 1, tables A.1.1
and A.1.2, pp. 412 and 413.

in all of the Americas. In Spanish America, cities founded on commod-

ity production and commerce, as opposed to the pre-Columbian cities that
originated from religious worship and administration of centralized, milita-
ristic states, emerged, around which predominantly market-based domestic
economies developed (Bakewell 1984, 150; Burkholder and Johnson 2004,
171–3; Hoberman 1991). Similar developments occurred in the plantation
and mining economies of Brazil, the Caribbean, and North America. hey
dominated commodity production (raw materials, foodstufs, and bullion)
for Atlantic commerce, and provided the foundation upon which the inte-
grated nineteenth-century Atlantic economy was built. he evidence in
Table 2.1 makes this clear.
As the table shows, by 1850 the population of the Americas at 53,573,000
was just about equal to the more acceptable 1492 population (53,904,000).7
But the distribution was very diferent. he sparsely populated regions of
1492 (the areas that became the United States and Brazil) had more people
(United States, approximately 23 million in 1850 and 40 million in 1870;
Brazil, 7 million in 1850 and 10 million in 1870) than the more densely
populated regions of 1492 (Mexico, approximately 8 million in 1850 and
9 million in 1870; the Andean region, comprising Ecuador, Peru, Bolivia,
and Chile, 6 million in 1850 and 7 million in 1870) (Bulmer-homas 2006,

7
Bulmer-homas’s Latin American population does not include the population of the non-
Latin Caribbean (English- and Dutch-speaking Caribbean). he populations of Jamaica
(1844), Barbados (1851), Trinidad (1851), St Kitts (1844), Nevis (1844), and the Bahamas
(1851) add up to 643,248 (Watts 1987, 459, table 10.2).
68 Inikori

412, table A.1.1). he populations of Mexico and the Andean region in

1850 were far below those of 1492. he regional distribution of the pop-
ulation was a function of the regional distribution of commodity produc-
tion for Atlantic commerce, by far the most important development of the
period from 1492 to 1850: the phenomenal expansion that started in the
mid-seventeenth century and accelerated from 1850 to 1912 (evidenced by
Table 2.1) was central to population movement (including the development
of new urban centers) and socioeconomic development. Relative to the
1492 situation discussed earlier, it is clear that all regions of the Americas
shared in the great economic transformation associated with the phenom-
enal growth of commodity production for Atlantic commerce. But the rev-
olutionary impact was exceptional for one region – the United States of
America – which had become an industrialized region by 1912 (Chase-
Dunn 1980, 189–230; Engerman and Sokolof 1997, 260–304; North 1961,
1965, 673–705; heberge 1968).
Comparatively, West Africa had fallen far behind the Americas in
domestic market development and in commodity production for Atlantic
commerce by the “long 1850.” he annual average value of West Africa’s
commodity production for Atlantic commerce in the 1830s was ₤821,337
and that of transatlantic captive exports was ₤976,173, making a total of
₤1,797,510 (Inikori 1986, 56–7). his combined total is about 24 percent
of Atlantic commodity exports by Brazil alone in the late 1840s (Inikori
2002a, 181, table 4.4). During the last decade of the nineteenth century,
West Africa’s commodity exports to Great Britain (the region’s overwhelm-
ingly dominant trading partner at the time) was ₤2,357,900 in 1891 and
₤3,106,300 in 1899 (Inikori 2008, table 2). he 1891 igure would be about
2 percent of the Latin American igure for 1890. Even if we double the West
African igures to allow generously for exports to other Atlantic markets,
this still gives a mere 4 percent.
More signiicant, the development of markets and the market economy
in the Americas by the “long 1850” far exceeded that of West Africa, where
subsistence (nonmarket-oriented) production was extensive and dominant
in the domestic economies. For this reason, West Africa’s domestic econ-
omies were very weakly connected to the commodity production chain
of the nineteenth-century Atlantic economy, in contrast to the domestic
economies of the Americas. All this would suggest that West Africa’s Gross
Domestic Product (GDP) and GDP per capita were considerably less than
those of the Americas at this time. Looking back at the relative positions of
the two regions in the “long 1400,” there can be little doubt that a reversal of
fortune of immense proportions had occurred. he next section attempts to
 Reversal of Fortune and Socioeconomic Development 69

show the contribution of the transatlantic slave trade and the employment
of enslaved Africans in large-scale commodity production in the Americas
for Atlantic commerce to this reversal of fortune.

2.2 HISTORICAL ORIGINS OF THE REVERSAL OF FORTUNE

What factors explain the development path of the economies of the
Americas, relative to those of West Africa, in the period 1492–1850?
Clearly, the prime mover was not internal. Given the “rude state” (as the
classical economists would say) of those economies in 1492, market devel-
opment originating primarily from internal factors would have to depend
on sustained long-run internal population growth, as opposed to migration
from other regions of the globe. We know there was no internally generated
population growth in the Americas during the period. On the contrary, the
indigenous populations in all areas of the region experienced a phenome-
nal demographic catastrophe during the sixteenth century. he indigenous
population of central Mexico fell by about 85 percent between 1520 and
1620; the situation in the Andean area was similar. he experience was rep-
licated in the rest of the Americas that had been sparsely populated in 1492;
in the Caribbean islands, the indigenous population virtually disappeared
in the sixteenth century (Bakewell and Holler 2010, 196–7). Internal pop-
ulation growth was, therefore, not the source of market development and
the evolution of the market economy in the Americas during the period.
As we have already implied, commodity production for Atlantic commerce
was the prime mover. Our irst analytical task in this section, it follows, is
to show the contribution of enslaved Africans and analyze the relationship
between exports and the development of domestic markets and the market
economy in the Americas.
In the period 1500–1850, commodity production in the Americas for
Atlantic-wide commerce faced two major problems that were related at
some point – transportation cost and the availability of labor for large-
scale production far beyond the scope of family labor. Given the rudi-
mentary nature of transportation technology of the period, the cost of
moving bulky goods across the oceans was extremely high. To be able
to move such goods across the Atlantic and still secure for them large
markets that included not just the rich but the middle class and the poor,
the cost of production in the Americas had to be very low, in the irst
instance. Without meeting this precondition, very little was likely to hap-
pen in this sphere.
70 Inikori

he story of rice production in the Americas narrated by Peter Coclanis

conirms the point. Before Columbus reached the Americas in 1492, the
production of rice had been well developed in Asia, Africa, and Europe.
When entrepreneurs found out that rice was in demand in the Americas,
there was enough rice in Asia, Africa, and Europe that could have been
shipped to the Americas. But this did not happen because of transportation
cost. Instead, the Americas had to develop their own rice production. As
the American producers achieved economies of scale and production cost
fell in the seventeenth and eighteenth centuries, they were able to export
their rice to Europe. Subsequently, the new transportation technologies
of the Industrial Revolution in England allowed Asian producers to take
over the European markets and threatened to do the same in the Americas,
until high-tech rice production in Arkansas confronted Asian producers
(Coclanis 2007).
he problem of labor for large-scale production in the Americas arose
from factor proportions – the abundance of land in relation to population,
evidenced by the generally low population densities shown earlier. he
problem was compounded by the catastrophic decline of the indigenous
population in the sixteenth century that let huge tracts of land vacant for
European immigrants to occupy at little cost, politically and inancially.
his made large-scale production, with a gang-labor organization of legally
free workers, virtually impossible, because such workers preferred to secure
their own land and engage in small-scale independent production. What
is more, the proletarianization (transformation of the masses into workers
owning little or no property) process in Europe, Africa, and Asia was yet to
generate the kind of pressure that pushed millions of voluntary migrants
from Europe and Asia to the Americas during the late nineteenth and twen-
tieth centuries. Yet small-scale production of the valued bulky commodi-
ties, such as sugar, cofee, raw cotton, rice, indigo, and so on, was high cost,
and large-scale production was needed to secure extensive economies of
scale. he importation of indentured servants from Europe was tried as a
solution with very limited and unsatisfactory outcome, as these servants
picked up land and became independent small-scale producers at the end
of their ive-year contract. Only the massive importation and enslavement
of captives from western Africa proved an efective solution to the labor
problem.
Between 1500 and 1820, ive out of every six persons crossing the
Atlantic to the Americas were African captives transported for enslavement
(Manning 1990, 37). Our knowledge of the total numbers transported con-
tinues to improve. he “New Database” by Eltis and Richardson has revised
 Reversal of Fortune and Socioeconomic Development 71

their earlier conirmation of Curtin’s 11 million to 12.5 million (Eltis and

Richardson 2008, 40–1, table 1.6; 2004b, 182, table 1).8 While the improved
knowledge is important for our understanding of what happened in west-
ern Africa, as will be shown shortly, what is important for the Americas is
the demographic evidence showing the dominance of African labor in the
production of the major commodities that fueled the growth of Atlantic
commerce during the period (Inikori 2002a, 156–214).
Brazil was the irst large-scale producer of sugar in the Americas. Its
sugar-producing regions were populated by Africans and their descendants
from the seventeenth to the nineteenth century: in Pernambuco, they were
68.2 percent of the total population in 1798 and 64 percent in 1872; Bahia,
78.6 percent and 72.3 percent; Rio de Janeiro, 64.3 percent and 48.8 percent;
for the gold-producing regions, Minas Gerais, 74.6 percent and 57.6 percent,
and Mato Grosso, 80.4 percent (1798); for the cofee-producing region of
Sao Paulo, 43.5 percent (1872). People of African descent were 61.2 percent
of Brazil’s total population in 1798 and 58 percent in 1872. Note that because
enslaved Africans were the forced specialized producers in plantation agri-
culture and gold production, they constituted virtually the total labor force
in those sectors during the period (Inikori 2002a, 189 and 190, table 4.5.A
and table 4.5.B). For purposes of large-scale commodity production for
Atlantic commerce, Brazil was an extension of Africa at the time.
During the eighteenth century, leadership in plantation agriculture
devoted to production for Atlantic commerce shited from Brazil to the
Caribbean, where it exploded. With that explosion, a demographic trans-
formation followed that made the Caribbean even more African than
Brazil. By the late eighteenth century, enslaved Africans employed in plan-
tation agriculture, mostly sugar plantations (with cotton, cofee, and indigo
as minor crops), made up at least 80 percent of the total population of most
islands. In the irst half of the nineteenth century, the growth of raw cotton
production in the southern states of the United States of America became
the epicenter of the Atlantic economy (Inikori 2002a; North 1961). Again,
the cotton plantations depended entirely on enslaved Africans for their gang
labor. he only sector of production for Atlantic commerce in the econo-
mies of the Americas in which the labor of enslaved Africans was not over-
whelmingly dominant during the period was bullion production in Spanish

8
As I argued elsewhere, the problem of incomplete data, though now reduced, remains; the
new database, like the old, “must be treated as a large sample, rather than a comprehen-
sive listing, of all slaving voyages” (Inikori 2003a). With all the emotions and bad blood
the controversy unnecessarily generated, readers may now be able to tell who won and
who lost.
72 Inikori

America. Here, the indigenous population supplied much of the labor in

silver mining. Yet the labor of enslaved Africans was not insigniicant. he
mine census of 1570 listed thirty-seven hundred enslaved Africans in the
mining camps of New Spain, making up about 45 percent of the laboring
population (Inikori 2002a, 181–2). In 1590, enslaved Africans provided on
average approximately 14 percent of the labor employed in the major silver
mines of “Nueva Espana,” Zacatecas, Guadalajara, and Guadiana (Tandeter
2006, 318, table 9.1). On the other hand, enslaved Africans completely
dominated gold production in Spanish America during the period, the
same way that they did in Brazil (Inikori 2002a, 184–5). As Michael Taussig
noted, enslaved Africans, not “Andean Indians,” provided the labor force in
the chief gold-producing region of the Spanish American empire located in
the Cauca Valley of Colombia (Taussig 1977, 397).
hus, it is beyond reasonable doubt that enslaved Africans and their
descendants overwhelmingly dominated the gang labor employed in
large-scale commodity production in the Americas for Atlantic commerce
between 1500 and 1850. he below-subsistence cost of African slave labor
(the slaves produced part of their own subsistence in their provision plots
worked during their limited free time) and the economies of scale in large-
scale plantation agriculture brought down production cost greatly. Hence,
those commodities sold at considerably reduced prices, the high cost of
transportation across the Atlantic notwithstanding. For example, in the
second decade of the seventeenth century the price of tobacco in England
was twenty to forty shillings (sterling) a pound. he phenomenal expansion
of plantation production in the third decade brought down the plantation
price in the Americas to less than a penny a pound by 1630, bringing down
the price in England to one shilling or less in the 1670s (Davis 1954, 80). In
general, plantation production of the other commodities experienced sim-
ilar developments during the period. he magnitude of the price reduction
created rapidly growing markets for the American commodities in Europe,
thereby fueling the rapid growth of Atlantic commerce from 1650 to 1850
(Davis 1967; Inikori 2002a).
he explosive growth of highly specialized plantation agriculture and
mining propelled the growth and development of interregional division
of labor and intra-American trade, which stimulated the development of
markets and the geographical spread of the market economy across the
Americas. In regions that had less than one person per square mile in 1492,
and market development was, therefore, at an extremely low level, such as
Brazil and North America, the developments were truly astonishing. In
Brazil, the expansion of specialized, large-scale plantation agriculture and
 Reversal of Fortune and Socioeconomic Development 73

gold mining, all devoted to production for Atlantic markets, created large
domestic markets for foodstufs, drat animals, and other goods. Because
the domestic demand for these goods was very large, it created the market
opportunity for a second round of specialized production and pushed fur-
ther the frontier of market-oriented production. he details of the process
are beginning to emerge from research (Barickman 1998; Luna and Klein
2003, 2004). Particularly signiicant in the Brazilian case, enslaved Africans
were also employed in the production of these products for the domestic
market (Luna and Klein 2004).
In what became the United States of America that was very much like
Brazil in 1492, the growth of specialized, large-scale plantation agricul-
ture in the Caribbean islands from 1650 to 1800 ofered the initial market
opportunity for the growth of market-oriented production and the gradual
evolution of the market economy. his was particularly important for New
England and the Middle Atlantic colonies that were locked up in a sub-
sistence production cul-de-sac, because they lacked the natural resources
needed to compete efectively in the production of plantation commodities
for Atlantic commerce.9 he Caribbean market for foodstufs, lumber, drat
animals, and shipping and other business services ofered them the irst
major opportunities to enter the evolving Atlantic economy. he phenome-
nal expansion of specialized, large-scale plantation agriculture in the south-
ern states of the United States during the irst half of the nineteenth century,
particularly the spectacular growth of the cotton economy, multiplied the
opportunities a hundredfold. Somewhat similar to the Brazilian case, but
on a much larger scale, the interaction between the specialized southern
plantation economy and the maritime trading and shipping economy of
the northeast opened up a large domestic market for foodstufs that stim-
ulated the growth of commercial food production in the western region of
the United States. Ultimately, this gave rise to the growth of complementary
economies in the southern, western, and northeastern regions of the United
States. his tripartite interregional division of labor, centered on the south-
ern cotton economy, was central to the subsequent industrialization of the
northeastern region, which, in the long-run, set the economy of the United
States apart from all the others in the Americas. he latter development has
given rise to arguments attempting to explain the United States’ divergence

9
While new and increasingly sophisticated research on the subject continues to grow, an
extensive literature already exists, of which the following is just a sample (Inikori 2002b;
Khan and Sokolof 1993; McCusker and Menard 1985; North 1961; Shepherd and Walton
1972; Sokolof 1988; Sokolof and Khan 1990).
74 Inikori

in a manner that fails to accurately identify the real prime movers in the
long-run development of the U.S. economy. his is particularly the case for
the factor endowments paradigm and the institutional framework.
he factor endowments argument classiies the economies of the Americas
into three according to their factor endowments – those with climates and
soils particularly suited for growing highly valued crops with extensive
economies of scale, making the employment of gang slave labor economic;
those with dense indigenous populations forced by European colonizers to
provide coerced labor; and those with climates and soils unsuited for plan-
tation agriculture employing gang slave labor and with no dense indigenous
populations that could be forced to provide coerced labor. New England
and the mid-Atlantic region of the United States (the northeastern region)
belonged to the last category. Without an Atlantic perspective, the factor
endowments paradigm explains virtually in isolation the industrialization
of the northeastern United States strictly in terms of its factor endowments.
he plantation and mining slave economies are presented, again virtually
in isolation, as performing relatively poorly in the long run because they
employed coerced labor (Engerman and Sokolof 1997). he institutional
framework Acemoglu, Johnson, and Robinson employed, though located
within the broader context of the Atlantic world, also fails to show elab-
orately the causal links between the slave economies of the Americas and
industrialization in the northeastern United States (Acemoglu et al., 2002;
2005a).
he factor endowments argument surprisingly overlooks (as does,
in diferent ways, the institutional analysis by Acemoglu and his coau-
thors) the well-documented fact that the northeastern region’s commer-
cial exploitation of its natural resource endowments depended, in the
irst instance, on the market opportunities created by the specialized,
large-scale plantation and mining economies of the Americas employing
coerced labor: Without the market opportunities arising from the divi-
sion of labor centered on the slave economies of the Atlantic world, New
England and the mid-Atlantic economies would not have been able to
escape from their subsistence production cul-de-sac at the time they did.
his subject has been treated elaborately elsewhere (Inikori 2002a, 156–
214, 2007, 75–9).10 here is no compelling reason to repeat the details
here. A summary treatment will suice.

10
A large and growing literature on the subject, a sample of which is presented in footnote 9,
provides direct and indirect support for the argument, empirically and logically.
 Reversal of Fortune and Socioeconomic Development 75

he comparative per capita income of the free population in the British

American colonies around 1774 testiies to the general economic predica-
ment of New England and the mid-Atlantic region at the beginning of the
last quarter of the eighteenth century: New England, ₤38.2; mid-Atlantic,
₤45.8; southern mainland, ₤92.7; West Indies, ₤1,200.00. Consistent with
the economic data, demographic evidence shows unambiguously that New
England was not attractive to European immigrants between 1680 and 1780,
as it sufered a net loss of thirty-one thousand during the period (Engerman
and Sokolof 1997, 266). Only ater the market opportunities created by the
slave-based economies of the Americas had become suiciently large for
the mid-Atlantic and New England economies to extensively exploit their
factor endowments were they able to break out of their subsistence produc-
tion cul-de-sac and create the conditions for industrial take-of between
the late eighteenth and the mid-nineteenth century. Douglass North’s cen-
tral argument, that “an analysis of the United States economic development
must necessarily be put into the context of the expansion of the Atlantic
economy,”11 is hard to fault.
Once the factor endowments analysis is appropriately located within the
evolving Atlantic economy, the main arguments become more intelligible.
While large-scale production and specialization in the slave-based econ-
omies of the Americas were central to the expansion of Atlantic markets,
on the basis of which revolutionary developments occurred in the econ-
omies of Old England and New England, the pattern of income distribu-
tion associated with the exploitation of slaves and the political economy
that developed with it over time tended to impose limits on the long-run
development of those economies. One major factor behind divergent devel-
opments in the northeastern United States and the slave-based plantation
economies of the Americas was the relative economic and political power
of large landholders. he overwhelming dominance of large landholders in
the plantation economies led to state policies that favored free trade at the
expense of industrialization. In contrast, the economic and political power
of merchants, maritime shipping entrepreneurs, inanciers, and incipient
industrialists in the northeastern United States, relative to planters in the

11
As North put it, in the 1830s, “it was cotton that initiated the concomitant expansion
in income, in the size of domestic markets, and creation of the social overhead invest-
ment (in the course of its role in the marketing of cotton) in the Northeast which were to
facilitate the subsequent rapid growth of manufactures. . . . [W]hen income from cotton
exports, including shipments to textile mills in our own Northeast, grew from $25 million
in 1831 to $70 million in 1836, it set in motion the whole process of accelerated expansion
which culminated in 1839” (1961, 66–9, 68–9).
76 Inikori

south, led to state policies that provided incentives for the development
of manufacturing.12 Ultimately, as stated earlier, industrialization set the
United States apart from the rest of the Americas on the eve of World War
I. he division of labor between industrial and nonindustrial economies
in the global economic system of 1914 would continue to hold back pro-
gress in the nonindustrial economies for decades ater the two world wars.
But the preponderance of evidence makes it abundantly clear it cannot be
reasonably argued that the slave-based economies of the Americas would
have been better of in 1850 without the labor of enslaved Africans. On
the contrary, the evidence shows beyond reasonable doubt that all the
economies of the Americas beneited immensely from the employment of
enslaved Africans in large-scale commodity production for Atlantic com-
merce. Paradoxically, as the evidence makes clear, the nonslave economies
of North America beneited even more in the long run.
One important comparative insight generally overlooked in Kenneth
Pomeranz’s study of China strengthens this point further. he region of
China with the greatest potential for industrialization in the nineteenth
century, the Yangzi Delta, failed to realize its potential because the other
provinces of China, with resource endowments similar to those of the
northeastern United States, had no economic reason to develop comple-
mentary economies with the Yangzi Delta (Pomeranz 2000). Robert Allen’s
recent argument, similar to that of North cited earlier, is also pertinent.
Writing about the contribution of coal energy to the Industrial Revolution,
he says:
We habitually describe coal as a “natural” resource. It is true that there would have
been no coal trade had there been no coal in the ground. hat much was a fact of
nature. But the mere presence of coal was not suicient to cause the coal trade. It
was only activated by the growth of the international economy. Coal was a social
artifact as well as a natural fact. (Allen 2009, 90)

Examining the reversal of fortune as it concerns West Africa, the irst

question we need to answer is whether the export of captives and the
employment of those captives to produce commodities in the Americas
for Atlantic commerce had adverse efects on the competitiveness of West
Africa’s economies in the development of commodity production for
Atlantic commerce during the period. his is an important issue, because

12
hose policies were unsuccessfully resisted by planters in the southern United States,
whose short-term interests, like their counterparts in the rest of the Americas who suc-
cessfully blocked them, favored free trading of primary commodities for imported manu-
factures from Europe.
 Reversal of Fortune and Socioeconomic Development 77

global trade in products was one of the major factors in the commercializ-
ing process in West Africa up to the sixteenth century. he other factor was
demography. he interaction between these factors gave rise to the process
of enlargement of state systems (especially in the Niger Bend), which, in
turn, supported the commercializing process, as we saw in the earlier part
of this chapter.
As the evidence shows, trade in products derived from West Africa’s labor
and natural resources was dominant during the irst two hundred years of
European commercial enterprise in the region. Apart from Senegambia,
most West African regions traded almost entirely in products. With all the
needed caution in using the available data, the indication is that there was
practically very little or no slave trading in Sierra Leone, the Windward
Coast, the Gold Coast, and the Bight of Benin until the mid-seventeenth
century (Eltis and Richardson 2008, 46–7, table 1.7). During this period,
West Africa’s Atlantic commerce in products grew steadily. In the late
1530s, a Portuguese oicial, writing about Portuguese trade in western
Africa, declared:
I do not know in this kingdom [Portugal] any yoke of land, toll, tithe, excise or any
other royal tax which is more certain in each yearly return than is the revenue of
the commerce of Guinea. It is, besides, so peaceful a property, quiet and obedient,
that – without our having to stand at the touch-hole of the bombard with lighted
match in one hand, and lance in the other – it yields us gold, ivory, wax, hides, sugar,
pepper, and it would produce other returns if we sought to explore it further.13

It is signiicant that the Portuguese oicial did not mention revenue from
slave trading. Clearly, this was not an oversight. From Senegambia to south-
eastern Nigeria, the product trade was overwhelmingly dominant – gold,
hides and skins, ivory, and pepper (the main products from Senegambia
through the Upper Guinea Coast to the Windward Coast); gold, the main
product from the Gold Coast (southern modern Ghana); red pepper and
cotton cloths, the main products from the Bight of Benin; ivory, the main
product from southeastern Nigeria (Bight of Biafra).
he early focus of European traders on product trade reinforced the com-
mercializing process in West Africa that was ongoing before the European
contact. he hides and skins trade in the Upper Guinea Coast stimulated
the growth of stock breeding and cattle trade, which enriched the Fulani
cattle rearers economically and empowered them politically in the Futa
Jallon plateau (Upper Guinea Coast) (Barry 1988, 38–9, 1992, 265, 289).

13
Quoted in Vogt (1979, 93).
78 Inikori

Ultimately, the focal point of the product trade from the mid-iteenth to
the mid-seventeenth century was the Gold Coast. he gold and kola trade
had made the region a major part of the globalization process led by the
Niger Bend merchant diaspora. Having no textile industry of its own, it was
a major market for textiles from other parts of West Africa before European
contact. hus, a specialist in early Portuguese trade on the Gold Coast, John
Vogt, noted:
Despite a large population concentration on the Mina coast [Gold Coast] since a
very early period, animal and vegetable ibers with which to weave cloth remained
in short supply. he small quantities of fabric imported into the region from the
Sudanese empires or from Nigeria before the Portuguese arrival brought high prof-
its. However, this pre-European trade never fully satisied local demands. When the
Portuguese arrived at Mina in the late 1460s, they discovered that a ready market
awaited them for any cloth that they wished to carry there. (Vogt 1979, 67)

Further expansion of the gold trade by European traders extended and

intensiied the preexisting trade links between the Gold Coast and other
West African regions, especially the trading area of the powerful Benin
kingdom (southwest Nigeria), which exported to the Gold Coast cotton
textiles (“Benin cloth” in European records) partly produced locally, but
largely procured from northeastern Yorubaland (Akintoye 1965, 544–6;
Inikori 2009a, 92). he European traders entered this intra-West African
trade as carriers during the sixteenth and early seventeenth century, rais-
ing the volume of the trade considerably and stimulating the growth of
Benin’s long-distance trade with its northeastern neighbors (Ryder 1965,
203–4). he multiplier efects of the general growth of trade in products in
the Gold Coast region during the period created favorable conditions for
the Akan merchants (referred to in European records as Akani or accan-
ists) to invest their growing proits from trade in commercial agriculture
aimed at the expanding domestic market that was fueled by the growth of
urban centers (Kea 1982, 85–91; Wilks 1977). To meet the expanding labor
needs of these developments, the Gold Coast imported labor from other
regions of western Africa during the sixteenth century and the irst half
of the seventeenth century, with European traders as carriers (Kea 1982;
Vogt 1979, 59–92; Wilks 1977). he general expansion of domestic trade
and the deepening of the market economy in all regions of West Africa at
this time is relected in the ratio of currencies in the imports brought by
the European traders of the period (Inikori 2007b).
hen came a radical shit of European traders’ demand from products
to captives as the growth of demand for captives to be enslaved in the
Americas grew manifold from the mid-seventeenth century. From this
 Reversal of Fortune and Socioeconomic Development 79

time period, the Gold Coast was transformed from a region that imported
labor to one that exported captives. Captives exported from the Gold Coast
to the Americas in 1651–1700 were about and forty-three times (in num-
ber) those exported in the one hundred and ity years from 1501 to 1650
(Eltis and Richardson 2008, 46–7, Table 1.7). his was a general phenome-
non across all regions of West Africa, although it was more in some than in
others. here is a general consensus among historians, consistent with the
evidence, that demand drove this development. But there is some confusion
concerning factors explaining supply.
For economic historians who believe the needs of European colonists
in the Americas for coerced labor able to withstand the epidemiological
environment of the region explain the growth of demand, what has to
be explained is the generally low prices African suppliers were willing to
accept, prices that, ater adding transatlantic transportation cost, still made
it economic to employ enslaved Africans to produce commodities in the
Americas for Atlantic commerce. A political economy explanation appears
the most logical and persuasive, particularly when presented in a global per-
spective. he operating factor in this explanation is political fragmentation
that exposed communities without the protection of relatively strong states
to capture (at little cost) by individuals or organized groups responding to
export demand. Under such conditions, the main determinant of export
prices was internal transportation and bulking cost, and the level of com-
petition among the European exporters. he empirical evidence for this in
western Africa is strong; it is also consistent with the evidence showing the
operation of the same factor in the rise and demise of captive exports in
Europe, the global perspective (Inikori 2003b).
A competing explanation is based on the productivity of agricultural
labor in western Africa. As Manning argued, “prices were low because the
low level of agricultural productivity [in western Africa] set by hoe agri-
culture limited the value of labor even when it was scarce” (Manning 1990,
21). Gareth Austin has recently elaborated the point. Austin’s argument is
nuanced and painstakingly articulated. But, in the end, Austin believes:
“Relatively low labour productivity was a premise of the external slave
trades [in sub-Saharan Africa]” (Austin 2008b, 996), the explanation for
which rests with the physical environment that makes intensive agriculture
virtually uneconomic (Austin 2007, 2008a, 2008b. 2009). his environmen-
tal argument, presented in diferent ways by other historians (Ilife 1995),
ignores the comparative case of the rise and demise of captive exports in
Europe. No historian has argued, for example, that low labor productivity in
agriculture or manufacturing explains captive exports from the British Isles
80 Inikori

during medieval times. What is more, it is not based on scientiic collection

and analysis of empirical evidence. Soil scientists, studying tropical agricul-
ture across the globe, who have done that, draw contrary conclusions. heir
data show that intensive agriculture based on a high input of purchased
fertilizers in sub-Saharan Africa is a function of factor ratios (population
to land ratios) and access to markets, and not the physical environment
(Vanlauwe et al. 2002, 12, 38–9). As will be concluded later, the factor pro-
portions and the level of development of markets for agricultural products
resulted from the adverse efects of the transatlantic slave trade.
Placed in an Atlantic world history framework, there is enough evidence
(some of which we presented earlier) to uphold the proposition that the
transatlantic slave trade and African slavery in the Americas adversely
afected the competitiveness of West Africa, relative to the Americas, in the
production of commodities for Atlantic commerce. Clearly, labor productiv-
ity in the pre-Columbian Americas, whether in agriculture or manufactur-
ing, was not higher than it was in West Africa of the same period. Given the
evidence presented earlier, and other available evidence,14 there is no rea-
sonable basis to argue that the Americas could have outcompeted western
Africa in the production of commodities for Atlantic commerce without
the employment of enslaved Africans made available by the transatlantic
slave trade. Had all of western Africa experienced the same levels of state
enlargement and military organization that protected citizens against cap-
ture and export as the late medieval and early modern European states
did – as also did some of the states that emerged in West Africa during the
crisis of the transatlantic slave trade (such as the Futa Jallon Islamic state,
the Asante state, the Dahomean state, and the earlier Benin state in Nigeria
(Inikori 2003b)) – the Americas would have had no enslaved Africans to
employ and West Africa would have continued to develop its production
of commodities for Atlantic commerce that began in the mid-iteenth cen-
tury (Inikori 1992, 44–53).
he Gold Coast (modern southern Ghana) best illustrates the conse-
quences of aborting the development of commodity production for Atlantic
commerce and the substitution of captive exports. Urban development was
reversed, as was the division of labor between town and country (Kea 1982,
11–32, 53–5, 85–91, 105–7, 136–68). Interregional trade between the Gold
Coast and other West African regions was also adversely afected as the
middlemen slave traders got European and Asian manufactures from the

14
For more details, see Inikori (1992, 39–53).
 Reversal of Fortune and Socioeconomic Development 81

European traders in direct barter exchange for captives, and turned the
coastal societies into enclave economies, only violently connected to their
hinterlands as sources for captives forcefully procured without commer-
cial intercourse, in the irst instance. his happened in all the West African
regions as their hinterlands became victim sources of captives rather than
suppliers of manufactures for the coastal societies (Inikori 2009a).
he addition of the demographic and political impact of captive exports
largely accounts for the low level of market development and the limited
geographical spread of the market economy in West Africa by the “long
1850” noted earlier. As a general order of magnitude, there is now enough
evidence to uphold the inference that, but for the impact of the transatlan-
tic slave trade, the population of West Africa would have been at least twice
what it was in 1850 (Manning 1990, 85).15 he research of soil scientists on
tropical agriculture across the globe shows unambiguously that the agricul-
tural resources of the region were more than adequate to support the addi-
tional population. Ater the considerable growth of population since 1850,
the soil scientists still hold that:
Although some areas are densely populated, the overall pressure on the land in sub-
Saharan Africa is low and the application of inputs is minimal. . . . he use of inputs
is higher where the crops are grown for export. . . . Although improved fallows work
well in the humid tropics of Cameroon there is no demand for them as land is plen-
tiful and fertility is yet to be depleted. (Vanlauwe et al. 2002, 12, 38–9)

hus, the evidence presented here supports strongly the conclusion that
the reversal of fortune established earlier, and the lowly position occupied
by West Africa in the nineteenth-century Atlantic economy, resulted from
the adverse efects of the transatlantic slave trade and African slavery in the
Americas on the competitiveness of West Africa in commodity production
for Atlantic commerce, together with the demographic and sociopolitical
impact of captive exports in the region.

References
Acemoglu, Daron, Simon Johnson, and James A. Robinson. (2002). “Reversal of
Fortune: Geography and Institutions in the Making of the Modern World Income
Distribution.” Quarterly Journal of Economics, 117: 1231–94.

15
Manning’s view that the sub-Saharan African population in 1850 would have been
100 million instead of 50 million is based on Curtin’s igure of 11 million for the entire
transatlantic captive exports. he new estimate of 12.5 million mentioned earlier would
make Manning’s view a conservative one.
82 Inikori

(2005a). “Institutions as the Fundamental Cause of Long-Run Growth.” In Handbook

of Economic Growth, Volume 1A, edited by Philippe Aghion and Steven N. Durlauf,
385–472. Amsterdam: Elsevier.
Acemoglu, Daron, Simon Johnson, and James A Robinson. (2005b). “he ‘Rise of
Europe’: Atlantic Trade, Institutional Change and Economic Growth.” American
Economic Review, 95: 546–79.
Akintoye, S. A. (1965). “he North-Eastern Yoruba Districts and the Benin Kingdom.”
Journal of the Historical Society of Nigeria, IV(4).
Allen, Robert C. (2009). he British Industrial Revolution in Global Perspective.
Cambridge: Cambridge University Press.
Austin, Gareth. (2007). “Labour and Land in Ghana, 1874–1939: A Shiting Ratio and
an Institutional Revolution.” Australian Economic History Review, 47(1): 95–120.
(2008a). “Resources, Techniques, and Strategies South of the Sahara: Revising the
Factor Endowments Perspective on African Economic Development, 1500–2000.”
Economic History Review, 61(3): 587–624.
(2008b). “he ‘Reversal of Fortune’ hesis and the Compression of History:
Perspectives from African and Comparative Economic History.” Journal of
International Development, 20: 996–1027.
(2009). “Cash Crops and Freedom: Export Agriculture and the Decline of Slavery in
Colonial West Africa.” International Review of Social History, 54: 1–37.
Bailyn, Bernard. (2005). Atlantic History: Concept and Contours. Cambridge, MA:
Harvard University Press.
Bakewell, Peter. (1984). “Mining in Colonial Spanish America.” In Cambridge History of
Latin America, Vol. II, edited by Leslie Bethell. Cambridge: Cambridge University
Press.
Bakewell, Peter and Jacqueline Holler. (2010). A History of Latin America to 1825. 3rd
ed. Chichester, UK: Wiley-Blackwell.
Barickman, B. J. (1998). A Bahian Counterpoint: Sugar, Tobacco, Cassava and Slavery in
the Reconcavo, 1780–1860. Stanford, CA: Stanford University Press.
Barry, Boubacar. (1988). Senegambia and the Atlantic Slave Trade. Cambridge: Cambridge
University Press, 1998; original French edition, 1988.
(1992). “Senegambia from the Sixteenth to the Eighteenth Century: Evolution of
the Wolof, Sereer and ‘Tukuloor.’” In General History of Africa, V: Africa from the
Sixteenth to the Eighteenth Century, edited by B. A. Ogot. Oxford, Berkeley, Paris:
Heinemann, University of California Press, UNESCO.
Baudin, Louis. (1967). “he Inca Empire was Socialistic.” In History of Latin American
Civilization: Sources and Interpretations, Volume 1: he Colonial Experience, edited
by Louis Hanke, 101–10. Boston, MA: Little, Brown & Company.
Benjamin, homas. (2009). he Atlantic World: Europeans, Africans and heir Shared
History, 1400–1900. Cambridge: Cambridge University Press.
Borah, Woodrow Wilson and Sherburne F. Cook (1960). he Aboriginal Population of
Central Mexico in 1548. Berkeley : University of California Press.
(1963). he Aboriginal Population of Central Mexico on the Eve of the Spanish Conquest.
Berkeley : University of California Press.
Brooks, Francis J. (1993). “Revising the Conquest of Mexico: Smallpox, Sources, and
Populations.” Journal of Interdisciplinary History, XXIV(1): 1–29.
 Reversal of Fortune and Socioeconomic Development 83

Bulmer-homas, Victor, John H. Coatsworth, and Roberto Cortes Conde, eds. (2006).
he Cambridge Economic History of Latin America, Volume I: he Colonial Period.
Cambridge: Cambridge University Press.
Burkholder, Mark A. and Lyman L. Johnson. (2004). Colonial Latin America. 5th ed.
Oxford: Oxford University Press.
Canizares-Esguerra, Jorge and Erik R. Seeman, eds. (2007). he Atlantic in Global
History, 1500–2000. Upper Saddle River, NJ: Pearson.
Carmack, Robert M. (1996). “Mesoamerica at Spanish Contact.” In he Legacy of
Mesoamerica: History and Culture of a Native American Civilization, edited by
Robert M. Carmack, Janine Gasco, and Gary H. Gossen, 83–6. Upper Saddle River,
NJ: Prentice Hall.
Carmack, Robert M., Janine Gasco, and Gary H. Gossen, eds. (1996). he Legacy of
Mesoamerica: History and Culture of a Native American Civilization. Upper Saddle
River, NJ: Prentice Hall.
Chase-Dunn, Christopher. (1980). “he Development of Core Capitalism in the
Antebellum United States: Tarif Politics and Class Struggle in an Upwardly Mobile
Semiperiphery.” In Studies of the Modern World System, edited by Albert Bergesen.
New York: Academic Press.
Cissoko, S. M. (1984). “he Songhay from the 12th to the 16th Century.” In General
History of Africa, IV: Africa from the Twelth to the Sixteenth Century, edited by D.
T. Niane, 187–210. London, California, Paris: Heinemann, University of California
Press, UNESCO.
Coclanis, Peter A., ed. (2005). he Atlantic Economy during the Seventeenth and
Eighteenth Centuries: Organization, Operation, Practice, and Personnel. Charleston:
University of South Carolina Press.
(2007). “Reorienting Atlantic History: he Global Dimensions of the ‘Western’ Rice
Trade.” In he Atlantic in Global History, 1500–2000, edited by Jorge Canizares-
Esguerra and Erik R. Seeman, 111–28. Upper Saddle River, NJ: Pearson.
Cook, Sherburne F. and Lesley B. Simpson. (1948). he Population of Central Mexico in
the Sixteenth Century. Berkeley : University of California Press.
Davis, Ralph. (1967). A Commercial Revolution: English Overseas Trade in the Seventeenth
and Eighteenth Centuries. London: Historical Association.
(1973). he Rise of the Atlantic Economies. London: Weidenfeld and Nicolson.
(1954). “English Foreign Trade, 1660–1700.” Economic History Review, 2nd
series, VI.
Denevan, William M., ed. (1976). he Native Population of the Americas in 1492.
Madison: University of Wisconsin Press.
(1992). he Native Population of the Americas in 1492, 2nd ed. Madison: University of
Wisconsin Press. 1st ed. 1976.
Diie, Bailey W. (1945). Latin American Civilization: Colonial Period. Harrisburg, PA:
Stackpole Sons.
Eltis, David, Frank D. Lewis, and Kenneth L. Sokolof, eds. (2004). Slavery in the
Development of the Americas. Cambridge: Cambridge University Press.
Eltis, David and David Richardson. (2004). “Prices of African Slaves Newly Arrived
in the Americas, 1673–1865: New Evidence on Long-Run Trends and Regional
Diferentials.” In Slavery and the Development of the Americas, edited by David
84 Inikori

Eltis, Frank D. Lewis, and Kenneth L. Sokolof. Cambridge: Cambridge University

Press.
(2008). “A New Assessment of the Transatlantic Slave Trade.” In Extending the
Frontiers: Essays on the New Transatlantic Slave Trade Database, edited by David
Eltis and David Richardson. New Haven, CT: Yale University Press.
Engerman, Stanley L. and Kenneth L. Sokolof. (1997). “Factor Endowments, Institutions,
and Diferential Paths of Growth among New World Economies.” In How Latin
America Fell Behind: Essays on Economic Histories of Brazil and Mexico, 1800–1914,
edited by Stephen Haber. Stanford, CA: Stanford University Press.
Hagen, Victor Wolfgang Von. (1961). he Ancient Sun Kingdoms of the Americas: Aztec,
Maya, Inca. Cleveland, OH: he World Publishing Company.
Hanke, Lewis. ed. (1967). History of Latin American Civilization: Sources and Interpretations,
Volume 1: he Colonial Experience. Boston, MA: Little, Brown & Company.
Hoberman, Louisa S. (1991). Mexico’s Merchant Elite, 1590–1660: Silver, State, and
Society. Durham, NC and London: Duke University Press.
Hogendorn, Jan. (2006). “Cowries.” In History of World Trade since 1450, Volume 1, edited
by John J. McCusker, 176–78. Farmington Hills, MI: Macmillan, hompson Gale.
Hogendorn, Jan and Marion Johnson. (1986). he Shell Money of the Slave Trade.
Cambridge: Cambridge University Press.
Hopkins, A. G. (2009). “he New Economic History of Africa.” Journal of African
History, 50(2): 155–77.
Hunwick, John O. (1999). Timbuktu and the Songhay Empire: Al-Sadi’s Ta’rikh al-sudan
down to 1613 and other Contemporary Documents. Leiden, Boston, Koln: Brill.
Ilife, John. (1995). Africans: he History of a Continent. Cambridge: Cambridge
University Press.
Inikori, Joseph E. (1979). “he Slave Trade and the Atlantic Economies, 1451–1870.” In
he African Slave Trade from the Fiteenth to the Nineteenth Century, 56–87. Paris:
UNESCO.
(1986).” West Africa’s Seaborne Trade, 1750–1850: Volume, Structure and Implications.”
In Figuring African Trade: Proceedings of the Symposium on the Quantiication and
Structure of the Import and Export and Long Distance Trade in Africa in the 19th
Century (C. 1800–1913), St. Augustin 3-6 January 1983, edited by G. Liesegang,
H. Pasch, and A. Jones. Berlin: Dietrich Reimer Verlag.
(1992). he Chaining of a Continent: Export Demand for Captives and the History of
Africa South of the Sahara, 1450–1870. Kingston, Jamaica: Institute of Social and
Economic Research, University of the West Indies, UNESCO Project.
Inikori, Joseph E. (2002a). Africans and the Industrial Revolution in England: A Study in
International Trade and Economic Development. Cambridge: Cambridge University
Press.
Inikori, Joseph E. (2002b). “he Development of Entrepreneurship in Africa:
Southeastern Nigeria during the Era of the Trans-Atlantic Slave Trade.” In Black
Business and Economic Power, edited by Alusine Jalloh and Toyin Falola. Rochester,
NY: University of Rochester Press.
(2003a). “he Known, the Unknown, the Knowable, and the Unknowable: Evidence
and the Evaluation of Evidence in the Measurement of the Trans-Atlantic Slave
Trade.” In Ghana in Africa and the World: Essays in Honor of Adu Boahen, edited by
Toyin Falola, 535–65. Trenton, NJ: Africa World Press.
 Reversal of Fortune and Socioeconomic Development 85

Inikori, Joseph E. (2003b). “he Struggle against the Trans-Atlantic Slave Trade: he
Role of the State.” In Fighting the Slave Trade: West African Strategies, edited by
Sylviane A. Diouf, 170–98. Athens: Ohio University Press.
Inikori, Joseph E. (2007a). “Africa and the Globalization Process: Western Africa, 1450–
1850.” Journal of Global History, 2(1): 63–85.
Inikori, Joseph E. (2007b). “Changing Commodity Composition of Imports into West
Africa, 1650–1850: A Window into the Impact of the Transatlantic Slave Trade
on African Societies.” In he Transatlantic Slave Trade: Landmarks, Legacies,
Expectations: Proceedings of the International Conference on Historic Slave Route
held at Accra, Ghana on 30 August – 2 September, 2004, edited by James Kwesi
Anquandah, 57–80. Accra: Sub-Saharan Publishers, 2007.
Inikori, Joseph E. (2008). “Africa and Globalization: Transition from the Atlantic Slave
Trade to Commodity Production for Export in West Africa.” Paper Presented at
the Conference, Orb and Sceptre: Global and Imperial Histories, he European
University Institute, Florence, May 14–16.
Inikori, Joseph E. (2009a). “English versus Indian Cotton Textiles: he Impact of Imports
on Cotton Textile Production in West Africa.” In How India Clothed the World: he
World of South Asian Textiles, 1500–1850, edited by Giorgio Riello and Tirthankar
Roy, 85–114. Leiden: Brill.
Inikori, Joseph E. (2009b). “Atlantic Perspectives on European Economic History, 1650–
1850.” Paper Presented at the XVth World Economic History Congress, Utrecht,
Netherlands, August 3–7.
(2010). “West Africa’s Niger Bend in Global Perspective, 1000–1591 C.E.” Paper
Presented at the American Historical Association Annual Meeting, San Diego, CA,
January 7–10.
Johnson, Marion. (1970). “he Cowrie Currencies of West Africa.” Journal of African
History, XI, Parts I and II.
Katz, Friedrich. (1972). he Ancient American Civilizations. New York: Praeger, original
German ed., Kindler Verlag, 1969.
Kea, Ray A. (1982). Settlements, Trade, and Polities in the Seventeenth-Century Gold
Coast. Baltimore, MD: Johns Hopkins University Press.
Khan, B. Zorina and Kenneth L. Sokolof. (1993). “‘Schemes of Practical Utility’:
Entrepreneurship and Innovation among ‘Great Inventors’ in the United States,
1790–1865.” he Journal of Economic History, 53(2): 289–307.
Lockhart, James and Stuart B. Schwartz. (1983). Early Latin America: A History of
Colonial Spanish America and Brazil. Cambridge: Cambridge University Press.
Luna, Francisco Vidal and Herbert S. Klein. (2003). he Evolution of the Slave Society
and Economy of Sao Paulo, from the 1760s to the 1850s. Stanford, CA: Stanford
University Press.
(2004). “African Slavery in the Production of Subsistence Crops: he Case of Sao
Paulo in the Nineteenth Century.” In Slavery in the Development of the Americas,
edited by David Eltis, Frank D. Lewis, and Kenneth L. Sokolof, 120–49. Cambridge:
Cambridge University Press.
Maddison, Angus. (2001). he World Economy: A Millennial Perspective. Paris:
Organisation for Economic Co-Operation and Development.
(2003). he World Economy Historical Statistics. Paris: Organisation for Economic
Co-Operation and Development.
86 Inikori

Mancke, Elizabeth and Carole Shammas, eds. (2005). he Creation of the British Atlantic
World. Baltimore, MD: Johns Hopkins University Press.
Manning, Patrick. (1990). Slavery and African Life: Occidental, Oriental, and African
Slave Trades. Cambridge: Cambridge University Press.
(2006). “African Connections with American Colonization.” In he Cambridge
Economic History of Latin America, Volume I: he Colonial Period, edited by
Victor Bulmer-homas, John H. Coatsworth, and Roberto Cortes Conde, 43–71.
Cambridge: Cambridge University Press.
McCusker, John J. and Russell R. Menard. (1985). he Economy of British America,
1607–1789. Chapel Hill: University of North Carolina Press.
McCusker, John J. and Kenneth Morgan, eds. (2000). he Early Modern Atlantic
Economy. Cambridge: Cambridge University Press.
McIntosh, Roderick James. (1998). he Peoples of the Middle Niger: he Island of Gold
Oxford: Blackwell.
(2005). Ancient Middle Niger: Urbanism and the Self-Organizing Landscape.
Cambridge: Cambridge University Press.
(2008). “Before Timbuktu: Cities of the Elder World.” In he Meanings of Timbuktu,
edited by Shamil Jeppie and Souleymane Bachir Diagne. Cape Town, South Africa:
HSRC Press.
Metraux, Alfred. (1967). “Despotism or Socialism?” In History of Latin American
Civilization: Sources and Interpretations, Volume 1: he Colonial Experience, edited
by Louis Hanke, 110–18. Boston, MA: Little, Brown & Company.
Moraes Farias de, P. F. (2003). Arabic Medieval Inscriptions from the Republic of Mali:
Epigraphy, Chronicles, and Songhay-Tuareg History. Oxford: Oxford University
Press.
Niane, D. T. (1984a). “Mali and the Second Mandingo Expansion.” In General History of
Africa, IV: Africa from the Twelth to the Sixteenth Century, edited by D. T. Niane,
117–71. London, California, Paris: Heinemann, University of California Press,
UNESCO.
(1984b). “Relationships and Exchanges among the Diferent Regions.” In General
History of Africa, IV: Africa from the Twelth to the Sixteenth Century, edited by D.
T. Niane, 614–34. London, California, Paris: Heinemann, University of California
Press, UNESCO.
(1984c). “Conclusion.” In General History of Africa, IV: Africa from the Twelth to
the Sixteenth Century, edited by D. T. Niane, 673–86. London, California, Paris:
Heinemann, University of California Press, UNESCO.
Newson, Linda A. (1985). “Indian Population Patterns in Colonial Spanish America.”
Latin American Research Review, XX: 41–74.
Newson, Linda A. (2006). “he Demographic Impact of Colonization.” In he Cambridge
Economic History of Latin America, Volume I: he Colonial Period, edited by V.
Bulmer-homas, John H. Coatsworth, and Roberto Cortes Conde, 143–84.
Cambridge: Cambridge University Press.
North, Douglass C. (1961). he Economic Growth of the United States, 1790–1860. New
York: W. W. Norton, 1966, irst published by Prentice Hall, 1961.
North, Douglass C. (1965). “Industrialization in the United States.” In he Cambridge
Economic History of Europe, Vol. VI Part II, edited by H. J. Habakkuk and M.
Postan. Cambridge: Cambridge University Press.
 Reversal of Fortune and Socioeconomic Development 87

Nunn, Nathan. (2007). “Historical Legacies: A Model Linking Africa’s Past to Its Current
Underdevelopment.” Journal of Development Economics, 83(1): 157–75.
(2008). “he Long Term Efects of Africa’s Slave Trades.” Quarterly Journal of
Economics, 123(1): 139–76.
Nunn, Nathan and Leonard Wantchekon. (2011). “he Slave Trade and the Origins of
Mistrust in Africa.” American Economic Review, 101: 3221–52.
O’Rourke, Kevin H. and Jefrey G. Williamson. (1999). Globalization and History: he
Evolution of a Nineteenth-Century Atlantic Economy. Cambridge, MA: MIT Press.
Pomeranz, Kenneth. (2000). he Great Divergence: China, Europe, and the Making of the
Modern World Economy. Princeton, NJ: Princeton University Press.
Prados de la Escosura, Leandro. (2006). “he Economic Consequences of Independence.”
In he Cambridge Economic History of Latin America, Volume I: he Colonial
Period, edited by V. Bulmer-homas, John H. Coatsworth, and Roberto Cortes
Conde, 463–504. Cambridge: Cambridge University Press.
Ryder, A. F. C. (1965). “Dutch Trade on the Nigerian Coast during the Seventeenth
Century.” Journal of the Historical Society of Nigeria, III(2).
Shepherd, James F. and Gary M. Walton. (1972). Shipping, Maritime Trade, and the
Economic Development of Colonial North America. Cambridge: Cambridge
University Press.
Sokolof, Kenneth L. (1988). “Inventive Activity in Early Industrial America: Evidence
from Patent Records, 1790–1846.” he Journal of Economic History, XLVIII(4):
813–50.
Sokolof, Kenneth L. and B. Zorina Khan. (1990). “he Democratization of Invention
during Early Industrialization: Evidence from the United States, 1790–1846.” he
Journal of Economic History, L(2): 363–78.
Solow, Barbara L., ed. (1991). Slavery and the Rise of the Atlantic System. Cambridge:
Cambridge University Press.
Storey, Rebecca and Randolph J. Widmer. (2006). “he Pre-Columbian Economy.” In
he Cambridge Economic History of Latin America, Volume I: he Colonial Period,
edited by Victor Bulmer-homas, John H. Coatsworth, and Roberto Cortes Conde,
73–106. Cambridge: Cambridge University Press.
Tandeter, Enrique. (2006). “he Mining Industry.” In he Cambridge Economic History
of Latin America, Volume I: he Colonial Period, edited by Victor Bulmer-homas,
John H. Coatsworth, and Roberto Cortes Conde. Cambridge: Cambridge University
Press.
Taussig, Michael. (1977). “he Evolution of Rural Wage Labour in the Cauca Valley
of Colombia, 1700–1970.” In Land and Labour in Latin America: Essays on the
Development of Agrarian Capitalism in the Nineteenth and Twentieth Centuries,
edited by Kenneth Duncan and Ian Rutledge. Cambridge: Cambridge University
Press.
heberge, James D., ed. (1968). Economics of Trade and Development. New York: John
Wiley.
Trigger, Bruce G. and Wilcomb E. Washburn, eds. (1996). he Cambridge History of
the Native Peoples of the Americas: Volume I, North America, Part 1. Cambridge:
Cambridge University Press.
Vanlauwe, B., J. Diels, N. Sanginga, and R. Merckx, eds. (2002). Integrated Plant
Nutrient Management in Sub-Saharan Africa: From Concept to Practice. New
88 Inikori

York: CABI Publishing in Association with the International Institute of Tropical

Agriculture.
Vogt, John. (1979). Portuguese Rule on the Gold Coast, 1469–1682. Athens: University
of Georgia Press.
Watts, David. (1987). he West Indies: Patterns of Development, Culture and Environmental
Change since 1492. Cambridge: Cambridge University Press.
Wilks, Ivor. (1977). “Land, Labour, Capital and the Forest Kingdom of Asante: A Model
of Early Change.” In he Evolution of Social Systems, edited by J. Friedman and M.
J. Rowlands. London: Duckworth.
 3

he Impact of Malaria on African Development

over the Longue Durée

David N. Weil

An extensive literature examines the impact of disease generally, and

malaria in particular, on economic development in the world today. Malaria
is especially a focus in Africa because of the high prevalence of the disease.
here are some 300 million malaria cases annually on the continent, and it
is the leading cause of under-ive mortality. A widely quoted estimate is that
malaria reduces growth of GDP per capita by 1.3% per year in the African
countries most alicted (Gallup and Sachs, 2001). Africa is also the histori-
cal home of malaria, and the region with a physical environment most sup-
portive of the disease.
In this chapter, I ask how malaria inluenced economic development in
Africa in the period before European contact. In particular, I ask whether
malaria played a role in holding back African development and whether
malaria shaped the geographic pattern of development. Knowing the role
of malaria is part of the much larger project of understanding the deter-
minants for Africa’s early development in comparison to other parts of the
world and also understanding heterogeneity within the continent itself.
Answering the question I pose requires overcoming a number of very
serious obstacles. he irst is the the absence of data. How prevalent was
malaria in Africa? Was it more or less important, in either absolute terms
or relative to other diseases, than it is today? Our understanding of the role
played by malaria in the contemporary world relies on data generated by
government statistical agencies and NGOs as well as on ield surveys by

[email protected]. I am grateful to Fred Piel, Ronald D. Lee, Andrew Mason, and

Gordon McCord for sharing data; to Emilio Depetris Chauvin, Federico Droller, Evan
Friedman, and Scott Weiner for research assistance; and to Quamrul Ashraf, seminar par-
ticipants at Indiana University and Pennsylvania State University, and participants at the
Weatherhead Center for International Economics conference on “Understanding African
Poverty in the Longue Durée” for helpful comments.

89
90 Weil

trained professionals. Even modern doctors cannot accurately diferenti-

ate malarial illness from numerous other tropical maladies without using
microscopes or rapid diagnostic tests. No statistical compilations exist from
the period before Europeans started coming to Africa, and what textual
sources are available do not provide suicient information to make a seri-
ous quantitative analysis.
Much of the contribution of this paper lies in trying to overcome this
irst obstacle, that is, to measure the extent of malaria in the pre-European
period. Although there are no paper records, malaria let a lasting imprint
in human genes that can be interpreted in order to gauge the extent of the
disease. Speciically, I look at the prevalence of the gene that causes sickle
cell disease. Individuals who carried one copy of this gene were aforded
a signiicant measure of protection against malaria, but those who car-
ried two copies died before reaching reproductive age. he equilibrium
prevalence of the gene therefore depended on how severe the selective
pressure of malaria was – more concretely, on the fraction of children
who didn’t carry the gene who died from malaria before adulthood. One
can thus infer the severity of malaria from the prevalence of the sickle
cell gene in modern populations. I carry out such calculations, and con-
struct estimates of the severity of malaria in Africa in the period before
European contact. I ind that in the more alicted regions, malaria low-
ered the probability of surviving to adulthood by about ten percentage
points. he reduction in the rate of malaria mortality over the last 500
years has been of approximately the same magnitude as the reduction in
mortality from other causes.
Even with an estimate of malaria’s health impact in the pre-contact
period, establishing how malaria afected African economic development
is no easy matter. As will be seen below, economists do not agree on how
malaria afects economic growth today, despite far superior knowledge of
the extent of the disease, data on the productivity of workers who are ill
or healthy, and concrete information on the functioning of the economy.
hus an assessment of how malaria afected African economic develop-
ment 500 years ago is necessarily highly speculative. I use simple economic
models and modern data to explore two channels through which malaria
might have afected growth: through loss of life (mortality) and through its
efect on the productivity of workers and the learning ability of children
(morbidity). I ind relatively little evidence that malaria would have had a
very signiicant efect on economic outcomes through either of these chan-
nels. Because malaria mortality is concentrated among the very young, the
 he Impact of Malaria 91

cost to society from malaria deaths is not large. In the case of morbidity,
we know from modern data that the burden of malaria among adults is
relatively light compared to other conditions. It is also notable that some of
the areas of highest development prior to European contact, such as West
Africa and the Niger Bend, had environments that were especially support-
ive of malaria transmission.
he rest of this paper is organized as follows. In Section 13.1, I paint a pic-
ture of the level of development in Africa prior to the arrival of Europeans,
using a variety of indicators. I also discuss diferent theories about the fac-
tors that inluenced early development in Africa, with a focus on the role of
disease. Section 13.2 discusses the impact of malaria today and the biology
of the relationship between malaria and the sickle cell trait. In Section 13.3,
I construct estimates of the historical burden of malaria, based on current
prevalence of the sickle cell trait. Section 13.4 turns to an examination of
the potential economic impact of the malaria burden estimated in the pre-
vious section. Section 13.5 concludes.

3.1 AFRICAN ECONOMIC GROWTH OVER

THE LONGUE DUR É E
I begin with an examination of variation in the level of development, com-
paring Africa to the rest of the world as well as looking at variation within
Africa itself. here is of course no single measure that summarizes the state
of economic development, and there are notable problems with all of the
data that are available. I thus look at a variety of indicators. I take as my
point of reference the year 1500, at least to the extent that this is practical,
to assess Africa’s standing prior to interaction with Europeans.

3.1.1 Population Density

The first measure I examine is population density. Density can be used,
within reason, as a measure of technological development of societies
that are living close to subsistence, which is a fair characterization of
pretty much the entire world during the period of interest. According to
a simple Malthusian model (see discussion in Section 3.4.1), countries
with more advanced agricultural technologies will not have better fed
people, just more of them. Within a given country, changes in popu-
lation density over time are a good measure of technological change
(Kremer 1993).
92 Weil

Log of Population Density in 1500

-5.9 - 0.2

0.3 - 0.4

0.5 - 1.8

1.9 - 3.8

No Data

Figure 3.1. Population density in 1500.

Figure 3.1 shows a measure of population density. he data are from

McEvedy and Jones (1978). he data have been mapped to the borders of
current countries, most of which did not exist at the time. he map divides
the countries into four groups of equal size (the underlying data are in the
appendix to this chapter). he igure shows that the highest levels of popu-
lation density were in the parts of Eurasia generally viewed as the most
developed: Europe and Turkey, South Asia, China, Japan, and Indonesia.
he second densest group of countries includes the Middle East, the Aztec,
Maya, and Inca empires, and Egypt. Within sub-Saharan Africa, the coun-
tries that fall within this group include a large swathe of West Africa, and
East Africa from Ethiopia and Sudan to Tanzania. Central Africa makes
up most of the third most dense group, while the fourth most dense group
is composed of the Americas (with the exceptions of the areas mentioned
earlier), Australia, the countries of the Sahara, and the southern part of the
African continent.

3.1.2 Land Quality

One obvious problem with using population density to measure the extent
of development is that land quality may vary. For a given level of agricultural
technology, population density will be higher, the more productive is the
agricultural environment. hus density will not be a good measure of the
level of agricultural technology. To address this issue, I use data on the suit-
ability of land for agriculture data from Ramankutty and colleagues (2002),
which is in turn constructed from high-resolution data sets for croplands,
 he Impact of Malaria 93

climate, and soil characteristics. heir index represents the probability that
a particular 0.5-degree grid cell (about 50 x 50 km at the equator) may be
cultivated. hus, each grid cell observation takes a value between zero and
one. he authors assume that land suitability for cultivation is a function
only of climate (temperature, precipitation, and potential sunshine hours)
and soil properties (total organic components measured by carbon density
and nutrient availability based on soil pH). Other biophysical and socio-
economic factors such as topography and irrigation or market price and
incentive structure are omitted. It is important to note that the index does
not account for the productivity of a particular piece of land, but simply
whether the characteristics of the land are favorable for crop cultivation. To
give some sense of what these data look like, within the United States, the
Midwest region has a mean of 0.77, the South 0.54, the Northeast 0.38, and
the West (which includes Alaska) 0.22. he country with the highest value
in the world is Uruguay, at 0.95. Within Africa, the highest levels of aver-
age land quality are found in Malawi (0.77), Lesotho (0.69), and Tanzania
(0.64). he full set of values is available in the appendix.
To estimate the density of population conditional on the quality of land,
I regress the country-level population density measures described previ-
ously on average land quality and the within-country variance of land qual-
ity. he results are shown in Table 3.1. I experiment with using both the
level and the logarithm of population density as the dependent variable,
although this seems to make little diference. he irst and fourth columns
of the results show mean density by continent, relative to Africa, which
is the omitted category. Europe and Asia have signiicantly higher density,
while Oceana and the Americas have lower densities. he second and ith
columns show that land quality has signiicant predictive power for popula-
tion density. However, as the third and sixth columns show, adjusting for
land quality has only a small efect on the continental measures of density.

3.1.3 Urbanization
Beyond population, a related measure of development over long histori-
cal periods is urbanization. he presence of large cities is taken as an
indicator of the existence of a signiicant agricultural surplus, a devel-
oped transportation infrastructure, and/or an efective ruling entity.
Cities served as loci for development in industry and knowledge. Finally,
by their nature, cities are relatively well observed by contemporary trav-
elers and historians.
94 Weil

Table 3.1. Cross-country regressions for population density in 1500

Independent Pop density in 1500 Log of pop density in 1500

Variables (1) (2) (3) (4) (5) (6)

Mean of Land 9.440*** 7.554*** 1.682*** 1.396***
Quality (2.22) (1.95) (0.56) (0.50)
Variance of 41.27** 28.12** 5.004 3.061
Land Quality (19.02) (12.25) (4.27) (3.01)
Asia 6.578*** 5.882*** 1.153*** 1.032***
(2.09) (1.94) (0.27) (0.24)
Europe 10.72*** 9.265*** 1.485*** 1.248***
(2.44) (2.19) (0.33) (0.28)
Oceania −2.515*** −3.088*** −1.949* −2.016**
(0.78) (0.73) (1.01) (0.95)
America −2.066*** −2.423*** −0.860*** −0.929***
(0.69) (0.88) (0.27) (0.28)
Constant 3.157*** 0.436 −0.306 0.543*** −0.103 −0.0309
(0.69) (0.64) (0.89) (0.16) (0.18) (0.22)
Observations 145 145 145 145 145 145
R-squared 0.316 0.141 0.396 0.361 0.113 0.43
Robust standard errors in parentheses.
*** p<0.01, ** p<0.05, * p<0.1.

Table 3.2. World city sizes by region in 1500

Sub-Saharan North Asia Europe N. America S. America

Africa Africa
20,000–49,999 22 17 28 78 19 4
50,000–199,999 2 3 58 19 4 2
200,000+ 0 1 8 0 0 0
Source: See text.

I use a newly constructed database that combines data from six sources
(although some borrow from others): Bairoch, Batou, and Chevre (1988),
Chandler (1987), De Vries (1984), Eggimann (2000), Modelski (2003), and
Rozenblat, Lennert, and Van Hamme (2011). he data are on population
for historical cities, matched with their locations. I use data on all available
cities in the year 1500.
Table 3.2 shows a breakdown of city sizes for the world as a whole. I
divide Africa into sub-Saharan and northern regions. Figure 3.2 shows the
geographic distributions of cities in Africa in 1500. By this metric, early
development in Africa is strongly concentrated in the west of the continent:
Mali, Niger, Burkina Faso, and most of all Nigeria. Outside this region,
 he Impact of Malaria 95

Large African Cities in 1500

City Population
(thousands)
16 – 35
36 – 65
66 – 130
131 – 400

Figure 3.2. Distribution of cities in Africa in 1500.

extraordinarily few cities appear. Overall, the level of urbanization in Africa

is similar to that of North America (where all but one city is in present-day
Mexico or Guatemala). North Africa has roughly the same level of urban-
ization as sub-Saharan Africa (although in a much smaller area), while
Europe has a signiicantly higher level and Asia higher still.

3.1.4 State Antiquity

While cities are an indirect measure of the existence of state power, one
can also try to look at this dimension of development directly. Chanda
and Putterman (2007) present an index of “state antiquity” that attempts
to measure the degree of a country’s political development. For every ity-
year period between the years 1 CE and 1500 CE, a country is awarded
a score based on three factors: the existence of a government at a supra-
tribal level, the proportion of the current territory of the country that the
96 Weil

State Antiquity Index in 1500

0.0
0.1 - 0.4
0.5 - 0.7
0.8 - 1.0
No Data

Figure 3.3. State history in 1500.

government controlled, and whether the government was indigenous or

externally imposed. hese half-century scores are then aggregated for
the full iteen-hundred-year period. he underlying sources for Chanda
and Putterman’s data are country entries in Encyclopaedia Britannica.
Putterman and Weil (2010) show that the Chanda and Putterman index is
a good predictor of modern economic development. One disadvantage of
this data relative to other measures presented here is that rather than look-
ing at a single point in time, it aggregates information from a millennium
and a half. Further, to the extent that available documentary sources are
slanted toward coverage of currently developed countries, the index will
be similarly biased.
he data are presented in Figure 3.3, again broken down into four country
groups – see the appendix for the underlying data. he highest group for this
measure includes a broad arc in Asia from Japan and China through Iran,
Turkey, Western Europe with the exception of Spain and Ireland, Bolivia,
Peru, and, on the African continent, Egypt, Morocco, Tunisia, and Ethiopia.
he second group includes much of Southeast Asia and Scandinavia,
Mexico, Nigeria, Cameroon, Ghana, Botswana, and several states in the
Sahara. he third group includes most of the rest of West Africa, while
the lowest group in terms of state development includes America north of
Mexico, the Amazon basin, Southern Africa, Australia, and New Zealand. It
is also important to note that there are good number of countries for which
there is no information at all for this index, although presumably there is
a bias toward these being countries with relatively undeveloped states that
let behind few records.
 he Impact of Malaria 97

Average Technology Adoption Index in 1500

0.0 - 0.2
0.3 - 0.5
0.6 - 0.8
0.9 - 1.0
No Data

Figure 3.4. Technology in 1500.

3.1.5 Technology
As a inal measure of development, I look directly at production technol-
ogy. While observing actual quantities of production at a lag of ive hun-
dred years is close to impossible, historical sources are better suited to
determine whether a particular piece of technology was used in a given
area. Figure 3.4 shows an index of technology adoption in 1500, created
by Comin, Easterly, and Gong (2010). Again, the underlying values are in
the appendix. he index is an aggregate of ive sub-indices in the areas of
communications, agriculture, military, industry, and transportation. For
example, in the agriculture sub-index, a country gets a score of four if the
primary form of subsistence is plow cultivation, three if it is hand cultiva-
tion, two if it is pastoralism, and one if it is hunting and gathering.1 In the
transportation sub-index, points are awarded for possession of the wheel,
horse-powered vehicles, the magnetic compass, and ships capable of cross-
ing the Indian, Atlantic, and Paciic oceans.
he highest values in the technology index are those given to Japan,
China, Turkey, and almost all of Europe. he second highest group includes
South and Southeast Asia, the Middle East and North Africa, Ethiopia,
Nigeria, Mali, and Burkina Faso. Most of the rest of Africa is in the third
highest technology group, along with Mexico, Nepal, and Mongolia. he
rest of the Americas, Australia, New Zealand, New Guinea, Botswana, and
Namibia make up the lowest group.

1
However, Hopkins (1973) argues that environmental conditions made West Africa unsuit-
able to animal-powered plow agriculture. hus the nonuse of plows is not necessarily a
measure of available technology.
98 Weil

3.1.6 Discussion
he use of a variety of perspectives to measure early economic development
is motivated both by the multidimensional nature of development itself and
by the diiculties of measuring any particular aspect of development from
a distance of ive hundred years. Economists like assigning numerical val-
ues to quantities, making statistical comparisons, and so on. But we should
remember that in the case of many of the things measured here, the fac-
tual base on which these numerical comparisons are built consists of a few
historical records, the observations of travelers, and bits of archeological
evidence.
Subject to these caveats, the overall picture painted by these data shows,
irst, a great deal of heterogeneity of development within Africa. Ignoring
North Africa, which was generally part of the Mediterranean world, the
highest levels of development by many indicators are found in Ethiopia (also
relatively isolated geographically and culturally from much of the rest of the
continent) and in the broad swathe of West African countries running from
Cameroon and Nigeria eastward along the coast and the Niger River. In this
latter region, the available measures show a level of development just below
or sometimes equal to that in the belt of Eurasia running from Japan and
China, through South Asia and the Middle East, into Europe. Depending
on the index used, West Africa was above or below the level of development
in the Northern Andes and Mexico. Much of the rest of Africa was at a
signiicantly lower level of development, although still more advanced than
the bulk of the Americas or Australia.
Several theories have been proposed to explain Africa’s level of develop-
ment on the eve of the modern era. Many of these have focused on sources
of low agricultural productivity, which in turn prevented the accumulation
of surpluses that could support urbanization, specialization, and technolog-
ical advance. Diamond (1997) stresses the availability of plants and animals
that could be easily domesticated (for example, sub-Saharan Africa had
only four large seeded grasses that could be developed for cultivation com-
pared to thirty-nine in Eurasia), and also the absence of Africa of a single
large agro-climactic zone over which food production technologies could
be shared. Other authors stress the impediments to agricultural productiv-
ity in tropical zones, including periods of torrential rain, leading to erosion
of cultivated soil, and the absence of a temperate zone pattern of seasonal
sunlight, which encourages development of easily stored seeds. Masters and
McMillan (2001) point to the absence of frost, called “nature’s great execu-
tioner” for its role in suppressing agricultural pests. Webb (2006) describes
 he Impact of Malaria 99

West Africa as characterized by poor soils and low caloric yields from the
principal grain crops. Looking beyond agricultural productivity, Gallup,
Sachs, and Mellinger (1999) argue that African development was held back
by lack of ocean-navigable rivers penetrating the continent and the location
of population centers away from coastal lowlands, both of which impeded
the development of waterborne trade.
Among the theories proposed to explain Africa’s relatively low level of
development, several authors have pointed to the continent’s severe disease
environment. As the historical cradle of humanity, African organisms that
found humans to be congenial hosts had a signiicant head start over those
in other parts of the world. Humans who migrated out of Africa let many
of their parasites behind. In particular, two of the most signiicant tropi-
cal diseases, malaria and yellow fever, did not make the jump from Africa
to the Americas until Europeans started sailing between the two areas
(McNeill, 1977). Gallup, Sachs, and Mellinger (1999) ind that beyond the
expected efect of tropical location in predicting disease intensity, there is
a further efect particular to sub-Saharan Africa. Discussions of the role of
disease in shaping African development generally focus on two diseases in
particular: malaria and trypanosomiasis (sleeping sickness). Webb (2006)
describes these as the two “most signiicant of the microparasitic infections
that inluenced human settlement patterns and culture.”
As discussed in Akyeampong (2006), much of our knowledge of health
in Africa comes from accounts of travelers. Arab chronicler Al-Bakri, who
wrote during the eleventh century, described a large town in northern
Ghana as an area “where the inhabitants have yellow complexions because
they sufer from fever and splenetis. here is hardly one who does not com-
plain of one or the another.” An enlarged spleen is a symptom of chronic
malaria.
However, the theory that the health environment in Africa was particu-
larly bad is not universally accepted. Premodern society was not healthy in
any part of the world. People living in temperate zones, though freed from
tropical parasites, faced many other health challenges brought on by cli-
mate, including nutritional deiciencies, diseases spread by being indoors,
and diseases that made the jump from livestock with whom people were
forced to share living quarters. Mabogunje and Richards (1985) point out
that European travelers who described Africa’s harsh disease environment
may have been overly inluenced by the terrible toll disease took on their
own cohorts (who had none of the immunity acquired by Africans as chil-
dren) as well as by the efects of new pathogens introduced by Europeans
themselves. hey conclude: “he grounds for supposing tropical Africa was
100 Weil

any more badly afected by epidemic diseases in the past than, say, western
Europe or North America are yet to be established.”

3.2 MALARIA
3.2.1 he Impact of Malaria Today
In the World Health Organization AFRO region, which roughly corre-
sponds to Africa south of the Sahara, in 2004 the annual death rate from
malaria for children under ive was 0.59 percent per year, implying that
roughly 3 percent of children would die of malaria before their ith birth-
day. Malaria accounted for 15.8 percent of the total mortality burden in
this age group. Malaria deaths are highly concentrated among the young.
Ater age ive, people who live in malaria endemic areas have developed
partial immunity to the disease (the exception is pregnant women, par-
ticularly women in their irst pregnancy, who lose their immunity). Among
people aged iteen to twenty-nine, for example, the death rate from malaria
was just 0.006 percent per year, and only 1.2 percent of deaths were due to
malaria (Mathers, Fat, and Boerma 2008).
Beyond its efect on mortality, malaria also exacts a cost in terms of ill
health and low productivity among those who survive. Table 3.3A shows
data on the incidence, prevalence, and average duration of malaria in sub-
Saharan Africa. Incidence is the number of cases per year, while prevalence
is the fraction of the population that has the condition at any point in time.
In the zero to four age group, there is an average of more than one case per
child per year. Among adults, however, only 17.5 percent of individuals are
alicted each year, and because of the short duration of episodes, less than a
quarter of one percent of individuals are ill at the average point in time.
Table 3.3B shows similar data for the neurological sequelae of childhood
malaria. A small fraction of children who recover from cerebral malaria suf-
fer from lifelong brain damage, including epilepsy, spasticity, and learning
disabilities. Because the condition lasts a lifetime, the prevalence of these
symptoms among the adult population is actually higher than the preva-
lence of episodic malaria.
Economic efects resulting from malaria may follow many channels. he
direct efect is that adults are unable to work during episodes of the disease,
and may be signiicantly weakened for a period aterward. Repeated infec-
tion with malaria is associated with anemia in children and adults. here
are also efects on the human capital accumulation of children. Beyond the
obvious efects in cases of neurological sequelae, children who sufer from
 he Impact of Malaria 101

Table 3.3A. Malaria episodes

Age Incidence Prevalence Average duration Death rate

(per 100,000) (per 100,000) (years) (per 100,000)
0–4 120,000 1,644 0.01 559
5–14 17,500 240 0.01 42
15–44 17,500 240 0.01 33
45–59 17,500 240 0.01 36
Source: Murray and Lopez (1996).

Table 3.3B. Malaria – Neurological Sequelae

Age Incidence Prevalence Average duration Death rate

(per 100,000) (per 100,000) (years) (per 100,000)
0–4 164 365 37.1 0
5–14 0 701 – –
15–44 0 617 – –
45–59 0 474 – –
Source: Murray and Lopez (1996).

malaria have reduced cognitive functioning and increased absenteeism

from school. High childhood mortality from the disease may also induce
parents to plan for extra births, and consequently invest fewer resources
in each surviving child. Malaria may also afect the seasonal and spatial
pattern of economic activity. Gallup and Sachs (2001) give examples of
regions in Europe that were largely uninhabited prior to the eradication
of malaria, and became productive agricultural regions aterward. he
need to shit the harvest cycle away from seasons in which malaria is most
prevalent may alter the distribution of crops that farmers plant away from
those that would otherwise be the most proitable. Malaria also occasions
direct use of resources. In heavily burdened countries, up to 40 percent
of public health expenditures, 30–50 percent of inpatient hospital admis-
sions, and 60 percent of outpatient health clinic visits are attributable to
malaria (Aregawi 2010).
he overall magnitude of these economic efects of malaria is the subject
of contentious debate. Most famously, the Abuja declaration of 2005, signed
by ity-three African heads of state claimed that “malaria has slowed eco-
nomic growth in African countries by 1.3% per year as a result of which
GDP for African countries is now 37% lower than it would have been in
102 Weil

the absence of malaria.” his conclusion is based largely on cross-country

regression studies, particularly Gallup and Sachs (2001). Ashraf, Lester, and
Weil (2008) argue that this sort of regression sufers from severe identi-
ication problems. Instead, they use a simulation model with key param-
eters based on well-identiied microeconomic studies. heir conclusion is
that eliminating malaria in sub-Saharan Africa would increase income per
capita by a total of only 2 percent, with that beneit taking more than half a
century to arrive.2

3.2.2 Malaria and Sickle Cell Disease

Malaria is caused by the plasmodium parasite, which is transmitted to
humans through the bite of a female anopheles mosquito. Within the
human body, the parasite follows a complex life cycle. In part of that cycle,
plasmodium merozoites invade and multiply within the body’s red blood
cells, which eventually burst and release merozoites to infect new cells. Early
symptoms of malaria include fever, chills, severe headache, and vomiting.
In severe cases, these are followed by respiratory distress, severe anemia,
or cerebral malaria. Infants are protected from the disease in the irst few
months of life by a combination of maternal antibodies and characteristics
of the structure of fetal hemoglobin. In malaria endemic areas, most chil-
dren have developed substantial immunity by the age of ive.
Four types of malaria infect humans, of which the two most common are
Plasmodium falciparum and Plasmodium vivax. P. falciparum is the most
deadly. Africa accounts for 85 percent of world malaria cases and 90 per-
cent of world malaria deaths. he geographical pattern of malaria’s severity
is largely determined by the climactic conditions that support mosquito
breeding as well as by the mix of mosquito species present. here are signif-
icant diferences in the vectorial capacity among the approximately twenty
species of anopheles that transmit malaria, based on factors such as the mos-
quito’s preferred targets, biting frequency, and lifespan. he most efective
vector, Anopheles gambiae, is the principal vector in Africa.
Kiszewski and colleagues (2004) construct an index of “malaria ecol-
ogy” that takes into account both climactic factors and the dominant vector
species to give an overall measure of how congenial the environment is to
the spread of malaria. he index does not take into account public health
interventions such as swamp draining or diferences among countries in
economic development or health care infrastructure. In other words, it

2
See Weil (2010) for a review of this literature.
 he Impact of Malaria 103

Malaria Ecology Index

38.1
15
0

Figure 3.5. Malaria ecology in Africa.

represents the component of malaria variation among countries that is

exogenous to human intervention. he index is calculated for grid squares
of one half degree longitude by one half degree latitude. Average values of
the index for countries are shown in the appendix. With the exception of
New Guinea and some areas of Southeast Asia, Africa is the only part of the
world in which the index reaches its highest levels. Areas in which malaria
played a signiicant role historically but has now been eradicated, such as
Greece, southern Italy, and the American South, all have relatively low
values for the malaria ecology index. Even within Africa, there is substan-
tial variation in the index, with peak values appearing in West Africa as well
as in Chad and Sudan. Figure 3.5 shows values of the index at the grid-cell
level in Africa. he mean value is 8.4 and the standard deviation is 9.7.
he current geographic distribution of malaria impact may not accurately
represent the historical burden of the disease. Packard (2007) gives numer-
ous examples of how human activity has changed the intensity of malaria.
Clearing forests and introducing irrigation oten produces the marshy envi-
ronment needed for breeding of Anopheles mosquitoes. When a small num-
ber of landowners consolidate and control large tracts of land, people living
104 Weil

on that land are oten deprived of the tools for properly tending to it and
for avoiding malaria infection. his problem is exacerbated when seasonal
labor is used, as migrants may come from non-malarious zones and lack
immunity, then return to these zones carrying malaria with them. As for
Africa, he claims that “while climate and the presence of highly eicient
vectors contributed to the persistence of malaria, conditions of produc-
tion played an equal if not greater role.” Webb (2008) similarly describes
how the introduction of agriculture led to malaria endemicity in lowlands
throughout the Eurasian landmass. Population density is also important
to the nature of malaria burden. Packard argues that in the African con-
texts, low population densities associated with mobile hunter-gatherer
populations would have been unable to support the P. Falciparum infection
because of the speed with which the disease exits or kills the human host.
he emergence of Anopheles gambiae is also believed to be closely associ-
ated with the development of agriculture. For all these reasons, the current
distribution of malaria in Africa is potentially a poor proxy for the histori-
cal distribution.
Several mutations have arisen in human populations that provide resis-
tance to malaria. hese include the mutation causing thalassemia, which
is present in Mediterranean, Arab, and Asian populations; the absence of
the Dufy blood group in West Africa (which speciically blocks P. Vivax);
hemoglobin E in Southeast Asia; and hemoglobin C in West Africa (Allison
2002; Nelson and Williams 2006). he most important such mutation is the
one that causes sickle cell disease.
he sickle cell trait is a mutation in the gene that produces hemoglobin,
the oxygen-carrying component in red blood cells. Individuals carry two
copies of this gene, one received from each parent. Individuals who carry
one normal copy of the gene (referred to as A type allele) and one copy with
the sickle cell mutation (S type allele) are carriers of the disease. In indi-
viduals of the AS genotype, a fraction of the hemoglobin in their red blood
cells has an abnormal structure. In individuals who have two copies of the
sickle cell allele (SS genotype), almost all hemoglobin molecules are of the
abnormal type.
In conditions of inadequate oxygen supply (hypoxia), hemoglobin pro-
duced by the S gene becomes rigid, leading to a characteristic sickle shape of
red blood cells. Carriers of the sickle cell trait generally do not sufer many
adverse efects.3 However, negative consequences can arise from sickling

3
Williams and colleagues (2005) show the absence of any signiicant efect of carrier status
on a wide range of childhood diseases.
 he Impact of Malaria 105

in conditions of low oxygen such as unpressurized airplane lights and

extremely rigorous exercise (Motulsky 1964). In individuals of the SS geno-
type, such sickling of red blood cells is far more common, leading to acute
episodes of disease in which abnormally shaped cells restrict blood low
to organs, cause severe damage to the spleen, and trigger anemia. In 1994,
life expectancy for SS children in the United States was forty-two years for
males and forty-eight years for females. In the absence of modern medical
care, individuals of the SS genotype are not able to survive to adulthood.
he sickle cell mutation is relevant to malaria because infection of a red
blood cell with the malaria parasite leads to hypoxia. In individuals of the
AS genotype, such blood cells sickle and are then eliminated by the body’s
immune system, lessening the burden of infection. Carriers of the sickle
cell trait are particularly resistant to severe malarial episodes; they are less
resistant to mild cases. he mechanism by which AS carriers are protected
from malaria is diferent than the acquired immunity that both AA and AS
individuals achieve following repeated exposure to the disease.
he beneit that possessing a single copy of the sickle cell gene conveys
counterbalances the biological cost incurred when homozygous SS children
are stricken with sickle cell disease. An individual of the AS genotype is
more likely to reach adulthood than is an individual of the AA genotype,
but the former is also more likely to see his/her child die of sickle cell dis-
ease. his is known as a heterozygote advantage or balanced polymorphism.
he stronger the pressure of malaria is on survival, the more advantaged are
individuals who carry the S gene, and in equilibrium, the higher the per-
centage of the population who will be carriers.
Figure 3.6 shows data on the fraction of alleles that are of the S type
among native African populations, using data from Piel and colleagues
(2010). Because individuals have two alleles, the fraction of individuals who
are carriers of the sickle cell trait (that is, the AS genotype) is slightly less
than twice as large as the fraction of alleles that are of the S type.4 he igure
shows signiicant areas with an allele frequency of 15 percent or greater,
implying that 25 percent or more of the population carries the sickle cell
trait. According to Nelson and Williams (2006), the prevalence of AS hemo-
globin in tropical Africa ranges from 16–29 percent in adults. Prevalence
rates of up to 40 percent have been found in particular populations, for

4
More precisely, if a fraction ρ of alleles are of the S type, then assuming no diferential
mortality, the fraction of people who are carriers will be 2ρ 2ρ2. So, for example, 10 per-
cent of alleles being of the S type would correspond to 18 percent of individuals being car-
riers. To the extent that mortality is higher for individuals of the SS genotype, the fraction
of carriers in the population will be even closer to exactly 2ρ.
106 Weil

Prevalence of Sickle Cell Gene

High : 0.18

Low : 0

Figure 3.6. Frequency of the S Allele in Africa.

example in western Uganda and Tanzania south of Lake Victoria (Allison

2002). Figure 3.6 also shows a striking correlation of sickle cell prevalence
with malaria ecology. Indeed, it was the correlation of high prevalence of
the sickle cell gene and the presence of malaria that irst led scientists to
understand the protective role of the sickle cell mutation.
It is of interest to compare malaria ecology and sickle cell prevalence, on
the one hand, to the indicators of early development discussed in Section
3.1, on the other. Notably, one of the regions with the highest level of early
development, West Africa and the Niger Bend, also had one of the high-
est burdens of malaria, although this is not true of Ethiopia, which was
the other locus of early development. he lack of a correlation between
malaria burden and early development might suggest that malaria was not
important in holding back development. However, such correlations must
be interpreted cautiously, because many other factors that inluence devel-
opment are omitted.
 he Impact of Malaria 107

3.3 MEASURING THE HISTORICAL BURDEN OF MALARIA

USING DATA ON THE SICKLE CELL TRAIT
A related paper, Depetris-Chauvin and Weil (2013), presents a mathemat-
ical analysis of the historical impact of malaria as derived from data on
sickle cell prevalence. he underlying logic, as described previously, is that
the sickle cell trait is a double-edged sword: it protects against malaria, but
raises the probability that a child will die from sickle cell disease. he frac-
tion of children who will die from sickle cell disease depends on the frac-
tion of adults who are carriers. he beneit from being a carrier depends on
how severe the efect of malaria on noncarriers is and how much protection
is ofered by being a carrier. In equilibrium, the percentage of the adult
population who are carriers of the sickle cell trait will depend on the sever-
ity of malaria in a region: the worse malaria is (in terms of the fraction of
unprotected children who will die of it), the higher the fraction of adults
who will be carriers of the sickle cell trait.
he bottom line measure of malaria’s impact that I use is the relative sur-
vival into adulthood of carriers versus noncarriers. his relects purely the
protective efect of the sickle cell gene, because the available evidence is that
carrier status has at most a trivial efect on survival from conditions other
than malaria. Call this relative survival, which will be a number between
zero and one, β. he lower β is, the more valuable the protective efect of
carrying the sickle cell trait is. Depetris-Chauvin and Weil (2013) show how
the value of β can be derived from the observed prevalence of the sickle cell
trait in adult populations, which I label as π (which also ranges between
zero and one). he value of β determines the equilibrium level of π. hus,
if one can observe the equilibrium level of π, one can back out how bad
the selective pressure from malaria is, that is, β.5 he second column of
Table 3.4 shows the implied value of β for diferent values of π, which are
shown in the irst column. I consider values of adult prevalence ranging
from zero to 40 percent, which is the highest level observed among speciic
populations. A prevalence of 20 percent, for example, implies that noncar-
riers are 89 percent as likely to live to adulthood as carriers.
One complication in using data on the frequency of the sickle cell trait
to assess the historical burden of the disease is that most measures of the
prevalence of sickle cell come from the past sixty years. By contrast, my
basic analysis considers a population that is in equilibrium in terms of the

⎛ β⎞
5
he key equation is π = (1 − β ) 1−
⎝ 2⎠
108 Weil

Table 3.4. Calculations based on prevalence of sickle cell trait

Fraction of Survival of Probability of Total Probability Malaria

adults who noncarriers dying of malaria death rate of dying and SC
are carriers relative to or sickle cell (preadult) of malaria deaths as
(π) carriers conditional on or SC fraction
(β) not dying of (preadult) of total
something else (preadult)
0 1.000 0.000 0.444 0.000 0.000
0.05 .974 0.025 0.458 0.014 0.032
0.10 .947 0.050 0.472 0.030 0.064
0.15 .919 0.075 0.486 0.047 0.096
0.20 .889 0.100 0.500 0.064 0.128
0.25 .857 0.125 0.514 0.082 0.159
0.30 .824 0.150 0.527 0.100 0.190
0.35 .788 0.175 0.541 0.119 0.220
0.40 .750 0.200 0.555 0.139 0.250

selective impact of malaria and the prevalence of the sickle cell trait. Such
a steady state presumably existed in Africa in the period before European
contact. One might worry that modern prevalence rates are not the same
as those that held historically, because the health environment has been
changing over time. In practice, there is little reason to believe that contact
with Europeans did anything to reduce the impact of malaria in Africa
until the second half of the twentieth century, in which case prevalence
rates would not have had much time to change before being measured.
Depetris-Chauvin and Weil (2013) use the case of African Americans as
a check on the basic model of selective pressure on the sickle cell trait.
Slaves transported to what is now the United States found themselves in
an environment in which the selective pressure of malaria was far lower.
One would thus expect the prevalence of the sickle cell trait to fall over
time. We show that the current prevalence of the trait among African
Americans (approximately 8 percent) is consistent with sickle cell preva-
lence in the known source countries of the slaves who came to the United
States and a time span of roughly ten generations since the average ances-
tor of current African Americans was removed from a highly malarious
environment.
In calculating the burden of malaria and the protective efect of the sickle
cell trait, one must also consider the cost from children who die from sickle
cell disease. For example, if the prevalence of the trait among the adult
 he Impact of Malaria 109

population is 20 percent, then one in twenty-ive children will have both

parents be carriers, and of these children one in four will get two copies
of the sickle cell trait – thus one percent of children will die from sickle
cell disease before adulthood (if they do not die from something else irst).
More generally, if π is the fraction of adults who are carriers, then the frac-
tion of children who sufer from sickle cell disease will be π 2 / 4 . In the
calculations presented later in this chapter, I measure the burden of malaria
as constituting both deaths directly from the disease as well as deaths from
sickle cell disease.
To turn estimates of relative survival of carriers into a measure of the over-
all burden of malaria, one needs a further piece of information. To see why,
note that relative survival of 80 percent for noncarriers would be consistent
with 0 percent of carriers and 20 percent of noncarriers dying of malaria,
but it would also be consistent with 50 percent of carriers and 60 percent
of noncarriers dying of malaria. Depetris-Chauvin and Weil (2013) discuss
solving this problem by using data on the probability that carriers will die
of malaria relative to noncarriers. Available data are sketchy, and further
only apply to a modern health environment (where the probabilities of both
carriers and noncarriers dying of malaria are lower than they were histori-
cally). hese data show the relative death probability from malaria for car-
riers ranging from zero to 8 percent. In other words, the sickle cell trait is
almost completely protective against a fatal case of malaria. For the calcula-
tions, I use a value of zero for this relative probability.
Beyond knowing mortality from malaria and sickle cell disease, to paint
a complete picture of the burden of malaria, I also need to know the proba-
bility of dying from other diseases, or alternatively, the total rate of mortal-
ity. Part of the reason for this is that all of the previous calculations refer to
the probability of dying of malaria or sickle cell disease conditional on not
dying of something else. In a society in which there is more mortality from
other sources, malaria deaths will necessarily be lower.
Systematic data on life expectancy in Africa is widely available only start-
ing in the 1950s. For the period 1950–5, the United Nations estimate of
life expectancy at birth in sub-Saharan Africa is 37.8 years (United Nations
2009). Acemoglu and Johnson (2007) date the beginning of the “interna-
tional epidemiological transition,” driven by more efective public health
measures, the discovery of new chemicals and drugs, and international
interventions, to 1940. Although the transition came late to Africa, it is very
likely that the 1950–5 igure represents an improvement relative to previous
decades. Clearly ater 1955 the pace of change was rapid. he UN estimates
that life expectancy in sub-Saharan Africa rose by two years in each of the
110 Weil

subsequent ive-year periods. Further evidence that health improvements

were already under way by 1950 comes from data on total population size.
Africa’s population grew at a rate of 1.0 percent per year between 1900 and
1950, compared to a growth rate of 0.2 percent per year over the previous
century (United Nations 1999).
Data for the period prior to 1950 are very sparse. Acemoglu and Johnson
(2007, appendix C) pull together disparate sources to present a few esti-
mates for the period before 1950. hese are, Angola in 1940: 35 (both sexes);
Mozambique in 1940: 45 (both sexes); Ghana in 1948: 38 (both sexes); and
Mauritius, 1942–6: 32.3 (male) and 33.8 (female). Riley (2005) estimates
that prior to the “health transition” that began in Africa in the 1920s (he
uses a diferent deinition than Acemoglu and Johnson), life expectancy at
birth averaged 26.4 years.6 In Asia, life expectancy prior to the health transi-
tion, which started there between 1870 and 1890, was 27.4. In Europe, the
transition started in the 1770s, and prior to it life expectancy was 34.3.
Riley comments that available estimates of African mortality prior to the
health transition all came from European colonies in Africa. here is a rea-
sonable basis for thinking that life expectancy may have been higher prior
to colonization, the arrival of Arabic-speaking merchants, and the dislo-
cations produced by the slave trade. Unfortunately, little information for
this period is available. Steyn (2003) looks at mortality during the preco-
lonial period in northern South Africa through an examination of skeletal
remains. She estimates life expectancy in the period 1000–1300 CE at 23.2,
with the probability of surviving to age twenty being 48 percent. Remains
for the post-1830 period show a slight decline in life expectancy ater the
expansion of European inluence.
In the UN Model Life Tables for developing countries (1982), life expec-
tancy at birth of thirty-ive (general model, for females) implies that
61.4 percent of girls will survive to age twenty (and only 53.0 percent to age
30). he majority of premature deaths are concentrated at very low ages:
22.5 percent of girls would not reach age ive. Unfortunately, the model
life tables that are available, to the extent that they relect African data at
all, certainly do not relect the pattern of age-dependent mortality that
existed in the period before the modern health environment of both dis-
ease and treatment was in place. It is likely that the pattern of mortality,

6
his is the mean of twelve estimates, which range from 22.5 to 31.0. Some representative
values are Angola: 27 years in 1940; Egypt: 30–3 years in the 1930s; Ghana: 28 years in
1921; Kenya: 23.9 years in the 1930s; South African black population: 38.1–40 years in
1935–40; Tunisia: 28.8 in the 1920s; Uganda: 23.9 in the 1930s; Zimbabwe: 26.4 in the
mid-1930s.
 he Impact of Malaria 111

and in particular the ratio of deaths at diferent ages, difered from what is
observed today, but a priori there is no basis for guessing the nature of this
diference. Allowing for the fact that life expectancy at birth was almost
certainly lower than thirty-ive (the lowest value available in the UN model
tables), and that “childbearing age” is not a single number but a range, a rea-
sonable estimate for a value to use in our model is that 50 percent of births
did not reach childbearing age.
My interest is in inding the fraction of people who do not live to repro-
ductive age, which I call “preadult mortality.” I make the assumption that
the level of malaria mortality is uncorrelated with the level of non-malaria
mortality. I also make the assumption that preadult mortality from malaria
has the same approximate age proile as other sources of preadult mortality.
I benchmark the model by assuming that in areas with sickle cell prevalence
of 20 percent, total preadult mortality would be 50 percent. his allows
me to back out the probability of death from non-malaria causes (in the
case where an individual does not die of malaria), which turns out to be
44 percent.7
he last three columns of Table 3.4 use these assumptions about non-
malaria mortality to paint a picture of the burden of malaria. For each value
of sickle cell prevalence, π, I calculate the total rate of preadult mortality
as well as the fraction of preadult deaths that would be due to malaria and
sickle cell disease. he table shows that, going from an area where malaria
was absent (π = 0) to one of reasonably high prevalence (π = 0.2), the overall
preadult death rate rises from 44 percent to 50 percent. In the worst alicted
areas (π = 0.4), the probability of preadult death was 56 percent and fully
25 percent of deaths were attributable to malaria and sickle cell disease.

3.3.1 Comparison to Modern Malaria Mortality Rates

he death rate from malaria derived from the historical data can be com-
pared to modern rates. As mentioned earlier, for the WHO AFRO region,
the under-ive death rate from malaria is 0.59 percent per year. In the 2008
World Malaria Report (data for 2006), the ive most severely burdened
countries are Niger (1.1%), Mali (0.98%), Burkina Faso (0.96%), Nigeria
(0.88%), and Chad (0.87%). Multiplying these numbers by ive gives an
approximation of the probability of dying from malaria in the irst ive years
of life, which is very close to the probability of dying from malaria before
reproductive age. Although I cannot assign values of sickle cell prevalence

7
See Depetris-Chauvin and Weil (2013) for details.
112 Weil

(π) to particular countries, it appears that the malaria death rate in the
pre-European contact period was about twice as high.
A diferent and somewhat cleaner way to do the comparison of histori-
cal malaria mortality to mortality today is to focus on the probability of
dying of malaria (or sickle cell disease) conditional on not dying of some-
thing else. his is the measure constructed from historical data on preva-
lence, and although it is not generally reported, it can be constructed from
modern data as well. Consider the case of Nigeria, which is a very heavily
alicted country. Annual malaria deaths for children under ive are esti-
mated to be 0.88 percent. his implies that roughly 4.4 percent of chil-
dren will die of malaria before their ith birthday. he life table for Nigeria
for 2006 shows that the probability of a woman surviving to age twenty-
ive is approximately 0.75. From these two pieces of information, I can
back out the probability of dying from malaria conditional on not dying
of something else (see Depetris-Chauvin and Weil, 2013, for details). he
result is 5.4 percent. Measured this way, the burden of malaria is far lower
in Nigeria today than it was in countries with a sickle cell prevalence of
20 percent in the historical period, where, according to Table 3.4, the ig-
ure was 10 percent.
It is also of interest to compare the decline in malaria mortality to the
decline in mortality from other causes. In the previous calculation, the
probability of malaria death falls by slightly less than half. By contrast,
the probability of dying of some other cause (conditional on not dying of
malaria) falls from 44 percent to 20.7 percent – a decline of slightly more
than half. hus the probability of death from other causes has fallen slightly
more than the probability of death from malaria, although the magnitudes
of the two declines are close. Certainly, given the uncertainty in so many
aspects of measurement, one cannot reject the possibility that the percent-
age of decline in malaria mortality has been the same as the decline in other
types of mortality.

3.4 ASSESSING THE IMPORTANCE OF MALARIA TO

EARLY AFRICAN DEVELOPMENT
Now that I have an estimate of how large a burden malaria imposed in
terms of survival, I can turn to the question of whether malaria had a
signiicant impact on African development. I divide my discussion into
two parts, looking irst at the efects of malaria deaths (mortality) and
then at the efects of ill health among those who survived the disease
(morbidity).
 he Impact of Malaria 113

3.4.1 Direct Efect of Malaria Mortality

Current analyses of the burden of disease focus on measures such as years of
life lost or disability-free life years lost. From this perspective, the death of
a young child is particularly costly because he or she had so many potential
life years. he ethical considerations regarding the allocation of scarce life-
saving resources, and implicitly the cost of death and disease experienced at
diferent ages, are quite complex (see Persad, Werthheimer, and Emanuael
2009). In assessing the role that disease played in afecting development his-
torically, however, it seems reasonable to take a purely instrumental view of
life and health, in which the primary considerations are how much society
has invested in an individual and that individual’s potential to produce ser-
vices for society in the future. Under this sort of view, the most costly death
is that of a young adult, who has consumed a good deal of resources (food,
child care, education), and who has many years of potentially productive
labor. he death of an infant or small child, by contrast, is far less costly,
because fewer resources have been invested; and the death of an old person
is similarly less costly because less productive potential is lost. Under this
interpretation, malaria deaths are relatively low cost, with the exception of
deaths of women in their irst pregnancies, who are near the peak of their
value as assets to society in terms of the balance between resources invested
in them and services they can deliver.
To formalize this idea, consider a simple model of production and con-
sumption with individuals of diferent ages. Let ci be the consumption of an
individual of age i, and similarly wi be labor income. I ignore income from
nonlabor sources, and assume that there is no storage of output between
periods. Let Ni be the number of people in age group i. he social budget
constraint is
T

∑ N (w
i i ci ) = 0
0

where T is the maximum lifespan. I assume that consumption at each age

is determined by two things: a consumption level of individuals at some
benchmark age (for example, prime age adults), which I call c , and some
age-varying relative consumption coeicient c#i .

ci cc#i

he values of c#i presumably relect both changing biological needs for con-
sumption over the course of the life cycle as well as the arrangements by
114 Weil

which consumption is divided up among diferent groups in society. One

would not necessarily expect the pattern of consumption to be the same in
all societies at all times.
I assume that wages (or output produced via labor) at each age group are
determined in a similar fashion:

wi ww i

Where w is the return to some standard unit of labor. Again, the pattern
of relative labor input across age groups relects both biological difer-
ences and diferences among societies in the economic value of difer-
ent characteristics (for example, strength versus wisdom versus manual
dexterity).
Combining these equations, we can solve for the consumption level of
the benchmark age group as a function of the age structure of the popula-
tion as well as the relative wage and consumption weights.

∑N w i i
0
c w T

∑ N c#i i
0

3.4.1.1 Consumption and Income Proiles

A number of sources provide data on the life cycle proiles of consump-
tion and labor input. Mueller (1976) synthesizes data from nine socie-
ties practicing what she calls “peasant agriculture,” by which she means
agricultural systems that use primarily traditional methods of cultiva-
tion, small landholdings, and low capital inputs. Her data are primarily
from South and Southeast Asia, from the middle of the twentieth century,
but they may not be a bad approximation to the most developed parts of
Africa in the periods of interest in this chapter. he proiles are shown in
Figure 3.7. Note that Mueller’s data on labor input apply to production
of output as it would appear in measured GDP but exclude home pro-
duction. Much of the latter is done by women, so in her data, productiv-
ity by prime age women is only 30 percent of the level of similarly aged
males. Because I am interested in both sorts of output, I focus on the
male proiles. For consumption, Mueller provides two proiles in which
consumption of people at each age (and of each gender) is compared to
males aged twenty to ity-four. For the medium consumption proile,
 he Impact of Malaria 115

children aged zero to four have a value of 0.32; for the low consumption
proile, the value is 0.12 (prime aged women get a value of 0.80). I use the
medium proile. Lee and Mason (2011) show very similar-looking proiles
for hunter-gatherer populations.

3.4.1.2 Population Age Structure

he other piece of data used in this equation is Ni, the number of people
in each age group. In general, this will be a function of both the probabil-
ity of survival to each age and the history of births or population growth.
However, for the long historical periods that I am considering, popula-
tion growth must have been very close to zero, which implies a constant
number of births per year. We can thus approximate the age structure
of the population with the fraction of survivors at each age from the life
table.
My approach to assessing the role of disease in afecting consumption
possibilities is to start with a baseline life table and then consider alterna-
tions that would result from eliminating or adding particular sources of
mortality. To construct these alternative life tables, I use information on
the age of death from diferent diseases. As my baseline, I use the United
Nations (1982) model life table for a population with life expectancy at
birth of thirty-ive years (male and female combined).
To apply this framework to diferent diseases, one needs a proile of the
mortality efect of the disease at diferent ages. Table 3.3 shows such data
for malaria in Africa today. To model the efect of malaria on the historical
life table, I proceed as follows. I start with the UN life table (life expectancy
of thirty-ive) discussed earlier. I take this as a benchmark. To this I add
additional mortality at each age proportional to the current age proile of
malaria mortality. Speciically, I take the current death rates from malaria
(0.00559 for ages zero to four, 0.00042 for ages ive to fourteen, etc.) and
multiply them by a scaling factor, and add these death rates to the death
rates in the UN life table. he scaling factor is chosen to match the magni-
tude of the change in malaria deaths I want to model. Speciically, I showed
previously that for the case of going from zero malaria to the case where
sickle cell prevalence is moderately high (π=.2), the total preadult death
rate rose from 44 percent to 50 percent. In the UN data, the probability of
living to age twenty-eight is .555, which is close enough to a 44 percent pre-
adult death rate. (In the model “adult” is the age at which children are pro-
duced. In reality, there is no single age that is appropriate, but twenty-eight
is roughly in the middle of the relevant range.) he scaling factor required
116 Weil

Consumption and Production in Peasant Agriculture (Males)

2.5

1.5

Production
1 Consumption

0.5

0
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
Age

Figure 3.7. Production and consumption over the life-cycle.

to set survival to age twenty-eight equal to 50 percent is two.8 Life expec-

tancy falls by three and a half years from this additional malaria mortality.
his increase in malaria mortality lowers steady state consumption by
only 0.63 percent. his is obviously an extremely small efect for such a
large change in mortality. he reason is clear once one has seen Figure 3.7
and noted that malaria deaths are concentrated at very low ages. In this cal-
culation, deaths beyond age ive account for only one-third of the reduction
in life expectancy due to malaria, but for two-thirds of the economic cost
of the disease.
his analysis shows that, at least in terms of the extra costs associated
with raising children who were subsequently going to die of malaria, the
efect of the disease on the standard of living in Africa should have been
relatively low. However, it is possible that there are other channels through
which high malaria mortality may have mattered. Death from disease may
impose an obstacle to economic development beyond the instrumental cost
discussed here. Stone (1977) argues that in the face of high child mortal-
ity, premodern Europeans avoided forming emotional bonds with their

8
Of course the UN table already includes deaths due to malaria. Another approach would
be to take as the baseline a current life table stripped of malaria mortality, and then to add
malaria mortality back in. his procedure would yield very similar results.
 he Impact of Malaria 117

Population
Size
Equilibrium population size
with healthy environment
Equilibrium population size
with unhealthy environment

Population
Growth
Rate Healthy Environment

Unhealthy Environment

Standard of Living
Equilibrium Equilibrium
standard of standard of
living with living
healthy with unhealthy
environment environment
Figure 3.8. he Malthusian model.

young children, even to the point of giving the same name to two living
children, expecting only one to survive. In such an environment, one might
also expect to see less investment of tangible resources in the human cap-
ital (both health and education) of their children. hus even if high child
mortality from malaria were not important in the strict sense of draining
resources from the economy, it might have contributed to reduced invest-
ment in children. Several recent theories of economic growth have stressed
the importance of such investments to long-term development.

3.4.1.3 he Malthusian Perspective

his analysis takes the perspective that deaths are costly in social terms,
because they represent the loss of potential productive workers. An alterna-
tive view points to a lip side of deaths, which is that they decrease the size
of population and thus lower pressure on ixed resources.
Consider a simple Malthusian model, as represented in the two panels
of Figure 3.8. In both panels, the horizontal axis measures the standard of
living. In a modern economy, this would be GDP per capita, but the model
should be applicable even in contexts where there are no markets or prices
118 Weil

with which to measure GDP. In the top panel, the vertical axis measures the
size of the population. he downward sloping line in the panel represents
the efect of higher population size in reducing living standards through the
crowding of ixed resources such as land. his crowding could take the form
of shorter fallow periods, higher pressure on grazing land, or increased
levels of animal disease, all of which would reduce per capita yields.
he bottom panel shows the relationship between the growth rate of pop-
ulation and the standard of living. I assume that there is an upward sloping
relationship between these. In the classical Malthusian model, this positive
relationship comes from either the “preventive check” (reduced desired fer-
tility) or the “positive check” (higher mortality or lower fecundity) when
income is low. Another mechanism that will produce very similar dynam-
ics is if higher population density raises the mortality rate by facilitating
the transmission of disease. In the panel, this upward sloping line is drawn
twice, at a higher level to represent a place where the health environment is
good, and at a lower level to represent a place where the health environment
is bad. For a given standard of living, population growth will be higher in a
healthier environment.
One of the deining characteristics of the Malthusian model is that it pro-
duces a steady state in which, as long as the quantity of land is constant and
the production technology is ixed, population size is constant (or, more
realistically, the population size and standard of living are mean reverting
in response to stochastic shocks). he near constancy of human popula-
tions over long historical eras suggests that such a mechanism must have
been operating in most places for most of human history (Galor and Weil
2000). Although there are clearly times when Malthusian constraints were
lited (such as the peopling of the Americas in prehistoric times), a simple
analysis of the efect of compounding shows that most of the time, aver-
age population growth rates must have been quite close to zero. Without
a homeostatic model of the Malthusian sort, it is vanishingly unlikely that
such near constancy of population would obtain.
As the igure shows, a worse health environment leads to lower popu-
lation and a higher standard of living in equilibrium. hus malaria would
have actually raised the standard of living in Africa. While counterintui-
tive, this conclusion is in line with the arguments in several recent anal-
yses of growth. Voigtlander and Voth (2009) argue that in Europe before
the Industrial Revolution, income rose as a result of increased mortality
due to plague, urbanization, and warfare, and that this rise in income was
instrumental in knocking the continent onto the path of industrialization.
Acemoglu and Johnson, in their analysis of the worldwide epidemiological
 he Impact of Malaria 119

transition of the mid-twentieth century, ind that countries that experi-

enced greater increases in life expectancy saw slower growth in income per
capita – a result that they attribute to the Malthusian efect of population
growth. And Young (2005) claims that higher mortality and lower fertil-
ity due to HIV in South Africa will more than compensate for declines in
worker productivity due to the disease, so that income per capita will actu-
ally rise as a result of the epidemic. Ashraf, Lester, and Weil (2008) examine
the Malthusian channel as part of a broader analysis of how health improve-
ments afect growth. hey ind that higher survival indeed pushes income
lower through the Malthusian channel (as well as through capital dilution),
although this efect is more than compensated for in the long run by higher
labor productivity due to better health.
Whether this sort of Malthusian model applies to Africa in the period
before European contact is unclear. Several authors stress that land shortage
was not generally a problem in Africa. For example, Herbst (2000) argues
that the abundant land was a persistent characteristic of African economies,
as a result of which it was the control of people, rather than territory, that
was of primary interest to rulers.
However, historians are oten too quick to dismiss the Malthusian model
without having another explanation for the near constancy of popula-
tion over time. It is conceivable that, rather than land being the binding
resource, the role of equilibrating population size was played by density-
dependent disease. his would be consistent with the oten-expressed view
that labor, rather than land, was the binding constraint African economies
faced. If so, the pressure from disease must have been extraordinarily high
indeed, however, because human capacity to reproduce is very high. he
levels of malaria indicated by the prevalence of the sickle cell trait would
not be suicient to hold down population growth in the absence of a land
constraint, although it is possible that malaria in combination with other
diseases did so.
It is also possible that malaria did indeed raise the standard of living in
the short run by reducing population density, but at the same time low pop-
ulation density (due both to malaria and low agricultural productivity) pre-
vented the development of large agglomerations that in other places were
the loci of specialization, gains from trade, and technological advance.

3.4.2 Economic Efects of Malaria Morbidity

To pursue the question of how much malaria afects labor input of adults,
I use data from the World Health Organization on Years Lost to Disability
120 Weil

(YLDs) from particular diseases. A country’s YLD for a given disease is

constructed as:
YLD = I × DW × L
where I is the number of incident (newly arising) cases in a period, DW is
the disability weight attached to the disease, and L is the average duration of
the disease until remission or death. he crucial parameter here is the dis-
ability weight, which is intended to be a cardinal measure of the severity of
diferent diseases or impairments, on a scale from zero, indicating perfect
health, to one, indicating death. Disability weights are constructed by pan-
els of healthcare providers and medical experts using a “person trade-of ”
protocol, which establishes utility equivalences between years of life lived
in diferent states of health. One year lived with a disability provides the
same utility as (1-DW) years lived disability free (Murray 1996). Disability
weights are therefore not primarily intended as a measure of labor supply.
Nevertheless, these estimates provide at least some basis for comparing the
efects of diferent diseases.9
To give an example of the interpretation of YLD: in the data presented
later, the average YLD from all causes for Africa is 0.12. his means that
in the average year, the average man sufers disease episodes that deprive
him of the equivalent of 12 percent of a year’s disability-free life. his could
mean being fully disabled for 12 percent of the year, 50 percent disabled for
24 percent of the year, and so on. It could also mean sufering an incident
that leaves him, say, 1 percent disabled for the next twelve years.
Table 3.5 shows the YLDs by disease type and age (for males) for the
WHO’s AFRO region (Mathers, Fat, and Boerma 2008). YLDs are counted
at the age in which a disease incident occurs. hus, for example, the neuro-
logical sequelae of malaria are counted as years lost to disability in the age
zero to four age group, even though the actual years lost are spread through
the individual’s whole life. (Note that the “total” column is based on the
distribution of the current population among age groups. hus disabilities
afecting the elderly, who are a very small percentage of the current African
population, play a very minor role in determining the total igure.)
he table shows that overall, malaria accounts for only 5 percent of total
years lost to disability. Further, almost all of these lost years were due to inci-
dents in the irst ive years of life. Some of the years lost to disability in this

9
Some examples of disability weights are blindness (0.600), deafness (0.216), HIV (0.136),
AIDS (0.505), tuberculosis sero-negative for HIV (0.264), severe iron-deiciency anemia
(0.093), malaria episodes (0.172), and neurological sequelae of malaria (0.473).
 Table 3.5. Years lost to disability per capita, WHO AFRO region

0–4 5–14 15–29 30–44 45–59 60–69 70–79 80+ total

All Causes 0.174 0.068 0.133 0.134 0.141 0.126 0.115 0.103 0.122
Communicable, maternal, 0.112 0.028 0.040 0.037 0.021 0.015 0.013 0.010 0.046
perinatal, and nutritional
conditions
Infectious and parasitic 0.047 0.019 0.037 0.035 0.021 0.014 0.012 0.010 0.031
diseases
Tuberculosis 0.000 0.001 0.002 0.004 0.002 0.002 0.001 0.001 0.002
121

HIV/AIDS 0.000 0.000 0.017 0.015 0.004 0.001 0.000 0.000 0.007
Diarrheal diseases 0.004 0.002 0.001 0.001 0.000 0.000 0.000 0.000 0.001
Childhood cluster diseases 0.007 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001
Malaria 0.027 0.001 0.002 0.001 0.001 0.001 0.001 0.001 0.006
Tropical cluster diseases 0.001 0.009 0.011 0.010 0.009 0.004 0.003 0.003 0.008
Noncommunicable Diseases 0.047 0.019 0.066 0.074 0.108 0.105 0.098 0.089 0.056
Injuries 0.015 0.021 0.027 0.023 0.011 0.006 0.004 0.003 0.021
Source: Global Burden of Disease, 2002 revision.
122 Weil

case afected adults, through the neurological sequelae of the disease, but
much of the disability burden fell directly on children. Although disability
(which in this case really just measures sufering) among children does not
directly afect production, one could argue that it might have afected their
accumulation of skills or human capital. In any case, however, the malaria
burden is relatively small in comparison to that of other diseases.
As mentioned previously, another potential channel through which
malaria might have afected economic development is through shiting
population away from potentially productive areas. Gallup and Sachs give
examples of this occurring in Europe. In the context of Africa, I do not know
of evidence that particular regions were not settled because of malaria, and
there is certainly evidence of people settling in areas of tremendously high
malaria transmission. Given that malaria deaths were concentrated among
the very young, and that mortality in this group was very high from other
causes in any case, it is hard to see how malaria would have had a great
inluence on settlement patterns.

3.5 CONCLUSION
In this chapter, I set out to assess the role that malaria played in afecting
African development in the period before Africans and Europeans came
into contact. he largest contribution of this chapter is in assessing the
health burden that malaria placed on Africans some ive hundred years ago.
I used data on the modern prevalence of the sickle cell mutation to estimate
the intensity of malaria historically. In areas of high malaria transmission,
20 percent of the population carries the sickle cell trait. My estimate is that
this implies that historically 10 percent of children died from malaria or
sickle cell disease before reaching adulthood. Such a death rate is roughly
twice the current burden of malaria in such regions. Comparing the most
afected to least afected areas, malaria may have been responsible for a ten
percentage point diference in the probability of surviving to adulthood.
In terms of its burden relative to other causes of mortality, malaria appears
to have been perhaps slightly less important historically than it is today,
although I certainly can’t rule out the decline in malaria mortality has been
proportional to the decline in mortality from other diseases.
hus malaria imposed a heavy mortality burden. Did it hold back eco-
nomic development? As a starting point, it is worth noting that many
 he Impact of Malaria 123

of the areas with highest malaria ecology and highest prevalence of the
sickle cell trait, most notably coastal West Africa, also had the highest
levels of economic development by the various measures discussed in
Section 3.1. Obviously the lack of correlation between malaria and low
development is not irm evidence for a lack of a causal link. hose parts
of Africa could have been well developed for other reasons, and were
it not for malaria might have been more developed still. Nonetheless,
simply looking at the locations of development and malaria suggests
that malaria was not an insuperable obstacle to economic development
within Africa.
My more detailed examination of the possible channels through which
malaria might have afected economic outcomes also does not point to a
strong role of the disease in holding back economic development. Malaria
mortality occurs at very young ages, which implies a low social cost. In a
Malthusian model, higher child mortality could even raise the standard of
living. he morbidity burden of malaria on adults is low relative to other
diseases. he fact that malaria bulked so large in the view of foreigners
may be attributable to their lack of immunity, which Africans acquired as
children.
My indings leave open several possibilities regarding the relationship
between malaria and economic development in Africa. First, there may
be some channel by which malaria afected precolonial development that
is let out of my analysis. One possibility is that the overall disease envi-
ronment, of which malaria was only a part, held back economic growth
on the continent. If malaria was not, in fact, an important determinant
of development before the arrival of Europeans, then could it have been
important aterward? he analysis of Ashraf, Lester, and Weil (2008)
suggests that the direct efects of malaria on productivity are not large
enough to explain much of Africa’s economic backwardness. However,
it is possible that malaria held back growth through other channels dur-
ing the colonial period. For example, Acemoglu, Johnson, and Robinson
(2001) have famously suggested that malaria and other diseases had a
negative efect on development in the tropics through their role in shap-
ing the structure of institutions that Europeans established. Finally,
of course, malaria may simply not be a part of the story of African
underdevelopment.
124 Weil

APPENDIX

Country Population State Average Malaria Average

Density in Antiquity Technology Ecology Land
1500 in 1500 in 1500 Quality
Afghanistan 3.08 0.63 0.62 0.03 0.23
Albania 6.67 0.01 0.81
Algeria 0.64 0.60 0.78 0.01 0.06
Angola 1.50 0.13 0.17 4.37 0.35
Antigua and Barbuda 1.25
Argentina 0.11 0.01 0.03 0.07 0.45
Armenia 0.01 0.74
Aruba 1.25
Australia 0.03 0.00 0.00 0.06 0.22
Austria 25.00 0.71 0.90 0.00 0.40
Azerbaijan 0.03 0.74
Bahamas 1.25 0.00 0.49
Bangladesh 23.70 0.42 0.63 6.11 0.82
Barbados 1.25 0.00
Belarus 0.00 0.38
Belgium 41.67 0.68 0.90 0.00 0.59
Belize 1.54 0.23 0.71 0.76
Benin 4.23 0.00 0.17 17.46 0.49
Bermuda 1.25
Bhutan 0.11 0.25
Bolivia 0.82 0.70 0.16 1.21 0.47
Bosnia and 0.83 0.00 0.64
Herzegovina
Botswana 0.14 0.21 0.10 5.77 0.10
Brazil 0.12 0.00 0.13 1.93 0.36
Brunei Darussalam 3.25 0.38
Bulgaria 7.27 0.01 0.87
Burkina Faso 4.23 0.05 0.51 31.55 0.29
Burundi 25.00 0.00 1.57 0.52
Cambodia 8.33 0.75 8.57 0.58
Cameroon 1.50 0.24 0.17 13.18 0.39
Canada 0.02 0.00 0.13 0.00 0.06
Cape Verde Islands 0.50 0.02 3.04 0.01
Cayman Islands 1.25 0.00
Central African 1.50 0.00 0.23 17.99 0.36
Republic
Chad 0.60 0.18 0.40 21.46 0.15
Chile 0.79 0.02 0.16 0.00 0.14
China 11.46 0.91 0.88 0.30 0.35
Colombia 0.88 0.00 0.13 1.16 0.46
 he Impact of Malaria 125

Country Population State Average Malaria Average

Density in Antiquity Technology Ecology Land
1500 in 1500 in 1500 Quality
Comoros 5.00 4.77 0.37
Congo, Dem Rep 1.50 0.23 12.06 0.31
Congo, Rep 1.50 0.05 0.23 12.92 0.27
Costa Rica 1.54 0.00 0.13 0.04 0.63
Cote d’Ivoire 4.23 0.06 0.17 19.70 0.40
Croatia 0.03 0.85
Cuba 1.25 0.13 0.14 0.86
Cyprus 22.22 0.62 0.06 0.73
Czech Republic 0.83 0.00 0.55
Denmark 15.00 0.61 0.90 0.00 0.48
Djibouti 10.97 0.02
Dominica 1.25 0.13
Dominican Republic 1.25 0.00 0.93
Ecuador 2.14 0.09 0.16 0.20 0.54
Egypt 4.00 0.76 0.78 1.73 0.00
El Salvador 1.54 0.00 0.23 1.05 0.90
Equatorial Guinea 1.50 0.17 12.96 0.32
Eritrea 2.00 4.71 0.05
Estonia 0.00 0.33
Ethiopia 1.67 1.00 0.53 2.03 0.50
Finland 0.29 0.14 0.69 0.00 0.04
France 27.27 0.74 0.93 0.01 0.74
French Guiana 1.91 0.54
Gabon 1.50 0.00 0.17 13.03 0.36
Gambia 4.23 0.00 15.56 0.53
Georgia 0.01 0.64
Germany 25.00 0.76 0.90 0.00 0.55
Ghana 4.23 0.21 0.38 23.33 0.43
Greece 7.69 0.52 0.93 0.02 0.81
Grenada 1.25
Guadeloupe 1.50
Guatemala 1.54 0.42 0.23 0.70 0.89
Guinea 4.23 0.09 0.40 18.68 0.60
Guinea Bissau 4.23 0.27 10.35 0.58
Guyana 0.21 0.00 0.13 3.00 0.54
Haiti 1.25 0.00 0.16 0.93
Honduras 1.54 0.13 0.23 1.04 0.72
Hong Kong 0.09 0.91 0.88 2.71
Hungary 13.89 0.90 0.02 0.94
Iceland 0.60 0.21 0.00 0.00
India 23.70 0.69 0.76 0.11 0.59
Indonesia 5.17 0.55 0.66 3.18 0.41

(continued)
126 Weil

Country Population State Average Malaria Average

Density in Antiquity Technology Ecology Land
1500 in 1500 in 1500 Quality
Iran 2.42 0.68 0.81 0.06 0.15
Iraq 2.27 0.78 0.16 0.13
Ireland 10.00 0.55 0.80 0.00 0.34
Israel 0.50 0.03
Italy 33.33 0.70 0.90 0.07 0.76
Jamaica 1.25 0.00 0.15 0.50
Japan 45.95 0.83 0.83 0.00 0.58
Jordan 2.22 0.51 0.07 0.04
Kazakhstan 0.01 0.60
Kenya 2.63 0.00 0.23 4.80 0.35
Korea, Rep 18.18 0.90 0.85 0.00 0.77
Kuwait 0.23 0.00
Kyrgyzstan 0.01 0.26
Laos 1.67 0.58 0.75 2.63 0.43
Lebanon 7.50 0.01 0.90
Lesotho 0.47 0.00 0.17 0.00 0.69
Liberia 4.23 0.17 17.34 0.38
Libyan Arab 0.28 0.78 0.02 0.01
Jamahiriya
Lithuania 0.85 0.00 0.44
Madagascar 1.19 0.00 0.33 5.03 0.50
Malawi 0.79 0.02 6.25 0.77
Malaysia 1.18 0.59 0.72 1.69 0.40
Mali 0.60 0.38 0.51 30.10 0.11
Malta 0.90
Mauritania 0.60 0.31 0.17 4.17 0.01
Mauritius 0.00 6.82 0.00
Mexico 2.50 0.53 0.26 0.45 0.58
Moldova 0.01 0.97
Mongolia 0.38 0.41 0.00 0.08
Morocco 2.50 0.82 0.41 0.06 0.36
Mozambique 1.28 0.00 10.54 0.44
Myanmar 5.88 0.84 0.75 5.87 0.55
Namibia 0.14 0.10 3.72 0.07
Nepal 14.29 0.84 0.30 0.02 0.45
Netherlands 30.00 0.63 0.90 0.00 0.55
Netherlands Antilles 1.25
New Zealand 0.37 0.00 0.13 0.00 0.32
Nicaragua 1.54 0.00 0.13 0.80 0.59
Niger 0.60 0.17 0.44 21.93 0.01
Nigeria 4.23 0.45 0.47 21.62 0.42
Norway 1.25 0.53 0.90 0.00 0.02
 he Impact of Malaria 127

Country Population State Average Malaria Average

Density in Antiquity Technology Ecology Land
1500 in 1500 in 1500 Quality
Oman 0.95 0.01 0.00
Pakistan 23.70 0.81 0.73 0.05 0.10
Panama 1.54 0.00 0.13 0.17 0.47
Papua New Guinea 1.53 0.00 0.13 9.57 0.51
Paraguay 0.49 0.00 0.00 2.33 0.61
Peru 1.55 0.66 0.16 0.54 0.36
Philippines 1.67 0.00 0.58 1.42 0.64
Poland 12.90 0.85 0.00 0.57
Portugal 13.89 0.72 0.97 0.05 0.62
Puerto Rico 1.25 0.27
Qatar 0.01 0.00
Romania 8.33 0.70 0.00 0.74
Russia 0.80 0.80 0.01 0.12
Rwanda 25.00 0.00 0.46 0.48
Sao Tome and 1.50 8.42 0.23
Principe
Saudi Arabia 0.48 0.00 0.00
Senegal 4.23 0.40 0.33 15.34 0.25
Serbia 0.83
Sierra Leone 4.23 0.17 16.72 0.45
Singapore 1.18 0.13 0.72 11.03
Slovakia 0.00 0.62
Slovenia 0.01 0.58
Solomon Islands 4.86 0.01
Somalia 1.25 8.36 0.04
South Africa 0.47 0.00 0.17 0.97 0.28
Spain 13.00 0.56 1.00 0.10 0.73
Sri Lanka 15.15 0.90 0.04 0.77
St Kitts and Nevis 1.25
St Lucia 1.25
St Vincent and the 1.25
Grenadines
Sudan 1.60 0.38 19.12 0.27
Suriname 0.21 1.47 0.47
Swaziland 0.47 0.00 3.09 0.63
Sweden 1.78 0.31 0.90 0.00 0.07
Switzerland 20.00 0.68 0.83 0.00 0.34
Syria 7.50 0.56 0.70 0.06 0.32
Tajikistan 0.02 0.31
Tanzania 1.97 0.23 7.62 0.64
hailand 3.92 0.53 0.75 6.30 0.48
Togo 4.23 0.00 21.14 0.57

(continued)
128 Weil

Country Population State Average Malaria Average

Density in Antiquity Technology Ecology Land
1500 in 1500 in 1500 Quality
Trinidad and Tobago 1.25 0.00 0.21
Tunisia 5.00 0.76 0.78 0.12 0.30
Turkey 8.18 0.81 0.83 0.02 0.79
Turkmenistan 0.07 0.09
Uganda 7.14 0.00 0.26 7.51 0.51
Ukraine 0.85 0.02 0.81
United Arab Emirates 0.01 0.00
United Kingdom 16.13 0.71 1.00 0.00 0.40
United States 0.09 0.00 0.13 0.06 0.41
Uruguay 0.00 0.00 0.00 0.00 0.95
Uzbekistan 0.45 0.05 0.24
Vanuatu 1.93 0.00
Venezuela 0.44 0.00 0.13 0.54 0.49
Vietnam 6.06 0.59 0.80 6.42 0.51
Virgin Islands US 0.12
Yemen 4.69 0.79 0.01
Zambia 0.79 0.00 0.19 5.68 0.54
Zimbabwe 0.79 0.00 0.17 4.08 0.41

References
Acemoglu, Daron and Simon Johnson. (2007). “Disease and Development: he Efect of
Life Expectancy on Economic Growth.” Journal of Political Economy, 115: 925–85.
Acemoglu, Daron, Simon Johnson, and James Robinson. (2001). “he Colonial Origins
of Comparative Development: An Empirical Investigation.” American Economic
Review, 91: 1369–1401.
Akyeampong, Emmanuel Kwaku. (2006). “Disease in West African History.” In hemes
in West Africa’s History, edited by Emmanuel Kwaku Akyeampong, 186–207.
Athens: Ohio University Press.
Allison, A. C. (2002). “he Discovery of Resistance to Malaria of Sickle-Cell
Heterozygotes.” Biochem. Mol. Biol. Educ., 30: 279–87.
Aregawi, Maru. (2010). World Malaria Report 2010. Geneva: World Health
Organization.
Ashraf, Quamrul, Ashley Lester, and David N. Weil. (2008). “When Does Improving
Health Raise GDP?” NBER Macroeconomics Annual.
Bairoch, Paul, Jean Batou, and Pierre Chevre. (1988). Population des Villes Europeennes
de 800 la 1850. Volume 2. Geneva: Droz.
Chanda, Areendam and Louis Putterman. (2007). “Early Starts, Reversals and Catch-up
in the Process of Economic Development.” Scandinavian Journal of Economics, 109:
387–413.
Chandler, Tertius. (1987). Four housand Years of Urban Growth: An Historical Census
(revised edition). Edward Mellen.
 he Impact of Malaria 129

Comin, Diego, William Easterly, and Erick Gong. (2010). “Was the Wealth of Nations
Determined in 1000 BC?” American Economic Journal – Macro, 2(3): 65–97.
De Vries, Jan. (1984). European Urbanization 1500–1800. Taylor & Francis.
Depetris Chauvin, Emilio and David N. Weil. (2013). “Malaria and Early African
Development: Evidence from the Sickle Cell Trait.” NBER Working Paper 19603,
October.
Diamond, Jared. (1997). Guns, Germs, and Steel: he Fates of Human Societies. New
York: W. W. Norton and Co.
Eggimann, Gilbert. (2000). “La Population des villes des tiers mondes de 1500 la 1950.”
Universite de Geneve, Centre d’histoire economique comparee.
Gallup, John Luke and Jefrey D. Sachs. (2001). “he Economic Burden of Malaria.”
American Journal of Tropical Medicine and Hygiene, 64(1, supplement): 85–96.
Gallup, John Luke, Jefrey D. Sachs, and Andrew Mellinger. (1999). “Geography and
Economic Development.” International Regional Science Review, 22(2): 179–232.
Galor, Oded and David N. Weil. (2000). “Population, Technology, and Growth: From
Malthusian Stagnation to the Demographic Transition and Beyond.” American
Economic Review, 90: 806–28.
Hopkins, A. G. (1973). An Economic History of West Africa. New York: Columbia
University Press.
Kiszewski, Anthony, Andrew Mellinger, Andrew Spielman, Pia Malaney, Jefrey Sachs,
and Sonia Ehrlich Sachs. (2004). “A Global Index of the Stability of Malaria
Transmission.” American Journal of Tropical Medicine and Hygiene, 70(5): 486–98.
Kremer, Michael. (1993). “Population Growth and Technological Change: One Million
B.C. to 1990.” Quarterly Journal of Economics, 108(3): 681–716.
Lee, R. D. and A. Mason. (2011). “Generational Economics in a Changing World.”
Population and Development Review, 37: 115–42.
Mabogunje, Akin L. and Paul Richards. (1985). “he Land and Peoples of West Africa.”
In History of West Africa, volume 1, 3rd ed., edited by J. F. A. Ajayi and Michael
Crowder. New York: Longman Publishing Group.
Masters, William A. and Margaret McMillan. (2001). “Climate and Scale in Economic
Growth.” Journal of Economic Growth, 6(3): 167–86.
Mathers, Colin D., Doris Ma Fat, and J. T. Boerma. (2008). he Global Burden of Disease:
2004 Update. World Health Organization.
McEvedy, Colin and Richard Jones. (1978). Atlas of World Population History. London:
Penguin.
McNeill, William. (1977). Plagues and Peoples. New York: Anchor Books.
Modelski, George. (2003). World Cities:-3000 to 2000. Washington, DC: Faros2000.
Motulsky, Arno. (1964). “Hereditary Red Cell Traits and Malaria.” Am. J. Trop. Med.
Hyg., 13 (Part 1): 147–58.
Mueller, Eva. (1976). “he Economic Value of Children in Peasant Agriculture.” In
Population and Development: he Search for Selective Interventions, edited by
Ronald Gene Ridker, 98–153. Baltimore, MD: Johns Hopkins University Press.
Murray, Christopher J. D. and Alan D. Lopez. (1996). he Global Burden of Disease: A
Comprehensive Assessment of Mortality and Disability from Diseases, Injuries and Risk
Factors in 1990 and Projected to 2020. Cambridge, MA: Harvard University Press.
Nelson, Kennard and Carolyn M. Williams. (2006). Infectious Disease Epidemiology:
heory and Practice. 2nd ed. Jones & Bartlett Publishers.
130 Weil

Packard, Randall M. he Making of a Tropical Disease: A Short History of Malaria. Johns

Hopkins University Press, 2007.
Persad, Govind, Alan Wertheimer, and Ezekiel J. Emanuel. (2009). “Principles for
Allocation of Scarce Medical Interventions.” he Lancet, 373: 423–31.
Piel, Frédéric B. et al. (2010). “Global Distribution of the Sickle Cell Gene and
Geographical Conirmation of the Malaria Hypothesis.” Nature Communications,
2(1): 104.
Putterman, Louis and David N. Weil. (2010). “Post-1500 Population Flows and the
Long Run Determinants of Economic Growth and Inequality.” Quarterly Journal
of Economics, CXXV: 4.
Ramankutty N., J. A. Foley, J. Norman, and K. McSweeney. (2002). “he Global
Distribution of Cultivable Lands: Current Patterns and Sensitivity to Possible
Climate Change.” Global Ecology and Biogeography, 11: 377–92.
Riley, James C. (2005). “Estimates of Regional and Global Life Expectancy, 1800–2001.”
Population and Development Review,31(3): 537–43.
Rozenblat, C., M. Lennert, and G. Van Hamme. (2011). “European Cities in Networks.”
United Nations – Habitat report.
Steyn, Maryna. (2003). “A Comparison between Pre- and Post-Colonial Health in
the Northern Parts of South Africa, a Preliminary Study.” World Archaeology,
Archaeology of Epidemic and Infectious Disease, 35(2): 276–88.
Stone, Lawrence. (1977). he Family, Sex and Marriage in England, 1500–1800. New
York: Harper & Row.
United Nations, Department of Economic and Social Afairs, Population Division.
(1982). Model Life Tables for Developing Countries. New York: United Nations.
(1999). he World at Six Billion. New York: United Nations.
(2009). World Population Prospects: he 2008 Revision. New York. (advanced Excel
tables).
Voigtlander, Nicola and Joachim Voth. (2009). “he hree Horsemen of Riches: Plague,
War and Urbanization in Early Modern Europe,” mimeo, UCLA.
Webb, James L. A. (2006). “Ecology and Culture in West Africa.” In hemes in West
Africa’s History, edited by Emmanuel Kwaku Akyeampong. Athens: Ohio University
Press.
Webb, James L. A. (2008). Humanity’s Burden: A Global History of Malaria. Cambridge,
UK: Cambridge University Press.
Weil, David N. (2010). “Endemic Diseases and African Economic Growth: Challenges
and Policy Responses.” Journal of African Economies, 19 (supplement 3): 81–109.
Williams T. N., T. W. Mwangi, S. Wambua, N. D. Alexander, M. Kortok, R. W. Snow,
and K. Marsh. (2005). “Sickle Cell Trait and the Risk of Plasmodium Falciparum
Malaria and Other Childhood Diseases.” Journal of Infectious Diseases, 192(1):
178–86.
Young, Alwyn. (2005). “he Git of the Dying: he Tragedy of AIDS and the Welfare of
Future African Generations.” he Quarterly Journal of Economics, 120(2): 423–66.
 4

African Population, 1650–2000: Comparisons

and Implications of New Estimates

Patrick Manning

New and comprehensive estimates of African population, at regional and

continental levels, suggest that from the seventeenth to the twentieth cen-
tury the continent’s population was much larger in size yet growing at a
slower rate than previously thought. In a project that is nearing completion,
Scott Nickleach and I, with the assistance of Yun Zhang, Brian McGill, and
Bowen Yi, are projecting African populations from 1950 back to 1650.1 he
analysis relies on a combination of methods that account for decennial and
regional estimates of net growth rates, efects of various social and envi-
ronmental factors, and especially the demographic impact of enslavement
and the attendant migration and mortality. Two hypotheses – a dense early
modern African population and a decline in African population because of
slave trade – are linked tightly together.
he present chapter focuses on the social implications of this new demo-
graphic argument. It begins with a concise summary of the new estimates
at continental and regional levels, a comparison with previously accepted
igures, and a brief exploration of the plausibility of the new igures. he
bulk of this chapter discusses global comparisons of African population
estimates and focuses especially on the main economic and social implica-
tions of these new estimates for our understanding of African history.

4.1 SUMMARY OF THE DEMOGRAPHIC ESTIMATES

he analysis begins with the dependable igures for the late twentieth cen-
tury and works stage by stage to earlier times. For this reason, the discus-
sion of the overall results, summarized in Table 4.1, moves from the most
recent period to earlier periods.

1
Work in process is to appear as Manning and Nickleach (forthcoming).

131
132 Manning

Table 4.1. Estimated African population by region, 1700–2000, in millions

West Central Southern East Northeast Northern Total

2000 226 93 50 172 112 154 807
1950 60 26 15 42 33 44 220
1890 41 19 10 26 20 29 145
1790 46 17 12 22 22 23 141
1700 50 22 10 20 19 17 138

he period from 1950 to 2000 was one of record-setting population

growth, at net growth rates consistently more than 2.5 percent per year.
his growth rate resulted especially from rapidly falling rates of mortality:
from 1950 to 1990, the expectation of life at birth for the average African
rose from thirty-ive years to ity years. his dramatic decline in mortality
accompanied the accession of African nations to independence and sub-
stantial expansion in public health expenditures by the new governments
and, in preceding years, by late colonial governments. In addition, a world-
wide decline in mortality resulted from the spread of antibiotics and many
other improved medical treatments. (hen, in a devastating change, average
expectation of life decreased drastically in southern and eastern Africa dur-
ing the 1990s as a result of the AIDS crisis, and declined to a lesser degree
elsewhere on the continent.)
In the preceding sixty years, from 1890 to 1950 – the colonial era for most
of Africa – African populations gradually shited from stagnation to demo-
graphic transition. Populations rose at a very modest rate from 1890, then
accelerated from 1920 to 1950. he most rapid decline in mortality rates
took place at the end of this period in the 1940s, as DDT spraying reduced
malaria incidence in many areas, enabling rapid population growth. African
growth rates were much smaller than those assumed by colonial oicials,
however, so that the African populations were substantially larger than the
estimates prepared by colonial regimes. Population growth was held down
by the violence that continued on a substantial scale from 1890 to 1920:
enslavement continued in the regions not yet under European control and
the establishment of European control meant wars of conquest, repression
of rebellions, and forced recruitment of laborers.
he era from 1790 to 1890 included the peaks of three waves of enslave-
ment – the 1790s peak of the transatlantic slave trade, the 1870s peak of
the trans-Saharan and Indian Ocean trades, and the 1850s peak in enslave-
ment of people held within the continent. he combination of these factors
 African Population, 1650–2000 133

created serious negative pressure on African regional populations – with

the worst efects felt irst in West Africa, then in Central Africa, then in East
Africa, and inally along the northern savanna. Overall, African population
remained stagnant slightly from 1790 to 1890.
From 1450 to 1790, the Atlantic slave trade grew irregularly. From 1650,
the drain of enslavement became suicient to reduce population size in var-
ious regions of the western coast of Africa, and by 1730, a general decline
had begun in the population of coastal West and Central Africa, contin-
uing until the 1850s. Because most of the captives sent across the Atlantic
were male, a signiicant shortage of adult males developed within West
and Central Africa and brought a substantial transformation in gender
relations.
he methods used for these population estimates rely on projections
backward in time. he projections start from 1950 and trace population for
seventy regions of Africa corresponding to national and provincial bound-
aries of today, selected in such a way as to be representative of economic
and social regions of earlier times. For the twentieth century, the projec-
tions rely on crude rates of growth drawn from study of a wide range of fac-
tors. For the nineteenth century and before, the analysis relies especially on
a simulation that accounts for the age and sex composition of free popula-
tions and of those enslaved. he work includes new estimates of the volume
and direction of the transatlantic slave trade, based on statistical analysis of
the well-known “Voyages” dataset (Eltis).

4.2 NEW VERSUS PREVIOUS ESTIMATES

In general, the new estimates difer from previous estimates in that they
were constructed through a comprehensive analysis. Previous studies took a
piecemeal approach, investigating restricted regions, short periods of time,
a limited range of social interactions, or simple continental guesstimates.
Addressing the fuller scope of issues permitted the location of contradic-
tions, lacunae, and errors in analysis. he efort to address the full range of
issues in African population history has resulted in a more internally con-
sistent argument. he groundwork for this analysis was the research per-
formed on African populations under the auspices of the United Nations.
For 1950–2000, the current decennial estimates of African population
published by the United Nations Population Oice are broadly depend-
able, thanks to years of review and analysis by demographers at the UN and
elsewhere. he population totals for 1950 and 1960 are well above those
irst reported in censuses and surveys at the moment of African national
134 Manning

independence. hus, the true size of the African population in recent

decades has only gradually come to be recognized.
Documenting African population for the years before 1950 will require
an approach analogous to that successfully applied to the post-1950 years:
large-scale, collaborative work, piecing together the many available shreds
of evidence into a coherent global pattern. he program of estimation
reported on here is an initial stage in the comprehensive analysis of the
African demographic past that needs to be pursued.
Meanwhile, the literatures on African history, economic history, and
demographic history have generally been skeptical of the notion that Africa
had a large and dense population from early modern times, skeptical of the
notion that African rates of population growth were lower than those of
other regions, and skeptical of the notion that enslavement caused decline
in African population. hus this argument for the plausibility of the new
estimates requires commentary on the preceding threads of skepticism
about the idea of a dense pre-slave-trade African population.
An initial basis for skepticism about large early modern African popula-
tions arose from methodology – from the piecemeal approach of studies in
demographic history. Perhaps understandably, scholarly studies from the
1960s through the 1980s worked on short periods of time and portions of
the African continent, so they were never able to compare the igures they
were using with continental igures. West Africanists studied the Atlantic
trade; East Africanists, ater a delay, studied the Indian Ocean trade; a con-
tinental picture did not yet exist.2 For the precolonial era, analysts have
given insuicient attention to the diference between the nineteenth cen-
tury, when enslavement raged over most of the African continent, and the
previous centuries, when the Atlantic slave trade was the main engine of the
patterns of migration and social disruption that afected West and Central
Africa most severely.
A second skepticism is based on a habitual reliance on colonial statisti-
cal records. During the years before World War II, it was widely assumed
that the African continental population had risen from about 100 million
in 1900 up to about 130 million in 1930 (Carr-Saunders 1936; Kuczynski
1934; Willcox 1931). It has only recently been noted that, to reach from

2
And this despite the fact that continental estimates of African population had been
advanced since the seventeenth century and were prominently published during the twen-
tieth century. My 1990 attempt at a continental picture fell short in at least two ways: it was
not linked to post-1950 populations and it did not enable readers to verify its calculations
(Manning 1990).
 African Population, 1650–2000 135

130 million in 1930 to 220 million in 1950, Africa’s population would have
had to grow at 2.7 percent per year through depression and war. Global
comparison shows that no large population worldwide was growing at such
a rate before 1950 – the growth rate must have been lower and the earlier
populations must have been higher. hat is, African populations of the early
twentieth century turn out to have been signiicantly underreported. Such
underreporting is understandable from the viewpoint of African communi-
ties. Presumably, African communities had little incentive to present them-
selves for enumeration when the main results of enumeration were taxation
and recruitment rather than provision of social services. Only from the
1950s, once social services began to appear in the national era – and when
political representation began to be determined by population size – did the
response to population surveys grow more positive.
A third basis for skepticism about the size of African populations and
the efects of slave trade came from comparison with European migration.
his was a comparison of European transatlantic migration (1850–1940)
with transatlantic trade in African captives (mostly from 1650 to 1850).
Because it was known that Europe, in the years ater 1850, was able to grow
in population even as large numbers of migrants let their home countries,
it seemed to many observers that Africa ought also to have been able to
avoid population decline. Ater all, Africa’s 10 million emigrants compared
to Europe’s 50 million. Nonetheless, a detailed analysis shows that Africa’s
export slave trade was indeed suicient to reduce the population of major
regions and, for more than a century, of the continent as a whole. he nega-
tive factors that made the diference were the high general levels of mortality
for Africa, the additional mortality brought by the violence of enslavement,
the loss of young adult females and their ofspring, and the high levels of
maritime mortality in the days before steamships. For the reproduction of
African population, the key group was fertile females, generally those in the
age group from iteen to forty-ive. As soon as enough of them died or were
exported, the result would be suicient to halt population growth.
A fourth basis for skepticism, questioning the negative impact of enslave-
ment on African population, argued that African nutrition improved with
the arrival of American crops (maize, manioc, peanuts, etc.), and that
African population therefore expanded. his argument – focusing on the
era from the sixteenth to the nineteenth centuries – while plausible, has not
yet come close to veriication. hat is, we have yet to verify when American
crops became a signiicant portion of African diets, and have yet to ver-
ify that they were substantially more productive than the previous African
136 Manning

crops (Manning 1982a; Wigboldus 1986).3 It is possible, therefore, that

Africans of the nineteenth century had a broader diet but consumed no
more calories than their predecessors of the sixteenth century.
All of these reasons apply to the post-1450 period of Atlantic maritime
voyages. An additional element of the overall African demographic debate is
that of whether African population in earlier – “medieval” – times was dense
or sparse. Some researchers claim that the population of coastal regions of
West Africa and Central Africa had been extremely sparse in these times,
and that they were settled from further inland only in early modern times.
For instance, scholars have argued that the prevalence of malaria and other
tropical diseases prevented population from becoming dense. his type of
argument lacks plausibility because the presence of infectious disease, while
a negative factor in the short term, is a clear indication that a dense popu-
lation has existed in a region over the long term, as such a population was
necessary to sustain the disease. Language distribution is another key to the
historical depth and density of population: for West Africa in particular, the
distribution of subgroups of the Niger-Congo language group indicates that
populations have lived and migrated – in both the forested and savanna
zones – for thousands of years, with a time depth far greater than that of
populations in Europe (Williamson and Blench 2000, 11–42).4 his does
more to demonstrate that West African population was ancient than that it
was dense, but it dismisses the notion that any regions of West Africa were
unsettled (Figure 4.1).
Walter Willcox, A. M. Carr-Saunders, and R. R. Kuczynski, writing in
the early 1930s, argued for a relatively static African population of about
100 million from 1650 to the late nineteenth century, and for a population
of roughly 130 million in 1930 (Carr-Saunders 1936; Kuczynski 1934, 240–
8; Willcox 1923). Caldwell and Schindlmayr argued forcefully against this
vision, on the grounds that Africans had expanded into previously under-
populated areas from the sixteenth century, relying on new crops from the
Americas.5 Caldwell and Schindlmayr note approvingly “a collapse of the
earlier African consensus” ater 1970, as Durand, McEvedy, and Jones and

3
On maize: (McCann 2005; Miracle 1966). hese works, while highly valuable for the nine-
teenth and twentieth centuries, provide little dependable information on the initial adop-
tion and spread of maize in Africa.
4
For the methods of this analysis, see Christopher Ehret, History and the Testimony of
Language (Berkeley: University of California Press, 2010).
5
“Kuczynski was admittedly a demographic conservative, and as late as 1944 . . . was arguing
that Africa and probably most of the developing world was then not more populated than
it had been 200 years earlier” (Caldwell and Schindlmayr 2002, 186). See also Kuczynski
(1944).
 African Population, 1650–2000 137

250

200 Manning 2011

McEvedy & Jones

150 Maddison 2003

Willcox 1931

100 Carr-Saunders 1936

Durand 1967

50 Caldwell 1985

0
1500

1529

1560

1590

1621

1651

1682

1713

1743

1774

1804

1835

1866

1894

1925
Figure 4.1. New and previous estimates of African population, 1500–2000.
Sources: Previous estimates from McEvedy and Jones (1978), Willcox (1931), Carr-
Saunders (1936), Maddison (2001), Durand (1965), Biraben (1979), and Caldwell and
Schindlmayr (2002).

Maddison each set African population at the much lower igure of about
50 million in 1500.6 John C. Caldwell remains unsurpassed in his knowl-
edge of African demography in late colonial and early postindependence
Africa, and his analysis was crucial in unraveling the “consensus” among
demographic historians that prevented an appropriate linkage among pre-
colonial, colonial, and postcolonial demographic studies. But his willingness
to treat Africans of the eighteenth century as just entering the Neolithic,
combined with his neglect of the severity of enslavement throughout the
continent, led him to assume rapid African population growth from the
sixteenth through the nineteenth century when such growth was quite
impossible.7

6
Caldwell and Schindlmayr argue an extreme version of this hypothesis, claiming that the
loss to enslavement “probably did not translate into a comparable diminution of the rate
of natural increase, and certainly not to a population decline, for three reasons. he irst
is that . . . the loss of some of the population probably allowed others to survive. he sec-
ond reason is that only one-third of those transported as slaves were females. . . . he third
point is that the period of slavery coincided with the spread of new foodstufs introduced
from the Americas, fostering slow but continuing population growth. . . . It is extremely
doubtful whether either the absolute or relative decline of Africa’s population ever took
place” (2002, 196–7).
7
Caldwell and Schindlmayr argue that, “he spread of the Neolithic revolution in the form
of hunting giving way to farming continued throughout the nineteenth and twentieth
centuries and is still not complete.” his argument that Africans were just reaching the
Neolithic is particularly surprising given the ample evidence that the African iron age
138 Manning

4.3 GLOBAL COMPARISONS

he new estimates of African population, when compared with populations
elsewhere in the world, are shown to be diferent indeed from previous esti-
mates. Table 4.2A shows Maddison’s population estimates for Africa and
other world regions; Table 4.2B shows the new African estimates along with
Maddison’s estimates for other regions. he new igures in Table 4.2B show
that, from 1700 to 1900, African population was roughly stagnant, while
the populations of all other major world regions grew at accelerating rates:
by 1900, African population had declined to no more than one tenth of the
global total. African populations began growing during the twentieth cen-
tury as growth rates slowed in some areas and accelerated in others. In the
last half of the twentieth century, African populations grew at a very high
rate, and by the time these growth rates began to decline, African popula-
tion had again reached one seventh of the world total.
Along with the planetary comparison of Table 4.2, here are some addi-
tional comparisons of population size and density. Comparing Africa with
Eurasia reveals a remarkable parallel in population density. Africa, in its
surface area, has some 30 million square kilometers, while Eurasia has some
53 million square kilometers. Africa’s current population is roughly one bil-
lion, while that of Eurasia is roughly four billion. hus, Africa’s population
density is now just short of half that of Eurasia.8 For 1700, Maddison shows
Eurasian population at 530 million and African population at 60 million;
the new estimates show an African population of 140 million in 1700. hus
the new estimates show an African population density that was half that
of Eurasia – the same ratio as at present – while Maddison’s igures pro-
pose an African population density in 1700 that was less than one fourth
that of Eurasia. hat is, the old estimates argue that Africa was consistently
a sparsely populated continent until very recent times; the new estimates
argue that Africa was not far behind Eurasia in density and has maintained

began at the same time as that of Eurasia, some two thousand years before the Atlantic
slave trade expanded (Caldwell and Schindlmayr 2002, 197). Twenty years earlier, Caldwell
had acknowledged the validity of my assertion that slave trade had the capacity to reduce
African population, though at that time too he preferred the argument that American
crops had overcome all obstacles to African demographic growth (Caldwell 1982); see also
Manning (1981 and 1982b).
8
his comparison in continental areas can be adjusted by accounting for underpopulated
areas – the Sahara in Africa; Arabia and Siberia in Eurasia – yet the results tend to cancel
out. Still another version of the comparison is to compare populations to areas of arable
land for Africa and Eurasia – these too tend to show an African density about half that of
Eurasia.
 African Population, 1650–2000 139

Table 4.2A. Size and proportion of world population, 1700–1950

Maddison estimates by world region (showing % of global population)

1700 % 1820 % 1913 % 1950 %

Africa 61 10 74 7 123 6 228 9
Asia 402 67 710 68 978 55 1382 55
L. Am 12 2 21 2 81 5 166 7
Europe 125 21 224 22 497 28 572 23
W. ofshoots 2 0 11 1 111 6 176 7

Table 4.2B. New estimates for Africa, Maddison estimates elsewhere

1700 % 1820 % 1913 % 1950 %

Africa 140 21 140 13 145 8 228 9
Asia 402 59 710 64 978 54 1382 55
L. Am 12 2 21 2 81 4 166 7
Europe 125 18 224 20 497 25 572 23
W. ofshoots 2 0 11 1 111 6 176 7
Source: (Maddison 2001, 175, 183). Maddison’s “Western Ofshoots” include the United
States, Canada, Australia, and New Zealand; “Europe” includes the entire territory of
the former Soviet Union. I calculated the growth rates shown.

or periodically regained its relative position in population density over the

centuries.
Comparing Africa and Europe reveals a remarkable parallel in popula-
tion size over the long term and shows both parallels and contrasts in the
history of migration. European and African populations are roughly equal
today, but Europe’s population density is three times that of Africa; Europe
(including Russia) is only 10 million square kilometers in area. he two
regions had equal populations of some 140 million in 1700. European pop-
ulations rose at a steady rate, however, while African population remained
stagnant until the mid-twentieth century. In migration, Europeans out-
numbered Africans as out-migrants from 1500 to 1600, Africans outnum-
bered Europeans as migrants from 1600 to 1850, Europeans outnumbered
Africans again until 1960, and African migrants grew in number thereater.
Among other diferences in the migratory patterns were that, in the era of
slave trade, there were few slaves among European migrants, and European
migrants were able to return home. Ater 1850, European migrants trav-
eled on low-mortality steamships, while the ongoing African slave trade
140 Manning

continued to exact a high toll in mortality. Some factors worked in the

opposition direction. In monogamous Europe, the absence of males let
women unmarried or marrying later in life, thus afecting their fertility. In
polygynous Africa, the remaining women still had children in marriage or
concubinage; indeed, it seems that the long-term shortage of males brought
by slave trade expanded the frequency of concubinage (if not of formal
marriage) in Africa.9
Africa and the Americas provide remarkable contrasts in population his-
tory. he Americas have 41 million square kilometers in area, but if one
subtracts the 10 million square kilometers of sparsely populated Canada,
the result is equal to the area of Africa. At present, the populations and
therefore the population densities of Africa and the Americas are roughly
equal. In 1700, however, the Americas had only some 13 million inhabit-
ants (at least two million of them of African descent), so that the population
density of the Americas was one tenth that of Africa.10 Among the varia-
tions in patterns of population were that American populations declined
sharply because of disease during the sixteenth and early seventeenth cen-
turies, while from 1650 to 1850 African populations declined through slave
trade as American populations rose through African immigration and
recovery of the indigenous population. American populations grew rap-
idly through European migration ater 1850 while African populations
remained stagnant.11

4.4 AFRICAN IMPLICATIONS

he new demographic estimates for Africa, when combined with a recon-
sideration of the continent’s economic, social, cultural, and ecological
conditions, yield a substantially revised picture of African life. his sec-
tion begins by laying out the main elements of the revised picture, then
turns to explaining the logic of the new picture and providing recommen-
dations on ways to pursue the reinterpretation of precolonial and colonial
Africa.

9
he large-scale, steamship-borne migrations from 1850 to 1940 transported migrants from
other densely populated regions (Europe, East Asia, South Asia), but not from Africa.
Only by 2000, when Africa had again become relatively densely populated, did overseas
African emigration again become signiicant (McKeown 2004).
10
Or an American population density one fourth that of Africa, using Maddison’s popula-
tion estimate.
11
During the sixteenth and seventeenth centuries, African populations too may have been
limited by the global circulation of diseases in that era.
 African Population, 1650–2000 141

African populations from 1500 to 1700 were relatively dense, linked with
one another in various networks of exchange, relying on local moneys.
Except along the fringes of the Eurasian commercial system, African states
were small. A sort of “informal economy” enabled many to participate, if
marginally, in the continental system of exchange. Mortality was high, many
were poor, and reliance on external trade was minimal. Yet to treat this as
a system of relatively dense population, we must assume that people were
linked through networks of exchange rather than reliant on self-suiciency
in isolated households. From 1700 to 1900, slavery expanded in region ater
region of the continent, bringing violence, refugee life, increased mortality,
new sorts of hierarchy (notably an expanded gender hierarchy), a growing
proportion of the population in slavery, and the rise of states that relied sig-
niicantly on warfare and enslavement. External trade – for which exports
relied especially but not only on slaves – grew steadily from 1700 to 1900.
From 1900 to 1950, African populations recovered from the pressure of
enslavement and the shock of colonial conquest, and gradually entered an
era of lower mortality. African economic growth (measured in terms of
domestic product or external trade) remained relatively slow during the
nineteenth and twentieth centuries except for occasional bursts of positive
or negative growth; demographic growth became strongly positive from the
1940s to the 1990s.
As a framework for assessing African socioeconomic life, we may begin
with that ofered in 1973 by A. G. Hopkins, whose groundbreaking analy-
sis of West African economic history crystallized the analysis of African
economies through his distinction of “the domestic economy” from the
“external trade” that crossed the desert and the ocean (Hopkins 1973). In
these terms, African total output or Gross Domestic Product is the sum of
the output of the domestic economy and the net value of external trade.
Hopkins’s two-sector model, although proposed for the interpretation of
precolonial African economies, has in fact remained the default system for
analyzing African economies from earliest times until the present. While
this formulation is entirely logical, we will see that it has severe practical
disadvantages. hat is, while Hopkins proposed the two-sector model in an
efort to draw attention to the size and diversity of the domestic economy,
in practice, the historical and contemporary analysis of African economies
has focused overwhelmingly on external trade. he domestic economy is
commonly let out of discussion; further, it is widely presumed to have been
homogeneous and rudimentary.
Viewed in this two-sector framework, the new estimates of African pop-
ulation imply a major rethinking of African economic structures. he new
142 Manning

population igures suggest that the “domestic economy” was much larger
than has been realized, and that “external trade” was a smaller proportion
of GDP than previously realized. hat is, the historical levels of sub-Saharan
African external trade – across the oceans and desert – are known, within
an order of magnitude. Now that we have new and higher estimates of
African population, it is necessarily the case that estimates of output in the
domestic economy for sub-Saharan Africa are larger than before, and that
the ratio of intercontinental trade to GDP will appear smaller.12
For present purposes, therefore, the framework of the “domestic econ-
omy” appears vague and inadequate in that it avoids specifying the levels
of exchange and economic interaction within the continent. he estimate
of a larger population leads by demographic logic to the presumption of
an expanded, continent-wide domestic economy. he same estimate of a
larger population also leads, by the logic of economic productivity, to the
presumption of substantial sectors of regional and local exchange networks.
his expanded domestic sector is best treated in terms of a range of subsec-
tors: self-suicient production (including foraging), local exchange (with
attention to occupational specialization), cross-regional and monetized
commerce within sub-Saharan Africa, and the secondary and tertiary
sectors of governance, knowledge, and cultural production. he “external
trade” sector, the best-documented aspect of African economies, can then
be linked more explicitly to the interacting subsectors of the domestic econ-
omy. In the paragraphs that follow, this expanded notion of sectors within
the domestic economy is applied successively to the era from 1500 to 1700,
the slave trade era from 1700 to 1900, and the colonial and neocolonial
period of the twentieth century.

4.4.1 Rethinking Africa from 1500 to 1700

Such a framework should facilitate the creation of a new picture of Africa
from 1500 to 1700, when African populations were presumably growing
yet already dense. he elements of such a reconsideration proposed here
include land-labor ratios, systems of money, the nature of ethnicity, the

12
For an earlier discussion on the balance of the domestic economy and overseas trade in
the Bight of Benin, see the works of Peukert and Manning: for the kingdom of Dahomey,
Peukert assumed export value reached only 4 percent of GDP; for the larger Bight of
Benin, Manning assumed export value reached 15 percent of GDP. Manning proposed
a three-century series on export revenue and per capita export revenue, with rougher
estimates of GDP (Manning 1982a, 2, 44; Peukert 1978).
 African Population, 1650–2000 143

functions of states, the role of poverty, and the contemporary notion of the
“informal economy” as applied to precolonial times.
We need to rethink land-labor ratios for Africa. Should Africa before the
twentieth century continue to be seen as a continent of land surplus and
labor shortage? Or should it be viewed as a region of relatively dense popu-
lation where land was better seen as a scarce resource than a free good? he
interpretive choice, highlighted clearly by the new estimates, draws renewed
attention to the old diferences between Hopkins’s assumption of African
labor shortage, the Gemery and Hogendorn assumption of African labor
surplus, and Ester Boserup’s analysis of behavior change resulting from
increased population density (Boserup 1965; Hopkins 1973; Gemery and
Hogendorn 1974).13 Regional and global comparisons of the labor intensity
of production and the social organization of labor will surely reveal a need
to study African labor more fully.
Study of the forms and dynamics of African money can facilitate the
required rethinking of exchange in African markets. It is noteworthy that
gold served as currency in the large portions of West Africa where it was
mined, and gold also served as currency in southeast Africa and along the
Swahili coast. Silver presented a contrast: silver was the main international
currency from the late sixteenth century, but silver was little used as cur-
rency in sub-Saharan Africa. Partly this was a response to the scarcity of
silver in Africa, but it also suggests that Africa was outside the main cur-
rents of global trade until the mid-nineteenth century, when silver coins
became widely used in many parts of Africa. (An exception to this pattern
is Ethiopia and northeast Africa generally – silver was traded and coined
in that region, relecting its integration into Middle East and Indian Ocean
trading circuits.) Overall, a denser African population means a further
magniied density in marketing. It means that there is much more to be
learned about African currencies, credit, and other means of exchange.
We need to rethink the handling of ethnicity in Africa. Too oten, analy-
sis has focused on ethnic groups as if they were autonomous and self-sui-
cient groupings. Among the approaches to ethnicity that provide promise
for developing an interactive approach to a multi-sector domestic economy
is the work of Frederik Barth and Frederic Pryor. Barth’s analysis of ethnic
groups and boundaries, showing how families crossed ethnic boundaries at
times when it was economically necessary, became useful in understanding

13
he Gemery-Hogendorn interpretation of slave exports in terms of the “vent for surplus”
theory was one of the few arguments of its time that land rather than labor was in short
supply in the era of slave trade.
144 Manning

the response to drought in the central Sudan, as families shited from

agriculture to pastoralism and back (Barth 1994). Pryor’s global study of
peasant distribution systems, drawing on ethnographic accounts from the
Human Relations Area Files, analyzed four types of distribution systems in
peasant economies and found correlations suggesting surprisingly complex
processes of economic choice and evolution (Pryor 1977).14 As with ethnic-
ity, so also with the spatial distribution of African population; attention to
zones of high economic activity and high population density may reveal
the African dynamics of interchange and division of labor that have been
neglected in interpretations that have conceded too much to the notion of a
poor, underpopulated, and backward African countryside.
he new estimates raise parallel questions about states and statecrat in
Africa. While Africans had states throughout the continent, these states in
general did not develop the bureaucracies, armies, and monuments of those
in Eurasia. Indeed, the scarcity of such great states is one of the reasons
scholars have tended to assume that African populations were sparse. Quite
a diferent possibility is that African societies found ways to develop and
nourish dense populations without giving such support to elite classes and
state institutions.
Similarly, we should consider the issue of poverty in long-term perspec-
tive. It may be that signiicant insights will come from a rereading of John
Ilife’s remarkably broad review, he African Poor. In this continental study
of poverty in the nineteenth and twentieth centuries, Ilife distinguished
between the poor and the very poor, where the latter never escaped the
fringes of hunger (Ilife 1987). he question is how far into the past this
analysis of poverty might be carried.
It is unlikely that new investigations will unearth previously unsuspected
concentrations of great wealth and centralized power in Africa. We should
look instead for an Africa of the past that can be linked to the Africa of
the present – that is, societies with relatively dense populations, with con-
siderable socioeconomic inequality, with an immense range of communi-
ties and activities, but in an environment of high mortality and widespread
poverty. For instance, we should consider the beneits of extending to the
more distant past the recently developed logic of the informal economy. A
substantial literature has grown up centering on contemporary economic
activity that takes place outside of the formal sector of wages, proits, and
regulation – work and transactions that enable people to survive and some-
times prosper, but that escape formal economic accounting (MacGafey

14
A parallel study relying on HRAF records was that of Patterson (1982).
 African Population, 1650–2000 145

1991; Portes, Castells, and Benton 1989). he phenomenon of the informal

economy, while generally treated as a spinof of the globalizing capitalist
economy, may instead be the current version of what was earlier the domes-
tic and regional market of Africa. hat is, the lessons learned from studying
the contemporary informal sector might be applied to earlier times, and
might help explain how a large African population could have sustained a
lively regional economy that nonetheless had little interaction with global
commerce.

4.4.2 African Social Dynamics, 1700–1900

he multi-sectoral approach to Africa’s domestic economy can then be
applied to the era from 1700 to 1900. During these two centuries, population
stagnated overall and luctuated up and down in most regions according to
luctuations in slave trade and other factors, and external trade expanded
in tandem with domestic enslavement. he new population estimates pro-
vide a new basis for analyzing the main historical questions of this era. For
instance, did the conditions of African life worsen in the era of slave trade?
Did general mortality rates rise in West and Central Africa from the sev-
enteenth to the nineteenth century and in nineteenth-century East Africa?
It is certainly the case that enslavement, drought, famine, and epidemic
reinforced one another periodically to the point that it becomes diicult to
distinguish the various causes of misery (Miller 1982). he savanna regions
of West Africa and Angola were particularly vulnerable to such downturns,
but social tumult and the creation of refugee populations brought periodic
turmoil to many African regions. Investigators of the emergence of the HIV
virus in humans continue to explore the question of whether it was the
social disorder of late nineteenth-century enslavement that led to the trans-
fer of viruses among species and the formation of the human-carried ver-
sion of HIV.
It may be that the character of African statecrat changed signiicantly
as enslavement grew in signiicance. hat is, the earlier focus of localized
states on the welfare of the monarch’s constituents (as in Ife and the Kuba
kingdom) came to be replaced by more warlike and hierarchical states (such
as Dahomey, Imbangala, and Segu). Meanwhile, the expansion of enslave-
ment put a premium on efective defense and on the maintenance of refu-
gee communities, which again undermined respect for large states.
What was the size of slave populations in Africa? his analysis has argued
that the enslaved populations of Africa grew sharply in the eighteenth cen-
tury and then grew again even more sharply in the nineteenth century. It
146 Manning

seems certain that the worldwide peak in enslaved populations came in

roughly 1850, and that those in Africa were the majority of those held in
slavery. hese are demographic projections, however, rather than descrip-
tions of individual lives, so much more research will be necessary to verify
the validity and the meaning of this assertion.
Another measure of the long-term efects of enslavement can be taken
through comparison of the two-plus centuries of enslavement in West
Africa with the one century of enslavement in East Africa. In eighteenth-
century West Africa, transatlantic slave trade expanded steadily and trans-
Saharan slave trade expanded to a lesser degree; eighteenth-century East
Africa remained in a situation of relative autarky. In nineteenth-century
West Africa, the Atlantic trade declined while the trans-Saharan slave trade
expanded somewhat and enslavement within the region expanded as well.
In nineteenth-century East Africa, Indian Ocean slave trade grew rapidly to
a peak late in the century, then came almost to an end at the end of the cen-
tury. In West Africa, the nineteenth-century decline in Atlantic slave trade
coincided with a rapid growth in commodity exports; commodity exports
from East Africa expanded considerably later. Overall, it appears that West
Africa rebounded irst from the negative demographic efects of slave trade,
followed by Central Africa and then by East Africa. But the East African
slave trade, though brutal in its impact, lasted for a shorter time than that for
West and Central Africa, so the continuing social markers of the slave past
seem clearer in West and Central Africa. In East Africa, Mozambique and
Malawi are the regions that show the most ater-efects of enslavement.

4.4.3 Twentieth-Century Africa

From 1900, enslavement had nearly stopped; those in slavery gradually
achieved non-slave or ex-slave status. Export agriculture and expanding
cities provided new opportunities for some, and refugee populations grad-
ually came into the open. here were exceptions to this positive picture:
for instance, the turmoil of the Congo basin and north Central Africa in
the early colonial years, under King Leopold’s Congo Free State and in the
French Congo, each succeeding a period of several decades of intensive
slave raiding. Whatever their growth rate, colonial African populations
are now known to have been far larger than European rulers understood
at the time. he inlection point was in the 1940s, when a world distracted
by war let Africa in a near-autarkic situation. From that time, the pop-
ulation of Africa began expanding at an unprecedented and unrelenting
rate until the present.
 African Population, 1650–2000 147

How long-lasting were the social efects of large-scale enslavement? Here

is one suggestion that some efects may have been long-lasting indeed.
Today, in the early twenty-irst century, the African regions with the high-
est incidence of polygyny and the lowest level of female education are the
same regions as those from which the export of male slaves was most con-
sistent and longest lasting: West Africa and Central Africa (Tabutin and
Schoumaker 2002). Can it be that centuries of slave exports and uneven
adult gender balance created a self-reproducing system that continues to
rely on the subordination of women?
To further test and document this line of argument, we need to expand
research on long-term African population patterns, working especially in
archaeology, linguistics, social and anthropological studies, and with writ-
ten documents. For these demographic questions and for the accompanying
social questions, it will take a while to clarify and document the implica-
tions of this new understanding of African population history.

4.5 AFRICA IN GLOBAL CONTEXT

he demographic analysis developed here contests sharply the notion –
which has grown especially in the past forty years – that African population
has consistently grown at a rate more rapid than the average of human pop-
ulation growth. Such a “rapid African growth” thesis has led to projections
of a small African population base in medieval times and steady growth
until the mid-twentieth century. his analysis also contests the colonial-era
estimates of African population, because they were generally too low to be
consistent with the earliest dependable African population igures during
the 1950s and 1960s.
It is deinitely the case that African population since 1950 has grown at
rates higher than the average for humanity, but that pattern cannot reason-
ably be projected into earlier times, and for three main reasons. First, there
is no reason to assume that African growth rates were among the highest
in the world before, during, or ater the era of the slave trade. Second, the
efects of enslavement seriously reduced the ability of African populations
to grow. hird, general African mortality was relatively high, so that preco-
lonial and colonial African growth rates may well have been lower than for
other regions. he old ideas about Africa’s demographic past simply cannot
stand up to a comprehensive and comparative analysis of African popula-
tion in global context.
he African continent, viewed from the outside, appeared to have partici-
pated only to a modest degree in the global interactions and transformations
148 Manning

of the era from 1500 to 1700, though the continent may have undergone
major changes that are not yet fully understood – impact of disease is one
possibility. hen, from 1650 to 1850, the continent was seriously drained
in population through enslavement.15 African migration continued ater
1850 – as enslavement sent many captives to destinations within Africa and
across the Sahara and into the Indian Ocean – though expanding migra-
tion from Europe and Asia overshadowed African migration ater 1850.
he adoption of American crops by African farmers is known to have taken
place from 1500, but we do not yet have precise information on the pace of
adoption or on the nutritional or demographic results of that agricultural
change, and cannot simply assume that it overcame all other factors to gen-
erate steady population growth. From 1850 to 1950, Africa fell increasingly
under European colonial rule, with its mix of repression, taxation, and eco-
nomic transformation. hroughout these successive periods, African popu-
lations, though relatively dense, met consistent limits that prevented any
long-term growth.
With independence ater 1950, African nations experienced bursts of
development in health, education, and economic growth, but fell into a pat-
tern of neocolonial subordination to international organizations, as with
the World Bank’s structural adjustment programs. During this period, how-
ever, African populations escaped their previous limits and grew steadily at
rates more than 2.5percent per year, a growth rate high enough to create its
own problems.
While Africans overcame the limits on their demographic growth, the
limits on African economic growth remained in place. From 1850 if not
before, African prices and wages in international markets have been held
at extremely low levels. Investment in the continent has been overwhelm-
ingly out of domestic funds – yet the African funds available for investment
were restricted by the low incomes that Africans gained in international
transactions. Nevertheless, African economies have undergone transfor-
mations and even growth over the past century, a pattern that tends to
argue that the domestic economy continues to have a viable mechanism
of accumulation (Manning 1982a, 218–19). he population of the African
continent is now – and presumably once again – near to one sixth of the
total of humanity, and the distinctive long-term history of its demographic,
economic, and social life may have important lessons for the continent itself
and for regions beyond it.

15
he guesses of Willcox and Carr-Saunders – that Africa remained in demographic stasis
during the era of slave trade – are conirmed in general by this new research.
 African Population, 1650–2000 149

In sum, this assembly of evidence and analysis suggests that we should

develop a modiied perspective on the overall characteristics of African
population from the sixteenth century through the colonial era. hat is, we
should envision African populations that were relatively dense, in compari-
son with others of their time. heir application of diversiied technologies
permitted relatively intensive exploitation of varying environments. hey
developed elites at local but not at imperial scales. Nonetheless, the disease
environment kept mortality at a high level. he expansion of slave trade and
a concomitant expansion in domestic slavery, in region ater region of the
continent, transformed and further marginalized this socioeconomic sys-
tem but did not destroy it. It may be that the same pattern has been propa-
gated into the present, though it cannot yet be said by what mechanisms the
system was reproduced over time. As this pre-1950 demographic pattern
clariies, Africa should be compared more systematically with tropical South
Asia and Southeast Asia, in both demographic and social terms. And Africa
before 1950 should be compared systematically with Africa since 1950.

References
Barth, Fredrik, ed. (1994). Ethnic Groups and Boundaries: he Social Organization of
Culture Diference. Oslo: Pensumtjeneste.
Biraben, Jean-Noël. (1979). “Essai sur l’évolution du nombre des homes.” Population,
34: 13–25.
Boserup, Ester. (1965). he Conditions of Agricultural Growth: he Economics of Agrarian
Change under Population Pressure. London: G. Allen & Unwin.
Caldwell, John C. (1982). “Comment on Manning, ‘he Enslavement of Africans: A
Demographic Model.’” Canadian Journal of African Studies, 16: 127–30.
Caldwell, John C. and homas Schindlmayr. (2002). “Historical Population Estimates:
Unraveling the Consensus.” Population and Development Review, 28: 183–204.
Carr-Saunders, A. M. (1936). World Population: Past Growth and Present Trends. Oxford:
Clarendon.
Durand, John D. (1965). “World Population Estimates, 1750–2000.” In Proceedings of the
World Population Conference, Belgrade, 30 August – 10 September 1965, II: 17–22.
New York: United Nations.
Eltis, David, ed. “Voyages: he Transatlantic Slave Trade Database.” http://www
.Slavevoyages.org/tast/index.faces.
Gemery, Henry A. and Jan S. Hogendorn. (1974). “he Atlantic Slave Trade: A Tentative
Economic Model.” Journal of African History, 15: 223–46.
Hopkins, A. G. (1973). An Economic History of West Africa. London: Longman.
Ilife, John. (1987). he African Poor: A History. Cambridge: Cambridge University
Press.
Kuczynski, R. R. (1934). “Population.” In Encyclopedia of the Social Sciences, Vol. 12,
edited by E. R. A. Seligman and A. Johnson, 240–8. New York: Macmillan.
(1944). “World Population Problems.” International Afairs, 20: 449–57.
150 Manning

MacGafey, Janet. (1991). he Real Economy of Zaire: he Contribution of Smuggling

& Other Unoicial Activities to National Wealth. Philadelphia: University of
Pennsylvania Press.
Maddison, Angus. (2001). he World Economy: A Millennial Perspective. Paris:
Organisation for Economic Co-Operation and Development.
Manning, Patrick. (1981). “he Enslavement of Africans: A Demographic Model.”
Canadian Journal of African Studies, 15: 499–526.
(1982a). Slavery, Colonialism and Economic Growth in Dahomey, 1640–1960.
Cambridge: Cambridge University Press.
(1982b). “Response to Comments by Caldwell and Inikori.” Canadian Journal of
African Studies, 16: 138–9.
(1990). Slavery and African Life: Occidental, Oriental, and Atlantic Slave Trades.
Cambridge: Cambridge University Press.
Manning, Patrick and Scott Nickleach (forthcoming). African Population, 1650–1950:
he Eras of Enslavement and Colonial Rule.
McCann, James. (2005). Maize and Grace: History, Corn, and Africa’s New Landscapes,
1500–1999. Cambridge, MA: Harvard University Press.
McEvedy, Colin and Richard Jones. (1978). Atlas of World Population History. New York:
Facts on File.
McKeown, Adam. (2004). “Global Migration, 1846–1940,” Journal of World History, 15:
155–89.
Miller, Joseph C. (1982). “he Signiicance of Drought, Disease and Famine in the
Agriculturally Marginal Zone of West-Central Africa.” Journal of African History,
22: 17–61.
Miracle, Marvin P. (1966). Maize in Tropical Africa. Madison: University of Wisconsin
Press.
Patterson, Orlando. (1982). Slavery and Social Death: A Comparative Study. Cambridge,
MA: Harvard University Press.
Peukert, Werner. (1978). Der Atlantische Sklavenhanden von Dahomey, 1740–1797.
Wiesbaden: Steiner.
Portes, Alejandro, Manuel Castells, and Lauren A. Benton, eds. (1989). he Informal
Economy: Studies in Advanced and Less Developed Countries. Baltimore, MD: Johns
Hopkins University Press.
Pryor, Frederic. (1977). he Origins of the Economy. New York: Academic Press.
Tabutin, Dominique and Bruno Schoumaker. (2002). “he Demography of Sub-
Saharan Africa from the 1950s to the 2000s. A Survey of Changes and a Statistical
Assessment.” Population, 59: 457–555.
Wigboldus, Jouke. (1986). “Trade and Agriculture in Coastal Benin c. 1470–1660: An
Examination of Manning’s Early-Growth hesis.” A.A.G. Bijdragen, 28: 299–380.
Willcox, Walter F., ed. (1931) “Increase in the Population of the Earth and of the
Continents since 1650.” In International Migrations, vol. 2, Interpretations, edited
by. W. F. Willcox, 33–82. New York: National Bureau of Economic Research.
Williamson, Kay and Roger Blench. (2000). “Niger-Congo.” In African Languages:
An Introduction, edited by Bernd Heine and Derek Nurse, 11–42. Cambridge:
Cambridge University Press.
 PART II

CULTURE, ENTREPRENEURSHIP,
AND DEVELOPMENT
 5

Redistributive Pressures in Sub-Saharan Africa:

Causes, Consequences, and Coping Strategies

Jean-Philippe Platteau

5.1 INTRODUCTION
A widespread idea in the development economics literature of the past
decades is that personalized relationships and social networks play an
important role in the spontaneous evolution of institutions required for
rapid growth, such as contract enforcement mechanisms and property
rights (see, e.g., Aoki 2001; Bates 2001; Fafchamps 2004; Greif 2006; Hayami
and Kawagoe 1993; North 1990; Platteau 2000). An intriguing question,
which is rarely raised, is how to reconcile such a view with the well-known
Weberian/Marxian thesis that the development of market societies is based
on the rising autonomy of the individual and his (her) emancipation from
the matrix of erstwhile social agencies.
his question has several aspects. Although personalized relationships
allow good circulation of information, they do not necessarily create the best
conditions for the impartial meting out of sanctions against rule violators,
thereby undermining the efectiveness of reputation mechanisms (Baland,
Somanathan, and Wahhaj 2013; Hof and Sen 2005; Platteau 2000, chap-
ter 6). Moreover, if traditional codes of honor prevail, even the informational
advantage of personalized relationships can be disputed: thus, a lot of secrecy
surrounds economic transactions if, in the event of rule breaking (think of
loan delinquency), people are wary of not bringing shame on resident fami-
lies with whom they are in continuous interaction, or if they are eager to con-
ceal acts of malfeasance of which they have been the victims because such
a revelation would undermine their reputation as strong persons (Platteau
2000, chapter 6). A cheated person may also refrain from leaking out the
information if he (she) does not wish to protect other community members,
say, because he (she) wants them to be harmed in the same way as he (she)
has been. Finally, to the extent that norms based on values of generosity,

153
154 Platteau

sociability, and hospitality tend to prevail among closely related persons, rel-
atively successful members of social groups or networks would face internal
pressures to redistribute their incomes, which would create strong disincen-
tives to apply efort, take risks, and accumulate capital.
he way out of this dilemma seems evident: kinship or place of origin
should play little or no role in the formation of business and other social
networks. his is what Arne Bigsten and his colleagues (Bigsten et al. 2000)
actually found in their survey of indigenous trading and manufacturing
irms located in a number of African countries. Note, however, that the
implied process of emancipation still requires that entrepreneurial indi-
viduals are strong enough to withstand the redistributive pressures that are
likely to be exerted from without by their kith and kin.
his is precisely the problem addressed in this chapter with special refer-
ence to lineage-based societies, from sub-Saharan Africa in particular. We
argue that, indeed, personalized relationships are not the magical solution
to contract enforcement problems that they are oten portrayed to be. In the
context of newly emerging economic opportunities, redistributive norms
prevailing in the sphere of personalized relationships tend to undermine
economic incentives and to slow growth. hese pressures can nevertheless
be overcome, at least partly, but always at a cost. Finally, we explore the
diicult question of how to explain the pervasiveness and persistence of
sharing norms in African countries compared with countries belonging to
other continents.
Our endeavor clearly belongs to a strand of literature that stresses the
role of culture for development and that has been out of fashion for a long
time because of widespread criticisms against the excesses of the so-called
modernization theory. his is perhaps the time to adopt a more nuanced
view that recognizes that cultural factors cannot be assumed to be strictly
endogenous − various ways lead to modernization, and they are inlu-
enced by historical and cultural factors − or that, at least, they are not so
malleable as to rapidly evolve when the economic environment changes;
social norms and codes of behavior are somewhat sticky (Basu 2000; Basu,
Jones, and Schlicht 1987; Kuran 2011; Platteau 2011; Platteau and Peccoud
2011). he analytical perspective followed in this chapter does not rest on
the view that African societies face cultural impediments that doom them
to remain economically underperforming. Instead, it draws attention to
speciic obstacles that these societies encounter in their transition to what
Simon Kuznets (1966) once called “modern economic growth,” and that
therefore shape their speciic development path. In particular, it also helps
us solve the following puzzle: Why is it that Senegalese people, for example,
 Redistributive Pressures in Sub-Saharan Africa 155

can be quite successful entrepreneurs in New York while being much more
discreet in their own countries?
he remainder of this chapter is organized as follows. In Section 5.2, the
various forms that acts of git giving can take in lineage-based societies are
reviewed with a view to highlighting the speciic nature of the transfers that
redistributive norms govern. In Section 5.3, attention is then focused on
the nature of the sanctions that support these norms, while in Section 5.4 it
shits to the rationales underlying them. Section 5.5 then proceeds by dis-
tinguishing among the various disincentive efects redistributive pressures
cause. In Section 5.6, individual strategies available to cope with these pres-
sures are examined, whereas, in Section 5.7, it is the lack of broad-based
forces to counteract them in the speciic case of sub-Saharan Africa that
is discussed in the light of the recent history of that region. Section 5.8
concludes.

5.2 GIFT GIVING AND REDISTRIBUTIVE NORMS

Gits can obey diferent motives. A kind of git that has received a lot of
attention among social scientists, especially economists, during the past
decades, are reciprocal, state-contingent transfers made on a voluntary
basis (see, e.g., Coate and Ravallion 1993; Dercon and Krishnan 2000;
Fafchamps 1992; Kimball 1988; Ligon, homas, and Worrall 2002; Morduch
1999; Townsend 1994; Udry 1994). hey fulill the function of informal
mutual insurance and involve a quid pro quo inasmuch as, in the words of a
renowned anthropologist, “everybody is thereby insured against hunger . . .
he who is in need today receives help from him who may be in like need
tomorrow” (Evans-Pritchard 1940, 85).
Acts of symmetrical git giving do not necessarily answer the need for
protection against various types of risks in highly uncertain environments.
As many anthropologists have stressed, they may also help to cement social
ties and fashion alliances with strangers who are thereby transformed into
surrogate relatives. hese alliances or friendship relationships may fulill dif-
ferent objectives, among which the building of trust required for economic
exchanges is oten mentioned. Masahiko Aoki (2001) has thus shown rigor-
ously that the git may serve both as a signal to communicate a willingness
to cooperate to a potential partner, and as a commitment device because,
once the git has been made, the donator’s interest is to abide by a contract
provided that the partner also does it.
In these two instances, the git and the counter-git are a manifestation of
enlightened self-interest, or of selishness with foresight. In other words, they
156 Platteau

appear to be motivated by exchange rather than by altruism. Furthermore,

they are made between equals, that is, persons occupying similar positions
in the social structure and/or endowed with similar amounts of wealth.
When we say that redistributive or sharing norms enjoin prosperous per-
sons to share the fruits of their economic success with their kith and kin,
we obviously refer to a diferent kind of git: it is now wealthier people who
transfer income to poorer ones and, to the extent that being more wealthy
is not a transient state, a git is unlikely to be reciprocated. Moreover, if the
git is voluntary, one may suspect it to be driven by noneconomic or non-
tangible considerations that include symbolic advantages. When income
is thus traded against social prestige or political power, the material git is
reciprocated by a demonstration of esteem, loyalty, allegiance, subordina-
tion, and devotion.
Git exchanges of the latter sort can be termed asymmetrical, because
they take place between persons of unequal wealth, rank, or social status.
Insofar as generosity and hospitality are highly praised behaviors in tra-
ditional rural communities, successful individuals gain social prestige and
esteem when they redistribute their surplus. here may be more at play,
however, than the “warm glow” efect that git giving sparks of among the
attendance when made in public (Andreoni 1990). As we have learned from
the works of many anthropologists (Belshaw 1965; Bourdieu 1990; Levi-
Strauss 1969; Malinowski 1922; Mauss 1925; Polanyi 1944, 1968; Sahlins
1960, 1963, 1974), indeed, the whole point of the game of asymmetric
exchange may actually consist for the dominant party of making sure that
the tangible beneits or services that he renders (including insurance against
the risk of hunger) can never be (fully) repaid. Being in his debt on the mate-
rial level, the donee inds himself compelled to return the favor on another
level, in ways that inluence the donor’s rank or status. Subordination is
created and perpetuated because the obligation to reciprocate, which is a
burden, cannot be relieved by means of a return git equivalent to the initial
git (Ofer 1997, 455). Upon this reading, social prestige and political power
originate in asymmetric git exchanges. Economists have recently enriched
this approach by highlighting the conditions under which unequal rank or
power positions may be permanently established through asymmetric git
exchange when a git brings pride, social esteem, or power to the donor
and shame or subordination to the donee (Gaspart and Seki 2003; Platteau
and Sekeris 2010; Platteau and Seki 2007). his kind of asymmetric git
exchange is characteristic of class-based agrarian societies dominated by
patron-client ties, such as are commonly found in large parts of rural Asia
(Breman 1974; Scott 1976).
 Redistributive Pressures in Sub-Saharan Africa 157

Transfers from rich to poor are also voluntary when they are governed
by the desire to follow well-internalized norms of social solidarity. A cen-
tral feature of lineage-based societies is, indeed, that other-regarding val-
ues are inculcated in every individual from early childhood and carefully
nurtured all throughout his or her life; hence Albert Hirschman’s (1958)
characterization of them as “group-focused” societies.1 Unlike what we ind
in the “Invisible Hand” doctrine, which is at the root of the market-based
view of the economic system, the presumption is that the pursuit of selish
ends runs counter to the collective good in the general case. Evidence to
the contrary must therefore be adduced before exceptions to this rule can
be granted.
Other-regarding values and codes of conduct are stressed not only in
each and every possible moment of ordinary life, but also under excep-
tional circumstances. Rituals and ceremonies during the course of which
the unity and harmony of the group are strongly asserted and celebrated
provide these exceptional moments (see, e.g., Godelier 1974). hey are
privileged occasions during which social norms and values stressing the
collective good and the necessity for the individual to sacriice his self-
interest for the beneit of the community are put in the foreground and
intensely felt through dances, songs in unison, and the sharing of abundant
food and drinks (see, e.g., Gyekye 1996, 36). During such events, all par-
ticipants are called on to manifest regularly and openly their concern for
the collective good and their willingness to contribute to it. To the extent
that such internalization of other-regarding values through nurturing and
rituals succeeds, any reference to self-centered considerations tends to be
suppressed from consciousness so that community members can think of
themselves as moral individuals (Polanyi 1977, 60; Wright 1994).
Internalization processes are never complete, however, and this is why
external reward and sanction mechanisms are needed to complement the
work of collective rituals and education. Sanctions meted out to individu-
als who promote their own interest at the presumed expense of the group
are probably becoming more important when a society is opened to new
economic opportunities that certain individuals are more prompt than oth-
ers to seize on. It is then, according to Arthur Lewis, that a member of an

1
In Gambia, for example, the concept of badingya, which represents harmony, cooperation,
and shared progress (or shared decline), is continuously emphasized in contrast to the
concept of fadingya, which refers to selish ambitions and competitiveness. Individualistic
proclivities are accepted only to the extent that they are considered necessary for the
advancement of the group as a whole (Von Braun and Webb 1989, 515–16; see also Sylla
1994, 170–6, for the Wolof in Senegal).
158 Platteau

extended family system whose income increases “may be besieged by corre-

spondingly increased demands for support from a large number of distant
relations” (1955, 114). Transfers that are driven by redistributive norms in
the sense meant in this chapter are arguably made under some degree of
coercion. Coercion is directly exerted when, without warning, individuals
settle down in the house of a prosperous relative so as to obtain free lodging
and boarding for an indeterminate period of time; or when they drop some
of their children into the house of such a relative, again for an indeinite
time, without bothering to ofer him (her) a compensation or leaving him
(her) the choice to refuse to adopt these children2; or when they consistently
evade repayment of their debts to close acquaintances. In many instances,
redistributive norms are backed by severe sanctions, foremost among them
ideological intimidation. he nature of these sanctions is highlighted later
in this chapter with a special focus on countries of sub-Saharan Africa.

5.3 SANCTIONS IN SUPPORT OF REDISTRIBUTIVE NORMS

Norms that enjoin economically rising individuals to share their surplus
with their kith and kin are justiied by the belief that success is due to luck
rather than to personal efort, risk taking, and talent. While transient luck
is believed to result from the ordinary course of natural events, persisting or
exceptional luck is attributed to the obscure manipulation of supernatural
forces. his sort of accusation is extremely grave inasmuch as witchcrat is
thought to cause the treacherous exploitation of other villagers and to jeop-
ardize the survival of the entire group. Because they can easily degenerate
into active hatred unless assuaged by corrective gits, the aversive emotions
of envy and jealousy are instrumental in bringing about a redistribution of
income between (presumably) lucky and unlucky individuals. he hatred of
fellow community members is something successful people try to suppress
because it is liable to cause psychological oppression and physical harm
(Jewsiewicki 1993).
In the words of an inhabitant of a South African village: “You build
a beautiful house. hey say that you are a witch. hey come and burn it
down. hey say they are burning a witch’s house. You buy a car. hey come

2
he author has collected stories of relatively well-to-do West Africans who upon their
return from business travel found that the number of their “children” had suddenly
increased. Exploiting their temporary absence, parasitic relatives (half-brothers or broth-
ers-in-law, for example) had come to their house and entrusted the children whom they
wanted to rid themselves of to the wife before disappearing to a rather distant location
(typically, in a neighboring country).
 Redistributive Pressures in Sub-Saharan Africa 159

and shoot you, just because you are working and they are not” (Golooba-
Mutebi 2005, 947). he main problem with witchcrat accusations is that
they are impossible to shake of because they are backed by a system of
self-reinforcing beliefs and sanctions: “Even people who might be con-
vinced of a person’s innocence dare not come to his or her defense, lest
they too be accused. Once accused, therefore, a person is shunned by all
villagers, save for the most committed friends and kin” (Golooba-Mutebi
2005, 953).
Witch beliefs thus constitute a conservative force that helps sustain a
rough egalitarianism, or a given distribution of ranks and status, by act-
ing as a check on undue individual efort (homas 1973, 643–4, cited in
Elster 1989, 261). As suggested by the example from South Africa, a rich
man knows that if he is stingy with his relatives or fails to dispense gener-
ous hospitality to all, he is likely to be spoken of as a witch and to arouse
suspicion that he has accumulated riches by robbing others (Elster 1989,
260). As Evans-Pritchard (1937) explained on the basis of his research
on the Nuer tribe of Sudan, the evil “soul” of witchcrat is set to work by
antisocial feelings such as envy, spite, jealousy, anger, and hatred. If these
feelings are absent, the witchcrat remains “cool” and harms no one (see
also Badini 1994, 141, for the Mossi of Burkina Faso). In rural communi-
ties of Rwanda (prefecture of Kibuye in the western part), according to a
more recent account, only gits can appease jealous feelings (ishari) and
dismiss accusations of misplaced pride (ishema), which cause fear in pros-
perous individuals. Such fear is nurtured by the all-pervasive belief in poi-
sonings (uburozi) that designate the many forms of baleful spells (Lame
1996, 182).
In Njombe district (Tanzania), prosperous traders and farmers were said
to use secret powers to enrich themselves, in some cases by scattering the
blood and lesh of victims on their farms so as to obtain plentiful harvests.
he suspicion is that entrepreneurial people are ready to kill their own chil-
dren and their relatives in the quest for riches. he truth is that individuals
illed with greed, envy, hatred, and deep jealousy are busy day and night
searching for poisons able to kill wealthy people so as to appropriate or
inherit their wealth (Giblin 2005a, 198, 200).
Witchcrat accusations therefore appear as a powerful means of ideolog-
ical intimidation aimed at suppressing deviance in general and economic
self-advancement in particular. It is within the kinship group that witchcrat
is summoned with special force, because jealous feelings are more intense
as relations between people are closer: the gravest accusation against a suc-
cessful individual is that he is ready to sell his parents in order to accumulate
160 Platteau

riches.3 Repressive measures are also observed in tribal societies built on a

ranked hierarchy of constituent groups and possessing a strongly central-
ized structure focused on the chief. According to many reports about life
in African villages, the chief or the elders do not easily accept that the rank
and ile rise above them by acquiring new or ancient symbols of wealth and
status. All eforts to accumulate such symbols are unavoidably viewed as
conscious attempts to overturn the existing social order (in which chiefs
and elders are richer but also stand as the main agents of redistribution),
and they immediately arouse insinuations or open accusations of sorcery
(Bruce 1993, 39–40; Colin 2004, 284–5; Geschiere 1995, 210; Peters 1994,
32). In Botswana, for example, someone a Tswana chief considered a dan-
gerous competitor was accused of witchcrat and stripped of his property.
In such cases, Pauline Peters argues, “the chief was able to justiiably ‘eat’ the
miscreant” (1994, 32). Witchcrat accusations against social superiors are
unthinkable, however, inasmuch as the latter are thought to be protected by
powerful spirits able to launch efective counter-sorcery (Evans-Pritchard
1964, 243; Geschiere 1995, chapter 4).4
Evidence from anthropological writings converges to show that, far
from diminishing, witchcrat accusations and practices continue to play
an important role in repressing individual acquisitiveness and mobility in
African (rural) societies, and the belief in occult forces is very common
at diferent levels of education and among various religious ailiations
(Kohnert 1996, 1348). For instance, René Bureau speaks about “a mas-
sive resumption of witchcrat accusations” (1996, 123–4) in the coastal
area of Cameroon where the immense majority of those accused of being
balemba (sorcerers) have an enviable place, function, fortune, or inluence.
In Cameroon, South Africa, Benin, and other African countries, as many
new opportunities for economic diferentiation arise, successful people are
said to be sorcerers able to transform their victims into kinds of zombies
whom they force to work at night in their invisible plantations in order
to enrich themselves. he victims complain of waking up in the morning

3
Sometimes, however, the feelings of jealousy, rivalry, and covetousness, which give rise
to witchcrat, can nevertheless be considerably extended. For instance, they can afect the
competition among tribes, clans, and villages for getting development projects. he belief
is then that the tribe, clan, or village that wins a project does so because its spirits won a
decisive battle against the spirits of rival tribes, clans, or villages (Geschiere 1995, 197–8;
Harragin 2004, 310–11; Platteau 2000, 199–200).
4
In the Green Valley (South Africa), for example, witchcrat accusations reinforced rather
than challenged social privilege, because persons of relatively greater status and inluence
used them manipulatively to their own advantage (Niehaus 2001a, 198).
 Redistributive Pressures in Sub-Saharan Africa 161

feeling extremely tired (Ardener 1970; Ellis and Ter Haar 2004, 123; Fisiy
and Geschiere 1991; Geschiere 1994, 1995, 175, 192–8; Rosny 1981, 1992;
Rowlands and Warnier 1988).5
In Rwanda, and also in Tanzania, Uganda, and Mozambique, we hear
about the increasing incidence of spirit possession and poisoning threats
as a result of social and economic diferentiation or, to use the words of a
Rwandese traditional healer (cited in Migeotte 1995), of “recent changes in
economic conditions that arouse people’s spite since nobody accepts that
someone else surpasses him in wealth.” Rich individuals are accused of pos-
sessing obscure powers and entertaining privileged relations with the invis-
ible world, which enable them to appropriate wealth from others (Ellis and
Ter Haar 2004, chapter 5; Lame 1996, 183; Platteau 2000, 202; Rennie 1984,
180–2). James Giblin has reported in detail the diiculties encountered in
rural Tanzania by people who want to build a family business, and their
accounts are particularly insightful. Economic success is invariably consid-
ered the result of exploitation of fellow villagers and “the proof that one uses
witchcrat” (2005a, 199). hus, when in the late 1950s, an individual from
Njombe district obtained unusually abundant maize harvests from hybrid
seeds, people said that he used mkwego to draw maize away from others
(mkwego is a medicine used to capture for one’s own use another person’s
ability to prosper), which eventually compelled him to share seeds with
neighbors (Giblin 2005a, 198). Prosperous individuals know that behind
the blame of maliciously using magic and witchcrat lies the grave accusa-
tion that they have violated the obligations of kinship (Giblin 2005a, 199).
In South Africa, witchcrat practices seem to be lourishing more than
ever before, and social and economic diferentiation provides fertile ground
for suspicion and accusation “when the well-to-do fail to fulill the expec-
tations of poorer neighbours and kin for support.” In Tiko village, witch-
crat is feared by almost everybody, and it includes the use of poisons and
other dangerous substances or “medicines” that bear a special local name
(Golooba-Mutebi 2005, 939–44). Elsewhere, those named and killed as
witches were oten doing well − they had large numbers of cattle, or had

5
Among the Douala, people believe that witchcrat powers are acquired in an ultrasecret
society admittance to which is conditioned on the killing of a person, preferably a close
relative such as a mother or a son. he most common way to harm people consists of
removing their “soul of life” and to reduce them to slavery ater their funeral. he victims
then work for the sorcerer on the Coupé and Manenguba mountains. he sorcerer uses his
lying powers to go to his secret plantation and collect the fruits of his slaves’ labor (Bureau
1996, 107–8).
162 Platteau

worked in the cities and earned a pension for their old age − jealousy
appearing as the most common motivation for the killings (Niehaus 2001a,
2001b). Increasingly, young people, who were not traditionally allowed to
accuse others of witchcrat, make accusations against older, typically richer
persons (Ellis and Ter Haar 2004, 152).
he following accounts, one from Ramon Sarro (2009) for the Upper
Guinea Coast (Guinea Conakry), and the other from Angelique Haugerud
(1993) for central Kenya, provide two additional, vivid illustrations of the
idea that witchcrat accusations have a tendency to increase with the rising
economic diferentiation that accompanies growth:
A man asked his classiicatory sister, who was living in Kamsar, to bring him new
shoes from that city. She refused, and a few days later she died in Kamsar. People in
the village thought she was killed by her envious brother. . . . Conrad . . . [worked] for
the Compagnie des Bauxites Guinéennes. He lived in Kamsar for several years, but
always felt that his relatives in the village, only iteen kilometres away, were very
demanding, asking him far too oten for money and goods from Kamsar. He opted
to move further away to feel more at ease and therefore went to Conakry, where he
was employed at the harbour. . . . He was very aware of the envy that his status could
awake in the village . . . every time he went to the village to spend some time build-
ing the house he took a fair largesse of money and presents to share among people
from his and his wife’s descent groups (most of the villagers helping him build the
house were his in-laws, and were doing it for free). . . . In 1999, when the house was
virtually inished . . . Conrad died suddenly in Conakry. Despite all his eforts to
avoid arousing envy while building his house, his death provoked all sorts of gossip
and even a serious allegation of deser [witchcrat] against a young member of his
own descent group who was accused of having killed his classiicatory father out of
sheer envy. (Sarro 2009, 156, 159)
As opportunities to accumulate wealth grew in central Kenya, so too did the stakes
for individual displays of “generosity” as opposed to “selishness” . . . enmities have
arisen due to wealth. Several people spoke of the narrowing range of kin among
whom one shares goats on ceremonial occasions, such as those associated with
marriage and childbirth. . . . And an Embu man explained his decision not to open
his own small shop, saying that such businesses oten failed within one year, that
customers asked for credit and called the owner “proud” if he or she refused to
extend the credit, and that sorcery against the owner might follow. Less well-of
individuals also termed “proud” wealthy individuals whom they believed did not
assist others enough in inding jobs or places in secondary schools. . . . Treating visi-
tors well includes preparing food for them rather than hoarding it for one’s own
family. Selishness and social isolation, in contrast, may carry associations of witch-
crat and sorcery, and at the very least invite negative gossip. Longstanding norms
of sociality and patronage include the expectation that the less well-of will seek
assistance from the wealthy, and that the latter indeed will respond with good will
to such requests. (Haugerud 1993, 133–4)
 Redistributive Pressures in Sub-Saharan Africa 163

5.4 THE RATIONALE BEHIND REDISTRIBUTIVE NORMS

he oicial reason invoked to justify redistributive norms is that when an
individual becomes too successful economically compared to fellow group
members, he (she) is likely to harm them: individual prosperity, according
to this account, is bought at the expense of collective prosperity. Two mutu-
ally reinforcing rationales may underlie this assertion.
he irst motive arises from the need to protect the group against the
threat of erosion of its ability to shoulder risks. Indeed, if a prosperous indi-
vidual were allowed to stop making gits to fellow villagers, the size of the
informal insurance network formed by the community would be reduced
and the group’s ability to spread risks would be correspondingly dimin-
ished.6 Private wealth accumulation is perceived as antisocial behavior
precisely because it is an attempt to break away from traditional solidarity
networks, and the exit of the (best) members would make the rest of the kin
group worse of (Fafchamps 1992, 160–3; Hof and Sen 2005, 173; Platteau
1991, 160–1, 2000, chapter 5). According to Guy Hunter, we are dealing
with “levelling societies, in which attempts by equals to gain individual
advantage are constantly suspected and bitterly resented,” and, at the root
of this suspicious atmosphere, lies the “fear that the fundamental security
of the village will slowly be lost if one individual ater another can reach a
platform of prosperity from which he might not need the help of the com-
munity and could therefore excuse himself from helping them” (Hunter
1969, 40).
Following the logic of this insurance-based argument, the establishment
of sharing norms can be rationalized as the outcome of individual decisions
by selish community members. Assuming that idiosyncratic shocks are suf-
iciently severe or frequent and randomly distributed (and their efects are
possibly ampliied by collective shocks), and also assuming that community
members are equally (or not too unequally) endowed in the initial situa-
tion, they will not only decide to join a mutual insurance scheme based on
contingent reciprocal transfers, but they may also support the laying down

6
Let us assume that individual incomes, which are independent and have identical vari-
ances (=σ2), are entirely redistributed through a perfect pooling scheme. he variance of
individual incomes ater pooling is then simply equal to σ2/n, from which it is evident that
any reduction in the size of the pooling group (=n) increases the volatility of incomes for
those who remain within it. Under our assumptions, indeed, the variance of aggregate
income (=Y) is equal to the sum of the individual variances (=nσ2), and the variance of
individual incomes ater pooling is, therefore, Var(Y/n)= nσ2/n2= σ2/n.
164 Platteau

of social norms. his implies that they will vote in favor of rigidly enforced
rules aimed at preventing members from quitting when favorable circum-
stances make it worthwhile for them to do so (that is, when the expected
payof from participation becomes smaller than the payof from nonpartic-
ipation). his is because, ex ante, they ignore what their income trajectory
will be and they therefore fear the breaking down of the risk-pooling mech-
anism over time. “Forced mutual help,” as Raymond Firth (1951) called it,
is then used to maintain the status quo, which has the potential efect of
discouraging exceptional individual performances on the grounds that they
can only take place “at the expense of other members and of the cohesive-
ness of the group” (Hirschman 1958, 23).
he second motive underlying redistributive norms is of a quite difer-
ent order: because the efort of any individual to improve his lot generates
positional externalities that negatively afect the welfare of fellow villagers,
redistributive norms that enjoin economically successful individuals to
share their surplus appear as a form of taxation designed to curb positional
race for status. his may be thought of as a desirable outcome if eforts
to improve one’s lot are mutually ofsetting (everyone wants to “keep up
with the Joneses”), and ineicient equilibria arise precisely because invest-
ment in status enhancement is more attractive individually than collectively
(Congleton 1980; Frank 1985, 1998).
hat the insurance-based explanation is insuicient to account for the
pervasive presence of redistributive norms in lineage-based societies is evi-
dent from the aforementioned fact that the hierarchy of ranks may not be
called into question. Typically, while one of their functions is to redistribute
wealth toward the needy (see, e.g., Malinowski 1922; Sahlins 1960, 1963,
1972, the chief or the elders do not accept that commoners rise above them
by acquiring old or new symbols of wealth and status (like cocoa planta-
tions, mechanical devices, or roofs made of corrugated iron). All eforts to
accumulate such symbols are unavoidably viewed as conscious attempts to
compete with traditional leaders and to overturn the existing social order.
As a consequence, they are strongly condemned and the prosperous com-
moner is immediately coerced into handing over to the chief his newly
acquired riches. hus, in a village Geschiere studied in Cameroon, a man
who accumulated wealth and used part of it to buy a new local suit (a beau-
tiful boubou) was severely rebuked by the local chief, to whom he was forced
to hand over his new garment. his was done without any compensation,
just on the ground that “he should not have displayed such a luxury while
the chief had to go without it” (Geschiere 1995, 210, note). If the new sym-
bol of wealth is a productive asset, the successful individual is badgered to
 Redistributive Pressures in Sub-Saharan Africa 165

either give it away to the chief or to use it for the latter’s beneit, for example,
by using his tractor to plow the chief ’s land (for other examples, see Platteau
and Abraham 2002, 114–18 and Platteau and Abraham 2004).
he young generations oten resent the situation as oppressive and may
be prompt to seize any opportunity to oppose it. his is one of the cen-
tral lessons to be learned from Ramon Sarro’s analysis of the rebellion that
spread through the villages of the Baga-speaking people of coastal Guinea
during the colonial period. Initiated by a Muslim preacher (Asekou Sayon),
this iconoclastic movement aimed at the destruction or abandonment of
many customary ritual practices and objects perceived as negative for the
development of the community. Educated people and war veterans who
had been abroad (and many of whom converted to Islam) identiied these
animist practices with the despotic rule of local chiefs and elders, itself rein-
forced by the French colonial policy of appointing powerful canton chiefs.
Chiefs and elders were, indeed, perceived as doing nothing “but drink the
palm wine that youths had to tap for them” (Sarro 2009, 94), and custom-
ary ceremonies (the so-called masquerades), which involved a lot of hard
work for the latter, were deemed far too numerous and costly. Under these
conditions, writes Sarro, “it is not surprising that Islam, being a religious
discourse that not only forbade masquerades and sacred bushes but also the
tapping and drinking of palm wine, was increasingly attractive for the strata
of society that felt (and were) abused by the elders” (2009, 95). Interesting,
we are also told that the elders insisted on receiving gits (including mar-
riage payments) in the form of palm wine instead of money because palm
wine being given in public and in large quantities, it could not be unduly
appropriated by the recipient but had to be shared with other members of
the ruling strata.
To sum up, lineage-based societies are prone to resist any diferentiation
process whereby relative status positions are modiied. Social reshuling
is perceived as a dangerous force susceptible of jeopardizing the fragile
social equilibria typical of small-group societies anchored in highly per-
sonalized relationships. As pointed out earlier, the redistributive pressures
through which diferentiation is repressed are oten backed by powerful
sanctions that include social pressures, constant harassment, and the use
of efective mechanisms of ideological intimidation, most notably witch-
crat accusations and practices. A direct implication of this analysis is that
the phenomenon highlighted should not be restricted to sub-Saharan
Africa, but be observed in all lineage-based societies dominated by per-
sonalized relationships. his prediction is supported by evidence from
Asian villages, and even from Europe. In Malaysia, for example, James
166 Platteau

Scott noted that the rich helped the poor “less from a spirit of liberal-
ity than as a response to the palpable pressures their neighbours and kin
brought to bear upon them” (1985, 192). With reference to Europe, the
presence of redistributive norms in Balkan villages is attested by the old
Bosnian saying according to which: “If the whole society is prosperous,
each of its members gains from it, but when one individual is too power-
ful, he harms his fellow members, whether he likes it or not. When a hand
is too big, swollen and painful, it is because it has been bitten or invaded
by bad spirits. In order to restore the health of the hand’s owner, the swell-
ing must be reduced or the hand will need to be cut of ” (Karahasan 2000,
191, our translation).
he historical experience of Western Europe itself shows that redistrib-
utive pressures and the accompanying witchcrat accusations have become
particularly widespread during the phase of transition from the egalitarian
mutually dependent village community of preindustrial times to the more
individualistic and socially diferentiated structure of market-based socie-
ties. According to a British historian, “one of the major reasons for the rise
of witchcrat accusations in the sixteenth century was the tension caused by
a conlict between a traditional ethic of mutual responsibility and charity,
the norms of self-contained and subsistence villagers, and the new acquisi-
tive and individualistic spirit of capitalism and Protestantism” (Macfarlane
1978, 59).
Before looking at the various disincentive efects redistributive norms
cause, it is worth emphasizing that the idea of redistributive pressures
grounded in feelings of envy toward successful individuals has strong
micro-theoretic foundations. he idea is that people’s utility is inlu-
enced not only by absolute income levels, but also by their relative posi-
tion in the (local) distribution of income, thus causing preferences to be
interdependent (see Veblen, 1899, and Duesenberry, 1949, for the irst
formulations of this idea).7 he component of the utility function that
relects the inluence of an external reference point is interpreted as the
“status return” from income, or the positional or conspicuous consump-
tion aspect of income. Andrew Clark, Paul Frijters, and Michael Shields
(2008) have recently concluded that the available evidence largely sup-
ports the hypothesis of relative income efects. In particular, works that

7
his idea, coupled with the fact that people also compare their present income with their
past income, is now used to explain the so-called Easterlin’s Paradox – signiicant increases
in real incomes in Western countries over the past ity years have not been accompanied
by corresponding rises in reported happiness levels (Easterlin 1974).
 Redistributive Pressures in Sub-Saharan Africa 167

use a geographic approach to reference groups – social comparisons are

made with people living in the same neighborhood − suggest that life sat-
isfaction is totally relative to income. In rural China, for example, accord-
ing to one survey based on ninety-two hundred households, 68 percent of
individuals see their village as their reference group, while only 11 percent
stated that their main comparison group consisted of individuals from
outside of the village. Moreover, respondents who say that their income
was much above the village average report far higher happiness than those
who say that their income was much below that average (Knight and Song
2006, cited in Clark et al. 2008, 107, 112).
Perhaps more convincing, a study using respondent-deined (rather
than assumed) reference groups reveals that individuals who associate
with higher earners need more money to describe their income as good
or adequate (Melenberg 1992, cited in Clark et al. 2008, 109–10). Another
source of evidence on the importance of social comparisons to others in
actual choice behavior may be found in the experimental economics liter-
ature, whether it deals with fairness (such as laboratory evidence on ulti-
matum games) or asks individuals to rank hypothetical outcomes. A series
of papers using the latter approach have adduced evidence of strong posi-
tional concerns over income, in that individuals say that they are willing
to give up absolute income to improve their relative position or gain status
(Clark 2008; Clark et al. 2008, 120).
As hinted at previously, some economists refer to this kind of efect to
advocate taxation of positional goods in all dynamic societies: because
changes in rank have no social beneit even though they yield private
beneits, and the additional beneit of positional goods to an individual
leads to overconsumption of such goods at society level to the detriment
of non-positional goods, the production of the latter type of goods needs
to be encouraged through the taxation of the former type. If the positional
good is equated to relative income and the non-positional good to leisure,
there is thus a ground for income taxation to promote leisure (Frank 1985;
Layard 2005).
he similarity between this position and the existence of redistributive
norms in African countries is plain: to the extent that individual preferences
are translated in utility functions in which absolute income and relative
position are substitutes, a possibility to exert pressures on dynamic indi-
viduals to repress their upward income mobility may be used by less efort-
prone or more-leisure oriented individuals even at the price of achieving
a lower income than they could have otherwise received (say, because the
growth generated by the eforts of the successful individuals may produce a
168 Platteau

trickle-down efect).8 In other words, the availability of redistributive norms

enables the individuals who do not want to forsake leisure to prevent the
sparking of a chain reaction of emulating eforts.

5.5 THE DISINCENTIVE EFFECTS OF

REDISTRIBUTIVE NORMS
he trade-of between absolute income and status, understood as a posi-
tional concern, is reminiscent of the well-known trade-of between equity
and eiciency highlighted in the economic literature. Excessive taxation of
the rich, by reducing eiciency too much, may result in lower incomes not
only for the rich, but also for the poor. Eiciency losses are the outcome
of various disincentive efects redistributive pressures cause. In the par-
ticular context of lineage-based societies, several such efects need to be
distinguished.
First, redistributive pressures discourage efort. To make matters simple,
just think of an individual who produces an output that is perfectly observ-
able and anticipates that his (her) income will have to be shared with n
kinsfolk. His (her) equilibrium level of efort will then be determined by
the equality between the marginal disutility or cost of efort and the mar-
ginal productivity of efort divided by n. he efort applied will therefore be
suboptimal, and the larger n the wider the gap between actual and optimal
efort levels. Note incidentally that this disincentive efect is formally iden-
tical to that produced in an alternative situation in which output is collec-
tively produced and perfectly observable, but individual efort is not. Joint
output is then divided equally among the n participants who apply the same
suboptimal amount of efort as that indicated earlier (under the assumption
of constant returns to scale). In this instance, a dynamic individual who
would like to improve his (her) lot would be prevented from doing so as a
result of the so-called moral-hazard-in-team problem.
Second, redistributive pressures discourage entrepreneurship and risk
taking. his is because these pressures operate in an asymmetrical manner:
if the investment project fails, the risk taker will be the only one to bear

8
hink of a runner who has to choose between the two following situations: (a) another
member of his team wins the race and transfers a (small) portion of the beneit earned to
him; and (b) nobody in the team wins the race and the prize ofered to the winner goes
to another team. If the runner has the sort of preferences highlighted here, he will prefer
(b) to (a): his absolute income is smaller than it could have been yet his relative position
has not deteriorated (for a development-related illustration, see Abraham and Platteau,
2004, 220).
 Redistributive Pressures in Sub-Saharan Africa 169

the burden of the ensuing loss, while, if the project is successful, the risk
taker will have to share the beneits with his or her kith and kin. Given a
certain degree of risk aversion, a dynamic individual will therefore refuse to
embark on a risky project that he (she) would have attempted in the absence
of redistributive norms. As a consequence, redistributive norms are likely
to discourage savings and investment or to stimulate comparatively safe
investments and invisible expenditures.
hird, redistributive pressures encourage misallocation of human
resources. he dynamic individual subject to such pressures is typically
compelled to hire relatives and friends when he (she) has reached an
employer position (whether in the state bureaucracy or the private sector),
a practice known as nepotism. here are two distinct reasons nepotistic
behavior gives rise to ineiciencies. On the one hand, it amounts to favorit-
ism toward relatives and friends that involves a cost in terms of the quality
of employees hired (Hof and Sen 2005, 175).9 On the other hand, moni-
toring labor eforts and sanctioning bad behavior are particularly diicult
for an employer or a manager who deals with people from his (her) own
fold. his diiculty is actually reinforced if the employees believe they have
a right to misbehave vis-à-vis a superior from whom they want to extract
more beneits on the grounds that he (she) has not been suiciently gener-
ous and “fair” with them (Platteau 2000, 313). he disciplining efect of the
reputation mechanism that is made possible by the repeated, personalized
nature of human interactions is then cancelled by the adverse efect of the
attendant sharing norms.
Note that the third problem does not apply only to the sphere of labor
relations, but to any kind of economic transactions in which a choice of
partner is involved. For example, if the dynamic individual is a banker or
a moneylender, or even a businessman, he (she) may be forced to extend
credit to kith and kin who are unreliable borrowers. An almost natural
experiment testifying to the importance of this nefarious efect was cre-
ated by the short war between Senegal and Mauritania in 1989. Tension
between the two countries led to the expulsion of all Mauritanian shop-
keepers from Dakar, the Senegalese capital. heir shops were spread all
throughout the city and were opened day and night to provide customers

9
If there is competition among job seekers who have redistribution obligations toward rela-
tives and those who do not, and if the former can be identiied by employers who would
bear the adverse consequences of their redistributive behavior (say, because the employee
to be hired is a senior worker expected to choose junior workers for teamwork), they
will sufer from discrimination in the modern sector labor market (Hof and Sen 2005,
175–7).
170 Platteau

with basic necessities. Ater their seizure by inhabitants of Senegalese ori-

gin, these businesses rapidly declined and became bankrupt until they were
taken over by their previous Mauritanian owners when they were allowed
to return to the country.
Two answers were given to me when I inquired into the reasons for the
failure of the Senegalese experiment in shopkeeping. First, the shops were
not opened in the evening and the night like they used to be when the
Mauritanians ran them. Second, redistributive pressures compelled the
Senegalese shopkeepers to sell on credit to their relatives and friends, and
the loans were typically not returned. A common rationale for not return-
ing overdue loans is that because the lender/shopkeeper is richer than the
borrower, the latter has a perceived right to consider the loan taken as a
(forced) git. By contrast, Mauritanian shopkeepers followed strictly the
rule of selling only against cash payments. Customers who did not com-
ply were always rebuked. his explains why the presence of Mauritanian
shopkeepers was both resented − they are doing well and are “selish” − and
felt necessary − their business is eiciently run and therefore provides a
valuable service.
hat such stories are not rare in sub-Saharan Africa is attested by the
failures of Kenyans and Ugandans in their running of shops and businesses
taken away from Indians. Regarding Kenya, we have the following account:
“Within a short period of time a good many of the African businessmen
who had taken over from the Asian shopkeepers had lost most of their cap-
ital and turned thriving shops into bars and lodging rooms instead. Others
were inviting back previous owners and partners” (Himbara 1994, 89). In
Congo-Brazzaville, the Popo ishermen from Benin, who were an active
and economically successful community operating in Pointe-Noire, were
harassed and, without any forward notiication, eventually expelled by the
Congolese government in September 1977 on the pretext that they did not
have their formal stay permits. heir outboard engines were coniscated on
the grounds that proper import authorizations were lacking. here quickly
ensued a severe lack of ish in the markets in town, which forced the gov-
ernment to reverse its decision and recall the Popo, who slowly returned to
Pointe Noire (Jul-Larsen 1993).
Finally, we may bear in mind the aforementioned account of Haugerud
regarding the pervasive presence of redistributive pressures in central Kenya
(see the citation provided at the end of Section 5.3), which implicitly testi-
ies that such pressures may easily cause losses of eiciency, bankruptcies,
and the forfeiting of growth opportunities. his is conirmed by another
 Redistributive Pressures in Sub-Saharan Africa 171

African analyst who writes, with the experience of Kenya in mind, that
“[African] entrepreneurs are expected to share their incomes with other
members of the extended family. hey are expected to come to aid in case
of inancial hardships and to employ relatives regardless of whether they are
eicient. hese considerations lead to a withdrawal of a substantial amount
of capital from an enterprise and its eventual failure” (Kamau 1965, cited in
Himbara 1994, 89).

5.6 INDIVIDUAL REDISTRIBUTION COPING STRATEGIES

Ordinary individuals are generally reluctant to start accumulating wealth
and deviate from customary norms of behavior. his is especially so if they
are afraid of being killed by envious local witches or of being accused and
harassed as witches themselves (because they have acquired their wealth by
using occult powers), a risk enhanced by the fact that the police are usu-
ally unable to investigate and punish witchcrat-related crimes because of
the local people’s refusal to reveal the names of the perpetrators (Golooba-
Mutebi 2005, 951).10 For those dynamic individuals who are strong enough,
emotionally, to break from the traditional society, two strategies that involve
distancing from the pressures of community life are nevertheless available:
migration and religious conversion. For those who do not wish to take such
a drastic step, a more defensive strategy consists of concealing incomes and
assets. We start by examining the latter possibility before considering the
other two, bolder strategies.11

5.6.1 Concealing Incomes and Assets

Concealing incomes and assets is the most evident way to escape demands
for gits emanating from relatives, friends, and neighbors. Yet it involves
transaction costs that may turn out to be signiicant. In Burkina Faso, for
example, aluent individuals may take wives and children outside of the
home to eat so as to eschew the obligation to cater to unwelcome visitors
(Platteau 2000, 210). In this instance, the cost of escaping redistributive

10
A household survey conducted in urban Ivory Coast revealed that households who
accommodate collateral kin dependents (deined as inactive people) are twice as wealthy
as households who do not (Laiglesia and Morrisson 2008). Unfortunately, that study does
not allow us to infer the direction of causality between wealth and accommodation of kith
and kin.
11
For the latter two subsections, I draw largely on Platteau (2009).
172 Platteau

pressures takes the form of additional expenditures on transport and

restaurants. he cost may also take the form of suboptimal allocation of
wealth resulting from the need to opt for relatively inconspicuous uses of
the savings accumulated. In Tanzania, we are thus told that better-of vil-
lagers avoid displays of wealth such as modern houses because the villagers
believe a successful farmer is maliciously using supernatural forces (Giblin
2005a, 199).
Unfortunately, very few systematic empirical studies exist document-
ing redistribution-coping behavior in the context of sub-Saharan Africa.
Two that I am aware of deserve close attention, one by Jean-Marie Baland,
Catherine Guirkinger, and Charlotte Mali (2010) on Cameroon, and the
other by Vincent Somville (2010) on Benin. While almost all the evidence
presented thus far on the existence of redistributive pressures or norms in
Africa has been drawn from the anthropological literature, these two stud-
ies have been produced by economists and ofer us indirect evidence of this
phenomenon.
he irst study is based on a ield survey of rural credit cooperatives
belonging to the Cameroon Cooperative Credit Union League. Looking at
the credit operations and deposits of the members of four cooperatives dur-
ing the years 2004–6, Baland and colleagues (2010) uncovered a puzzling
fact of what they term “excess borrowing” behavior. Around one-ith of the
loans taken from rural credit cooperatives are fully collateralized by sav-
ings available on the borrowers’ savings accounts, and the amount of sav-
ings available to the borrowing member is on average twice as large as the
amount of the loan. Still more intriguing is the fact that the behavior of these
excess borrowers is costly because they could save on substantial interest
payments by self-inancing their investment projects. he authors adduce
two kinds of evidence in support of the idea that the desire to escape redis-
tributive pressures lies at the heart of the puzzle. he irst kind of evidence
is qualitative and consists of the explanations invoked by the excess borrow-
ers themselves. heir statements can be summarized as follows: excess bor-
rowing, and the consequent need to fulill reimbursement obligations, are
a way to signal inancial diiculties to relatives in search of inancial assis-
tance. As a matter of fact, people believe that when an acquaintance takes a
loan, it is because he (she) has no money. he explanation is credible: if the
existence of savings accounts and credit contracts is common knowledge,
the amount of the cooperative members’ savings is unanimously hidden.
Members actually report showing their credit book to suspicious relatives
to prove their state of inancial duress.
 Redistributive Pressures in Sub-Saharan Africa 173

he second line of argument proceeds by dispelling competing explana-

tions of excess borrowing. hus, the authors show that there are no likely
indirect beneits resulting from such a behavior. In particular, it is not true
that the loan serves as a vehicle of risk diversiication for risky investment
projects under conditions of limited liability. he borrower’s savings can
be seized in the event of default, and there can be no risk sharing with the
lender owing to full liability. In addition, the authors do not verify that
excess borrowing is the outcome of time-inconsistent preferences, peo-
ple engaging in illiquid saving to protect their money from their future
selves while relying on credit to meet immediate cash expenses. In reality,
local cooperatives explicitly allow their borrowers to draw on their savings
accounts to pay back their loans.
In the second study, also based on irsthand data collection, Somville
(2010) examines the role of so-called daily collectors in Benin. hese oper-
ators, as their name suggests, collect small savings on a daily basis from
their clients and keep them safe. In exchange for this service, they charge
a monthly fee that is equal to the amount of one deposit. his fee is sub-
tracted from the savings recovered at the end of each month by the client:
amounting to 3.3 percent, it is equivalent to a negative annual interest rate
of 54 percent. How to account for the people’s willingness to use the ser-
vices of such costly operators? One obvious answer is their desire to put
their money in a safe place so as to protect it against the risk of loss, thet,
or other hazards. According to the author, however, this explanation does
not stand up to the scrutiny of the facts because it does not explain inter-
individual variations in the use of daily collectors.
he need for protection against claims by household members and
acquaintances is the motivation that appears to best match the two years
of panel data collected in 2004 and 2006 in three diferent districts of the
outskirts of Cotonou, the country’s capital city. In particular, econometric
testing bears out the hypothesis that, other things being equal, individuals
use the services of a daily collector to contribute less to the household pub-
lic goods and increase their own private consumption: men make smaller
contributions to joint household expenditures such as electricity, house
repairing and maintenance, and school expenses, while women contribute
less to school expenses. Both men and women spend more on personal
clothing when using the service. It also appears that these users transfer less
money to those presenting requests considered as illegitimate, and more
money to recipients of their own choice. Results thus show that women
who use the services of a collector transfer less money to their husband
174 Platteau

but more to their children and their friends. As for men, they transfer less
money to their spouse, their children, their friends, and the members of
their own parish.12
In contrast to these two empirical studies, Jakiela and Ozier (2011) have
conducted a controlled laboratory experiment in western rural Kenya to
measure the economic efects of social pressures to share income with
kin and neighbors. An important feature of the experiment is that the
distribution of the sizes of the (variable) endowments given to the sub-
jects at the start of the game is common knowledge, which creates an
incentive for those receiving the larger amount to invest in a manner that
keeps the size of their budget hidden. he authors found convincing evi-
dence that women subjects are willing to reduce their expected income
to avoid making investment returns observable to their kin. More pre-
cisely, women receiving the large endowment, who may wish to hide their
advantageous position from the family, are signiicantly more likely to
invest an amount no larger than the smaller endowment when returns are
observable. Such an efect appears to be driven primarily by the behavior
of women with relatives attending the experiment, who would be able
to observe income from the experiment directly. Estimates indicate that
these women invest much less when investment income is observable
than when it is hidden. hey are also signiicantly more likely to invest
no more than the amount of the small endowment, suggesting that their
strategy is designed to keep the size of their endowment hidden. Another
stage of the experiment consists of ofering the subjects the possibility, at
a positive cost, to keep their investment returns secret. he results show
that a signiicant proportion of those able to aford the cost of hiding
income choose to do so.

12
he problem is modeled as a two-step game. In step 1, both agents receive some income
and simultaneously decide how much they deposit with the collector. hey bring home
whatever money is let. In step 2, they decide how much to contribute to a public good,
given that each agent’s contribution cannot exceed the income that the agent brought
home. Finally, they recover part of their deposits and consume it as they wish. In this set-
ting, the agents may make deposits with a view to reducing the amount of money taken
back home, decrease their equilibrium contribution to the public good, and compel the
other agent to increase his (her) contribution. he main result of the model is that because
of the strategic interaction that takes place between them, the agents will follow the deposit
strategy only if they receive similar incomes. If one agent receives a very low income com-
pared to the other, he will not contribute to the public good in step 2. Simultaneously, the
agent who received the high income is not making deposits either because the level of
deposits that would be necessary to prompt the low-income agent to contribute is high, or
the cost of the deposits is proportional to the level of deposits.
 Redistributive Pressures in Sub-Saharan Africa 175

5.6.2 Migration and Estrangement

To escape the incriminatory and jealousy-ridden atmosphere that pervades
many African villages, entrepreneurial individuals may choose to move a
comfortable distance from their kith and kin. Yet even migration to cities
will not necessarily enable them to start up a business and earn proits they
can use for themselves rather than to support “an endless string of social
parasites.” hey also need to sever their ties with relatives, neighbors, and
friends so as to prevent them from settling in their new home and drawing
on their incomes (Poewe 1989). Furthermore, they will seldom return to
their native area, so as to evade customary rules of greeting and extended
family pressures for showing generosity and compassion.
One account that deserves special attention here is drawn from the per-
ceptive work on African capitalism by Paul Kennedy (1988). It is actually
reminiscent of several points made in the previous section. According to
Kennedy, few African irms can be labeled as genuine family businesses:
there is, indeed, a marked reluctance on the part of African entrepreneurs
to establish pooling arrangements with partners from the family for fear
that relatives will seize opportunities to cheat. Any “disappearance” of their
money into a common fund tends to be regarded as evidence that the main
partner has squandered or stolen their assets. In addition, they are almost
as unwilling to delegate authority to kin-related supervisors and managers
as they are to share ownership, again because of deep distrust about peo-
ple’s honesty. Involving relatives in one’s business undertakings, whether as
supervisors or workers, gives rise to intractable problems and tensions.
On the one hand, relatives do not respond to the same discipline as other
employees because they usually demand special treatment, causing discon-
tent among the workforce. Also, jealous of the owner’s success and resentful
of his authority, they are more prone to dishonest and unreliable behavior
in the workplace. On the other hand, the owner tends to be overwhelmed
by continuous kinship obligations, such as requests for cash donations or
gits, demands to inance the education of nephews, nieces, and younger
siblings, or obligations to provide more or less permanent support for wid-
owed or deserted sisters. True, family connections may help in the early
stages by providing start-up capital, yet these shortcomings largely out-
weigh this advantage. If they are strong enough, entrepreneurs will there-
fore resist demands to provide jobs for a wide range of kin, irrespective of
their qualiications. hey know that a business cannot grow unless a clear
separation exists between the business and the kinship spheres (Kennedy
1988, 166–70; in the same vein, see also Beveridge and Oberschall 1979,
176 Platteau

57–9; Ellis and Fauré 1995; Himbara 1994, 85–93; Marris and Somerset
1971, 131–45, 226–7; Nafziger 1977, 192–3).
Asians and Levantines, whose undertakings have typically taken the
form of family irms, have apparently done much better than Africans (and,
in the other way around, small native entrepreneurs from Central and West
Africa have been successful in South Africa and even in New York).13 Two
factors explain their relative success. First, since colonial times, these two
groups have lacked political power and occupied a position of social and
cultural marginality, so that business ofered them the only real opportu-
nity to succeed in life. Much like the Levantines of West and Central Africa,
the Asians in East Africa were willing to operate at low proit margins and
had a high propensity to save. Many of them were also ready to live in rural
outposts for long periods, taking the trouble to learn the vernacular and
operating at low levels of turnover.14 he second factor is more directly rele-
vant for this discussion: by virtue of being strangers, Levantines and Asians
are not subject to the institutions and customs that stile the initiative of
enterprising natives of their adopted country. In particular, they are not
subject to extended family pressures for redistribution. Moreover, magi-
cal repression does not work against them, because they do not believe in
the same spirits as the indigenous population. However, they tend to be
widely resented by the local population, who sees them as strangers usurp-
ing the business opportunities that rightfully belong to indigenous people
(Kennedy 1988, 47–9; Platteau 2000, 224).
he family irm provides foreign entrepreneurs in Africa with a low-cost,
disciplined, and trustworthy labor force whose members all have a direct
interest in the success of the joint endeavors (Kennedy 1988, 48). he fact
that immigrants from the Middle East and Asia belong to small families
(formed essentially around monogamous parents and their children) rather
than to the extended kinship units typical of low-density African societies
implies that the moral hazard-in-team and the incentive dilution problems
that plague collective undertakings is relatively moderate in the former.
his diagnosis has been largely conirmed by recent studies of indigenous

13
Interesting, some of the mulatto descendants of European traders who operated from
the ports and towns of the southern Gold Coast during the mid-nineteenth century were
competing successfully with Europeans (Kennedy 1988, 34–5).
14
As Kennedy pointed out, “where individuals or groups face a situation of multi-depriva-
tion, strong personal motivations may generate an ascetic orientation towards economic
activity . . . this asceticism and the energy, determination and rationality it sustained, did
not stem from a set of shared cultural values existing in their own right, but originated as a
response to the prevailing political, social and economic climate of inequality and power”
(Kennedy 1988, 47–9).
 Redistributive Pressures in Sub-Saharan Africa 177

and nonindigenous (immigrant) irms located in certain African countries,

and based on detailed interviews of a sample of trading and manufacturing
irms (see Fafchamps, 2004, for a synthesis of these indings). he central
results are as follows.
(i) Kinship or place of origin play little or no role in the formation
of business networks: individuals from the same village constitute
a minute portion of the suppliers and clients reported by African
manufacturers (Bigsten et al., 2000; Fafchamps 2002). Buying from
and selling to family members is rare and, in a survey of ity-eight
irms in Accra and Kumasi (Ghana), no sales on credit to rela-
tives were recorded (Fafchamps 1996, 2004, 454). his contrasts
sharply with indings for Asia (Hayami and Kawagoe 1993). In
Africa, the relationship between ethnicity and business networks
is quite loose even when the latter are ethnically concentrated, in
the sense of being located in areas dominated by one ethnic group
(Fafchamps 2004, 307). More speciically, once networks are con-
trolled for, the measured efect of ethnicity on access to supplier
credit falls dramatically (Fafchamps 2000). In other words, busi-
ness networks, which have strong returns,15 essentially result from
interaction among simple business acquaintances, and commercial
relationships are nurtured through business meetings and sociali-
zation outside of work. Ethnic concentration “seems to result from
nothing else than historical accident and socialization patterns that
are reinforced by the practice of business itself ” (Fafchamps 2004,
308; Fafchamps and Minten 2002). Such a conclusion also follows
from the examples cited by Moore (1997), where it appears that suc-
cessful African businessmen have used personal connections devel-
oped at an earlier stage of their professional career, yet outside the
sphere of their kith and kin relations (oten as salaried employees of
European irms).16
(ii) Not only is trade with relatives and friends extremely rare in Africa,
but, whenever it happens, it harms irm performance. Entrepreneurs
complain that it is diicult to keep business with relatives within

15
Typically, trust is based on repeated interaction. Continuing business with reliable sup-
pliers with whom one has had satisfactory dealings in the past is the best way of avoiding
problems. he same applies to trade loans (Fafchamps 2004, chapter 9, 1996; Fafchamps
and Minten 2002).
16
It is odd, therefore, that Moore uses those examples to criticize what he calls the “family/
kin drag” thesis, thereby confusing kinship with other types of social connections.
178 Platteau

the conines of an economic transaction. For example, it is hard to

collect payment from relatives, whether for a loan or delivery of a
good, typically because borrowers do not feel morally obliged to
repay debts incurred from a prosperous relative. More generally,
payment problems are frequent because friendship and family ties
get in the way of exerting pressures on clients (Fafchamps 1996,
441, 2004, 104, 109, 173). In addition, irms buying from family and
friends encounter more late delivery problems (Bigsten et al. 2000).
he conclusion is that involving relatives in business is “the surest
way to go out of business,” while selling on credit to relatives and
neighbors amounts to “signing the death warrant of the irm” and is
therefore carefully avoided (Fafchamps 2004, 173).
(iii) Family background has some efect on start-up working capital,
yet this efect is small, and no systematic efect of family networks
on factor accumulation ater enterprise creation can be discerned
(Fafchamps 2004, 325–6; Fafchamps and Minten 1999). By the
same token, relations based on family, friendship, or ethnicity/kin-
ship make it easier for irms to solve disputes, yet they also raise
the incidence of contract nonperformance. his may result from
the Africans’ preference for nonconfrontational methods of dispute
resolution, which oten end up in face-saving compromises.
(iv) here is no evidence of systematic mechanisms whereby infor-
mation about clients’ trustworthiness is shared among African
manufacturing irms other than direct recommendation by com-
mon acquaintances (Fafchamps 2004, 173, 256–7, 295). Among
agricultural traders, too, trust-based relationships constitute the
dominant contract enforcement mechanism, implying that trust is
established primarily through repeated interaction with little role
for referral by other traders. Information on bad clients does not
circulate widely, which severely limits collective punishments for
opportunistic breach of contract (such as nonpayment). African-
managed irms face more cases of nonpayment than other irms,
and they also complain more frequently about deicient quality
(Fafchamps 2004, 92, 109, 117, 135). heir transaction costs are
consequently higher. hus, the overwhelming majority of Malagasy
traders (in Madagascar) and their clients respond to quality risk by
inspecting each and every purchase, because the task is “virtually
never delegated to family helpers, employees, or collecting agents”
(Fafchamps 2004, 117; see also Fafchamps and Minten 1999). In
other words, many African-managed irms fall back on a “lea
 Redistributive Pressures in Sub-Saharan Africa 179

market” mode of transacting: “inspect the good on the spot, pay

cash and walk away with it” (Fafchamps 1996, 441–4). Firm size
and irm growth are unavoidably restricted in such circumstances.
By contrast, within alien communities, information circulates rather
freely and client referral is a common practice. As a result, nonindigenous
irms operating in Africa are at an advantage. In Kenya, for example, it is
only within the Indian community that irst-time customers are able to
obtain trade credit from the date of their irst purchase. Moreover, indige-
nous irms (in Kenya and Zimbabwe) are less likely to socialize with suppli-
ers, and they have more restricted knowledge about them and their supplies
than do the immigrant irms, suggesting that ethnic barriers are more lim-
iting for the former. Because better business contacts allow irms to enforce
contracts and economize on screening costs, immigrant irms tend to be
more proitable (Fafchamps 2004, 252–3, 258, 300).
(v) Foreign irms hesitate to enter into business relations with indige-
nous irms that they generally deem unreliable. In particular, they
ind fault with African managers for continuously trying to rene-
gotiate delivery and payment terms ex post (Fafchamps 2004, 110).
his bad reputation of unreliability and dishonesty seems to date
back to colonial times (Kennedy 1988, 41, 46).
In the light of the inding reported under (iv), it bears emphasis that the
multilateral reputation mechanism that, according to Avner Greif (1994),
characterizes so-called collectivist cultures is conspicuously absent in sub-
Saharan Africa. What we ind, instead, is the bilateral reputation mechanism
typical of “individualistic cultures.” his is a rather paradoxical conclusion,
yet it is perfectly congruent with our contention that in this region, kin-
ship/ethnic ties and their associated obligations are more an impediment
to private capital accumulation than a social capital susceptible of reducing
transaction costs.

5.6.3 Religious Conversion

To escape the moral and social proscriptions against self-enrichment, will-
ing entrepreneurs need not necessarily leave their native area physically.
hey can become “native outsiders” (Hagen 1975, 279) or “internal strang-
ers,” as converts or urban migrants (ressortissants) are termed in Guinea-
Conakry (Sarro 2009, 160), that is, people who have chosen to distance
themselves radically from their native network so as to be able to disregard
180 Platteau

the attendant obligations without much fear of hostility.17 Such a transfor-

mation occurs when they symbolically adopt a kind of “stranger” status
by converting to a diferent religion: “this places them efectively under an
alternative set of obligations and links them to a spiritual community whose
members encourage private accumulation and economic experimentation
whilst providing resources such as technical knowledge, credit or labour”
(Kennedy 1988, 141–2). In the new spiritual community, emphasis is put on
virtues of self-restraint and consumption moderation (e.g., abstaining from
drinking alcohol and gambling), and on the need to concomitantly do away
with traditional feasts where lavish expenditures on food, drinks, and other
goods are the rule.
In most cases, the necessary escape from community loyalty is secured
through conversion to Islam or Christianity, the rise of which is paralleled
by the decay of indigenous religions. hese were, indeed, communal reli-
gions that could lourish only within local environments, and were doomed
to retreat under the inluence of increased market penetration, new social
strains, and the development of modern communication means. Syncretic
religions and independent churches combining faith in a High God with
beliefs in magic and witchcrat that borrowed elements of Christian escha-
tology gradually supplanted them as they were better adapted to the new
world of the Africans (Ilife 2007, 234, 237). In particular, the churches
“provided institutional models for people who had few relevant indigenous
institutions” (Ilife 2007, 235).
Especially successful in Africa (and other parts of the developing world,
such as Central and South America) are charismatic churches that ofer
“places for retreat and personal transformation” while allowing people to
engage in village life in new ways, opening up new career paths, and hold-
ing out good prospects for the forthcoming future (Jones 2009, 92, 97–8).
Pentecostalism, for example, ofers the individual “a chance to break with
the past,” and its various prohibitions − on alcohol, traditional medicine,
and polygamy − “serve as outward signs of this inward rupture” (Jones
2009, 153). Presented here are a few examples destined to illustrate more
vividly the transformation logic at work.
Among the Fra Fra people of northern Ghana, when successful indi-
viduals are unwilling to share their proits generously with members of

17
In the Upper Guinea Coast of Guinea-Conakry, we are told: “More and more numerous
are the young men who abandon their village for a nearby urban environment, where they
hope to escape all control. hose who remain marry strangers, convert to Islam, and reject
their former identity” (Sarro 2009, 97).
 Redistributive Pressures in Sub-Saharan Africa 181

their ethnic group and simultaneously keen to avoid the grave accusation
of being “swindlers,” they convert to Islam or Christianity, which removes
the need to move to an impersonal city (Hart 1975). In the Serenje district
of Zambia, Jehovah’s Witnesses were disproportionately represented in the
commercial farmer and shopkeeper categories, although they were some-
what younger and had less urban experience than their economic rivals.
hey also tended to live outside the traditional matrilineal village in their
own independent settlements, and in their spare time practiced such trades
as bricklaying, carpentry, or tailoring from which they earned extra cash
(Long 1968, 146). In addition, they systematically avoided hiring kinsmen-
workers who did not share in the ownership of the farm or its equipment
because their employment “is likely over time to lead to diicult farmer/
worker relations which centre around the problem of reconciling the rights
and obligations of matrilineal kinsmen with their role as workers on the
farm.” he fact of the matter is that Jehovah’s Witnesses “have little inter-
est in traditional status criteria and espouse an ethic which emphasizes the
spiritual and moral dangers of associating too freely with non-believers,
even if kin” (Long 1977, 138–9). In their struggles to disentangle themselves
from the demands of the matri-kin, they can rely on religious justiication,
spiritual protection, and practical assistance provided by their church, thus
improving their ability to concentrate on their business and nuclear family
interests (Kennedy 1988, 142).
Parkin’s study of the Giriama people of coastal Kenya (1972, chapters 2–5)
provides another vivid illustration that shows, inter alia, how God can be
summoned to defeat witchcrat. As Kennedy aptly summarized:
[T]he conversion of some young entrepreneurs to the Islamic faith might follow
a long period of psychological tension and physical illness induced by the possi-
bility of conlict with the elders whose status and power were threatened by the
younger men’s activities. Such “illness” could be diagnosed as caused by powerful
Islamic spirits whose appeasement required nothing less than the religious conver-
sion of those unfortunate enough to become possessed. Once this had occurred,
the Islamic ban on the consumption of alcohol and certain foods, as well as the
need to follow a partly separate ritual and social life, all provided the opportunity
for entrepreneurs to reduce their level of involvement in traditional society. Yet this
behaviour no longer incurred community displeasure since it was now judged to be
religiously determined rather than the result of selish individualism. At the same
time, the converts were still reasonably close to community afairs and so could
use their social connections with the elders in order to gain access to land, reliable
information, business contracts and so on. (Kennedy 1988, 142)

In eastern Uganda, the discourse and practice of “born-again” Christianity

as they are observed within the Pentecostal Assemblies of God aim at helping
182 Platteau

church members to separate themselves from society. In the words of Ben

Jones, “born-again Christians challenge accepted notions of community,
kinship and tradition. Becoming ‘born-again’ means joining a new family,
one bound together by faith. Within the community of the ‘saved,’ there is
oten a strong opposition to past relationships, as these relationships are
oten seen as a vehicle for sin or a sense of personal misfortune. . . . Church
members are permitted to cut familial and kinship ties in the knowledge
that they are leaving behind the worship of ancestors, gods or the use of
traditional medicines. his leads to oten radical social actions, precipitat-
ing some sort of retreat from wider society” (2009, 97–8). he same analy-
sis of the role of Pentecostalism as a progressive force for the development
of Africa has been made for other countries, such as Ghana (Meyer 1998,
2004) or Burkina Faso, to which our attention is now turned.
In Burkina Faso, a dynamic individual who had converted to the
Pentecostal church expressed the view that: “If people see that someone is
going to fare better, they use magic tricks to kill him, and this is something
that inhibits our development because jealousy is great . . . there are those
who want to harm me, but I also know that, thanks to my love for God, they
are made powerless since the power of God surpasses that of the evil local
spirits” (Fancello 2006, 124). Conversion to Islam or to Assemblies of God
also proved to be the most efective weapon available to a youth group who
wanted to embark on a development project in their (burkinabé) village,
but they were subject to threats of magical attacks manipulated by local
elders (Laurent 1998, 108–9). Among the Orma pastoralists living near the
Tana River (Kenya), likewise, it seems that the irst shopkeepers and traders
were young Islamic converts who chose to challenge the authority of the
elders and who found in Islam codes of law and behavior particularly con-
ducive to sound business relations (Ensminger 1992, 48–62). Poewe (1989)
ofers a similar analysis of the motives underlying the spread of evangelical
churches in Zambia (see also Shillington 1989).
If the religious sodality or association serves as a business network,
the convert enjoys a double advantage: besides becoming immune to the
demands of kinsmen, friends, and neighbors, he (she) is able to econo-
mize on transaction costs because of the availability of efective contract
enforcement mechanisms. his is because the sharing of the same faith
ensures that people abide by the same codes of behavior, adopt the same rit-
uals, and use the same language, resulting in a common identity grounded
in the same symbolic universe. If the brotherhood is suiciently small for
information transmission to be efective, the mutual trust that it nurtures
and the repeated interaction framework that it provides allow the members
 Redistributive Pressures in Sub-Saharan Africa 183

to solve many incentive problems (of the moral hazard and adverse selection
kinds) arising in commercial relationships. Moreover, religion performs the
function of signaling or identifying members. In fact, the religious associa-
tion works as a club: entry into it involves a ixed cost (the initial fee) under
the form, useless for outsiders and therefore not transferable, of learning the
language as well as the doctrine and the rituals. It replicates a git exchange
relationship by a group of agents, and is efective to the extent that any vio-
lation of the standard of behavior (honest trading) within the club is pun-
ished by the termination of club membership (Aoki 2001, 64–7).
In West Africa, for example, long-distance trading communities or “trad-
ing diasporas” dating back to the thirteenth and fourteenth centuries have
been built on elaborate and successful networks. An important feature of
these networks is their openness to new entrants on the (obviously restric-
tive) condition that they shared or accepted the essential cultural require-
ments for participation in the moral community that distinguished the
members from the host society: Islam and the appropriate trading language
(Austin 1993, 115).18 he adoption of Islam thus spurred the economic inte-
gration of West African regions and their integration into trans-Saharan
trade thanks, in part, to increasing safety of the caravans and smaller con-
tract enforcement costs (Brooks 1993, 117; Cohen 1969, 1971; Hopkins 1973,
58–65; Levtzion 1973). To enforce contracts and rules of commerce along
the Juula and Hausa inland networks, as well as at the Sahelian entrepôts
along the Niger bend, a parallel diaspora of clerical specialists, both judges
(qadis) and legal scholars (mutis), was created. Interesting, however, they
did not always demand strict religious adherence on the part of the local lay
Muslim community (Lydon 2009; see also Launay 1992, 191, for the Dyula
of Ivory Coast).
Especially worthy of note is the lexibility shown regarding the most
growth-limiting Islamic rules, such as the law of commercial partnerships,
which limited enterprise continuity and intergenerational persistence, and
inheritance prescriptions, which encouraged wealth fragmentation and
restrained capital accumulation.19 hus, according to Austen, descent could
be manipulated so as to avoid dispersion of business assets by selecting one
unique successor among slaves/clients (rather than relatives) recruited into
the trading organization as junior partners. In fact, speciic arrangements

18
he nonfulillment of this condition explains why European (French) merchants failed to
make inroads in cattle trade within what is now the republic of Niger: they were unable to
participate as equals in indigenous systems of guaranteeing credit and enforcing contracts
that emphasize a common religion (Islam) and language (Hausa) (Austin 1993, 117).
19
On these aspects, see Kuran (2003, 2004a, 2004b).
184 Platteau

“depended almost entirely upon arrangements made within a modiied ver-

sion of the secular kinship idiom” (1987, 42–4).20 he major social function
of Islam was, in fact, “to provide merchants with an identity which rein-
forced their occupational role.” Muslim merchants of Sudan acted as “repre-
sentatives of a cult that had material and spiritual connections to a universe
larger than the parochial world of local villages, or even savanna empires.”
his helped them to gain respect from other West African peoples. As is
evident from the processes of formation of the Juula and Hausa groups,
the Sudanic identity of Muslim merchants was based on multiple levels
of ethnicity “beginning with their use of speciic West African languages
and receding into their consciousness of descent from various population
groups strategically placed to enter the occupation of trade.”21
he spread of Christianity and Islam may play a progressive role by over-
coming traditional barriers to change in sub-Saharan Africa. Not only do
the imported religions emphasize the need to ight against alcoholism, drug
use, cheating at university exams, prostitution, corruption, nepotism, and
delinquency (Fancello 2006, 103–4; Miran 2006, 447), but they also ref-
use to recognize ethnic identities and the customary obligations associ-
ated with them, especially when they involve lavish expenditures. In the
words of Aboubacar Fofana, an inluential reformist Muslim cleric of Ivory
Coast, “what matters is not where an individual comes from, but what he
or she does.” Muslim brotherhoods put emphasis on personal merits and
engagement, not on social status and ethnic identity: they are based on elec-
tive membership (Miran 2006, 450, 472–3). Generalized morality, without
which an efective market and polity are hard to develop, precisely requires
that the ability to see things from another’s viewpoint becomes based on
identity or loyalty feelings toward a large reference group accessible to all.
To put it in another way, identity must be anchored in the abstract rather
than the concrete other (Platteau 1994, 2000, chapter 7).
An important caveat must be mentioned. Acceptance of Islam may have
been a collective, politically motivated move rather than an individual
step of emancipation from erstwhile community control. Societies that are

20
Sophisticated inancial instruments, such as letters of credit and bank drats, do not appear
to have circulated farther away than the northern entrepôts of the Sahara itself, not a sur-
prising inding because most Sudanic Muslim merchants remained functionally illiterate
despite some training in Quranic liturgy (Austen 1987, 41, 43).
21
he Juula groups were Soninké with a Mauritanian homeland. hey took on their Manding
linguistic identity during the height of the medieval Mali and Songhai empires. During
the fourteenth and iteenth centuries, they penetrated into the central Sudan, where they
founded the irst Hausa-speaking merchant lineages (Austen 1987, 42–3).
 Redistributive Pressures in Sub-Saharan Africa 185

apparently Islamic may then have fundamentally retained their erstwhile

kinship structure and organization, thus replicating the existing social
structure under the new system. his outcome is more likely to be observed
if religious conversion had been imposed on a people by foreign invad-
ers. For example, strongly organized polities, such as the kingdoms of the
Mossi, Dagomba, and Akan, in which central power was irmly based on
the kinship organization of the people, resisted iniltration of Islam in the
late eighteenth and early nineteenth centuries, and “continued to maintain
the old Sudanic traditions of government which might use but which were
never wholly committed to the new religion” (Fage 1995, 194–5).
Across the Hausa lands (in what is now northern Nigeria), the conquest
by the Fulanis, who adhered to a rather strict version of Islam and pos-
sessed a highly developed class of Muslim clerics (torodbe), had quite dif-
ferent efects, depending on the strength of local political organizations
and their ability to resist outside inluences.22 For example, ater the Fulani
conquest the Hausa peasantry remained submitted to the rule of powerful
landlords, but because the new landlord class was, initially at least, an alien
one less subject to the dictates of local custom, its erstwhile rights were less
efectively protected. Yet “both landlords and peasants were now Islamised
and operating within a framework of general law and principle, which mil-
itated against arbitrary oppression, and made for greater security and ei-
ciency,” hence the increasing prosperity of the trade and industries of the
Hausa cities ater the Fulani-led jihad (Fage 1995, 207–8).23
In the Songhai empire of Gao, bureaucracies might appear to have been
Islamized, but the principles of Sudanic royal power remained pagan. he
same holds true for the central Saharan Berbers, who, despite Islamization,
still cling to their language and many of their customs, and have succeeded
in absorbing the Arab groups as tributaries into their own tribal system. As
for the Massassi (of mixed Soninke and Fulani descent), although they had
Muslim clerics in their entourages, they acted in distinctly Islamic ways,
but only when it suited their interests. Rather than pursuing the aim to

22
With the general decline of Mande inluence following the Moroccan invasion, Fulani
clerics joined with those of the Tuareg in the leadership of Muslim scholarship in the
Sudan (Fage 1995, 197).
23
In the Futa Jalon, where the jihad actually started, Muslim Fulani leaders succeeded in
completely overturning the existing sociopolitical structure. Indeed, local Jalonke chiefs
initially had the right to tax Fulani herders and to control the allocation of land to them. By
the middle of the eighteenth century, however, as a result of the Fulani Islamic upheaval,
they had become tributary to Fulani overlords throughout the territory. Moreover, they
ultimately converted to Islam and became accepted members of Fulani society (Fage 1995,
199–200).
186 Platteau

propagate Islam (that is, to preach and establish the truth of the one God
and of the universal brotherhood of man subordinate to his will alone), the
Massassi actually wanted to convert the clan and age-grade structures of
traditional Bambara society into associations of serfs and clients subordi-
nate to their will as war leaders (Fage 1995, 187–9, 192, 194, 196–208).
When Islam was adopted as a result of a collective strategy of the domi-
nant society or social group, appeals to strict interpretations of the faith may
be the way disgruntled individuals or marginal groups chose to emancipate
themselves from local customs and rules. A striking illustration is found in
the attacks of Wahhabites on the relatively tolerant forms of Islam repre-
sented by the Sui orders and practiced in the towns of Maradi (Niger) and
Koko (Nigeria). he younger educated elite who espouse this puritanical
version of Islam are attracted by its asceticism and its proclaimed rejection
of any ostentatious redistribution of wealth. hey want to set their members
apart from the older Islamic community by rejecting the authority of local
scholarly traditions and adopting distinctive behaviors and rituals (Cooper
1997, 130–4; Gregoire 1993, 109–10; Launay 1992, 104–31). he appeal of
a strict Islam also seems to be strong for groups that were either excluded
from the traditional sociopolitical structure, or felt that their position in it
was, on the whole, disadvantageous. For these groups, indeed, pure Islam
“provided a means by which the legality of the traditional system might
be challenged,” because this system could be shown to be “a denial of the
essential brotherhood and equality of man, irrespective of his ethnic and
social origins” (Fage 1995, 195).
he conditions under which Wahhabism rose in Ivory Coast are espe-
cially interesting: this puritanical doctrine started to spread during a
period of increasing disruption of the traditional society as a result of
quickly growing urbanization and migration movements ater the Second
World War. It attracted rich merchants, whom it supplied with an ideology
that was both antiestablishment and bourgeois. Wahhabism was anties-
tablishment in the sense of being opposed to the feudal-like elite of the
marabouts, who are at the heart of traditional Islam, but are viewed as
impostors illegitimately interposing between God and the faithful. And it
was bourgeois in the sense of being emancipated from the constraints of
the traditional aristocratic system (Miran 2006, 250).24 What bears most

24
It is true that the egalitarian discourse of Wahhabism also appealed to low-caste people
willing to liberate themselves from the yoke of the traditional system of hierarchy, but only
in the 1980s did it start recruiting among the urban poor. For them, Wahhabism ofered
a partial response to their quest for social protection and spiritual advice in a context of
urban economic precariousness (Miran 2006, 285).
 Redistributive Pressures in Sub-Saharan Africa 187

emphasis is that behind the screen of religious antagonisms lay genuine

social and political conlicts:
[T]he battleground was less religious and doctrinal than social and political. As
a matter of fact, the recourse to the religious sphere through wahhabite sectari-
anism allowed dissatisied people to express diferences that could not be overtly
declared in the political realm. Accusations of intolerance, dogmatism and narrow-
mindedness against the Wahhabites were therefore partly correct: since their sep-
aratism was grounded elsewhere, no discussion of a doctrinal nature was possible
and no practical compromise could be reached with traditionalist Muslims. (Miran
2006, 259, my translation)

Clearly, the authenticity or depth of an act of religious conversion depends

largely on the position of an individual or a group in the existing social
and political structure, as well as on the opportunities and constraints for
change that are perceived in particular historical circumstances. It is there-
fore not surprising that signiicant and robust relationship between reli-
gious ailiations and economic growth performance does not emerge from
cross-country econometric studies, and that Islam, or Protestantism, some-
times appears to have a positive, sometimes a negative efect on growth.
he complexity of historical processes and the crude manner in which reli-
gious ailiations are measured account for these disappointing economet-
ric results (Platteau 2009; see also Austin 2008).

5.7 THE LACK OF BROAD-BASED COUNTERACTING

FORCES IN SUB-SAHARAN AFRICA
If these highlighted strategies remain marginal attempts by a few individu-
als to escape redistributive pressures, the underlying social norm persists. If
they are followed by a suiciently large number of individuals, however, the
norm of sharing is likely to fall into decay. Note that the same efect would
be achieved, yet without incurring the cost involved in defensive or circum-
venting strategies, if a critical mass of deviant, “selish” entrepreneurs could
make a coordinated move to resist redistributive pressures (Platteau 2000,
chapter 5). It is, indeed, reasonable to assume that the individual cost of
being punished for noncompliance falls as the number of violators increases,
so that there must exist a critical mass of violators above which this cost
becomes smaller than the individual beneit from emancipation (assumed
to be independent of the number of violators). Such a scenario seems to it
the situation observed in southern (yet not northern) Ghana today, where
many rural migrants have not severed ties to their kith and kin. In fact,
relationships with the latter are good enough to enable migrants to put up
188 Platteau

beautiful houses in their native villages without having to fear retribution

for (partial) noncompliance with the norm of sharing.25
he overall picture nevertheless suggests that redistributive pressures
remain a pervasive feature of many parts of Africa. his may be explained
not only by the lack of coordinated moves by “selish” entrepreneurs or the
limited scale of religious conversions and tie-severing migrations, but also
by the weakness of broad counteracting forces operating at the societal
level. Following Karla Hof and Arijit Sen, two types of drivers of cultural
transformation can be distinguished: factors that tend to increase the mar-
ket opportunities foregone as a result of the extended family system, such
as the expansion of the domain of the market, on the one hand, and, on
the other hand, factors that tend to lower the return to participation in
this system, including a fall in non-covariate risk, an increase in mobility
that raises the cost of enforcing kinship-based obligations, and exposure
to new ideas that weakens the hold of supportive traditional norms and
beliefs (Hof and Sen 2005, 183). In the following discussion, attention is
focused on two broad forces that are expected to yield these efects simul-
taneously, namely, the twin processes of industrialization and urbaniza-
tion, and the emergence of nationwide ideological movements. hey are
examined in turn.

5.7.1 he Dilemma Born of the Lack of Industrialization

As Karl Marx rightly emphasized, the development of industrial capitalism
and the accompanying concentration of a growing part of the workforce
in large economic establishments go hand in hand with a process of social
class formation epitomized by the rise of employers’ associations and work-
ers’ trade unions. By creating new ties of solidarity and mutual help based
not on the primordial ties of the family, village, and clan, but on work rela-
tionships woven in an alien environment, industrialization allows both the
workers and the managers or capitalists to emancipate themselves from the
web of reciprocal obligations inherent in their traditional universe. Social
classes, as opposed to kin-based relationships, are wide reference groups,
each with its own subculture. hey provide their members with a new and
enlarged vision of their own interests as well as a concrete basis on which to

25
Noncompliance is only partial because among the Akan and the Ashanti ceremonials
are still currently organized that are essentially aimed at containing successful migrants,
prompting them to bring beneits to their native community (I thank Emmanuel
Akyeampong for drawing my attention to this point).
 Redistributive Pressures in Sub-Saharan Africa 189

anchor new patterns of brotherhood and socialization, as well as motivation

for participating in society-level struggles.
A vicious circle is nevertheless set in motion if, as argued earlier, the
existence of redistributive norms has precisely the efect of hampering the
dynamic of capital accumulation, risk taking, and enterprise creation in the
industrial sector, in particular. Such norms would then be self-supporting
or self-reinforcing. Foreign direct investment, irm creation by ethnic or
stranger communities, and/or the formation of a critical mass of indigenous
entrepreneurs who have migrated or converted to a monotheistic religion
are possible ways of breaking out of this nasty equilibrium. Yet the situation
of sub-Saharan Africa is actually more intricate because other obstacles
than redistributive norms are at work that discourage industrial investment
and make for an economic landscape in which the overwhelming majority
of African entrepreneurs operate very small businesses. Prominent among
these additional obstacles are considerable transaction costs caused by
poor physical infrastructure, lack of competition, and deicient macroeco-
nomic management (Minten and Kyle 1999; Platteau 2000, chapter 2), lack
of familiarity of African entrepreneurs with the conditions prevailing else-
where both on the technological and institutional or organizational levels
(Fafchamps 2010, 152–6), as well as political factors that include political
instability, widespread state corruption and rent seeking, and subversion of
political choices by ethno-regional forces (Bates 2008; Fosu 2008, 137–72;
Ndulu and O’Connell 2008, 55–6; Platteau 2009, 671–8; Van de Walle 2001,
chapters 3–4). All these factors point to the important role of politico-insti-
tutional factors for economic development, a role Jan de Vries underlined
in the light of Western European history:
he key to understanding the growing power of capital in the European economy
is not to be found by searching for esoteric sources of capital; rather, it rests with
the solution to the problem of preserving and keeping productive the capital stock
already in existence. In other words, a major weakness of the European economy
that had to be overcome was not the inadequacy of capital, but the misinvestment
and dissipation of capital. (1976, 213–14)

5.7.2 he Lack of Deep-Rooted Nationalist Ideologies

We have seen earlier that the rapid rise of religions combining faith in a
High God with beliefs in magic and witchcrat plays a progressive role in
helping Africans to construct a collective identity based on the abstract
rather than the concrete other. Social struggles, as has been pointed out
previously, could generate the same efect of undermining particularized
190 Platteau

loyalties, yet they are still in their infancy in much of sub-Saharan Africa.
Nationalist ideologies – in the sense of ideologies proposing a symbolic
framework within which to give form and meaning to emerging nations,
so as to create a difuse sense of common destiny − could also broaden
the Africans’ vision of public purpose and image of social reality in which
to anchor a new collective subject articulated to an independent state. In
the words of Cliford Geertz, the formative stage of nationalism essentially
consists of “confronting the dense assemblage of cultural, racial, local, and
linguistic categories of self-identiication and social loyalty that centuries
of uninstructed history had produced with a simple, abstract, deliberately
constructed, and almost painfully self-conscious concept of political eth-
nicity – a proper ‘nationality’ in the modern manner” (1973, 239). In any
nation-building process, the new states, or their leaders, must contain or
domesticate primordial attachments instead of wishing them out of exis-
tence or belittling them. his implies that they are able to reconcile them
“with the unfolding civil order by divesting them of their legitimizing force
with respect to governmental authority, by neutralizing the apparatus of the
state in relationship to them, and by channeling discontent arising out of
their dislocation into properly political rather than para-political forms of
expression” (Geertz 1973, 277).
he problem is again that nationalist movements are of a rather recent
origin in Africa, having started in the context of comparatively short anti-
colonial struggles. Moreover, political mobilization since independence has
centered on competition for access to the privileges and rents associated
with power, following a logic of patronage based on ethno-regional fac-
tionalism rather than on broad-based political movements and allegiances.
hese causes of ill development of nationalist ideologies in Africa are
explored in two successive subsections.

5.7.2.1 Short-Lived Independence Movements

Compared to Latin America and Asia, sub-Saharan Africa has been sub-
ject to a rather short spell of colonial rule. his is largely the consequence
of the fact that the African continent was swept along in the whirlwind of
decolonization that surged throughout the world ater the Second World
War. Many African countries thus obtained their independence shortly
ater Asian countries that went through a much longer period of colonial
domination. Unlike what was observed in Latin America and Asia, proin-
dependence political movements in Africa had therefore little time to
develop and consolidate. An important upshot of this situation is that they
had to abruptly shit from the comparatively easy task of claiming national
 Redistributive Pressures in Sub-Saharan Africa 191

freedom and sovereignty by uniting against an external power to the much

more complex task of building a nation in the aforementioned sense. A
nation-building efort can succeed only if a sustainable compromise can
be struck between the need of a country to modernize its social, political,
and economic structures so as to adapt to the world environment, on the
one hand, and the need to win a mass-based legitimacy and involve tra-
ditional interests that play on primordial attachments, on the other hand.
he following diagnosis proposed by Geertz thus applies with especial force
to sub-Saharan Africa:

[T]he very success of the independence movements in rousing the enthusiasm of

the masses and directing it against foreign domination tended to obscure the frailty
and narrowness of the cultural foundations upon which those movements rested,
because it led to the notion that anti-colonialism and collective redeinition are the
same thing. (Geertz 1973, 239)

In many cases, the new African political leaders opted for a swit move to
some form of top-down, centralized modernization policies heavily bor-
rowed from the experiences of other continents or directly taken over from
their own colonial models: import substitution, state planning, marketing
boards, socialist cooperatives and consolidated village communities, and
so forth. In other words, relying on a blend of African nationalism and
some version of Marxism, the new educated elite gave little thought to the
question of the suitability of these policies to the African context and to
the importance of inding a common ground with local interests and tra-
ditional constituencies.
Tanzania is an interesting case to consider because the strength of the
independence movement, the Tanganyka African National Union (TANU),
was probably greater there than in any other African colonial territory. John
Ilife (2007) attributes the exceptionally widespread support TANU enjoyed
to its use of the widely spoken Swahili language and the absence of strong
tribal politics, conditions largely inherited from Tanganyka’s nineteenth-
century experience (Ilife 2007, 256). Revealingly, Sir Richard Turnbull,
governor of Tanganyka, confessed that he did not foresee that TEMO,
the parliamentary caucus of TANU, would develop into “the monolithic,
strictly disciplined structure” that it had become, and that the situation in
Tanganyka was unique because he could not think of any other colonial
territory where such a situation had occurred (Ilife 2005, 189). he British
authorities were still determined to refuse too quick an access of Tanganyka
to independence both because this would trigger all other East African ter-
ritories (Kenya, Uganda, Ethiopia, and Somalia) to follow suit, and because
192 Platteau

local human resources were deemed insuiciently developed to answer the

needs of an independent state.
Under the growing inluence of the radical wing of the TANU (espe-
cially ater Oscar Kambona took its command in October 1959) to which
even Julius Nyerere had to give in, the British were nevertheless forced to
abandon their plan and concede independence to Tanganyka (in December
1961) although it was not ripe for it.26 “he idea of a long pause between
responsible government and independence” had become patently unrealis-
tic in the eyes of British authorities overwhelmed by the fearful prospect of
a combination of Mau Mau and Maji Maji rebellion (Ilife 2005, 189, 192).
It was considered wiser to try to maintain a good relationship with Nyerere
to save the long-term interests of Britain in East Africa. In the words of
Turnbull:
We are, ater all, under an obligation to make Tanganyka self-governing, and it
would be better to reach the consummation too early with the people on our side
than ater a campaign with the people irrevocably against us. Our irst interest must
surely be to maintain peaceful conditions . . . so that, when the inal change comes
about, Tanganyka will look to us and not to the Soviet bloc to keep the country sup-
plied with technicians and as a source of manufactured articles. . . . It is essential for
us to use Nyerere whilst he is still powerful; if we wait too long, he will be ousted by
the extremists; and with him will go all hopes of an enduring European inluence in
Tanganyka. (Ilife 2005, 190)

Strategic considerations born out of the Cold War context therefore

prompted old European colonial powers to grant political independence
to African countries that had too little time to evolve a genuine sense of
nationhood.

5.7.2.2 Factionalism and Tribal Patriotism

If we follow the analysis of John Ilife in his ambitious book Africans –
he History of a Continent (2007, 239–41), a key feature of many tropical
African colonies lies in the fact that they were purely European creations.
hey therefore lacked precolonial elites, which in North Africa, Asia,
and Latin America were able to propose colony-wide perspectives and

26
Radical leaders and activists of TANU, particularly of the Youth League, were remarkably
efective in constructing a nationalist discourse and deining the content of a national
identity and common destiny. In pursuing their aims they did not hesitate to openly blame
the ruling clans for protecting unqualiied oiceholders and denying positions to edu-
cated individuals, for preventing appeals against the decisions of the chiefs’ courts, and for
perpetuating ineiciency, nepotism, and corruption (Giblin 2005b, 143–4; Monson 2005,
109–10).
 Redistributive Pressures in Sub-Saharan Africa 193

implementation strategies. he modern politics of sub-Saharan Africa

actually began at two levels, one above and the other below the territorial/
national level. At the supra-territorial level, pan-African ideologies, such
as that of negritude (which was born in the Caribbeans), inspired many
African intellectuals, and pan-African political movements arose, such as
the National Congress of British West Africa (NCBWA) with branches in
the Gold Coast, Nigeria, Sierra Leone, and the Gambia. At the infra-ter-
ritorial level, tribal patriotism prevailed. Africans strove to defend their
locality against European intrusion, oten resulting in the formation of
innumerable local associations by traders, commercial farmers, teachers,
clerks, and clergymen. Although their oicial purpose was to promote the
development of the country, in actual practice they defended particularis-
tic interests and did not hesitate to foster tribal identities to better achieve
their objective.
Nationalism was therefore a veil behind which many parochial interests
could conceal themselves. In the words of Ilife, “as predominantly local
people, most Africans saw nationalism in part as a new idiom for ancient
political contests” (2007, 258). As a result, nationalist movements were fre-
quently plagued by factional conlicts that put local issues in the forefront
while minimizing the importance of national party ailiations. In these con-
ditions, it is not surprising that “nationalism only partially aroused many of
Africa’s deepest political forces,” and that “responses to it depended on local
circumstances” (Ilife 2007, 257). his said, it is precisely because nation-
alism was thus absorbed into local political rivalries rather than driven by
broad social conlicts that it gathered the support needed to remove colo-
nial rule. Unfortunately, for the same reason, ater independence people
continued to perceive national issues in terms of local interests and to judge
their representatives and the state on the basis of their contributions to local
advancement (Ilife 2007, 258, 267).
Although less true than in other territories, this observation regarding
the political motives underlying the anticolonial struggle was even veri-
ied in some districts of Tanganyka, such as in Mahenge, where people
decided to join TANU in the hope of enlisting the support of the nation-
alist party in their “tribal” war. In particular, Ndamba activists hoped to
obtain redress of a long list of grievances against the occupation of their
(Ulanga and Kilombero) valleys by members of the Bena ethnic group.
hey became especially outraged when in the mid-1950s a Bena “big man”
named Kiwanga required them to list their ethnicity as “Bena” in govern-
ment books, which implied that they would have to abandon any claim to
sovereignty over the valley (Monson 2005, 107–10).
194 Platteau

In Uganda, a country much more subject than Tanzania to the divisive

forces of local factionalism and tribal patriotism, it is impossible to under-
stand the Teso rebellion if one overlooks the fact that its political dimen-
sions were “parochial rather than regional in orientation” (Jones 2009, 50).
Revealing, the rebels killed local “big men” (such as the village council chair-
men, the parish chief, or the leaders of traditional associations) rather than
district or regional oicers, in spite of the rhetoric that their struggle was sup-
posed to be fought against the central government. he fact of the matter
is that the Teso insurgency was not a coordinated attack against the central
government or an assault on the regional elite, but a local conlict opposing
younger and older men. his conlict took center stage under harsh condi-
tions where younger people felt threatened by the sudden fall of cattle stocks
in the hands of the elders, thereby undermining the resources on which they
depended for access to marriage and education (Jones 2009, 51–2, 61).
he famous Mau Mau revolt that erupted in central Kenya in the 1940s
and 1950s was also driven by local interests, so much so that Kenyatta, pres-
ident of the Kenya Africa Union (KAU) party, felt the need to denounce the
movement’s reactionary character in February 1951 (Kanogo 1987, 136–43).
Violent acts committed by Mau Mau, such as the massacre of Lari (a village
in the Central Province), were linked to land feuds that had their origins in
the land alienation schemes for settlers. Local land conlicts in the Kikuyu
region were actually the main force behind the Mau Mau revolt, and these
conlicts were not necessarily ignited by land expropriation to the beneit
of the British colonizer. A good part of rural tension also arose between
families that had lost what they claimed were their ancestral lands and the
new indigenous landed gentry who had acquired such lands through intim-
idation and corrupt means, including lengthy litigation processes and the
associated payments of bribes. he loss of prestige of traditional elders as
chiefs appointed by the colonial administration emerged as new centers of
power, and the resulting severe clash of generations that caused the disinte-
gration of traditional cultural obligations, were other critical factors in the
chaotic situation epitomized by the Mau Mau movement (Maloba 1998,
27–34, 100–1).
Other examples readily come to mind. he Nigerian Youth Movement,
which Ilife considers the irst truly nationalist program in tropical Africa,
divided ater a few years of existence (in the early 1940s) amidst rivalry
between Yoruba and Igbo. It is telling that at that time British oicials
saw no danger in Africa of the nationalism threatening them in India.
When elections were later granted by the Richards Constitution in 1946,
fear of domination by the Igbo-led National Council of Nigeria and the
 Redistributive Pressures in Sub-Saharan Africa 195

Cameroons (NCNC) led western and northern politicians to create regional

parties, the Action Group, claiming to represent Yoruba, and the Northern
People’s Congress standing for the Hausa (Ilife 2007, 242–3). In Ghana,
the Convention People’s Party of Nkrumah faced serious opposition in the
mid-1950s from the National Liberation Movement. Both parties courted
the support of Asante and Ahafo chiefs by taking sides in the rivalries that
opposed them, particularly over the delicate question of the right to col-
lect tribute from cocoa farmers. Like in Nigeria, “such strategies of political
mobilisation served to link debates over customary jurisdictions and land
rights to emerging patterns of party conlict, ensuring that the politicisa-
tion of rural property rights continued ater the end of colonial rule” (Berry
1993, 124).
In Congo, the nationalist movement was hampered by deep divisions
between small, unitarist ethnic groups and the large, federalist Kongo and
Lunda groups. When political responsibilities were transferred to nation-
alist politicians (President Kasavubu and Prime Minister Lumumba) while
control of the army and the civil service remained in Belgian hands, these
centrifugal tendencies were cynically exploited by foreign powers (Belgium
and the United States), and central power soon collapsed. Four regional
armies came into being, and regional rebellions, millenarian movements,
and tribal wars engulfed the whole country, creating an awful legacy of
chaos, ethnic hatred, and violence (Berry 1993, 268).
he end outcome of these dynamics has been the emergence of societies
stratiied in complex ways, and where class alignments are crisscrossed by
kinship, age group, ethnic and regional ailiations, as well as by patron-
client relationships relecting the Big Man syndrome (First 1970, 97, 103).
Here is a world of factions understood as groups or networks that use par-
enthood, friendship, and clientelistic ties to conquer power and obtain
material advantages (Redield 1962, 42). he bureaucracies that emerged
from independent African states were all to varying degrees patrimonial,
in the sense that oice was conferred in return for personal loyalty and
service to the ruler. he regimes were typically held together by personal
relationships among a small elite that coalesced into a single ruling group.
Each member of this elite group stands at the apex of a patronage network
typically based on ethnic or regional identities and linking him to a locality
(Ilife 2007, 269–70).
As for political parties, they have been aptly described by Jaindi Kisero, a
Kenyan journalist in the Daily Nation (Nairobi, December 2000). According
to him, indeed, they are “beret of platforms and ideology and are no more
than institutions whose only reason for existence is the intention to divide
196 Platteau

the spoils of state among the ethnic communities that support them. hese
attitudes are captured well by the Yoruba saying that ‘an elder who is brought
a plate does not lick it clean, or he will not ind young ones to run errands for
him.’ . . . Instead of getting annoyed about cases of corruption, we Kenyans
get more enraged when fruits of corruption are not shared out according
to settled expectations” (cited in Berman, Eyoh, and Kymlicka 2004, 49). It
follows that, despite oten glaring social inequalities, social class identities
do not emerge in societies with pervasive clientelism: by reproducing ties of
reciprocity and allegiance across class lines, patron-client networks and fac-
tions reinforce ethno-regional identities and undermine the development
of class consciousness and conlict (Bourdieu 1990, chapter 8; Chabal and
Daloz 1999, 41–3; Van de Walle 2001). he aforementioned lack of urban-
ization and industrialization reinforces this process because it makes peo-
ple strongly dependent on privileged and personalized links to patrons for
their daily livelihood and protection against all sorts of misfortunes.

5.8 CONCLUSION
When dealing with the institutional prerequisites of an efective market sys-
tem, economists typically focus their attention on property rights and con-
tract enforcement mechanisms. Other social scientists, on the other hand,
have put emphasis on the features of the social fabric needed to sustain the
savings predisposition at the heart of the surge of capitalism. he Weberian
distinction between Gemeinschat and Gesellschat is the classical approach
to characterizing such social features. Because it deals with redistributive
norms as they are applied in (traditional) societies dominated by highly
personalized relationships, the present contribution clearly belongs to the
second strain of thought, although it also points to the existence of interde-
pendent links between the factors highlighted under the two approaches.
hat redistributive pressures exist in sub-Saharan African countries is
evident from a large body of anthropological observations documenting
the widespread fear of ideological intimidation and more painful sanctions
inlicted on successful people who refuse to share the fruits of their economic
success with their kith and kin. Additional evidence is provided by the fact
that when such pressures cannot be resisted bankruptcy tends to follow.
Finally, there is solid ground to believe that African entrepreneurial individu-
als are ready to incur signiicant costs to escape the demands for help emanat-
ing from close and distant relatives. In the same line, it is striking that trade
with relatives and friends is extremely rare in African countries, especially
 Redistributive Pressures in Sub-Saharan Africa 197

because of the risk of serious contract enforcement problems when a deal is

made with an acquaintance. What needs to be emphasized is that the exis-
tence of redistributive or sharing norms has the efect of complicating these
problems to the extent that customers or debtors can easily invoke them as
an excuse for reneging on payment or repayment obligations. In other words,
the disciplining efect of the reputation mechanism that is made possible by
the repeated, personalized nature of human interactions is cancelled by the
adverse efect of the attendant sharing norms.
hat individuals attach a lot of weight to their relative income position
seems to be an increasingly accepted fact, even in economically advanced
countries. It has solid microeconomic foundations, too. Moreover, redis-
tributive pressures exerted on prosperous individuals seem particularly
strong in times of growing socioeconomic diferentiation and when the
degree of personalization of social relationships is high. he Weberian
notion of Gesellschat, which implies anonymous relationships and the use
of rational, eiciency-based criteria, as opposed to the traditional world of
the Gemeinschat, takes on all its meaning here. What African people need
is emancipation from the circle of kinship ties and the associated moral
obligations. hat is precisely what happens when they choose to migrate
physically or symbolically out of their native universe.
When African entrepreneurs or traders avoid dealing with members
of their family or clan, as pointed out earlier, it is typically because they
have put a physical distance between them and their native area. Symbolic
migration occurs when individuals become “internal strangers” vis-à-vis
their own community as a result of conversion to a monotheist religion.
his move enables them to escape their primordial attachments and loyalty
obligations, as well as to adhere to new values more conducive to individ-
ual advancement. More precisely, the new converts can avail themselves of
a formidable shelter against redistributive pressures emanating from kith
and kin, while simultaneously practicing the virtues of frugality, honesty,
and hard work that encourage saving, risk taking, and improved contract
enforcement and lexibility.
Universal religions with a powerful God as igurehead have thus opened
up to entrepreneurial Africans a way to fulill the new expectations that
they have themselves raised. It is a well-recognized fact, indeed, that edu-
cation, which was pioneered by Christian missionaries, has provided the
chief dynamic of colonial change: by bringing in new skills and securing
opportunities for social diferentiation, it became the main generator of
both mobility and stratiication. It expanded so quickly that most pupils
198 Platteau

were of relatively low status, even in strongly stratiied precolonial socie-

ties. herefore, and in contrast to India, for example, “African education at
this time did more to foster social mobility than to entrench old privileged
classes, largely because tropical Africa had no long-established literate elites
except Muslim clerics” (Ilife 2007, 230). Christianity’s direct association
with literacy and education, which included the learning of European lan-
guages and the acquisition of other skills required for employment in the
state sector, explained much of its success during the irst half of the twenti-
eth century. he promise of social success and economic advancement was
especially attractive for the young, the poor, and the female, who saw in it
a refuge and an avenue for self-emancipation: like communism in twenti-
eth-century Asia, Christianity represented an opportunity for generational
revolt and a way out of the repressive customs of the precolonial society
(Ellis and Ter Haar 2004, 138; Ilife 2007, 232). Islam, on the other hand, has
historically been associated with successful trading and a long tradition of
Arabic education that gave rise to intellectual elites.
Universal religions in Africa do not only instill into people values and
norms that violate customs with adverse efects on capital accumulation
and entrepreneurship. hey also nurture feelings of loyalty and identity
toward a large reference group that exceeds ethnic boundaries, and by
doing so they may help to gradually establish the basis for a generalized
morality and an inclusive civic polity backed by pan-lineage and pan-ethnic
institutions. In principle, this efect could also arise from other sources, in
particular industrialization and urbanization, that lead to the creation of
large companies, the formation of social class identities, and the unfolding
of social struggles, or nationalist ideologies and movements that broaden
the people’s image of social reality and help fashion a new collective subject
articulated to an independent state and belief in a common destiny. For
complex reasons, however, these alternative mechanisms of emancipation
from the web of obligations inherent in the traditional universe of social
relationships could not be efectively used in most countries of sub-Saharan
Africa. Various obstacles to investment in addition to redistributive norms
and pressures, on the one hand, and the short duration of many indepen-
dence movements and their domination by local and factional interests, on
the other hand, help explain this situation.
A direct consequence of the failure of nationwide social and political
movements or ideologies is that a big void is let for religions to ill up in
Africa. It would nonetheless be a mistake to think that the recent rapid
lowering of Pentecostal churches and other revivalist or charismatic move-
ments born outside the ambit of established churches is necessarily a change
 Redistributive Pressures in Sub-Saharan Africa 199

that goes into the right direction.27 It might just relect the ability of unscru-
pulous and cynical entrepreneurs posing as inspired preachers or prophets
to exploit the gullibility of poor, illiterate people who are especially vulner-
able to charismatic leadership in times of prolonged economic and political
crisis. Far from the ascetic morality that characterized mainline churches a
few decades earlier, the new evangelists tend to be people who, by cunning
and trickery, have succeeded in becoming immensely rich while deceptively
promising to their followers that the same luck is soon going to befall them.
heir success has much to do with their professed belief in the active pres-
ence of the Holy Spirit, and their alleged ability to cast out demons. Indeed,
for people brought up in magical universes inhabited by spirits who are
represented as real individual beings and with whom some form of efective
communication is possible and essential, this feature ofers almost irresist-
ible attraction (Platteau 2009, 686–7).
Finally, it bears stressing that individual advancement does not require
complete autonomization because the individual needs to join new net-
works of people (church parishes, religious sodalities, migrant or business
associations) to efectively respond to the challenge of modernization and
development. As pointed out previously, a process of individual emanci-
pation from erstwhile kin-based circles and all the attendant obligations
must take place. To promote rather than hinder modern economic growth,
networks and associations must be freely formed and joined by individuals
instead of being imposed on them.

References
Abraham, A. and J. P. Platteau. (2004). “Participatory Development: Where Culture
Creeps In.” In Culture and Public Action, edited by V. Rao and M. Walton, 210–33.
Stanford, CA: Stanford University Press.
Andreoni, J. (1990). “Impure Altruism and Donations to Public Goods: A heory of
Warm-Glow Giving.” Economic Journal, 100: 464–77.
Aoki, M. (2001). Toward a Comparative Institutional Analysis. Cambridge, MA and
London: MIT Press.
Ardener, E. (1970). “Witchcrat, Economics and the Continuity of Belief.” In Witchcrat
Confessions and Accusations, edited by M. Douglas, 141–60. London: Tavistock
Publications.
Austen, R. (1987). African Economic History. London and Portsmouth: James Currey
and Heinemann.

27
Today, more than 400 million Africans, about one-third of the whole African popula-
tion, describe themselves as born-again Christians, compared to 17 million in 1970 (he
Economist 2010, 35).
200 Platteau

Austin, G. (1993). “Indigenous Credit Institutions in West Africa, c.1750-c.1960.” In

Local Suppliers of Credit in the hird World, 1750–1960, edited by G. Austin and K.
Sugihara, 93–159. New York: St. Martin’s Press.
(2008). “he ‘Reversal of Fortune’ hesis and the Compression of History: Perspectives
from African and Comparative Economic History.” Journal of International
Development,20: 996–1027.
Badini, A. (1994). Naître et grandir chez les Moosé traditionnels. Paris & Ouagadougou:
Sépia-A.D.D.B.
Baland, J. M., C. Guirkinger, and C. Mali. (2011). “Pretending to be Poor: Borrowing
to Escape Forced Solidarity in Credit Cooperatives in Cameroon.” Economic
Development and Cultural Change, 60(1): 1–16.
Baland, J. M., R. Somanathan, and Z. Wahhaj. (2013). “Repayment Incentives and the
Distribution of Gains from Group Lending.” Journal of Development Economics,
105: 131–9.
Basu, K. (2000). Prelude to Political Economy − A Study of the Social and Political
Foundations of Economics. Oxford: Oxford University Press.
Basu, K., E. Jones, and E. Schlicht. (1987). “he Growth and Decay of Custom: he
Role of the New Institutional Economics in Economic History.” Explorations in
Economic History, 24(1): 1–21.
Bates, R. H. (2001). Prosperity and Violence. New York: W. W. Norton & Company.
(2008). When hings Fell Apart – State Failure in Late-Century Africa. Cambridge:
Cambridge University Press.
Belshaw, C. (1965). Traditional Exchange and Modern Markets. Englewood Clifs, NJ:
Prentice-Hall.
Berman, B. (1984). “Structure and Process in the Bureaucratic States of Colonial Africa.”
Development and Change,15: 161–202.
Berman, B., D. Eyoh, and W. Kymlicka. (2004). “Ethnicity and the Politics of Democratic
Nation-Building in Africa.” In Ethnicity and Democracy in Africa, edited by B.
Berman, D. Eyoh, and W. Kymlicka, 1–21. Oxford: James Currey and Athens: Ohio
University Press.
Berry, S. (1993). No Condition is Permanent – he Social Dynamics of Agrarian Change
in Sub-Saharan Africa. Madison, Wisconsin and London: University of Wisconsin
Press.
Beveridge, A. A. and A. R. Oberschall. (1979). African Businessmen and Development in
Zambia. Princeton, NJ: Princeton University Press.
Bigsten, A., P. Collier, S. Dercon, M. Fafchamps, B. Gauthier, J. W. Gunning, A. Isaksson,
A. Oduro, R. Oostendorp, C. Patillo, M. Soderbom, F. Teal, and A. Zeufack (2000).
Contract Flexibility and Dispute Resolution in African Manufacturing. Journal of
Development Studies, 36(4): 1–37.
Bourdieu, P. (1990). he Logic of Practice. Stanford, CA: Stanford University Press.
Breman, J. (1974). Patronage and Exploitation: Changing Agrarian Relations in South
Gujarat. Berkeley : University of California Press.
Brooks, G. (1993). Landlords and Strangers: Ecology, Society and Trade in Western Africa
1000–1630. Boulder, CO: Westview.
Bruce, J. W. (1993). “Do Indigenous Tenure Systems Constrain Agricultural
Development?” In Land in African Agrarian Systems, edited by T. J. Bassett and D.
E. Crummey, 35–56. Madison: University of Wisconsin Press.
 Redistributive Pressures in Sub-Saharan Africa 201

Bureau, R. (1996). Le peuple du leuve – sociologie de la conversion chez les Douala. Paris:
Karthala.
Chabal, P. and J. P. Daloz. (1999). Africa Works: Disorder as Political Instrument. Oxford:
James Currey.
Clark, A. (2008). “Happiness, Habits and High Rank: Comparisons in Economic and
Social Life.” Paris School of Economics, mimeo.
Clark, A., P. Frijters, and M. A. Shields. (2008). “Relative Income, Happiness, and Utility:
An Explanation for the Easterlin Paradox and Other Puzzles.” Journal of Economic
Literature, 46(1): 95–144.
Coate, S. and M. Ravallion. (1993). “Reciprocity without Commitment: Characterization
and Performance of Informal Insurance Arrangements.” Journal of Development
Economics, 40(1): 1–24.
Cohen, A. (1969). Custom and Politics in Urban Africa: A Study of Hausa Migrants in
Yoruba Towns. Berkeley : University of California Press.
(1971). “Cultural Strategies in the Organization of Trading Diasporas.” In he
Development of Indigenous Trade and Markets in West Africa, edited by C.
Meillassoux, 266–81. London: Oxford University Press.
Colin, R. (2004). Kénédougou – Au crépuscule de l’Afrique coloniale. Paris: Présence
Africaine.
Congleton, R. (1980). “Competitive Process, Competitive Waste, and Institutions.” In
Toward a heory of the Rent-Seeking Society, edited by J. Buchanan, R. Tollison, and
G. Tullock. Texas A&M Press.
Cooper, B. M. (1997). Marriage in Maradi – Gender and Culture in a Hausa Society in
Niger, 1900–1989. Portsmouth: Heinemann and London: James Currey.
Dercon, S. and Krishnan, P. (2000) “In Sickness and in Health: Risk-Sharing within
Households in Ethiopia.” Journal of Political Economy, 108(4): 688–727.
Duesenberry, J. (1949). Income, Saving, and the heory of Consumer Behavior. Cambridge,
MA and London: Harvard University Press.
Easterlin, R. (1974). “Does Economic Growth Improve the Human Lot? Some Empirical
Evidence.” In Nations and Households in Economic Growth: Essays in Honor of Moses
Abramovitz, edited by R. David and R. Reder, 89–125. New York: Academic Press.
Economist, he (2010). “Slain by the Spirit” (pp. 35–6). July 3.
Ellis, S. and J. Y. Fauré. (1995). Entreprises et Entrepreneurs Africains. Paris: Karthala-
ORSTOM.
Ellis, S. and G. Ter Haar. (2004). Worlds of Power. Religious hought and Political Practice
in Africa. London: Hurst & Company.
Elster, J. (1989). he Cement of Society − A Study of Social Order. Cambridge: Cambridge
University Press.
Ensminger, J. (1992). Making a Market – he Institutional Transformation of an African
Society. Cambridge: Cambridge University Press.
Evans-Pritchard, E. E. (1937). Witchcrat, Oracles and Magic among the Azande. Oxford
and New York: Oxford University Press.
(1940). he Nuer: A Description of the Modes of Livelihood and Political Institutions of
a Nilotic People. Oxford: Clarendon Press.
(1964). Closed Systems and Open Minds. Chicago: Aldine Publishing Company.
Fafchamps, M. (1992). “Solidarity Networks in Preindustrial Societies: Rational
Peasants with a Moral Economy.” Economic Development and Cultural Change,
41(1): 147–74.
202 Platteau

(1996). “he Enforcement of Commercial Contracts in Ghana.” World Development,

24: 427–48.
(2000). “Ethnicity and Credit in African Manufacturing.” Journal of Development
Economics, 61: 205–35.
(2003). “Ethnicity and Networks in African Trade. Contributions to Economic
Analysis and Policy.” Berkeley Electronic Press, 2(1): Article 14.
(2004). Market Institutions in Sub-Saharan Africa – heory and Evidence. Cambridge,
MA and London: MIT Press.
(2010). “Markets and the Difusion of Institutional Innovations.” In Culture,
Institutions, and Development – New Insights into an Old Debate, edited by J. P.
Platteau and R. Peccoud, 143–59. London and New York: Routledge.
Fafchamps, M. and B. Minten (1999). “Relationships and Traders in Madagascar.”
Journal of Development Studies, 35: 1–35.
Fafchamps, M. and B. Minten. (2002). “Returns to Social Network Capital among
Traders.” Oxford Economic Papers, 54: 173–206.
Fage, J. D. (1995). A History of Africa, 3rd ed. London and New York: Routledge.
Fancello, S. (2006). Les Aventuriers du Pentecôtisme Ghanéen. Paris: Editions Karthala.
First, R. (1970). he Barrel of a Gun – Political Power in Africa and the Coup d’Etat.
London: Allen Lane & the Penguin Press.
Firth, R. (1951). Elements of Social Organization. London: Watts.
Fisiy, C. and P. Geschiere. (1991). “Sorcery, Witchcraft and Accumulation: Regional
Variations in South and West Cameroon.” Critique of Anthropology, 11:
251–78.
Fosu, A. K. (2008). “Anti-Growth Syndromes in Africa: A Synthesis of the Case Studies.”
In he Political Economy of Economic Growth in Africa, 1960–2000, edited by B. J.
Ndulu, S. A. O’Connell, R. H. Bates, P. Collier, and C. C. Soludo, 137–72. Cambridge:
Cambridge University Press.
Frank, R. (1985). Choosing the Right Pond – Human Behavior and the Quest for Status.
New York and Oxford: Oxford University Press.
(1998). “Social Norms as Positional Arms Control Agreements.” In Economics, Values,
and Organization, edited by A. Ben-Ner and L. Putterman, 275–95. Cambridge:
Cambridge University Press.
Gaspart, F. and E. Seki. (2003). “Cooperation, Status Seeking and Competitive Behaviour:
heory and Evidence.” Journal of Economic Behaviour and Organization, 51(1):
51–77.
Geertz, C. (1973). he Interpretation of Cultures. New York: Basic Books.
Geschiere, P. (1994). “Parenté et argent dans une société lignagère.” In La réinvention du
capitalisme, edited by J. F. Bayart, 87–113. Paris: Karthala.
(1995). Sorcellerie et politique en Afrique – La viande des autre. Paris: Editions
Karthala.
Giblin, J. L. (2005a). A History of the Excluded – Making Family a Refuge from State
in Twentieth-Century Tanzania. Oxford: James Currey, Dar Es Salaam: Mkuki na
Nyota, and Athens: Ohio University Press.
(2005b). “Some Complexities of Family and State in Colonial Njombe.” In In Search
of a Nation – Histories of Authority and Dissidence in Tanzania, edited by G. H.
Maddox and J. L. Giblin, 128–48. Oxford: James Currey, Dar Es Salaam: Kapsel
Educational Publications, and Athens: Ohio University Press.
 Redistributive Pressures in Sub-Saharan Africa 203

Godelier, M. (1974). “Anthropologie et économie. Une anthropologie économique est-

elle possible?” In Un domaine contesté: l’anthropologie économique, edited by M.
Godelier, 285–345. Paris: Mouton.
Golooba-Mutebi, F. (2005). “Witchcrat, Social Cohesion and Participation in a South
African Village.” Development and Change,36: 937–58.
Gregoire, E. (1993). “Islam and the Identity of Merchants in Maradi (Niger).” In Muslim
Identity and Social Change in Sub-Saharan Africa, edited by L. Brenner, 106–15.
Bloomington: Indiana University Press.
Greif, A., 1994. “Cultural Beliefs and the Organization of Society: A Historical and
heoretical Relection on Collectivist and Individualist Societies.” Journal of
Political Economy, 102(5): 912–50.
Greif, A. (2006). Institutions and the Path to the Modern Economy – Lessons from
Medieval Trade. Cambridge: Cambridge University Press.
Gyekye, K. (1996). African Cultural Values − An Introduction. Accra, Ghana: Sankofa
Publishing Company.
Hagen, E. E. (1975). he Economics of Development. Homewood, IL: Richard D. Irwin.
Harragin, S. (2004). “Relief and an Understanding of Local Knowledge: he Case of
Southern Sudan.” In Culture and Public Action, edited by V. Rao and M. Walton,
307–27. Stanford, CA: Stanford University Press.
Hart, K. (1975). “Swindler or Public Benefactor? he Entrepreneur in his Community.”
In Changing Social Structure in Ghana: Essays in Comparative Sociology of a New
State and an Old Tradition, edited by J. Goody. London: International African
Institute.
Haugerud, A. (1993). he Culture of Politics in Modern Kenya. Cambridge: Cambridge
University Press.
Hayami, Y. and T. Kawagoe. (1993). he Agrarian Origins of Commerce and Industry:
A Study of Peasant Marketing in Indonesia. London: Macmillan and New York: St.
Martin’s Press.
Himbara, D. (1994). Kenyan Capitalists, the State, and Development. Nairobi: East
African Educational Publishers.
Hirschman, A. O. (1958). he Strategy of Economic Development. New Haven, CT and
London: Yale University Press.
Hof, K. and A. Sen. (2005). “he Extended Family System and Market Interactions.”
In he Social Economics of Poverty – On Identities, Communities, Groups, and
Networks, edited by C. B. Barrett, 171–87. London and New York: Routledge.
Hopkins, A. G. (1973). An Economic History of West Africa. London: Longman.
Hunter, G. (1969). Modernizing Peasant Societies: A Comparative Study in Asia and
Africa. New York and London: Oxford University Press.
Ilife, J. (2005). “Breaking the Chain at its Weakest Link – TANU and the Colonial
Oice.” In In Search of a Nation – Histories of Authority and Dissidence in Tanzania,
edited by G. H. Maddox and J. L. Giblin, 168–97. Oxford: James Currey, Dar Es
Salaam: Kapsel Educational Publications, and Athens: Ohio University Press.
(2007). Africans – he History of a Continent. Cambridge: Cambridge University
Press, irst published 1995.
Jakiela, P. and O. Ozier. (2011). “Does Africa Need a Rotten Kin heorem? Experimental
Evidence from Village Economies.” Working Paper, Department of Economics,
Washington University, St. Louis and University of California.
204 Platteau

Jewsiewicki, B. (1993). Naître et Mourir au Zaïre – Un Demi-siècle d’Histoire au Quotidien.

Paris: Editions Karthala.
Jones, B. (2009). Beyond the State in Rural Uganda. Edinburgh: Edinburgh University
Press.
Jul-Larsen, E. (1993). “he Micro-state of the Popo in Pointe Noire. An Analysis of
Local Governance, State-Society Relations and Access to Resources.” Centre for
Development Studies, University of Bergen Norway, mimeo.
Kamau, J. (1965). “Problems of African Business Enterprise.” Center for Economic
Research, University of Nairobi, Discussion Paper N° 6.
Kanogo, T. (1987). Squatters and the Roots of Mau Mau. London: James Currey and
Nairobi: East African Educational Publishers.
Karahasan, D. (2000). L’âge de sable. Paris: Robert Lafont.
Kennedy, P. (1988). African Capitalism – he Struggle for Ascendency. Cambridge:
Cambridge University Press.
Kimball, M. S. (1988). “Farmers’ Cooperatives as Behavior toward Risk.” American
Economic Review, 78(1): 224–32.
Kohnert, D. (1996). “Magic and Witchcrat: Implications for Democratization and
Poverty-Alleviating Aid in Africa.” World Development,24: 1347–55.
Knight, J. and L. Song. (2006). “Subjective Well-being and Its Determinants in Rural
China.” University of Nottingham, mimeo.
Kuran, T. (2003). “he Islamic Commercial Crisis: Institutional Roots of Economic
Underdevelopment in the Middle East.” Journal of Economic History,63: 414–46.
(2004a). “Cultural Obstacles to Economic Development: Oten Overstated, Usually
Transitory.” In Culture and Public Action, edited by V. Rao and M. Walton, 115–37.
Stanford, CA: Stanford University Press.
(2004b). “Why the Middle East Is Economically Underdeveloped: Historical
Mechanisms of Institutional Stagnation.” Journal of Economic Perspectives, 18(3):
71–90.
(2011). he Long Divergence – How Islamic Law Held Back the Middle East. Princeton,
NJ and Oxford: Princeton University Press.
Kuznets, S. (1966). Modern Economic Growth. New Haven, CT: Yale University Press.
de Laiglesia, J. R. and C. Morrisson. (2008). Household Structures and Savings: Evidence
from Household Surveys. OECD Development Centre, Working Paper N° 267.
de Lame, D. (1996). “Une colline entre mille ou le calme avant la tempête –
Transformations et blocages du Rwanda rural.” Musée Royal de L’Afrique Centrale,
Annales Sciences Humaines 154. Tervuren (Belgium).
Launay, R. (1992). Beyond the Stream: Islam and Society in a West African Town. Berkeley :
University of California Press.
Laurent, P. J. (1998). Une association de développement en pays mossi. Paris: Karthala.
Layard, R. (2005). Happiness: Lessons from a New Science. London: Allen Lane.
Levi-Strauss, H. (1969). he Elementary Structures of Kinship. London: Social Science
Paperbacks (revised edition).
Levtzion, N. (1973). Ancient Ghana and Mali. New York: Africana Publishing Press.
Lewis, W. A. (1955). he heory of Economic Growth. London: George Allen and
Unwin.
Ligon, E., J. P. homas, and T. Worrall. (2002). “Informal Insurance with Limited
Commitment: heory and Evidence from Village Economies.” Review of Economic
Studies, 69(1): 209–44.
 Redistributive Pressures in Sub-Saharan Africa 205

Long, N. (1968). Social Change and the Individual. Manchester: University of Manchester
Press.
(1977). An Introduction to the Sociology of Rural Development. London and New York:
Tavistock Publications.
Lydon, G. (2009). “A Paper Economy of Faith without Faith in Paper: A Relection on
Islamic Institutional History.” Journal of Economic Behavior and Organization,
71(3): 647–59.
Macfarlane, A. (1978). he Origins of English Individualism. Oxford: Blackwell.
Malinowski, B. (1922). Argonauts of the Western Paciic − An Account of Native Enterprise
and Adventure in the Archipelagoes of Melanesian New Guinea. London: Routlege
(reprinted 1999).
Maloba, W. O. (1998). Mau Mau and Kenya – An Analysis of a Peasant Revolt. Oxford:
James Currey and Bloomington: Indiana University Press.
Marris, P. and A. Somerset. (1971). African Businessmen. London: Routledge and
Kegan Paul.
Mauss, M. (1925). he Git. London: Routledge and Kegan Paul.
Melenberg, B. (1992). “Micro-econometric Models of Consumer Behaviour and
Welfare.” Tilburg University, mimeo.
Meyer, B. (1998). “Make a Complete Break with the Past: Memory and Post-Colonial
Modernity in Ghanaian Pentecostalist Discourse.” Journal of Religion in Africa,
27(3): 316–49.
(2004). “Christianity in Africa: From African Independent to Pentecostal-Charismatic
Churches.” Annual Review of Anthropology, 33: 447–74.
Migeotte, F. (1995). “Innovation institutionnelle et tensions sociales: le cas de
l’intensiication de l’élevage sur une colline rwandaise.” Musée Royal d’Afrique
Centrale. Tervuren (Belgium).
Minten, B. and S. Kyle. (1999). “he Efect of Distance and Road Quality on Food
Collection, Marketing Margins, and Traders’ Wages: Evidence from the Former
Zaïre.” Journal of Development Economics,60: 467–95.
Miran, M. (2006). Islam, Histoire et Modernité en Côte d’Ivoire. Paris: Editions
Karthala.
Monson, J. (2005). “he Tribal Past and the Politics of Nationalism in Mahenge District,
1940–60.” In In Search of a Nation – Histories of Authority and Dissidence in
Tanzania, edited by G. H. Maddox and J. L. Giblin, 103–13. Oxford: James Currey,
Dar Es Salaam: Kapsel Educational Publications, and Athens: Ohio University
Press.
Moore, M. (1997). “Societies, Polities and Capitalists in Developing Countries: A
Literature Survey.” Journal of Development Studies, 33(3): 287–363.
Morduch, J. (1999). “Between the Market and the State: Can Informal Insurance Patch
the Safety Net?” World Bank Research Observer, 14(2): 187–207.
Nafziger, E. W. (1977). African Capitalism: A Case Study of Nigerian Entrepreneurship.
Stanford, CA: Hoover Institution Press.
Ndulu, B. J. and S. A. O’Connell. (2008). “Policy Plus: African Growth Performance
1960–2000.” In he Political Economy of Economic Growth in Africa, 1960–2000,
edited by B. J. Ndulu, S. A. O’Connell, R. H. Bates, P. Collier, and C. C. Soludo,
3–75. Cambridge: Cambridge University Press.
Niehaus, I. (2001a). Witchcrat, Power and Politics: Exploring the Occult in the South
African Lowveld. Capetown: David Philips.
206 Platteau

(2001b). “Witchcrat in the New South Africa: From Colonial Superstition to

Postcolonial Reality.” In Magical Interpretations, Material Realities: Modernity,
Witchcrat and the Occult in Post-Colonial Africa, edited by H. L. Moore and T.
Sanders, 184–205. London and New York: Routledge.
North, D. C. (1990). Institutions, Institutional Change and Economic Performance.
Cambridge: Cambridge University Press.
Ofer, A. (1997). “Between the Git and the Market: he Economy of Regard.” Economic
History Review, 50(3): 450–76.
Parkin, D. J. (1972). Palm, Wine and Witnesses. London: Intertext Books.
Peters, P. E. (1994). Dividing the Commons − Politics, Policy, and Culture in Botswana.
Charlottesville and London: University Press of Virginia.
Platteau, J. P. (1991). “Traditional Systems of Social Security and Hunger Insurance: Past
Achievements and Modern Challenges.” In Social Security in Developing Countries,
edited by E. Ahmad, J. Drèze, J. Hills, and A. K. Sen, 112–70. Oxford: Clarendon
Press.
(1994). “Behind the Market Stage Where Real Societies Exist – Part II: he Role of
Moral Norms.” Journal of Development Studies, 30(4): 753–817.
(2000). Institutions, Social Norms and Economic Development. London: Routledge.
(2004). “he Gradual Erosion of the Social Security Function of Customary Land
Tenure Arrangements: he Case of Tribal Societies in SubSaharan Africa.” In
Insurance against Poverty, edited by S. Dercon, 247–78. Oxford: Clarendon Press.
(2009). “Institutional Obstacles to African Economic Development: State, Ethnicity,
and Custom.” Journal of Economic Behavior and Organization, 71(3): 669–89.
(2011). “he Role of Culture in Development: An Overview.” In Culture, Institutions,
and Development – New Insights into an Old Debate, edited by J. P. Platteau and R.
Peccoud, 3–19. London: Routledge.
Platteau, J. P. and A. Abraham (2002). “Participatory Development in the Presence of
Endogenous Community Imperfections.” Journal of Development Studies, 39(2):
104–36.
Platteau, J. P. and R. Peccoud. (2011). “Revisiting the Role of Culture.” In Culture,
Institutions, and Development – New Insights into an Old Debate, edited by J. P.
Platteau and R. Peccoud, 247–58. London: Routledge.
Platteau, J. P. and P. Sekeris. (2010). “On the Feasibility of Power and Status Ranking in
Traditional Setups.” Journal of Comparative Economics, 28(3): 267–82.
Platteau, J. P. and E. Seki. (2007). “Heterogeneity, Social Esteem, and Feasibility of
Collective Action.” Journal of Development Economics,83: 302–25.
Poewe, K. (1989). Religion, Kinship, and Economy in Luapula, Zambia. Lewinston:
Edwin Mellen Press.
Polanyi, K. (1944). he Great Transformation. New York: Rinehart.
(1968). Primitive, Archaic, and Modern Economics. New York: Doubleday.
(1977). he Livelihood of Man. New York and London: Academic Press.
Redield, R. (1962). he Little Community. Chicago: Aldine.
Rennie, J. K. (1984). “Ideology and State Formation: Political and Communal Ideologies
among the South-Eastern Shona, 1500–1890.” In State Formation in Eastern
Africa, edited by A. I. Salim, 162–94. Nairobi, London, and Ibadan: Heinemann
Educational Books.
 Redistributive Pressures in Sub-Saharan Africa 207

De Rosny, E. (1981). Les yeux de ma chèvre. Sur les pas des maîtres de la nuit en pays
douala. Paris: Plon.
Rowlands, M. and Warnier, J. P. (1988). “Sorcery, Power, and the Modern State in
Cameroon.” Man,23: 118–32.
Sahlins, M. (1960). “Political Power and the Economy in Primitive Society.” In Essays in
the Science of Culture in Honor of Leslie White, edited by G. Dole and R. Carneiro.
New York: Crowell.
(1963). “Poor Man, Rich Man, Big Man: Political Types in Melanesia and Polynesia.”
Comparative Studies in Society and History, 5: 285–303.
(1972). Stone Age Economics. Chicago: Aldine.
Sarro, R. (2009). he Politics of Religious Change on the Upper Guinea Coast – Iconoclasm
Done and Undone. Edinburgh: Edinburgh University Press.
Scott, J. C. (1976). he Moral Economy of the Peasant – Rebellion and Subsistence in
Southeast Asia. New Haven, CT and London: Yale University Press.
(1985). he Weapons of the Weak – Everyday Forms of Peasant Resistance. New Haven,
CT and London: Yale University Press.
Shillington, K. (1989). History of Africa. New York: St. Martin’s Press.
Somville, V. (2010). “Daily Collectors & Public Good Provision in Non-Cooperative
Households: heory and Evidence from Urban Benin.” Centre for Research in the
Economics of Development (CRED), University of Namur, mimeo.
Sylla, A. (1994). La philosophie morale des Wolof, IFAN, Université de Dakar.
homas, K. (1973). Religion and the Decline of Magic. Harmondsworth: Penguin
Books.
Townsend, R. M. (1994). “Risk and Insurance in Village India.” Econometrica, 62(3):
539–91.
Udry, C. (1994). “Risk and Insurance in a Rural Credit Market: An Empirical Investigation
in Northern Nigeria.” Review of Economic Studies,61: 495–526.
Van de Walle, N. (2001). African Economies and the Politics of Permanent Crisis, 1979–
1999. Cambridge: Cambridge University Press.
Veblen, T. (1899). he heory of the Leisure Class. London: George Allen and Unwin.
Von Braun, J. and P. Webb (1989). “he Impact of New Crop Technology on the
Agricultural Division of Labor in a West African Setting.” Economic Development
and Cultural Change, 37(3): 513–34.
De Vries, J. (1976). he Economy of Europe in the Age of Crisis, 1600–1750. Cambridge:
Cambridge University Press.
Wright, R. (1994). he Moral Animal – Why We Are the Way We Are: he New Science of
Evolutionary Psychology. New York: Pantheon Books.
 6

Accumulation and Conspicuous Consumption:

he Poverty of Entrepreneurship in Western
Nigeria, ca. 1850–1930

Ayodeji Olukoju

he failure to develop a concept of business as an impersonal and long-term occu-

pation helps to explain why, in spite of its many attributes, the African business
community in Lagos was unable to produce a capitalist class capable of competing
successfully with expatriate irms. (Hopkins 1964, 444)

6.1 INTRODUCTION
he underdevelopment of West African entrepreneurship, epitomized by
the near-total absence of an indigenous capitalist class, alluded to here
by Professor A. G. Hopkins, has been examined within the framework
of divergent explanations – cultural versus structural.1 In the same vein,
Gareth Austin, pondering the “basic question . . . [of] the apparent rarity
of transgenerational continuity of speciic businesses in West Africa,”2 sug-
gested any of three possibilities for further investigation: the institutional
(cultural issues, such as inheritance practices); the general instability of
economic life (the vicissitudes and risks associated with enterprises); and
the changes wrought by colonial intervention. Against the background of

I thank the editors of this volume and their anonymous readers, as well as Kenneth Nwoko,
for comments on an earlier drat of this chapter.
1
Hopkins (1964, 1978, 1988) and Akeredolu-Ale (1975) examine the conlicting interpreta-
tions. hough reference will be made to “the poor” (akuse, ol òsì, and tálíkà), the subject
of the John Ilife-John Peel exchanges (see Journal of African History, 32, 1991, 495–500;
501–6) on poverty in nineteenth-century Yorubaland, this chapter focuses on the mer-
cantile and political elite (bòròkìní, olówó, olólà, gbajúmò, etc.), rather than the “poor,” to
explain the poverty of development in the region. he underlying assumption in this chap-
ter, which it seeks to demonstrate, is that Yoruba entrepreneurs, unlike their counterparts
elsewhere, failed to seize the moment and drive the process of societal development to its
logical conclusion.
2
Austin (2002, 122).

208
 Accumulation and Conspicuous Consumption 209

the propositions by Hopkins and Austin, this chapter argues, irst, that the
cultural context was as important as the constraining structure of British
colonialism in explaining the underdevelopment of indigenous entre-
preneurship, and, second, that both dynamics should be taken together
in any meaningful explanation. he importance of the cultural dynamic,
highlighted in this chapter, is anchored on indigenous Yoruba conceptions
of accumulation, wealth, and social status.
First, the Yoruba distinguished among the poor (akuse, olòsì, and tálíkà),
the rich (olówó), the wealthy (olórò or olólà), and the honorable person
(olólá). Second, they esteemed the wealthy, who had property and assets
(olórò or olólà) above the possessor of mere cash (olówó). hird, as soci-
ety also prioritized honor, a person who combined honor and wealth with
generosity – the bòròkìní or gbajúmò – was the ideal entrepreneur (Falola
1984, 71).3 he proiles of nineteenth- and twentieth-century Yoruba entre-
preneurs in Lagos and its hinterland show that, with few exceptions, the
aspiring bòròkìní or gbajúmò eagerly sought to earn societal approval.
he second major concern of this chapter is the chronology of the decline
of indigenous entrepreneurship. It attempts to show how, why, and when
exactly this region missed the proverbial boat of development. It argues that
the period 1880–1920 was critical to the failure of Yoruba entrepreneurs to
take of to the stage of developmental entrepreneurship. hat period was
characterized by two concurrent developments: the transition to full colo-
nial rule – with its constraints and opportunities – and at least two cycles of
boom – 1906–14 and 1918–20. It was the failure to maximize the opportu-
nities of this period and to surmount the institutional and structural disad-
vantages of the colonial context, coupled with anti-entrepreneurial societal
expectations, that doomed Yoruba entrepreneurship under British rule. he
lamentation of a boom-era observer while capital was being frittered away
without sparing a thought for the future speaks volumes:
Do we realize that whatever money we get now on our palm oil and palm kernels is
our capital and that in spending it outright we are spending our capital? . . . [We] are
frittering away with a hopeless want of foresight, rather characteristic of our people,
that capital which alone can give us economic liberty in the future. . . . We sound the
grave warning that our future position in this country will depend on how much of
our own capital is invested in its development.4

3
hat such values and expectations were not peculiar to Yorubaland is indicated in Zakaria
(1997, 63).
4
Lagos Weekly Record, September 27, 1919, editorial: “Whither are we going?” here was an
undercurrent of economic nationalism in this remark. A recent study is Olukoju (2010).
See also Olukoju (2002b).
210 Olukoju

Against this background, this chapter examines the dimensions of the pov-
erty5 of indigenous entrepreneurship in western Nigeria, the homeland of
the Yoruba, between 1850 and 1930. It details how chiefs, warriors, and
traders (both male and female) accumulated and dissipated wealth (with
particular emphasis on elite lifestyle and investment patterns). hough the
politico-economic contexts were diferent – warfare in the irst and colo-
nialism in the second – this chapter highlights some striking parallels and
contrasts.
In all, this chapter grapples with the following questions: Why did indig-
enous accumulation (admittedly in the face of local and external counter-
vailing forces, especially domestic feudalism, oligarchic rule, and Western
imperialism) fail to generate the societal transformation witnessed in, for
example, pre-1952 Egypt, founded in part on mercantile enterprise? How
did these entrepreneurs respond to changes in the local and wider political
economy? What was the developmental and long-term impact of western
Nigerian entrepreneurs on their society – against the demands of social
obligations – in terms of diversiication into noneconomic enterprises? Did
the business elite really fail in view of the long-term consequences of their
careers?
his chapter deals with these questions by analyzing case studies of nota-
ble entrepreneurs in the Yoruba urban and commercial centers of Ibadan,
Ijaye, Abeokuta, and Lagos in their peculiar politico-economic settings.6
It draws on oicial sources, contemporary newspapers, and private papers
in the archives and libraries in Ibadan and Lagos. Newspapers and private
correspondence provide contemporary insights into the issues and person-
alities examined in the rest of this chapter.
Business biographies provide valuable insights into the dynamics of
entrepreneurship. Coupled with contemporary documentary and oral evi-
dence, they help in understanding the contexts in which the protagonists
operated, and how they coped with the challenges and utilized the oppor-
tunities available to them. his is amply illustrated by the career of Daniel
Conrad Taiwo (1810s?–1901) (a.k.a. “Taiwo Olowo” – “Taiwo the Rich”),
whose biographical sketch is one of the case studies that underpin this chap-
ter. Of immediate importance is his striking epitaph in a Lagos newspaper:

5
his refers to its failure to turn capital into an instrument of structural change in society
and to build commercial houses with a record of intergenerational succession and transi-
tion from the precolonial or colonial past to the present.
6
here are several biographical sketches and in-depth micro-histories of the business and
sociopolitical activities of many of the selected personalities. hey provide important
source material for this chapter.
 Accumulation and Conspicuous Consumption 211

Chief Taiwo . . . unquestionably wielded a greater inluence than any other individ-
ual member of the community. . . . [He] was a true type of African “big man.” With
him money became wealth in the fullest sense of the term, for he . . . utilized it . . . and
the result was that his name was a household word in all the surrounding country,
while he was respected and revered far and wide.7 (Italics added for emphasis)

he careers of Chief Taiwo and others cited in this chapter help to situ-
ate the central issue of the underdevelopment of indigenous entrepreneur-
ship – how money was made and utilized – in the longue durée.

6.2 ENTREPRENEURSHIP IN THE CONTEXTS OF

WARFARE AND POLITICAL COMPETITION IN
NINETEENTH-CENTURY YORUBALAND
As is well detailed in the literature, the fall of Old Oyo, the rise of the suc-
cessor states of Ibadan, Ijaye, and Abeokuta, and the protracted warfare
among Yoruba states during the nineteenth century dovetailed into the
British occupation of Lagos between 1851 and 1861 and culminated in the
colonization of Yorubaland (the hinterland of Lagos) by 1893.8 Initially,
these were parallel processes, but both intersected in the fueling of the slave
trade by the interstate conlicts in Yorubaland. Warfare between the south-
ern emirates of the emergent Sokoto Caliphate (Ilorin and Bida) and the
northern Yoruba successor states ater the fall of Old Oyo, and the inter-
state wars among the Yoruba between 1820 and 1893 generated population
movements and insecurity, and the massive export of slaves via the coastal
ports, including Lagos and Badagry. he abolition of the transatlantic slave
trade later provided a pretext for British intrusion into Lagos and, subse-
quently, the Yoruba hinterland. For our purpose, intermittent warfare in
Yorubaland, including the Ibadan-Ijaye, Ibadan-Egba, and Ibadan-Ekiti
Parapo wars, provided the context in which the processes of indigenous
entrepreneurship unfolded between the 1820s and 1893. he internal and
external trade in war captives, and the appropriation of the surplus gener-
ated by free and servile labor, underpinned the entrepreneurial activities
of notable Yoruba political, military, and commercial leaders. heir wealth
consisted in their slaveholdings and the exploitation of slave labor in agri-
culture, trade, and crats to build up chiely power.

7
Lagos Weekly Record, February 23,1901. his is a clear indication of popular perceptions
and deinitions of (the means and ends of) money, aluence and inluence, and what it was
to be a “big man.”
8
See, among others, Akinjogbin (1998), Atanda (1973), and Babayemi (1990).
212 Olukoju

he careers of the leading Yoruba entrepreneurs of the nineteenth cen-

tury – Kurunmi, the military autocrat of the short-lived state of Ijaye;
the Ibadan war chiefs and two notable women entrepreneurs (Efunsetan
Aniwura and Madam Tinubu, who also oiled the military machines of
Ibadan, Abeokuta, and Lagos) – conlate issues of gender, militarism, power
politics, and accumulation during the nineteenth century. In spite of the
diversity of gender, origins, and circumstance, these personalities shared
the following characteristics. First, they lived during the same epoch –
the nineteenth century. Second, they were active players in the interstate
warfare of the period. hird, they operated in the context of precolonial
indigenous autonomy in the Yoruba hinterland. Fourth, they shared similar
experiences of accumulation through slaveholding and through commer-
cial and agricultural enterprise. Fith, they all subscribed to the cultural
ethos of utilizing social networks to build economic and political power.
In Ijaye, the autocratic leader, Kurunmi, and his chiefs amassed wealth
in trade and farming. A British missionary, Henry Townsend, estimated
that Kurunmi alone had extensive plantations that covered all of 100 acres.9
Another contemporary observer, he Reverend S. H. Clarke, noted that
farmlands stretched some ten to twelve miles beyond the city walls. Indeed,
he rode on horseback for an hour through the farmlands cultivated in the
direction of the Upper Ogun River. Such large-scale farming turned Ijaye
into a food basket for the emerging megalopolises at Ibadan and Abeokuta,
havens of large communities displaced by the wars of the 1820s and 1830s.
Men and women engaged in the production of yam, corn, palm oil, veg-
etables, and fruits. Farming was complemented by weaving and cloth dye-
ing, carving, pottery, hunting, and, to a lesser extent, stock breeding. he
products of these economic activities were fed into the local and regional
commercial systems linking Ijaye with Ibadan, Oyo, Ogbomoso, Abeokuta,
Ilorin, and Lagos. he regional market at Ijaye was the destination of more
than twenty thousand people and caravans from the coast, which arrived
three times during the week (Falola 1991a, 26, 1991b).
he foregoing narrative is intended to highlight the basis of accumu-
lation in the indigenous Yoruba political economy during the nineteenth
century. Although Ijaye was short-lived, it was typical of the successor states
of Old Oyo (New Oyo, Ibadan, Ijaye, and the non-Oyo state of Abeokuta) in
terms of how military and political leaders harnessed the resources of state
to accumulate wealth. “On internal capital accumulation,” Falola notes with
reference to Kurunmi, “his strategy was to be the richest and weaken the

9
See Falola (1991a, 24) for the details in this paragraph.
 Accumulation and Conspicuous Consumption 213

access of the other power elite to resources” (Falola 1991a, 27). His wealth
derived from market dues and tolls on caravans, which were eiciently
collected by his oicials, war booty, heavy ines on ofenders, gits, heavy
taxation, and tributes. All proceeds went into the central cofers that were
under his personal control. He controlled the trade routes into the city-state
and monopolized arms and ammunition, which also represented a form of
wealth in material and military terms (Falola 1991a, 27–8).
Kurunmi was an atypical Yoruba military autocrat because he did not
share power with chiefs who represented lineages and quarters as in other
Yoruba kingdoms. But his style of accumulation paralleled that of Ibadan
war chiefs of the nineteenth century. he latter also built their wealth on
slaveholding, large-scale farming, control of trade, levying of tolls, trib-
utes, and court ines. Balogun Ibikunle, the commander-in-chief of the
Ibadan army in the 1850s, was reputed to be a rich man with extensive
farmlands beyond the city walls, much like Kurunmi at Ijaye (Awe 1973,
67, note 4). Awe noted that “one of the distinctive [attributes] . . . of any
(Ibadan) war chief of repute was [the ability] . . . to ight, farm and trade”
(Awe 1973, 67, note 4). his was how they sustained the large body of men
under arms and maintained large households in the metropolis. However,
unlike Kurunmi, the Ibadan war chiefs had greater opportunities to accu-
mulate and to build more enduring military and economic entities based
on their lineages, military forces, and the control of subject territories
because of the longer life span of the Ibadan state and the process of its
expansion into the Ekiti-Ijesa territories. hey employed the ajele system
to administer and plunder the subject territories. But it was the system that
underpinned the Ibadan state that also led to its fall as the excesses of the
ajele (intendants) alienated the subject peoples of eastern Yorubaland and
precipitated the protracted Ekiti Parapo resistance that sounded the death
knell to the empire.
While virtually all the rulers of Ibadan were wealthy in slaves, prop-
erty, and money, the case of a notable woman leader, Efunsetan Aniwura
(1820s–74), the Iyalode (leader of the women), is particularly striking for
its exceptionalism. She was the richest and most inluential woman entre-
preneur of her age, rivaled only by Madam Efunroye Tinubu (ca. 1805–87)
of Lagos, Badary, and Abeokuta fame.10 Both women were of Egba origins,
amassed wealth in the context of warfare, and were deeply involved in the

10
he only book-length account of Madam Tinubu’s life, admittedly written by an ama-
teur historian, is Oladipo Yemitan, Madame Tinubu: Merchant and King-Maker, Ibadan:
University Press Limited, 1987. It contains some useful details.
214 Olukoju

power politics of their era, a male-dominated sphere in which they were

given short shrit.
Married to Adele, a prince and later king (Oba) of Lagos, from 1835–7,
Madam Tinubu was a notable trader in slaves, arms, and ammunition and
a key player in succession politics in Lagos. She was also well established in
her natal home of Abeokuta, where she had a thriving business managed by
sixty slaves and a literate clerk (Awe and Olutoye 1998, 128). Such was the
scale of her involvement in the defense of Abeokuta (against Dahomey) that
she was made the irst Iyalode of the Egba. Of direct relevance is Madam
Tinubu’s accumulation of wealth in land, material assets, and people. She
had extensive landholdings on the island and mainland of Lagos, as well as
in Abeokuta. Many of these holdings survive to this day or have been alien-
ated by her descendants (her children died as minors). By 1855, Madam
Tinubu reportedly sat atop a vast commercial network and “had few rivals
among the traders in Lagos. She had vast properties and power” (Yemitan
1987, 24).
But she was heavily dependent on credit in the running of her business.11
Hence, when her political stock fell and she was expelled from Lagos in
1856, she was indebted to the tune of £5,000, a huge sum of money at the
time. One of her major creditors was Captain James Pinson Labulo (J. P. L.)
Davies, to whom reference is made later in this chapter. He foreclosed her
mortgaged properties on Lagos Island, but he too forfeited the land around
modern Tinubu Square to the government when he became bankrupt in
1875. Such were the vicissitudes of commercial life in nineteenth-century
Lagos and Abeokuta of which Tinubu, Davies, and others were victims. She
and others made and lost fortunes to disasters like ire outbreaks, commer-
cial slumps, and social debts. Tinubu’s biographer stated that she “made
money (from commercial enterprises) . . . but she spent the money on peo-
ple – never on herself alone” (Yemitan 1987, 28).
he details of Efunsetan’s career can be found elsewhere,12 but it is impor-
tant to highlight some aspects of it that illustrate the processes of accu-
mulation and consumption in nineteenth-century Yorubaland. A member
of important aristocratic families in Abeokuta, Ife, and Ibadan, Efunsetan
achieved some success as a trader in Abeokuta before moving to Ibadan.
At Abeokuta, she traded extensively with Ibadan, Lagos, Badagry, and
Porto Novo (modern Benin Republic). Efunsetan conducted trade with
Portuguese and Brazilians, to whom she sold slaves, indigenous textiles,

11
Ibid., for the details in this paragraph.
12
he details in the following paragraphs are derived from Awe (2001).
 Accumulation and Conspicuous Consumption 215

ivory, and palm produce. In exchange, she procured tobacco, beads, a wide
variety of textiles, and liquor, guns, and gunpowder. She also conducted a
considerable internal trade across Yorubaland and with neighboring states,
such as Ilorin and Nupeland, on the basis of the complementary needs and
products of the diferent regions. Among these were foodstufs, kolanuts,
woven mats and cloths, natron, and livestock. hese commodities were
conveyed by river transport (on the Ogun) and head porterage.
hough unlettered and without a modern bookkeeping system, and
operating in a milieu of traditional currencies and barter, Efunsetan built
an extensive commercial enterprise, which included extensive farms on the
outskirts of Ibadan. Her relocation to Ibadan demonstrated her keen busi-
ness acumen, for the burgeoning war camp conferred strategic advantages –
location on the forest-savanna fringe, extensive farmlands and production
of agricultural surplus for export, and a booming market for guns and gun-
powder and war captives. Efunsetan owned hundreds of slaves who served
her in various capacities – as farm labor, retainers, and artisans. At her death
in 1874, each of her three large farms contained 100 slaves, who planted
yams, corn, and vegetables. Her male slaves traded on her behalf to as far as
Bida (Nupe), manned her canoe leet on the Ogun, and were pressed into
battle as her contribution to the Ibadan ighting force. he female slaves
wove kijipa (country cloth in great demand in the Yoruba and neighboring
territories and in South America, where a large number of slaves had been
transported) and mats, and cultivated indigo used for dyeing cloth. hey
also produced various indigenous cosmetics from camwood and palm ker-
nel oil. Much of her commercial wealth derived from her virtual monopoly
of the tobacco import trade. She mixed politics with business by extending
generous credits in guns and gunpowder to the Ibadan war chiefs in return
for war captives. Her political stature rose with her stupendous wealth and
culminated in her appointment as Iyalode by either Basorun Ogunmola,
who died in 1867, or Aare Latosa, who became Ibadan’s supreme ruler in
1872. Not surprising, her political exposure and increasing assertiveness in
a male-dominated society led to her clash with Latosa and her assassina-
tion in 1874. In addition to other issues, the huge indebtedness of Latosa
and several Ibadan chiefs to Efunsetan and her refusal to grant them fur-
ther credit in arms and ammunition made possible their common cause
against her.
A signiicant point, to return to the idea of how wealth was utilized, was
the use to which the war chiefs of nineteenth-century Yorubaland put their
wealth in human and material resources (Awe 1973; Falola 1984). As auto-
cratic as he was, even Kurunmi was wise to the use of wealth for political
216 Olukoju

patronage – to sustain the support of his loyalists and to “placate the public”
(Falola 1991a, 28). Like his counterparts in Ibadan, he maintained “a gran-
diose style of living” in consonance with his position in society. His con-
spicuous consumption and opulent lifestyle found expression in his vast
compound that covered all of eleven acres, his harem of three hundred
wives, and his holdings of one thousand slaves (Falola 1991a, 28). With
her extensive farming, trading, and crats enterprises, Iyalode Efunsetan
Aniwura was extremely rich by the standards of the times. In addition to
three large farms beyond the city walls, she had two cattle farms in the city
managed by her Hausa slaves. As expected, she invested her wealth in an
opulent lifestyle (expensive wardrobe and ornaments), in the Ibadan war
efort, and in an elaborate political structure of her own to match her status
as the female equivalent of the Balogun of Ibadan. First, she had a large col-
lection of various types of expensive beads, silver trinkets and brass brace-
lets, and a rich wardrobe of the most expensive ceremonial cloth (alari,
sanyan, and etu), in addition to the durable, everyday wear kijipa. Second,
as the mogaji (head) of her lineage (Basorun Oluyole’s), she was responsible
for the upkeep of a large retinue of supporters and the funding of wedding,
child naming, and funeral ceremonies in the compound. Finally, as Iyalode,
her responsibility covered the entire community, which expected her to dis-
play opulence and generosity.
As illustrated by the careers of twentieth-century Ibadan entrepreneurs –
Salami Agbaje and Adebisi Giwa – the community expected the wealthy
to share their good fortune with the less privileged and to display their
wealth on public occasions and in other forms of conspicuous consump-
tion (Adeboye 2001). In the cases of the Ibadan war chiefs, Ijaye’s Kurunmi
and Iyalode Efunsetan, political power was accompanied by the dissipation
of wealth in the community. In the event of a fall from grace – Kurunmi
in Ijaye and Efunsetan in Ibadan – much of what was accumulated could
be lost in a single orgy of looting ordered by the political establishment to
obliterate a formidable adversary. In such cases, the wealth built up could
not be passed on.

6.3 CIVIL ENTREPRENEURSHIP IN THE COLONIAL

CONTEXT IN LATE NINETEENTH- AND EARLY
TWENTIETH-CENTURY YORUBALAND
A complementary group of indigenous entrepreneurs in Lagos, Abeokuta,
and Ibadan emerged from the mid-nineteenth century to the 1930s in
the context of British colonialism. Fairly representative of the group were
 Accumulation and Conspicuous Consumption 217

Salami Agbaje and Adebisi Giwa of Idikan, both of Ibadan; and D. C. Taiwo
(a.k.a. “Taiwo Olowo”), J. P. L. Davies, Peter homas, J. H. Doherty, S. H.
Pearse, W. A. Dawodu, and Charlotte Olajumoke Obasa, who were based in
Lagos. hese entrepreneurs were prominent in diferent sectors of the econ-
omy, including shipping (Davies), urban transport (Dawodu and Obasa),
farming (J. K. Coker and Davies), moneylending, and maritime commerce.
he rest of this chapter traces their diferential fortunes through the cycles
of boom and bust during the period between 1850 and 1939. We shall see
how they took advantage of openings in the colonial economy (in sectors
where European and Levantine presence or interest was marginal) and the
infrastructure of the colonial political economy – modern government and
law enforcement, standardized currencies, road, rail, and maritime trans-
port facilities, and banking services – to build their business enterprises.
We shall also see how they were constrained by, and attempted to tran-
scend, the structures of the colonial political economy.
D. C. Taiwo (“Taiwo Olowo”) was typical of the rise of indigenous entre-
preneurs in the context of the establishment of British rule.13 Originally
from Isheri, barely iteen miles north of Lagos Island, he had come to
Lagos in the early nineteenth century as a retainer in a chiely family. He
proceeded to amass considerable wealth from his investment in commerce,
land, and people. Taking advantage of changes in land tenure as part of the
new order imposed by British colonialism, Taiwo created a business enter-
prise based on trade, land acquisition, investment in real estate, and money-
lending. Between the 1840s and his death in 1901, he steadily acquired land
in strategic parts of Lagos Island and exploited the opportunities created
by the transition from the slave trade to the so-called legitimate trade. It is
signiicant that he utilized his contacts with inluential indigenous peoples
and the British to promote his commercial interests. He deployed credit
advanced by European traders to develop a considerable export trade in
palm produce and import business in textiles, especially during the 1890s,
a period of extreme commercial conditions. Taiwo also supplied arms to
Ibadan that were used in the Kiriji War.
Next, he took advantage of the issuance of Crown grants to buy and sell
land and to use land as collateral for credit. He regularized the ownership of
his existing land acquisitions and secured fresh ones under the new order.
His huge investment in land is attested by the fact that he bought twenty-one
properties in Lagos between 1870 and 1892. “Control of land and capital,”
Mann notes, “enabled the rising Lagosian to acquire wives, slaves, clients

13
he section on Taiwo is based on Mann (1991).
218 Olukoju

and other dependants” (Mann 1991, 101). He also manipulated the legal
system and his wealth to keep his subordinates in check, oten by litigation.
An adroit exploitation of the opportunities colonialism created enabled
him to mobilize human and material resources for the advancement of his
commercial and political interests. He did not acquire titles in Lagos, but
assumed the paramount chietaincy of his native Isheri. As his epitaph indi-
cated, he was the most inluential individual in Victorian Lagos on account
of his stupendous wealth, access to British oicials, and extensive networks
among the local political elite.
Yet, like the traditional Yoruba entrepreneurs in Ibadan and Ijaye, Taiwo
was careful to invest in people to maintain their loyalty and support. Such
“extension of inluence” in the community, Mann remarks, “consumed valu-
able resources” (Mann 1991, 101). In addition, Taiwo, like Salami Agbaje
and Adebisi Giwa in colonial Ibadan, was expected to dispense inancial
favors and write of debts to secure the goodwill of the community. In the
particular case of Taiwo, there were occasions when he waived his right
to foreclose the property of loan defaulters for purely cultural reasons, as
opposed to economic rationality. “He and other big men . . . may sometimes
have hesitated to foreclose for fear of alienating public opinion and losing
followers” (Mann 1991, 96). But he (and his ilk) did not sustain a dead loss
in such seemingly uneconomic transactions. First, indebtedness made the
debtors beholden to him for life. “Indebtedness in general and mortgages in
particular,” notes Mann, “gave Taiwo great power over people” (Mann 1991,
96). Second, valuable contacts and networks of inluence among Africans
and Europeans enabled him to “increase trade, collect debts, protect prop-
erty, and control subordinates” (Mann 1991, 102). By the 1880s, his inlu-
ence extended beyond Lagos and the Yoruba hinterland to the middle Niger
(Hopkins 1964, 41) and Kano in northern Nigeria, and beyond colonial
boundaries to Porto Novo in French Dahomey. According to Cole, Taiwo
was conferred with the sole privilege of presenting delegations from the
hinterland to the British colonial oicials (Cole 1975, 63, 170, 190).
A contemporary of Taiwo’s, who also achieved commercial success
and social inluence in Victorian Lagos, was J. P. L. Davies (1828–1906).14
Unlike Taiwo, Davies was a descendant of freed slaves who had returned
to his ancestral homeland in Yorubaland. His checkered career witnessed
his involvement in trade, agriculture, and shipping through the long nine-
teenth century until his death in 1906. He was credited with establishing the
irst commercially successful cocoa farm at Ijan on the outskirts of Lagos

14
Details on Davies in this paragraph are sourced from Hopkins (1978, 83–96).
 Accumulation and Conspicuous Consumption 219

in 1880. Davies had experimented with cofee and cocoa that he obtained
from a Brazilian ship and from the island of Fernando Po in 1879 and 1880,
respectively. He subsequently pioneered the shipment of cocoa, which was
to be the economic mainstay of western Nigeria from the 1920s. Davies
had earlier been a successful trader and ship owner,15 who was described
as “the most prominent of the African merchants in Lagos in the 1860s
and 1870s.”16 But his successes were punctuated by the economic turbu-
lence of the 1880s and 1890s. Indeed, he ventured into agriculture to recoup
his business losses with the aim of relaunching himself in the social life of
Lagos, where he had enjoyed great prominence.
In addition to his pioneering role in cocoa farming and exports and
attainment of mid-nineteenth century prominence in the commercial life
of Lagos, Davies was also signiicant for his connection with the career of
J. K. Coker, an important igure in agricultural entrepreneurship based at
Agege about eighteen miles north of Lagos. For it was from Davies that
Coker (whose brother married Davies’ daughter) obtained the cocoa seed-
lings with which he started of his cocoa plantations at Agege, where his
iteen-hundred-acre holdings dwarfed Davies’ two hundred acres at Ijan
(Hopkins 1978, 88). Coker was the driving force behind the formation
of the two-hundred-member Agege Planters Union (APU) in 1907. Like
Davies, Coker veered into agriculture to make up for losses in his commer-
cial enterprises in the late 1890s.
hough the APU was a collective, big planters like Coker who mentored
the smaller holders dominated it. he foray of members of the union into
agriculture had been induced by the economic downturn of the 1880s, when
many of them lost their fortunes in the commercial luctuations of the last
quarter of the century. hey were also keen to demonstrate the capacity of
Africans to run large-scale business enterprises – a blend of pecuniary and
nationalist motives. It can be argued that Coker and the APU succeeded in
their enterprise not merely in terms of accumulation, but in the spread of
the “gospel” of cocoa farming coupled with a brand of “African” Christianity
into large parts of eastern Yorubaland. his was to be the economic founda-
tion of the subsequent social and economic development of western Nigeria
from the 1920s onward.
Other notable African merchants in Lagos during the opening decades of
the twentieth century achieved social and commercial prominence beyond
the agricultural sector and in spite of the commercial vicissitudes of the

15
See Olukoju (1992, 259); Herbert Macaulay Papers, Box 30.
16
See Hopkins (1978, 91) for this quote and information in the rest of the paragraph.
220 Olukoju

period. Most of them did not survive the severe depression of 1920–1, but it
is important to highlight their careers for the insights they yield into the rise
and decline of African entrepreneurship. Representative of this group are W.
A. Dawodu, Nigeria’s pioneer motor importer and automobile mechanic; S.
H. Pearse; P. J. C. homas; and J. H. Doherty.
W. A. Dawodu has featured in the literature on aspects of the urban
and economic history of Lagos for his pioneering role in the automobile
import and repair business.17 Scion of a notable Lagos business family, the
Mabinuori Dawodus, his elder brother, B. C. Dawodu, was “one of the lead-
ing merchants in Lagos during the 1880s” (Hopkins 1964, 370). Ater his
primary and technical education in Lagos, he taught for a few years before
starting a bicycle sales and repair business in 1905. Foreseeing the prospects
for motor transport as the colonial government embarked on road con-
struction, he ventured into the motorcar import business and was helped by
the generous credit facilities his American and English suppliers provided.
his was the foundation of a business that encompassed the sale, hire, and
repair of lorries, motorcars, and bicycles. Indeed, he enjoyed the sole agency
of Ford cars and Humber cycles in the colony for a while.
he Dawodu automobile enterprise had a turnover of £26,000 on the eve
of World War I and enjoyed a boom in the atermath of the war. he range
of Dawodu’s operations extended beyond Lagos to the Yoruba hinterland
and as far as Kano in northern Nigeria. Messrs W. A. Dawodu & Co. oper-
ated a transport service in Ilesha in central Yorubaland, where an expan-
sion in the produce business in the context of the post–World War I boom
stretched it to the limit. A report of September 1920 stated that its transport
service had been “unable to cope with the transport business” in the local-
ity.18 hough the irm was acknowledged to have “done their best in the
past,” it was overwhelmed by the massive demand. Recourse was therefore
made to the Railway motor service to make up for the shortfall in supply.
At its height in 1921, the business had three retail outlets (showrooms) in
Lagos and a large yard in Osogbo in the heartland of the Yoruba country
and employed two hundred ity persons. It declined steadily from the mid-
1920s in the face of global economic adversity, competition from European
irms, and his domestic crisis. By 1920, it had been joined in the motor
import business by Miller Brothers (Ford cars – Touring and Runabout
brands), John Holt (Willy’s Knight Overland from the United States of

17
Details of Dawodu’s life and career in the following paragraphs are found in Hopkins
(1964, 371–2).
18
Lagos Weekly Record, September 11, 1920, Osogbo Notes & News.
 Accumulation and Conspicuous Consumption 221

America), L. Ambrosini, British American Produce Supply Company

(Victor cars), and hornton’s Lagos Company. Other competitors were
Nigerian Motors Limited and the African and Eastern Trade Corporation.
he latter sold Dodge, Studebaker, and Reo cars and obtained its motor
spares from Nigerian Motors Limited.19
Family sources were reticent about the details of the terminal crises of
Dawodu’s private and business lives.20 An indication of the domestic crisis
was the (probably forced) resignation of his irst son, Alabi, as manager
of Messrs W. A. Dawodu & Co. in August 1929. He was replaced by Mr.
Frank Ade-John.21 W. A. Dawodu himself died in January 1930, weeks
ater the death of his kinsman, James Patrick Dawodu, on December 31,
1929. he latter had been a “one time successful businessman in the Warri
District,” but had been “ousted by excessive competitions of irms with
big capitals [sic].”22 W. A. Dawodu’s death inally ended his business enter-
prise, but he had lost the import of Ford cars to Nigerian Motors Limited
well before then.23 Described as “the Pioneer of the Motor Industry in
Nigeria,” and “a self-made . . . and very shrewed [sic]” man, Dawodu was
reported to have “conducted a very remunerative business which became
at last too big for him.”24
he scale of Dawodu’s business enterprise can be deduced from the
magnitude of assets that were auctioned to ofset liabilities occasioned
partly, according to oral sources, by mismanagement by his employees,
including close kin. His business assets in Lagos were auctioned on June
28, 1930 ater a newspaper advertisement invited engineers, mechanics,
and general traders to make daily bids. he auctioneer, J. Aleshinloye
Williams, listed the following items in stock: Brockway lorries, Hand
tricks, crane, Willy’s Knight cars, Federal lorries, Outer Boat engine, Ice
chest, Indiana car engines, planks, iron bars, pieces of angle and lat irons,
Go-cart springs, cutting and punching machines, portable lathe machines,
machine driving lathes, emery machine, disc machine, band circular saw,
playing machines, sawing machines, drilling machines, wooden window
frames, corn mills, wooden chairs, iron plates, motor boat engine, rolling
machine, iron anvils, fanning machine, pumping machine, motor engine,

19
Lagos Weekly Record, November 20 and 27, 1920.
20
Interview: Alhaji R. O. Dawodu, Mokola, Ibadan, 1984.
21
Lagos Daily News, August 26, 1929.
22
Nigerian Pioneer, January 10, 1930, Random Notes & News.
23
Lagos Daily News, May 19, 1930.
24
Nigerian Pioneer, January 10, 1930, Random Notes & News, italics added. It is not clear
what the commentator meant by the remark in italics.
222 Olukoju

shats, pulleys, three-face alternating current motor with starter, one motor
boat under construction, circular saw with bench wood bending machine,
new motor boat engine, propeller shat, and “articles too numerous to
mention.”25 In July 1930, the stock in his extensive warehouse in Osogbo
was also auctioned.26
Another exceptional Lagos entrepreneur of the irst third of the twen-
tieth century was S. H. Pearse (1866–1953), who was described as “that
rare type of African businessman, the produce merchant” (Hopkins 1988,
433). His career epitomized the vagaries of the maritime trade of Lagos and
Nigeria as he experienced mixed fortunes in the export business. But it also
demonstrated the lexibility and adaptability of African entrepreneurs to
changing circumstances and challenging opportunities. Pearse reportedly
took to business in emulation of his father’s successful business friends –
R. B. Blaize, J. P. L. Davies, and Z. A. Williams (Hopkins 1988, 434). He
launched his own irm, Pearse and hompson, in 1888 in partnership with
a Sierra Leonean merchant ater serving as a clerk in the European irm
of W. B. McIver and Co. and as an apprentice under the African irm of
Williams Brothers. Pearse’s irm, however, folded in 1894 as a result of the
trade depression of the 1890s and the Anglo-Ijebu war of 1892.
With a liability of £1,000, he briely retired from business and spent some
time in England, ater which he became a Lagos agent of the African and
Gold Coast Trading Company. In 1897, he entered into a new partnership
in Calabar with a fellow Lagosian, E. O. Williams, who died only a year
later. Pearse achieved commercial success by combining an export business
in palm produce, ivory, and rubber with an import trade in various staple
items. he proitability of the ivory trade in Calabar enabled him to pay of
his old debt by 1902. His business expanded considerably to the point that
he appointed an agent for his operations in Calabar and acquired an impres-
sive head oice on Broad Street, aptly named Elephant House in appreci-
ation of the source of his wealth. Pearse conducted his European business
through the Liverpool irm of Taylor and Company and diversiied his busi-
ness into other enterprises. He bought shares in three mining companies on
the Gold Coast and opened the irst hotel in Lagos in 1907 (Hopkins 1988,
435). By 1914, Pearse was already a man of considerable social and political
inluence in Lagos, in the mold of D. C. Taiwo, on account of his business
success. Fitting, he became a member of the Nigerian Legislative Council
in 1915.

25
Lagos Daily News, June 27, 1930.
26
Lagos Daily News, July 9, 1930.
 Accumulation and Conspicuous Consumption 223

On the outbreak of World War I, Pearse had accumulated wealth in prop-

erty and liquid assets for his produce business. He also invested more than
£10,000 in war bonds (Hopkins 1988, 436). However, he soon encountered
diiculties owing to the decline of the ivory trade and the vicissitudes of
the produce trade, both induced by wartime exigencies. he severe postwar
slump of 1920–1 forced him to restructure his business by divesting in the
hotel business and disposing of his war bonds. He then ploughed his capital
into the rubber trade, where he also made a decent fortune until the out-
break of the Great Depression in 1929 practically decimated his business.
Undaunted, Pearse diverted his capital and energy into the import of stock-
ish and the export of hides and skins. However, his prosperous business
in these commodities was disrupted by the outbreak of World War II in
1939. A wartime accident incapacitated him and his business subsequently
petered out well before his death in 1953.
Meanwhile, constitutional changes introduced in 1922 by Governor
Hugh Cliford had been duly followed by the institution of the Legislative
Council of Nigeria. Pearse was adopted by the Alake and Council as the
representative of Egba Division in the Legislative Council. As a irst step, he
visited Abeokuta in April 1923 where the Alake gave a dinner in his honor
and later installed him as the successor to the late Oluwo of Itesi Ake, whose
stool had been occupied by Pearse’s proxy that he had personally appointed
and invested in Lagos in 1921. he Alake stated that Pearse had taken “great
interest in his township” and oten attended to the pressing needs of his
people. He reasoned that, as Pearse “could not come home to ill the oice
himself, it was thought that he could better serve them in another capacity
in Lagos by representing the Egbas on the Legislative Council at Lagos.”27
his move generated a huge controversy as it was alleged that Pearse had
procured a letter of introduction from the lieutenant governor to the resi-
dent of Abeokuta Province, which he used “to trick the Egbas into believ-
ing it (was) an instruction from the Government to get him appointed as
their representative on the Legislative Council of Nigeria.”28 But this was
refuted by the Nigerian Pioneer, which backed Pearse, by pointing out
that he was indeed a native of Abeokuta domiciled in Lagos. “His trading
successes which are evidently the result of foresight and care,” the news-
paper editorialized, “have brought him into prominence, and in religious,
social and political life of this country, he is not obscure.”29 Pearse’s other
27
Nigerian Pioneer, May 4, 1923.
28
N. Kuforiji, Abeokuta, letter to the editor, Nigerian Pioneer, May 25, 1923.
29
Nigerian Pioneer, June 8, 1923, editorial: “Abeokuta Native Administration: Mr. S. H.
Pearse, the Egba’s Representative at the New Legislative Council.”
224 Olukoju

engagements – as a representative at the local synod, member of the District

(Church) Council, member of the Lagos Chamber of Commerce, member
of the defunct Legislative Council, and member of unspeciied “important
Committees” – were also listed to buttress his suitability for the position.
At the end of the day, he went on to occupy the Legislative Council seat.
In 1929, Pearse became the Odoin of Itesi and in 1936 was granted the
Member of the British Empire (MBE) by the British government.30
S. H. Pearse was the colonial-era gentleman par excellence. He lived in
aluence and invested in the education and vocation of his children. Two
of his daughters, Ayo and Remi, attended the Ryford Hall Ladies’ College
in England and excelled at violin and piano, respectively. Both were “highly
commended,” spent some time on the Continent, and were reputed to have
been “the irst from West Africa, Africans, whether male or female, who
have lown in an Aeroplane.”31 he marriage of Ayo Pearse to Reverend S.
C. Phillips on January 26, 1922 was celebrated in a major newspaper, which
also published the full list of wedding presents and their donors.32 Pearse’s
aluence earned him respect in business circles and made him one of the
irst Africans readmitted into the Lagos Chamber of Commerce when it
was desegregated in 1921. His election was hailed by a Lagos newspaper
for the valuable experience that he would bring to that body. It described
Pearse as having “been in contact with the trade of Lagos, or for that matter
Nigeria before some of the members of the present chamber were born.” It
urged the Chamber to consider the inclusion of Messrs J. H. Doherty and
David Taylor at the “very earliest opportunity.”33 Pearse was unanimously
elected vice president of the Chamber in 1922.34
Pearse’s career was signiicant for several reasons. First, he remained in
the export business much longer than any of his contemporaries, whose ten-
ure oten coincided only with the cycles of boom in the global economy.35
Second, his career witnessed a succession of booms and slumps, including
bankruptcy or near bankruptcy on several occasions. his made him coun-
sel his son, who qualiied as a medical doctor, not to dabble in business
(Hopkins 1988, 437), a vocation that contemporary observers regarded

30
Nigerian Pioneer, April 19, 1929; Nigerian Daily Times, June 26, 1936.
31
Nigerian Pioneer, September 10, 1920, Random Notes & News.
32
Nigerian Pioneer, January 27, 1922.
33
Nigerian Pioneer, October 7, 1921, Random Notes & News.
34
Nigerian Pioneer, April 28, 1922, minutes of meeting of Lagos Chamber of Commerce of
April 11, 1922.
35
A Lagos newspaper remarked caustically that “[w]ith the decline in the prices of Produce,
we hope to see a great rush into clerical appointments.” Nigerian Pioneer, June 11, 1920,
Random Notes & News.
 Accumulation and Conspicuous Consumption 225

as nothing better than gambling.36 hird, he was as active in politics and

society as in business.
J. H. Doherty, another notable African entrepreneur, started on a small
scale.37 Ater serving as a clerk in one of the Lagos irms, he ventured into
business in 1891 with a capital of £47. He then gained the conidence of a
European irm that advanced him monthly credits ranging from an ini-
tial sum of £100 to £4,000. hrough careful management of the business,
Doherty was able to operate independently as a direct importer from 1898.
He dealt in cotton and silk goods, hardware, and other imports. As his busi-
ness expanded, he established branches in the hinterland of Lagos – Osogbo,
Lokoja, Zaria, and Kano. A notable feature of Doherty’s career was the
investment in the education of his children, two of whom were sent abroad
for university education: one to England for a degree in law and another to
Fourah Bay College in Freetown, Sierra Leone (an ailiate of University of
Durham in the United Kingdom) for a degree in the humanities to prepare
him for succession to his father in the family business. Ultimately, this was
an exceptional success story in business succession.
he inal biographical subject considered in this survey is Salami Agbaje,
an Ibadan-based entrepreneur of Iseyin (Oyo-Yoruba) parentage.38 Born
between 1865 and 1870, he migrated to Lagos with his Arabic teacher-
father and trained as a tailor. Looking for greener pastures, he relocated to
Ibadan, where he became a logger and, subsequently, a timber trader. He
made his fortune as a timber contractor to the irm of Messrs Shelford and
Sons, which constructed the railway from Lagos in the 1890s. Agbaje then
diversiied into the produce and import business on a commission basis
as a middleman between producers and expatriate irms. He accumulated
wealth by stockpiling cocoa during World War I and selling it at a substan-
tial proit during the post-boom period.
His success derived from his ability to take risks, his extensive networks
among and rapport with producers and European irms, and the opera-
tional strategies of advertising his business in the newspapers, especially the
Yoruba News, engaging his own lorries and trucks to convey his produce,

36
he Nigerian Pioneer, June 18, 1920, noted that: “Trade is like a gamble; it has its chances.”
his apparently popular view was shaped by the severity of the slump and the harsh
reversal of fortunes, rather than a carryover from the catastrophe of the slave trade era
or precolonial politically induced business reverses of the likes of Madam Tinubu and
Efunsetan Aniwura. It was reinforced by the constraining context of British colonialism
and the calamity of the Great Depression, which inally doomed colonial-era indigenous
commercial enterprises. See Olukoju (2002a, 176–98).
37
he entire paragraph is based on Macmillan (1993, 98–9).
38
he following paragraphs are based on Adeboye (2001).
226 Olukoju

and running a passenger transport service in Oyo Province and other parts
of Nigeria. His entrepreneurial acumen in operating the transport service
on “three important trade routes” was duly commended as “most progres-
sive” by the resident of his province in 1921 (Adeboye 2001, 6). Agbaje estab-
lished a mechanical garage to maintain his commercial leet, but this soon
developed into the largest commercial mechanical garage in the province,
second only to that of the government’s Public Works Department (P.W.D.).
Even the P.W.D. referred cases to his garage because of the eiciency of his
technicians, among them his own son, Seidu. he scale of Agbaje’s business
enterprises included small-scale manufacturing (mineral water and food
processing) and commercial printing.
Salami Agbaje’s career is signiicant for the following reasons. First, like
Pearse, he diversiied into several business lines, taking advantage of favor-
able circumstances (such as the postwar boom) and divesting where neces-
sary from some business lines, such as the cinema business that he sold to
the Levantines. Second, he adopted modern business methods – advertis-
ing his business in the newspapers, establishing a limited liability company
(Salami Agbaje and Company), and joining forces with other entrepreneurs
to form the Oke’Badan Trading Company in 1950. hird, Agbaje’s business
was transgenerational as his irst son, Dr. Saka Anthony Agbaje, took over
the commercial press and joined others to publish a newspaper, the Western
Echo, in the 1940s. Fourth, like most other entrepreneurs, he was involved
in the politics of his community and this involved dispensing patronage,
which he refused to do. Hence, he was perceived as tightisted, and this
robbed him of his entitlement in the chietaincy hierarchy in Ibadan.
Finally, on the important issue of how he utilized his wealth, Agbaje was
in a class by himself. he scale of his enormous wealth was patent to all.
he provincial report of 1922 stated that he owned “very large cement block
houses on European lines” for his business and personal uses, “a very hand-
some dwelling for himself,” and “a very large leet of motors and a repair
garage.”39 He had owned the irst car in Ibadan as early as 1915, was the irst
to own a multistory cement block house, the irst to connect electricity to his
house, and the irst private citizen to construct motorable roads in Ibadan
(linking the city to his farms). Like Doherty, he invested heavily in the edu-
cation of his children. His eldest son, the irst Ibadan man to do so, qualiied
as a medical doctor in 1931. His children assisted him directly in running the
business (as exempliied by Seidu) or ofered him legal services (two sons,
Mojeed and Ganiyu, qualiied as lawyers). To this day, his family remains

39
Cited in Adeboye (2001, 9).
 Accumulation and Conspicuous Consumption 227

inluential in Ibadan (and western Nigeria) on account of the number of

lawyers, judges, doctors, and other professionals among its scions.

6.4 CONCLUSION
his chapter has demonstrated that it was a combination of cultural and
structural impediments to indigenous entrepreneurship in the context
of colonialism and global economic dynamics that caused the nonemer-
gence of a capitalist class in the mold of contemporary Asian enterprises.
On the one hand, the colonial experience, while bringing modern trans-
port, currency, credit, and banking facilities and thus creating an enabling
environment for migrations and exchange, was equally constraining. he
demonetization of precolonial currencies and the abolition of slavery efec-
tively undercut the basis of the wealth and power of the category of entre-
preneurs whose wealth had derived from the slave trade and the use of slave
and indentured labor. Across West Africa, “businessmen and women lost
much of the value of their assets as a result of colonial intervention” (Austin
2002, 135). In addition, colonial rule facilitated the ingress of European
traders into the hinterland, thus reducing indigenous competitors to sub-
sidiary operators given the overwhelming superiority of the former in cap-
ital, organization, and links with the imperial power. he loss of control
over the hinterland stages of the emergent overseas trade proved fatal to the
aspirations of indigenous entrepreneurs in the long run.
On the other hand, the weight of cultural values proved too heavy for sus-
tained accumulation. For instance, societal expectations oten made it too
diicult for the entrepreneur to accumulate or act solely according to the
proit motive. Nineteenth-century Ibadan epitomized the Yoruba concep-
tion of true wealth, entrepreneurship, and the ideal entrepreneur. Taiwo’s
epitaph cited at the beginning of this chapter conirmed the resilience of
such views in colonial Lagos.
his was why Taiwo Olowo could aford to waive his right to foreclosure
of mortgaged property for fear of earning a reputation for meanness. he
failure to meet this societal expectation made the colonial Ibadan entre-
preneur, Salami Agbaje, unpopular with “a large section of the predomi-
nantly illiterate populace . . . [which] dislike[d] him” (Adeboye 2001, 11).
But Agbaje refused to bow to societal pressure and proceeded to expand
his business and invest in the education of his children. According to his
biographer, he “used his power derived from the modern economic sec-
tor without subjecting himself to the claims of tradition to which others
bowed with the largesse of the big man” (Adeboye 2001, 11). In contrast, the
228 Olukoju

Dawodus of Lagos chose to conform in 1930. Even in the throes of the Great
Depression, when two leading lights of the family (James Patrick Dawodu
and William Akinola Dawodu) died in rapid succession in December 1929
and January 1930, respectively, the family undertook the opulent “reburial”
of their long-dead forebear, Mabinuori Dawodu in June 1930!40 Economic
rationality succumbed to social pressure even in the face of global distress.
On the whole, the survey of the selected leading entrepreneurs of nine-
teenth- and twentieth-century Yorubaland indicates that the constraining
environment of warfare and colonialism provided a context for their ulti-
mate failure to emulate their Asian counterparts. First, with the exception
of Doherty and Agbaje, none had a clear succession plan within or outside
the lineage. In addition, those who died intestate and even those who let
a will hardly outlived their business as acrimonious disputes over sharing
of assets efectively destroyed their legacies. Hence, the dissipation of assets
and lack of cooperation among several branches of a typical polygynous
family ruined many worthwhile businesses. Second, Yoruba entrepreneurs
lacked a common focus and platform of action and were generally indi-
vidualistic. hough Salami Agbaje collaborated with his Ibadan kinsmen to
establish a joint stock company, this was the exception rather than the rule.
Hence, there was no enduring indigenous chamber of commerce to advance
their common interests. When these existed, they were short-lived and
fractious. Instead, the leading lights of the indigenous entrepreneurial class
strove to join the hitherto Europeans-only Lagos Chamber of Commerce.
hird, the basis of their wealth was fragile – consisting of limsy assets tied
to the luctuating economic circumstances. his made them vulnerable to
the vagaries of global economic dynamics as happened during the interwar
years. Fourth, the absence of a national bank for indigenous entrepreneurs
during the period up to 1930, unlike in Egypt, denied the entrepreneurs a
reliable source of capital for their business. he Egyptians established the
wholly Egyptian Bank Misr in 1920 to supply capital to national industry.
he charter of the bank limited shareholding and membership of board of
directors to native Egyptians. Bank Misr accordingly “played a major role in
mobilizing the savings of local people and using them to support a variety
of projects including the establishment of a number of public companies”
(Owen 1981, 5).
Finally, the concentration on commerce without a commensurate invest-
ment in manufacturing was the bane of African entrepreneurship. As
observed with respect to Egypt, “[e]conomic nationalism was part and parcel

40
Nigerian Pioneer, June 6, 1930.
 Accumulation and Conspicuous Consumption 229

of national ideology in Egypt from early on, and local industrialization was
meant to be its most important project.”41 his policy focus of Egyptian, as
opposed to Nigerian, nationalism was a critical issue in the divergence of
their experiences. For the Egyptians used economic nationalism to drive
the founding of their national bank, Bank Misr, and the promotion of indig-
enous manufacturing. Local manufacturers used advertisements to rally
their fellow nationals to “Buy Egyptian” to show how patronage of local
manufactures contributed to promoting Egyptian employment. But it must
be admitted that the Egyptians, unlike the Nigerians, had room to maneu-
ver as their country had gained “partial independence” by 1922, whereas
the interwar years were the heyday of colonial rule in Nigeria, where indig-
enous capacity for freedom of action was highly circumscribed. In addition,
Egypt was more culturally and politically coherent than was culturally and
politically diverse Nigeria.
hat said, several of the leading entrepreneurs invested in the education
of their children, who sustained their legacy outside the mercantile sector.
To that extent, the failure of indigenous entrepreneurship to sustain itself
and to generate societal development was not totally unmitigated.

References
Adeboye, O. A. (2001). “Salami Agbaje: he Life and Times of an Indigenous Entrepreneur
in Colonial Nigeria.” Lagos Historical Review, 1: 1–15.
Akeredolu-Ale, E. O. (1975). he Underdevelopment of Indigenous Entrepreneurship in
Nigeria. Ibadan: Ibadan University Press.
Akinjogbin, Adeagbo. Ed. (1998). War and Peace in Yorubaland, 1793–1893. Ibadan:
Heinemann.
Atanda, J. A. (1973). he New Oyo Empire. London: Longman.
Austin, Gareth. (2002). “African Business in Nineteenth-Century West Africa.” In Black
Business and Economic Power, edited by Alusine Jalloh and Toyin Falola, 114–44.
Rochester: University of Rochester Press.
Awe, Bolanle. (1973). “Militarism and Economic Development in Nineteenth Century
Yoruba Country: he Ibadan Example.” Journal of African History, XIV(1): 65–77.
(2001). “Iyalode Efunsetan Aniwura (Owner of Gold).” In Nigerian Women: A
Historical Perspective, edited by Bolanle Awe, 63–82. Ibadan: Sankore/Bookcrat.
Awe, Bolanle and Omotayo Olutoye. (1998). “Women and Warfare in 19th Century
Yorubaland: An Introduction.” In War and Peace in Yorubaland, 1793–1893, edited
by Adeagbo Akinjogbin, 121–30. Ibadan: Heinemann.
Babayemi, S. O. (1990). he Fall and Rise of Oyo, c. 1706–1905. Lagos: Lichield Nigeria
Limited.

41
See Shechter (2008, 573) for this quote and the following discussion.
230 Olukoju

Cole, P. D. (1975). Modern and Traditional Elites in the Politics of Lagos. Cambridge:
Cambridge University Press.
Falola, Toyin. (1984). he Political Economy of a Pre-Colonial African State: Ibadan,
1830–1900. Ile-Ife: University of Ife Press.
(1991a). “he Construction and Destruction of the Economy of Ijaye, c.1832–1862.”
Afrika und Ubersee, 74: 21–37.
(1991b). “he Yoruba Caravan System of the Nineteenth Century.” International
Journal of African Historical Studies, 24(1): 111–32.
Hopkins, Antony G. (1964). “An Economic History of Lagos, 1880–1914.” PhD Diss.
University of London.
(1978). “Innovation in a Colonial Context: African Origins of the Nigerian Cocoa-
Farming Industry, 1880–1920.” In he Imperial Impact: Studies in the Economic
History of Africa and India, edited by Clive Dewey and A. G. Hopkins, 83–96.
London: he Athlone Press.
(1988). An Economic History of West Africa, 2nd ed. London: Longman.
Macmillan, Allister. (1993). he Red Book of West Africa. Ibadan: Spectrum, reprint.
Mann, Kristin. (1991). “he Rise of Taiwo Olowo: Law, Accumulation, and Mobility in
Early Colonial Lagos.” In Law in Colonial Africa, edited by Richard Roberts and
Kristin Mann, 85–102. Portsmouth, NH: Heinemann.
Olukoju, Ayodeji. (1992). “Elder Dempster and the Shipping Trade of Nigeria during the
First World War.” Journal of African History, 33(2): 255–71.
(2002a). “he Impact of British Colonialism on the Development of African Business
in Colonial Nigeria.” In Black Business and Economic Power, edited by Alusine
Jalloh and Toyin Falola, 176–98. Rochester, NY: Rochester University Press.
(2002b). “‘Nigeria or Lever-ia?’: Nationalist Reactions to Economic Depression and
the ‘Menace of Mergers’ in Colonial Nigeria.” Journal of hird World Studies, 19(1):
173–94.
(2010). “Economic Nationalism and Decolonization: West Africa in Comparative
Perspective.” Hagar, 9(2): 137–56.
Owen, Roger. (1981). “he Ideology of Economic Nationalism in Its Egyptian Context:
1919–1939.” In Intellectual Life in the Arab East, 1890–1939, edited by Marwan
R. Buheiry, 1–9. Beirut: Center for Arab and Middle East Studies, American
University of Beirut.
Shechter, Relli. (2008). “he Cultural Economy of Development in Egypt: Economic
Nationalism, Hidden Economy and the Emergence of Mass Consumer Society
during Sadat’s Initah.” Middle Eastern Studies, 44(4): 571–83.
Yemitan, Oladipo. (1987). Madame Tinubu: Merchant and King-Maker. Ibadan:
University Press Limited.
Zakaria, Yakubu. (1997). he Cultural Context of Business: A Study of Firms in a Northern
Nigerian Society. Uppsala: Uppsala University Library.
 7

Commerce, Credit, and Mobility in Late

Nineteenth-Century Gold Coast: Changing
Dynamics in Euro-African Trade

Emmanuel Akyeampong

7.1 INTRODUCTION
he nineteenth century represents one of the most fascinating and dynamic
centuries in the history of West Africa in the second millennium in terms
of West Africa’s relations with the wider world. he century began with the
abolition of the export slave trade and the substitution (a return to in some
cases) of an export trade in primary products such as palm oil and ground-
nut oil. he simple technology behind African agriculture – the hoe, cutlass,
and pick – and the centrality of the family as the social and economic basis of
production and reproduction expanded the ranks of the economically active
with the abolition of the slave trade. Chiefs and “big men,” who dominated
the slave trade as they had monopoly over the means of capture, remained
major players in the era of “legitimate trade,” deploying their capital and
labor resources in new ways. It was a century of political revolution with
Islamic jihads in the West African sahel and savanna and political exper-
imentation along the coast based on the fusion of African and European
ideas. In the social sphere, Christian missionary work and Western edu-
cation from the 1820s created “new men and women,” while kin and clan
systems responded to the new opportunities of trade during the eighteenth
and nineteenth centuries through social engineering in structures such as
the “house system” in the Niger Delta. Technological advances added to
the excitement of the century, and in mid-century, the introduction of the
steamship and the demonstrated utility of quinine opened up West Africa

An early version of this chapter was presented at the July 2010 international conference in
Accra on “Understanding African Poverty over the Longue Durée.” I am grateful to Robert
Bates for his insightful comments on earlier drats.

231
232 Akyeampong

to European traders and adventurers. Banks and commercial credit linked

West African merchants with European merchant houses, and European
commission houses served as brokers for West African traders, selling West
African produce and purchasing European manufactured goods on their
behalf.1 Palm oil, palm kernel, and rubber from West Africa underpinned
this buoyant trade during the nineteenth century. he enormous opportu-
nities for social mobility these economic factors presented received a severe
setback from the 1870s with a glut in the world market of fats and oils, the
onset of economic crisis, and the constriction in credit facilities. he scram-
ble for and partition of Africa would end a remarkable era of socioeco-
nomic opportunity, and British imperial historian David Hargreaves noted
with irony how European military force had been summoned to contain
the socioeconomic revolution in West Africa that free trade had unleashed
(Hargreaves 1986).
his chapter focuses on the last quarter of the nineteenth century (or the
period from 1870), when the ground shited under the feet of West African
entrepreneurs and merchant houses collapsed. his followed two decades
of economic prosperity in the 1850s and 1860s driven by European tech-
nological developments and inancial investments that had a trickle down
efect globally. Economists have argued that the period from 1873 to 1896
was technically not an economic “depression,” as output increased consid-
erably even as prices dropped. But 1873 marked a sharp economic down-
turn, initiating two decades of economic distress in industrial economies
in Europe and North America. Adding to the confusion of the period, to
“contemporaries, the nature of the problem was unclear” (Gourevitch 1986,
73). hese closing decades of the nineteenth century witnessed bankrupt-
cies on both sides of the Atlantic with dire implications for extensive com-
mercial networks, introduced suspicion into relations between established
Gold Coast and European merchants, and extinguished the hope of small

1
Commission houses, sometimes referred to as consignment houses, were essentially brokers
based in Britain or Europe, who for a commission of between 2 and 5 percent would receive
produce from African traders on board steamships, sell the produce at prevailing market
prices, and purchase European manufactured goods requested by the African traders and
consign them to steamships bound for Africa. his facility had been made possible with
the rise of steamship service to the western African coast from 1852, calling regularly at
designated stops. It ended the monopoly of large expatriate traders who owned their own
sailing ships, disconnecting trade and transport ownership. Small African traders were
able to trade directly with Europe, bypassing expatriate merchants on the West African
coast. Commission houses also extended credit to trusted clients. hey kept accounts with
their African clients. In the latter part of the nineteenth century, the more established
commission houses had their own agents based in West Africa.
 Commerce, Credit, and Mobility 233

traders who aspired to the economic success of the “merchant princes,” who
also served as mentors for these small traders. hrough the private papers of
the established Gold Coast family businesses such as the Ocanseys and the
Caesars of Ada, especially their correspondence with their European sup-
pliers and brokers and the smaller African traders who depended on them
for credit and supplies, I shed light on the local ramiications (economic,
social, cultural) of shits in the global economy, deepening an understand-
ing of markets that transcend the abstract forces of demand and supply. I
use the checkered career of Christian Jacobson, an aspiring young trader
of Keta, to examine the closing of economic opportunities in the late nine-
teenth century and its impact on mobility and social aspiration. I draw on
the insights of Pomeranz and Topik that markets are “socially constructed
and socially embedded,” and that “it is necessary to understand both the
local speciicity of a transaction or event as well as its international context
to appreciate its importance” (Pomeranz and Topik 2006, xii–xiii). Access
to information and credit was severely skewed against Gold Coasters and
other West Africans during the nineteenth century, incorporated into com-
mercial networks not of their creation, lacking the dispersed but connected
international networks and solidarities of trading diasporas, and in the last
resort lacking the military support of colonially minded European states
that followed the interests of their merchants. hough West African chiefs
were deeply invested in trade, their networks were local. heir combina-
tion of political and economic power would identify them as the major
protagonists for European merchants. Colonial rule initially undermined
chiely power and enterprise, but a consolidated colonialism would in turn
empower colonially endorsed chiefs, even seeking to establish them in busi-
ness. At the end of the long nineteenth century, as the forces unleashed dur-
ing the century played out into the immediate World War I era, it was the
new hierarchy of colonial chiefs that remained standing in business with
colonial backing.

7.2 NINETEENTH-CENTURY CONTEXT: THE

INFRASTRUCTURE OF “LEGITIMATE” TRADE
7.2.1 Steamships and Quinine
In 1792, Denmark abolished the export slave trade for its citizens with
implementation from 1803, becoming the irst European power to abolish
the Atlantic slave trade. Britain followed suit in 1807, and the motives for
the British abolition have attracted scholarly attention over the twentieth
234 Akyeampong

century, beginning with Eric Williams’s inluential Capitalism and Slavery

(1944).2 A trade in cash crops such as palm oil and groundnuts, and older
commodities like ivory and gold, was revived under the rubric of “legitimate
trade” in contradistinction to the proscribed export trade in slaves. Palm
oil was used in an industrializing Europe for greasing factory machines,
lubricating new railway engines, and manufacturing soap and candles.
Groundnut oil was used as cooking oil and in soap making. Palm oil trees
grew wild in the West African forest from Sierra Leone to the Cameroons,
with a dense concentration in the Niger Delta and its hinterland, earning
the Delta the epithet of “oil rivers.” British imports of palm oil from West
Africa grew from 2,233 Cwt in 1807 to 213,467 Cwt in 1830, 315,458 Cwt
in 1840, and 731,659 Cwt in 1854 (Lynn 1998, 13). he Senegambia was the
groundnut basin in the second half of the nineteenth century, but Senegal
exported virtually no groundnuts in the 1840s. hese important develop-
ments intensiied European interest in penetrating the interior of West
Africa, an endeavor that united European statesmen, merchants, military
men, explorers, and missionaries. he advent of the iron gun boat in the
mid-nineteenth century was timely, and the British naval squadron that
patrolled the West African waters to monitor the slave trade was instructed
to make periodic forays up the Gambia and Niger Rivers to underscore
British naval might and its commitment to protecting British merchants in
these places.
It was the stranglehold of African chiefs and traders along the Niger Delta
and their resistance to any attempts by foreigners to travel into the inte-
rior that the British hoped their Niger Expedition of the 1850s and 1860s
would break. Touted as a scientiic and exploratory mission, its economic
implications were clear and profound. British trade ambitions coupled with
scientiic curiosity played itself out in West Africa irst with the expedi-
tion of Mungo Park (1795–7) under the auspices of the Africa Association.
Mungo Park’s travels and the publication of his Travels in the Interior
Districts of Africa (1799) became the basis for a domestic argument for an
interventionist policy of setting up trade stations in West Africa backed by

2
Williams probed the relationship between the West Indian plantation economy and the
rise of British capitalism, and the subsequent elevation of wage labor over slave labor by
the supposed beneiciaries of the plantation economy. Williams’s thesis and the question
of whether Britain was an economic beneiciary of abolition has elicited major works such
as Anstey (1975), Drescher (1977), and Eltis (1987). More recent works on British abo-
litionism have highlighted how central the loss of the North American colonies was to
British rethinking of empire, and how the British came to value the moral capital of aboli-
tionism in the remaking of empire. See Brown (2006) and Jasanof (2011).
 Commerce, Credit, and Mobility 235

military power (Marsters 2000). When Mungo Park revisited West Africa
in 1805, it was under the sponsorship of the British government. In 1830,
two other explorers, brothers John and Richard Lander, traversed the Niger
Bend and discovered that the river continued to low southward for another
twenty-six hundred miles into the delta.
On their return to England, the Lander brothers were contacted by
Macgregor Laird of the Birkenhead shipbuilding family. Together they cre-
ated the African Inland Commercial Company with the objective of explor-
ing and developing the commerce of the River Niger. heir ill-fated mission
up the Niger in 1832–3 ended with heavy losses in life, but Laird retained
his interest in Africa (Davies 1977). he 1832 expedition involved the irst
use of an iron steamer, Alburkah, on a West African river (Headrick 1981,
61). Subsequent expeditions up the Niger would be a partnership of the
British government, mercantile, and missionary interests. he Rev. J. F.
Schön and Samuel Crowther on behalf of the Church Missionary Society
accompanied the 1841 expedition led by Dr. T. R. H. homson. he death
rate among Europeans on the Niger Expedition of 1841 was equally hor-
rendous. homson himself took regular doses of quinine on this trip and
remained healthy, and his article in the Lancet in 1846 advocated the regu-
lar use of quinine in West Africa. he 1854 Niger expedition followed the
medical prescription of quinine meticulously, and Dr. William Baikie, cap-
tain of he Pleiad, ensured that his crew took the appropriate dose daily.
His ship stayed for one hundred twelve days on the Niger and returned
with all its European crew intact (Headrick 1981, 68–9). he Niger expedi-
tion was contracted to merchant Macgregor Laird from January 1857, who
received a British government subsidy to send a commercial steamship up
the Niger annually with the backing of a British naval ship for a ive-year
period. A similar arrangement was already in place for the Gambia River.3
Baikie also accompanied the 1857 expedition and became the embodiment
of the Niger expedition through his death in 1864.
An Admiralty letter in 1862 summed up precisely what was at stake for the
British in the Niger expedition. It pitted British mercantile interests against
those of African trading chiefs and coastal merchants, insinuating – perhaps

3
National Archives of Britain (NAB): FO 97/434. See Memorial to Viscount Palmerston,
presented March 15, 1860. he memorialists were pleased by the government’s arrange-
ment with Laird and the presence of British naval ships on the Niger, and commented on
how the similar presence of naval ships on the Gambia had been conducive for British
commerce. Signatories to this memorial included the Earl of Shatesbury, the Earl of
Harrowby, Lord Viscount Middleton, the Lord Bishop of Oxford, Captain the Honorable
F. Maud, and others.
236 Akyeampong

rightly – that peasants in the interior would welcome direct British com-
merce and the removal of the centuries-old trade control by the chiefs.
It has been shown by the reports received from Dr. Baikie and from the command-
ers of the several Expeditions that have at various times ascended the Niger that the
Chiefs and People in the Upper part of the River are peaceably disposed and very
anxious to trade with Europeans. he produce of the country is valuable and can be
increased to any extent that there may be a demand for it, in fact the supply is only
limited by the quantity of suitable European goods that can be brought into their
markets. he river moreover is navigable for vessels of 300 tons further for a dis-
tance of upwards of 500 miles thus opening a highway into the very heart of Africa.
It has been shewn also that from the regions visited by Dr. Baikie in the neighbour-
hood of the river, large caravans depart and arrive annually, carrying European
goods that ind their way up from the coast into the centre of Africa. In fact the
trade that used to pass from the northern shores of the Mediterranean to central
Africa is now being diverted to the Western Coast, sensibly diminishing the supply
of slaves that used formerly to be taken to the Mediterranean Coast. If therefore
it is found more proitable for the caravans to procure their supplies of European
goods from the Western Coast, when these supplies have to be iltered through the
hands of several chiefs and are enhanced in value by a long and dangerous land
carriage how much more proitable would it be if the supplier they required were to
be found in markets established in the upper banks of the Niger. But it is impossi-
ble that these markets should be established and the trade of the river developed as
long as the navigation in the Delta is practically stopped by the hostility of some of
the native chiefs. Render the navigation safe and our merchants will soon push their
way up the river and establish trading posts.4

he display of hostility of towns along the lower Niger to British merchants

was interpreted as the machinations of chiely economic interest, and the
British government and merchants recommended the use of naval force
to protect British trade. Free trade was seen as a right for British traders
and African peasants who supposedly yearned for that same right. British
pioneers of the African frontier like Baikie painted a picture of numerous
disparate “tribes” in the interior distrustful of each other, oten speaking
unintelligible languages and who would rather deal with Europeans than
their neighboring Africans.5
Trade, then, was the key vision that drove British interest in accessing the
West African interior from the early nineteenth century. his vision had
been facilitated by technological innovations, and the advent of the steam-
ship bore revolutionary consequences for the West African trade. In 1852,
regular steamship service commenced between Britain and West Africa,

4
Ibid.
5
NAB: FO 2/20.
 Commerce, Credit, and Mobility 237

and under royal charter the African Steam Ship Co. was incorporated in
August 1852 to transport mail to West Africa. his company was headed by
Macgregor Laird. he African Steam Ship Co. would move its headquarters
from London to Liverpool in 1856. A second shipping line would join the
African Steam Ship Co. in 1869, the British and African Steam Navigation
Co., which operated initially from Glasgow and Liverpool. From 1874, it
operated only from Liverpool, which had emerged as the center of British
shipping and palm oil trade with West Africa. By 1880, the age of sail was
efectively over (Lynn 1989, 1998, 105–10).
hese advances in technology, the expanding European economic
vision in Africa, and other accompanying developments also created new
social and economic opportunities for African producers and traders. So
the period from the 1860s, for example, witnessed the increasing ranks of
African small traders, who capitalized on the advent of regular steamship
service, the proliferation of European consignment houses who represented
the interests of African merchants in Europe for a small commission, and a
growing familiarity with local consumer demand that oten came through
employment as clerks in European commercial irms. In 1866, barely a
decade ater the introduction of steamship service to West Africa, Charles
Livingston, British consul, wrote from Fernando Po to Lord Stanley of the
Foreign Oice on how the era of the big commercial houses that had dom-
inated the slave and ivory trade and the early palm oil trade was coming to
an end.
he monthly line of mail steamers has broken up the monopoly of the large traders
on the West Coast of Africa. Business was formerly conducted on the principle of
large proits and slow returns, but since the steamers furnished such facilities, great
numbers of small black traders have sprung up whose motto evidently is small prof-
its and speedy sales, many being satisied with only 5% proit. he captain of one of
the mail steamers informs me that a few years since a steamer used to bring 300 to
400 packages of goods to Sierra Leone all for white traders, but now each steamer
brings 1500 to 1600 packages three fourths of which are for black traders and the
same is true, he says, of Lagos and other parts of the coast.6

In the Gold Coast, major African merchants such as the Ocanseys of Ada
at the mouth of the Volta River and John Sarbah with several trading posts
along the coast emerged from the mid-nineteenth century. Both would enjoy
remarkable success through the 1870s, and their trading networks would
include several coastal and up-country traders who depended on these

6
NAB: FO 2/47
238 Akyeampong

African merchant princes for credit and supplies of European manufac-

tured goods in return for palm oil and later palm kernel and rubber.

7.2.2 Credit and Trust in Euro-African Trading Networks

Lovejoy and Richardson in several works have underlined the centrality of
pawnship or debt bondage in credit mechanisms that underpinned the slave
trade in many parts of West Africa, and the city-states of the Niger Delta
and the Gold Coast were two such areas (Lovejoy and Richardson 2001).
Slaving privileged the political aristocracy in West Africa who possessed
the means of capture, and chiefs were oten big traders. In the Gold Coast,
noted for its wealth in gold, which constituted another form of security
in addition to human pawns, it has been observed for the eighteenth cen-
tury that most credit advances Europeans gave were to people with political
status. In Old Calabar, where gold was not mined, pawnship was the basis
of European mercantile credit to African traders (Lovejoy and Richardson
1999, 333–55). In Bonny, where a strong monarchy existed, the state was
the guarantor of credit (Lovejoy and Richardson 2004). he abolition of
the slave trade and the rise of legitimate trade raised two key challenges
where credit in Euro-African trading relations was concerned. First was the
demise of pawnship as a means of security. Pawns represented ideal collat-
eral for European merchants on the West African coast during the era of the
slave trade, especially if the pawns were close relatives of West African mer-
chants. hese merchants then did everything within their means to redeem
their kinsmen. Where pawns could not be redeemed, they became slaves
and were carried away in the holds of slave ships. Indeed, ship captains
noted that pawns, being insiders of coastal communities, were healthier and
more inclined to survive the rigors of the middle passage than captives who
had been marched from the interior to the coast. A second challenge related
to the expansion of the ranks of African traders with several newcomers
and the need for criteria to determine who could be trusted with credit.
Abolition meant that pawns were no longer acceptable security for British
merchants. Emancipation in the last quarter of the nineteenth century and
the abolition of pawnship forced a rethinking of the nature of labor. Old
social networks provided some basis for trade and credit and provided
a template for new social relations in business. Personal ties became an
important basis of credit during the nineteenth century, and European
merchants then used established African businessmen as references for
new applicants for credit. Here there were continuities from the past. A
perusal of the diary of Antera Duke from the late eighteenth century reveals
 Commerce, Credit, and Mobility 239

the level of socializing between Eik merchants and chiefs and ship captains
as both wined and dined each other (Duke 1956). It helped that a small
group of merchants or mercantile families dominated trade in the Niger
Delta city-states such as Calabar on both the British and African sides,
embedding credit in social networks (Lovejoy and Richardson 1999). It is
out of this context that British merchants such as Ambrose Lace brought
the son of Robin John Ephraim (alias King George) of Old Town, Calabar,
to Liverpool for schooling in 1767. he son spent two years in a school at
Liverpool at Lace’s expense before returning to the Niger Delta. Both sides
expected such products to facilitate Euro-African trade to their mutual
beneit. At the beginning of the nineteenth century, it was observed that
several Africans from Calabar came to Liverpool expressly to learn English
(Lovejoy and Richardson 1999, 342).
hese strategies facilitated the transition to legitimate trade and the
maintenance of credit networks. here were also novel developments dur-
ing the nineteenth-century era of legitimate trade, and in Calabar in the
1820s it was noted that British ship masters had begun to join the Ekpe
secret society that dominated political and economic social life. he society
served as a guarantor of credit. Lovejoy and Richard see this innovation as
an outcome of the palm oil trade and the changing political and commer-
cial conditions ater British abolition (Lovejoy and Richardson 1999, 349).
hey note how Ekpe served as a “merchants’ guild.” It is clear that over the
eighteenth and nineteenth centuries merchants on both sides accumulated
a body of knowledge of their opposite societies and cultures in the interest
of the functioning of commerce. While this promoted commerce, it also
had the potential to undercut the other side in trading. We have noted how
increasing knowledge of the Niger Delta interior convinced British trad-
ers that interior peasants would rather deal directly with Europeans than
through coastal middlemen chiefs. Likewise, African coastal merchants
came to the conclusion that they could deal directly with Europe rather than
through the European merchants on the West African coast, who proited
through facilitating both African exports and imports. Consul Livingstone
of Lagos had anticipated this development in 1866 in the same letter to
Lord Stanley with his observations of the radical changes steamship service
brought in West African trade.
Another and greater change in the palm oil trade seems inevitable. he shrewd
African chiefs, like those of Bonny will soon begin to ship oil themselves by steamer
direct to Liverpool, there Commission Agents, for a small percentage, are ready to
sell their oil and return by next steamer any kind of goods the natives may order.
hey ofer likewise to insure the oil as soon as it is aloat, the natives in this way get
240 Akyeampong

50% more for their oil and yet [are] able to sell it in Liverpool 50% cheaper than it
is now sold.7

he extension of inancial and brokerage services to West African traders

by British commission houses, together with steamship service, reshaped
trade in West Africa by leveling the ground in signiicant ways. It undercut
the monopolistic advantages of the large expatriate companies that owned
or chartered the sailing ships that brought goods to West Africa and took
oil away from the coast.
Lynn notes how between 1830 and 1880 the number of British trad-
ers who dealt in West African palm oil at the ports of Liverpool, London,
Bristol, and Glasgow increased from twenty-one to one hundred ive.
Numerous commission houses entered the palm oil trade in this period, of which
Irvine and Woodward (later James Irvine & Co.) of Liverpool, John Holt’s suppliers
between 1867 and 1875, and John Walkden & Co. of Manchester, prominent in the
1870s and 1880s, were among the most notable. Even long-established irms like
Tobins felt the pull of this trend: it became a commission house, the Company of
African Merchants, in 1863. (Lynn 1998, 138)

Lynn comments on how the “increase in numbers of commission houses

was simply a symptom of the emergence of smaller-scale palm oil traders
that occurred during the third quarter of the century” (Lynn 1998, 138).
Now West African traders could choose to use the long-established expa-
triate merchants on the coast or bypass them to send their produce direct
to Europe and source their manufactured goods through commission
houses. In the Gold Coast, the Ocanseys sometimes dealt with the irm of
F & A Swanzy, an established mercantile company in the Gold Coast, and
other times dealt directly with Liverpool using the commission house of
Hickson and Sykes. John Sarbah used J. F. Hutton & Co. of Manchester and
Grimsdale and Ridley as his commission houses (Lynn 1998, 141). It was
to facilitate these developments that Alfred Jones – a former staf member
of the African Steam Ship Co., who rose to become in 1884 the managing
partner in the Elder Dempster shipping line in Liverpool – expanded into
banking. In 1894, he set up the Bank of British Bank West, which domi-
nated banking in West Africa until 1917, when the Colonial Bank (now
part of Barclays) began operating in West Africa (Davies 1977, 13). Jones
was deeply appreciative of how steamship service, banks, and cable could
revolutionize West African commerce.8 West Africa was linked to Europe
by submarine cables in 1885 (Headrick 1981, 161).

7
Ibid.
8
Personal communication, Peter Davies, Liverpool, May 12, 2010.
 Commerce, Credit, and Mobility 241

7.3 BOOM AND BUST: THE COLLAPSE OF THE MARKET

IN OILS AND FATS
hese auspicious developments in the mid-nineteenth century concealed
major contradictions in the commercial environment. Paradoxically, the
seeds of West Africa’s commercial decline in the late nineteenth century
were embedded in the very processes that brought prosperity in mid-cen-
tury. Just as the advent of iron steamships and steamer services had incor-
porated West Africa more deeply into European commercial networks as
suppliers of raw materials for industrial purposes in Britain and elsewhere,
so did these same processes incorporate other tropical sites in Asia into
European networks as suppliers of oils and fats. West Africa had an initial
advantage as an early supplier of oils, but not for long. Within West Africa
itself, the huge expansion in supply depressed prices. While just about a
dozen ports exported palm oil from West Africa to Britain in the 1830s and
1840s, in 1885 some one hundred thirty-three ports in West Africa were
exporting oil to Britain (Lynn 1998, 240). Also, continued technological
innovation generated the need for specialized oils for particular produc-
tion processes, such as the manufacture of tinplate. For these specialized
processes, an all-purpose oil like palm oil was unsuitable. he outcome was
twofold: steamship service increased the number of palm oil suppliers from
West Africa and Asia with a dampening efect on palm oil prices; com-
peting oils emerged, such as Australian tallow (now made more accessible
ater the opening of the Suez Canal in 1869) and petroleum, discovered in
the United States in 1859 with its by-product kerosene. Gradually, mineral
oil displaced palm oil as the lubricant for machinery and railways from the
1860s (Lynn 1998, chapter 5). he 1850s peak of palm oil prices that galva-
nized trade in West Africa and Europe and underpinned the proliferation
of commission houses thus marked the beginning of a downward trend in
palm oil prices. he general prosperity of the 1850s and 1860s concealed
this shit in the ground that became more apparent from 1873.
From a high point of £43.50 per ton of palm oil in the United Kingdom in
1855, palm oil prices dropped to £33.67 in 1873 and to £20.58 in 1897 (Lynn
1998, 112). German prices for palm oil continued to rise through the 1880s,
but the German market was insigniicant compared to the British one. he
growth of an export market in palm kernels, a by-product of the oil palm,
brought some relief from the 1860s. Palm kernel was used initially in soap
manufacture, and increasingly in the making of margarine. Germany was
the major importer of West Africa’s palm kernels, explaining the broaden-
ing of the commercial networks of Gold Coast traders to include merchant
houses in Hamburg. he Gold Coast’s palm kernel exports from producers
242 Akyeampong

in Krobo and Akuapem went primarily to Germany (Lynn 1998, 125). Lynn
concluded that the palm kernel trade moderated the worst efects of the
slump from the 1870s, but it did not make up or compensate for the decline
in palm oil prices on both sides of the Atlantic between 1870 and the 1890s
(Lynn 1998, 126). In the Gold Coast, African producers would shit to rub-
ber – which soon witnessed a similar glut on the world market – and even-
tually cocoa.
he pattern of technological innovation and investment in the mid-nine-
teenth century creating over capacity and then resulting in plummeting
prices was repeated globally. he development of refrigeration meant that
steamships could bring produce and meat from distant places, encourag-
ing industrialized countries to specialize in high-end agricultural products
such as dairy, meat, fruits, and vegetables, while depending on imports for
their other food needs such as grains. he impact of the crisis of 1873–96
was varied in Britain. Among the four main sectors of the British economy
in 1870 – agriculture, industry, banking and inance, and shipping – nota-
ble disconnections occurred that shaped how the crisis was experienced.
Britain’s agricultural sector continued to decline, especially as Britain as the
pioneer industrialized nation believed it had comparative advantage in free
trade. Britain’s industries, such as iron and steel and textiles, remained prof-
itable by cutting back investments in machinery, but lost their competitive
edge ater the crisis was over, setting the context for British decline in these
industries. British banking and shipping in their international focus were
to some extent disconnected from internal or national industries. Banking
and shipping performed well in the crisis of the late nineteenth century,
while several merchant houses collapsed (Gourevitch 1986, 76–83).
he British merchants Horsfalls, pioneers in the palm oil trade, exited
West Africa in 1875. Other established companies that disappeared in the
1870s included R. & G. Benn, Tyson, Richmond and Jones, and Wilson
and Dawson. he bankruptcies continued in the 1880s with John Capper
and Co. (1880), Irvine and Woodward (1881), and Moritz Herschell (1886).
Other commercial giants collapsed in the 1890s: the Gold Coast household
name F & A Swanzy was taken over by Millers in 1894, Taylor, Laughland
& Co. folded in 1896 and W. B. MacIver in 1900 (Lynn 1998, 147–8). Lynn
notes that in a sense this was the end of the age of the older family busi-
nesses, which collapsed in spectacular fashion because of the economic
pressures of the period or because the original founder had died. hose
that survived merged to create joint-stock companies or found other ways
of reorganizing such as through amalgamations. he collapse of these fam-
ily businesses were replicated on the Gold Coast, and by the end of the
 Commerce, Credit, and Mobility 243

nineteenth century, the Ocanseys, Ghartey, Sarbah, and other notable

merchants who had emerged in the mid-century prosperity had also col-
lapsed. Austen comments on the unwillingness of even the most success-
ful Euro-African family businesses to adopt “the organizational and legal
structure of modern European businesses” (Austen 1987, 90). he beneit
of hindsight sheds light on the important need for these family businesses
to have restructured; caught up in the uncertainty of the late nineteenth-
century crisis, the causes of which remained unclear to contemporaries,
the way out was not that clear cut. heir collapse underscores how invested
they were in Euro-African trade, and how their commercial existence, like
the European merchants operating in West Africa, had become dependent
on the continuance of this commerce. Not all African merchants were able
together with producers to “simply drop out of the market until conditions
provided better incentives for commercial eforts” (Austen 1987, 102).
Olukoju (Chapter 6, this volume) provides us with contrasting examples
of African businesses that utilized the exit strategy described by Austen in
southwest Nigeria. Here some Yoruba merchants, whose businesses col-
lapsed in the economic crisis of the late nineteenth century, branched into
agriculture and other ventures or local strategies of accumulation not so
heavily dependent on Atlantic trade to recoup their losses and recapitalize
for a more opportune reentry into commerce. But then Yorubaland was not
annexed to British rule until the 1890s. Land and real estate represented
important avenues of investment and diversiication for merchants in the
early colonial period.
Having sketched the context of boom and bust during the nineteenth
century on both sides of the Atlantic within the global economy, it is time
to examine the repercussions on commercial relations in the Gold Coast.
he crisis marked the rise of tensions between Gold Coast merchants and
their European counterparts, as they struggled to make sense of the contin-
uous decline in palm oil prices compared to the stable prices of European
imports; extended the impact of British bankruptcies to the Gold Coast;
altered relations between large Gold Coast merchants and small traders in
a period of trade constriction; and dashed the hopes of small traders for
social mobility. We can now examine the local speciicity of transactions
and events and the social ramiications. How did merchants in the Gold
Coast understand or interpret the economic crisis of 1873–96? It is note-
worthy that the crisis began just before the commencement of colonial rule
in the Gold Coast in 1874. Did merchants in the Gold Coast have suicient
information to understand the economics of the crisis and not to see the
sudden adverse terms of trade as a product of the extension of colonial rule?
244 Akyeampong

Trust was a central pillar of Euro-African trade. What was the impact of
the crisis on trading relations? Can we discern the impact of the numerous
bankruptcies in Britain from the 1870s on African businesses in the Gold
Coast? And what was the impact of the economic downturn in the Gold
Coast on the social aspirations of traders?
I have chosen to focus on the most notable trading family from Ada
in southeast Gold Coast, the Ocanseys. With extensive trading relations
to Britain and Germany, the Ocanseys were without a doubt the “mer-
chant princes” of Ada. Ada was the outlet for Krobo palm oil producers,
and this was the hub of the palm oil trade. Dumett counts the Ocanseys
among the twenty-ive or so large merchants that emerged in the Gold
Coast between 1865 and 1900. he business documents – letter books, led-
gers, and accounting records (invoices, cash and wage books, etc.) – relat-
ing to the Ocansey business have been deposited at the Public Records
and Archives Administration Department (PRAAD) in Accra. heir sadly
deteriorating state makes it diicult to follow developments for a contin-
uous number of years from a single collection. Consequently, I draw on
the correspondence of a second Ada trading family, the Caesars, though
they did not operate on the grand scale of the Ocanseys. Correspondence
between J. H. Caesar and Sons with their British partners illumines some
of the questions raised on the local interpretation of a global economic cri-
sis. And I draw on the private papers of a third trader, Christian Jacobson,
who operated on a much smaller scale. hough based at Keta, a little to
the east of Ada, he was very connected to the commercial world of Ada
and was a protégé of the Ocanseys. Clearly knowledgeable about trade on
the coast, as evident from a long line of former expatriate and Gold Coast
employers and partners, Jacobson’s futile attempts to set himself up during
the 1880s and 1890s as an independent trader relects the closing of the
gates of opportunity in the late nineteenth century, compared to the pros-
perous 1850s and 1860s that had witnessed an explosion in the numbers of
small, independent traders.

7.4 THE SOCIAL WORLD OF COMMERCE IN THE GOLD COAST

7.4.1 William Narh Ocansey of Ada: A Merchant Prince
Ada, at the estuary of the Volta River and an important trade artery into
Krobo oil palm country and upcountry markets such as Kete Krachi,
was the base of the prominent irm of William N. Ocansey and Sons in
 Commerce, Credit, and Mobility 245

the second half of the nineteenth century.9 Formerly part of the Danish
sphere of inluence on the Gold Coast, the Danes had withdrawn in 1850,
selling their trading forts or lodges with the “spheres of inluence” to the
British. Ada thus came under British inluence. hat this transition coin-
cided with years of prosperity in the palm oil trade, for which Ada was a
major port, strengthened the place of Britain in the social world of Ada
merchants and traders. Anlo territory to the east, with its leading trad-
ing town of Keta, had also been transferred from Danish inluence to the
British (Nørregård 1966).
William Narh Ocansey was born in the 1820s, and Dumett, who inter-
viewed his son, described him as a “self-made man.” hough illiterate, he
built up an impressive business based on the export of palm oil to Britain
and Germany and the import of European manufactures of cloth, liquor,
and tobacco. Exploiting his strategic base at the mouth of the Volta, his
canoes also sent local products of salt and ish to central and northern
Ghana (Dumett 1983, 673). From the 1870s he was in correspondence with
mercantile irms in London, Liverpool, Glasgow, and Hamburg, who val-
ued his business and used him as a referee for other Gold Coast merchants
in Ada, the lower Volta region, and east of the Volta who desired credit
from European merchant houses. In 1877–8, a irm of Liverpool produce
brokers, Hickson and Sykes, referred to William N. Ocansey and Sons as
the “best traders on the coast” because of the irm’s ability to buy palm oil
cheaper than any other trader on the Gold Coast (Dumett 1983, 675). he
nucleus of Ocansey’s business seemed to have been in place even in the
1860s, when he is reported to have a network of twelve trading stations each
with a store manager.
At the height of its prosperity the Ocansey irm controlled branch factories at the
towns of Keta, Beh Beach, and Danu [Denu] Beach on the eastern (Anlo) shore of the
Volta estuary, as well as establishments at Agrave, Adidome, Battor, Weycoomabe,
Amanor, Tolesem, Adafoa, Prampram, Akuse and Kpong in the western of Ada[n]
gbe zone of inluence. (Dumett 1983, 675)

he Ocansey papers help us to understand the infrastructure of commerce

on the Gold Coast in the late nineteenth century.

9
Ghana Public Records and Archives Administration Department at Accra (PRAAD): SC
8: Ocansey Papers. Also important to this section are a number of pioneering articles writ-
ten by Raymond Dumett during the 1970s and 1980s on Gold Coast entrepreneurs, such
as the Ocanseys, and Gold Coast commercial history at the turn of the twentieth century
(Dumett 1971, 1973, 653–79, 1983, 661–93).
246 Akyeampong

he key exchange products were palm oil from the Gold Coast and the
addition of palm kernels from the 1870s and textile, liquor, and tobacco
imports from Europe with an assorted range of other minor items. William
Ocansey shipped his palm oil direct to Europe, though he also shipped
through the large European merchant houses such as the Basel Mission
Factory and Alexander Miller (Dumett 1983, 675). In the Gold Coast, it
has been noted that many of the major nineteenth-century merchants were
from royal lineages or were ailiated to royal lineages in some way (Dumett
1983, 671, 680; see Dumett 1972). In nineteenth-century Anlo, several chiefs
were prominent traders during the waning years of the slave trade and in
legitimate trade (Akyeampong 2001, 58). Chiefs who were prominent trad-
ers in Anlo from the mid-nineteenth century included Chief Anthony of
Addaia, Chief Amegbor of Klikor, Chief Hoku of Blekuso, Chief James
Ocloo of Keta, and Chief Akolatse, who we will encounter again in the
career of Christian Jacobson (Dumett 1983, 671). hough being a chief or
coming from a chiely family did not necessarily start late nineteenth-cen-
tury Gold Coast merchants of with inancial capital, they had enormous
social capital that facilitated their hiring of local labor and their sourcing of
produce. But membership in a chiely family also had its exactions in terms
of traders’ personal wealth. In the legal disputes that marked chietaincy in
Ada during the colonial period, it was pointed out during one of the inqui-
ries in the 1920s that the Ocanseys of the Kababiawe clan had provided the
manche (king) of Ada for the preceding two hundred years (Sutton 1984,
53). William Ocansey was the younger brother to the king of Ada. Ada’s
most important natural resource was salt from the Songor lagoon. It is not
surprising that William Ocansey and Sons was an important salt trader to
markets upcountry. hat the major oil producers of the Gold Coast, the
Krobo, were situated in the lower Volta region and were Adangme speak-
ers was an asset in Ocansey’s ability to secure large quantities of palm oil at
attractive prices (Dumett 1983, 673–4).10
Being a member of the royal family in Ada brought heavy inancial impo-
sitions on William Ocansey in the 1860s and 1870s. As probably the wealth-
iest merchant in Ada and brother to its king, William Ocansey came to the
support of his brother when the neighboring Anlo state attacked Ada in the
1860s. Ocansey reportedly lost £5,000 in this war, according to the personal
account of his son John E. Ocansey. It is not clear whether this loss was a
10
F. and A. Swanzy’s agent Alfred Triggs spent a fortnight traveling in the interior of the
Gold Coast in March 1870 going into Krobo country. He describes the “immense palm
forests which aford us the staple of our palm oil supplies at Accra and Prampram” (he
United Africa Company and Unilever Archives[UAC]: 2/33/AG/1/2/3, 1870).
 Commerce, Credit, and Mobility 247

contribution to the war efort or a result of damages to trade. And he almost

lost his life from a gunshot wound in this very war (Ocansey 1989, 10).11
Another “corporate” liability cost Ocansey $2,000 in 1875–6, when a ship
of F & A Swanzy, prominent British traders in the Gold Coast, was wrecked
of the coast of Ada. Swanzy’s agent claimed that the wrecked ship was pil-
laged by the residents of Ada and sued the king of Ada in the colonial court
for not protecting the ship. Awarded damages of $5,000, the residents of
Ada paid $2,000 and refused to pay any more as they claimed the Ada agent
for Swanzy (Charles Cassah) had removed goods from the ship before the
ship sunk, hence they contested the value the agent had put on the goods
aboard the ship. he colonial government put pressure on the king of Ada
to make good the rest, and the king paid $1,000 and his brother the trader
paid $2,000 (Ocansey 1989, 11–13). he war between Ada and Anlo took
place before colonial rule was declared in the Gold Coast. he court case on
the shipwreck showed how colonial rule had changed the operation of trade
and understandings of liability. During the precolonial period, shipwrecks
were considered fair game for looting. Residents in Ada found out just a
year ater colonial imposition that the rules of the game had changed.

7.4.2 Commerce and Credit in the Gold Coast in

an Era of Economic Crisis
he boom years of the 1850s and 1860s had increased considerably the
ranks of traders in the oil palm trade on both the European (Britain and
Germany) and West African sides. As market prices declined from the 1850s,
the investments European and African traders had made in their trade and
their dependence on the trade made it diicult for them to pull out of the
oil trade. European merchants continued to prime the pump, hopeful that
expanded trade would make up for the decline in prices. his explains on
the African side the expansion in the supply of palm oil – and later palm
kernels – even in the face of declining prices. In the second half of the nine-
teenth century, African merchants on the Gold Coast had stations or stores
in the interior with managers and agents who purchased produce such as
palm oil, palm kernels, and rubber, and who sold an inventory of European
manufactures. Such inland storekeepers advanced credit in goods to trad-
ers who dealt directly with producers, though store managers could also
source produce directly from producers. Coastal merchants could sell their
produce to European companies on the coast such as F. and A. Swanzy,

11
Written by John E. Ocansey during his short stay in Liverpool, England in the year 1881.
248 Akyeampong

Pickering and Berthoud, or Miller Brothers or send their produce directly

to Europe to a consignment house via steamship service. hese consign-
ment houses then sold the produce on behalf of the African merchant,
purchased and shipped their list of requested European manufactures, and
kept an accounting of these imports and exports. Consignment houses
charged between 2.5 and 5 percent commission on the net invoice value
of goods they purchased and shipped to African merchants from Europe
(shipping and insurance extra), and another 2.5 to 5 percent commission
for selling palm oil, palm kernels, and rubber on behalf of these African
merchants and traders (Dumett 1973, 664). Dumett notes the liberal exten-
sion of credit in the mid-nineteenth century by British consignment houses
to small traders throughout the British Empire, and how this in turn eased
local credit terms and facilities on the Gold Coast, providing a boost in gen-
eral to trade (Dumett 1973, 655).
he role of large merchants such as William and John Ocansey as refer-
ees and guarantors for small traders needs to be emphasized as an impor-
tant factor in the expansion of trade in the late nineteenth century, as
European merchant houses utilized them as supervisors and even enforc-
ers of goods given in trust, especially for traders these African merchants
had recommended. As palm oil prices on European markets continued
to drop, while prices for European manufactures remained constant or
rose, many African merchants and traders were pushed into debt with
consignment houses, breeding mistrust concerning European handling of
African accounts. African merchants were oten compelled to request an
extension of credit from consignment houses, and the latter were willing
to oblige at an added interest rate of about 8 percent per annum (Dumett
1973, 665–6).
John E. Ocansey, the literate adopted son of William Ocansey and mar-
ried to Ocansey’s daughter Salome, handled business correspondence on
behalf of William N. Ocansey and Sons, and served as credit enforcer for
minor merchants and African traders. He seemed to have guaranteed credit
with consignment houses for Christian Jacobson of Keta, and for smaller
Ada traders like J. T. Amenyar, Tetteh Dosoo, D. J. Adoajee, heo Taye,
and other members of the Ocansey family such as W. B. Ocansey and J. C.
Ocansey.12 he purchases of small traders were sometimes put on William
N. Ocansey’s bills of lading, which Ocansey and Sons would then deliver

12
See, for examples, PRAAD: SC 8, Ocansey Papers, Donner and Callenburg, 1889; PRAAD:
SC 8, Ocansey Papers, Hart and Co., Oct. 1889; PRAAD: SC 8, Ocansey Papers, Hart and
Co., Sept. 1888; PRAAD: SC 8, Ocansey Papers, Donner and Callenburg, 1888.
 Commerce, Credit, and Mobility 249

(PRAAD: SC 8/64, Edward Bros.). Christian Jacobson’s business was strug-

gling inancially in the 1880s as discerned from the Ocansey correspon-
dence. In August 1889, Jacobson acknowledged the accuracy of Donner
and Callenburg of Hamburg’s accounting that he owed them £217.1.4 [?].
Jacobson proposed a scheme by which he would gradually pay of his debt
and improve on his business with the reduction in his liabilities (PRAAD:
SC 8, Ocansey Papers, Jacobson). Donner and Callenburg forwarded a copy
of Jacobson’s letter to John E. Ocansey.

Certainly we agree to this arrangement with Mr. Jacobson only when you pro-
longate your letter of guaranty given us for the debt of Mr. Jacobson to which no
doubt you agree to. We enclose a letter for Mr. Jacobson and beg to hand it over to
him when you agree to the above when you ind it more advantageous to stop all
business with him you will kindly return the letter to us. (PRAAD: SC 8, Ocansey
Papers, Jacobson, Donner and Callenburg 1889)

Jacobson’s account with Lionel Hart and Co in Liverpool was also bad, and
the irm reminded John E. Ocansey in regard to Jacobson “that you per-
sonally in the irst instance asked us to assist him for [your] sake we did so,
now we must ask you to use your unlimited inluence upon him to remit,
or we will have to quarrel about his ac[count] and you only can prevent
this unpleasantness happening” (PRAAD: SC 8, Ocansey Papers, Jacobson,
Hart and Co., Dec. 1888). Eight months later, Lionel Hart and Co was blam-
ing John Ocansey “for having introduced such a bad man to us” (PRAAD:
SC 8, Ocansey Papers, Jacobson, Hart and Co., Aug. 1889).
Bad faith operated on both sides, as admitted in the letter of the Liverpool
irm Fletcher and Fraser, as they made a bid for the business of Ocansey and
Sons in 1887.

We have no fear in saying that we should give you satisfaction, and of this be sure
for every sovereign we might be fortunate to earn from you in the way of commis-
sion. Firms which shall remain nameless would extract ive pounds by cutting your
clothes, lowering the qualities, charging you for bags in which rice is shipped, not
giving you the full price paid for your produce, and other [un]pleasant little tricks for
which they are famous. (PRAAD: SC 8, Ocansey Papers, Jacobson, Fletcher 1887)

Mistrust clouded African interpretation of the economic crisis of 1873–96,

and responsible European merchants did their best to allay the fears of their
African partners that the decline in palm oil prices was not a result of price
manipulation by European merchants.
Frequently in the 1880s, commission houses would dispatch together
with the goods ordered by their African partners weekly produce prices
on the European market to inform their clients of price luctuations on the
250 Akyeampong

market. hese dispatches also listed prices for competing oils and fats such
as Australian tallow. Manchester merchants Pickering and Berthoud in
their letter to J. H. Caesar and Sons on July 13, 1888, noted the long silence
from their client in the Gold Coast.
We trust we may soon hear again from you that we shall soon have further and reg-
ular remittances so there may be no delay with your orders. It is now some [time]
since we had the pleasure of hearing from you, but the “Congo” from your front is
now due and we look for news by her. We enclose the usual weekly produce prices
current. Palm oil today is nominally about £16. 15/- per ton. Market inactive with
buyers holding back. (PRAAD: SC 13/11, J. H. Caesar’s Trading Papers)

As of September 21, 1888, when the Manchester merchants wrote again to

J. H. Caesar and Sons, they still had not heard from their Gold Coast cli-
ents. hey however reported a slight improvement in palm oil prices in the
previous fortnight (PRAAD: SC 13/11, J. H. Caesar’s Trading Papers: SC 8,
Pickering Sept. 1888).
Pickering and Berthoud’s letter to Caesar and Sons in October the fol-
lowing month acknowledged inally hearing from their Gold Coast clients
and made suggestions on how to improve the eiciency of their business,
including using the local agent of Pickering and Berthoud in the Gold
Coast to make out orders for the irm instead of Caesar and Sons using their
own clerks. he letter relected the strained atmosphere of Euro-African
trading relations during this period. In this particular instance, Caesar and
Son’s order for a surboat through Pickering and Berthoud was yet to be
fulilled.
Surboat. On the 14th April last you sent us an order for this: on the 23rd idem you
cancelled the order, and on the 28th of the same month you reordered it. Now this
sort of thing is to say the least extremely confusing as we have an agent at your front
whose duty is speciically to attend to the needs of our various clients. We cannot
understand why you do not take advantage of his services. . . . .
We would however repeat the necessity of getting Mr. Vanderpuye to make out
your orders as the manner in which this last indent is made out is very deicient and
misleading, should any error in its execution be made by us, though being misled
by your clerk. You would no doubt [put] the whole blame on us, as you threaten, on
certain conditions send the goods back to us. [emphasis added] (PRAAD: SC 13/11,
J. H. Caesar’s Trading Papers, Oct 1888)

hey reported an improvement in palm oil prices to £20. 5/ to £20. 10/ per
ton for Ada quality, and that the market remained strong.
Part of the response of Gold Coast traders in this era of economic crisis
was to search for other commission houses that would be more faithful
in the conduct of business or provide better prices and services. Popkin
 Commerce, Credit, and Mobility 251

notes how customer markets in produce (e.g., rubber) as distinct from

auction markets (e.g., rice) tend to have higher middleman margins, so
Gold Coast palm oil suppliers were not wrong in seeking better deals.
But Popkin also emphasizes the importance of reputation in produce
markets such as rubber – and by extension palm oil – in securing top
value for quality produce. Hence, established traders with long relations
with European consignment houses who trusted the quality of their pro-
duce were unlikely to switch brokers unless deeply provoked (Popkin
1981). Caesar and Sons began to explore from 1889 a new relationship
with Radclife and Durant in Liverpool. hrough their ailiate Christian
Jacobson of Keta, who also operated in Ada, Akuse, and the River Volta
district, Caesar and Sons sent palm oil on a trial basis to Radclife and
Durant. he response of the Liverpool irm was enthusiastic: “We shall
do all in our power to merit your conidence and hope by so doing we
shall be able to do an extensive business with you to our mutual advan-
tage” (PRAAD: SC 13/11, Radclife and Durant). Radclife and Durant
from their correspondence seem to have worked hard to earn the trust
of Caesar and Sons. heir correspondence addresses the volatility of the
market in the Gold Coast in this time of crisis. Caesar and Son’s attempt
to cancel an order for guns and matchets in August 1890 as local demand
had shited could not be obliged as Radclife and Durant had already
shipped the guns aboard the steamer Africa, and “the matchets are well
in work and cannot now be cancelled” (PRAAD: SC 13/11, Radclife and
Durant, Aug. 1890). In the insecure times of the late nineteenth century, a
missed detail such as opting not to insure one’s merchandise could be the
tipping point to irrecoverable inancial losses. Radclife and Durant were
able to replace twenty cartons of headkerchiefs (scarves?) that Caesar and
Sons had reported lost and had sent a certiicate of loss.
We have claimed on the insurance company for the value and no doubt we shall
receive payment in due course. In the meantime we have enclosed another lot of
20 cts to be made and they will be shipped in a few weeks. (PRAAD: SC 13/11,
Radclife and Durant, July 1890)

When the irm of Ocansey and Sons declined in the 1890s, it was a result
of a combination of losses, including inancial embezzlement by agents on
the coast of West Africa, and bankruptcies sufered by partners in Europe.
African businesses sought to reorganize in the crisis of 1873–96 to become
more eicient. he ability to deliver produce safely to coastal ports from
inland rivers was important, considering the rough surf that characterized
the coast of Guinea. We noted earlier the attempt by Caesar and Sons to
purchase a surboat through Pickering and Berthoud. he success of the
252 Akyeampong

Swanzys and George Watts on the west coast of Africa was partly based on
their efective use of steam launches to open up new sources of palm oil
(Lynn 1998, 146). A major inancial blow was dealt to Ocansey and Sons
when the Liverpool irm of Messrs. Hickson, Sykes & Co declared bank-
ruptcy in 1881. William Ocansey had paid a substantial amount to Robert
W. Hickson for the purchase of a steam launch for his business on Ada. It
is the impact of bankruptcies in Britain on Gold Coast traders that we turn
to in the next section.

7.4.3 Bankruptcies in England and the Gold Coast:

Interconnectedness of Businesses
In 1878, Ocansey sufered a loss of £630 when forty butts of palm oil loated
down the Volta River could not be stopped at the estuary at Ada and were
lost to the sea because of rough surf. his incident – and the death of his son
Gordon Samuel on April 20, 1880, who managed his upriver factory – lay
behind Ocansey’s decision to purchase a steam launch for his palm oil busi-
ness on the Volta (Ocansey 1989, 14–15). he steam launch would improve
the conduct of his business up the river, eliminate losses that resulted from
loating palm oil casks down the river, and reduce dependence on agents
who had cost him huge sums of money in the past through the extension of
credit in goods. William Ocansey turned to Robert Hickson of the commis-
sion house of Messrs. Hickson, Sykes & Co. to purchase the steam launch
estimated at about £2,200. Ocansey sent produce worth £2,678. 11s. 8d. to
Hickson, who in his correspondence assured Ocansey that the launch would
arrive in Ada by July or August 1880. he steam launch never arrived, and
following a string of broken promises, Ocansey, suspecting foul play, sent
his son John to Liverpool aboard the Mayumba in late April 1881 to ascer-
tain the state of afairs.
John on arrival in Liverpool would discover that the irm of Hickson,
Sykes & Co. had declared bankruptcy some months earlier, but lack of
information had prevented Ocansey and sons from knowing this.
On the day following my arrival, I made inquiries about Hickson, Sykes & Co., and
then I found they had been declared bankrupt, and that the 17th of June had been
appointed for the meeting of their creditors. his was terrible news. I felt my heart
sink, and I could not eat for three days. I thought of dear father and the heavy loss
he would sustain of £2,678, besides all the delay in the enlargement of his business
for want of the steam launch, for which he had already paid, and which ought to be
out and in full working order. I now found out that I could do nothing but wait and
attend the meeting of the creditors. (Ocansey 1989, 26)
 Commerce, Credit, and Mobility 253

John E. Ocansey sued Robert Hickson, but the trial was delayed until the
end of July, detaining him in Liverpool. In the interim, John visited busi-
ness contacts in London, Manchester, and elsewhere, some of these meet-
ings of which were arranged by a long-standing partner of his father’s irm,
Moritz Herschell.
he trial was a travesty of justice, as counsels for the prosecution and
defense, Mr. Carver and Mr. Aspinall, seemed to have reached some pri-
vate agreement not to press the more damning charge of defrauding the
Ocanseys under false pretenses, though the correspondence between the
two irms supported that charge. he judge’s questioning led clearly to the
fact that Hickson misled the Ocanseys into believing that the launch had
been ordered and was on the way, when even the order for the launch had
not been placed with the shipbuilders. Meanwhile, according to Hickson’s
defense, Mr. Aspinall, Hickson’s produce broker appropriated the proceeds
of the Ocansey shipments against debts owed him by Hickson. Justice
Lopes’s sentencing remarks sum up the case.
Robert William Hickson, you have pleaded guilty to having misappropriated
a large sum of money that had been entrusted to you as agent. You were also
indicted for obtaining a very large sum of money by false pretences. No evidence
was ofered on that indictment, and you were acquitted. I desire to say that, if that
indictment had been proceeded with, and you had been convicted, I should have
felt it to be my duty to direct you to be placed in penal servitude for a consider-
able time, because the ofence would be a most serious one. he ofence to which
you have not pleaded guilty is also a most serious one. Everything that can be said
in your favour was urged by your counsel in palliation of the charge against you,
and that charge is that you misappropriated something like 2,600. It has been said
that you got into diiculties, and that you drited into this act of dishonesty. It
appears that in the month of May, 1880, you received special orders to purchase a
steam launch for your employers in West Africa, and it also appears that you did
not order it until December of that year. In the interval money enough had come
into your hands to pay for it, the whole of which you did not apply to the purpose
you were directed to apply it, but devoted it instead to your own purposes. I am
desirous, as far as I possibly can, to give efect to any extenuating circumstances
which I can ind in your favour; but I cannot help saying that I ind great diiculty
in discovering any in this case, and I can pass upon you no other sentence than
that you be imprisoned with hard labor for iteen calendar months. (Ocansey
1989, 58)13

13
(Ocansey 1989, 58). he huge disappointment to William Ocansey of this incident in the
history of interracial, transatlantic trade relations has been reiterated in a recent book by
Caryl Phillips on the Atlantic’s historic connections (Phillips 2001).
254 Akyeampong

Later that same decade, Moritz Herschell, who was a pillar of strength to
John Ocansey in the latter’s stressful visit to Britain, declared bankruptcy in
1886 (Lynn 1998, 148). At this moment I am unable to ascertain the impact
of Herschell’s bankruptcy on Ocansey’s business, but this is a line of inquiry
worth exploring as we can see in the case of Hickson.
hese unfortunate incidents compelled William Ocansey and Sons to
make changes in its conduct of business, including a decision that it would
only ship produce on receipt of irm goods. he Hamburg irm of Wolber
and Brohm complained to John in 1887:
We know that your senior [William N. Ocansey] has made bad experience with
certain Merchant irms, but who has not during his life and how much money is
lost in Africa; we understand that since you have adopted a policy of expecting irm
goods to be there, before shipping against their pay, not so much because you may
require credits given to your good selves as with the open introduction, so to speak
to be constantly on the safe side. his manner of business involves distrust, which
misplaced, creates at once ill feelings. (PRAAD: SC 8/64, Wolber and Brohm)

And following the hard-nosed business practices of the consignment irms

in Europe, who held John E. Ocansey partially liable for bad debts incurred
by traders he had recommended, John also insisted on collecting commis-
sion on produce originating from traders he had recommended, such as
Christian Jacobson and other Ocansey relatives who operated their own
businesses (PRAAD: SC 8, Hart and Co., Nov. 1888).
By the end of the nineteenth century, the irm of William N. Ocansey
and Sons was in decline and William Ocansey is said to have exhausted his
wealth by the time of his death. Dumett explains the collapse of the impres-
sive Ocansey business:
he Ocansey irm’s slow decline can be traced to a combination of natural disas-
ters and external factors, including loss by ire, shipwreck, robbery, and involve-
ment in warfare against the neighboring Anlo people. William Ocansey showed
amazing resilience in rebounding from many of these misfortunes, but in the end
succumbed to general loss from the declining palm oil trade coupled with default
on credit by local middlemen and oppressive treatment by European competitors.
(Dumett 1983, 676)

he liberal era of European trade credit to small traders throughout the

British Empire that we have noted in the mid-nineteenth century had come
to an end. he advent of colonial rule and the political economy of colo-
nialism remapped the context and the nature of trade in West Africa. Not
only were African merchants and traders displaced in Atlantic trade net-
works, regional long-distance trade systems that had predated colonial rule
 Commerce, Credit, and Mobility 255

for generations were undermined. We look at the closing of these gates of

opportunity for the career of young Christian Jacobson as we bring this
section to a close.

7.4.4 Mobility Denied: he Frustrated Ambitions

of Christian Jacobson, ca. 1880–1912
Christian Jacobson was by all accounts a talented young trader who in the
1870s and 1880s had a bright future in trade. Jacobson grew up during the
prosperous 1850s and 1860s and had done the usual apprenticeship as an
agent or a clerk for an expatriate mercantile company, the common route
at this time for acquiring experience and knowledge in trade, as well as a
network of both European merchants and African traders and producers.
Young Jacobson was the agent or factor in charge of the lagoon shop of
the Bremen Factory in Keta in the 1870s (PRAAD: SC 14/3). He certainly
seemed to have made a favorable impression on many, and an established
Anlo trader, Chief Akolatse, sought to woo Jacobson away from the
Bremen Factory. Jacobson let the Bremen Factory in 1879 ater ive years
employ to work with Chief Akolatse on terms that were unclear and would
be disputed in an 1892 legal suit. What can be discerned from the court
proceedings is that Akolatse lent Jacobson several large sums of money
to go into business partnerships. Jacobson assumed that Akolatse was his
partner. Akolatse argued that these were loans. In these arrangements,
Akolatse managed to receive his imported goods through Jacobson’s busi-
nesses at invoice prices, depended on Jacobson for his business corre-
spondence, and even at one time rented one of his houses for Jacobson to
be used as the premises for his business while retaining personal use of
a section of the house. Clearly, Akolatse was an astute businessman who
sought to beneit from the talents of the young Jacobson without being
liable for any debts or losses Jacobson would incur in this volatile eco-
nomic period.
he irst major setback the young Jacobson sufered was the bankruptcy
of the British commission house he used. Jacobson had been trading with
Messrs. Taylor & Co. of 72 Virginia Street in Glasgow. He sent the irm
African produce valued at £183 and the irm did not return the money or
send the goods Jacobson had ordered. Rather, it declared bankruptcy, and
unfairly instructed Jacobson to go and collect his money from its debtors in
Grand Bassam in the Ivory Coast. Jacobson was John Ocansey’s best friend
when they were youths, and a relative of the Ocansey family. John Ocansey
256 Akyeampong

relected on the impact of this early setback on the young Jacobson’s trading
career.
Now, Mr. Jacobson is a young man just venturing out into trade; and it does seem
very cruel that he should meet with such heartless conduct, and sufer so great a
loss, on the doorstep of his life. To ask him, who lives at Quitta [Keta], to collect
another man’s account from the people of Grand Bassar, is adding insult to injury.
(Ocansey 1989, 38)

Jacobson seemed to have followed the example of his friend John, who had
gone to Liverpool in pursuit of the investment his father had made in a
steam launch. Jacobson boarded a ship to Liverpool at the personal expense
of £54 for his return ticket, with the intent to continue to Glasgow to collect
his money from Messrs. Taylor & Co. John was still in Liverpool awaiting
the outcome of his case against Hickson and was able to provide lodging for
Jacobson. Jacobson wrote to the irm of Taylor & Co. from Liverpool and
received a very duplicitous response.
But they sent a letter back at once, saying they had nothing to do with their past
afairs; that they had paid a part or composition to those who would receive it and
as to his account they still expected it would be got from the people of Grand Bassar,
and they would say no more as to the matter, but if Mr. Jacobson would send them
any more produce they would do well for him. (Ocansey 1989, 39)

Jacobson returned empty-handed to Keta, saddled with a huge debt very

early in his career that Chief Akolatse would insist was a loan and not a
partnership. He set a historical record, though, as the irst Anlo to have vis-
ited Europe, albeit under unhappy circumstances. Bankruptcies by British
businesses worked against their African clients who oten lacked even the
information to know that bankruptcy had been declared and a settlement
was to be reached with creditors.
Jacobson’s subsequent string of unsuccessful business partnerships
speaks to the unpredictable economic conditions of the 1880s. From 1880
to 1881, Jacobson went into partnership with Mr. Wood, agent for Edwards
and Bros. Akolatse lent Jacobson £245 for this endeavor. Next came a part-
nership with a Mr. Bellois of Bremen from 1881 to 1883, and again Akolatse
lent £201 for the business and received his imported goods at invoice
prices. In addition he charged Jacobson and Bellois £10 a month for rent
for the house from which they ran their business. hen came business with
one Duncan, agent for Taylor and Co. Akolatse seemed to have retained
Jacobson in some capacity between 1880 and 1886, and several witnesses,
including trader Chief James Ocloo, attested that in their trade dealings
with Akolatse, Jacobson had represented Akolatse (PRAAD: SC 14/3). As
 Commerce, Credit, and Mobility 257

of 1892, Jacobson was still paying for the debts he had incurred to Akolatse
from the early 1880s. Indeed, it appears that Akolatse even worked against
the aspirations of his protégé. Jacobson recalled how he or the defendant
had proposed around 1887 that they build a house for the business at
the lagoon side, the house to be built under the supervision of Jacobson.
Akolatse secured building materials on credit from the local agent of the
Hamburg irm of Rotterman for the project. Jacobson recounted the demise
of the project.

It was to be a storey house but I only built the walls. One day det. [Chief Akolatse]
told me that he withdrew and transferred the house to me. He said a boy should not
build a storey house. hen Rotterman sent a power of attorney to his nephew . . . to
demand from det the amount for which he was surety for me. But W. N. Ocansey
paid the debt for me. Det. charged me £101.14.1 for the materials that he had got
from Rotterman at Adda in ’87 [1887]. (PRAAD: SC 14/3)

he traditional antagonism between chiefs or elders and young men had

reared its head. In this instance, William Ocansey, the established merchant
at Ada, had intervened to bail out his young relative by paying Rotterman
the amount involved. Old Ocansey may also have discerned the manip-
ulation of Chief Akolatse and his exploitation of young Jacobson, and
instructed his son John to assist Jacobson in his business. John introduced
Jacobson to British and German commission houses, serving as his guar-
antor. As the letters from British and German businesses to John indicate,
these initiatives failed and they blamed John for recommending an unreli-
able trader to them. It is clear from the circumstances of Jacobson’s career
that the fault perhaps was due more to the unpredictable business environ-
ment than a character law.
A peeved Jacobson sued Chief Akolatse in the Supreme Court of the
Gold Coast (Eastern Province) in 1892 for wages as Akolatse had not paid
him wages for the ive or six years that he served Akolatse in the capacity
of a clerk. Jacobson’s reasoning was that if he was not a partner to Akolatse,
hence making Akolatse jointly liable for debts incurred, then he was an
employee entitled to remuneration for his services. Akolatse in his defense
argued that Jacobson let his employ in 1884, at which time he had given
Jacobson a git of $410. To be more precise, he reduced the amount Jacobson
owed him by this amount. He also claimed that Jacobson wrote a single let-
ter on his behalf to McIver of Glasgow in June 1884, but that was in the
capacity of a friend and not a clerk. he case was dismissed with costs. As
late as 1912, Jacobson was still struggling to ind his feet in the world of
commerce. In that year, he was appointed agent or representative for the
258 Akyeampong

Volta River Districts of an American trading company based in Brooklyn,

New York, the Akim Trading Company. his was probably an African
American business. he times had certainly changed, but not the fortunes
of Christian Jacobson.

7.5 DENOUEMENT: COLONIAL RULE AND THE DEMISE

OF WEST AFRICAN MERCHANT PRINCES
he role of many European traders as advocates for European colonial rule
in the late nineteenth century relected their belief that they could dictate
trading conditions and that their proits would increase under a political
system in which they were dominant. Hopkins noted the importance of
the economic downturn at the end of the nineteenth century in generating
momentum for the colonization of Africa and of the 1930s depression in
the decolonization of Africa (Hopkins 1973). Indeed, in many instances
in West Africa, the treaties of protection that became the basis of formal
European colonial rule had been entered into by European merchants with
African chiefs on behalf of their national governments. George Goldie’s
treaties with chiefs of the lower and middle Niger was instrumental to
the drawing of the colonial map of Nigeria. Big British traders like John
Holt openly showed their displeasure when British tardiness resulted in
the German annexation of the Cameroons (NAB, FO 84/1660). European
merchants like Alexander Miller who believed that Jaja of Opobo in the
Niger Delta had been treated unfairly in his defense of his political and eco-
nomic sovereignty by his deposition and exile in 1887 were rare.14 Colonial
rule deinitely changed the environment of trade. With the protection of
colonial rule, European mercantile houses increasingly showed a prefer-
ence for European agents on the West African coast. When Lebanese trad-
ers proved a suitable and cheaper alternative, they, not African merchants,
would become the recipients of European credit, as racial lines hardened
with the advent of scientiic racism at the end of the nineteenth century
(Akyeampong 2006; Bierwith 1999; Boumedouha 1990; Falola 1990; Van
der Laan 1975).

14
Jaja (1991) reproduces voluminous correspondence in which Miller Brothers petitioned
the Foreign Oice about the unjust manner in which Consul Johnston had treated Jaja,
counseled Opobo chiefs on the right demeanor so as not to prejudice the Jaja case, and
even ensured that Jaja was extended funds through F. and A. Swanzy while resident in
Accra for his trial.
 Commerce, Credit, and Mobility 259

But the collapse of West African merchants in the late nineteenth

century cannot be attributed only to the onset of colonial rule and eco-
nomic discrimination. First, this chapter has shown that West Africans
were incorporated into productive processes or systems in Europe based
on technological developments beyond their control. he industrial pro-
cesses that created a need for vegetable oils were based in Europe. Africans
were happy beneiciaries of this development but not critical to its ori-
gin or sustenance. Second, it appears that coastal merchants in the Gold
Coast had become overly dependent on Atlantic commerce compared to
Yoruba merchants in the nineteenth century, who tapped into Atlantic
commerce, but also farmed extensive acres to supply food to the large
urban populations in Yorubaland. hey were enmeshed in international,
regional, and local markets, shiting their emphasis as economic circum-
stances demanded. hird, West Africans represented one supply source in
an increasingly competitive market to supply Europe with oils. And palm
oil, for that matter, was not indispensable to the productive processes for
which it was required, increasingly becoming limited where specialized
processes required nongeneric oils. Austen points to how at the end of
the nineteenth century, the relative value of West Africa’s trade within the
Atlantic economy was already declining: “Within this market, the capac-
ity of Africans to purchase the growing output of European products
remained limited compared to more populous and accessible portions of
the world” (Austen 1987, 112). he economic situation of West Africa at
the end of the nineteenth century foreshadowed Africa’s economic pre-
dicament in the twentieth century. As Dumett contemplated the decline
of the African merchant princes at the end of the nineteenth century, he
could not escape the conclusion that “[i]nherent distortions in the world
market for tropical commodities plus biases in the political economy of
colonialism, which favored Europeans, lie at the root of this discussion”
(Dumett 1983, 693).
One feature of nineteenth-century trade in West Africa survived into
the twentieth century and the political economy of colonialism: the prom-
inence of chiefs in trade or business. In southern Nigeria and the Gold
Coast, chiefs were prominent traders during the era of the slave trade,
and moved their capital and control over labor into the export of cash
crops during the era of legitimate trade. Colonial rule was initially inim-
ical to chiefs, viewed as the major opponents to European rule. But as
colonial rule became consolidated, the British shored up the authority of
chiefs, whose collaboration became essential to the extension of European
260 Akyeampong

hegemony. Colonial administrations would even sponsor chiefs in busi-

ness who lacked business acumen. So, during World War I, when wartime
shipping shortage led to oicial restrictions on cocoa exports, the governor
of the Gold Coast, Hugh Cliford, in 1917 gave export licenses to the nota-
ble paramount chiefs of Accra, Akuapem, and Akyem Abuakwa to export
cocoa. hat they sought to use their stools as collateral with the Colonial
Bank to charter a ship perhaps relects their lack of business savvy. he
resulting outcry led to the destoolment of the paramount chief of Accra
(Parker 2000, 211). In Lagos, “white cap” chiefs, or those who owned the
land in Lagos, were instrumental in the establishment of British colonial
rule. One study underscores the economic beneits that came out of this
relationship as the white cap chiefs became the trading partners of the
British during the colonial era, which positioned them as businessmen
and captains of industry at the end of colonial rule (Oyo 1989). For other
African traders in early colonial West Africa, the surest route to success
was to be extended credit facilities by an expatriate company or to be made
agent – particularly sole agent – to distribute an import that was in great
demand. With the consolidation of European rule and the expansion of
expatriate economic activities, these opportunities became more rare.
hus W. A. Dawodu’s pioneering role in motor transport, his sole agency
for a period of Ford cars and Humber cycles, and automobile servicing in
the 1910s and 1920s brought him great wealth in his operations in Lagos
and other major Yoruba cities, but he would fall victim to global economic
adversity from the mid-1920s and ierce competition from expatriate irms
with access to greater capital (Olukoju, Chapter 6, this volume). By the early
twentieth century, land had become one of the most important sources of
indigenous capital in a context of declining credit for Africans, facilitating
the resurgence of chiely wealth and inluence. he century that began with
the abolition of the slave trade and the advent of legitimate trade, a cen-
tury that had witnessed ordinary commoners become economic producers
and consumers, that had occasioned the lourishing of African merchants
and traders, that had seen the political and economic foundations of chief-
taincy challenged, ended with the chiefs in a new relationship with colo-
nial rulers and being groomed as businessmen and the indigenous voice
for the colonial agenda.

References
Akyeampong, Emmanuel. (2001). Between the Sea and the Lagoon: An Eco-Social History
of the Anlo of Southeastern Ghana c.1850 to Recent Times. Oxford: James Currey.
 Commerce, Credit, and Mobility 261

(2006). “Race, Identity and Citizenship in Black Africa: he Case of the Lebanese in
Ghana.” Africa, 76(3): 297–323.
Anstey, Roger. (1975). he Atlantic Slave Trade and British Abolition, 1760–1810. Atlantic
Highlands, NJ: Humanities Press.
Austen, Ralph A. (1987). African Economic History: Internal Development and External
Dependency. London: James Currey.
Bierwith, Chris. (1999). “he Lebanese Communities of Côte d’Ivoire.” African Afairs,
98(390): 79–99.
Boumedouha, Said. (1990). “Adjustment to West African Realities: he Lebanese in
Senegal.” Africa, 60(4): 538–49.
Brown, Christopher L. (2006). Moral Capital: Foundations of British Abolitionism.
Chapel Hill: University of North Carolina Press.
Davies, P. N. (1977). “he Impact of the Expatriate Shipping Lines on the Economic
Development of British West Africa.” Business History, 19(1): 3–17.
Drescher, Seymour. (1977). Econocide: British Slavery in the Era of Abolition. Pittsburgh:
University of Pittsburgh Press.
Duke, Antera. (1956). “he Diary of Antera Duke, an Eik Slave Trading Chief of the
Eighteenth Century.” In Eik Traders of Old Calabar, edited by Darryl Forde, 27–78.
London: Oxford University Press.
Dumett, Raymond E. (1971). “he Rubber Trade of the Gold Coast and Asante in the
Nineteenth Century: African innovation and Market Responsiveness.” Journal of
African History, 12(1): 79–101.
(1972). “Compilation of African Merchant and Traders’ Agents of Major Towns in
Ghana.” Transactions of the Historical Society of Ghana, 13: 260–4.
(1973). “John Sarbah, the Elder, and African Mercantile Entrepreneurship in the
Gold Coast in the Late Nineteenth Century.” Journal of African History, 14(4):
653–79.
(1983). “African Merchants of the Gold Coast, 1860–1905 – Dynamics of Indigenous
Entrepreneurship.” Comparative Studies in Society and History, 25(4): 661–93.
Edward Brothers, Liverpool, to Messrs William N. Ocansey, March 17, 1887.
Eltis, David. (1987). Economic Growth and the Ending of the Transatlantic Slave Trade.
New York: Oxford University Press.
Falola, Toyin. (1990). “Lebanese Traders in Southwestern Nigeria, 1900–1960.” African
Afairs, 89(359): 523–53.
Ghana Public Records and Archives Administration Department at Accra: SC 8, Accra.
Ghana Public Records and Archives Administration Department at Accra: SC 8,
Schedule of Ocansey Papers.
Ghana Public Records and Archives Administration Department at Accra: SC 8/64,
Accra.
Ghana Public Records and Archives Administration Department at Accra: SC 8/64,
Accra, Letter Book.
Ghana Public Records and Archives Administration at Accra: SC 13/11, Radclife and
Durant to Messrs. J.H. Caesar and Sons, Liverpool.
Ghana Public Records and Archives Administration at Accra: SC 14/3, Christian
Jacobson Papers, Accra.
Gourevitch, Peter. (1986). Politics in Hard Times: Comparative Responses to International
Economic Crises. Ithaca, NY: Cornell University Press.
262 Akyeampong

Hargreaves, John. (1986). “he European Partition of West Africa.” In History of West
Africa, edited by J. F. A. Ajayi and Michael Crowder, 408–28. Essex: Longmans,
2nd edition.
Headrick, Daniel R. (1981). Tools of Empire: Technology and European Imperialism in the
Nineteenth Century. New York: Oxford University Press.
Hopkins, Antony G. (1973). Economic History of West Africa. London: Longman.
Jaja, S. O. (1991). Opobo since 1870: A Documentary Record with an Introduction. Ibadan:
University of Ibadan Press.
Jasanof, Maya. (2011). Liberty’s Exiles: American Loyalists in the Revolutionary World.
New York: Alfred A. Knopf.
Lovejoy, Paul E. and David Richardson. (1999). “Trust, Pawnship and Atlantic History:
he Institutional Foundations of the Old Calabar Slave Trade.” American Historical
Review, 1049(2): 333–55.
(2001). “he Business of Slaving: Pawnship in Western Nigeria, c.1600–1810.” Journal
of African History, 42(1): 67–89.
(2004). “`his Horrid Hole’: Royal Authority, Commerce and Credit at Bonny, 1690–
1840.” Journal of African History, 45(3): 363–92.
Lynn, Martin. (1989). “From Sail to Steam: he Impact of the Steamship Services on the
British Palm Oil Trade with West Africa, 1850–1890.” Journal of African History,
30(2): 227–45.
Lynn, Martin. (1998). Commerce and Economic Change in West Africa: he Palm Oil
Trade in the Nineteenth Century. Cambridge: Cambridge University Press.
Marsters, Kate Ferguson. (2000). Travels in the Interior Districts Africa by Mungo Park.
Durham, NC: Duke University Press.
National Archives of Britain(NAB): FO 2/20. Dr. Balfour Baikie to the Secretary of the
Cotton Suppliers’ Association of Manchester, Lukoja, April 3, 1861.
NAB: London, FO 2/47. Consul Charles Livingstone to Lord Stanley, Fernando Po,
October 1, 1866.
NAB: London, FO 84/1660. FO note by Lister, December 3, 1884.
NAB: London, FO 97/434. Niger Expedition, 1862–4.
Nørregård, Georg. (1966). Danish Settlements in West Africa, 1658–1850, Trans., Sigurd
Mammen. Boston, MA: Boston University Press.
Ocansey, John E. and Kwame Arhin. (1989). An African Trading; or he Trials of William
Narh Ocansey of Addah, West Coast of Africa, River Volta. 3rd ed. Accra: Ghana
Academy of Arts and Sciences; First edition 1881, with Robert Hickson, Liverpool:
J. Looney.
Oyo, Akinfe. (1989). “he White Cap Chietaincy Institution of Lagos: A Historical
Survey.” BA hesis, Department of History, University of Ibadan.
Parker, John. (2000). Making the Town: Ga State Society in Early Colonial Accra.
Portsmouth, NH: Heinemann.
Phillips, Caryl. (2001). he Atlantic Sound. London: Vintage.
Pomeranz, Kenneth and Steven Topik. (2006). he World that Trade Created: Society,
Culture, and the World Economy 1400 to the Present. New York: M. E. Sharpe.
Popkin, Samuel L. (1981). “Public Choice and Rural Development – Free Riders,
Lemons, and Institutional Design.” In Public Choice and Rural Development,
edited by Cliford S. Russell and Norman K. Nicholson, 43–80. Washington, DC:
Resources for the Future, INC.
 Commerce, Credit, and Mobility 263

Sutton, Inez. (1984). “Law, Chietaincy and Conlict in Colonial Ghana: he Ada Case.”
African Afairs, 83(330): 41–62.
United Africa Company and Unilever Archives: Liverpool, UAC/2/33/AG/1/2/3, F. and
A. Swanzy Papers, Alfred Triggs to Andrew Swanzy, Cape Coast, March 29, 1870.
Van der Laan, H. L. (1975). he Lebanese Traders in Sierra Leone. he Hague: Mouton
Publishers.
 8

he Textile Industry of Eastern Africa

in the Longue Durée

William G. Clarence-Smith

8.1 INTRODUCTION
Historians are divided over the issue of the role of textiles in the economic
development of eastern Africa, deined here as stretching from the Sudan to
South Africa. Ravi Palat and Immanuel Wallerstein claim that India “dein-
dustrialized” its Indian Ocean periphery by exploiting advanced textile tech-
niques and cheap labor (Palat and Wallerstein 1999, 36–7). Similarly, John
Mack and Pedro Machado contend that Indian imports fatally undermined
a fragile textile sector in eastern Africa, long before Lancashire and New
England cottons looded in from the mid-nineteenth century (Machado
2005, 91–151, 2009, 167; Mack 1993, 303–4). Michael Pearson recognizes
that Indian cloth penetrated deep into the markets of early modern eastern
Africa, but he poses a crucial counterfactual question: Would the perfor-
mance of African textiles have been better or worse without imports from
South Asia? (Pearson 1998, 121–2).
he fact that textile production grew in South Asia’s oceanic periphery
from the mid-seventeenth to the mid-nineteenth century, including in
eastern Africa, argues against the position adopted by Palat, Wallerstein,
Mack, and Machado. Trade proved to have generally positive rather than
negative consequences. While locally produced cloth was only episodi-
cally competitive overseas, it had substantial internal markets in Africa.
South Asian artisans emigrated little, but they provided models to imi-
tate. Specialized trade diasporas, oten South Asian in composition,
supplied information, credit, shipping, commercial facilities, and inter-
mediate goods for further transformation, such as colored yarn and plain
cloth (Clarence-Smith 2009, 2011). hus eastern Africa was not a captive
market for South Asia, but rather a partner in the development of textile
production.

264
 he Textile Industry of Eastern Africa 265

hat said, eastern Africa’s textile industry remained far weaker than that
of South Asia, and there is absolutely no indication that any part of Africa
was heading for an industrial revolution on European lines (Parthasarathi
2011). here is some evidence for tendencies toward proto-industrializa-
tion, marked by the specialization of workers, the creation of workshops,
and market relations of a putting-out type. However, eastern Africa was
very sparsely populated, with a few exceptions, and neither the produc-
tivity of agriculture nor per capita incomes were favorable to widespread
proto-industry.
Colonial and postcolonial leaders usually failed to recognize the strengths
of this budding artisanal sector, and their policies ensured that the proto-
industrial path to economic development was even less likely to be fol-
lowed. All their energies eventually went into developing modern mills that
proved to be “white elephants.” Nevertheless, rather surprisingly, the arti-
sanal sector survived this misguided modernizing onslaught, and is at last
being given some oicial attention in our own times.

8.2 ARTISANS OF EARLY MODERN MADAGASCAR

European newcomers in the sixteenth century considered Malagasy textiles
to be the inest in the region (Grandidier 1928, 168; Rita-Ferreira 1999,
117–18). he great island displayed more diverse techniques, wove a greater
variety of textile ibers, and employed a wider palette of colors than any
other part of eastern Africa (Fee 2002; Mack 1987, 1989). Cotton cultiva-
tion dominated in the dry southwestern zone of Madagascar, while forested
lowland zones were chiely the domain of raia. Wild silk prevailed in many
places, and hemp, banana, spun bark, and plaited reeds played a minor role.
Fibers were sent from the coastlands to the densely peopled central plateau
for weaving, and cloth moved the other way (Campbell 2005, 31–2; Fee
2002, 40, map 2, 2005, 94).
Pedro Machado claims that South Asian competition was negatively
afecting Madagascar’s textile economy by the 1780s, citing sales of South
Asian cloth on the central plateau at that time (Machado 2009, 168).
However, the mere mention of markets for Indian textiles in the interior
does not mean that local production was being undermined. Yemen may
have ceased to purchase Malagasy cloth from the thirteenth century, but
East Africa remained a signiicant external market (Baldry 1982, 17; Kent
1970, 69; Newitt 1995, 28). Moreover, European vessels took local cloth
to the newly settled Mascarene Islands (Mauritius and Réunion) from the
mid-seventeenth century (Ellis 1838, I, 68; Fee 2005, 98; Larson 2000, 52).
266 Clarence-Smith

In any event, Madagascar’s internal market for its own cloth showed no
signs of weakening in the late eighteenth and early nineteenth centuries.
Contemporary texts and oral traditions depict a lourishing and vital textile
sector. A French traveler on the Malagasy central plateau in 1777 opined
that women were so busy weaving that they let domestic chores to men,
pointing to a growing division of labor. An 1826 report noted that local
cloth was competing well with the imported variety, and that every woman
in the country was busy weaving. Indeed, many women wove full time,
ranging “from the King’s wives to the slaves” (Fee 2005, 94; Larson 2000,
124, 128–9, 145). In 1838, one observer wrote about textiles as “the most
general employment of the people,” ater rice cultivation, and gave a detailed
description of spinning and weaving (Ellis 1838, I, 277, 282, 324–7).
Changing Malagasy attitudes toward rituals and politics boosted weav-
ing. Ater Radama II had been murdered, partly for ordering his subjects
to dress in Western clothes in 1861, it became mandatory to wear locally
woven lamba for ritual occasions (Fee 2002, 61, 90, note 29). Burial cere-
monies became more elaborate, involving greater quantities of cloth. In the
Merina and northern Betsileo lands, there was a second solemn burial of
dried remains of the deceased, which were wrapped in cloth. Only hand-
spun and handwoven textiles appear to have been acceptable for these pur-
poses (Bronson 2004, 166; Deschamps 1947, 121; Fee 2002, 62–3; Larson
2000, 128–9; Schaedler 1987, 428). Local cloth was also entrenched in the
giving of gits up and down the social hierarchy (Fee 2002, 64–7).

8.3 ARTISANAL TEXTILES IN EARLY MODERN

NORTHEASTERN AFRICA
Northeastern Africa had a narrower range of products than Madagascar,
almost entirely restricted to white cottons. However, these were of ine quality
and were produced in considerable and growing quantities. Moreover, value
was added by interweaving or embroidering with imported colored yarn.
he origins of cotton weaving were most ancient in Nubia, although cot-
ton cultivation itself probably only dated from the beginning of the second
millennium CE. Sudan’s Blue Nile regions of Sennar and Damar were noted
for their output of dammar, plain white cloth (Coquet 1993, 141–2; Lamb
2005, 107–8; Werbelof 1987, 155). By the mid-nineteenth century, dammar
was sold, given for marriage payments, paid as tribute, and used as cur-
rency (Ewald 1990, 72–3, 76, 87, 147–8). On the eve of colonial rule, women
did the spinning, with Darfur enjoying a particular reputation, while men
wove. he Jazirah (Gezira) region supplied dammar to Omdurman, where
 he Textile Industry of Eastern Africa 267

tailors turned it into the national costume. However, the best dammar was
imported from Ethiopia, while printed cottons came from the wider world
(Wingate 1986, 210, 380–2).
Ibn Battuta noted as early as the fourteenth century that Mogadishu’s
“unequalled” cotton textiles were exported “to Egypt and elsewhere” (Gibb
1962, 374). Although exports to Egypt had ceased by early modern times,
Somalia’s Benadir coast continued to send cotton wares to other parts of
eastern Africa into the twentieth century. Exports by sea were estimated at
three hundred sixty thousand to three hundred eighty thousand pieces a
year in the 1840s, and much cloth also went inland into southern Ethiopia
(Alpers 1983, 77–89; Reese 1996, 95–6). Mogadishu contained around a
thousand weaving households in the 1840s, with smaller numbers in other
Benadir towns, and raw cotton was imported from India. hese textiles
found their main market in southern Ethiopia, but were still regularly sent
by sea to Mombasa and north Somali ports (Guillain 1856–7, II, 531–2, 535
and III, 148–9, 172, 323).
Highland Christian Ethiopia produced similar white cottons for its own
densely settled territory (Pankhurst 1968, 257–60; Schaedler 1987, 93–9,
396–423). Gondar and Axum, in the north, were traditional centers of
production, together with Adowa, Ankobar, and Harar (Lamb 2005, 108;
Pankhurst 1968, 257). Reports of travelers from the sixteenth century sug-
gest that cotton cloth was slowly becoming the normal form of dress, grad-
ually replacing skins and hides. Men were said to have adopted cloth before
women, and “Semitic” peoples before others (Simoons 1960, 186–7).
Ethiopia’s plain white material was made into undergarments, shawls, and
turbans, whereas more colorful and elaborate stufs were typically imported
from South Asia (Burton 1894, II, 16–17; Wylde 1901, 246–8). Although
vegetable dyes were employed to a limited extent, Ethiopians commonly
unraveled imported Indian cloths to obtain colored wet, which they used to
decorate the borders of their cloth (Pankhurst 1968, 260). Europeans were
impressed by the “closeness, warmth and durability” of this Ethiopian cloth.
Male weavers did not usually grow crops, pointing to labor specialization
(Darkwah 1975, 145–6). Women spun, as well as embroidering plain cloth
(Pankhurst 1968, 32). By the 1840s, some six hundred women were spin-
ning in the imperial palace to supply court weavers (Darkwah 1975, 146).

8.4 ARTISANAL TEXTILES IN THE “SWAHILI WORLD”

To the south, the production of cotton textiles and a little silk was best
established in Swahili port cities, from Pate (modern Kenya) to Sofala
268 Clarence-Smith

(modern Mozambique) (Newitt 1987, 203, 206–7). Spindle whorls occur

in the archaeological record of Muslim coastal settlements from the tenth
century (Horton 1996, 337–41; Kjekshus 1977, 106). Portuguese docu-
ments suggest a growth spurt during the iteenth century, and that even
slaves wore a piece of locally woven blue or white cloth from waist to knees.
he Portuguese also noted that imported Gujarati stufs, especially blue
ones, were unraveled to obtain colored yarn, for example in Sofala and
Pate, although indigo dyeing was practiced in places (Newitt 2002, 31–3;
Prestholdt 1998, 24–5, 33).
he Kirimba (Querimba) islands, of the coast of northern Mozambique,
produced a milwani cloth that was dyed with local indigo, and cotton was
sometimes mixed with imported silk yarn. Milwani cloth was sold widely
in eastern Africa during the sixteenth century, and was one of the “cloths
of the land” that Portuguese traders eagerly sought for local commer-
cial operations. However, it was no longer mentioned in a report of 1634
(Newitt 1995, 189–92, 2002, 125, 127–30; Prestholdt 1998, 27–30). hat
said, its demise was probably due to raiding, provoking the migration of
artisans (Newitt, personal communication). As weaving cotton in the wider
Rovuma region continued, artisans simply seem to have moved out of the
vulnerable islands (Kjekshus 1977, 107). his episode does not, therefore,
prove the thesis of underdevelopment.
Ofsetting collapse in one area was the rise of new industries elsewhere.
hus Zanzibar only became a center of textile production ater the sultans
of Oman moved their main residence to the island in the early nineteenth
century (Guillain 1856–7, III, 342). Cotton weaving grew there from the
1850s, conducted in the open air. he wares of local weavers could be
inspected on the pavements of the town in the 1860s (Kjekshus 1977,
105–6).
Most East African output was machira, which was made into hammocks,
litters, blankets, cloaks, and belts (Davison and Harries 1980, 181–91, note
23). his was a coarse, unbleached cloth, a kind of cheap and durable home-
spun. Richard Burton criticized its “loose texture” and averred that “when
it is dry it is rough and unpleasant, when wet, heavy, comfortless as leather,
and it cannot look clean, as it is never bleached” (Burton 1860, II, 311, cited
in Kjekshus 1977, 108). However, machira was widely traded, notably in
exchange for gold dust on the Zimbabwean plateau (Alpers 1975, 24–5, 55;
Newitt 1995, 28, 66, 75, 78, 94, 141, 214, 232, 239). he cloth was also used
for currency, tax, bride price, and booty, and as an index of wealth and pres-
tige (Davison and Harries 1980, 187).
 he Textile Industry of Eastern Africa 269

Isolated tax igures suggest that the cost of machira roughly halved in the
eighteenth century, pointing to increasing productivity (Rita-Ferreira 1999,
118). Portuguese reports indicate that output was rising, and that the cloth
continued to ind a ready market on the Zimbabwean plateau (Bhila 1982,
122, 131).
Most striking, machira production survived deliberate Portuguese
attempts to sufocate it. From 1750, colonial oicials considered that local
weaving was reducing potential cloth imports from India, and thus nega-
tively afecting tax revenues (Isaacman 1972, 66, 73–5, 199, 201). In 1750,
the Junta do Comercio (Board of Trade) therefore suggested banning the
cultivation of cotton in the lower Zambezi valley. his was judged to be
unenforceable, as was an equally impractical scheme of 1753 to buy up
all available raw cotton to sell it in India and China (Lobato 1957, 241–
2; Machado 2005, 110). In the event, the Portuguese proved incapable of
stiling the production and sale of machira cloth (Mudenge 1988, 187).
Moreover, the mixed-race lords of the lower Zambezi’s prazo estates under-
mined the stance of Portuguese oicials, as they demanded machira cloth
as tribute from their African “serfs” (colonos) (Isaacman 1972, 66; Newitt
1995, 232, 239).
In any event, machira cloth production was still lourishing in the prazo
region in the mid-nineteenth century, when David Livingstone noted that
it was preferred to imports for many purposes (Alpers 1975, 25, 35, note
86). Indeed, imported yarn, initially from North America, allowed artisans
to weave “lighter and iner cloths.” By the 1890s, imported colored yarn
enabled the production of new intricately patterned designs, perhaps inlu-
enced by Malagasy traditions (Davison and Harries 1980, 181–3, 189; Mack
1993, 303–4).
Cotton in the interior of East Africa faced much competition from
bark, raia, leather, and skins (Alpers 1975, 16, 21; Beinart 1982, 24–5;
Cuypers 2004; Darish 1989, 117–40; Gillow 2003; Picton and Mack 1989,
33–43, 83–90, 131; Vansina 1998). Nevertheless, there was patchy spin-
ning and weaving of cotton, from the Great Lakes down to what are today
the northern Transvaal and Natal. Wild cotton frequently provided the
raw material, although there was some cultivation (Alpers 1975, 16, 21–2,
25; Davison and Harries 1980, 175–81, 187). Richard Burton reported
in 1860 that he found looms in every village, and another traveler wrote
that he found two to three looms in each village in 1891. Weavers in the
center of today’s Tanzania produced striped and checked materials in yel-
low and black. Around 1900, the Haya, to the west of Lake Victoria, had
270 Clarence-Smith

only just learned how to weave from their Nyamwezi neighbors, and were
experimenting with new types of cloth (Kjekshus 1977, 107–9).

8.5 EARLY MODERN TECHNOLOGY AND

PRODUCTION SYSTEMS
Textile technology remains the least researched aspect of the early mod-
ern story. he ot-repeated mantra that methods of production stagnated
in Africa is clearly incorrect, even if economic historians ind few or no
mentions of techniques in documents, and archaeologists lament the per-
ishable nature of wood. Technical change did not necessarily imply more
intricate designs or weaves, the interest of art historians, but rather faster
production, in greater quantities, more cheaply, and in more varied for-
mats. However, such techniques required some investment and training,
and the equipment generally became less portable.
Traders were one probable source of technical change. hey speciied
designs, provided market information, controlled quality, purchased in-
ished textiles, and shipped them to near and distant markets. hus Indian
and Portuguese traders commercialized machira together with Indian cot-
tons (Antunes 1992, 127; Newitt 1995, 28; Prestholdt 1998, 283, note 111).
Nyamwezi traders, from today’s central Tanzania, were active over a wide
area during the nineteenth century, spreading techniques of cotton weaving
to new regions (Kjekshus 1977, 108).
here were occasional examples of putting-out arrangements, and of
merchant-owned workshops. In sixteenth-century Madagascar, where
Portuguese and Arab merchants competed keenly for local textiles, cloths
were woven to order in individual homes in the north of the island (Kent
1970, 69; Newitt 1995, 28; Prestholdt 1998, 30). Wealthy Arab and Somali
merchants of the Benadir coast owned textile workshops, although they
were far from dominating production (Cassanelli 1988, 312).
Coerced labor was common in workshops. Slavery had advantages in
terms of training and retaining skilled labor, and Muslim scholars frowned
on other forms of compulsion (Clarence-Smith 2006). Slaves thus manned
the royal workshops of Tasi, in the Sudan, in 1839 (Ewald 1990, 87).
However, slavery interacted with other forms of labor. Most Somali
weavers in the 1840s were free, albeit of low caste, weaving full time and
buying their food on the market, and assisted by slaves, family members,
and clients (Alpers 1983, 81, 84; Guillain 1856–7, II, 531; Reese 1996, 96–8).
Unlike spinning, a prestigious occupation, weaving in Christian Ethiopia
was allocated to Muslims, Jews, and those of low caste (Levine 1974, 56–7;
 he Textile Industry of Eastern Africa 271

Pankhurst 1968, 32, 40–1, 257–9). Men wove on the African mainland,
whereas women did so in Madagascar, and women tended to dominate
spinning everywhere (Kjekshus 1977, 109; Levine 1974, 113, 153; Mack
1989, 21; Schaedler 1987, 396–423).
Spinning wheels speeded up and cheapened the production of yarn,
albeit at some cost in quality (Kuhn 1988). he evidence is sparse for east-
ern Africa, but the dynamism of Somalia’s Benadir workshops was due partly
to wheels, which produced four types of yarn in the 1840s. No date is given
for their adoption, and they were probably simple spindle-wheels, lacking
the crank, treadle, and lyer that enhanced the productivity of Saxon wheels
(Guillain 1856–7, II, 532). As for Ethiopia, East Africa and Madagascar, they
remained wedded to the hand spindle (Alpers 1975, 24–5; Grandidier 1928,
167; Kjekshus 1977, 106, 108; Pankhurst 1968, 258). In the Rukwa valley in
1899, in modern Tanzania, men spun with a “long spindled spool,” rubbed
between the palm of the hand and the thigh (Kjekshus 1977, 107–9). Christian
missionaries tried to introduce spinning wheels to Madagascar in 1822, but
the technique failed to take root ater an English artisan returned home (Ellis
1838, I, 324, 327). Ambitious water-driven machines proved ill adapted to
Malagasy conditions, and were rapidly abandoned (Fee 2002, 90, note 14).
Foot-operated treadles, liting and lowering two ixed heddles, freed the
weaver’s hands to work more rapidly, and pit looms itted with such treadles
appeared in the settled communities of the Middle East and South Asia
from around the beginning of the Common Era. Pits reduced the size of the
wooden frame containing the loom and kept the yarn moist, thus less likely
to break. As pit looms in northeastern Africa are generally of the Persian
rather than Egyptian type, they probably came from the east (Lamb 2005,
98–107, 111–56; Roth 1977). Pit looms were standard by the nineteenth
century in settled regions of northeastern Africa, spreading down into what
is today northern Kenya (Alpers 1983, 80; Gillow 2003, 160–1; Lamb 2005,
107; Pankhurst 1968, 259–60; Roth 1977, 62–3).
King Tewodros (heodore) of Ethiopia sought to improve looms, asking
Queen Victoria to send artisans. A Protestant missionary arrived in 1855
for this purpose, and Swiss artisans followed (Bates 1979, 30, 55, 68). A
photograph, taken in Gondar around 1900, depicts a large “tie-back” form
of pit loom, typical of greater Syria and Egypt, which allowed for the weav-
ing of longer and wider pieces (Lamb 2005, 107–10, 140, and plate 96). Such
improved pit looms were probably introduced by diasporic traders rather
than by European experts.
A kind of double-heddle loom was adopted in eastern Madagascar,
seemingly in the eighteenth century. he northeastern coast employed such
272 Clarence-Smith

looms by around 1800, and the technique spread some way down the coast.
Moreover, Merina weavers of the island’s central plateau displayed knowl-
edge of this technique of weaving at a later period, even though they no
longer practiced it. When the Betsimisaraka of the east coast employed a
single ixed heddle, they lashed it to the raters, as in the Persian Gulf (Mack
1987, 84–5, 1989, 21–2, 29–31). A youth trained in England ran a guild that
“wove cotton on mechanical looms” from 1826 on Madagascar’s plateau,
but this missionary initiative failed (Fee 2002, 88–90, notes 14–15).
East Africa, south of the Horn, as well as much of Madagascar, employed
the simple horizontal ground loom, adapted from that of Middle Eastern
Bedouins. With its ixed single heddle, it was quick and easy to disassemble
(Hecht 1989, 61–5; Lamb 2005, 1–6, 62). Ground looms were not necessarily
as rudimentary as they are sometimes depicted, however. When the White
Fathers sought to revive weaving in the Great Lakes region, one diiculty
they faced was that local looms were “very complicated and of a delicate
construction” (Kjekshus 1977, 80). Moreover, ground looms wove cloths
that were wider than those of West Africa (Davison and Harries 1980, 181–
7; Lamb 2005, 31–3; Roth 1977, 40). Possibly relecting the prevalence of
raiding, miniature looms appeared in the interior of Central Africa, raised
from the ground, and worked from the side (Lamb 2005, 32).
Other types of loom were occasionally found. Southern Madagascar had
Southeast Asian looms, with the warp tensed by the use of a backstrap it-
ted on a seated weaver (Ellis 1838, I, 325; Fee 2005, 94; Grandidier 1928,
167; Mack 1987, 84–6, 1989, 22–32). Vertical or oblique looms, with a sin-
gle heddle, existed in forested regions, generally conined to weaving raf-
ia (Schaedler 1987, 65; Vansina 1998, 267).hey may have been related to
Egypt’s vertical “Coptic loom,” warp-weighted and with one ixed heddle
(Lamb 2005, 61–3, 69, 72–3, 83, 86–8, 91).
Little colored cloth was produced outside Madagascar, but indigo dyeing
was practiced in places, perhaps relecting Arab traditions (Newitt 2002,
31–3; Prestholdt 1998, 24–5, 33). More signiicant, an anonymous author
noted in the mid-eighteenth century that the Malagasy were imitating cloth
from Bengal and the Coromandel Coast, suggesting incipient painting or
printing (Peers 2004, 145).

8.6 THE PERSISTENCE OF ARTISANAL TEXTILES IN THE

“AGE OF EMPIRE”
David Edgerton has argued persuasively that new technologies rarely simply
displace old ones. Rather, novel techniques lead to resistance and adaptation
 he Textile Industry of Eastern Africa 273

(Edgerton 2008). Tax policies led to the underreporting of Africa’s artisanal

output, while indirect rule and local rituals maintained a culturally pro-
tected niche for handmade textiles (Renne 1995, 128–43). Whereas hand
spinning was adversely afected by imports of factory-made yarn, weavers,
dyers, and tailors viewed imports of yarn, plain cloth, and chemical dyes
as a godsend. Moreover, new modes of transport extended local markets
(Johnson 1978; Munro 1984, 62–3). Intermediate goods for textile artisans
only came from the West for a brief period. Favored by the provisions of
the Congo Free Trade Zone set up in 1885, Indian yarn and cloth looded
into eastern African markets from the 1890s, followed by Japanese products
ater 1914 (Clarence-Smith 1989a, 176, 1989b).
New or reconstructed crats emerged during the colonial period (Munro
1984, 63). Tailoring was not entirely novel, but it changed out of all rec-
ognition in response to treadle-powered and electric sewing machines,
employed in households and workshops (Godley 2001). he embellishment
of cloth by hand also lourished, as embroidery and lace came to be traded
on a large scale around the globe (Clarence-Smith 2004, 449–51).
he Benadir coast of southern Somalia is the best-studied case of arti-
sanal production in the Age of Empire. From around the 1880s, imported
machine-made yarn, increasingly chemically dyed, became a signiicant
intermediate good for local weavers (Alpers 1983, 87–92). Somalis still
preferred locally woven to imported cloth in the 1890s, and weaving in
Mogadishu was “lourishing,” with substantial imports of yarn (Cattelani
1897, 67, 72–3, 88–9, 135; Robecchi-Bricchetti 1899, 618–20). South Asian
yarn quickly replaced the Western product. Aden received nearly 2.5 mil-
lion pounds of yarn from British India in 1894–5, much of which was redis-
tributed along the Somali coast (Great Britain 1896, 46, 56). Exports of
Benadir cloth continued, even as imports of European cloth grew (Chiesi
1909, 347–53).
Commercial disruptions caused by the First World War led to a renais-
sance in Somali weaving, with men taking up the crat in substantial num-
bers. India was the chief source of imported yarn, with Japan of growing
importance (Alpers 1983, 87–93). A report of 1920 noted that “a few hun-
dred” looms were active in Mogadishu alone, worked by men. Other towns
along the Benadir Coast also produced cloth, notably Merka, and there
were exports to Mombasa and Zanzibar. Furthermore, women could still
be seen spinning, despite imports of machine-made yarn (Stefanini 1922,
39). As late as the 1950s, more than one thousand households wove on the
Benadir coast. Local textiles had become prestige goods for special occa-
sions. Merchants provided yarn in a putting-out system, decried by Marxists
274 Clarence-Smith

as rank exploitation (Alpers 1983, 93–7). Somalia also exported artisanal

caps and shawls to neighboring countries (Administration Italienne 1954,
79–80).
From the late nineteenth century, India supplied yellow and blue yarns
for Ethiopian weavers, while grey, unbleached cloth came sequentially from
North America, India, and Japan (Pankhurst 1968, 261, 402–4, 407–8, 416).
he sector remained buoyant ater the devastation caused by the Italian
occupation of 1935–41 (Levine 1974, 113). In the early 1950s, women of
the northwest spun white cotton yarn with spindles, while colored yarn was
imported. Male indigenous Muslims worked in pit looms, but there were
also itinerant weavers, who carried a light loom on their shoulder (Simoons
1960, 187). In Jimma, in the southwest, plain white shawls for men and
women were the main product, with weaving relegated to men of low caste
(Lewis 2001, 53–4).
Sudanese weavers continued to produce plain white dammar. Indeed,
the introduction of new strains of cotton, aimed at boosting exports of raw
materials, may have had the unintended consequence of improving the
quality of yarn for local weavers (Issawi 1966, 479, 484). Local cloth held its
own despite rising imports of cotton textiles, which accounted for about a
quarter of the colony’s total import bill at the time of the First World War
(Great Britain 1920, 136, 148).
French oicials predicted the imminent demise of handweaving in
Madagascar ater the conquest of 1896, but some French people sought
to revive local crats. A school of applied arts, set up in 1925, preserved
cloths that were no longer fashionable, and contributed to the prestige of
handweaving (Fee 2002, 68–9, 74–5, 81). A revival of weaving during the
interwar years also served the purpose of stocking colonial exhibitions (Peers
1995, 47). Simultaneously, wearing local lamba became a sign of opposition
to Westernization (Fee 2002, 61). Malagasy weavers worked with imported
yarns in the 1920s, as well as the locally spun article (Kusimba, Odland, and
Bronson 2004, 34, introduction by Kusimba, citing Ralph Linton).
Various authors similarly proclaimed the death of handmade textiles in
East Africa (Davison and Harries 1980, 189–91; Pearson 1969, 124; Silver
1984, 44–5; Swainson 1980, 26). However, Helge Kjekshus notes that Rukwa
cotton textiles, in Tanganyika, “deied extinction largely through the patron-
age of the Christian missions.” Oicials also stimulated production to com-
bat shortages of cloth during the Second World War (Kjekshus 1977, 109
(quote), and 80, 106–9). Machira cloth production persisted in the lower
Zambezi valley, and photographic and descriptive evidence of weaving
across the Central African plateau was common until the First World War
 he Textile Industry of Eastern Africa 275

and even later (Davison and Harries 1980, 181, 185). Moreover, Christian
missionaries made considerable eforts to teach Western methods of spin-
ning, weaving, and sewing, for example in Southern Rhodesia (Zimbabwe)
(Jeater 2007, 105, 146, 158).
Indian mills became major suppliers of grey and bleached plain cotton
cloth to eastern Africa from the 1890s (Pirio 1982, 174–5). From the 1890s,
some of the plain cloth was locally printed, as South Asian merchants in
Swahili towns employed wooden blocks, on the Indian model, to print pop-
ular kanga. hese were two-piece rectangular wrappers for women, with
written mottos (Fair 2001, 68, 80; Hilger 1995, 44–5; Zawawi 2005, 16).
Printing blocks were increasingly consigned to museums ater 1918, as for-
eign imitations of kanga looded in (Fair 2001, 76–83). Nevertheless, old
wooden blocks still exist in smaller settlements such as Lamu, where printed
textiles have been produced until “quite recently” (Mack 1993, 304).
Embroidery also achieved a new prominence, with men, oten Islamic
scholars, embroidering skullcaps in coastal Swahili towns (Bhacker 1992,
134; Martin 1971, 538–41; Reese 1996, 94). However, these caps, “delicately
worked with ine stitching,” met severe competition from cheap imports in
Dar es Salaam by the 1950s (Leslie 1963, 110).
South Asian immigrants initially dominated workshops making garments
(Chauleur 1979, 111; Kilby 1975, 477). Some tailors had a background in
textiles, notably Hindus from the Khatri caste of weavers (Enthoven 1920–
2, II, 205; Morris 1968, 139–41). In 1952, ive hundred seventy-ive tailor-
ing units were counted in Tanganyika, scattered around the territory. here
were only forty-seven in 1961, but this relected changing criteria for statis-
tical purposes (Silver 1984, 44, 90, 94–5, 103).
he irst tailoring “factory” in South Africa was recorded in 1907
(Werbelof 1987, 139). Ashkenazi Jews from Lithuania exploited their skills
as tailors and their knowledge of local markets as peddlers to develop a
fragmented and small-scale garments industry, employing poor Afrikaner
farm girls driting into cities. hey beneited greatly from protectionist
South African import tarifs ater 1925 (Saron and Hotz 1955, 362–5). hey
then extended their activities into Central Africa (Cinammon 2004). he
Bulawayo Clothing Factory, which appeared in the records in 1920, may have
been the pioneer (Wolfgang Döpcke, personal communication). Exports to
neighboring territories were noted in 1930 (Mosley 1983, 209). In 1938,
seventeen small clothing enterprises represented 3.2 percent of Southern
Rhodesia’s industrial output by value, and their production reached 17 per-
cent of net imports of textiles and clothing (Mlambo, Pangeti, and Phimister
2000, 25; Southern Rhodesia 1946, 63–5). Ater 1945, modern industrial
276 Clarence-Smith

enterprises came to the fore in inishing South Asian intermediate imports

(Kilby 1975, 483).

8.7 MODERN COLONIAL MILLS

Modern textile factories faced colonial hesitation rather than outright hos-
tility (Kilby 1975). Metropolitan workers were perhaps those most con-
sistently opposed to “exporting jobs” to the colonies, and they allied with
small metropolitan textile entrepreneurs in declining regions, who feared
the loss of protected markets. Colonial oicials feared the social and politi-
cal consequences of proletarianization and urbanization. In contrast, com-
petitive Western textile irms were keen to cut costs by producing overseas.
Moreover, metropolitan bureaucrats at times backed manufacturing for
military reasons, or to appease nationalist leaders, and were increasingly
prepared to grant a measure of tarif protection (Clarence-Smith 1985;
Drummond 1974; Havinden and Meredith 1993; Marseille 1984).
Large modern integrated textile mills sufered from excessive protection,
as well as diseconomies of scale. Skilled labor and capital were exceedingly
costly in Africa, and even unskilled factory labor was expensive, as low
productivity generally more than ofset low wages. A supply of local raw
ibers was a mixed blessing, because of poor quality and irregular deliver-
ies. Entrepreneurs were chiely attracted to the business by cheap land and
hydroelectricity, and by high levels of tarif protection. Colonial regimes
frowned on diasporic entrepreneurs, notably South Asians, while showering
metropolitan irms with protectionist tarifs, quotas, licenses, and monop-
olies. hese were double-edged swords, for any initial spur to investment
came at the cost of structural ineiciency, most obviously relected in a lack
of exports beyond the region (Clarence-Smith 1989a; Kilby 1975).
he rising tide of Japanese imports into the Congo Free Trade Zone in
the interwar years acted as a precocious spur to the establishment of mod-
ern mills (Austen 1987, 136). In 1925, two Belgian textile producers set up
the Société Textile Africaine, or Texaf, in Léopoldville (Kinshasa), at the
head of the Congo River navigation system into eastern Africa (Heyse 1938,
7–12; Lacroix 1967, 19–20). In 1927, Texaf obtained a concession to gener-
ate its own hydroelectric power, setting up a subsidiary for this purpose in
1930 (Heyse 1936, 329–34). his was a state-of-the-art integrated spinning
and weaving mill. By 1930, the mill consumed about a ith of the raw cot-
ton grown in the Congo (Vellut 1985).
he impact of the early 1930s recession on Texaf was severe, but the
company survived (Heyse 1938, 7–12). Texaf became a holding company
 he Textile Industry of Eastern Africa 277

in 1934, and the textile business was renamed Société Usines Textiles de
Léopoldville, or Utexléo (Heyse 1938, 7–12). An unspeciied Belgian bank
acquired a controlling stake, and substituted labor for capital. Existing
machines were not replaced, and real wages were compressed (Strihou
1961, 51–2, 62, 65, 78). Output went from 930 tons in 1931 to 1,640 tons
in 1940, accounting for about a third of Congolese consumption (Belgium
1949, 576).
he Second World War inaugurated a period of rapid expansion in the
Congo. Utexléo took on more workers, and acquired some good spindles
and some indiferent secondhand looms in 1942–3. Production rose by
63.4 percent between 1940 and 1945, and the company imported foreign
cloth to meet excess demand (Strihou 1961, 52–4). Utexléo’s output reached
2,670 tons by 1945 (Belgium 1949, 576). Ater the war, Utexléo invested
more and improved the quality of its inished goods (Strihou 1961, 53–5).
Concurrently, Flemish industrialists established FILTISAF in Albertville
(Kalemie) in 1946 (Vellut 1985).
he 1925 Belgian initiative sent shock waves through the Portuguese col-
onies. Although only northern Angola and northern Mozambique lay in
the Congo Free Trade Zone, Léopoldville was the hub of a web of water and
rail transport links across the continent. Expressing fears that the products
of the Congolese factory would invade the Mozambican market, a leading
Portuguese entrepreneur requested permission in 1926 to establish a simi-
lar industry in Mozambique. However, his plea was rejected (Macedo 1939,
280). Indeed, Antonio Salazar’s conservative and authoritarian regime,
shaken by the Great Depression, passed a decree in 1932, which guaranteed
high ixed prices for colonial raw cotton in Portugal in return for formally
prohibiting the creation of textile mills in the colonies (Clarence-Smith
1985, 84–6, 163–5).
Although British East Africa lay entirely within the Congo Free Trade
Zone, and the proportion of textile imports from Japan rose to 70 percent
during the interwar years, Britain failed to support initiatives to set up local
textile mills. It was to no avail that imports had grown beyond a notional
threshold for import substitution, and that a free trade area covered Kenya,
Uganda, and Tanganyika from 1924. Nicola Swainson blames the Lancashire
lobby for this outcome (Swainson 1980, 26–7, 43–5, 124).1 In 1937, when an
interdepartmental committee examined the question, the Colonial Oice
poured cold water on arguments for a protected “infant industry.” Oicials

1
For the general British colonial position, see Havinden and Meredith 1993, 159, 168–74.
278 Clarence-Smith

insisted on levying an excise tax on local production that would be equal to

tarifs on imports (Drummond 1974, 441–2).
Shortages in the Second World War led to a British reappraisal. A
restrictive licensing system, introduced in 1943, applied to cotton yarn,
piece goods, and blankets. In 1959, a single textile mill had been autho-
rized for Kenya and Uganda respectively, although several licenses were
granted thereater (International Bank 1962, 275; Swainson 1980, 116–21,
125). By 1963, the combined textile, apparel, and footwear sector employed
3,701 people in Kenya, and a further 3,250 in Uganda (Kilby 1975, 477–8).
Uganda’s Nyanza Mills, or Nytil, built in Jinja in 1954, integrated spinning,
weaving, and inishing. It was built to consume cheap hydroelectricity from
the Owen Falls dam, as well as local raw cotton (Elkan 1961, 5–7, 59–60;
International Bank 1962, 275; Kilby 1975, 478; Pearson 1969, 125). A textile
mill was set up in hika, Kenya, toward the end of the 1950s, and exported
within British East Africa (Mosley 1983, 212, 222; Swainson 1980, 125, 128).
In 1952, Tanganyika had two dyeing irms employing eighty people and a
cotton-weaving shed with thirty workers. By 1961, there were ive weaving
mills in the Dar es Salaam area, all small and processing imported yarn
(Silver 1984, 74–5, 90–1, 94–5, 102–3. See also Werbelof 1987, 163–4).
he British generally backed metropolitan capitalists against South
Asian rivals, even though the latter had a proven historical track record
in manufacturing (Clarence-Smith 1989b). Indians repeatedly asked for
permission to manufacture textiles, for which much expertise existed in
Gujarat, and the only textile mill in Kenya in the 1950s was a small Indian
concern (Swainson 1980, 26–7, 125). Nevertheless, it was Calico Printers of
Manchester who developed the large Jinja mill in Uganda. Despite beneit-
ing from a local monopoly and cheap hydroelectricity, the irm lost so
much money that the Uganda Development Corporation was obliged to
buy the mill in 1957. Astonishingly, Calico Printers was retained as manag-
ing agent, turning a proit only because tarif protection was raised to a stif
30 percent (International Bank 1962, 275; Kilby 1975, 478; Pearson 1969,
125–6). However, in 1963, during the decade of independence, the South
Asian irm of Mehta was building a new textile mill, symbolically situated
on the opposite bank of the Nile from the Jinja plant (Pearson 1969, 127).
Economic dislocation in the Second World War impelled Salazar to set
up textile industries in Angola and Mozambique, mainly owned by met-
ropolitan capitalists (Clarence-Smith 1985, 164–5). In 1950, a large textile
factory was nearing completion in central Mozambique (Pitcher 1993,
163–4). Located at Vila Pery (Chimoio), on the railway inland from Beira,
it was close to the frontier with Southern Rhodesia (Zimbabwe), and about
 he Textile Industry of Eastern Africa 279

twenty miles from an abundant source of cheap hydroelectric power on the

Revue River. he skilled workers were immigrants from Portugal (Spence
1951, 83). In 1958, this factory consumed 1,774 tons of raw cotton for
14,500 spindles and 460 looms, and in 1960, it employed some 2,000 peo-
ple. It was reputed to be the largest cotton textile factory in sub-Saharan
Africa. In 1954, three textile establishments in Mozambique accounted
for more investment than any other economic sector, apart from sugar
(Bravo 1963, 82–5; Gersdorf 1958, 66–7). Artur Cupertino de Miranda’s
Banco Português do Atlântico, founded in 1942 on the basis of the remit-
tance business from Brazil, provided much of the capital for the Sociedade
Algodoeira de Fomento Colonial, with textile mills in both Mozambique
and Angola. Concession companies, which purchased compulsorily culti-
vated colonial raw cotton, were junior partners (Clarence-Smith 1985, 169;
Pitcher 1993, 163–5).
Some Western entrepreneurs functioned in a more diasporic mode.
Roberto Barattolo, an Italian trader in Eritrea before the Second World
War, was imprisoned by the British in 1941 and repatriated to Italy. He
returned to Asmara to set up the Cotonoicio Barattolo in 1950. From an
initial investment in spinning, he moved to weaving and inishing, and then
to clothing, while simultaneously going upstream into raw cotton and gin-
ning. By the time the Ethiopian Derg nationalized the enterprise in 1975,
this was perhaps the most fully integrated textile company in eastern Africa.
It had grown up gradually and organically, under the watchful eye of an
entrepreneur who took a leading position in the resident Italian commu-
nity.2 In contrast, Somalia’s Société de Manufactures Cotonnières Somalies,
located in the capital Mogadishu (Muqdisho) and providing about a sixth of
internal consumption in 1954, seems to have been owned by an Italian irm
(Administration Italienne 1954, 81–2).
Southern African settlers wielded a political power that generally stim-
ulated precocious industrial development, but textiles were an exception.
British Central Africa’s white miners and farmers opposed protection for
consumer goods that might force up the wage rate and thus damage their
export prospects. In addition, settler communities were determined to
frustrate the economic advance of South Asian communities, most likely
to prosper in the textile sector. Furthermore, Lancashire lobbied vigorously
against protective tarifs (Brett 1973, 266–81; Mlambo et al. 2000, 12–13).
At the imperial Ottawa Conference of 1932, the Southern Rhodesian
regime made speciic concessions on imports of British cotton piece goods

2
http://www.quirinale.it/elementi/DettaglioOnoroicenze.aspx?decorato=612.
280 Clarence-Smith

in return for British guarantees on Southern Rhodesian exports of mineral

products (Cole 1968, 29).
he Southern Rhodesian government thus only intervened to encour-
age the cotton textile industry from 1941, as the war afected supplies. he
authorities developed spinning at Gatooma from 1943, with 17,500 spindles
installed by 1948. David Whitehead and Sons, from Lancashire, set up the
irst large weaving mill in 1951, with production starting two years later.
A second spinning mill followed in 1952, with attached weaving and knit-
ting units. he government privatized the Gatooma mill by 1960, seeing
public ownership as a temporary aberration (Guillie 1995, 52; Kilby 1975,
483; Mlambo et al. 2000, 31). Exports were stimulated by free trade agree-
ments with South Africa and by the market provided by the Federation of
Central Africa from 1953 to 1963 (Mlambo et al. 2000, 37–8, 42–5; Mosley
1983, 212, 218–19). he textile and garments industries grew at an aver-
age annual rate of 20.8 percent from 1946 to 1953. By 1957–8, spinning
and weaving employed 6,574 people in twenty-six establishments. Up to a
third of Southern Rhodesia’s textile output was destined for South Africa in
the 1950s (Mlambo et al. 2000, 31, 45–7). By 1964, exports of textiles were
worth roughly half as much as imports (Mosley 1983, 218–19).
Although South Africa industrialized early and extensively, textiles were
not the leading sector. Secondary industries were a preserve for employing
poor whites, but they were particularly costly to employ in textiles (Cronje
1952, 24–5; Saron and Hotz 1955, 364–5). A Cape blanket factory was
founded in 1891, and woolen and cotton blankets were the country’s only
signiicant textile products for a long time. In 1933–4, there were twelve
textile establishments employing 2,358 workers. Once again, the real boom
came during and ater the Second World War, with high levels of tarif pro-
tection, so that by around 1960 there were 104 factories and 29,142 workers
(Beinart 1982, 24–5; Cronje 1952, 24; Du Toit 1978, 33, 37–8, 43). In 1967,
71,000 employees were recorded (Chauleur 1979, 113).
Fickle metropolitan capital came to the fore in this late spurt in South
Africa. Textile irms from Britain, France, and Italy lacked electricity and
other inputs at home to meet shortages due to the war in 1945, and relo-
cated to South Africa. he low productivity of African labor was a disincen-
tive, however, and investment fell drastically ater 1951, as postwar global
shortages turned to glut (Cronje 1952, 26–7; Kilby 1975, 483). Local Jewish
concerns in the garments business, moving upstream to produce cloth,
showed a more long-term commitment to the industry (Saron and Hotz
1955, 362–4).
 he Textile Industry of Eastern Africa 281

8.8 THE ARTICULATION OF MODERN COLONIAL FACTORY

PRODUCTION WITH ARTISANS
Modern textile factories in eastern Africa displayed little or no synergy with
small producers (Pearson 1969, 124). Both Western and settler industrial-
ists were unaccustomed to working with artisans. Indeed, many managers,
isolated in their gleaming new factories and cosseted by oicials, seem not
to have realized that there were textile artisans at all. Discrimination on
grounds of culture and race and weak or nonexistent representative institu-
tions made it all the more diicult to bridge the cultural gap. he strongest
linkages were in South Africa, where Jewish garment entrepreneurs and
workers were legally deemed white.
Overall, entrepreneurs in Africa deliberately turned their backs on
artisans, seeking out isolated hydroelectric sites where they could obtain
cheap power and control the entire production process from start to inish.
In particular, factories rarely supplied yarn to weavers, in stark contrast to
the situation in Asia. Tellingly, the International Bank for Reconstruction
and Development strongly criticized the Jinja mill in Uganda for not sell-
ing yarn or plain cloth within the colony. British managers seem to have
feared competition from those who might purchase semi-manufactures
from their mill (International Bank 1962, 275). hat said, Utexléo’s out-
put of yarn rose twice as fast as that of cloth in the 1950s, suggesting that
this Congolese enterprise may have been selling yarn to artisans (Strihou
1961, 55).
Clothing workshops had somewhat more positive links with modern
mills, for integrated factories rarely extended to producing garments. Small
workshops, making items of clothing mainly for the African market, thus
purchased cloth locally (Döpcke 1992, 104–5; Mlambo et al. 2000, 25; Wilde
1971, II, 7, 3–4).

8.9 THE IMPACT OF INDEPENDENCE ON MODERN MILLS

Independence from around 1960 was accompanied by policies of import
substitution, which were particularly ill suited to small extroverted econ-
omies (Chauleur 1979; Werbelof 1987). Allied policies of modernization
and state ownership made matters worse. Ineicient and corrupt indepen-
dent governments interfered incessantly, whether they owned or regulated
mills. Far from encouraging diasporic entrepreneurs, they sought to reduce
their share of manufacturing (Morris 1968, 145).
282 Clarence-Smith

From the 1960s to the 1980s, the general trend was thus for states to
encourage large, integrated, and highly protected textile mills to meet
the demands of restricted national markets. hought was rarely given to
exporting competitively within regions, let alone on the world market.
Such factories seemed to be quite successful for a while, but only at the
cost of ramping up protectionist tarifs, which already stood at an efective
average of 269 percent for textiles in Tanzania in 1966. Foreign exchange
shortages then throttled imports of inputs, fuel, and spare parts. Power
shortages multiplied, labor discipline collapsed, maintenance ceased, pro-
ductivity plummeted, location of new plants obeyed political criteria,
debts ballooned, capacity utilization fell, and smuggled imports of foreign
textiles soared. Private mills did better than state-owned ones, but only
marginally (Chauleur 1979, 103–10, 109–13; Valk 1996, 146–52, 216–42;
Werbelof 1987).
As the “white elephants” of modernization theory and Marxism became
increasingly discredited, export markets for textiles were created by multi-
iber agreements from 1973, and international agencies pushed strongly for
export substitution on Asian lines from the 1980s. However, international
agreements provided highly artiicial stimuli, liable to sudden and devastat-
ing changes as international politics shited. Foot-loose international entre-
preneurs were then all too quick to move on to new pastures (Valk 1996,
107–10, 128, 137). he ending of the World Trade Organization’s textile
and clothing quota system in 2005 was a particularly heavy blow for Africa
(Maliti 2005).
An alternative, albeit weak, reaction to the disastrous policies adopted
at independence was to foster neglected and long-sufering textile artisans.
his policy shit was aided by a gloriication of precolonial “authenticity.”
However, this all too easily degenerated into folklore, which was desultorily
marketed to sparse foreign tourists (Guillie 1995, 51; Polakof 1982, 4–10).
Ater the Derg toppled Haile Selassie in Ethiopia in 1974, the “command-
ing heights” of the textile industry, at that time the third largest in sub-
Saharan Africa, passed into the hands of the state-owned National Textile
Corporation. In 1982, this body controlled twenty-one mills and employed
some twenty-nine thousand people. Only some thirty small-scale knit-
ting workshops and two small spinning and weaving enterprises fell out-
side the state’s net. he sector sufered from poor and uneven productivity.
Some two-thirds of total output came from the Diredawa mill, an old mill
founded in 1943, but which was eiciently managed by a Japanese enter-
prise (Werbelof 1987, 93–5).
 he Textile Industry of Eastern Africa 283

A tale of two mills in Tanzania similarly indicates how state-owned

textile plant could come in very diferent forms. Following the Arusha
Declaration’s call for economic self-suiciency in 1967, the govern-
ment set up two factories. Mwanza Textiles, only partially state owned,
installed the most modern automated machinery that the West had to
ofer, at immense cost. he factory proceeded to lose money by the buck-
etful. In contrast, Friendship Mill, entirely state owned, was equipped
with antiquated secondhand Chinese machinery, some of which had been
designed in the 1930s. Employing far more workers, and thus soaking
up urban unemployment, it was quite consistently proitable. Indeed, the
Friendship Mill at times outperformed privately managed mills (James
1995, 35–6).
International sanctions constrained settler-ruled South Africa and
Southern Rhodesia (Zimbabwe), but they obtained capital from Israel and
Taiwan, and exported some of their output. By the mid-1980s, the Frame
Group was South Africa’s eighth largest company in terms of share capital,
employed some twenty-two thousand people, and was allegedly the largest
producer of blankets in the world. Not only did it spin, weave, and inish
cloth, but it also produced garments, which was quite a common pattern
in South Africa. he Tongaat Group, the country’s second largest textile
producer, even integrated upstream into cotton ginning (Werbelof 1987,
140–2, 180). In contrast, Zimbabwe’s textile sector declined rapidly ater
real independence had been achieved in 1980 (Valk 1996, 142–4; Werbelof
1987, 181–2). Cone Textiles, the largest employer, went into liquidation
in 1995, and David Whitehead, a Lonrho subsidiary, struggled to survive
(Guillie 1995, 52).
he factory sector turned to copying imported inished cloth specially
made for the African market, with Malagasy factories turning out brightly
colored kanga with proverbs in Malagasy (Fee 2002, 69). Producers in
Kenya, notably the integrated Rit Valley Textiles (Rivatex) of Eldoret, ini-
tially succeeded in ousting foreign kanga (Hilger 1995, 44). he same was
true in Tanzania, but prices then rose above those of imports (Zawawi 2005,
29). Maurer Textiles of Switzerland managed Sotexki in Kisangani, in the
eastern Congo, producing chitenge (kitenge), but political uncertainty and
Asian copies undermined the enterprise (Guillie 1995, 51). Zambian facto-
ries developed “imiwax” in the 1970s to replace foreign copies of wax-dyed
batik from Southeast Asia. However, by the 1990s, Mulungushi Textiles was
under such pressure from imports that it switched to exporting undyed
yarn and grey cloth (Guillie 1995, 51–2).
284 Clarence-Smith

Imports of cheap, or even free, secondhand clothes from the West became
a new and major threat to textile output in eastern Africa from the 1980s,
although such imports were generally underreported or even unreported
(Valk 1996, 116, note 25, 216). hey undercut most kinds of local business,
even if T-shirts, printed with topical local material, were able to compete
(Guillie 1995, 52; Hilger 1995, 45). Zambian enterprises, unable to com-
pete, sold their sewing machines for a song and witnessed the erosion of
skills (Matheson 2000).
he alternative strategy promoted by multilateral organizations, namely
export-substituting industrialization, remains to prove its worth. Italy’s
Zambaiti Group, an old textile company, bought Asmara’s Cotonoicio
Barattolo for the symbolic sum of one dollar in 2004. With experience of
producing in China, the irm invested E13 million to revive the company.
Integrated from spinning to clothing, Zambaiti Eritrea (Zaer) has 85 per-
cent of its output earmarked for export.3
However, the experience of Mauritius casts a shadow on this strategy.
he small island country experimented with Export Processing Zones
from 1970, in order to lessen dependence on sugar. Liberal tax, customs,
capital transfer, and labor regulations attracted foreign textile companies.
hey also obtained cheap water, electricity, and rents, as well as guarantees
against nationalization. he policy was intensiied from 1982, leading to
a boom in the production of cloth and garments for export to the West,
with about one hundred thousand workers employed by the mid-1990s.
Unfortunately, this artiicially generated sector could not stand up to grow-
ing Asian competition and global trade liberalization, and was in a serious
crisis by 2003, which shows no signs of abating (Neveling 2010, 76–82; Valk
1996, 138–41).

8.10 THE PERSISTENCE OF ARTISANAL TEXTILES

IN THE ERA OF INDEPENDENCE
Despite this oicial love-in with large integrated modern textile mills, the
neglected and even despised artisans of eastern Africa survived (Valk 1996,
165). hey also had to overcome major new hurdles, such as shortages of
imported yarn, due to protectionism and foreign exchange crises, and out-
ages in electricity supply (Renne 1995, 146). South Asian preponderance in
tailoring, for example in Uganda, engendered additional oicial hostility
(Wilde 1971, XII, 2, 6).

3
http://www.eritreaeritrea.com/zambaiti_eritrea_plc.htm.
 he Textile Industry of Eastern Africa 285

he Horn of Africa retained the healthiest artisanal sector. Ethiopia’s

weavers kept producing through the Marxist rule of the Derg ater 1974,
although makers of garments were forced into inefective cooperatives
(Werbelof 1987, 94, 96). By the 1990s, weaving was in full swing again,
with traditional double-heddle pit looms. Indeed, Ethiopia was declared
an exception to the overall decline of artisanal textiles in eastern Africa
(Gillow 2003, 157–60, 164).
Sudan was unusual in formally acknowledging the existence of an arti-
sanal textile sector ater independence in 1956. In addition to a program
of modern mills, the government initiated weaving sheds containing up to
100 looms. In 1970, it was estimated that “handicrat producers” met about
half the domestic demand for textiles, and the sector remained dynamic in
the 1990s (Valk 1996, 103; Werbelof 1987, 156). Tailoring was of further
importance, oten in the hands of West African immigrants (Yamba 1995,
98–9, 112, 140).
Tanzania’s brand of humanistic socialism favored artisans, at least on
paper. he National Small Industries Corporation was set up in 1967, pro-
viding subsidized rents for tailoring and other workshops, albeit with poor
facilities and a reluctance to eject defaulting tenants (Wilde 1971, X, 3–7).
By the 1980s, there was a conscious policy, largely borrowed from India, of
developing small-scale textile manufacturing. On the one hand, there was
to be a “decentralized” sector, typically consisting of “semi-urban” units,
with knitting machines and ten to forty semiautomatic looms. On the other
hand, there was to be a “manual” sector in rural areas. hree “handloom
manufacturing units” were charged with distributing looms, of an unspeci-
ied type. Tanzania preached this model to fellow members of SADCC, but
with indiferent success (Werbelof 1987, 164, 166). Although it is unclear
how these policies panned out on the ground, Gillow seems excessively pes-
simistic in asserting that handloom weaving has become merely a matter of
“historical interest” (Gillow 2003, 158).
Ater an initial stress on modernization, Madagascar tipped the balance
back toward supporting ancient traditions from 1992. Handweaving was
concentrated in specialized quarters, small towns, and villages, especially
in the highlands and the south. Weavers exported their wares to other
parts of the island through middlemen, while regular fashion shows and
sales to tourists helped to gain markets in the West. Output ranged from
resurrected noble cloths to cheap loose weaves, the latter competing with
imports to meet the needs of poor households. Men wove more oten, at
least in the capital city, breaking with the ancient gender division of labor.
Tailors also worked to express ethnic identities in cloth. However, attempts
286 Clarence-Smith

to use handweaving to relieve extreme rural poverty failed, as did coopera-

tives, for some capital was needed for weaving, and iercely independent
households had little desire to cooperate (Fee 2002, 67–70, 74–86. See also
Peers 1995, 46–7).
he revival of artisanal eforts in Madagascar was accompanied by tech-
nical adaptations, even if Valk criticized “outdated technologies” (Valk
1996, 100). Double-heddle treadle looms still existed in the northeast in the
1980s, and there were eforts to spread the technique, build larger looms,
and introduce “European looms” (Fee 2002, 78; Gillow 2003, 214; Mack
1989, 31–2; Peers 1995, 47). he most expensive cloths were made with
hand-spun yarn on old ground looms, and were colored with natural dyes.
here were experiments with new natural ibers, for example a sisal warp.
However, most artisans employed chemical dyes and factory-made yarn,
including shiny imported synthetic ibers for cheaper wares (Fee 2002,
74–9, 85–6).
Davison and Harries consider that weaving and spinning skills were
“almost forgotten” by the 1950s in southeastern Africa, although their own
fascinating evidence clearly does not support this pessimistic conclusion.
hus, Shona and Chewa weavers of the 1970s, in Zimbabwe and Malawi
respectively, wove belts for ritual purposes, such as spirit possession cults
(Davison and Harries 1980, 175, 181, 185, 187–91). Similarly, weavers in
Barue were producing machira as late as the 1970s (Davison and Harries
1980, 181–2; Lobato 1957, 242). As Africans gradually moved into textiles
in South Africa, they embellished machine-made cloth with Ndebele and
Zulu beadwork (Guillie 1995, 54).

8.11 CONCLUSION
Textiles formed a sector of modest dimensions in eastern Africa, and yet
production persisted throughout the early modern and modern periods.
Handicrat techniques progressed and proto-industrial structures emerged,
although far too little research has been carried out on the artisanal sec-
tor. Modern mills appeared in the twentieth century, but were generally
too capital-intensive, overly protected, and poorly articulated with artisanal
production. hat said, the mills of well-rooted diasporas, notably South
Asian and Jewish, fared better than those of Western capitalists, white set-
tlers, or independent governments.
his story gives the lie to extreme theories of underdevelopment, in
which Africa was reduced to a source of raw materials and a captive
 he Textile Industry of Eastern Africa 287

market for foreign manufactures. In reality, South Asian textiles failed to

deindustrialize early modern eastern Africa, probably because of rising
African productivity. he model proposed here is that merchants reduced
costs and uncertainties and spread best technical practice. Western and
Asian textile industries more strongly afected the region from the late
nineteenth century, but again without fulilling Dependentista predic-
tions. he existing artisanal sector did not die, not even hand spin-
ning, and new workshops sprang up producing garments with sewing
machines.
hat said, the consequences of creating a disarticulated cotton textile
sector were negative. Large mills had few or no moorings in civil society
and were incapable of stimulating artisanal production. Once colonial or
settler rule ended, such enterprises either lost the crutches of state sup-
port or became “white elephants,” draining resources from the wider econ-
omy. his was an important factor holding back the emergence of “African
tigers.” Indeed, the bitter irony of independence has been that nationalists
have been passionate about achieving industrialization, and yet have sys-
tematically killed the geese that laid the industrial eggs. For the historian,
however, the era since independence is little more than the blink of an eye-
lid. he resilience of eastern Africa’s artisanal textile sector in the past is a
good augury for its prosperity in the future, and it should certainly act as a
spur for better studying its history.

References
Administration Italienne de Tutelle de la Somalie. (1954). Plans de développement
économique de la Somalie, années 1954–1960. Rome: Istituto Poligraico dello
Stato.
Alpers, Edward A. (1975). Ivory and Slaves in East Central Africa. London: Heinemann.
(1983). “Futa Benaadir: Continuity and Change in the Traditional Cotton Textile
Industry of Southern Somalia, c.1840–1980.” In Entreprises et Entrepreneurs en
Afrique, XIXe et XXe siècles I: 77–98. Paris: L’Harmattan.
Antunes, Luis Frederico Dias. (1992). “A Actividade da Companhia de Comercio
Baneanes de Diu em Moçambique, 1688–1777,” MA hesis, Universidade Nova
de Lisboa.
Austen, Ralph. (1987). African Economic History, Internal Development and External
Dependency. London: James Currey.
Baldry, John. (1982). Textiles in Yemen: Historical References to Trade and Commerce in
Textiles in Yemen, from Antiquity fo Modern Times. London: British Museum.
Bates, Darrell. (1979). he Abyssinian Diiculty: he Emperor heodorus and the Magdala
Campaign, 1867–68. Oxford: Oxford University Press.
288 Clarence-Smith

Beinart, William. (1982). he Political Economy of Pondoland, 1860–1930. Cambridge:

Cambridge University Press.
Belgium (1949). Ministère des Colonies, Plan Décennal Pour Le Développement
Économique et Social du Congo Belge, Brussels: De Visscher.
Bhacker, M. Reda. (1992). Trade and Empire in Muscat and Zanzibar: Roots of British
Domination. London: Routledge.
Bhila, H. H. K. (1982). Trade and Politics in a Shona Kingdom: he Manyika and heir
African and Portuguese Neighbours, 1575–1902. London: Longman.
Bravo, Nelson Saraiva. (1963). A Cultura Algodoeira Na Economia do Norte de
Moçambique. Lisbon: Junta de Investigações do Ultramar.
Brett, E. A. (1973). Colonialism and Underdevelopment in East Africa, 1919–1939.
London: Heinemann.
Bronson, Bennet. (2004). “Wrapping Up Malagasy Textiles.” In Unwrapping the Textile
Traditions of Madagascar, edited by Chapurukha M. Kusimba, J. Claire Odland,
and Bennet Bronson, 164–9. Los Angeles, CA: Fowler Museum.
Burton, R. F. (1860). he Lake Regions of Central Africa. London: Longman, Green,
Longman, and Roberts.
(1894). First Footsteps in East Africa or an Exploration of Harar. London: Tylston and
Edwards, 2nd ed.
Campbell, Gwyn. (2005). An Economic History of Imperial Madagascar, 1750–1895: he
Rise and Fall of an Island Empire. Cambridge: Cambridge University Press.
Cassanelli, Lee V. (1988). “he Ending of Slavery in Italian Somalia: Liberty and the
Control of Labor, 1890–1935.” In he End of Slavery in Africa, edited by Suzanne
Miers and Richard Roberts, 308–31. Madison: University of Wisconsin Press.
Cattelani, G. (1897). L’avvenire Coloniale D’Italia Nel Benadir. Naples: F. Giannini e
Figli.
Chauleur, Pierre. (1979). L’Afrique Industrielle. Paris: G. P. Maisonneuve et Larose.
Chiesi, Gustavo. (1909). La Colonizzazione Europea Nell’ Est Africa: Italia, Inghilterra,
Germania. Turin: Unione Tipograico-Editrice Torinese.
Cinnaman. John. (2004). “he Jews of Manicaland,” http://www.zjc.org.il/showpage
.php?pageid=155.
Clarence-Smith, William G. (1985). he hird Portuguese Empire, 1825–1975: A Study in
Economic Imperialism. Manchester: Manchester University Press.
(1989a). “he Efects of the Great Depression of the 1930s on Industrialisation in
Equatorial and Central Africa.” In he Economies of Africa and Asia in the Inter-
War Depression, edited by Ian Brown, 170–202. London: Routledge.
(1989b). “Indian Business Communities in the Western Indian Ocean in the
Nineteenth Century.” he Indian Ocean Review, 2(4): 18–21.
(2004). “Middle Eastern Migrants in the Philippines: Entrepreneurs and Cultural
Brokers.” Asian Journal of Social Science, 32(3): 425–57.
(2006). Islam and the Abolition of Slavery. London: Hurst & Co.
(2009). “he Production of Cotton Textiles in Early Modern South-East Asia.” In he
Spinning World: A Global History of Cotton Textiles, 1200–1850, edited by Giorgio
Riello and Prasannan Parthasarathi, 127–42. Oxford: Oxford University Press.
(2011). “he Expansion of Cotton Textile Production in the Western Indian Ocean,
c1500–c1850.” In Indian Ocean Worlds: Essays in Honour of Prof. K. N. Chaudhuri,
 he Textile Industry of Eastern Africa 289

edited by Stefan Halikowski-Smith, 84–106. Newcastle-upon-Tyne: Cambridge

Scholars Publishing Ltd.
Cole, R. L. (1968). “he Tarif Policy of Rhodesia, 1899–1963.” Rhodesian Journal of
Economics, 2(2): 28–47.
Coquet, Michèle. (1993). Textiles Africains. Paris: Adam Biro.
Cronje, F. J. C. (1952). “he Textile Industry in the Union of South Africa.” South African
Journal of Economics, 20(1): 23–30.
Cuypers. Jan-B. (2004). L’habillement du Rwanda Ancient. Tervuren: Musée Royal de
l’Afrique Centrale.
Darish, Patricia. (1989). “Dressing for the Next Life: Raia Textile Production and Use
among the Kuba of Zaïre.” In Cloth and Human Experience, edited by Annette B.
Weiner and Jane Schneider, 117–40. Washington, DC: Smithsonian Institution
Press.
Darkwah, R. H. Koi. (1975). Shewa, Menelik, and the Ethiopian Empire, 1813–1889.
London: Heinemann.
Davison, Patricia and Patrick Harries. (1980). “Cotton Weaving in South-East Africa: Its
History and Technology.” Textile History, 11: 175–92.
Deschamps, Hubert. (1947). Madagascar. Paris: Berger Levrault.
Döpcke, Wolfgang. (1992). Das Koloniale Zimbabwe in der Krise; Eeine Wirtschats- und
Sozialgeschichte, 1929–1939. Hamburg: Lit.
Drummond, Ian. (1974). Imperial Economic Policy, 1917–1939: Studies in Expansion and
Protection. London: George Allen and Unwin Ltd.
Du Toit, Bettie. (1978). Ukubamba Amadolo: Workers’ Struggles in the South African
Textile Industry. London: Onyx Press.
Edgerton, David. (2008). he Shock of the Old: Technology and Global History since 1900.
London: Proile.
Elkan, Walter. (1961). he Economic Development of Uganda. London: Oxford University
Press.
Ellis, William. (1838). History of Madagascar. London: Fisher, Son & Co.
Enthoven, Reginald E. (1920–2). he Tribes and Castes of Bombay. Bombay : Government
Central Press.
Ewald, Janet. (1990). Soldiers, Traders and Slaves: State Formation and Economic
Transformation in the Greater Nile Valley, 1700–1885. Madison: University of
Wisconsin Press.
Fair, Laura J. (2001). Pastimes and Politics: Culture, Community and Identity in Post-
Abolition Urban Zanzibar, 1890–1945. Oxford: James Currey.
Fee, Sarah. (2002). “Madagascar’s Textiles through History.” In Objects as Envoys: Cloth,
Imagery and Diplomacy in Madagascar, edited by Christine M. Kreamer and Sarah
Fee, 33–93. Washington, DC: Smithsonian Institution.
(2005). “Ze Mañe Aze: Looking for Patterns in Malagasy Cloth.” In Textiles in Indian
Ocean Societies, edited by Ruth Barnes, 85–109. London: RoutledgeCurzon.
Gersdorf, Ralph von. (1958). Moçambique. Bonn: Kurt Schroeder.
Gibb, H. A. R. (1962). he Travels of Ibn Battuta, A.D. 1325–1354, Volume 2. Cambridge:
Cambridge University Press.
Gillow, John. (2003). African Textiles: Colour and Creativity across a Continent. London:
hames and Hudson.
290 Clarence-Smith

Godley, Andrew. (2001). “he Global Difusion of the Sewing Machine, 1850–1914.”
Research in Economic History, 20: 1–45.
Grandidier, Alfred and Guillaume Grandidier. (1928). Histoire Physique, Naturelle et
Politique de Madagascar, volume IV, Ethnographie de Madagascar, Tome quatrième.
Paris: Hachette.
Great Britain. (1896). “Statement of the Trade of British India with British Possessions
and Foreign Countries for the Five Years 1890–1 to 1894–5.” Parliamentary Papers,
1896, Accounts and Papers, C. 7997.
Great Britain, Foreign Oice. (1920). Anglo-Egyptian Sudan. London: HMSO.
Guillain, Ch. (1856–7) Documents sur L’histoire, La Géographie et Le Commerce de
l’Afrique Orientale. Paris: A. Bertrand.
Guillie, Jackie. (1995). “Southern African Textiles Today: Design, Industry, and
Collective Enterprise.” In he Art of African Textiles: Technology, Tradition, and
Lurex, edited by John Picton, 51–4. London: Barbican Art Gallery.
Havinden, Michael and David Meredith. (1993). Colonialism and Development: Britain
and Its Tropical Colonies, 1850–1960. London: Routledge.
Hecht, Ann. (1989). he Art of the Loom: Weaving, Spinning and Dyeing across the World.
London: British Museum Publications.
Heyse, T. (1936). “Cessions et Concessions Foncières du Congo.” Congo, 2: 329–38.
(1938). “Cessions et Concessions Foncières du Congo.” Congo, 1: 7–12.
Hilger, Julia. (1995). “he Kanga: An Example of East African Textile Design.” In he Art
of African Textiles: Technology, Tradition, and Lurex, edited by John Picton, 44–5.
London: Barbican Art Gallery.
Horton, Mark. (1996). Shanga: he Archaeology of a Muslim Trading Community on the
Coast of East Africa. London: British Institute in Eastern Africa.
International Bank for Reconstruction and Development. (1962). he Economic
Development of Uganda. Baltimore, MD: Johns Hopkins University Press.
Isaacman, Allen. (1972). Mozambique: he Africanisation of a European Institution: he
Zambezi Prazos 1750–1902. Madison: University of Wisconsin Press.
Issawi, Charles. (1966). he Economic History of he Middle East, 1800–1914: A Book of
Readings. Chicago: University of Chicago Press.
James, Jefrey. (1995). he State, Technology and Industrialization in Africa. New York:
St. Martin’s Press.
Jeater, Diana. (2007). Law, Language and Science: he Invention of the “Native Mind” in
Southern Rhodesia, 1890–1930. Portsmouth, NH: Heinemann.
Johnson, Marion. (1978). “Technology, Competition and African Crats.” In he Imperial
Impact: Studies in the Economic History of Africa and India, edited by Clive Dewey
and A. G. Hopkins, 259–69. London: Athlone Press.
Kent, Raymond K. (1970). Early Kingdoms in Madagascar, 1500–1700. New York: Holt,
Rinehart and Wilson.
Kilby, Peter. (1975). “Manufacturing in Colonial Africa.” In Colonialism in Africa,
Volume 4, he Economics of Colonialism, edited by Peter Duignan and L. H. Gann,
470–520. Cambridge: Cambridge University Press.
Kjekshus, Helge. (1977). Ecology, Control and Economic Development in East African
History: he Case of Tanganyika 1850–1950. Berkeley : University of California
Press.
 he Textile Industry of Eastern Africa 291

Kuhn, Dieter. (1988). Science and Civilisation in China, Volume 5, Part IX: Textile
Technology: Spinning and Reeling. Cambridge: Cambridge University Press
(section 31).
Kusimba, Chapurukha M. J., Claire Odland, and Bennet Bronson, Eds. (2004).
Unwrapping the Textile Traditions of Madagascar. Los Angeles, CA: Fowler
Museum.
Lacroix, Jean-Louis. (1967). Industrialisation au Congo: La Transformation des Structures
Économiques. Paris: Mouton.
Lamb, Venice. (2005). Looms Past and Present: Around the Mediterranean and Elsewhere.
Hertingfordbury : Roxford Books.
Landen, Robert G. (1967). Oman since 1856. Princeton, NJ: Princeton University Press.
Larson, Pier. (2000). Becoming Merina in Highland Madagascar: History and Memory in
the Age of Enslavement. Oxford: James Currey.
Leslie, J. A. K. (1963). A Survey of Dar es Salaam. London: Oxford University Press.
Levine, Donald L. (1974). Greater Ethiopia: he Evolution of a Multiethnic Society.
Chicago: Chicago University Press.
Lewis, Herbert S. (2001). Jimma Abba Jifar: An Oromo Monarchy, Ethiopia, 1830–1932.
Lawrenceville, NJ: he Red Sea Press, 2nd ed.
Lobato, Alexandre. (1957). Evolução Administrativa e Económica de Moçambique, 1752–
1763. Lisbon: Agência Geral Ultramarina.
Macedo, Alberto Cardoso Martins de. (1939). “Relações Entre a Indústria Nacional e a
Agricultura Colonial.” Boletim da Sociedade de Geograia de Lisboa, 57: 267–90.
Machado, Pedro. (2005). “Gujarati Indian Merchant Networks in Mozambique, 1777-c.
1830.” PhD hesis, University of London.
(2009). “Awash in a Sea of Cloth: Gujarat, Africa and the Western Indian Ocean,
1300–1800.” In he Spinning World: A Global History of Cotton Textiles, 1200–1850,
edited by Giorgio Riello and Prasannan Parthasarathi, 161–79. Oxford: Oxford
University Press.
Mack, John. (1987). “Weaving, Women and the Ancestors in Madagascar.” Indonesia
Circle, 42: 76–91.
(1989). Malagasy Textiles. London: Shire Publications.
(1993). “Sub-Saharan Africa and the Ofshore Islands.” In 5000 Years of Textiles, edited
by Jennifer Harris, 295–305. London: British Museum Press.
Maliti, Tom. (2005). “Rise and Fall of African Textile Industry Provides Cautionary
Tale.” http://www.busrep.co.za/general/print_article.php?fArticleId+2447442&Se
ctionId+552&fSetId=662 (consulted July 5, 2010).
Marseille, Jacques. (1984). Empire Colonial et Capitalisme Français, Histoire d’un Divorce.
Paris: Albin Michel.
Martin, Bradford G. (1971). “Notes On Some Members of the Learned Classes of
Zanzibar and East Africa in the Nineteenth Century.” African Historical Studies,
4(3) 525–45.
Matheson, Ishbel. (2000). “Charity Killing Zambia’s Textile Industry.” BBC News Online.
December 25.
Mlambo, A. S., E. S. Pangeti, and Ian Phimister. (2000). Zimbabwe, a History of
Manufacturing, 1890–1995. Harare: University of Zimbabwe Publications.
Morris, H. S. (1968). he Indians in Uganda. London: Weidenfeld and Nicolson.
292 Clarence-Smith

Mosley, Paul. (1983). he Settler Economies: Studies in the Economic History of Kenya
and Southern Rhodesia, 1900–1963. Cambridge: Cambridge University Press.
Mudenge, Stanley I. G. (1988). A Political History of Munhunmutapa, c. 1400 to 1902.
Harare: Zimbabwe Publishing House.
Munro, J. Forbes. (1984). Britain in Tropical Africa, 1880–1960, Economic Relationships
and Impact. London: Macmillan.
Neveling, Patrick. (2010). “Von Nutzen der Geschichte, vom Wissen der Akteure
und vom Nachteil der Multi-Sited Ethnography: Welthandel, Wirtschatskrise
und Standortwettbewerb in Mauritius Anfang des 21. Jahrhunderts.” Sociologus:
Zeitschrit für empirische Ethnosociologie und Ethnopsychologie, 60(1): 71–97.
Newitt, Malyn. (1987). “East Africa and Indian Ocean Trade.” In India and the Indian
Ocean, 1500–1800, edited by Ashin Das Gupta and M. N. Pearson, 201–33. Calcutta:
Oxford University Press.
(1995). A History of Mozambique. London: Hurst.
(2002). East Africa: Portuguese Encounters in the World in the Age of Discoveries.
Aldershot: Ashgate.
Palat, Ravi A. and Immanuel Wallerstein. (1999). “Of What World-System Was Pre-
1500 ‘India’ A Part?” In Merchants, Companies and Trade: Europe and Asia in
the Early Modern Era, edited by Sushil Chaudhury and Michel Morineau, 21–41.
Cambridge: Cambridge University Press.
Pankhurst, Richard. (1968). Economic History of Ethiopia, 1800–1935. Addis Ababa:
Haile Selassie I University Press.
Parthasarathi, Prasannan. (2011). Why Europe Grew Rich and Asia Did Not: Global
Economic Divergence, 1600–1850. Cambridge: Cambridge University Press.
Pearson, D. S. (1969). Industrial Development in East Africa. Nairobi: Oxford University
Press.
Pearson, M. N. (1998). Port Cities and Intruders: he Swahili Coast, India and Portugal in
the Early Modern Era. Baltimore, MD: Johns Hopkins University Press.
Peers, Simon. (1995). “Weaving in Madagascar.” In he Art of African Textiles: Technology,
Tradition, and Lurex, edited by John Picton, 46–7. London: Barbican Art Gallery.
(2004). “History and Change in the Weaving of Highland Madagascar.” In Unwrapping
the Textile Traditions of Madagascar, edited by Chapurukha M. Kusimba, J. Claire
Odland, and Bennet Bronson, 142–53. Los Angeles, CA: Fowler Museum.
Picton, John and John Mack. (1989). African Textiles. London: British Museum,
2nd ed.
Pirio, Gregory R. (1982). “Commerce, Industry and Empire: he Making of Modern
Portuguese Colonialism in Angola and Mozambique, 1890–1914.” PhD hesis,
University of California at Los Angeles.
Pitcher, M. Anne. (1993). Politics in the Portuguese Empire: he State, Industry and
Cotton, 1926–1974. Oxford: Clarendon Press.
Polakof, Claire. (1982). African Textiles and Dyeing Techniques. London: Routledge and
Kegan Paul.
Prestholdt, Jeremy G. (1998). As Artistry Permits and Custom May Ordain: he Social
Fabric of Material Consumption in the Swahili World, circa 1450 to 1600. Evanston,
IL: Northwestern University, PAS Working Paper No. 3.
 he Textile Industry of Eastern Africa 293

Reese, Scott S. (1996). “Patricians of the Benaadir; Islamic Learning, Commerce, and
Somali Urban Identity in the Nineteenth Century.” PhD hesis, University of
Pennsylvania.
Renne, Elisha P. (1995). Cloth hat Does Not Die: he Meaning of Cloth in Bùnú Social
Life. Seattle: University of Washington Press.
Rita-Ferreira, António. (1999). African Kingdoms and Alien Settlement in Central
Mozambique, c. Fiteenth to Seventeenth Century. Coimbra: Universidade de
Coimbra.
Robecchi-Bricchetti, Luigi. (1899). Somalia e Benadir; Viaggio de Esplorazione Nell’Africa
Orientale. Milan: Carlo Aliprandi.
Roberts, Edmund. (1972). Embassy to the Eastern Courts of Cochin-China, Siam and
Muscat in the US Sloop-of-War Peacock, during the Years 1832–3–4. Wilmington,
DE: Scholarly Resources Ltd., reprint of 1837 ed.
Roth, H. Ling. (1977). Studies in Primitive Looms. Bedford: Ruth Bean, 4th ed.
Saron, Gustav and Louis Hotz. (1955). he Jews in South Africa: A History. London:
Oxford University Press.
Schaedler, Karl-Ferdinand. (1987). Weaving in Africa South of the Sahara. Munich:
Panterra.
Silver, M. S. (1984). he Growth of Manufacturing Industry in Tanzania: An Economic
History. Boulder, CO: Westview Press.
Simoons, Frederick J. (1960). Northwest Ethiopia: Peoples and Economy. Madison:
University of Wisconsin Press.
Southern Rhodesia. (1946). Report of the Committee of Enquiry into the Protection of
Secondary Industries in Southern Rhodesia. Salisbury : Government Printer (also
known as the “Margolis Report”).
Spence, C. F. (1951). he Portuguese Colony of Mozambique: An Economic Survey. Cape
Town: A. A. Balkema.
Stefanini, Giuseppe. (1922). In Somalia: Note e Impressioni di Viaggio. Florence: Le
Monnier.
Swainson, Nicola. (1980). he Development of Corporate Capitalism in Kenya, 1918–
1971. London: Heinemann.
Strihou, Jacques van Ypersele de. (1961). “Variations des Coeicients de Fabrication
dans Une Entreprise Congolaise, et Équilibre économique, 1931–1958.” Bulletin de
l’Institut de Recherches Économiques de Louvain, 27 (2): 47–96.
Valk, Peter de. (1996). African Industry in Decline: he Case of Textiles in Tanzania in the
1980s. Basingstoke: Macmillan.
Vandewalle, G. (1966). De Conjuncturele Evolutie in Kongo en Ruanda-Urundi, van 1920
tot 1939, en van 1949 tot 1958. Gent: Rijksuniversiteit te Gent.
Vansina, Jan. (1998). “Raia Cloth in West Central Africa, 1500–1800.” In Textiles:
Production, Trade and Demand, edited by Maureen Fennell Mazzaoui, 263–81.
Aldershot: Ashgate.
Vellut, Jean-Luc. (1985). “Originalités et Limites de L’industrie Manufacturière au Congo
Belge, c.1920–1960.” Unpublished longer version of a piece published under this
title in Les Belges à L’étranger: 150 Ans de Réalisations dans Le Tiers-monde, cata-
logue for an exhibition held in Brussels March 21 to April 24.
294 Clarence-Smith

Werbelof, Arnold. (1987). Textiles in Africa: A Trade and Investment Guide. London:
Alain Charles.
Wilde, John C. de. (1971). he Development of African Private Enterprise. Washington,
DC: International Bank for Reconstruction and Development.
Wingate, F. R. (1986). Ten Years’ Captivity in the Mahdi’s Camp, 1882–1892, from the
original manuscript of Father Joseph Ohrwalder. London: Darf, 2nd ed.
Wylde, Augustus B. (1901). Modern Abyssinia. London: Methuen.
Yamba, V. Bawa. (1995). Permanent Pilgrims: he Role of Pilgrimage in the Lives of West
African Muslims in the Sudan. Washington, DC: Smithsonian Institute Press.
Zawawi, Sharifa M. (2005). Kanga: he Cloth hat Speaks. New York: Azaniya Hills
Press.
 9

Explaining and Evaluating the Cash Crop

Revolution in the “Peasant” Colonies of
Tropical Africa, ca. 1890–ca.1930: Beyond
“Vent for Surplus”

Gareth Austin

he fundamental economic change of the colonial era in the so-called

peasant colonies of tropical Africa – those in which no major appropria-
tion of agricultural land for use by European settlers or concessionaires
occurred – was the massive expansion of export agriculture in areas with
favorable land quality and location. he timing varied among and within
individual colonies, but in general the fastest growth in the volume of
agricultural exports took place during the early decades of colonial rule,
ca. 1890 to ca. 1930, especially during the twenty years or so immedi-
ately preceding the First World War. In many areas, African agricultural
exports had begun before the colonial occupation; in a few, such as north-
ern Nigeria, with groundnuts (peanuts), it followed it. But even where it
was already established, export agriculture expanded greatly during colo-
nial rule. his expansion was associated with infrastructural innovation,
especially the introduction of mechanized transport. It was also associated
with institutional changes. he most important of these was the emergence
of a migrant-labor system that channeled workers into the areas suited to
export agriculture from the more distant and/or drier parts of the same or
neighboring colonies.
he cash crop revolution is evidently a key episode in the economic his-
tory of Africa, deining the structure of most of the nonsettler economies.
It is also particularly worthy of attention for students of comparative and

his chapter was written speciically for the July 2010 conference in Peduase, Ghana. In revis-
ing it, I have beneited from feedback there and on versions presented the same year to the
Department of Economic History and Institutions, Universidad Carlos III, Madrid, and at
the annual meeting of the African Studies Association of the United Kingdom, in Oxford.

295
296 Austin

long-term economic development. Sub-Saharan economies have oten been

thought of as static: at least before colonial rule, and perhaps during and
since. Specialists know that there have been episodes of economic growth,
albeit oten sustained for just a few decades, in parts of the region during all
these eras (e.g., Austin 2008; Berthélemy and Söderling 2001; Jerven 2010;
Manning 1982). In this respect, the cash crop revolution is especially signif-
icant for several reasons. First, it was unusually widespread geographically.
Second, in some areas for certain commodities, the speed of “takeof ” was
dramatic. hird, production was carried out almost entirely by Africans;
and, where they existed at all, European rivals failed in competition (e.g.,
Austin 1996). hus research into this episode may be particularly useful as a
means of testing long-term generalizations about the sources of stagnation
or change in indigenous economic behavior.
he causes and efects of the expansion of African export agriculture
became the subject of continuous research and lively controversy from the
1960s well into the 1980s. he “vent-for-surplus” (VFS) model(s), pioneered
by Hlya Myint and Robert Seresewski at the London School of Economics,
postulated that Africans expanded their output, in response to price incen-
tives newly provided by overseas markets, by making fuller use of previ-
ously underutilized stocks of land and labor (Myint 1958, 1964/1973, 1977;
Szereszewski 1965). hus it means exporting without opportunity cost: pro-
duction for domestic consumption (whether subsistence or marketed) is
unafected. In this perspective, African economic behavior was rational, but
only minimally innovative. During the 1970s, dependency theory became
popular in the academy, in Africa and beyond, with its radically pessimis-
tic model in which, while Africans increasingly produced for export, they
did so only because the colonial government prodded them into it, either
through the direct coercion of compulsory cultivation and forced labor or
through the indirect pressure of taxes that had to be paid in cash (Freund
and Shenton 1977; Rodney 1972). he result was that the export growth had
an opportunity cost in reduced subsistence output, because of the diversion
of labor from food to cash crops. Both VFS and dependency interpretations
became increasingly hard to sustain as research continued. heir accounts
of the mechanisms and motives of the cash crop revolution were found,
respectively, at best inadequate and at worst to a large extent wrong in the
light of a series of studies of the creative and decisive role of African entre-
preneurship in the faster and larger export expansions (see Berry 1975; Hill
1963, 1966; Hogendorn 1979, Hopkins 1973, 1978).
But in social science, even the most heavily criticized paradigms rarely
lose all adherents, especially when intellectual markets are segmented
 Explaining and Evaluating the Cash Crop Revolution 297

by disciplinary and ideological boundaries.1 Both VFS and dependency

continue to attract occasional support (e.g., respectively, Teal 2002 and
hompson 2006). Indeed, when the debate went quiet in the later 1980s, it
was not because the arguments had been deinitively settled – certainly no
new synthesis was established – but rather because the intellectual climate
had changed. he hold of dependency theory on let-leaning scholars had
been challenged by the critiques of a small but active group of heterodox
economists sympathetic to the old Marxian expectation that, ultimately,
colonialism could do no other than promote capitalist institutions, and
thereby the potential for long-term economic expansion (Sender and Smith
1986; Warren 1980). Politically, the 1980s brought the ideal of the unfet-
tered market “back in” at the level of policy, notably through “Structural
Adjustment.” Intellectually, the renewed interest in markets was animated
by a concern with the institutional frameworks surrounding transactions
(the seminal contributions, in the interdisciplinary study of Africa, being
Bates 1981, 1983). Indeed, the trickle of studies of the cash crop revolution
during the past quarter century was distinguished from the steady stream of
the previous quarter century not only in frequency, but by a tendency to give
more attention to institutional aspects (Austin 2005, 2007, 2009; Firmin-
Sellers 1996; from a social history perspective, Fall 1993; Manchuelle 1997).
his is not at all to say that recent work has completely overlooked issues
of resources and technology, as is illustrated by Bassett’s (2001) study of
Ivorian cotton growing, and Ruf ’s (1995a, 1995b) exploration of the long-
term dynamics of fertility and proitability in cocoa production. But an
unmistakable shit of emphasis has taken place. An exception that “proves
the rule” is Swindell and Jeng’s work on groundnut production and trade
in the Gambia. his gives a relative weight to resource issues that recalls an
earlier generation of scholarship; and indeed, though completed recently,
this important study had its roots in research that the authors began in the
very diferent intellectual setting of the 1970s (Swindell and Jeng 2006).
On the causes of the cash crop revolution, given the general emphasis on
institutional matters in recent economic historiography and related litera-
tures, this chapter will take up the uninished business of the sources and
scale of resource inputs and how they were combined. he main discussion
is organized in four sections. he irst describes the growth of agricultural
exports in more detail. he second section introduces the VFS approach,
its variants, and the many lines of criticism directed at it from the 1960s

1
What Hopkins called “the persistence of imperfections in the market for knowledge”
(Hopkins 1987, 122; cf. Hopkins 1986, 1473).
298 Austin

to the early 1980s. he third section proposes a more nuanced framework

for explaining the expansion. In what was a predominately labor-scarce
region of the world, the issue turns on the mobilization, use, and produc-
tivity of labor. It will be argued that both VFS and dependency approaches
it at least part of the evidence from parts of the region, but that both miss
a phenomenon that was both important at the time and signiicant for
long-term economic development in Africa: expansion achieved by raising
productivity.
he fourth section discusses the welfare consequences of the growth of
export agriculture, and relects on the long-term economic implications.
he welfare outcomes were probably determined by the institutions as much
as the size of income. On one hand, the capture of resource rents – by con-
trolling access to the areas where export agriculture could most proitably
be pursued – was a source of continuing inequality among diferent parts of
the region. On the other hand, the transition from slavery to free migrant
labor had more wide-ranging efects than previously recognized: it enabled
welfare gains from increased purchasing power to be difused, unevenly
but much more widely than is usually supposed. Finally, this chapter will
relect, among other things, on the implications of colonial-era cash crop
expansion for further structural and productivity change: on whether it
was, as dependency writers argued and as VFS implied, a story of “growth
without development,”2 and one-of growth at that.

9.1 THE GROWTH OF AGRICULTURAL EXPORTS

he most striking characteristic of the expansion of export agriculture was
its unevenness, over space as well as time.3 Table 9.1 summarizes the growth
of output of the major crops in the major centers of production in West
Africa, which accounted for the majority of population and foreign trade in
tropical Africa, and for the great majority of population and foreign trade
from the “peasant” colonies. In contrast, central Africa included substantial
foreign plantation sectors, while East Africa included a major “settler” col-
ony, Kenya. Outside West Africa, the largest example of export agriculture
was Uganda. In contrast to West Africa, there was apparently no precolo-
nial cotton industry in Uganda. Following initial experiments in 1902–3,

2
he phrase was used as the title of Clower, Dalton, Harwitz, and Walters (1966).
3
his section briely describes what subsequent sections try to explain. For fuller intro-
ductions, see Hopkins (1973, 167–236); Hogendorn (1975); and Harrison, Ingawa, and
Martin (1987, 517–34).
 Explaining and Evaluating the Cash Crop Revolution 299

Table 9.1. Main agricultural exports from main producing

colonies in West Africa, 1890–1910 (metric tons)

Colony & Crop 1890 1900 1914

Senegal groundnuts 27,221 140,922 280,526
Gambia groundnuts 18,554 36,378 67,924
Ghana cocoa 0 0.5 53,746
Nigeria cocoa 6* 205 5,018
Nigeria palm oil – 46,236 73,659
Nigeria palm kernels – 86994 165050
Nigeria groundnuts – 608 17,271
* Lagos only. For comparison, in 1900 Lagos accounted for 56 percent of
the total (calculated from Berry 1975, 221).
Note: “–” means no igure available. In Nigeria during this period, exports
of cocoa beans, palm oil (and palm kernels), and groundnuts came from the
southwest, southeast, and (at least ater 1912) north, respectively.
Sources: From Austin 2009, 5; Berry (1975, 221); Helleiner (1966, table
IVA8); Hill (1963/1997, 177); Kay and Hymer (1972, 336); Martin (1988,
148); Searing (2002, 199); and Swindell and Jeng (2006, 23, 101, 149).

production rose to nearly six thousand metric tons in 1914–15 (Engdahl

1999, 51–3, 73).
Unevenness over time was manifest in the contrast between the rel-
atively gradual multiplication of export volumes of certain crops from
areas where they had already been produced for export for several decades
(notably groundnuts in Senegambia and palm oil – and then also palm
kernels – in southeastern Nigeria) and the rapid takeof from nil exports
in other cases (cocoa in Ghana, groundnuts from northern Nigeria, cot-
ton in Uganda). he older export crop cultures went back to the transition
from the Atlantic slave trade to “legitimate commerce” during the early
and middle nineteenth centuries; groundnuts and palm oil were grown
for export in areas not far from water transport, such as the Senegambian
river basins and the Niger Delta. he dramatic new trades responded to
the construction of railways to the interior (Uganda, northern Nigeria)
and the new proitability of a relatively recently imported exotic crop,
cocoa beans. he expansion of the area under export crops generally
slowed greatly during the period from the middle of the First World War
until at least the end of the Second World War, primarily because, while
prices continued to luctuate, they did so between lower bands than dur-
ing most of the preceding quarter century (Martin 1989). It took a new
era of higher prices to help induce a further major expansion of output
300 Austin

during the last decade or so of colonial rule in most of tropical Africa: an

episode beyond the scope of this chapter.
he unevenness over space was stark in its extremes: the contrast between
the richest cash-cropping areas and the areas whose direct participation in
the agricultural export economy never went much beyond the export of
(mostly male, mostly migrant) labor to the production zones and market-
ing centers. he diference was a function of soils and transport costs.4 he
crops were not of equal value; cocoa beans were clearly the most proitable
option where they would grow well, which was only in certain forest zones.
he geographical variation, however, was graduated in two ways. First,
colonial administrations generally encouraged, and – especially in the drier
areas – in many cases tried to compel, savanna farmers to grow cotton.
Despite compulsion, production in such areas remained relatively mod-
est, as for instance in northern Côte d’Ivoire during this period (Bassett
2001). Hence migrant labor developed as a more realistic way by which the
inhabitants could obtain the cash to pay tax and buy goods (Asiwaju 1976).
Second, there were intermediate areas where crops were grown for export
giving returns to labor that were high enough to avert any major exodus of
labor but low enough not to attract an immigrant labor force. his was, for
instance, the situation in much of the palm oil belt in southeast Nigeria.

9.2 VENT-FOR-SURPLUS MODELS AND THEIR CRITICS

By deinition, VFS growth of exports occurs as result of putting to work a
previously unused “surplus productive capacity” (Myint 1964, 321). It is
growth with an unchanged production function: inputs are combined in
the same proportions and using the same methods of production as for
existing domestic markets. he only diference is that now they are applied
for export as well. hough the intellectual lineage of the idea may be traced
to Adam Smith, the VFS model proximately and decisively belongs to the
same generation as W. A. Lewis’s (1954) much more famous model of “eco-
nomic development with unlimited supplies of labour.” Like Lewis (from St
Lucia in the West Indies), Myint was a native of a tropical colony (Burma)
who became a professor in Britain. Like Lewis, he wrote in the classical
rather than the neoclassical tradition, in that his model envisaged a stable
equilibrium in which at least one factor of production was less than fully
employed. Like Lewis, who had just written a important report on the (as

4
On the latter see Chaves, Engerman, and Robinson (Chapter 10, this volume); Jedwab and
Moradi (2012); and Austin (2007, 100–3).
 Explaining and Evaluating the Cash Crop Revolution 301

he saw it, not particularly encouraging) prospects for industrialization in

Ghana (Lewis 1953), Myint was struck by the scarcity of labor relative to
land in West Africa, as in his native Southeast Asia. Yet Myint postulated
that labor was underemployed at the beginning of the growth of agricul-
tural exports, in the sense that there was a “leisure” reserve. He envisaged
an absolute deiciency of efective demand, such that Africans had no incen-
tive to use even their very limited supply of labor, despite the availability
of abundant cultivable land. he growth of agricultural exports was trig-
gered by the arrival of mechanized transport, in the form of the steamship
and then the railway, which greatly improved the barter terms of trade for
African exporters (the ratio of units of imports they could obtain in return
for one unit of exports). More than that, Myint envisaged the steamships
bringing to Africa a range of new commodities, which greatly increased
the desire of Africans to produce for export in order to be able to buy the
expanded range of imports (Myint 1958, 1964, 1977).
hough the VFS idea is oten referred to as “Myint’s model,” a series of
other authors soon formulated diferent versions of it (notably Szereszewski
1965; Helleiner 1966; Elliott 1969). hese revisions (none of whose authors
were from the hird World, as it happens), were all in the neoclassical tra-
dition. hey replaced the underemployment equilibrium with a trade-of,
but retained the fundamental proposition that export growth was achieved
without reducing output for existing consumers. hey did this by ascribing
a positive value to leisure in the minds of producers, and in Elliott’s case,
also to risklessness. hus the assumed initial idleness of part of the stocks
of labor and land was the result of a trade-of between the incentive of pos-
sible but very limited rewards from trade (relatively expensive purchases)
and the disincentive of losing leisure. While Elliott noted that farmers in
some areas would risk their food supply if they went into export farming,
he thought that, precisely for that reason, they would not want to do so.
Where export agriculture took of, on his analysis, food supply was not at
risk; though he recognized that the long gestation period of investment in
cocoa trees made cocoa growing inherently high risk in terms of the efort
and money involved (Elliott 1969).
he VFS approach drew a variety of criticisms. Some of the critics’
observations could be accommodated in a modiied version of the model;
others were incompatible with it. One potential criticism was anticipated
by Szereszewski himself: that, at least in the case of a tree crop like cocoa,
the expansion involved capital formation on a wide scale. hus one could
hardly claim it was growth with the same production function as in “tradi-
tional” agriculture. Szereszewski proceeded to estimate the scale of capital
302 Austin

formation involved, but noted that it consisted mainly of tree planting using
simple tools, hence “the formation of cocoa capital was simply the capitali-
sation of current local labour” (1965, 75). hus he argued that the produc-
tion function was efectively a single-factor one, in that the size of output
was essentially a function of labor inputs. Given this, the VFS framework
could be retained, providing that its basic premise held: that land and labor
were initially under used.
One obvious absence from the models was the important contribution of
migrant labor to the expansion of the major centers of export agriculture.
his falsiied the notion of growth being achieved purely, or even largely, by
the mobilization of previously idle hands. But it did so only at a relatively
local level: the VFS model could still apply to the economic region con-
cerned, if that is deined as comprising both the labor-importing area and
the areas that supplied it with seasonal migrant workers.
Again, it was observed that cash crop zones became net importers of
foodstufs (e.g., Freund and Shenton 1977, 193). But the evidence for this
comes from ater export agriculture had already “taken of ” in the area con-
cerned. It is not surprising that a population swollen by seasonal migrants
(both laborers and itinerant traders), and containing a high proportion of
households whose purchasing power had been boosted by their involve-
ment in the export economy, should provide a new market for foodstufs
from neighboring areas. Let us take a Ghanaian example. In Ejura district,
in the savanna just to the north of the forest zone that had become the cocoa
belt, commercial rice growing developed during the interwar years, supply-
ing the market in Kumasi, the commercial center of Ashanti (the heartland
of the former Asante kingdom),5 which comprised the northern half of the
Ghanaian forest zone (Austin 2005, 54–5, 470–1). his does not necessarily
mean that food production within the cocoa belt actually declined, only
that food consumption within it increased. hat a cash crop center might
eventually cease to be self-suicient in foodstufs does not contradict the
basic VFS claim that the cash crop takeof did not involve a reduction in
production of foodstufs within the area itself.
A criticism that posed greater diiculty for the model was that it erred by
omitting any role of African entrepreneurship. he supply response envis-
aged in VFS models was purely responsive, requiring no innovation and
amounting simply to doing more of the same thing. his stylized story took
no account of the actual history of the adoption of cocoa cultivation. Cocoa
was an exotic crop, and the people responsible for its initial spread in Nigeria

5
Consisting of what are now the Ashanti and Brong-Ahafo regions of Ghana.
 Explaining and Evaluating the Cash Crop Revolution 303

were not peasants but Creole merchants who, faced with diicult years in
trade in Lagos, decided to venture a long-term investment in this novel
tree crop. One of them, J. K. Coker, the leading igure in the Agege Farmers
Union (founded 1907), also had a political motive: he and his colleagues
spread what they apparently called “the cocoa gospel” so that others might
join them in proving that Africans could be successful in “modern” activi-
ties without becoming European in outlook (Hopkins 1978). he adoption
of the crop was not the only innovation of the early Nigerian cocoa plant-
ers. According to Agiri, the annual wage contract in Nigerian agriculture
was pioneered on Agege farms (Agiri 1983–4, 102). In Ghana, though the
Basel Mission had already been experimenting with cocoa growing, the
colonial government recognized a blacksmith named Tetteh Quashie as its
key pioneer. he most plausible version of the Tetteh Quashie story is that
a young male dependent of his went to the Portuguese island of Fernando
Po as a laborer on a cocoa plantation and smuggled cocoa pods back with
him on his return. Quashie used them to establish a small cocoa nursery
in Mampong on the Akwapim Ridge “from which he sold, or distributed,
pods or seedlings to Akwapim farmers” (Hill 1963, 172–3). Hill documented
the process by which cocoa farming in what became Ghana moved from an
experimental stage on the Akwapim Ridge, to investment in the purchase of
land in the Akim Abuakwa district, the planting of trees, and eventually the
production of many thousands of tons of beans a year. As she emphasized,
this was a story of risk taking and long-term investment (Hill 1963). Again,
Hogendorn showed that the export trade in groundnuts from northern
Nigeria began at the initiative of Hausa merchants. Confounding the expec-
tations of British merchants and the hopes of the British government and
Lancashire cotton textile manufacturers, well-capitalized Hausa merchants
did not respond to the opportunity presented by the arrival of the railway at
Kano by supplying cotton. Instead, Alhassan dan Tata and others responded
by giving credit to peasants in return for promises to plant groundnuts that
season to sell to the Hausa middlemen, who would sell them to the European
buyers (Hogendorn 1979). he active role of African entrepreneurs does not
necessarily conlict with the letter of the VFS models. On the other hand, the
evidence of Africans as innovators, risk takers, and investors stands in con-
trast to the spirit of the VFS approach, with its depiction of the African role
as purely responsive. Again, the likes of Tete Quarshie, Coker, and dan Tata
are hard to reconcile with the VFS sketch of static, subsistence-dominated
precolonial economies. hus, the entrepreneurship issue suggests that VFS
is at least insuicient as a framework for explaining the dynamics of the cash
crop revolution.
304 Austin

Drawing together the available literature in the 1970s, Hopkins (1973,

231–6) and Hogendorn (1975, 1976) used VFS as their framework, but
qualiied and embellished it to take account of these kinds of criticisms. A
major step toward refuting rather than just modifying the model was taken
by Tosh, with a powerful argument restricting VFS to the forest zones. He
noted that the assumption of a reserve of leisure was incompatible with
the well-known brevity of the planting season in tropical savannas, which
would have entailed a trade-of between the production of food crops such
as millet for household subsistence and that of potential export crops such
as cotton and groundnuts (Tosh 1980). While groundnuts were also part of
the local diet, Tosh’s point makes it hard to sustain the VFS premise that the
export crop expansion involved no reduction in food crop output.6
During the 1970s and the early 1980s, writers attracted to dependency
theory attacked the VFS approach wholesale for overlooking the role of colo-
nial coercion. Forms of such coercion ranged from compulsory cultivation
of cotton to the use of direct taxation, payable only in money, which obliged
farmers to obtain more cash, either from their own farming or as migrant
laborers (notably Freund and Shenton 1977; Mbodj 1980; Muhtar 1988;
Watts 1983; and, earlier, Suret-Canale 1964/1971). Tosh’s aforementioned
argument clariied in principle why savanna farmers might be reluctant to
respond to agricultural export markets, and why colonial administrations
might therefore believe that coercion was necessary to induce them to do
so. As a general framework for analyzing the agricultural export expansion
in tropical Africa, an obvious weakness of dependency theory was (and is)
that no serious reason exists to doubt that in many cases – including by
far the largest cases – of export expansion, the decision to plant crops for
export was voluntary. his applied to groundnuts, which was already the
major export crop from the savanna before the colonial Scramble for Africa
began in 1879. Groundnut exports expanded many times during the early
colonial period. In southern Ghana (cocoa) and southern Nigeria (palm
oil products and cocoa), there was not even any direct taxation during
the early colonial decades. With coercion applicable to only some of the
cases, the explanatory power of dependency theory is let to rely on the
notion of unequal exchange.7 Market economists and economic historians
tend to be skeptical of this as a general explanation (particular monopolies
excepted), to the extent that it relies on Marx’s labor theory of value. Later I

6
Elliott noted the seasonal labor bottleneck, but did not consider that it applied as broadly
as Tosh did (1969, 136–7).
7
For a brief discussion of the internal debate on this theory, see Frank (1978).
 Explaining and Evaluating the Cash Crop Revolution 305

will consider briely more limited versions of the idea of unequal exchange
that do not rest on the labor theory of value.
Ironically, the ideological tension between the VFS and dependency
interpretations of the colonial-period expansion of export agriculture over-
lay a partial similarity in their assessment of the dynamics of long-term
economic development in Africa. Both agreed that the cash crop expan-
sion constituted “growth without development,” in the sense that the tech-
nologies and production functions remained the same. For the future,
Myint envisaged two possibilities, once the land and leisure surpluses were
exhausted. he optimistic scenario was a transition to what we might now
call “Smithian growth”: a process of rising specialization, with the export
crops increasingly produced by farmers who produced nothing else, per-
mitting higher allocative eiciency in the economy as a whole. But Myint
also took seriously the pessimistic prospect of Geertzian agricultural invo-
lution: as the (in his view, exogenous) growth of population induced farmers
to decrease production for the market in order to concentrate on growing
enough food for the household’s consumption on plots of gradually dimin-
ishing size. hus, for Myint, the growth of export agriculture might prove
to be either a bridge to structural change and further growth or an interval
between a demand-constrained poverty amid underused resources and a
supply-constrained poverty with increasing pressure on resources (Myint
1958, 329–31, 1964, 36–9).
In any case, by the mid-1980s, the accumulation of suggested supple-
ments and partial refutations meant that the VFS approach had lost its ini-
tial aura of elegance and wide applicability across the various crops and
places involved in the export agriculture “revolution.” Dependency theory
had proved even more lawed empirically. As noted earlier, the next step in
the evolution of the literature was sideways: with some exceptions, schol-
arly interest shited to other problems. he next section suggests ways in
which we might take forward the uninished business, especially in the light
of research done on various precolonial and colonial contexts and themes
over the past scholarly generation.

9.3 BEYOND “VENT FOR SURPLUS”

Rereading the writings of VFS authors, and indeed of dependency writers, I
am struck by four assumptions that, whether reasonable or not at the time,
have now been overtaken by later research. Let me begin from these.
First, the VFS models assumed that the starting point of the growth of
export agriculture during the early colonial period was the ending of a
306 Austin

more or less deep isolation from international trade. Of course, Myint and
his colleagues were aware of the Atlantic slave trade, but in modeling what
happened some decades ater the beginning of the end of it, they made no
allowance for the antiquity and continuity of Africans’ commercial links
with other continents, and for the growth of “legitimate” commerce in
palm oil and groundnuts from the coast of West Africa, from well before
the introduction of steamship services in the 1850s. Again, Uganda is oten
said to have been more removed from maritime trade, but that had changed
greatly from the 1860s. A speciic continuity from the external trade of these
economies, from before until ater colonization, is in the kinds of commod-
ities that were imported. Cotton cloth, alcohol, and metal goods remained
prominent.8 To be sure, new products such as the bicycle arrived between
colonization and the First World War, but their appearance came piecemeal
rather than in one transformative moment. hus the evidence for Myint’s
idea of the emergence of new consumer wants as an important motive for
the growth of export agriculture is slight. On the contrary, colonial gov-
ernments’ imposition of their own currencies and restrictions on Africans’
purchases of guns and gunpowder presumably constricted the demand for
two of the main precolonial imports: materials for commodity currencies
(such as cowrie shells) and irearms.
Second, the formulation of the VFS models, and the surrounding expo-
sition and commentary, implicitly underestimated the extent of market
activity within late nineteenth-century precolonial economies. his is not
a criticism of the authors writing during the 1960s; that we can make this
observation relects other research that was under way then and synthe-
sized soon aterward (Hopkins 1973), and augmented by subsequent work
(including, for Uganda, on the Buganda kingdom in the nineteenth century:
Reid 2002).9 VFS writers, and their dependista critics, tended to treat com-
modity markets as if they had been pretty insigniicant before colonization.
Research has increasingly documented precolonial market exchange that
was not only more widespread, but more important in the economic and
social life of the populations than scholars previously recognized (Austin,
forthcoming). A telling indicator of this is Inikori’s observation that West
Africans’ importation of currency materials was intended not to facilitate
trade with Europeans, but to lubricate intra-African trade: European mer-
chants did not accept cowries and other commodity currencies back in

8
For a convenient if rather inadequately disaggregated example, see the tables of imports in
Kay and Hymer (1972, 327–31).
9
Despite its title, Reid’s book mainly concerns economic history.
 Explaining and Evaluating the Cash Crop Revolution 307

payment; they were currencies only in internal markets (Inikori 2007, 84).
For the precolonial decades of the nineteenth century, Lovejoy and others
have shown that, besides the growth of “legitimate commerce” in place of
the slave trade on the coast, there was also market expansion in the interior,
centered on the Sokoto Caliphate (1804–1900) and its commercial hub, the
city of Kano (e.g., Lovejoy 2005). he cotton textile industry and trade was
central to the economy of the Caliphate and to West Africa as a whole.
his means that the potential opportunity costs of export agriculture were
incompletely realized in the VFS models.10 Again, crops exported to other
continents were not everywhere the irst crops sold. Towns on the inner
bend of the Niger river, in the latitude of the desert, had long depended on
grains grown at a distance and brought by trading canoes.11 In the Sokoto
Caliphate, raw cotton and vegetable dyes, the raw materials of cotton textile
manufacturing, were produced and sold from estates (e.g., Lovejoy 2005).
Meanwhile on the other side of the continent, hundreds of miles from the
ocean, the regional trade included such cash crops as sesame (Tosh 1978).
In the Kigezi district of southwestern Uganda, before and during colonial
rule, “food crops were cash crops” (Carswell 2007, 25).
he third, and perhaps starkest, of the implicit assumptions in the VFS
literature about the starting point of the export takeof is that the agri-
cultural labor force comprised owner-cultivators and their families: no
slaves or pawns, no labor market. he supposed absence of the latter fol-
lows from the underestimation of market activity: if commodity produc-
tion was only a relatively small supplement to subsistence production,
there would be no or negligible demand for extra-familial labor. Since
the 1970s, research has cumulatively shown evidence of increasing use of
slaves within West African economies, during the Atlantic slave trade and
especially ater it. he nineteenth century saw increased trading in slaves
who stayed within West Africa, and were purchased primarily to pro-
duce commodities: whether for the maritime “legitimate commerce” or
for intraregional trade (Austin, forthcoming). To a lesser extent, this also
happened in parts of East Africa, including what became Uganda (Médard
and Doyle 2007; Reid 2002). Slavery was not a marginal phenomenon
during that century, especially toward the end of it. Klein estimates that

10
In a very perceptive 1976 article, Sheila Smith considered the implications of including
an existing handicrats sector in the analysis (something of course omitted by the VFS
models). But her article did not present enough evidence to convert her theory into an
alternative historical interpretation (Smith 1976).
11
he situation early in the colonial period was the subject of a report by a French oicial
(Givieu 1911).
308 Austin

more than 30 percent of the population of French West Africa were slaves
at the turn of century (Klein 1998, 252–6). Lovejoy and Hogendorn put
the equivalent igure for the Sokoto Caliphate, the most populous preco-
lonial state in West Africa, as “between a quarter and a half ” (Lovejoy and
Hogendorn 1993, 305). Pawning (debt bondage), too, was a major source
of extra-familial labor recruitment in non-Muslim parts of West Africa
(Lovejoy and Falola 2003).
he integral importance of slavery to most of the economies in which
the cash crop revolution was to occur on a large scale is incompatible with
the basic VFS proposition of surplus labor in the form of a prior leisure
reserve. Again, the decline of slavery and pawning ater colonization was
oten slow: colonial governments oten, though not always, took immedi-
ate action against slave raiding and trading but in many cases were hesitant
to take decisive measures against an institution on which they oten con-
sidered chietaincy and social control to depend. Not least, the economic
conditions under which labor coercion was proitable to owners had not
been abolished. Colonial governments acted on this when they used forced
labor. Slave masters appreciated the point too. Hence, “‘When the slaves let,
the owners wept’” (Ohadike 1999). Before slaves ran away or walked out
(as many, but far from all, did), they were oten put to work establishing or
expanding cash crop farms (Austin 2005, 240–2, 513–14, 2009; Salau 2010).
Again, when credible state action was taken against slaveholding in south-
western Nigeria, the supply of pawns proved remarkably elastic: pawn labor
was substituted for slaves and for some years was the main source of labor
on the Agege cocoa farms (Oroge 1985).
he inal VFS assumption is not wrong, though it is too simple: the notion
that land was relatively abundant, in the sense that the cost and availabil-
ity of it was not a constraint on the expansion of output. I have reviewed
this issue elsewhere in detail, over a much longer period (Austin 2008).
Qualiications need to be made regarding the heterogeneity of land and
the diiculty of reproducing soil fertility, especially where trypanosomiasis
denied the opportunity to use animal manure. Even today, when chemical
fertilizers exist (at relatively high cost in energy and money), the most ei-
cient way to maintain fertility in much of West Africa seems to be the tra-
ditional one of fallowing (de Rouw 1995; Goldstein and Udry 2008). In the
case of the most proitable of the early colonial cash crops, Ruf ’s long-term
international comparisons and micro research have shown that the cost of
producing cocoa when the trees are replanted, rather than planted on land
newly cleared from forest, is high enough to have an oten decisive inluence
on long-term proitability (Ruf 1995a, 1995b). But, for the irst generation of
 Explaining and Evaluating the Cash Crop Revolution 309

cocoa farmers, that was not necessarily a problem (intergenerational trade-

ofs are considered in Austin 2005, 348–50, 533). Again, in Ghana perhaps
more than anywhere, the adoption of cocoa stimulated buying and renting
of land. While land was generally abundant, this was not so for farmers
whose chietaincies owned relatively little land suitable for cocoa farming.
Such investors bought or rented, oten in anticipation of future rather than
current needs (Austin 2007; Hill 1963/1997). A inal qualiication is that
land abundance may not have been true of the vicinity of Kano when the
groundnut trade took of. Hogendorn conceded that additional groundnuts
were grown on plots previously used for millet, though he maintained that
land as such was not a constraint (Hogendorn 1979). His pioneering study
focused on the groundnut trade; further research is needed on production.
We are now in a position to set out a range of forms taken by the growth
of export agriculture, in the hope that this will provide a useful framework
for thinking about it in the context of the weaknesses of both existing frame-
works, VFS and dependency. Given that labor was generally the scarce fac-
tor, the most pertinent way to propose this is in terms of the use of labor.
First, there is at least one case the VFS framework still appears to it: the
oil palm belt of southeast Nigeria. here the growth of agricultural exports,
at least until 1918, seems to have involved no opportunity cost: no reduction
in existing forms of production, whether for export or domestic markets,
according to Martin’s study of Ngwa district, as well as Usoro’s macro-level
assessment (Martin 1988; Usoro 1974).
Second, there are many cases in the savannas of West Africa where the
growth of exports constrained or would have constrained the planting of
food crops. It is not entirely surprising that colonial governments and met-
ropolitan textile manufacturers hoped that the African colonies would come
to rival India, or even the southern United States, as suppliers of cotton: it
was already very widely grown for the indigenous handicrat industry. But
given that expanded production of output risked the food security of the
producing households, it is also not surprising that there was reluctance
and resistance. he use of terror to induce the collection of wild rubber in
King Leopold’s Congo had no parallel in the forests of Nigeria, Ghana, Côte
d’Ivoire, and Guinea, where African producers and traders collected rubber
freely (Arhin 1980; Dumett 1971; Fenske 2011; Osborn 2004). Even with
cotton cultivation, the use of coercion by colonial authorities was more sus-
tained and systematic in the concession colonies than in the “peasant” colo-
nies (Isaacman and Roberts 1995, Likaka 1997). But in parts of the West
African savannas, unlike with rubber collection in the forest, there were
attempts at coercion by the administration. In northern Nigeria, as noted
310 Austin

earlier, Hausa merchants interposed an alternative, groundnut exporting,

which was so successful commercially that cotton ceased to be the prior-
ity for government revenue or for European merchants. In Uganda, it was
diferent again. In Buganda, it appears that the plantain and banana diet
could be supplied so eiciently, in labor terms, that the usual food security
constraint on planting more cotton did not apply (Elliott 1969, 136–7). In
western Uganda, the story was more complicated. Tosh reports that Lango
farmers initially declined to expand their cotton output. But by the 1920s
they changed their minds, partly because higher cotton prices increased
the incentive, and partly because government measures to improve food
security by establishing a system of communal granaries reduced the risk
(Tosh 1978).
hird, there is at least one important case where cultivation for export
was adopted at the expense, not of the farmers’ food security, but of a local
market. It seems that many of the Hausa peasants who planted additional
groundnuts, to sell them into the export trade, found time and space to
do so by reducing their planting of millet. Previously, they had sold millet
in the Kano market to obtain cash for their own purchases and to pay tax
(Hogendorn 1979). Note that the Sokoto Caliphate had a land tax payable in
cash (cowrie shells) before the colonial conquest (Lovejoy and Hogendorn
1993, 165–6, 338). he trade-of between growing food crops for sale and
growing groundnuts was one of the major sources of the 1916 famine in
Kano (Watts 1983, 283–5). Further research may reveal other cases where
export production created or contributed to food scarcities in local mar-
kets. he Lango farmers’ eventual adoption of cotton was at the expense of
reducing their cultivation of sesame seeds for regional markets (Tosh 1978).
Carswell shows, however, that the failure of colonial attempts to promote
export crops in Kigezi was to a large extent because farmers stuck to grow-
ing food crops, which they regularly sold but could also eat themselves if
necessary (Carswell 2007, 24–48).
Finally, one major crop can be considered as a case of innovation and
improved total factor productivity (Austin, in press, 2014). he adoption of
cocoa farming in Ghana and southwestern Nigeria was new. As a “perma-
nent” tree crop, its cycle was completely diferent from that of the annual
crops that farmers were used to growing. Planting cocoa trees created capi-
tal goods. Its labor requirements in the harvesting and marketing phase
extended through the dry season. his hugely increased labor productivity
during the season in which, previously, there was very little to do in agri-
culture and therefore the opportunity cost of labor was very low (Austin
2005, 73–9, 475–6; Austin 2008, 609). It is not surprising that the story in
 Explaining and Evaluating the Cash Crop Revolution 311

both major cocoa-producing colonies (of this period) was one of African
enterprise rather than colonial coercion. For in the supply of enterprise,
and of cash to buy land (though much of that was on credit), pay rents,
and support the workforce until the irst cocoa trees came into bearing,
Ghanaian and Nigerian producers reinvested the proceeds of earlier, preco-
lonial market activities.
he adoption of cocoa entailed serious opportunity costs. It is oten
remarked that there was no risk to food supply, because West African farm-
ers (especially women, in the Ashanti case) would plant tall food crops such
as plantain to shade the young cocoa trees that they (and particularly their
husbands) had planted. Once the cocoa trees formed a shade canopy, how-
ever, food crops could no longer survive on the farm. Farmers deliberately
planted cocoa trees close together, because the shade also prevented weeds
from growing. So once the cocoa stock stopped expanding, there was a risk
to food security. his was why the Ashanti Confederacy Council of Chiefs
came to ban new cocoa farms in 1938, for what proved to be eight years
(Austin 2005, 325–36, 529–31). But more immediately, during the cocoa
takeof, existing forms of production for the market, rather than for sub-
sistence, were dropped – at the choice of the producers – to make way for
cocoa. Southwest Nigeria and much of the south of Ghana turned from
palm oil production to cocoa farming (an option not available in southeast
Nigeria). Further inland, in Ashanti, producers abandoned artisanal gold
mining, again because cocoa paid better on average (Austin 2005, 78, 476).
Of the four categories of cases of export crop adoption or expansion, the
VFS may well have been the least widespread. More commonly, the growth
process involved opportunity costs. However, these varied. Where export agri-
culture expanded fastest, what was reduced was existing forms of production
for the market: whether of food or gold. he overall implication is that the cash
crop revolution of the early colonial period needs to be viewed in the context
of the prior history of market production in West Africa and Uganda.

9.4 INSTITUTIONS, IMPLICATIONS, AND OUTCOMES

I have deliberately given priority to resources, products, and techniques,
because their history has been neglected recently, compared to institutions.
But politics and rule making is a necessary dimension of any explanation of
economic change. he economic historiography written during the 1960s
and 1970s was highly impressed with the lexibility and economic eiciency
of indigenous institutions. Hill, above all, documented the social organiza-
tion of collective land purchase by the stranger farmers in Akim Abuakwa,
312 Austin

and pointed to its evident success (Hill 1963). his impression is supported
by the account given earlier, with the diference that not all the institutions
were good for everybody. In southwest Nigeria, pawning proved remark-
ably efective at quickly providing a supply of laborers when slavery was
banned, but this lexibility evidently came at a cost to the pawns. Again,
slavery was a major source of early cash crop labor in some areas.
Focusing on the colonial context, we have noted that the most pros-
perous “peasant” colonies were those where the colonial government did
not impose direct taxation, still less compulsory cultivation. Again, where
African farmers were let either free of European competition (as in Nigeria)
or free to beat it without government interference (as in Ghana), the out-
come was much more rapid growth of exports than where European settlers
enjoyed privileged access to African labor through the French system of
corvée, a form of forced labor. It is surely no coincidence that the Ivoirian
cocoa takeof began following the abolition of forced labor (Hopkins 1973,
218–19). he latter case shows what the former does not: that the European
authorities did indeed have choice. It was politically possible to try to turn
“peasant” economies into European planter colonies; and where that choice
was taken, several decades passed before the policy was abandoned – ater
a campaign led by African capitalist cocoa farmers (Cooper 1996, 186–97,
527–31). For the British, the availability of settler colonies in Kenya and in
the south may have reduced the demands on the administrators in West
Africa and in London to concede to the demand of soap manufacturer W.
H. Lever for large land concessions in Nigeria on which his company would
produce palm oil. Given that degree of policy autonomy, the British in
Ghana, in particular, could see how the “experiment” of African-led export
agriculture worked out before they institutionalized it in the rhetoric of the
“West Africa Lands Policy” (Austin 1996; Hopkins 1973, 209–16).
his is not to suggest that, even in British West Africa, colonial sub-
jects were free to compete on fully equal terms with European irms. he
European cartelization of the service sector of the export economy is
well known, and on the whole was tolerated by the colonial governments
(Austin and Uche 2007; Olukoju 1992, 2001–2). he attempts of European
merchants to replace intense competition with price ixing oten enjoyed
some success, even though confounded in 1937–8 in Ghana by the most
comprehensive and sustained of a series of cocoa “hold-ups” (Austin 1988;
Miles 1978). Monopoly is one of the ways “unequal exchange” became
reality, even though African brokers could occasionally exploit informa-
tion asymmetries to extract economic rents from European irms (Austin
2005, 366–73, 535–6). Again, even when the markets were free of collusion,
 Explaining and Evaluating the Cash Crop Revolution 313

Table 9.2. Foreign trade of major West African economies, selected years
(£ million, rounded down, at 2005 UK prices)

1897 1913 1919 1929 1932 1938 1945 1952 1960

Ghana 133 558 548 888 615 981 741 2,950 3,610
Nigeria 225 838 784 1,240 752 978 940 4,740 5,942
FWA 266 723 418 813 531 763 1,140 3,980 4,120
Uganda n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1,390 1,060
Key: FWA = French West Africa. he 1919 igure for FWA is actually for 1920. For the years marked
“not available,” the igures for Uganda’s exports were not separated from Kenya’s.
Source: Nominal data from Munro, Africa, Appendix 1, to two decimal places. Converted to real
terms using Oicer 2010. Accessed July 12, 2010.

W. A. Lewis’s argument is surely valid, that the prices paid for African pro-
duce, as opposed to the prices paid for agricultural produce from the likes
of Australia, were partly a function of the low average earnings of the rest of
the economy (Lewis 1978).12
Even so, recent and current research suggests that the outcomes of the
cash crop revolution, at least where it was on the kind of scale that might
be considered revolutionary, were more positive than previously thought.
Scholars have long recognized that export agriculture in “peasant” colonies
multiplied the per capita import purchasing power. Whether the popula-
tion of the colonies concerned more than doubled between 1900 and 1960,
as has oten been thought (a recent example is Tabutin and Schouemaker
2004, 590–1), or grew much more slowly as Manning (2010) argued,
Table 9.2 suggests a major increase in foreign trade per head (even though
the price index used, unfortunately, is the available but only loosely relevant
UK retail price index).
But the distribution of this purchasing power was highly unequal, and
there has been much debate about whether the migrant labor system dif-
fused more than a very small proportion of it to the distant villages from
which the seasonal workers came (Amin 1974b; Miracle and Bery 1970).
Recently, Moradi has pioneered the use of anthropometrics as a means
of studying changes in physical welfare during the twentieth century.
Current work on the heights of African soldiers recruited to the Gold Coast
Regiment provides strong evidence that people from all regions of Ghana
(and beyond) tended to become taller during the colonial period (Moradi
2008). A paper on which Moradi, Austin, and Baten are currently engaged

12
I thank Joseph Inikori for reiterating this in conversation.
314 Austin

includes evidence that the improvements in the average height of succes-

sive birth cohorts were correlated with cocoa income (output times price).
his is consistent with an earlier inding that migrant laborers on Ashanti
farms succeeded in obtaining gradually better terms of employment over
several decades of the colonial period (and ater) (Austin 2005, 316–21,
412–30, 528–9, 542–5). One cannot assume that the positive story revealed
by the changing heights of African soldiers in Ghana will be replicated
when the anthropometric spotlight shits to African recruits in the French
and Portuguese armies.13 At least, given the resource poverty of much of
French West Africa (the Ivorian forest zone and the Senegal groundnut belt
excepted), it would be surprising if heights rose as much as in a cocoa econ-
omy – except insofar as migrant laborers from French West Africa were a
major component of the hired labor force on Ghanaian cocoa farms.
he difusion of cash crop income to migrant laborers, speciically via
their capacity to obtain increasingly good terms from the owners of the
cocoa farms, was made possible by a fundamental institutional change: the
ending of slavery. he areas from which migrant laborers now came were
oten the same relatively poor places that had previously been the unwill-
ing source of large numbers of slaves. But this institutional reform was not
fully exogenous to cash cropping. Admittedly, the ideology of abolition was
exogenous in this respect, and ultimately colonial administrations had no
political alternative but to act – sooner or later – against the recruitment of
new slaves, which in turn undermined the position of existing slave owners.
But the areas where hired labor relatively quickly replaced slave labor were
those where the advent of a lucrative cash crop, notably cocoa in southwest
Nigeria and southern Ghana, enabled masters to aford to become employ-
ers. Simultaneously, even less proitable cash crops enabled former slaves to
set themselves up as free peasants, paying their taxes and buying a modest
range of items (Austin 2005, 236–49, 512–15, 2009).
Finally, let us relect briely on the signiicance of the cash crop revo-
lution for long-term economic growth and structural change. “Peasant”
economies, even when much of the output was really in the hands of small
capitalists rather than peasants, tend not to provide the kind of linkages
that facilitate early industrialization. he contribution of cofee proits
to the origins of the Sao Paulo cotton textile industry before 1914 owed
much to the concentration of those proits in the hands of large estate own-
ers (Dean 1969). hat was part of the logic – or at least, of the rhetorical
rationale – for the later state “socialization” of export revenue through the

13
he French data are currently being studied by Denis Cogneau and Alexander Moradi.
 Explaining and Evaluating the Cash Crop Revolution 315

marketing board system, to usually disappointing economic efect. Cash

crop earnings, private and public, however, paid for much of the formal
education that the next generations received. hey also expanded the mar-
ket for manufactures. Recent work has shown that this included specialist
African handicrat producers (Byield 2002; Kriger 2006).
he cash crop revolution in tropical Africa was made possible by the
industrialization of the West. he irst industrial revolution created the mass
markets for soap that made British and French merchants seek palm oil and
groundnuts in the mid-nineteenth century. he “second industrial revolu-
tion” followed through with the rubber tire and milk chocolate. One thing
the VFS approach got partly right is that, during the early colonial period,
exporting had a very low opportunity cost for at least the more food-se-
cure colonies, in the sense that producing cocoa or groundnuts for export
usually delivered higher returns than did operating within purely domes-
tic markets;14 though this was not true of all crops everywhere, as we have
seen. But the dynamic Western economies did not intersect with a purely
static set of African economies. he cash crop revolution was a dramatic
episode in the very long-term African path of responding to abundant but
fragile lands with land-extensive agricultural techniques and a willingness
to adopt innovations, notably exotic crops, providing they did not imperil
the adopters’ own food security (Austin 2008).

References
Agiri, B. A. (1983–4). “he Development of Wage Labour in Agriculture in Southern
Yorubaland 1900–1940.” Journal of the Historical Society of Nigeria 12(1&2):
95–107.
Amin, Samir. Ed. (1974a). Modern Migrations in West Africa. London: Oxford University
Press for International African Institute.
(1974b). Accumulation on a World Scale: A Critique of the heory of Underdevelopment,
2 vols. New York: Monthly Review Press.
Arhin, Kwame. (1980). “he Economic and Social Signiicance of Rubber Production
and Exchange on the Gold and Ivory Coasts, 1880–1900.” Cahiers d’études afric-
aines 20: 77–8, 49–62.
Asiwaju, A. I. (1976). “Migrations as Revolt: he Example of the Ivory Coast and the
Upper Volta before 1945.” Journal of African History 14: 577–94.
Austen, Ralph A. (1981). “Capitalism, Class and African Colonial Agriculture: he
Mating of Marxism and Empiricism.” Journal of Economic History 41: 657–63.
Austin, Gareth. (1988). “Capitalists and Chiefs in the Cocoa Hold-ups in South Asante,
1927–1938.” International Journal of African Historical Studies 21(1): 63–95.

14
I thank Leandro Prados de la Escosura for this observation.
316 Austin

(1996). “Mode of Production or Mode of Cultivation: Explaining the Failure of

European Cocoa Planters in Competition with African Farmers in Colonial
Ghana.” In Cocoa Pioneer Fronts since 1800: he Role of Smallholders, Planters and
Merchants, edited by W. G. Clarence-Smith, 154–75. London: Macmillan.
(2005). Labour, Land and Capital in Ghana: From Slavery to Free Labour in Asante,
1807–1956. Rochester NY: Rochester University Press.
(2007). “Labour and Land in Ghana, 1879–1939: A Shiting Ratio and an Institutional
Revolution.” Australian Economic History Review 47(1): 95–120.
(2008). “Resources, Techniques and Strategies South of the Sahara: Revising the
Factor Endowments Perspective on African Economic Development, 1500–2000.”
Economic History Review 61(3): 587–624.
(2009). “Cash Crops and Freedom: Export Agriculture and the Decline of Slavery in
Colonial West Africa.” International Review of Social History, 54(1): 1–37.
(in press, 2014). “Vent for Surplus or Productivity Breakthrough? he Ghanaian
Cocoa Take-of, c.1890–1936,” Economic History Review.
(Forthcoming). Markets, Slaves and States in West Africa. Cambridge: Cambridge
University Press.
Austin, Gareth and Chibuike Ugochukwu Uche. (2007). “Collusion and Competition in
Colonial Economies: Banking in British West Africa, 1916–1960.” Business History
Review 81: 1–26.
Bassett, homas J. (2001). he Peasant Cotton Revolution in West Africa: Côte d’Ivoire,
1880–1995. Cambridge: Cambridge University Press.
Bates, Robert H. (1981). Markets and States in Tropical Africa. Berkeley : University of
California Press.
(1983). Essays on the Political Economy of Rural Africa. Cambridge: Cambridge
University Press.
Berry, Sara S. (1975). Cocoa, Custom and Socio-Economic Change in Rural Western
Nigeria. Oxford: Oxford University Press.
Berthélemy, Jean-Claude and Ludvig Söderling. (2001). “he Role of Capital Accumulation,
Adjustment and Structural Change for Economic Take-of: Empirical Evidence from
African Growth Episodes.” World Development 29(2): 323–243.
Byield, Judith. (2002). he Bluest Hands: A Social and Economic History of Women Dyers
in Abeokuta (Nigeria), 1890–1940. Portsmouth, NH: Heinemann.
Carswell, Grace. (2007). Cultivating Success in Uganda: Kigezi Farmers and Colonial
Policies. London and Nairobi: British Institute in Eastern Africa.
Chaves, Isaías, Stanley L. Engerman and James A. Robinson (2014) “Reinventing the
Wheel: he Economic Beneits of Wheeled Transportation in Early Colonial British
West Africa” (this volume)
Clower, Robert W., George Dalton, Michael Harwitz, and A. A. Walters. (1966). Growth
without Development: An Economic Survey of Liberia. Evanston, IL: Northwestern
University Press.
Cooper, Frederick. (1996). Decolonization and African Society: he Labor Question in
French and British Africa. Cambridge: Cambridge University Press.
Dean, Warren. (1969). he Industrialization of São Paulo, 1880–1945. Austin: University
of Texas Press.
De Rouw, A. (1995). “he Fallow Period as a Weed-Break in Shiting Cultivation
(Tropical Wet Forests).” Agriculture, Ecosystems and Environment 54: 31–43.
 Explaining and Evaluating the Cash Crop Revolution 317

Dumett, Raymond E. (1971). “he Rubber Trade of the Gold Coast and Asante in the
Nineteenth Century.” Journal of African History 12(1): 79–101.
Elliott, C. M. (1969). “Agriculture and Economic Development in Africa: heory and
Experience 1880–1914.” In Agrarian Change and Economic Development, edited by
E. L. Jones and S. J. Woolf. London: Methuen.
Engdahl, Torbjörn. (1999). he Exchange of Cotton: Ugandan Peasants, Colonial Market
Regulations and the Organisation of the International Cotton Trade, 19014–1918.
Uppsala: Uppsala University.
Fall, Babacar. (1993). Le travail forcé en Afrique Occidentale française (1900–1945). Paris:
Karthala.
Fenske, James. (2011). “‘Rubber will not keep in this country’: Failed Development in
Benin, 1897–1921.” Working paper.
Firmin-Sellers, Kathryn. (1996). he Transformation of Property Rights in the Gold Coast.
Cambridge: Cambridge University Press.
Frank, Andre Gunder. (1978). “On Unequal Exchange.” In Dependent Accumulation
and Under-Development, edited by Andre Gunder Frank, 103–10. London:
Macmillan.
Freund, W. M. and R. W. Shenton. (1977). “Vent-for-Surplus heory and the Economic
History of West Africa.” Savanna 6(2): 191–6.
Givieu, I. H. (1911). “Etude commerciale sur les riz de la vallée du Niger.” Dakar, March
25, unpublished: ile Q.54, Archives national du Sénégal.
Goldstein, Markus P. and Christopher Udry. (2008). “he Proits of Power: Land Rights
and Agricultural Investment in Ghana.” Journal of Political Economy 118(6):
981–1022.
Harrison, C., T. B. Ingawa, and S. M. Martin. (1987). “he Establishment of Colonial
Rule in West Africa, c.1900–1914.” In History of West Africa, vol. 2, edited by J. F.
Ade Ajayi and Michael Crowder, 485–545. Harlow, UK: Longman, 2nd ed.
Helleiner, Gerald K. (1966). Peasant Agriculture, Government and Economic Growth in
Nigeria. Homewood, IL: R. D. Irwin.
Hill, Polly. (1963). he Migrant Cocoa-Farmers of Southern Ghana, 2nd edition with
preface by G. Austin. Hamburg: LIT (1997 edition; 1963 edition, Cambridge).
Hogendorn, Jan S. (1975). “Economic Initiative and African Cash Farming: Pre-colonial
Origins and Early Colonial Developments.” In Colonialism in Africa 1870–1960,
vol. IV, he Economics of Colonialism, edited by Peter Duignan and L. H. Gann,
283–328. Cambridge: Cambridge University Press.
(1976). “Vent-for-Surplus Models and African Cash Agriculture to 1914.” Savanna
5(1): 15–28.
(1979). Nigerian Groundnut Exports: Origins and Early Development. Zaria: Ahmadu
Bello University Press.
Hopkins, Antony G. (1973). An Economic History of West Africa. London: Longman.
(1978). “Innovation in a Colonial Context: African Origins of the Nigerian Cocoa-
Farming Industry, 1880–1920.” In he Imperial Impact, edited by Clive Dewey
and Antony G. Hopkins, 83–96, 341–2. London: Athlone Press for University of
London.
(1986). “he World Bank in Africa: A Historical Perspective.” World Development
14(12): 1473–87.
(1987). “Big Business in African Studies.” Journal of African History 28(1): 119–40.
318 Austin

Inikori, Joseph E. (2007). “Africa and the Globalization Process: Western Africa, 1450–
1850.” Journal of Global History 2(1): 63–86.
Isaacman, Allen and Richard L. Roberts. Ed. (1995). Cotton, Colonialism and Social
History in Sub-Saharan Africa. Portsmouth, NH: Heinemann.
Jedwab, Rémi and Alexander Moradi. (2012). “Revolutionizing Transport: Modern
Infrastructure, Agriculture and Development in Ghana.” Paper presented at
the Modern and Comparative Economic History Seminar, London School of
Economics.
Jerven, Morten. (2010). “African Growth Recurring: An Economic History Perspective
on African Growth Episodes, 1690–2010.” Economic History of the Developing
Regions 25(2): 127–54.
Kay, G. B. and Stephen Hymer. (1972). “Statistical Abstract.” In he Political Economy
of Colonialism in Ghana, edited by G. B. Kay, 305–419. Cambridge: Cambridge
University Press.
Klein, Martin A. (1998). Slavery and Colonial Rule in French West Africa. Cambridge:
Cambridge University Press.
Kriger, Colleen E. (1954). “Economic Development with Unlimited Supplies of Labour.”
Manchester School of Economic and Social Studies 20: 139–91.
(1978). “he Factorial Terms of Trade.” In he Evolution of the International Economic
Order, edited by W. A. Lewis, 14–20. Princeton, NJ: Princeton University Press.
(2006). Cloth in West African History. Lanham, MD: AltaMira Press.
Lewis, W. Arthur. (1953). Report on Industrialisation and the Gold Coast. Accra: Gold
Coast Government.
Lewis, W. Arthur. (1954). “Economic Development with Unlimited Supplies of Labour.”
he Manchester School, 22(2): 139–91.
(1978). he Evolution of the International Economic Order. Princeton, NJ: Princeton
University Press.
Likaka, Osumaka. (1997). Rural Society and Cotton in Colonial Zaire. Madison:
University of Wisconsin Press.
Lovejoy, Paul E. (2005). Slavery, Commerce and Production in the Sokoto Caliphate of
West Africa. Trenton, NJ: Africa World Press.
Lovejoy, Paul E., and Toyin Falola. Eds. (2003). Pawnship, Slavery, and Colonialism in
Africa. Trenton, NJ and Asmara: Africa World Press.
Lovejoy, Paul E. and Jan S. Hogendorn. (1993). Slow Death for Slavery: he Course of
Abolition in Northern Nigeria, 1897–1936. Cambridge: Cambridge University
Press.
Manchuelle, François. (1997). Willing Migrants: Soninke Labor Diasporas, 1848–1960.
Athens OH: Heinemann.
Manning, Patrick. (1982). Slavery, Colonialism and Economic Growth in Dahomey,
1640–1960. Cambridge: Cambridge University Press.
(2010). “African Population: Projections, 1851–1961.” In he Demographics of Empire:
he Colonial Order and the Creation of Knowledge, edited by Karl Ittmann, Dennis
D. Cordell, and Gregory Maddox, 245–75. Athens: Ohio University Press.
Martin, Susan M. (1988). Palm Oil and Protest: An Economic History of the Ngwa Region,
South-Eastern Nigeria, 1800–1980. Cambridge: Cambridge University Press.
(1989). “he Long Depression: West African Export Producers and the World
Economy, 1914–1945.” In he Economies of Africa and Asia in the Inter-War
Depression, edited by Ian Brown, 74–94. London: Routledge.
 Explaining and Evaluating the Cash Crop Revolution 319

Mbodj, Mohamed. (1980). “Sénégal et dépendance: le Sine-Saloum et l’arachide,

1887–1940.” In Sociétés paysannes du tiers-monde, edited by Catherine Coquery-
Vidrovitch, 139–54. Lille: Presses Universitaires de Lille.
Médard, Henri and Shane Doyle. Eds. (2007). Slavery in the Great Lakes Region of East
Africa. Oxford: James Currey.
Miles, John. (1978). “Rural Protest in the Gold Coast: he Cocoa Hold-ups, 1908–1938.”
In he Imperial Impact: Studies in the Economic History of Africa and India, edited
by Clive Dewey and Antony G. Hopkins, 152–70, 353–7. London: Athlone.
Miracle, Marvin P. and Sara S. Berry. (1970). “Migrant Labour and Economic
Development.” Oxford Economic Papers 22 (n.s.): 1, 86–108.
Moradi, Alexander. (2008). “Confronting Colonial Legacies – Lessons from Human
Development in Ghana and Kenya, 1880–2000.” Journal of International
Development 20(8): 1107–21.
Muhtar, Mansur. (1988). “Vent for Surplus Critique and the Development of Export
Crop Production in Northern Nigeria – Towards a Socio-Economic Interpretation
of History.” In Topics on Nigerian Economy, edited by I. A. Kiyawa, D. S. Hegde, M.
Sagagi, and S. Yahaya, 1–18. Kano: Bayero University, Department of Economics,
Occasional Papers 1.
Myint, Hla. (1958). “‘he Classical heory’ of International Trade and the Underdeveloped
Countries.” Economic Journal 68(2): 317–37.
(1964). he Economics of the Developing Countries. London: Hutchinson, 1973
reprint.
(1977). “Adam Smith’s heory of International Trade in the Perspective of Economic
Development.” Economica 44: 231–48.
Ohadike, Don C. (1999). “‘When the slaves let, the owners wept’: Entrepreneurs and
Emancipation among the Igbo People.” In Slavery and Colonial Rule in Africa,
edited by Suzanne Miers and Martin A. Klein, 189–207. London: Frank Cass.
Olukoju, Ayodeji. (1992). “Elder Dempster and the Shipping trade of Nigeria during the
First World War.” Journal of African History 33(2): 255–71.
(2001–2). “‘Getting too great a grip’: European Shipping Lines and British West
African Lighterage Services in the 1930s.” Afrika Zamani 9 &10: 19–40.
Oroge, E. Adeniyi. (1985). “Iwofa: An Historical Survey of the Yoruba Institution of
Indenture.” African Economic History 14: 75–106. Reprinted in Pawnship, Slavery,
and Colonialism in Africa, edited by Paul E. Lovejoy and Toyin Falola, 325–56.
Trenton, NJ, 2003.
Osborn, Emily Lynn. (2004). “‘Rubber Fever,’ Commerce and French Colonial Rule in
Upper Guinée, 1890–1913.” Journal of African History 45(3): 445–65.
Reid, Richard. (2002). Political Power in Pre-Colonial Buganda: Economy, Society and
Warfare in the Nineteenth Century. Oxford: James Currey.
Roberts, Richard. (1996). Two Worlds of Cotton: Colonialism and the Regional Economy
in the French Soudan, 1800–1946. Stanford, CA: Stanford University Press.
Rodney, Walter. (1972). How Europe Underdeveloped Africa. London: Bogle-
L’Ouverture.
Ruf, François. (1995a). Booms et crises du cacao: les vertiges de l’or brun. Montpellier:
Ministère de la Coopération CIRAD-SAR and Karthala.
(1995b). “From Forest Rent to Tree-Capital: Basic ‘Laws’ of Cocoa Supply.” In
Cocoa Cycles: he Economics of Cocoa Supply, edited by François Ruf and P. S.
Siswoputranto, 1–54. Cambridge: Woodhead.
320 Austin

Salau, Mohammed Bashir. (2010). “he Role of Slave Labor in Groundnut Production in
Early Colonial Kano.” Journal of African History 31(2): 147–65.
Searing, James F. (2002). “God Alone Is King”: Islam and Emancipation in Senegal: he
Wolof Kingdoms of Kajoor and Bawol, 1859–1914. Portsmouth, NH: Heinemann.
Sender, John and Sheila Smith. (1986). he Development of Capitalism in Africa. London:
Methuen.
Smith, Sheila. (1976). “An Extension of the Vent-for-Surplus Model in Relation to Long-
Run Structural Change in Nigeria.” Oxford Economic Papers 28(3): 426–46.
(1980). “he Ideas of Samir Amin: heory or Tautology?” Journal of Development
Studies 17(1): 5–21.
Suret-Canale, Jean. (1971). French Colonialism in Tropical Africa 1900–1945. London:
Pica Press, (French original, 1964).
Swindell, Kenneth and Alieu Jeng. (2006). Migrants, Credit and Climate: he Gambian
Groundnut Trade, 1834–1934. Leiden: Brill.
Szereszewski, R. (1965). Structural Changes in the Economy of Ghana, 1891–1911.
London: Weidenfeld and Nicolson.
Tabutin, Dominique and Bruno Schoumaker. (2004). “La démographie de l’Afrique au
sud du Sahara des années 1950 aux années 2000: Synthése des changements et bilan
statistique.” Population 59: 521–620.
Teal, Francis. (2002). “Export Growth and Trade Policy in Ghana in the Twentieth
Century.” World Economy 25(9): 1319–37.
hompson, Alvin O. (2006). Economic Parasitism: European Rule in West Africa 1880–
1960. Cave Hill, Barbados: University of the West Indies.
Tosh, John. (1978). “Lango Agriculture during the Early Colonial Period: Land and
Labour in a Cash-Crop Economy.” Journal of African History 19: 415–39.
(1980). “he Cash-Crop Revolution in Tropical Africa: An Agricultural Reappraisal.”
African Afairs 79: 314, 79–94.
Usoro, E. J. (1974). he Nigerian Palm Oil Industry. Ibadan: Ibadan University Press.
Warren, Bill. (1980). Imperialism: Pioneer of Capitalism, edited by John Sender.
London: NLB.
Watts, Michael. (1983). Silent Violence: Food, Famine and Peasantry in Northern Nigeria.
Berkeley : University of California Press.
 10

Reinventing the Wheel: he Economic Beneits

of Wheeled Transportation in Early Colonial
British West Africa

Isaías Chaves, Stanley L. Engerman, and James A. Robinson

10.1 INTRODUCTION
Today sub-Saharan Africa is the poorest part of the planet. hough scholars
and researchers debate just when the gap between Africa and the rest of
the world developed, it is clear that Africa lagged behind Eurasia in terms
of many of the key building blocks of economic growth. One can see this
in the factors that go to determine income, for example literacy and human
capital, but it is perhaps most evident in technology. he basis of the mod-
ern economic growth that emerged in Britain in the late eighteenth cen-
tury was technological innovation, and the Industrial Revolution had itself
built on a long incremental series of innovations in agriculture, transpor-
tation, and elsewhere in the economy. Many of these innovations did not
take place in Africa. For example, outside of Ethiopia, no African country
innovated the plow. Similarly, systems of writing were largely restricted to
the same region, though also encompassing the Sudan and Somalia. Also
absent was the wheel.

Department of Economics, Stanford University, 579 Serra Mall, Stanford, CA 94305–6044.

E-mail: [email protected].
University of Rochester, Department of Economics and Department of History, Harkness
Hall, University of Rochester, Rochester, NY 14627. E-mail: [email protected].
Harvard University, Department of Government, IQSS, 1737 Cambridge Street N309,
Cambridge, MA 01238. E-mail: [email protected].
We are particularly grateful to Gareth Austin, Emmanuel Akyeampong, Joel Mokyr, and Bill
Summerhill for their detailed comments and suggestions. We also thank seminar partici-
pants at the WCFIA Conference on African Development in the Longue Durée at Akwapim,
Ghana, and the Harvard Economic History Seminar, particularly Toke Aidt, Joseph Inikori,
Noel Maurer, Joseph Miller, and Peter Temin. We also thank María Angélica Baustista and
Leander Heldring for their invaluable research assistance.

321
322 Chaves, Engerman, and Robinson

he fact that wheeled transportation was not used in sub-Saharan

Africa until the early colonial period is paradoxical because it is well estab-
lished that African societies knew about the wheel from the early modern
period onward.1 hey did not have to reinvent the wheel, only adopt it.
Law (1980) documents many cases where Europeans gave gits of wheeled
transportation to diferent African kings. Wheeled carriages were in use
in Dahomey from at least the eighteenth century and were even produced
there. Nevertheless, wheeled vehicles did not spread out of ceremonial uses
with the exception of a small amount of military use.
Why did African societies not adopt a technology that seemed to prom-
ise huge economic beneits in terms of reduced costs of transportation?2
hat such cost savings existed was certainly believed by early colonial oi-
cials and Europeans of the nineteenth century who noted this failure to
adopt the wheel. In the absence of wheeled transportation, the majority of
goods were transported by head porterage. British colonial diplomat Sir
Gerald Portal noted in 1903:

As an animal of burden man is out and out the worst. He eats more, carries less,
is more liable to sickness, gets over less ground, is more expensive, more trou-
blesome, and in every way less satisfactory than the meanest four-footed crea-
ture that can be trained, induced, or forced to carry a load. (Quoted in Cliford
1920, 151)

he consensus view about the absence of wheeled transportation on the

development of West Africa is well summed up by the 1926 report that
Secretary of State for the Colonies William Ormsby-Gore made to
Parliament. Commenting on transportation diiculties in various parts of
Nigeria where he visited, Ormsby-Gore wrote:

he Province of Ogoja contains an estimated population of between 600,000 and

700,000 people who are producing little or nothing for export, and a low standard
of life obtains. Until Ogoja is opened up by a network of roads . . . there can be little
trade with its consequent stimulus to production, and the real development of the
Province has not yet begun. (Ormsby-Gore 1926, 24–5)

1
See Piggott (1983) for a history of the use of wheeled transportation and Bulliet (1975) for
how the introduction of the camel into the trans-Saharan trade led to the abandonment of
the use of wheeled transportation on that route.
2
As Goody (1971) and Austen and Headrick (1983) point out, this does not exhaust the
puzzles surrounding the non-adoption of the wheel in Africa because neither potters’
wheels nor spinning wheels were adopted either.
 Reinventing the Wheel 323

Later he noted:
In British West Africa there is still too much of the most obsolete and expensive form
of transport. I refer to the wide use of head porterage. (Ormsby-Gore 1926, 29)

Indeed, as recently as 1980 one could read in a textbook treatment:

At irst, head porterage had to be used for carrying imports; and palm oil was sent
to the coast by the curious and expensive method of barrel-rolling. Porterage was a
social evil, a political danger and an economic waste. (Church 1980, 152)

he use of such an incredibly labor-intensive system of transportation

seems all the more puzzling when one considers that Africa has historically
been very labor scarce (Austin 2008; Herbst 2000). Portal’s view was widely
shared by early colonial oicials, who not only took it for granted that the
absence of wheeled transportation was disastrously ineicient, but also rou-
tinely produced numbers to prove it.
Most of these oicials gave no explanation for why Africans chose not to
adopt a technology that they regarded as massively superior to the alterna-
tives in use. he irst attempt to provide such an explanation appears to
be McPhee (1926). hough Portal noted how unhealthy human porters
were, McPhee in essence argued that, at least in the forest zone of Africa,
draught animals could not be used because of the presence of the tsetse
ly. his, he claimed, made wheeled transportation, which would normally
have depended on draught animals, uneconomical. His argument is rather
equivocal, however, because the relevant chapter of his book attempts to
show in some detail that road and rail transportation were far superior to
head porterage. He seemed to take it for granted that Africans could not
have built motorable roads or railways. He also explicitly pointed out, on
the basis of Lugard’s 1919 report on the amalgamation of northern and
southern Nigeria for northern Nigeria, that:
Strangely enough, although there are over 3,000,000 cattle, 176,000 donkeys, 113,000
horses and 4,000 camels, yet such a thing as a cart may be said to be unknown.
(McPhee 1926, 121)

herefore, though the argument about the tsetse ly is obviously relevant, it

can at best only be a very partial solution to the puzzle. McPhee’s discussion
was augmented by Hopkins (1973), who argued that the cost of building
roads through the forest zones was so high that this also made wheeled
transportation uneconomical. Ogunremi (1975) also claims that head por-
terage was economically eicient because labor was not really scarce, and
he claims that the calculations made by Lugard and others are misleading
324 Chaves, Engerman, and Robinson

because they ignore the huge capital costs involved in constructing railways.
In essence, McPhee to an extent, but certainly Hopkins and Ogunremi,
respond to the puzzle of the non-adoption of wheeled transportation by
asserting that it is not a puzzle and that in fact it was an economically ratio-
nal decision given the circumstances.3
hese existing explanations present obvious problems. First, none of
them is based on any real calculation or what was or was not economically
rational. Second, while it is clearly correct that the impact of the tsetse ly
made it diicult to use draught animals in large parts of central Africa,4
wheeled transportation was not used in areas where there was no tsetse
either. his is true not just in southern Africa, but also in the Sahel or north-
ern Nigeria. hird, as Portal’s remark notes, humans were very unhealthy as
well, so what is relevant is the health of animals compared to humans, not
the absolute health of animals in the tsetse zone. Finally, Hopkins’s claim
that wheeled transportation was not adopted in the forest zone because
roads were uneconomical to build runs into the problem that African pol-
ities in the forest zone did indeed build such roads. Most notably, Wilks
(1989) discusses in detail the great roads of the Asante Empire in the Gold
Coast in the irst half of the nineteenth century,5 and Reid (2002) does the
same for the roads built by the Buganda state. Yet neither the Asante nor the
Buganda states used wheeled transportation.
In this chapter, we conduct the irst attempt, to our knowledge, at bring-
ing systematic evidence to bear on the question of whether wheeled trans-
portation was economically rational in sub-Saharan Africa. We focus on
the three British West Africa colonies of the Gold Coast, Nigeria, and Sierra
Leone, largely because the British colonial state recorded in great detail the
costs of constructing and maintaining diferent transportation systems in
these colonies. hough we examine the eiciency of various types of wheeled
transportation, including carts and motor vehicles, the bulk of our analysis
focuses on railways. Railways of course embody not just wheels, but also

3
See Basu and Weil (1998) and Acemoglu and Zilibotti (2001) for formalizations of the idea
of appropriate technology.
4
Witness, for instance, the enormous and costly lengths to which the Oyo Empire had to go
to keep its cavalry safe from the ravages of tsetse (see Law 1977).
5
he early colonial oicials who deplored the backward state of transportation in Africa
seem not to have been aware of these roads. In his discussion, McPhee, for example, notes:
“At the beginning of the last century no proper roads existed anywhere in West Africa. . . .
he earliest roads, not much better than rough tracks which were liable to be obliterated
within a year by forest growths, were military roads. hus Sir Garnet . . . Wolsley con-
structed a road into the heart of Ashanti during the War of 1873–1874. hroughout the
whole century very little progress was made” (1926, 106–7).
 Reinventing the Wheel 325

other technologies, such as iron smelting and casting and steam engines.
Our focus is motivated by the fact that we have from colonial records very
comprehensive information on the amount of goods and passengers that
the railways carried, which we do not have for any other method of trans-
portation. We also know a great deal about the capital expenditures associ-
ated with railway building, the neglect of which has been used as a critique
of earlier estimates that head porterage was ineicient. In addition, this
focus allows us to make comparisons with the rich literature in economic
history that has examined the economic impact of railways. Our methodol-
ogy is the canonical one based on that of Fogel (1964).6
Our basic indings are very contrary to the conventional wisdom. First,
all forms of wheeled transportation were economically rational in the sense
that they generated positive social savings, that is, that their adoption would
have increased national income. In the case of railways, the social savings
for goods traic as a percentage of GDP range from a low of 0.8 percent in
the Gold Coast in 1909 to a high of 7.8 percent for the same colony in 1934–
5. For passenger traic, the numbers are smaller, basically zero. Second,
and more important, the social rate of return on railway construction was
incredibly high, a calculation that explicitly takes into account the capital
expenditures. his rate of return was around 100 percent in Nigeria, imply-
ing that the social savings in any year were equivalent to the entire capital
outlays until then. Elsewhere they were lower, but were close to 50 percent
for both the Gold Coast and Sierra Leone for much of the period. Our esti-
mates therefore contradict the idea that wheeled transportation was not
adopted in sub-Saharan Africa because it was an inappropriate technology.7
Indeed, quite contrary to the conventional wisdom that railways were not
appropriate because they were too expensive to build in African conditions,
we ind that they were cheap to build – their cheapness, in fact, causes the
astonishing social rates of return calculations we present. his is consis-
tent with the recent indings of Jedwab and Moradi (2011), who ind that
colonial railway construction in the Gold Coast had a powerful efect on
exports and development.8

6
Tsey (1986), in his analysis of the expansion of railways in the colonial Gold Coast,
observed that one could undertake such an exercise but he chose not to do so.
7
At some level this is not very surprising. Technological diferences today, as captured by
total factor productivity, are at the heart of diferences in income per capita between Africa
and the rest of the world (Acemoglu, Johnson, and Robinson 2001; Hsieh and Klenow
2010; Hall and Jones 1999), but few believe that such diferences are eicient.
8
See also the innovative work by Donaldson (2010), who inds large positive efects of colo-
nial railroads in India using a general equilibrium model.
326 Chaves, Engerman, and Robinson

If wheeled transportation was economically eicient and generated very

high social rates of return, why did Africans not adopt it? For the case of
railways, we can identify three types of explanations. he irst is the obvious
point that it was very diicult to construct such large public works as rail-
ways without suicient political centralization. We illustrate this argument
in Sierra Leone. he second comes from thwarted attempts by a mission
from the Asante state in Ghana to contract British engineers in London
to build a railway in 1895: European powers had an interest in restricting
technology adoption by African polities. Asante’s attempt to build a railway,
part of a larger program of modernization embarked on ater 1874, was
blocked by British colonial secretary Joseph Chamberlain. he likely expla-
nation is that Britain did not want autonomous modernization of African
polities. African states considering adopting railways invariably needed to
rely on foreign capital, engineers, and expertise. his was normal during the
nineteenth century: railways in Latin America and the Middle East were
built with foreign capital and expertise. But at the time of the Scramble for
Africa, European powers wished to control this type of technology adop-
tion by African polities because it made them harder to control, a kind of
incentive that was absent in Latin America and the Middle East. he third
explanation comes from the rich evidence on the one independent African
polity that actually built a railway, Ethiopia. Namely, African polities that
could have proitably adopted railway technology had a political incentive
not to because doing so might hasten the loss of sovereignty to Europeans.
Here the evidence clearly shows that the Ethiopians were very concerned
that constructing a railway from Djibouti up to Addis Ababa would precip-
itate a process of colonial domination and conquest. Similar mechanisms
were at work elsewhere, for example in the Sultanate of Zanzibar.
his chapter proceeds as follows. In the next section, we discuss some
of the estimates of the costs of transportation that colonial oicials and
administrators made during the early twentieth century. his evidence,
which certainly suggests that head porterage was ineicient, is typically
very incomplete, because it is not clear exactly what is involved in the cal-
culations or how representative any of the information is. In Section 10.3,
we discuss the data that are available to undertake cost and beneit calcula-
tions in the British colonies of the Gold Coast, Nigeria, and Sierra Leone.
Section 10.4 then uses this data to provide estimates of the social savings
from the introduction of the railways in these three colonies, and it also
provides more speculative estimates of the social savings associated with
simpler forms of wheeled transportation. Section 10.5 then focuses on the
social rate of return of the railways. Section 10.6 discusses in more detail,
 Reinventing the Wheel 327

and presents evidence for, our explanation for why African states did not
adopt economically superior transportation technology during the nine-
teenth century. Section 10.7 concludes.

10.2 CONTEMPORARY DISCUSSION AND

EXISTING EVIDENCE
British colonial oicials did not doubt that the lack of wheeled transpor-
tation was a major impediment to economic progress in West Africa. To
demonstrate this, they produced a whole range of diferent numbers, which
were then constantly repeated over the years. Some of these are summa-
rized in Table 10.1.
Unfortunately, in all the cases we have found, it is never clear exactly how
these estimates were constructed or what considerations went into them.
Ormsby-Gore’s numbers, for example, were introduced by noting that:
At Zaria, in Northern Nigeria, I was provided with some carefully compiled igures
regarding the cost of diferent forms of transportation per ton-mile. Head porterage
in an area where labour is plentiful and cheap works out at 2/6 per ton-mile; motor
transport at 1/- per ton-mile; donkey transport at 11d; camel transport at between
9d and 10d; while the railway takes baled cotton from Zaria to Lagos at under 2d
per ton-mile. (1926, 29–30)9

hese numbers were widely reproduced, for example in Hailey’s famous

survey (1938, 1540). Yet Ormsby-Gore provided no further information
about who gave him these igures or what sources they used to calculate
them. Clearly, taken at face value, these numbers suggest that head porter-
age was extremely ineicient. According to these numbers, while the cost
of shipping freight by railway was two pence per ton-mile, the cost of head
porterage was two shillings and six pence, or iteen times more!
Sir Frederick Lugard similarly regarded head porterage as extremely
ineicient:
For uncounted centuries the African has been his own beast of burden, and a sim-
ple calculation shows that the cost of land transport by such means with a wage rate
of 9d. per day is about 3 shillings per ton mile. (1922, 461)

his number is also very widely reproduced. In a footnote, Lugard noted

that this calculation was based on assuming that a porter carried sixty-ive
pounds and could walk twelve miles a day and was then adjusted upward to

9
he notation d. means pence, s. means shillings; there were twelve pence in a shilling and
twenty shillings in a pound.
328 Chaves, Engerman, and Robinson

Table 10.1. Contemporary estimates of the relative cost of diferent

methods of transportation (cost in shillings and pence per ton-mile)

House of Commons Ormsby-Gore Ogunremi

(1909) (1926) (1982)
Cask Rolling 1s.21/2 d. – 1s.11d.
Hand Trucks 1s. 10d
Head Porterage 3s.1d – 5s. 2s.6d. 1s.10d. – 2s.
Motor Lorry 1s.8d. 1s.
Railway 2d.
Donkey 11d. 9d. – 10d.

allow for sickness and supervision. He also observed that “for bulky loads
the cost is much more.” Elsewhere Lugard remarked that “a railway train
of average capacity and engine power will do the work of 13,000 carriers at
one twentieth of the cost” (1922, 462–3).
Other calculations suggest similar things about the relative eiciency of
diferent forms of transportation. For example, the numbers taken from
the House of Commons (1909) come from an extensive survey of methods
of transportation in the entire British Empire undertaken by the secretary
of state for the colonies. In this survey, carried out in 1907, the governors
of the diferent colonies were requested to provide information to a stan-
dardized set of questions about the nature of transportation in their colo-
nies. he information provided for the Gold Coast, presented by E. F. W.
Wilkinson, acting director of public works, suggests that transportation by
head porterage cost between three shillings and one pence to ive shillings
per ton-mile (House of Commons 1909, 43). hese igures are as much as
twice those for Zaria, perhaps indicating the relative labor scarcity of the
Gold Coast during that period. hey are, however, consistent with Lugard’s
numbers.
For our purposes at the moment, the most interesting comparison is
between head porterage and the railways.10 hese ot-quoted numbers sug-
gest that head porterage was about iteen to twenty times as costly as the
railway. Yet it is not clear how these numbers were constructed. Most cru-
cially, it is not clear whether they factor in the large ixed cost of constructing

10
Forms of transportation other than head porterage were very important in diferent pre-
colonial contexts. For instance, Hill (1972) describes how tobacco produced in Katsina in
northern Nigeria was shipped as far south as Ilorin using donkeys, and the great kola trade
between Nigeria and Ghana was carried out mostly by donkey (Lovejoy 1980).
 Reinventing the Wheel 329

the railway or whether they are based just on the variable cost. Because it is
precisely the large cost of constructing modern transportation systems in
West African conditions that Hopkins (1973) argues made them economi-
cally irrational, we need to assess the eiciency of diferent systems of trans-
portation properly taking these costs into account – something we will do
by calculating the social rate of return.

10.3 THE DATA

10.3.1 Railways
he surveys for railway construction were carried out at diferent dates in
British West Africa. For Sierra Leone, they were undertaken in 1893–4, and
for Lagos, they were undertaken in early 1893 during the tour of Governor
Sir Gilbert Carter. Construction on the Sierra Leone and Lagos railways
started in 1896, while construction on the Gold Coast Railway started in
1898 (see Shelford 1920). he Sierra Leone railway ran southeast from
Freetown and reached Pendembu in 1908. he irst railway in the Gold
Coast went inland from the port of Sekondi toward the Asante goldields,
and it reached Kumase in 1903. he second line, linking Kumase to Accra,
was started in 1909 and completed in 1923. he irst railway line in Nigeria
started at Lagos and reached Jebba on the Niger river in 1909. he second
railway linked Kano to the port of Baro on the Niger, and it was completed in
1911. All the lines were built and operated by the colonial governments.11
Our main source of data on the costs of constructing, maintaining, and
running the railway, the amount of freight hauled, the number of passen-
gers carried, and the revenues generated from freight and passengers is the
various reports of the colonial governments to London, particularly the
reports of the railways and transportation departments. We focus on three
dates, 1909 (1911 for Sierra Leone because earlier reports did not present
the appropriate data), 1924/25, and 1934/35. he inancial year for all colo-
nies started on January 1 for 1909/11, but it switched to April 1 for the Gold
Coast and Nigeria during the 1920s and 1930s – whence the fact that for
our latter two colonies and dates, the data on railways straddles two years.
Our basic data on railways for the Gold Coast comes from the 1909 “Report
of the General Manager upon the Government Railways,” and the Gold
Coast Railway Administrative Reports for 1924–5 and 1934–5. For Nigeria,
we have the “Lagos Railway Annual Report 1909,” the “Nigerian Railway
11
For a contemporary discussion of the pros and cons of government versus private owner-
ship and operation of the railways, see House of Commons (1924).
330 Chaves, Engerman, and Robinson

and Udi Coalmines Administrative Report for the Year ending 31st March
1925,” and for 1934/35 the “Annual Report on the Government Railway
and Colliery of Nigeria for the Financial Year ending 31st March 1935.”
Finally, for Sierra Leone, we have the “Report on the Transport Department
for the Year 1911” and the Administrative Reports of the Railway for 1925
and 1935.
hese sources of information give us extensive data to calculate the
amount of freight and passengers the railways carried in these three years.
he reports also provide information on current receipts and expenditures,
as well as capital outlays to date on railway construction and maintenance.
A potentially important and obvious advantage of the railway is that it
moved people much faster than they could have moved by walking. We
can estimate the social savings associated with this change, because for
most years the railway reports tell us the total number of passenger miles
traveled. We can estimate how long it took to travel this number of miles
if we know how fast the trains went, and also how fast it took to walk.
Unfortunately, we have only discovered a railway timetable for Nigeria in
1921 (Burns 1921), but not for either the Gold Coast or Sierra Leone, and
the reports of the Railway Department never mention how long it took
trains to travel between stations (though they do assiduously report the
percentage that were late). he data in Burns (1921) implies an average
speed of iteen miles per hour. As a reality check, a rough calculation of
the speed of the trains can be backed out of the travel account of Alldridge
(1910), who visited Sierra Leone in the early twentieth century. Alldridge
traveled on the train from Freetown to Bo leaving at seven in the morning
and arriving at ive in the evening. Because the distance between the sta-
tions is 136 miles, we can say that the average speed in Sierra Leone was
13.6 miles per hour. his appears to be very slow, but it is partly justiied by
the fact that the Sierra Leone railway had a narrow gauge, which possibly
meant that it had to go slower than one would have expected. We have no
comparable account of travel from the Gold Coast or Nigerian railways.
hese used broader gauges, so the iteen miles per hour in Nigeria seems
completely reasonable. hus we choose iteen miles per hour as a conser-
vative estimate. For an estimate of how fast someone could walk in this
region, we use Wilks (1989, chapter 1), who has an extensive discussion of
travel times between diferent parts of Asante during the nineteenth cen-
tury. All his estimates are close to iteen miles per day, which is the modal
estimate of how far a head porter could walk, and we shall take this as the
relevant travel speed for passengers on foot. his implies a speed of two
miles per hour.
 Reinventing the Wheel 331

To value the time saved by travel on the railway, it is natural to take the
wage rate as the opportunity cost of time. Exactly what the correct wage to
use is depends on who the traveler was. For example, in 1924–5 in the Gold
Coast, of the 1,487,164 passengers who traveled by train, 14,851 traveled
irst class, 21,988 traveled second class, and the rest traveled third class. We
do not have the information necessary to decompose the total passenger
miles into components of diferent groups, so we could assume instead that
the passenger miles were allocated in proportion to the numbers of each
class (i.e., on average each type of person traveled the same distance). his
implies that of the 41,751,573 total passenger miles, 416,936 (1%) were irst
class passenger miles, 617,305 (1.5%) were second class passenger miles,
and the rest (97.5%) were third class passenger miles. To value the time
saved by the introduction of the railway, we need to impute some oppor-
tunity cost of time to these diferent groups. For third class passengers, we
could choose the unskilled wage rate of 9d. per day or 1d. per hour, assum-
ing a nine-hour workday. Skilled wages in the Blue Books range from 2s.6d.
to 5s. To calculate the value of time for irst and second class passengers, we
value the hours saved at the rate of 5s. per day. Unfortunately, this (1924–5)
is the only time for which our sources document the breakdown among the
diferent classes of passengers. And yet, as is evident, irst and second class
travel was relatively so unimportant that year-to-year changes in the com-
position of passenger miles are unlikely to drive our results. herefore, we
proceed with the simple assumption that time can be valued at the unskilled
wage rate. his is 9d. per day for all the colonies for the irst two dates and
then 12.d., 8.d., and 11.d. in the mid-1930s for the Gold Coast, Nigeria, and
Sierra Leone, respectively.

10.3.2 Head Porters

To compare to the cost of using head porters, we need data on how much a
head porter could carry and the rate of pay. hough information on this is
much less systematic than the data from the railways, the consensus seems
strong on what the right numbers are. Ormsby-Gore notes:
[T]here is a considerable body of labour temporarily employed on road and rail-
way construction. he supply of voluntary labour for the latter purposes has always
proved inadequate in Nigeria, and recourse is had to compulsory or “enlisted” –
sometimes called “political” – labour for these essential public works and services.
All the railways and most of the roads in Nigeria have involved the use of this com-
pulsory labour. . . . Such compulsory labour is recruited by the native authorities. It
is only called upon to work for a deinite period, usually, and never more than, one
332 Chaves, Engerman, and Robinson

month at a time. It is paid, usually at a rate of 9d. per day. . . . Unpaid compulsory
labour legalized under the Roads and Rivers Ordinance of Northern Nigeria is only
used for keeping clean roads and rivers within local boundaries when called upon
to do so by the Resident. (1926, 133)

Nine pence a day is the igure that is widely quoted from all over British
West Africa for the cost of a head porter from around 1910 until the middle
of the 1920s, though head porterage was surely much less prevalent in 1925.
Ormsby-Gore reported a higher number from his visit to Sierra Leone:
1s.3d. per day in the Colony (the capital Freetown and its environs) and
1s. per day in the Protectorate (the hinterland and interior of what is now
Sierra Leone) (1926, 58).
Just as there is a consensus on wages in the primary sources, there is
also a consensus that a head porter could carry about sixty pounds. Indeed,
this seems to have been more or less the oicial load colonial oicials used
when they hired porters for government work (Ogunremi 1975, 47). he
numbers on how far a fully loaded porter could walk in a day do vary, with
perhaps iteen miles being the consensus. For instance, in the Sierra Leone
case, Ormsby-Gore reported that a porter usually carried forty-ive to ity
pounds of weight and could walk twelve to iteen miles per day (1926, 58).
A command paper from the House of Commons on mechanized transpor-
tation suggested that in the Gold Coast:

he motor lorries carry about 1 ton to 2 1/2 tons; a cask of palm oil weighs 17
3/4 cwt.; a cask of cocoa weighs about 12 cwt.’ a hand truck carries from 15 to 20
cwt with 6 to 8 men to a truck; head loads are about 60 lbs. (House of Commons
1909, 42)

Moreover, “casks, hand trucks and head loads get over 20 miles per day.” he
chief commissioner for the Northern Territories of the Gold Coasts reported
that, although “native rates [are] impossible to gauge,” the government paid
“10d. a day for loads of 50 to 60 lbs., 1d. a day of which goes to the chief who
provided the carriers.” As for how far a porter could walk, the commissioner
noted, “At present natives are content to do 10 to 15 miles a day.” Northern
Nigeria presents a similar situation, with sixty pounds mentioned as the
normal head load and wages for hammock men listed at nine pence and
those for a laborer at nine pence to one shilling per day (House of Commons
1909, 22–3). In Sierra Leone, “Head loads 60 to 100 lbs. Hand carts 1 ton to
30 cwts” (House of Commons 1909, 102). hough this report provided no
information on the wage paid to head porters in Sierra Leone, it noted that a
barrel roller, a similarly unskilled worker, was paid one shilling per day. his
igure is identical to the wage Ormsby-Gore recorded for a head porter.
 Reinventing the Wheel 333

To judge if these wage rates are reasonable, we can compare them to

other readily available information. For example, Oyemakinde notes that
workers who were recruited by compulsion to build the railways in north-
ern Nigeria between 1911 and 1915 were paid nine pence per day, while
in Yorubaland, where workers freely took up such employment, they were
usually paid one shilling per day (1974, 318). he House of Commons com-
mand paper detailing the costs of railway construction in the Gold Coast,
Nigeria, and Sierra Leone does report the daily wage rate of unskilled work-
ers used in railway construction (House of Commons 1904). hese were ten
pence a day in Sierra Leone, one shilling a day in Lagos, and one shilling
three pence a day in the Gold Coast.
All in all, these scattered numbers are quite consistent with each other,
and for the period around 1909 they suggest that Lugard’s number of three
shillings per ton-mile is a reasonable igure for the cost of head porterage,
at least to the extent that he corrected it for the health of the porters, which
would explain why the cost is higher than other simple calculations. Ormsby-
Gore’s 1926 igures suggest that this number is reasonable for 1924–5 as well.
For our 1935 estimates, we do not have contemporary information on the
rates of pay of head porters. However, the Blue Books for the colonies report
unskilled workers’ wages. Because the wage rates we have for head porters
in earlier periods correspond closely to the wages for unskilled workers as
stated in Blue Books, we can use the latter data to get a counterfactual wage
for head porterage in 1935. In the Gold Coast, this wage ranged from nine
pence to iteen pence; in Nigeria, the range three pence to one shilling per
day is given; and in Sierra Leone, unskilled workers’ wages are reported as
eleven pence. We therefore used the wage rates of twelve pence for the Gold
Coast, eight pence for Nigeria, and eleven pence for Sierra Leone.
Rather than use Lugard’s three-shilling estimate, we assume in our calcu-
lations for 1909/11 and 1924/25 that head porters could walk iteen miles
in a day, carry sixty pounds, and were paid nine pence in all the colonies.
hese numbers imply a cost of one shilling and ten pence per ton-mile for
head porterage. For 1934/35, this cost increases in the Gold Coast by one
third to two shillings and six pence per ton-mile, it falls to one shilling and
eight pence per ton-mile in Nigeria, and it increases to two shillings and
three pence per ton-mile in Sierra Leone.
An important issue that can limit the applicability of the social savings
approach to colonial Africa is the nature of the labor market. Coerced labor
(or “political” labor, as colonial oicials called it) was used extensively to
build roads and railways; at the same time, slaves were also commonly used
as head porters, and some of these slaves were no doubt supplied by chiefs
334 Chaves, Engerman, and Robinson

to help railway and road construction as well. In the Gold Coast in particu-
lar there appears to have been a great labor shortage at the time the British
were building the railway, and the supply of Africans that were forthcom-
ing at the wages that the British were prepared to pay was insuicient to get
the work done. Colonial oicials therefore induced local chiefs to provide
labor (see Akurang-Parry 2000; Mann 1995; and homas 1973 for studies
of forced labor, and Mason 1978 and Swindell 1992 for speciic studies in
the context of railway construction all in the context of British West Africa).
One could argue that because labor could be coerced either to build and
run railways or to work as porters this should not inluence the relative
beneits of the two methods of transportation. Nevertheless, because head
porterage is much more labor intensive than railway construction, a natural
conjecture would be that the ability to repress labor would bias downward
the social savings from railways. We return to this when we discuss the
West African colonies in comparative perspective. We note, however, that
coerced labor was not everywhere used for head porterage. For example, in
East Africa, Rockel (2006) shows that there was a basically free labor mar-
ket for head porterage, so this problem may be much more important there
(see Coquery-Vidrovitch and Lovejoy, eds. 1985 for comparative studies).

10.3.3 GDP Estimates

To get some sense of how big the cost diferences between diferent methods
of transportation were, it is useful to have something to compare the costs
to. he most obvious normalization is with respect to GDP, and this is the
standard approach in the social savings literature. However, to our knowl-
edge, with the exception of Szereszewski’s (1965) pioneering construction
of GDP estimates for the Gold Coast in 1891, 1901, and 1911, no estimates
of GDP for the colonies of interest exist for this period. Szereszewski’s
approach was based on colonial Blue Books, which contain extensive infor-
mation about imports and exports and the public sector. He used a number
of assumptions to construct estimates of consumption and capital forma-
tion. Most speculatively, he also constructed estimates of the consumption
and investment made by the “native economy,” of which there are only pop-
ulation estimates during this period. Nevertheless, Szereszewski’s approach
uses the existing information in a very creative way and produces quite
believable estimates. We therefore used it to construct from the Blue Books
estimates of nominal GDP for the three colonies, for all of the relevant
periods. hese calculations may be of independent interest, and we discuss
them in detail in the appendix to this chapter.
 Reinventing the Wheel 335

10.4 WERE THE RAILWAYS ECONOMICALLY EFFICIENT?

A SOCIAL SAVINGS APPROACH
hough the calculations discussed previously are interesting, they fall very
short of a systematic treatment of the issue. Moreover, one could easily
imagine that colonial oicers, anxious to legitimize their “civilizing mis-
sion” in Africa, may have been inclined to overemphasize the technological
backwardness of Africa and therefore the advantages of “modern” methods
of transportation.
he conventional method for tackling the issues broached in this chapter
is that of social savings introduced by Fogel (1964) and Fishlow (1965). he
social savings of a given method of transportation, such as the railways, is
the diference between the actual cost of shipping the goods and people by
that method of transportation and the cost of shipping them without that
method of transportation. As Fogel (1979) points out, the social savings
associated with a particular eicient method of transportation is the loss of
national income associated with the substitution of an ineicient method
of transportation for the eicient one. In our case, we focus on the social
savings associated with using railways rather than head porterage to move
goods and people.
his approach has been heavily criticized, requires strong assumptions
about the nature of the economy, and fails to capture important impacts
of transportation innovations (see Crats 2004; Fogel 1979; Leunig 2010;
O’Brien 1977; and Summerhill 2003 for extensive discussions of the pros
and cons of this approach). For our purposes, one central problem with
this method is that it requires the assumption that one is studying a per-
fectly competitive industry in long-run equilibrium, so that price (aver-
age revenue) is equal to long-run average cost. Obviously railways, which
involve a huge ixed cost, cannot be in such an equilibrium. Moreover, in
all the West African colonies the prices charged for freight and passengers
were not determined by perfect or imperfect competition; they were set
by the colonial administration.12 Furthermore, basic social savings calcula-
tions cannot incorporate the potentially large externalities created by the
construction of the railways. It is clear from contemporary discussions that

12
Indeed, it is not just that the colonial administration regulated ton-mile prices; it also reg-
ulated other activities in ways that heavily inluenced the proitability of the railways. For
example, in 1936 the Gold Coast government passed Ordinance 38, which prevented the
carriage by roads of key export goods (such as cocoa) and key imports because road trans-
portation was diverting traic from the railways. Sierra Leone adopted a similar measure
the following year (Ordinance 6 of 1937) (see Hailey 1938, 1559–60 and Church 1956).
336 Chaves, Engerman, and Robinson

colonial oicials strongly believed these externalities were real and signii-
cant. Harry Johnston, an avid colonizer of Africa, noted in 1889 that:
here is no civiliser like the railway, and to build a railway through an uncivilized
country is to centuple its existing trade, or to create commerce if none exists: the rail-
way saps race prejudices and dissolves fanaticism. (quoted in McPhee 1926, 111)

British observers in many places echoed his views. McPhee argued that:
Slavery in Northern Nigeria found its chief buttress in the demand for cheap trans-
port in a region where animal transport was not feasible on account of the tse-
tse ly . . . the Government built railways, and slave carriage died a natural death,
because it became uneconomical. (McPhee 1926, 126–7)

Lugard observed about the construction of the railway in Nigeria that “it
has killed the slave trade” (1922, 463). Knowles extensively discussed posi-
tive externalities lowing from railway construction (1928, 138–52).
Despite these caveats, in the absence of suicient data to calibrate a gen-
eral equilibrium model, the social savings methodology does present us
with a simple method of looking at the economic impact of the railways;
and because these caveats would apply to all such studies, the comparison
between our results and those of others is in itself interesting. To apply this
methodology, we follow the simplest approach of Fogel (1964) in assuming
a zero elasticity of demand for transportation services. Maybe more impor-
tant in the present context, we assume that, in the counterfactual scenario
where the freight hauled by the railways is carried instead by head por-
terage, there is no impact on the labor market; with this assumption, we
can use the observed wages to calculate the counterfactual cost of moving
the freight by head porterage. Hence, we are assuming that labor supply is
completely elastic.

10.4.1 Results from the Gold Coast

Table 10.2 presents basic data on the Gold Coast railways for the three
dates of interest. In 1909, only 168 miles of track were open in the col-
ony, from Secondi on the coast up to the Asante capital of Kumase. By
1924–5, the railroad had expanded to 394 miles, because by then a line
linked Kumase to the colonial capital of Accra, and by 1934–5 it had grown
to 500 miles, as the grid had been extended to a number of smaller feeder
railways (Gould 1959 gives a good overview). Between 1909 and 1924–5,
the amount of freight hauled expanded dramatically, with the number
 Reinventing the Wheel 337

Table 10.2. Basic data on the Gold Coast railways

1909 1924–5 1934–5

Mile of Track Open 168 394 500
Total Freight Hauled (Tons) 56,454 796,888 730,382
Total Freight in Ton-Miles 3,763,552 43,170,885 42,952,026
Total Passengers Moved 215,729 1,487,164 1,822,093
First Class 14,851
Second Class 21,988
hird Class 1,450,325
Total Passenger Miles 41,751,573 44,704,654
Total Freight Revenues (₤) 146,845 850,238 633,525
Total Passenger Revenues ₤) 38,565 214,703 150,840
Total Expenditures (₤) 73,914 568,012 487,530
Total Capital Outlays (₤) 1,808,323 7,419,086 9,241,698

of ton-miles growing by more than 1,000 percent. From that point until
1934–5 there was a contraction, which no doubt relects the impact of the
Great Depression; the economic collapse in Britain and elsewhere severely
restricted the demand for tropical exports. he data for total passenger
miles is unfortunately incomplete and does not exist for the earlier date
(or dates near it), and we only have the breakdown of passenger traic into
diferent classes for the period 1924–5. We also record in Table 10.2 total
freight and passenger receipts, which rose sharply but then fell quite signif-
icantly between 1924–5 and 1934–5. Finally, we record total capital outlays
to date for the three dates.
Table 10.3 contains our three sets of estimates of social savings from
freight for 1909, 1924–5, and 1934–5. In all columns, the irst set of cal-
culations relates to our direct measure of the relative costs of the diferent
methods of transportation. he irst row reports total ton-miles of freight
transported in the diferent years; these numbers are taken from Table 10.2.
he second row contains the information from Table 10.2 on total freight
revenues. he third row shows how much it would have cost to move the
observed ton-miles of railway freight with head porters, given our assump-
tions that a head porter walked iteen miles a day and could carry sixty
pounds (there are 2,240 lbs in a British ton). As discussed earlier, for the
irst two columns we assume a wage rate of nine pence per day, increasing
to twelve pence a day in 1934–5. he social savings from railways are then
simply the diference between lines 3 and 2. In all cases, these are positive,
338 Chaves, Engerman, and Robinson

Table 10.3. Estimates of social savings from freight in the Gold Coast

1909 1924–5 1934–5

Total Freight Ton-Miles 3,763,552 43,170,885 42,952,026
Total Freight Revenues (₤) 146,845 850,238 633,525
Total Cost of Head Porterage (₤) 344,992 3,957,331 5,369,003
Social Saving (₤) 198,147 3,107,093 4,735,478
GDP (in current ₤) 26,266,269 53,108,142 61,068,569
Freight Social Saving (as a % of GDP) 0.8 5.9 7.8

suggesting that the introduction of the railways did indeed increase national
income. To get some relative quantitative sense of how big these numbers
are, we record our estimates of nominal GDP in row 5, and then we present
measured social savings as a percentage of GDP in row 6. In 1909, as the
railway was getting under way, social savings were negligible, but by 1924–5
they were up to 5.9 percent of nominal GDP, with the number increasing
further in 1934–5 to 7.8 percent of GDP. his latter increase shows up in
the data mostly as a consequence of the assumed rise in head porter wages
over time. he numbers are not disproportionately large but are neverthe-
less signiicant.
In Table 10.4, we calculate the social savings from passenger transpor-
tation. For 1924–5 and 1934–5, the colonial sources record the total pas-
senger miles traveled. he second row then calculates the total time used
up in moving these passengers by rail based on the assumption that trains
moved at iteen miles per hour. We then compare this to the amount of
time it would have taken for passengers to walk this far on foot, rather than
travel by train, on the assumption that they could walk at two miles per
hour. he third row shows how long it would have taken by foot – clearly
much longer than it would have taken by rail. To calculate the social sav-
ings, we then price the time diference between the two modes of travel
using diferent wage rates, 1d. per hour for 1909 and 1924–5 and 1.33d.
per hour for 1934–5, as we mentioned previously. Row 5 shows the value
of the time saved by moving people by rail instead of having them walk. To
calculate the social savings, we compare this to total passenger revenues in
the next line. It is clear that revenues were actually greater than the value
of the time saved, suggesting that the beneit of moving people by rail did
not compensate for the extra cost of doing so. Hence the social savings are
negative.
 Reinventing the Wheel 339

Table 10.4. Estimates of the social savings from passengers in the Gold Coast

1909 1924–5 1934–5

Total passenger miles 41,751,573 44,704,654
Total time required to travel by rail 2,783,438 2,980,310
(at 15 miles per hour) (hours)
Total time required to travel by foot 20,875,786 22,352,327
(at 2 miles per hour) (hours)
Travel time saved (hours) 18,092,348 19,372,017
Value of Time Saved (₤) 85,939 118,653
Total Passenger Revenues (₤) 38,565 214,703 150,840
Passenger Social Saving (₤) −128,764 −32,186
GDP (in current ₤) 26,266,269 53,108,142 61,068,569
Passenger Social Saving (as a % of GDP) −0.24 −0.05

Table 10.5. Basic data on the Sierra Leone railways

1911 1925 1935

Mile of track open 255.5 338 341
Total freight hauled (tons) 55,540 72,298 71,628
Total freight in ton-miles 5,971,693 9,437,472 10,569,611
Total passengers moved 339,332 587,944 450,707
Total passenger miles 11,047,266 11,377,080
Total Freight Revenues (₤) 82,086 167,687 107,868
Total Passenger Revenues (₤) 22,644 52,305 28,383
Total Expenditures (₤) 69,503 176,482 128,862
Total Capital Outlays (₤) 989,194 1,359,680 6,994,715

10.4.2 Results from Sierra Leone

Table 10.5 mimics for Sierra Leone the structure of Table 10.2. By 1911,
the railway grid in Sierra Leone was 255.5 miles in length, longer than the
amount of track open at the time in the Gold Coast, but it grew much more
slowly than the Gold Coast’s thereater. Row 3 shows that the growth in
freight haulage was also much smaller in Sierra Leone. hough in 1911 the
Sierra Leone railway hauled more freight than the Gold Coast railway, by
1925 it was hauling less than a quarter of the amount hauled by the Gold
Coast railway. his is a relection of the rapid expansion of the cocoa econ-
omy in the Gold Coast over this period, a development that had no analogy
in Sierra Leone. he Sierra Leone railway also moved far fewer passengers,
340 Chaves, Engerman, and Robinson

Table 10.6. Estimates of the social savings from freight in Sierra Leone

1911 1925 1935

Total Freight Ton-Miles 5,971,693 9,437,472 10,569,611
Total Freight Revenues (₤) 82,086 167,687 107,868
Total Cost of Head Porterage (₤) 547,405 865,102 1,189,081
Social Saving (₤) 465,319 697,415 1,081,213
GDP (in current ₤) 25,444,186 32,818,287
Social Saving as a % of GDP 2.7 3.3

about one-third by the mid-1920s, a pattern that, like the relative freight
haulage, continued into the 1930s. Unsurprisingly, freight revenues were
much lower in Sierra Leone.
hese diferences partially show up in Table 10.6, in the sense that total
social savings in pounds are considerably lower by the mid-1920s in Sierra
Leone and also make up one-half the size relative to GDP (we do not
have estimates for GDP in Sierra Leone in 1911) as compared to the Gold
Coast.
In Table 10.7, we move to the social savings for passengers in Sierra
Leone. he indings here are very similar to those for the Gold Coast. Social
savings are negative but hardly distinguishable from zero, as was the case in
Table 10.4. he methodology is identical in both cases, except that we apply
the wage rate of 1.22d. per hour to the last column to relect the increase in
nominal wages between the mid-1920s and the mid-1930s.

10.4.3 Results from Nigeria

Table 10.8 then records the basic data from Nigerian railways. hough in
1909 the length of track open in Nigeria was slightly larger than that in
Sierra Leone and around two-thirds greater than that in the Gold Coast, it
is already evident that Nigerian railways were much more active. he freight
ton millage, for example, is about three times greater and freight revenues
about double that of Sierra Leone. his divergence becomes more stark by
1924–5. he amount of ton-miles of freight is more than twenty times that
of Sierra Leone and more than ive times that of the Gold Coast. Moreover,
this gap is even bigger by 1934–5. By this time there were 2,184 miles of
track open in Nigeria, more than four times the amount in the Gold Coast
and about six and a half times the amount in Sierra Leone.
 Reinventing the Wheel 341

Table 10.7. Estimates of the social savings from passengers in Sierra Leone

1911 1925 1935

Total passenger miles 11,047,266 11,377,080
Total time required to travel by rail 736,484 758,472
(at 15 miles per hour) (hours)
Total time required to travel by foot 5,523,633 5,688,540
(at 2 miles per hour) (hours)
Travel time saved (hours) 4,787,149 4,930,068
Value of Time Saved (₤) 19,946 25,061
Total Passenger Revenues (₤) 22,644 52,305 28,383
Passenger Social Saving (₤) −32,359 −3,321
GDP (in current ₤) 25,444,186 32,818,287
Passenger Social Saving −0.13 −0.001
(as a % of GDP)

Table 10.8. Basic data on Nigerian railways

1909 1924–5 1934–5

Mile of Track open 272 1,220 2,184
Total freight hauled (tons) 165,150 680,107 866,681
Total freight in ton-miles 16,024,024 218,427,093 293,186,018
Total passengers moved 285,202 1,922,580 5,080,016
Total passenger miles 13,353,158 92,283,840 148,165,399
Total Freight Revenues (₤) 154,126 1,736,194 1,721,825
Total Passenger Revenues (₤) 46,387 290,639 230,270
Total Expenditures (₤) 131,820 970,446 1,038,758
Total Capital Outlays (₤) 1,377,284 14,978,225 23,014,851

Table 10.9 reports the basic social savings calculations for Nigerian
freight. In absolute terms, these are much larger than those of the Gold
Coast or Sierra Leone, but expressed relative to GDP they are quite similar.
For example, in 1909 they are 1.2 percent of GDP and ater that the num-
bers are very close to those of the Gold Coast.
Table 10.10 examines the social savings associated with passenger travel.
he indings here are very similar to those from the Gold Coast and Sierra
Leone, though for 1934–5 we do ind positive, but very small, social savings
on passenger transportation.
342 Chaves, Engerman, and Robinson

Table 10.9. Estimates of the social savings from freight in Nigeria

1909 1924–5 1934–5

Total Freight Ton-Miles 16,024,024 218,427,093 293,186,018
Total Freight Revenues (₤) 154,126 1,736,194 1,721,825
Total Cost of Head Porterage (₤) 1,468,869 20,022,480 24,432,170
Social Saving (₤) 1,314,743 18,286,286 22,710,345
GDP (in current ₤) 105,316,616 300,260,499 300,859,733
Social Saving as a % of GDP 1.2 6.1 7.5

Table 10.10. Estimates of the social savings from passengers in Nigeria

1909 1924–5 1934–5

Total passenger miles 13,353,158 92,283,840 148,165,399
Total time required to travel by rail 890,211 6,152,256 9,877,693
(at 15 miles per hour) (hours)
Total time required to travel by foot 6,676,579 46,141,920 74,082,699
(at 2 miles per hour) (hours)
Travel time saved (hours) 5,786,368 39,989,664 64,205,006
Value of Time Saved (₤) 24,109 166,624 238,094
Total Passenger Revenues (₤) 46,387 290,639 230,270
Passenger Social Saving (₤) −22,277 −124,015 7,824
GDP (in current ₤) 105,316,616 300,260,499 300,859,733
Passenger Social Saving (as a % of −0.02 −0.04 0.003
GDP)

10.4.4 A Comparative Context

To give some comparative context, Table 10.11, adapted from Leunig (2010,
791, table 1), records benchmark estimates of the social savings of railways
in diferent countries of the world using the methodology that we have
used here. In particular, these are estimates that assume zero elasticity of
demand for transportation services, so that the actual amount of passengers
and freight moved by the railways is taken as the amount that would have
been moved by the less eicient method of transportation in the absence
of railways. Table 10.11 shows that the types of numbers we have found
are fairly normal for the social savings literature. hey are, however, much
smaller than the numbers Summerhill calculated for 1913 Brazil, where
social savings were not less than 18 percent of GDP and quite likely more,
and much smaller than the even larger numbers Coatsworth (1979) found
 Reinventing the Wheel 343

Table 10.11. Estimates of social savings for various countries

Belgium
1846 Freight and passenger 1%
1865 Freight and passenger 2.5%
1912 Freight and passenger 4.5%
Brazil
1913 Freight 18%–38%
Passengers 4.6%
China
1933 Freight and passenger 0.5%
Colombia
1924 Freight 4.8%
England and Wales
1865 Freight 4.1%
1890 Freight 29.1%–31.6%
1843–1913 Passengers 1.5%–14%
France
1872 Freight 5.8%
Passengers 1.7%
Germany
ca. 1900 Freight <5%
Mexico
1910 Freight 24.9%–38.5%
Russia
1907 Freight 4.6%
Passengers 1.6%
Spain
1878 Freight 7.5%
1912 Freight 11%
United States
1859 Freight 3.7%
1859 Passengers 1.6%
1890 Freight 4.9%
1890 Passengers 4.8%

for Mexico in 1910. Both of these scholars attributed the large social sav-
ings to the very ineicient alternative means of transport in their coun-
tries of study. Because neither Mexico nor Brazil had a canal system like
the United States’, freight had to be moved by very ineicient (according
344 Chaves, Engerman, and Robinson

to their calculations) mule cart in the non-railroad counterfactual world.

Fogel himself argued that in the nineteenth-century United States, social
savings might have been as high as 30 percent of GDP, were it not for the
relative eiciency of the American canal system. Now, especially in light of
all the remarks by colonial oicers complaining about head porterage, one
would have anticipated that the same argument that worked for Mexico and
Brazil would apply to West Africa: head porterage, a highly ineicient pre-
railway method of transportation (even more so than mule carts), would
imply large social savings from railway adoption. As we described, however,
we ind small social savings, which is quite surprising. One likely reason for
this is the ability of both precolonial African rulers and colonial govern-
ments to coerce labor into head porterage at below market prices (some-
thing strongly suggested by the historical record). his coercion would then
make head porterage seem cheaper in our data than it would have been if
wages had been set at normal market clearing levels. here were no mecha-
nisms of this type in Brazil and Mexico that could artiicially keep down the
costs of alternative methods of transportation.

10.4.5 Social Savings from other Forms of Transportation

Having made these calculations for the railways, we can make simpler cal-
culations for other forms of transportation. From Table 10.12, we do have
estimates of the cost per ton-mile of freight carried with simpler methods
of transportation, in particular barrel rolling, handcarts, and motor lor-
ries. Using the data from the 1909 Parliamentary Report on these issues,
we can investigate the social savings of using these other “wheeled” meth-
ods of transportation as opposed to head porterage. We do this just for the
Gold Coast. For comparability, therefore, we take the data for the cost per
ton for head porterage from this report, which gives 48.5d. per ton-mile
(the midpoint of 3s.1d. and 5s.), which is signiicantly higher than our own
estimates based on the information we have. To calculate the social sav-
ings associated with cask rolling, for example, we take the midpoint of the
estimate presented in the Parliamentary Report on the cost per ton-mile
of cask rolling and assume that all of the freight moved by the railway was
moved by cask rolling (with the usual zero elasticity of demand for freight
and zero supply of labor assumptions). We then compare the cost of moving
all of this freight by cask rolling to the cost of moving it by head porterage.
hese simple calculations suggest that all forms of wheeled transportation
created positive social savings though the numbers are modest expressed
relative to GDP.
 Reinventing the Wheel 345

Table 10.12. Implied social savings relative to head

porterage of diferent methods of transportation in
1909 in the Gold Coast

Total freight carried by the railways 3,763,552

(in ton-miles)
GDP (in current ₤) 26,266,269
Total Cost of Head Porterage (₤) 760,551
Total Cost of Cask Rolling (£) 258,744
Social Savings of Cask Rolling (£) 501,807
Relative to GDP (%) 1.9
Total Cost of Hand Trucks (₤) 344,992
Social Savings of Hand Trucks (₤) 415,559
Relative to GDP (%) 1.6
Total Cost of Motor Lorry (₤) 313,629
Social Savings of Motor Lorry (₤) 446,922
Relative to GDP (%) 1.7

10.5 THE SOCIAL RATE OF RETURN

We now turn to a diferent way of looking at this issue. Instead of asking
what the social savings associated with the construction of railways was, we
ask what the social rate of return from these projects was. We can do this
in the case of railway construction, because we have detailed knowledge
of the capital expenditures involved in building the railways. his meth-
odology has the advantage that it explicitly takes into account the capital
costs associated with the construction of the railways; the fact that social
savings calculations ignore this has oten been cited as a major drawback
of this approach. he construction of the numbers needed here is straight-
forward; the social rate of return of the railways in any particular year is
simply the social savings in that year plus the net railway revenues, as a
percentage of the capital outlays to date. While results using this method
are not necessarily relective of the relevant decision margins, that is, they
are not marginal social rates of return for an additional unit of capital, they
serve the key purpose of illustrating the magnitude of the railways’ efect on
the economy.13 he indings of this calculation, reported in Table 10.13, are
very striking. hey show that the social rate of return was remarkably high
in these West African colonies, even exceeding 100 percent in Nigeria: in

13
For a diferent approach to measuring social rates of return, with a critique of the current
method, see Mercer (1970).
346 Chaves, Engerman, and Robinson

Table 10.13. Social rates of return %

1909/1911 1924–5 1934–5

Gold Coast 11.0 41.9 51.2
Sierra Leone 48.0 51.3 15.5
Nigeria 95.5 122.0 98.7

other words, the social savings in a single year were suicient to cover the
entire capital outlays until then. To give some sense of how large this rate of
return is, we present in Table 10.14 some estimates of social rates of return
for railroads in other countries.14 Estimates of social rates of return for his-
torical railways are less common than social savings calculations, so we
cannot be as certain in our comparative assertions about West Africa as we
were in Table 10.11. Nonetheless, as Table 10.14 makes clear, the social rate
of return in West African railways was, at its lowest, signiicantly larger than
the largest recorded social rate of return for Western European economies.
Even in the Brazilian example, which Summerhill (2003, 2005) has con-
vincingly argued shows a much larger efect of railway adoption than that
in the United States and Western Europe on account of the country being
a “small, laggard economy,” the social rate of return was between 17 and
23 percent, depending on various assumptions about the extent of social
savings. Only Coatsworth’s calculations for Mexico approach the magni-
tude of our estimates, and even then they are at their highest half the size of
the rates of return we ind in Nigeria.15
How could it be that while social savings in West Africa were so mod-
est, the social rate of return was so high? he obvious answer is that, quite
contrary to the argument that it was expensive to build railways in the for-
est zone of West Africa, building railways there was remarkably cheap by
international standards. his is not entirely surprising. hough building
these lines did require cutting down trees, West Africa is mostly very lat,
and in the three cases we consider, railway construction did not face the
14
For England and Wales, we use Hawke’s estimates of gross social returns without passen-
ger comfort considerations (1970, 406, table XV.01, column 5) combined with Kenwood’s
capital expenditure igures, net of depreciation, and maintenance expenses (1965, 322,
column 1), rather than Hawke’s own. Kenwood’s capital expenditure data have the advan-
tage of stretching back to 1825 and are separated into net and gross categories.
15
hese rates of return we ind are also outsized in comparison to other forms of infra-
structure investment. For instance, in a related calculation outside of the railway context,
Maurer and Yu (2011) showed that the social rate of return on the construction of the
Panama Canal oscillated between 3 percent and 16 percent, depending on the year.
 Reinventing the Wheel 347

Table 10.14. Estimates of social rates of return (%) for various countries

Country/ Social Rate Source

year of Return (%)
Brazil
1913a 17.9–23.1 Summerhill (2005, 87, table 7).
England and Wales
1854b 13.58 Constructed from Hawke (1970, 406, table XV.01,
column 5) and Kenwood (1965, 322, column 1).
1855 13.68 ibid.
1856 14.61 .
1857 14.54 .
1858 13.98 .
1859 15.15 .
1860 16.07 .
1861 15.77 .
1862 15.51 .
1863 15.98 .
1864 16.67 .
1865 16.46 .
1866 17.18 .
1867 17.62 .
1868 17.91 .
1869 18.43 .
1870 19.68 .
Mexico
1881c 1.8 Coatsworth (1972, 141, table IV.12).
1882 10.4 ibid.
1883 14.2 .
1884 9.7 .
1885 18.8 .
1886 14 .
1887 15.3 .
1888 18.9 .
1889 20.3 .
1890 26.6 .
1891 35.9 .
1892 34.9 .
1893 27.2 .
1894 25.3 .
1895 29.1 .
1896 30.3 .
1897 31.6 .
1898 33.4 .
1899 52

(continued)
348 Chaves, Engerman, and Robinson

Table 10.14. (continued)

Country/ Social Rate Source

year of Return (%)
USA
1859 15 Fishlow (1965)
1890d 12.3–15.8 David (1969, 522–3), social savings from Fogel
(1964)
1890e 15.8–20.4 ibid.
Notes: a) Uses demand elasticity of -1. b) See text. c) Refers only to the largest railroad company. d)
Returns net of maintenance expenditures. e) Returns gross of maintenance expenditures.

problem of bridging large rivers (the railways skirted or went around what
large rivers there were, such as the Volta or Niger). In the United States and
the British isles, the relative abundance of good waterways that made the
railways less appealing also made them more expensive to build. his topic
deserves further investigation.

10.6 WHAT STOPPED TECHNOLOGY ADOPTION?

If wheeled transportation, particularly railways, was economically eicient,
why did African societies not adopt it? Part of the explanation for this stems
from the nature of African polities. Most of the continent had relatively
low levels of political centralization and consolidation, which would make
it diicult to implement public works projects with such a large ixed cost.
In Sierra Leone, for example, Abraham (2003) shows how the south of the
country, Mendeland, was divided into a system of nine competing and war-
ring states in the second half of the nineteenth century. He makes a dis-
tinction between territorial states, such as the Sherbro, Lugbu, Gallinas,
Bumpeh, and Kpaa-Mende states, which had well-deined territories and
which were not identiied with a single person, and the Hegemonies, such
as the Tikongoh state of Makavoray and the Luawa state of Kai Londo,
which were. Nevertheless, even the territorial states were not bureaucra-
tized and did not collect systematic taxes from their inhabitants, though
they did collect tribute and organize compulsory labor and armies. It is
diicult to imagine that these states could have cooperated suiciently to
create large-scale public works like a railway, even if they had been able to
raise the capital and attract the expertise. his lack of political centraliza-
tion in much of sub-Saharan Africa is plausibly part of the explanation for
 Reinventing the Wheel 349

why railways were not built (though it does not explain the failure to adopt
much simpler wheeled technologies).
Nevertheless, this explanation can hardly apply to larger, more bureau-
cratic, and consolidated African states such as Asante and Ethiopia (see
Hopkins 2000a, 2000b; Warner 1999). In both these historical cases, we
have direct evidence relevant to the issue of railway adoption; Asante made
a belated attempt to construct a railway to Kumase, and Ethiopia did in fact
build a railway before its colonization. We also have evidence for Zanzibar,
where a similar project to build a railway was started by the independent
sultan and then collapsed.
In the case of Asante, Wilks notes, “from the reign of Mensa Bonsu
onwards, the Asante government began to explore the possibilities of utiliz-
ing European capital and skills to create a railroad system in Asante” (1989,
41). Asantehene (king) Mensa Bonsu came to the throne in 1874 and ruled
until being deposed in 1883. Agyeman Prempe took the throne in 1888.
Prempe’s chief of his “foreign afairs bureau” was John Owusu Ansa, who
took up the task of developing a railway. To this aim, he proposed setting up
the “Ashanti and Prah Mining and Trading Company.” On April 26, 1892,
the Asantehene signed an agreement with Dr. J. W. Herivel to start the com-
pany, which was to inance and manage the construction of railroads in
cooperation with the Asante government; the government even agreed to
immediately supply four hundred laborers to begin laying track. As Wilks
notes:
Governor Griith of the Gold Coast viewed the project with considerable alarm,
and deemed it expedient to deter Herivel from pressing forward with a scheme
which might greatly have strengthened the Asante economy. In 1893 the actual
agreement between the Asantehene and Herivel was impounded by the High Court
of the Gold Coast, and Herivel was harassed by the Customs Department until
inally in 1894 he was obliged to abandon the scheme. (1989, 636)

his setback did not deter the Asantehene, however. In early 1895, he sent
an embassy headed by Ansa to London. In October of that year, the Asante
entered into an agreement with George Reckless, giving him “a Charter for
the opening up of Ashantee to British enterprise and skill” (Wilks 1989,
650). Part of what Reckless agreed to do was to build a railway from the
coast to Kumase. As Wilks puts it, “Chamberlain’s reaction was to refuse
to accept the validity of the concession” (1989, 652). Ansa responded by
hiring a barrister, and Chamberlain was forced to concede the charter
was legal. Nevertheless, even before Ansa was on the boat back to West
Africa, Chamberlain had cabled Governor Maxwell to inform him that the
350 Chaves, Engerman, and Robinson

“expedition must go [to] Kumasi at all events” (Wilks 1989, 655). British
military intervention blocked any chance of autonomous adoption of the
railway.
he evidence from Asante suggests that one reason that African states
did not adopt railways during the nineteenth century was perhaps precisely
because they were economically eicient. European powers, embarked
on colonial expansion, had an interest in blocking the difusion of tech-
nology to places they considered valuable colonies. he modernization of
Asante would have made it more diicult to colonize, which explains why
the British opposed it. his mechanism seems likely to have operated quite
widely during the nineteenth century.
he case of Ethiopia is just as revealing about the mechanisms that pre-
vented African states from building railways during the nineteenth century,
and it independently paints a picture that is very consistent with the evi-
dence from Asante (see Gilmour 1906 for an entertaining contemporary
account). Plans to build a railway from the coast to the capital city of Addis
Ababa moved forward the moment that Menelik II came to the throne in
1889. In February 1893, Menelik empowered one of his advisers to cre-
ate a company, and Leon Chefneux, a French trader who had been living
in Ethiopia since 1882, was sent to Europe to look for capital to build a
railway between Djibouti, in French Somaliland, and Addis Ababa. By this
time, the Italians had declared Ethiopia a protectorate of Italy, which made
it diicult to raise capital and also made the French government reluctant
to allow the construction of a railway across French territory. However, the
military defeat of the Italians by the Ethiopians at the battle of Adowa in
1896 solved these problems. here was a great deal of internal opposition
to the railways, however. Ras Makonnen, a leading aristocrat, told Menelik,
“When the railway reaches Harar, Harar will no longer be yours and when it
reaches Addis Ababa, Shoa will no longer be yours” (Pankhurst 1968, 307).
Gleichen, a British diplomat, recorded that “a large number of the chiefs . . .
strongly object to such a new fangled idea . . . on the grounds that it would
introduce into the country the all-pervading white man” (Pankhurst 1968,
308). Work began on the line in October 1897, though it took until July
22, 1901 for the irst train to run. At that point the track going inland from
Djibouti was 106 kilometers long and had just passed into Ethiopian terri-
tory at Dire Dawa. he British, whose colony of Somaliland was just next
to Djibouti, disliked the fact that a French private company ran the rail-
way. With the encouragement of the colonial oice, British investors began
buying up shares in the company, and in response the French government
efectively nationalized it in 1902. In the process, the French unilaterally
 Reinventing the Wheel 351

rewrote the concession violating both the original details and the spirit of
the deal signed with the Ethiopian monarch. Menelik was incensed: “the
Emperor, arguing that his original concession had been violated, withdrew
various priviledges [sic] . . . [and he] refused to grant permission to start on
construction of the line between Dire Dawa and the capital” (Pankhurst
1968, 323). He announced that, while he still wanted the railway, “he would
not permit the line to be built by a foreign government or by a company
controlled by a foreign government” (quoted in Pankhurst 1968, 323). In
response to this crisis, various schemes were proposed; British and French
interests, for example, suggested the internationalization of the railway,
while Menelik attempted to raise the needed capital himself, directly from
international inancial markets. he reasoning behind the French takeover
of 1902 surfaced in a remarkably frank debate in the French Senate on April
1, 1905, when the Comte d’Aunay remarked:

[W]e were able to cherish the inest hopes for our position in Abyssinia. We had the
monopoly of the railway, which gave us a precious instrument of penetration; one
could say that the Empire of Menelik would become a colony for us from which we
could gather the beneits without assuming any of the responsibilities. (Quoted in
Pankhurst 1968, 327)

In mid-1906, an exasperated Menelik began the construction of the railway

from Addis Ababa to Dire Dawa. Regardless, by 1908, Chefneux’s company
went bankrupt. Menelik, unable to build the railway himself without exter-
nal inancing, gave a second concession to the French, who agreed that this
concession would again be a private company run through the Banque de
l’Indo-Chine. he new company was charged with extending the line up to
Addis Ababa, which it reached in 1917 ater the death of Menelik and dur-
ing the regency of Ras Tafai (later Haile Selassie).
he evidence from the Ethiopian case makes clear that a key obstacle
preventing African polities from adopting the modern technology of the
railway was that they feared this adoption would accelerate colonization.
Finally, we see similar, if not identical, political pressures afecting
the decision of African polities to introduce railways in the Sultanate of
Zanzibar. he Sultanate of Zanzibar was descended from the Arab mon-
archy that ruled Oman, which had established political supremacy along
the East African coast before the massive entry of European interests into
the region. Ater the abolition of slavery in the British colonies, when
British eforts to eliminate the slave trade shited to East Africa, the British
refrained from extending any formal political authority in the region, leav-
ing the Sultanate’s sovereignty intact. Moreover, colonial oicers tried to
352 Chaves, Engerman, and Robinson

strengthen the monarchy and its state apparatus, and then relied on infor-
mal inluence to pressure Zanzibari rulers into curbing the trade in their
domain. hus, by the 1870s, the Sultanate of Zanzibar claimed suzerainty
over wide swathes of the land extending from the coastline of modern
Tanzania, Kenya, and Somalia into the East African interior as far as south-
ern Sudan and the Great Lakes, making it one of the largest political entities
in Africa at the time.
Secondary sources describe strategic reasoning by the Sultan of Zanzibar
that strongly resembles the considerations facing Menelik in Ethiopia, when
proper allowances are made for the very diferent nature of the two king-
doms. In the 1870s, the Egyptian Khedive’s plans to expand his empire deep
into the Sudan threatened the territorial claims of the Sultan. At the same
time, the East African interior had begun to attract the interest of William
Mackinnon, a politically connected British shipping magnate (Galbraith
1972). In 1876, Sultan Bargash therefore seriously considered giving a con-
cession to a British company led by Mackinnon where, in exchange for
the right to function almost as an autonomous government (taxing, reg-
ulating trade in arms and liquor, garrisoning troops) in the Sultan’s inland
domain, the company committed to constructing roads and railroads on
the Sultan’s behalf and upholding his sovereign claims over the area. he
decision was far from straightforward. As Coupland described, the Sultan
reasoned that:

If he called European capital and enterprise to his aid, he would proit inancially, all
going well, from the increase in duties in the mainland trade, and he would proit
politically by the establishment of efective control in the interior. he inancial and
commercial interests he enlisted in his service would be mobilized to prevent any
more attempts like [the Khedive] Ismail’s to question his authority and violate the
integrity of his domains. But there was an obvious risk. His servants might become
his masters. hose European pioneers might prove to be more than philanthropists
and business men. hey might be converted into instruments of national aggrand-
isment; and, once they had got a irm grip on the interior in the Sultan’s name, they
might coolly hand it over to their Government. (Coupland 1939, 305)

Coupland argues that Bargash was willing to face this risk because the alter-
natives were quite likely the loss of sovereignty to the Khedive or eventual
colonization where none of his sovereign rights would be recognized. he
negotiations advanced considerably, with the Sultan demonstrating strong
commitment to the project, and in fact this proposal received approval at
the highest levels of the British government. Disagreement exists in the
historical literature on why exactly the negotiations collapsed, but at least
part of the reason was that “[Bargash’s] advisers remonstrated against his
 Reinventing the Wheel 353

inclinations to transfer his rights, even his sovereign power, to a European

company. he Europeans would preempt trade with the interior, and the
Arabs would be deprived of their livelihood” (Galbraith 1972, 67). hus, in
this case as well, the beneits of technology adoption were weighed against
the political risks arising from the threat of colonization and European
control.

10.7 CONCLUSIONS
In this chapter, we have undertaken preliminary calculations to examine
the economic beneits of introducing modern methods of transportation
into the British West African colonies of the Gold Coast, Nigeria, and Sierra
Leone during the early twentieth century. We did this in the context of an
academic literature that has argued that the reason that Africans did not
adopt modern technologies, such as wheeled transportation, was that they
were not appropriate technologies given the underlying factor endowments
and circumstances. he bulk of this chapter focused on the introduction of
railways. hough this is not the cleanest case for examining the economic
rationality of wheeled transportation, it is facilitated by the very rich data
that the British colonial state collected on the construction and operation of
the railways. he main question we ask is whether constructing the railways
was economically rational compared to moving goods by head porterage
and moving passengers by foot. We tackle this issue in two ways, irst by
using the social savings approach of Fogel (1964) and second by calculat-
ing the social rate of return. Our results are very contrary to the conven-
tional wisdom. hough it is true that the social savings railways created
were modest compared to estimates of national income, the more interest-
ing concept, the social rate of return on capital, was very high. In the case
of Nigeria, it averaged around 100 percent per annum, suggesting that the
social savings in one year were suicient to cover the entire capital expen-
ditures until then.
We would also argue that in a sense these calculations almost certainly
underestimate the economic impact of the adoption of the railway. his is
for the simple reason that the British colonial government built the railways
not simply as an economic activity, but also as part of a strategy of extending
colonial rule. For instance, Tsey (1986) points out that the irst railway built
in the Gold Coast, though it headed north into the Asante goldields, which
seems economically sensible, was extended to Kumase to allow the British
to extend their military domination into the heart of the Asante state, not
from any obvious economic motive. Similarly, in the case of Sierra Leone,
354 Chaves, Engerman, and Robinson

when construction of the irst railway started, the original plan was for the
tracks to run from Freetown to the north of the country. In 1898, however, a
massive rebellion, the “Hut Tax Rebellion,” (known in modern Sierra Leone
as the Bai Bureh War) broke out. hough it started in the north, it was
most intense and lasted the longest in Mendeland in the south. Ater they
defeated the rebellion, the British changed the planned route of the rail-
way line, and instead of building it to the north, they built it to the south,
right into the heart of Mendeland. To the extent that military and strategic
factors were important in determining the routes the railways took, and
to the extent that the relevant economic fundamentals were not perfectly
correlated with the political fundamentals that made some places harder
to govern, the railways could not have been built in the most economically
optimal places for them. herefore our indings surely underestimate the
economic potential for railway construction.
Nevertheless, the fact that railways were economically eicient but not
adopted by Africans does not imply in any sense that Africans were irratio-
nal or not able to ascertain the relevant costs and beneits. In fact, in the case
of railways, we showed that three circumstances inhibited African polities
from adopting such technologies. First, there was little political centraliza-
tion (so that private costs and beneits diverged from social costs and bene-
its); second, technology adoption was blocked by potential colonial powers
who did not want autonomous African development; third, the African pol-
ities feared that railway construction would in itself accelerate colonization.

APPENDIX: RECONSTRUCTING COLONIAL GDP

his appendix describes the procedure we followed to reconstruct the GDP
for the Gold Coast, Sierra Leone, and Nigeria. To the best of our knowl-
edge, no other systematic attempts to reconstruct GDP for the prewar era
have been constructed for these countries, except for the pioneering work
of Szereszewski (1965). We largely follow his method, which we outline for
the Gold Coast together with our deviations from his method. We present
a full result for the Gold Coast in 1909 in Table 10.A2. Finally, we briely
discuss diferences between reconstructing Gold Coast GPD and GDP for
Sierra Leona and Nigeria.

Szereszewski’s Work
Szereszewski tries to understand the structure of the Ghanaian econ-
omy to investigate the process that transformed the Gold Coast from an
 Reinventing the Wheel 355

Table 10.A1. Components of colonial GDP

Symbol from (1) Category

C (2) Private Consumption of Imported Goods
(4) Consumption of Public and Related Services
(6) Traditional Consumption
G (3) Consumption of Government Services
I (5) Gross Capital Formation
E (1) Export Production
M (7) Imports of Goods and Non-factor Services

agricultural society into the largest cocoa exporter in the world in 1911. For
1891, 1901, and 1911 (coinciding with census years), he constructs three
“successive accounting models which would assess quantitatively the struc-
tural changes of the economy over the two decades and the magnitude of its
growth” (1965, 128). He considers an introduced and an indigenous sector,
broadly corresponding to the coastal area and the hinterland in Ghana (and
the Colony and Protectorate in Nigeria and Sierra Leone).

Data and Methodology

All data are taken from the so-called Blue Books, which are colonial reports
submitted to the British government (Government of the Gold Coast,
1909, 1926, 1935; Government of Northern Nigeria, 1909; Government
of Southern Nigeria, 1909; Government of Nigeria, 1925, 1935; and
Government of Sierra Leone, 1909, 1925, 1935). hese reports contain,
aside from a short descriptive part, data on imports and exports and on
government activity. From these data we constructed an estimate of GDP
following a standard GDP from expenditure formula:
Y=C+G+I+E–M
where Y stands for GDP, C for consumption, G for government expendi-
ture, I for investment, E for exports, and M for imports. We break this for-
mula down into two building blocks, the introduced part of the economy
(G, I, (E-M) and the introduced part of C) and the traditional (indigenous)
consumption (the indigenous part of C). he letters in the equation corre-
spond to categories in our GDP calculation (see Table 10.A2). Table 10.A1
lists the inputs in the equations and their equivalents in our calculation.
he names of the equivalents have been chosen to conform to Szereszewski
(1965) and to match closely to the Blue Books.
356 Chaves, Engerman, and Robinson

Table 10.A2. Estimated Expenditures on GDP – Gold Coast 1909 – in pounds

Category from Table A1 Total

(1) Export Production 2,656,000
Seaborne 2,656,000
(2) Private Consumption of Imported Goods 3,490,178
Spirits 775,200
Textiles 1,048,000
Tobaccos 158,400
Provisions 779,850
Miscellaneous 728,728
(3) Consumption of Government Services 746,000
(4) Consumption of Public and Related Services 199,000
Post and Telegraph 5,000
Education 12,000
Railway 182,000
(5) Gross Capital Formation 746,240
Buildings and construction 820,000
Equipment 512,400
Cocoa 0
Net accumulation of specie 133,000
Changes in stock of imported goods 8,000
(6) Traditional Consumption 20,646,650
(7) Imports of Good and Non-factor Services 2,217,800
Seaborne 2,205,000
Services other than insurance and freight 12,800
New imports of specie 133,000
GDP is the total of categories 1–5 minus 7.
GDP 26,266,270
including (6)
GDP 5,619,618
excluding (6)
Total Unskilled
Population wage (pence)
1,696,985 8

he indigenous consumption part (category (6)) accounts for the con-

tribution of the indigenous population to GDP. Szereszewski computes a
typical consumption basket for the Gold Coast and multiplies this by the
population to arrive at the money value of traditional consumption. his
measure is hard to replicate, and it isn’t easily transferable across countries.
herefore, we assume that the minimum unskilled labor daily wage rate
is set such that the population was made indiferent between subsistence
 Reinventing the Wheel 357

farming and unskilled day labor. We then ind the minimum day labor
wage from the Blue Books and multiply this with the population to arrive
at a measure of the traditional expenditure on GDP. he other components
of our analysis, the categories in Table 10.A1, will be dealt with later in this
appendix.

Export Production
Because exports are paid for from abroad, they add to GDP. he total ig-
ures for exports, including imports duties and c.i.f. (cost, insurance, and
freight) costs, in one particular year can be taken directly from the Blue
Books. Overland exports were usually not measured in any coherent way.
herefore, we conine ourselves to seaborne exports.

Private Consumption of Imported Goods

Szereszewski computed ratios between the import and the consumption
of diferent classes of goods. With these ratios and the import igures from
the Blue Books we can compute domestic consumption of these goods. In
the Blue Books, all imported commodities are listed with their respective
import value and are taken from the “home consumption” categories in the
Blue Books. We have aggregated the individual commodities into ive cate-
gories; Spirits, Textiles, Tobacco, Provisions (food, paper, candles, etc.), and
Miscellaneous (tools, fuel, medicine, etc). From Szereszewski we take the
following conversion factors:
Spirits, 1.7
Textiles, 1.6
Tobacco, 1.6
Provisions, 1.5
Miscellaneous, 1.4.
his means that, for instance, in the Gold Coast in 1909 total spirit imports
were worth 456,000 pounds and total consumption of spirits was worth
775,200 pounds.

Consumption of Government Services

his category records the expenditures of the government, given directly in
the Blue Books. To this have been added the expenditures of the local gov-
ernments such as, in the case of the Gold Coast, the governments in Accra,
358 Chaves, Engerman, and Robinson

Cape Coast, and Secondi. Traditional government is not included. Also,

government expenditure on construction is considered separately.

Consumption of Public and Related Services

his category records consumption of the services ofered by the postal ser-
vice and the railway services. Both quantities can be recorded directly from
the Blue Books. For the Gold Coast in 1909, we were able to include expen-
diture on missionary schools as well.

Gross Capital Formation

As a means of accounting for investment, several capital formation catego-
ries are computed.
Buildings and construction. We have used government expenditure on
construction works as a measure of the capital investment in building and
construction. Szereszewski uses the money value of the imports of building
and construction materials times a “construction coeicient” meant to cap-
ture the relation between the money value of the imports and the eventual
investment value for GDP. He estimates this coeicient at four. We have
used this methodology only for the Gold Coast in 1909, lacking detailed
government expenditure data. To the government expenditure on buildings
and construction, the government expenditure on railway plant and rolling
stock is added.
Equipment. Spending on equipment is constructed from the import lists
in the Blue Books. However, only for the Gold Coast in 1909 and Sierra
Leone in 1925 could detailed igures on equipment imports be found. For
other country/years the ratio between construction spending and equip-
ment spending (0.26) for the Gold Coast in 1909 has been assumed constant
and has subsequently been applied to the construction igures to generate
the data for the equipment category.
Agriculture. Although all agricultural output was created by labor inputs,
there is one category, the establishment of new cocoa farms, that deserves
attention. We measure the investment/expenditure on GDP by the extra
acreage of cocoa planted. he acreage we get from the tonnage of cocoa
exported, which is mentioned in the Blue Books. Using the average yield
formula from Szereszewski of four hundred twenty pounds per acre we can
compute the total acreage that was used for growing the exported tonnage
of cocoa. his is subsequently multiplied by the number of days it takes to
 Reinventing the Wheel 359

bring an acre of cocoa to bearing age (170 days; sources are in Szereszewski).
Finally, we use the unskilled labor wage rate to assign a money value to
these days worked. his gives the money value of the investment needed
to grow the exported quantity of cocoa. his investment is recorded in the
year of export, although it was actually invested seven years before, which is
the time it takes for cocoa to grow to maturity. However, we record expen-
ditures as they enter output, so the data of the year of inquiry can be used
to “trace” the original investment and assign it to the current year. No other
crops or agricultural investment can be assessed in a similar way.
Accumulation of specie. he category records the diference between the
imports and exports of specie (foreign currency).
Changes in stocks of imported goods. his category records the diference
between total imports and the imports cleared through consumption (in
the “home consumption” categories in the Blue Books). he diference rep-
resents the goods stock in warehouses in the ports and includes mainly
spirits and textiles and can be seen as an stockpiling investment.

Traditional Consumption
his category records the traditional consumption as outlined earlier. It uses
the population igures and unskilled wage igures from the Blue Books to
assess the money value of the traditional consumption. We assume that the
unskilled wage rate was set such that people were made indiferent between
(subsistence) farming and unskilled labor.

Imports of Goods and Non-Factor Services

his category records the money value of imports which are to be deducted
from total GDP because the money to pay for the imports in added to the
exporting country’s GDP. Also, the imports of specie are deducted here
because the domestic currency to pay for the foreign currency accrues to
the other nation’s GDP. As a last element, the remunerations for services
performed by irms abroad are added here (f.i., the treatment of soldiers/
oicers in hospitals in England).
Table 10.A2 gives a complete reconstruction using the aforementioned
methodology for the GDP of the Gold Coast for the iscal year 1909. his
same methodology can be applied, mutatis mutandis, to Sierra Leone and
Nigeria. heir total GDP igures, together with the remaining igures for the
Gold Coast, are given in Table 10.A3.
360 Chaves, Engerman, and Robinson

Table 10.A3. Colonial GDP for the Gold Coast, Nigeria and Sierra Leone, in
pounds

(Fiscal) GDP with GDP without Population Unskilled

Year traditional wage rate
consumption (pence)
Gold Coast 1909 26,266,270 5,619,620 1,696,985 8
Sierra Leone 1909 2,330,890 1,398,250 76,655* 8
Nigeria (North) 1909 94,832,020 29,336,230 6,714,038 9
Nigeria (South) 1909 10,484,600 9,937,100 60,000* 6
Nigeria (Total) 1909 105,316,620 39,27,330 6,774,038
Gold Coast 1924/25 53,108,140 21,913,280 2,279,077 9
Sierra Leone 1924/25 25,444,1890 4,419,280 1,536,066 9
Nigeria 1924/25 300,26,500 14,561,340 18,900,391 9
Gold Coast 1934/35 61,068,570 17,969,230 3,163,568 9
Sierra Leone 1934/35 32,818,270 3,233,090 1,768,480 11
Nigeria 1934/35 300,859,730 28,225,050 19,918,516 9
* Population without Protectorate (Colony only).

Nigeria and Sierra Leone

he main diferences between the analysis for the gold coast and Sierra
Leone and Nigeria come from two sources. he irst source, which applies
only to Nigeria, is that Nigeria was for administrative purposes split into
Northern and Southern Nigeria in 1909. GDP igures for both colonies
plus a total are given in Table 10.A3. By 1925 the colonies had merged.
he second source, which applies to both countries, has to do with popula-
tion igures. In 1909 the population igures in the Blue Books give only the
population igures for the Colony and not for the Protectorate, causing an
underestimation of total GDP.

Extra References for the Appendix

Government of the Gold Coast. (1909). “Blue Book.” Accra: Government
Printers Oice.
Government of the Gold Coast. (1926). “Blue Book.” Accra: Government
Printers Oice.
Government of the Gold Coast. (1935). “Blue Book.” Accra: Government
Printers Oice.
 Reinventing the Wheel 361

Government of Northern Nigeria. (1909). “Blue Book.” Lagos: Government

Printers.
Government of Southern Nigeria. (1909). “Blue Book.” Calabar: Government
Printers.
Government of Nigeria. (1925). “Blue Book, Colony & Protectorate of
Nigeria.” Lagos: Government Printer.
Government of Nigeria. (1935). “Blue Book, Colony & Protectorate of
Nigeria.” Lagos: Government Printer.
Government of Sierra Leone. (1909). “Blue Book.” Freetown: Government
Printing Oice.
Government of Sierra Leone. (1925). “Blue Book.” Freetown: Government
Printing Oice.
Government of Sierra Leone. (1935). “Blue Book.” Freetown: Government
Printing Oice.

References
Abraham, Arthur. (2003). An Introduction to the Pre-Colonial History of the Mende of
Sierra Leone. Lewiston: he Edwin Mellor Press.
Acemoglu, Daron, Simon Johnson, and James A. Robinson. (2001). “he Colonial
Origins of Comparative Development: An Empirical Investigation.” American
Economic Review, December, 91(5): 1369–1401.
Acemoglu, Daron and Fabrizio Zilibotti. (2001). “Productivity Diferences.” Quarterly
Journal of Economics, 116: 563–606.
Akurang-Parry, Kwabena O. (2000). “Colonial Forced Labor Policies for Road-Building
in Southern Ghana and International Anti-Forced Labor Pressures, 1900–1940.”
African Economic History, 28: 1–25.
Alldridge, T. J. (1910). A Transformed Colony: Sierra Leone as It Was, and as It Is, Its
Progress, Peoples, Native Customs and Undeveloped Wealth. London: Seeley and Co.
Limited.
Austen, Ralph A. and Daniel Headrick. (1983). “he Role of Technology in the African
Past.” African Studies Review, 26: 163–84.
Austin, Gareth. (2008). “Resources, Techniques, and Strategies South of the Sahara:
Revising the Factor Endowments Perspective on African Economic Development,
1500–2000.” Economic History Review, 61: 587–624.
Basu, Susanto and David N. Weil. (1998). “Appropriate Technology and Growth.”
Quarterly Journal of Economics, 113: 1025–54.
Bulliet, Richard W. (1975). he Camel and the Wheel. Cambridge, MA: Harvard
University Press.
Burns, A. C. (1921). he Nigeria Handbook Containing Statistical and General Information
Respecting the Colony and Protectorate. Lagos: Government Printer.
362 Chaves, Engerman, and Robinson

Church, R. J. Harrison. (1956). “he Transport Pattern of British West Africa.” In

Geographical Essays on British Tropical Lands, edited by R. W. Steel and C. A.
Fisher, 30–61. London: G. Philip.
(1980). West Africa : A Study of the Environment and of Man’s Use of It, 8th edition.
New York: Longman.
Cliford, Sir Hugh. (1920). he Gold Coast Regiment in the East African Campaign.
London: J. Murray.
Coatsworth, John H. (1972). “Impact of Railroads on the Economic Development of
Mexico, 1877–1910,” PhD Dissertation: Department of History, University of
Wisconsin.
Coatsworth, John H. (1979). “Indispensable Railroads in a Backward Economy: he
Case of Mexico.” Journal of Economic History, 39(4): 939–60.
Coquery-Vidrovitch, Catherine and Paul E. Lovejoy, eds. (1985). he Workers of African
Trade. Beverly Hills, CA: Sage Publications.
Coupland, Sir Reginald. (1939). he Exploitation of East Africa, 1856–1890: he Slave
Trade and the Scramble. London: Faber and Faber Ltd.
Crats, Nicholas F. R. (2004). “Social Savings as a Measure of the Contribution of a New
Technology to Economic Growth,” London School of Economics, Department of
Economic History, Working Paper No. 06/04.
David, Paul A. (1969). “Transportation Innovation and Economic Growth: Professor
Fogel On and Of the Rails.” Economic History Review, 22: 505–26.
Donaldson, Dave. “Railroads of the Raj: Estimating the Impact of Transportation
Infrastructure.” forthcoming in the American Economic Review.
Fishlow, Albert. (1965). American Railroads and the Transformation of the Antebellum
Economy. Cambridge, MA: Harvard University Press.
Fogel, Robert W. (1964). Railroads and American Economic Growth: Essays in Econometric
History. Baltimore, MD: Johns Hopkins University Press.
(1979). “Notes on the Social Saving Controversy.” Journal of Economic History,
39: 1–54.
Galbraith, John S. (1972). Mackinnon and East Africa, 1878–1895: A Study in the “New
Imperialism.” Cambridge: Cambridge University Press.
Gilmour, T. Lennox. (1906). Abyssinia: he Ethiopian Railway and the Powers. London:
Alston Rivers Ltd.
Goody, Jack. (1971). Technology, Tradition and the State in Africa. New York: Cambridge
University Press.
Gould, Peter R. (1959). Transportation in Ghana. Northwestern University, Studies in
Geography No. 5.
Government of the Gold Coast. (1909). “Report of the General Manager upon the
Government Railways,” Accra: Government Printers.
(1926). “Gold Coast Railway Administrative Report for the Year 1924–1925,” Accra:
Government Printers.
(1935). “Gold Coast Railway Administrative Report for the Year 1934–1935,” Accra:
Government Printers.
Government of Nigeria. (1909). “Lagos Railway Annual Report 1909,” Lagos:
Government Printers.
(1925). “Nigerian Railway and Udi Coalmines Administrative Report for the Year
Ending 31st March 1925,” Lagos: Government Printers.
 Reinventing the Wheel 363

(1935). “Annual Report on the Government Railway and Colliery of Nigeria for the
Financial Year Ending 31st March 1935,” Lagos: Government Printers.
Government of Sierra Leone. (1911). “Report on the Transport Department for the Year
1911,” Freetown: Government Printers.
(1925). “Sierra Leone Railway Administrative Report for the Year 1925,” Freetown:
Government Printers.
(1935). “Sierra Leone Railway Administrative Report for the Year 1935,” Freetown:
Government Printers.
Hailey, Lord. (1938). An African Survey. London: Oxford University Press.
Hall, Robert E. and Charles I. Jones. (1999). “Why Do Some Countries Produce So
Much More Output per Worker than Others?” Quarterly Journal of Economics,
114: 83–116.
Hawke, G. R. (1970). Railways and Economic Growth in England and Wales, 1840–1870.
Oxford: Clarendon Press.
Herbst, Jefrey I. (2000). States and Power in Africa. Princeton, NJ: Princeton University
Press.
Hill, Polly. (1972). Rural Hausa: A Village and a Setting. New York: Cambridge University
Press.
Hopkins, Anthony G. (1973). An Economic History of West Africa. New York: Addison
Wesley Longman.
(2000a). “Asante and the Victorians: Transition and Partition on the Gold Coast.” In
Imperialism, Decolonization, and Africa, edited by Roy Bridges, 33–57. London:
Palgrave Macmillan.
(2000b). “Quasi-States, Weak States, and the Partition of Africa.” Review of
International Studies, 26: 311–20.
House of Commons. (1904). “Papers Relating to the Construction of the Railways in
Sierra Leone, Lagos, and the Gold Coast,” Cd. 2325.
(1909). “Mechanical Transport in the Colonies,” Cd. 4589.
(1924). “Private Enterprise in British Tropical Africa,” Cmd. 2016.
Hsieh, Chiang-Tai and Peter J. Klenow. (2010). “Development Accounting.” American
Economic Journal: Macroeconomics, 2(1): 207–23.
Jedwab, Remi and Alexander Moradi. (2011). “Transportation Technology and
Economic Change: he Impact of Colonial Railroads on City Growth in Africa.”
http://home.gwu.edu/~jedwab/Jedwab_Moradi_Oct2013.pdf.
Kenwood, A. G. (1965). “Railway Investment in Britain, 1825–1875.” Economica, 32:
313–22.
Knowles, Lilian C. A. (1928). he Economic Development of the British Overseas Empire.
London: George Routledge & Sons.
Law, Robin C. (1977). Oyo Empire, c.1600-c.1836: West African Imperialism in the Era of
the Atlantic Slave Trade. New York: Oxford University Press.
(1980). “Wheeled Transportation in Pre-Colonial West Africa.” Africa, 50:
249–62.
Leunig, Timothy. (2010). “Social Savings.” Journal of Economic Surveys, 24(5): 775–800.
Lovejoy, Paul E. (1980). Caravans of Kola: he Hausa Kola Trade, 1700–1900. Zaria:
Ahmadu Bello University Press Ltd.
Lugard, Frederick. (1922). he Dual Mandate in British Tropical Africa. London: William
Blackwood and Sons.
364 Chaves, Engerman, and Robinson

Mann, Kristin. (1995). “Owners, Slaves and the Struggle for Labour in the Commercial
Transition at Lagos.” In From Slave Trade to “Legitimate” Commerce: he Commercial
Transition in Nineteenth-Century Africa, edited by Robin C. Law, 144–71. New
York: Cambridge University Press.
Mason, Michael. (1978). “Working on the Railway: Forced Labor in Northern Nigeria.”
In African Labor History, edited by Peter C. W. Gutkind, Robin Cohen, and Jean
Copans, 56–79. Beverly Hills, CA.
Maurer, Noel and Carlos Yu. (2011). he Big Ditch: How America Took, Built, Ran and
Ultimately Gave Away the Panama Canal. Princeton, NJ: Princeton University
Press.
McPhee, Allan. (1926). he Economic Revolution in British West Africa. London: George
Routledge & Sons.
Mercer, Lloyd J. (1970). “Rates of Return for Land-Grant Railroads: he Central Paciic
System.” Journal of Economic History, 30: 602–26.
O’Brien, Patrick K. (1977). he New Economic History of the Railways. London:
Croom Helm.
Ogunremi, Deji. (1975). “Human Porterage in Nigeria in the Nineteenth Century.”
Journal of the Historical Society of Nigeria, 8: 37–45.
Ogunremi, Gabriel. (1982). Counting the Camels: he Economics of Transportation in
Pre-Industrial Nigeria. New York: Nok Publishers International.
Ormsby-Gore, W. G. A. (1926). Report by he Hon. W.G.A. Ormsby-Gore, M.P. on His
Visit to West Africa during the Year 1926. London: H. M. Stationary Oice.
Oyemakinde, Wale. (1974). “Railways Construction and Operation in Nigeria, 1895–
1911: Labour Problems and Socio-Economic Impact.” Journal of the Historical
Society of Nigeria, 7: 303–24.
Pankhurst, Richard. (1968). Economic History of Ethiopia 1800–1935. Addis Ababa:
Haile Selassie I University Press.
Piggott, Stuart. (1983). he Earliest Wheeled Transport from the Atlantic Coast to the
Caspian Sea. Ithaca, NY: Cornell University Press.
Reid, Richard J. (2002). Political Power in Pre-Colonial Buganda: Economy, Society, and
Warfare. Oxford: James Currey.
Rockel, Stephen J. (2006). Carriers of Culture: Labor on the Road in Nineteenth-Century
East Africa. Portsmouth, NH: Heinemann.
Shelford, F. (1920). “Transport in Africa by Road, Rail, Air, and Water.” Journal of the
African Society, 19: 165–75.
Summerhill, William R. (2003). Order against Progress: Government, Foreign Investment,
and Railroads in Brazil, 1854–1913. Stanford, CA: Stanford University Press.
(2005). “Big Social Savings in a Small Laggard Economy: Railroad-Led Growth in
Brazil.” Journal of Economic History, 65: 72–102.
Swindell, Ken. (1992). “he Struggle for Transport Labor in Northern Nigeria, 1900–
1912: A Conlict of Interests.” African Economic History, 20: 137–59.
Szereszewski, R. (1965). Structural Changes in the Economy of Ghana, 1891–1911.
London: Weidenfeld and Nicholson.
homas, Roger G. (1973). “Forced Labour in British West Africa: he Case of the
Northern Territories of the Gold Coast 1906–1927.” Journal of African History, 14:
79–103.
 Reinventing the Wheel 365

Tsey, Christian E. (1986). “Gold Coast Railways: he Making of a Colonial Economy,

1879–1929,” PhD thesis, University of Glasgow.
Warner, Carolyn M. (1999). “he Political Economy of ‘Quasi-Statehood’ and the Demise
of 19th-Century African Politics.” Review of International Studies, 25: 233–55.
Wilks, Ivor. (1989). Asante in the Nineteenth Century, 2nd edition. New York: Cambridge
University Press.
 11

Mbanza Kongo/São Salvador: Culture and the

Transformation of an African City, 1491 to 1670s

Linda Heywood

Since it is necessary to write many times about the city of São Salvador . . . Your
Excellency should not be led to believe that it is a city like those of Europe; I have
no doubt that in the past it was a lot better, given the ruins that exist.1

11.1 INTRODUCTION
During the early decades of the Atlantic slave trade, slaving ports located
on the Atlantic coast of Africa and localities far removed from the coast
served as centers of trade as well as cultural contacts between Africans
and Europeans. Although the economic impact of the trade in human
beings has been the subject of numerous scholarly studies, less is known
about the cultural dimensions of the trade. In fact, in no place in Atlantic
Africa were the cultural ramiications as pronounced and diferent as
in the kingdom of Kongo. Recent works by Kristin Mann on Lagos and
Robin Law on Whydah (Kristin 2007; Law 2004) have contributed signif-
icantly to this issue as they have brought into focus the social dynamics
that led to the rise and transformations of these two important Atlantic
slave ports. Law’s use of the concept of the “middleman community”
to explore the place of port cities in the Atlantic world is a good start-
ing point for the study of Mbanza Kongo, and it also was a slave trade
center. In some respects, however, Law’s “middleman community” does
not fully represent the complexity of the social and cultural dynam-
ics of Mbanza Kongo as the city not only served as an entrepôt for the
slave trade with the Portuguese, but was also the capital of the kingdom.

1
Informazione sul regno del Congo di Fra Raimondo da Dicomano (1798), ed. Arlindo
Correia, http://www.arlindo_correia.com/121281.html, note 3.

366
 Mbanza Kongo/São Salvador 367

Unlike Lagos and Whydah and other port cities whose character was
informed by the trade in slaves, Mbanza Kongo was the administrative
and political capital of one of the most centralized states in Africa at the
time. For more than a century ater the arrival of the Portuguese in 1487,
Kongo kings and nobles were guided not so much by the demands of
the trade in slaves, but by imagining Mbanza Kongo as a new city. he
ideas that they implemented transformed Mbanza Kongo into their ver-
sion of the “Christian city.” During this period, they collaborated with a
host of Portuguese religious and cultural personnel sent by the kings of
Portugal and missionaries coming from Rome to give Mbanza Kongo a
new life and identity. Over time, they so transformed the physical and
cultural landscape of the court that the city came to possess European-
style churches, the city supported a wide range of cultural institutions
that the churches promoted, and the life of the African elite as well as
commoners came to be deined by the rhythms of the Christian religious
calendar. In essence the presence of churches, other Christian institu-
tions, and a cultural life dominated by mixed Kongo-Christian obser-
vances and rites marked Mbanza Kongo as a unique cultural space in
Atlantic Africa. During the period that the city lourished as an Atlantic
city, Mbanza Kongo was not a replica of a medieval European city nor
did its cultural life mimic that of European Christian cities. Even though
the kings and nobles who lived in the city requested priests and Christian
ritual objects from Europe, developed literacy in European languages
with Portuguese as the language of instruction in the schools and con-
vents they built, conducted European-style court proceedings, under-
went Christian baptism, and adopted Christian names and feudal titles,
the European aspects of this new life were in many ways grated on to
Kongo religious ideas, political apparatuses, and preexisting cultural
norms. he Kongo experiment remained under Kongo direction because
the kings and nobles defrayed the economic cost. In the end they cre-
ated a unique Atlantic African Creole culture where free and slave and
noble and peasant were proud to assert the Christian Kongo identity of
the city. he experiment, however, did not last as toward the end of the
seventeenth century the city crumbled under the weight of civil war and
neglect. Dependent on European technical expertise, the model of the
“Christian city” was perhaps unsustainable in the long run. Exploring
the social investments that the Kongo elite put into the project might
ofer early lessons for longue durée understandings of wealth and poverty
in the Atlantic during the era of the slave trade.
368 Heywood

11.2 MBANZA KONGO AND ITS EARLY TRANSFORMATION

he Portuguese were the irst Europeans to establish some of the early
cultural centers in Atlantic Africa – Elmina in 1482 and Luanda in 1575.
Although the Portuguese visited African courts like Benin City in the 1470s
and Mbanza Kongo in 1483, the transformation of the city only began in
1491. When the irst Portuguese visitors arrived in Mbanza Kongo, which
means city (or large central place) of Kongo in Kikongo, the locality served
as the court of Nzinga Nkuwu, the reigning Kongo king. Mbanza Kongo’s
roots go back to the fourteenth century when Nzinga Nkuwu’s predeces-
sors conquered it. Oral traditions identify Mbanza Kongo as the irst place
that the conquerors of what became the Kongo Kingdom settled once they
crossed the River Zaire.2 he invaders chose Mbanza Kongo because of its
strategic location, for it is situated on a lat, well-watered mountaintop some
ten miles in circumference. From the time of its original occupation until
the kingdom became part of Portuguese Angola in 1914, Mbanza Kongo
(called São Salvador from the 1570s) served as both the real and the sym-
bolic capital of the Kongo Kingdom.
Mbanza Kongo already had a distinctive layout, which the earliest visi-
tors to the area noted. he irst description of the physical appearance of
the capital comes from a 1491 letter of the ambassador of Milan to Lisbon,
which noted that the king of Kongo had “a terra [land] encircled by a wall
as large as Evora and is inland, some 60 leagues from the sea, all full of
houses better than are found in all of Guinea.”3 Because Évora was the sec-
ond largest city in Portugal, the fact that Mbanza Kongo was compared to
it suggests that Portuguese visitors were impressed by Mbanza Kongo’s size
and appearance.
We know little more of the physical layout of the capital or of the reli-
gious, political, or economic life of the residents in the period before King
Nzinga Nkuwu greeted the irst Portuguese visitors who arrived at his court.
Mbanza Kongo, however, underwent signiicant physical transformation
from 1491 when the king and members of his court were baptized in a
lavish religious ceremony at his residence to celebrate their conversion to
the Catholic faith. By 1548, when we have the irst estimates of the density
of the population of Mbanza Kongo and the rural population it supported,

2
For the events leading to the establishment of Mbanza Kongo as recorded in modern oral
traditions, see Cuvelier (1934, 65, 85). For earlier traditions and for how these modern
traditions developed, see hornton (2001, 112, 119).
3
Letter of Milanese ambassador in Lisbon, November 6, 1491, in Adriano Capelli, “A
Proposito di Conquiste Africane.” Archivo Storico Lombardo (1896, 416).
 Mbanza Kongo/São Salvador 369

the population numbered about sixty thousand people. From 1491 to 1678,
when the city was temporarily abandoned, Kongo kings oversaw a physi-
cal transformation of the city that in some respects made it appear to have
more in common with Portuguese Luanda than with the capitals of other
African court cities such as Benin City or with port cities such as Whydah
and Lagos. he city’s physical transformation was most noticeable in the
new materials used to rebuild the wall around the court, in the new stone
palace, and in the many stone structures that were built. he density of
the city also increased as new settlements grew up around the court. In
addition to its physical transformation, Mbanza Kongo became a political
and religious ritual center as churches and institutions connected to them
replaced or were built in places that had served as the administrative and
religious centers in precontact Mbanza Kongo (hornton 2000, 67–85).
Mbanza Kongo’s physical transformation began immediately ater the
irst Portuguese cultural mission arrived early in 1491. his mission brought
Portuguese religious teachers and skilled workers, including carpenters,
masons, bakers, Christian women, and “farmers with their tools and a bell
for the church.” In the months following, the king and his nobles mobilized
thousands of Kongo laborers to build a Christian church of stone, the irst
major public work project in the city ater their conversion. housands of
Kongo laborers worked under the supervision of Portuguese cratsmen and
were able to complete the project in record time. hey began on May 6
and completed the building by June 1. he church, dedicated to Our Lady,
was the irst European-like structure in the city, and thousands of stones
and pounds of mortar were used to build it.4 Before this project, the peo-
ple in Mbanza Kongo had never used stone or mortar to construct their
houses or public buildings but instead used wattle and daub and thatch. A
rich trove of letters penned by Kongo oicials, Portuguese residents, and
visitors to the city beginning in the early 1500s describes the continuing
physical transformation of the city as more churches and other stone struc-
tures went up in the following years. he most spectacular of these was the
stone wall built around the court, and the construction of a stone palace
that the laborers began to build for Kongo’s most famous Christian king,
Afonso I (1509–42). During Afonso’s reign, many more stone houses were
constructed for the Portuguese who called the city home.5

4
Rui de Pina, Crónica d’El D. João II, Cap. 61 in Carmen Radulet, O Cronista Rui de Pina e
a “Relação do Reino de Congo” (Lisbon, 1992).
5
King Afonso to Manuel I, October 5, 1514, Monumenta Missionaria Africana (MMA) [Ser.
1, 15 vols. Lisbon, 1952–88] 1:294–323; King Afonso to Manuel I, June 13, 1517, MMA
1:410–11.
370 Heywood

he city’s physical transformation was really the brainchild of King

Afonso, who implemented elaborate plans for reconstructing the city dur-
ing his long reign. He largely inanced the building as well as helped recruit
or gave money to the cratsmen to recruit Kongo laborers, who he intended
would work alongside the Portuguese cratsmen and learn new skills. In
1509, for example, he welcomed masons and carpenters, who King Manuel
of Portugal had sent to build churches, as well as materials and other prod-
ucts that came from Portugal with the several Portuguese cultural missions
that included Portuguese farmers, teachers, and skilled workers such as
stone masons and carpenters. Moreover, he paid for upkeep of the skilled
workmen who also came from Portugal to rebuild his residence so that
stone buildings could replace the thatched ones. One of the most impres-
sive results of this Kongo-Portuguese cultural and technical collaboration
was the building of a mile-long stone wall, twenty feet high, that separated
the royal residences from the rest of the city. Although in 1514 Afonso
complained that the masons had not completed his stone residence despite
the money he paid them because they were more interested in trading in
slaves, he pushed for the building of more stone structures in the city. For
example, to make sure that his relatives and other nobles in his household
were exposed to the new knowledge the priests from Portugal brought, he
“gathered together all our brothers and children and cousins and the chil-
dren of our criados in such a way that there were 400 young men and chil-
dren and . . . we ordered a great wall made with many thorns on top so they
could not climb it and lee and we entrusted the fathers with teaching them,
and we also ordered other walls made . . . for the fathers to live together.”6
In addition, he ordered the construction of walled enclosures for priests
and a large walled enclosure with houses for the “1,000 sons of nobles with
their teachers and dependants so that the priests could teach them read-
ing, writing, grammar and the things of our holy faith.”7 In fact, Afonso
was so concerned that the Kongo students he had recruited to study under
the Christian foreigners found ways of escaping from the walled enclosures
that he asked the king of Portugal to turn over to him the island of São
Tomé so he could have schools built there where he could send “male and
female” Kongo students who could study and would not have the incentive
to run away and return to their families.8

6
King Afonso to Manuel I, October 5, 1514, MMA 1:299–306.
7
“De Algumas Coisas que Tocam ao Rei do Congo,” 1516, MMA 1:374.
8
King Afonso to Manuel I, October 5, 1514, MMA 1:299–306.
 Mbanza Kongo/São Salvador 371

Figure 11.1. Garcia II with the Dutch (1668) depicting the 1641 meeting.

Construction of stone buildings was not limited to Afonso’s palace; he

also gave permission to the Portuguese to build a somewhat smaller wall
around the stone houses that they were building. As a result of Afonso’s
building initiatives, the physical landscape of the city was rapidly chang-
ing. For example, between the walled enclosures that he had built he let
a large open space in the city “where the principal church was built.” Later
every nobleman would build his own city residence, although each of them
would also have villas in the villages outside the city (Pigafetta, Lopes, and
Grassi 1591, 39–40). Afonso was very concerned about the upkeep of these
buildings, and in 1517 and again in 1526 he asked the king of Portugal to
send him more stonemasons and carpenters to complete the churches that
were under construction. For example, in 1526, when only one carpenter
remained in the city, he wrote to King João III requesting ive masons and
twenty carpenters to inish one of the churches, called Nossa Senhora da
Victoria. He noted that he only had one carpenter let who “who can repair
and cover the churches”9 (Figure 11.1).

9
King Afonso to João III, August 25, 1526, MMA 1:475, 479.
372 Heywood

he basic layout of the city would remain the same in the decades follow-
ing Afonso’s reign, although later kings built additional churches, schools
and convents, cloisters, and other buildings for the missionaries and nobles
built residences to ensure that they had a place in the center of power. By
the late 1500s, Mbanza Kongo (by then known as São Salvador) contained
several churches, the most prominent being the Cathedral built in 1549
and located in the large square where outdoor masses were held to accom-
modate thousands of parishioners. he central place that the churches and
other religious institutions occupied set the city apart from other port cit-
ies in Atlantic Africa. Mbanza Kongo’s physical transformation continued
in the 1600s. In the early 1600s, for example, land was set aside to build a
monastery for the Dominican missionaries, while in 1632 Alvaro IV gave
the Jesuits, who were beginning their work in the city, a site for a college
as well as “other lands for their upkeep.” He requested that they build “a
church for the college that can be visited any time and continuous alms that
the people could give.”10
Afonso, and the kings ater him, allocated funds not only for the building
of the stone structures and the payment of the workers, but for the upkeep
of the churches and other buildings they had constructed. he income
to maintain these buildings and their stafs increased as the number of
churches grew. At the end of Afonso’s reign in 1542, there were already
between six and eight churches in the city, while by the late 1590s the num-
ber had risen to twelve. In 1595, Antonio Viera, the Kongolese ambassador
to Lisbon, noted that although there were only six churches in the city, many
nobles kept their own private chapels.11 hese churches were all dedicated
to Christian saints and were located throughout the city. Both the Church
of St. James and the Church of Our Lady of the Conception were located
within the palace grounds, while the Nossa Senhora de Ajuda Church was
located in the Portuguese quarters. he Jesuits also had three unidentiied
churches and a cloister in the part of the city that they occupied. By 1642,
when a Dutch embassy visited the city, they were able to list by name at least
ten churches, although some were already in ruins and others in need of
repair. At that time four of the churches – St. James Church, the Church of
the Holy Ghost, the Church of St. Michael, and the Church of St. Joseph –
were located within the palace walls while the others – the Cathedral, the
Church of Our Lady of the Conception, the Church of Victory, the Church

10
Alvaro IV to General of the Company of Jesus, October 25, 1632, MMA 8:199–200.
11
“Interrogatorio de Statu Regni Congensis Facta Ulissipone. Anno Domini 1591” MMA,
3:503.
 Mbanza Kongo/São Salvador 373

Figure 11.2. Dapper São Salvador (1668) situation in 1641.

of Seven Lamps, and the Church of St. Anthony – were located in various
parts of the city. Walls also separated some of the churches from the sur-
rounding area. For example, the Church of Our Lady of the Conception
was surrounded by a stone wall in an area that also contained two very
large residences built with thatch, giving the appearance of a convent.12 he
city by this time had named streets as well. Some, like St. James Street, took
their names from the churches that stood on them, while another street
called Bacas de las Almadias (Gulf of the Canoes) was located near a river
(Ogilby 1670, 525). Olfert Dapper’s engraving of the city as it appeared in
1641–2 depicted several of the prominent features of the city, including
the Cathedral, the two-story palace, the wall, and several other churches
(Figure 11.2).
he lifestyle of city residents also changed as celebrations connected to
the Christian religious calendar came to dominate city life. As the social
experiment continued, all aspects of life in the city, from politics to eco-
nomics and recreation, created demands for trained personnel and for the
goods required to support the activities of the church. Church personnel
from Europe, along with the religious items that informed church life, were

12
Relacão Dos Carmelitas Descalços (1584), MMA 4:396.
374 Heywood

Figure 11.3. Kongo clothing – Olfert Dapper (1668) depicting the situation in 1641.

high in demand, and the kings spared no efort in acquiring the where-
withal to pay the expenses of the European and African church personnel
that stafed the churches and for the variety of Christian ritual objects that
came from Europe. As the city grew, it became increasingly divided into two
social groups. On the one hand was the king, his court, nobles who were
recognized by their mixed European African clothing, their literacy, and
their memberships in the various religious institutions; on the other hand
was a large population (almost half of the city residents) who were slaves
and freemen only marginally integrated into the new lifestyle. Although
in many ways the social distance between the two groups diverged, in
many ways the social experiment brought slaves and freemen closer to the
members of the nobility. he European missionaries and secular European
(mainly Portuguese merchants) and church personnel who lived in the city
also added another social dimension to city life. hey had their own quarter
with a wall as well (Figure 11.3).
he city’s position as an Atlantic Creole center was evident in the promi-
nent role the Kongo Catholic Church and the institutions connected to it
had in the life of the residents of the city. Church services and rituals con-
nected with the Christian calendar dominated the culture of the city. Weekly
services and celebrations connected to saints’ days attracted large numbers
 Mbanza Kongo/São Salvador 375

of the city’s residents as did the yearly celebration associated with St. James
Day, which by 1584 had become Kongo’s national day. In fact, a 1619 a
report commenting on the life of the Christians in São Salvador noted that
so many people wished to attend weekly services that the Cathedral could
not accommodate the “innumerable multitude of inhabitants” (Cuvelier
and Jadin 1954, 47) and that the king along with the rest of the popula-
tion usually “stand outside the church in the vast square to hear mass.” his
Christian community was not limited to Africans, for the Portuguese popu-
lation had grown as well. Many of the one thousand Portuguese in the king-
dom had a residence in the city.13 By 1645, the mixed African-Portuguese
residents in the city had also increased and numbered a total of four hun-
dred people, several of whom were priests.
he expenses that went to maintain this Atlantic Creole center increased
as well, although it is impossible to come up with a detailed accounting of
the percentage of the kings’ revenue that went into building and maintain-
ing the institutions. Nevertheless, a range of descriptive evidence suggests
that the kings did not spare any resources. In the early years they spent
liberally to pay in kind for the satin and velvet and other clothes, books,
bed canopies, door curtains, and religious and cultural items that came
from Portugal, as well as for the wages of the skilled Portuguese workmen
who worked in the various construction projects. hese in-kind payments
included slaves and a wide variety of Kongo products and animals such as
silver and copper manilas, leopard skins, parrots, civet cats, and the like.
hey also covered the upkeep of the churches and their personnel. Because
Kongo had an actual currency before the Portuguese arrived, some of the
payment was made in it. In Kongo the people used zimbu shells ished from
the shallow waters of the island of Luanda, and the kings and nobles used it
to assess taxes and to pay for the goods and services they received. he high-
est unit of account was the cofu, which was worth twenty thousand zimbu
shells, while the lefuco was worth ten thousand zimbu shells (Cuvelier 1946,
307–10). he Portuguese workers who came during Afonso’s time were not
only paid in the local currency, but were paid in kind. Over the years, the
kings sent thousands of slaves, ivory, manilas, local cloths, and the like to
their Portuguese counterparts (and at times to Rome) to help cover the cost
for priests, the religious icons, and other items that came from Portugal and
Rome. hey also set up accounts in Portugal and Kongo to allow transfers
of money from one to the other.

13
For the activities of the Portuguese in Kongo, see Heywood and hornton (2007, 139).
376 Heywood

Some idea of the expenses that went into the building projects in the
early years of the city’s transformation comes from a letter that Afonso
wrote to King Manuel of Portugal in 1514 in which he complained that the
masons who came from Portugal refused to work and instead “demand pay
all the time [for] building the house for me and my queen.” Slaves who the
king gave to the representatives of the Portuguese as gits to the Portuguese
kings or free Kongo subjects who the Portuguese illegally enslaved also rep-
resented part of the payments from the kings’ revenues because slaves were
valuable commodities among both the Kongolese and the Portuguese. he
various complaints that Afonso and later kings made against the Portuguese
who came to work in the city but who oten found it more proitable to use
their income to unlawfully enslave or purchase Kongo subjects also repre-
sented part of the cost connected with the physical layout and cultural life
of the city.14
Part of the revenues that the Kongo kings used to cover the Christian
project also came in the form of the labor hours that Kongo workers put
into the building projects. hese laborers gathered the stones from several
miles outside the city and transported them to the city center where the
churches were built. Kings also handed over large amounts of land and the
laborers on them to the priests who managed the churches. Afonso and the
kings ater him oten referred in their letters to the thousands of laborers
they recruited to work on the building projects.
he resources that Kongo kings plowed into the building projects repre-
sented only a part of the expenses they invested to transform Mbanza Kongo
into an Atlantic Creole city. Over the years, as they deepened their cultural
ties with fellow Christian monarchs in Portugal and with the Papacy, Kongo
kings allocated a part of their revenues to cover the cost of the recruitment
and upkeep of the European experts – mainly priests and other religious
personnel who they believed were essential to implement their vision. All
of the foreign experts along with the many Kongolese they trained lived at
one time or another in the city and were in part supported by funds that the
kings and nobles provided.
During Afonso’s time when few institutions existed for training mem-
bers of the local elite, he provided the funds for the travel to and the
upkeep of his relatives in Portugal. In fact, in 1540, he actually arranged
to have ive thousand cruzados on credit from King João in Portugal to
cover the cost of his relatives who he had sent to Portugal to study and

14
King Afonso to Manuel I, October 5, 1514, MMA 1:306.
 Mbanza Kongo/São Salvador 377

to cover the expenses of an embassy he had sent to Rome.15 He promised

to pay him back in the local currency that the Portuguese agents could
use to purchase slaves. As more missionaries traveled to Kongo to man-
age the churches and to help in the education and training of local men
and women, kings and nobles continued to cover the cost. For example,
soon ater the arrival in the city of a group of missionaries from Europe
in 1548, one of them wrote back to his colleagues noting that King Diego
I ordered his nobles to “bring us great gits” to help with their transi-
tion.16 Moreover, Diego also paid for the salary of Portuguese linguist
Pedro Alverez and carriers who he supplied to missionary Father George
Vas in 1548. Several decades later, Alvaro II, writing to Pope Paulo V in
1613, attempted to outline his eforts to assess the cost of the upkeep of
the bishops and the canons at the Cathedral. He insisted that the king of
Portugal should bear some of the cost, because “the chaplains and curates
instituted by his grandparents are paid each year from his royal income.”17
Similarly Alvaro III, writing to Rome in 1615 to request missionaries,
took pains to assure papal oicials that he would be able to pay for their
upkeep as the “kings my predecessors” did. He noted that earlier kings
always gave the royal chapel to their confessors “with lands, rents, vas-
sals and Lordship . . . which renders a good 2,000 cruzados,” and prom-
ised that this amount would apply “from now for ever to that dignitary.”
Later kings made similar inancial commitments. In the 1623 invitation
that King Pedro II made to the Jesuits in his attempt to encourage them
to return to the city to work, he took pains to remind them of the many
churches that his predecessors had built, noting that they had “instituted
and endowed the church of St. James . . . to serve as a royal chapel.” He also
spelt out the funding he planned to spend on the mission, agreeing to set
aside “400 cofus of zimbo” in perpetuity for a chaplain major and for “suc-
cessors in the oice” to run the operation.18 Likewise, in 1632, King Alvaro
IV reminded the head of the Jesuits in Rome that the kings of Kongo had
funded the college that they had established and that he himself had given
them “not only the site for the college but also other lands for farms for
the rearing and producing of animals and for planting.”19 Later kings also
did not waver when it came to the inancial upkeep of the churches and

15
King Afonso I to João III, April 4, 1540, MMA 2:100–2.
16
Christvão Ribiero, August 1, 1548, MMA 15:162–3.
17
Alvaro II to Pope Paulo V, February 27, 1613, MMA 6:128–32.
18
“Instructions of King Pedro II, King of Congo. . .” July 5, 1623, MMA 7:33–8.
19
Alvaro IV to General of the Company of Jesus, October 25, 1632, MMA 8:199–200.
378 Heywood

Figure 11.4. Garcia II welcoming the Capuchins. his is from Cavazzi (1687) p. 336.

their personnel. In 1652, King Garcia blamed some of his immediate pre-
decessors for allowing four of the churches built in the early years to fall
into ruins and vowed to “restore them and rebuild them from their foun-
dation.” In fact, he impressed Father Giacinto da Vetralla, the Capuchin
prefect, with his commitment, even though the missionary concluded
that “this would cost a lot, because it would be necessary to make them
and then transport them here.” At the same time, Garcia employed a large
number of laborers to rebuild the Church of St. Michael for the use of the
Capuchins.20 Unfortunately, by the 1660s, because of the disastrous civil
war many more churches fell into ruin as tensions among the elite led to
the neglect of the churches. A 1660 Capuchin report relecting on the sev-
eral ruined churches blamed the Portuguese in Angola, who they accused
of preventing the Kongos from importing “lime” and other materials used
to repair the churches and other buildings because of the fear they would
use them to build “fortiications.” In 1678, when the city was temporarily
abandoned, São Salvador’s status as a showpiece of architecture and social
engineering was doomed as the city was already a shell of what it was dur-
ing its heyday (Figure 11.4).

20
Giacinto da Vatralla ao Procurador Geral, 20 August, 1652, MMA X1:221–3.
 Mbanza Kongo/São Salvador 379

11.3 ATLANTIC CREOLE CULTURE IN MBANZA KONGO

Despite all their eforts (Afonso even at one point wrote that he had sui-
cient numbers of Kongolese trained as carpenters and masons who could
work on repairing the church buildings), Kongo kings and nobles were
never able to train enough local experts to have a cadre of profession-
als who could maintain the physical and cultural infrastructure and the
cultural knowledge so necessary to maintain Mbanza Kongo’s place as a
vibrant Atlantic Creole city. hroughout the period when the city lour-
ished they continued to rely on foreign experts. Kings and nobles did all
they could to provide inancial support not only to the churches and the
many chapels, but also to defray the cost of the several dozen European and
Afro-European priests, the bishop, canons, and other church oicials who
came to the city from Europe. hey also paid the cost for the training of the
thousands of Kongo-born lay ministers (mestres de escola) who performed
the crucial work of bringing the new religious ideas to the population in
the city and elsewhere. Afonso’s commitment to the churches and the very
pious life he lived laid the foundations for shaping the Christian character
of the city and its inhabitants. Most of the Kongo kings who followed him
also worked to maintain the Christian character of the city and to promote
it as the Christian capital of the kingdom. Christianity took root in the city
with the annual arrival of missionaries from Portugal and Rome who con-
centrated on teaching the people, from kings and nobles to slaves, the core
principles of Christianity. he number of foreign priests and secular clergy
was never enough and luctuated from as high as ity to as low as ive (in
1649, there were sixteen European missionaries in the entire kingdom). To
ofset the shortfall, Kongo maintained a system that trained a great num-
ber of Kongolese as lay teachers. he kings provided the resources for the
training of this local laity who carried out the day-to-day teaching and
other duties of priests, even though they did not do baptisms. Although the
kings and nobles and members of their households were the irst people in
the city to be baptized, over the years the majority of the city’s residents,
whether slave or free, were baptized. In fact, Inquisition accounts from
the 1570s suggest that most Kongos who lived in the city (and indeed the
entire kingdom) were being instructed in the teachings of the church.21 In
the early years the baptisms were mostly mass conversions, as occurred in
1548 when on March 17 the Jesuit Father Vas reported that he had baptized

21
Arquivo Nacional de Torre do Tombo (ANTT),Inquisição de Lisboa, Processo 2522, fol.
144, Testimony of Francisco de Medeiros, 1584.
380 Heywood

twenty-seven hundred people who attended the several churches, some of

them elderly Kongos between sixty and eighty years old. All twenty-seven
hundred who received baptism lived in or immediately around the city. By
the 1650s, missionaries were baptizing as many as four thousand to ive
thousand babies per year, and by the 1670s, most people who lived in the
city in fact were baptized and considered themselves Christians.22 By the
1640s, although there were still mass baptisms in the city, most people who
were baptized would have received some form of religious education from
the many priests who oiciated at masses in the city’s several churches
or from the lay Kongo and mixed Kongo-Europeans who served as their
teachers. By the early seventeenth century, as the city grew in population
and social complexity, its layout and social makeup made Mbanza-Kongo
very diferent not only from Luanda, but also from the capitals of other
African kingdoms of the time. Seventeenth-century residents and visitors
oten commented on the layout and makeup of São Salvador. For example,
in 1609, Antonio Viera, the confessor of King Alvaro II, testiied in Lisbon
that the city had ten thousand houses and that the bishop’s residence was
located next to the Cathedral, which had a parish church, a college, and
a hospital attached to it (Cuvelier and Jadin 1954, February–April 1609).
By the mid-1620s, when Jesuit missionary João de Paiva arrived in São
Salvador, he described the Portuguese quarter of the city, noting that the
city was surrounded by “villages and neighboring hamlets” (Franco 1627,
249, paragraph 17). Descriptions from the early 1630s noted that the Jesuits
had their own quarters in the city as well. Olfert Dapper, describing the
city as it was when the Dutch embassy arrived in 1645, noted that it con-
tained streets and pointed out that the public square (mabzi), which was
located between the enclosed court of the king and the Portuguese enclo-
sure, had become a central place for celebrations. Even when the number
of residents began to decline in the 1650s, the city kept its status as a center
of Christianity in the kingdom. Antonio Cavazzi, writing about the city in
the early 1670s, put the population at a mere ive thousand people. Even by
the late 1670s, when most of the remaining population led because of the
increased ighting among royal factions, the city remained the symbolic
religious center of the kingdom and was eventually reoccupied in 1705
(Cavazzi 1965, 80).

22
“Carta do Padre Jorge Vaz ao Padre Simão Rodrigues,” August 1, 1548, MMA XV:151–2;
Giacinto da Vetralla ao Procurador Geral di Capucini, August 20, 1652, MMA X1:221–3;
“Descrição das necessidades do reino do Congo,” Instituto Historico e Geograico
Brasileiro, DL 846, 16.
 Mbanza Kongo/São Salvador 381

Mbanza Kongo’s culture was dominated by the religious and secular

culture that developed around the kings and nobles. In the early years the
pious Christian demeanor of King Afonso stood out against the Portuguese
experts, who he felt disrespected him, and the nobles, who were still deeply
attached to the old pre-Christian ways and rituals even though they had
converted. In time, however, as the Kongo nobility blended the Christianity
that the Europeans brought with elements from their pre-Christian ritu-
als and beliefs, life in the city comprised the vibrant cultural mixture that
characterized Mbanza Kongo’s Atlantic Creole character. he most impor-
tant aspects of this culture were the celebrations that were connected with
the coronation of kings, religious services that the churches, convents, and
lay brotherhoods organized on Sundays and other religious days in the
Catholic calendar, the new burial rites that included a mixture of Kongo
and Christian rituals, and the central place of religion in Kongo politics.
he city was the center of all of these developments. For example, the city
served as the venue for the coronation of every new king following Afonso
as the elaborate ceremony had to take place in the Cathedral with the
participation of the bishop or a European priest. In fact, throughout this
period no king was considered legitimate unless he came to the city to be
crowned. During the coronation ceremony, the new king was required to
take an elaborate oath in front of the thousands of city residents and Kongo
notables and their followers who traveled to the city from the provinces to
witness the event. Because European religious authorities played such a cru-
cial role in proclaiming the king’s legitimacy, reigning kings or nobles who
were eligible to be elected as king made a point of staging elaborate public
ceremonies in the city to welcome European missionaries. Furthermore,
the king, his nobles, and members of their households always attended the
weekly masses held in front of the Cathedral, led the many public penances
that the churches encouraged, and participated in the religious celebrations
to honor the feast days of various saints. Because the Cathedral was built
on lands used as the burial site during the pre-Christian period, burials
in the Cathedral were as important as coronations. Burials of kings and
nobles were major public events, as all kings and nobles, no matter where
they died, were brought back to the Cathedral to be buried in elaborate
Christian ceremonies (hornton 1992, 57–63). As important as the burials
were the annual Christian celebration of All Souls’ Day and celebrations
honoring the ancestors (pre-Christian and Christian), which also took
place in the Cathedral. he detailed descriptions of the burials of Pedro
II and the coronation of Garcia I, which both occurred in 1624 with the
mix of the knights of the Order of Christ as well as the presence of people
382 Heywood

carrying the traditional blacksmith whistle, spoke to the unique European

and Kongo cultural mix (hornton 1992, 57).
he experience of the irst Carmelite mission, which arrived on the out-
skirts of the city on October 28, 1583, provides an excellent example of
how important it was to the kings to use the city as a center of grandi-
ose political and religious pageantry. When Alvaro I, who was reigning at
the time, received the news that the members of the mission had arrived,
he ordered them to spend the night outside the city while he prepared a
great procession to welcome them and to receive the statue of the Virgin
Mary and other relics that they had brought from Rome. he missionar-
ies duly complied, and waited until the following day with “some relics
of the Eleven housand Virgins, and a bone of one of the Ten housand
Martyrs.”23 During that time Alvaro was able to assemble more than thirty
thousand persons, who formed a grand procession to lead the group into
the city. his all took place without the king being present, because he
had an ailment that prevented him from walking, although he was able to
view the procession and image from his palace.24 Such public gatherings
to welcome new missionaries or to attend religious festivals became com-
monplace for the residents of Mbanza Kongo. In 1619, for example, when
missionaries arrived from Rome bringing an altar and “the ten thousand
medals, crown and similar things,” the crowd that gathered to witness the
blessing of the medals was so large that most people had to remain outside
the Cathedral. In fact, by the irst decades of the 1600s, the Cathedral did
not have enough space to accommodate the large crowds that attended
regular Sunday services, and these took place in the square in front of the
Cathedral. In 1631, for example, missionary Francisco de Soveral reported
that Sunday services and religious festivals were always held outdoors.25
Again in 1651 King Garcia II was so pleased that Pope Innocent X had
sent him an oicial letter granting him several indulgences that he had
requested that he “decreed that there be great celebration in the city, which
lasted three months, all to show how pleased he was for the great extraor-
dinary favors he received from the Pope.”26

23
Diego del Sanctissimo Sacarmento, “Relation del viage de Gvinea que hiço el Padre Fray D
d S,” 1583, MMA 4:365.
24
“Relatione di quello che occorse, videro nel Regno di Congo tre Religiosi Carmelitani
Scalzi. . .,” MMA 4:395.
25
[Francisco do Soveral] Infra scriptam relationem de statu Cathedralis ecclesiae Sancti
Saluatoris . . . April 1, 1631, MMA 8:13–18.
26
“Interrogatorio de Statu Regni Congensis Facta Ulissipone. Anno Domini 1591,” MMA,
III:502–3.
 Mbanza Kongo/São Salvador 383

Another interesting development that made Mbanza Kongo diferent

from other Atlantic Creole cities was the important place of the various
religious brotherhoods in the city’s religious and political life. During this
period, every Kongo king and male member of the nobility belonged to one
of the various lay religious brotherhoods that missionaries established in
the city. he irst brotherhood in Mbanza Kongo was founded sometime
before 1585, and the numbers dramatically increased in the early 1600s
when several groups of Jesuit and Dominican missionaries arrived and the
king honored their request to open more brotherhoods. By 1609, Antonio
Viera noted that six confraternities operated in the city and the king and
nobles supported them and that “every day masses are celebrated for the
souls of the dead.”27 In fact, in 1612 King Alvaro II was so eager to show the
newly arrived Dominicans that he was a devoted Christian that he agreed
to their suggestion to found “the Brotherhood of the Rosary,” and actually
oversaw the irst procession of nobles and people in the city who paraded in
a way similar to the “way it was done in Portugal.” In time, the activities of
the various religious brotherhoods were intimately connected to royal pol-
itics as rival factions of the nobility attached themselves to one or another
of the brotherhoods. he experience of the Jesuits provides a telling com-
mentary on how the kings and nobles used the brotherhoods to settle polit-
ical rivalries. In the early 1620s King Pedro invited the Jesuits to return to
work in the city because, as he explained, he needed to strengthen Christian
morality, which he believed was necessary to protect the kingdom from
the Portuguese, who were sending armies into Kongo’s southern provinces.
One of these invasions had proven disastrous to the Kongo province of
Mbamba, where the invaders had killed and enslaved thousands of Kongo
subjects. According to Jesuit João de Paiva, who later became the rector of
the college the Jesuits built, “King Pedro . . . [believed] that only men of the
Society of Jesus could placate God for the Congo people. . . . [And] he gave
letters to our rector at Luanda, to send him some of the Society in order that
they erect a stable home in the royal city.”28
Pedro and the kings ater him gave their full support to the Jesuits, but
they soon realized that Pedro had an ulterior motive as he intended to use
their presence to support his rather weak claims to the Kongo throne. In
fact, when the Jesuits sent some of members of the order to the city to start
construction of the college, Pedro insisted that he should be considered the
founder of the college. He not only selected the best site, but ofered the

27
Ibid.
28
Franco, Synopsis, 1726, paragraph 4, p. 235.
384 Heywood

brothers food and supplies and “was oten present so that by seeing him the
workers might be urged in their work.”29 he irst church that the Jesuits built
in São Salvador was dedicated to “Jesus and King Garcia” who succeeded
his father Pedro and provided the funds to continue the construction.30
By the late 1620s, the Jesuits had erected their own residence, opened
schools for children of the nobility, and opened brotherhoods in almost all
eight churches that were functioning in the city. he presence of the Jesuits
actually led to more intense political divisions among members of the Kongo
elite, the Portuguese, and the rest of the city’s population. One such frater-
nity was the Brotherhood of the Holy Trinity, which restricted its mem-
bership to twelve-year-old unmarried “noble adolescents” and Portuguese,
while another Jesuit-supported brotherhood located in the Church of the
Holy Trinity was dedicated to St. Ignatius and restricted membership to the
nobility “either married or bereaved of a spouse.” Fraternities proved an
excellent way of spreading the Atlantic Creole culture of the nobility to the
rest of the population as the members of the brotherhoods were required
to take a vow “to teach the Christian doctrine to uncultured and ignorant
men especially to the members of their own household,” to take commu-
nion once a month, and to “engage in other practices of Christian piety.”31
Not to be outdone, noblewomen in the city pressed their husbands and sons
to organize a brotherhood for them, and by 1628 Mbanza Kongo boasted
a brotherhood whose members consisted of “whatever matrons are more
illustrious by birth . . . queens and wives of former kings of these and the
rest were moved to virtue and holy morals.”32 In time, the religious festivals
that the brotherhoods organized to honor their patron saints attracted large
crowds to the city center.
here was an underside to the crucial role that brotherhoods came to
have in the city. By the 1630s, Kongo members of the brotherhoods divided
themselves not only according to the various religious orders that funded
their brotherhoods (Jesuits, Dominicans, and Capuchins), but also accord-
ing to their political allegiance. During wars and other periods of politi-
cal inighting, members of the brotherhoods oten took political sides, and
might support one or another contender for the throne. Moreover, at times
the religious and political ailiations of leading members of the brother-
hoods led to open warfare in the city.33 For example, in 1628 members of

29
Ibid., paragraph 10, p. 235.
30
Ibid., paragraph 23, p. 242.
31
Ibid., paragraph 17, p. 249.
32
Ibid., paragraph 20, p. 249.
33
For an overview of the competition between Rome and Portugal, see hornton (1984). For
a general treatment of European missionaries in Kongo, see Saccardo (1982–3, vol. 1).
 Mbanza Kongo/São Salvador 385

one brotherhood whose members were suspected of plotting against the

king were chained and dragged through the city streets, while a member
of another brotherhood was among ity conspirators executed for plotting
against the king.34 Despite the political intrigues, the religious activities of
the brotherhoods contributed to the social fabric of the city.
he inancial support that Kongo kings put into the physical transfor-
mation of the city during more than a century and a half of transforma-
tion gave rise to two social groups. he city boasted a population made up
of European missionaries, other European residents (mostly Portuguese),
and members of the Kongo ruling elite, many of them descendants of past
kings, as well as a mixed race population. he Kongo members of the group
were literate in Portuguese and even in Latin, bore Portuguese and African
names that indicated their noble status, owned and sold slaves, dressed in
a mixture of European and Kongo clothing, and devoted a lot of their time
and resources to activities connected to the church. he members of the
other group, comprising the bulk of the city’s population, were the Kongo
commoners, both free and enslaved. hey provided the labor on which the
other groups relied, but were oicially excluded from the social and other
privileges that the members of the elite enjoyed.
Despite their diferences, both groups shared a common Atlantic Creole
Christian identity. City residents, whether they were noble or enslaved,
called themselves Moxicongo, a term that distinguished them from most of
their non-noble rural counterparts (mubhata) who lived beyond the city
or in the provinces. he markers of noble Moxicongo status were literacy
in Portuguese as well as the use of Kikongo, Christian baptism, regular
church attendance, church burials, and opportunities to participate in the
several celebrations that took place. Familiarity with these aspects of city
life separated Moxicongos from their rural and provincial counterparts
(hornton 1983).
he urban environment that nurtured this Moxicongo identity was largely
owing to the inancial investment that Kongo kings and nobles put into the
project to attract European secular and religious experts to Mbanza Kongo.
heir investment did not create a city that was a copy of its European coun-
terparts as the foreign missionaries hoped or that Afonso and some of the
other kings imagined. By the 1640s, the churches and other buildings in
the city represented a blend of Kongo and European elements, constructed
with stones and mortar but with thatched roofs. Moreover, the services held
in the Cathedral and other places of worship did not replicate services in

34
Franco, Synopsis, 1726, paragraph 22, p. 253 and paragraph 16, 1726, p. 256.
386 Heywood

Europe. In fact, on many occasions European priests substituted palm wine

for imported wine during communion, and candles made from palm oil
oten took the place of imports from Europe. he choirs that sung hymns
in Latin might be accompanied by musicians who played both European
and African musical instruments. Furthermore, more Kongolese in the
city learned church doctrine from Kongo lay preachers who were speaking
Kikongo than from the lips of the foreign religious experts as the Kongolese
served as interpreters when priests were around or as lay ministers for the
long years when the people saw no priests (Heywood and hornton 2007).
his cultural blend was most visible in the celebrations that took place
in the city. On St. James Day, the most important religious and cultural
celebration in the city, thousands of Kongos from the provinces locked to
the city eager to participate in the public celebrations to honor St. James,
Kongo’s patron saint, who they believed helped King Afonso defeat his
pagan brother in 1509. Despite its religious origin, the day was also one
when people paid respect to the kings by bringing the traditional tribute
and performing the ensanga/sangamento, the pre-Christian military dance,
and otherwise engaging in non-Christian rituals.35 Furthermore, on All
Souls’ Day when Kongo Christians gathered in the Cathedral and in the
many churches or went to outdoor localities with candles and other ritual
objects to honor the dead in Christ, they also honored the non-Christian
ancestors as the Cathedral and some of the churches were built on pre-
Christian burial sites.
he slave trade and slaveholding that became a part of the city’s eco-
nomic and social fabric slavery never dominated Mbanza Kongo as it did
the Atlantic port cities of Quidah and Lagos. Although slavery predated
the physical and cultural transformation of the city, between 1491 and the
1670s slavery in the city had more in common with the slaveholding pat-
terns found in other Atlantic Creole cities. Whereas in the early days of the
city’s transformation most of the slaves held there were captives brought in
from conquests, by the seventeenth century, slave trading and slavehold-
ing were commonplace. Most members of the elite – Kongos, Portuguese,
missionaries, and merchants – sold and owned slaves. Many commoners
also owned a slave or two. Slaves did not comprise a separate class in the
city and many found creative ways to participate in the city’s culture (da
Roma 1648, 69, 78). he inequalities associated with social transforma-

35
For a discussion of an eighteenth-century celebration of St. James Day, see hornton
(1998, 31–5).
 Mbanza Kongo/São Salvador 387

Figure 11.5. Ruins of the Episcopal See, Mbanza Kongo, 1940. (Antonio Brásio,
Monumenta Missionaria Africana, (1954) vol. IV, p. 192).

tion, however, were more pronounced in the 1670s than in 1491 when the
Portuguese arrived (Figure 11.5).

11.4 CONCLUSION
Although Afonso and the kings who followed him imagined and partially
succeeded in transforming Mbanza Kongo/São Salvador into their own ver-
sion of a Kongo Christian city, by the 1670s most of the stone structures they
had built and some of the cultural institutions they had supported were dis-
appearing from the landscape. In fact, modern critics of the Kongo experi-
ment oten dismiss the early initiatives of the Kongo elite, even arguing that
the Portuguese duped Afonso to such an extent that they turned Kongo into
one of the major emporiums for the slave trade in central Africa. In fact,
they oten regard the resources the rulers expended on the infrastructural
and cultural projects as early examples of Africa’s foreign dependency, as
a city such as Mbanza Kongo was not organic and could not be sustained
without the constant presence of foreign experts. his kind of critique has
been leveled at many postcolonial governments whose foreign experts built
many white elephants that now litter the landscape of many countries. his
thinking has been encouraged since Portuguese colonial control, which
began in the 1880s, also led to the physical erasure of the experiment as
388 Heywood

stones forming the ruined buildings (except the Cathedral) were removed
to build new colonial structures. Despite the disappearance of the city,
however, the Mbanza Kongo experiment does provide early lessons for
the longue durée understandings of wealth and poverty in the Atlantic. he
most remarkable impact of the experiment was not its eventual failure but
what it tells us about the imaginings of the Kongo nobility to inance and
spearhead a cultural revolution even as the slave trade, conlicts among the
nobility, and Portuguese conquest would obliterate their vision. Perhaps the
one lasting by-product of the Mbanza Kongo experiment was that it gave
rise to a new Kongo identity based on notions of kingship and Christianity.
he hundreds of thousands of Kongos who were enslaved and carried to
places such as Brazil brought this Kongo identity with them. his Kongo
identity survives as folklore in contemporary Brazil in the congada, which
has as its central activity the crowning of the Kongo king. Furthermore, an
even more politically important legacy of the Mbanza Kongo experiment
is the pride that modern Kongos, who are citizens of independent Angola,
place in their Kongo identity and the place of Mbanza Kongo in their his-
tory. hese Kongos are proud of Mbanza Kongo’s Atlantic Creole past, are
conscious of where the ruined churches used to be located, and would cer-
tainly disagree with the idea that the resources that the kings invested into
transforming Mbanza Kongo were misplaced.

References
Cavazzi da Montecuccolo, Giovanni Antonio. (1687). Istorica Descrizione de’ tre Regni
Congo, Matamba ed Angola. Bologna.
(1965). Descrição Histórica dos Três Reinos do Congo, Matamba e Angola. Lisbon.
Cuvelier, Jean. (1934). Nkutuma a Mvila za Makanda. Tumba, Congo: Diocèse de
Matadi.
(1946). L’ancien royaume de Congo; fondation, découverte, première évangélisation
de l’ancien royaume de Congo, règne du grand roi Afonso Mvemba Nzinga (1541).
Bruges: Desclée de Brouwer.
Cuvelier, Jean and Louis Jadin. (1954). Le ancien Congo d’après les archives romaines
(1518–1640). Brussels: Visita ad Limina, 1619.
da Roma, Giovanni Francesco. (1648). Breve relatione del svuccesso della missione de
Frati Minori Capuccini del Seraico P.S. Francesco al Regno del Congo. . . . Rome:
Sacra Congregatione de Propagande Fide.
Franco, R. P. Antonio. (1726). Synopsis Annalium. . .in Lusitania, 1540–1725. Augustae:
Veith, 1726.
Heywood, Linda and John hornton. (2007). Central Africans, Atlantic Creoles and the
Foundation of the Americas, 1585–1660. Cambridge: Cambridge University Press.
Mann, Kristin. (2007). Slavery and the Birth of an African City: Lagos, 1760–1900.
Bloomington: Indiana University Press.
 Mbanza Kongo/São Salvador 389

Law, Robin. (2004). Quidah, the Social History of a West African Slaving “Port,” 1729–
1892. Athens: Ohio University Press.
Ogilby, John. (1670). Africa. London: homas Johnson.
Pigafetta, Filippe, Duarte Lopes, and Bartolomeo Grassi. (1591). Relatione del Reame di
Congo e delle Circonvicine Contrade. Rome: Grassi.
Saccardo, Graziano. (1982–3). Congo e Angola con la storia del antica missione dei cap-
puccini, 3 volumes. Milan: Curia Curia Provinciale dei Cappuccini.
hornton, John K. (1983). Kingdom of Kongo: Civil War and Transition, 1641–1718.
Madison: University of Wisconsin Press.
(1984). “he Development of an African Catholic Church in the Kingdom of the
Kongo.” Journal of African History 25: 147–67.
(1992). “he Regalia of the Kingdom of Kongo, 1491–1895.” In Kings of Africa: Art
and Authority in Central Africa, Collection Museum für Völkerkunde Berlin, edited
by Ema Beumers and Hans-Joachim Koloss, 57–64. Berlin: Maastricht: Foundation
Kings of Africa.
(1998). he Kongolese Saint Anthony: Dona Beatriz Kimpa Vita and the Antonion
Movement. Cambridge: Cambridge University Press.
(2000). “Mbanza Kongo/São Salvador: Kongo’s Holy City.” In Africa’s Urban Past,
edited by David M. Anderson and Richard Rathborne, 67–84. Portsmouth, NH:
Heinemann.
(2001). “he Origins and Early History of the Kingdom of Kongo, c. 1350–1550.”
International Journal of African Historical Studies 34(1): 112–19.
 PART III

INSTITUTIONS
 12

he Fragile Revolution: Rethinking War and

Development in Africa’s Violent Nineteenth Century

Richard Reid

12.1 INTRODUCTION
he central concern of this chapter is sub-Saharan Africa’s nineteenth
century, and its place in the continent’s long-term development. I argue
that although war had long been both a driver and a product of inno-
vation and consolidation in African history, the nineteenth century is
of particular importance because it witnessed a major transformation in
the organization and deployment of violence. Much of the new warfare
involved imported technology, expanded strategic horizons, new tacti-
cal formations, and new forms of organization, with signiicant political
and economic implications. his was tantamount to a revolution in mili-
tary afairs, which also had political and economic facets. It was, in sum,
a golden age of extraordinary vitality and creativity, as well as a markedly
violent one. Scholars have increasingly neglected this period, however,
placing ever greater emphasis on modern conlict at the expense of deeper
understanding of the role and impact of warfare in Africa’s past. Popular
notions endure surrounding Africa’s violent savagery in some imagined
precolonial past, and have prevented a reconsideration of the role of war
in African development over the long term, as well as of the links between
past and present conlict. his chapter aims to make a contribution to that
debate. I suggest that in the course of the nineteenth century – in diferent
places at diferent times, and for a variety of reasons – there were indica-
tions that this transformation in military afairs was moving in the direc-
tion of new and enlarged forms of political cohesion, increased economic
capacity and diversiication, and heightened levels of social mobility. In
other words, the revolution in violence involved development, according
to the most basic deinition of the term.

393
394 Reid

However, this developmental process did not go far enough, especially

politically – political reform usually lagged behind military advances –
while economic systems remained fragile and vulnerable to both endog-
enous and exogenous pressures. In part, this must be seen in the context
of certain perennial limitations – notably, environmental and economic
challenges and uneven demographic spread – on the scope and scale of
revolutionary warfare in Africa that slowed the rate of change, and indeed
prevented wholesale internal reform in some cases. Ultimately, European
imperial incursions – facilitated in large part, ironically, by Africa’s own
violent transformation in the course of the preceding century – engen-
dered further instability, and sharpened, indeed institutionalized, internal
conlict.

12.2 DECOLONIZING WARFARE

In many respects, the history of African warfare is one of the last frontiers
of decolonization: the study of the role and impact of organized violence
in Africa’s deeper past, certainly over la longue durée, remains a minority
sport. While scholars have carried out ine work over the past forty years
or more on various aspects of precolonial war,1 it has never represented
the academic mainstream. here are various reasons for this, no doubt.
One letist tendency of the generation responsible for the establishment of
African history as a professional discipline during the 1960s and 1970s –
against the backdrop of an unpopular war in Southeast Asia – was its dis-
inclination to foreground violence and the military in its work. he shit
toward “history from below,” moreover, militated against the history of war,
with its emphasis – at least before its 1980s relaunch as “war studies” – on
great men and decisive battles, and African studies was in the cockpit of this
new historiography. But perhaps most important of all – and certainly most
germane to the argument in this chapter – was that to study war in African
history was to push against the developmental agenda in academic and pol-
icy circles, already building in momentum in the course of the 1960s and
1970s. A concentration on violence was inappropriate vis-à-vis more main-
stream work on political dynamism, social innovation and creativity, and
economic development – while at the same time, from the 1970s onward,
most of this kind of work was being done not on the precolonial past at all,
but rather on the colonial era, which increasingly dominated the scholarly

1
A very basic reading list would include Ogot (1972), Mazrui (1977), hornton (1999),
Smith (1976), and Reid (2007). For a representative sample, see also Lamphear (2007).
 he Fragile Revolution 395

horizon. Historians were more interested in heroic economies than military

heroism. his is not to say that no one worked on it: a number did, for a
while, as even a cursory glance at the irst decade of the Journal of African
History reveals. But it was not sustained. In the end, few self-respecting his-
torians of Africa could ponder warfare without the odorous associations of
nineteenth-century missionaries and humanitarians lingering about them,
for ater all, to look seriously at war would mean, presumably, having to
concur that sometimes nineteenth-century Africa was really rather violent.
hat is what the racialists of the 1880s and 1890s had emphasized, and so
it would not do.
Moreover, warfare, it seemed, was the one realm of activity on which the
European Scramble for Africa had imposed itself decisively: industrial-
era armaments swept aside an array of African armies and militias, and
dancing and painted warriors were thus consigned to distant and quaintly
tribal memory. In certain circles, some of them survived – the Zulu, for
example, or the Sudanese “Dervishes” – as examples of savage but some-
how honorable warriors bravely but foolishly waving their spears in the
direction of modernity. here was something compelling in this image,
for military history bufs, but not for most Africanists, who had long
since let that particular battleield behind. hey were now increasingly
interested in what came ater, in those aspects of creeping anticolonial
protest and identity that, some believed, were central to understanding
Africa’s own modernity, and that, not coincidentally, involved “history
from below.”
Not only has this development privileged the period of the Scramble for
Africa as a military moment, positioning it as a point of departure between
the past and the present in terms of African violence, but it also, by impli-
cation, denies the broader developmental role of warfare in Africa’s earlier
past. hose scholars who have worked on the subject have done enough
to indicate how important war was to African states and societies seeking
to accumulate people, create broader pools of factor endowments, access
commerce, and enhance social cohesion – in sum, to build stronger and
more competitive economies and polities. Oten, it was bloody and messy,
and it did not work, but sometimes it did, and either way it needs and
deserves much closer and more sustained attention. Meanwhile, the veri-
table explosion of literature on modern African wars since the mid-1990s2

2
In recent years, see, for example, Clapham (1998), Reno (1998), Ali and Matthews (1999),
Richards (2005), Kaarsholm (2006), Boas and Dunn (2007), Derman, Odgaard, and
Sjaastad (2007), and Nhema and Zeleza (2008a, 2008b).
396 Reid

is notable in one signiicant respect, namely the strikingly ahistorical, fore-

shortening approach of much of this material to the subject. A great deal
of it, indeed, appears to be driven by solution-oriented, presentist research
agendas. he overwhelming focus on the modern has enabled scholars to
restrict themselves to thinking in terms of speciically postcolonial mal-
aise. hey do so at the expense of a long-term approach – in other words,
a consideration of the role of war in Africa’s deeper past and of the links
between past and present. To be sure, this is a risky strategy, for it will
inevitably involve highlighting African failures as well as achievements,
and the destructive as well as the creative in Africa’s history. But without
it, our understanding of the recent past is at best one-dimensional. A per-
vasive presentism in intellectual culture is much in evidence across the
disciplines, but it is perhaps tautological to point out that it is in history
that it does the most damage.
he central focus of this chapter is the nineteenth century, an era of
extraordinary vitality in political, economic, and military afairs, which
has nonetheless sufered from a certain scholarly neglect in recent years.3
he nineteenth century was, in the 1960s, the centerpiece of much of
the new African historiography; since then, it has gradually become
mere prologue to the main action – namely, that of the colonial period
and beyond. My purpose here is to contribute to its rehabilitation as the
foundational era of modern African history, and also to resuscitate and
develop the idea of a nineteenth-century African military revolution, irst
proposed, in a somewhat diferent form and context, by Nigerian histo-
rian Godfrey Uzoigwe more than thirty years ago (Uzoigwe 1977). hat
argument quickly izzled out – the trend was against both military history
and indeed the study of the precolonial period – but a wealth of evidence
supports it.
Inspiration can be drawn from European models, in particular the
ongoing debate surrounding the nature of the putative “revolution in mili-
tary afairs” and its relationship with the rise of European statehood and
attendant economic development.4 A cursory glance at the literature on
European warfare reveals a deeply embedded and widely accepted thesis
that a dynamically reciprocal relationship existed between state formation,
economic development, and organized violence in European history. It

3
A notable, though somewhat disappointing, exception in recent times is Coquery-
Vidrovitch (2009).
4
For example: McNeill (1982), Tilly (1992), Parker (1996), Howard (2009), Kennedy
(1988).
 he Fragile Revolution 397

may ultimately be considered a successful combination, with the predict-

able caveat that of course success is always relative, and its measurement
debatable. he notion of the military revolution in European history has
long been the subject of debate among scholars;5 the precise details of that
debate need not detain us in the course of this chapter, but only those key
aspects – particularly in terms of the relationship between war and state
formation – that appear to have relevance for the African experience. he
dangers of Eurocentrism in drawing on this body of scholarship are clear
enough; I hope they have been avoided here. he most obvious risk is that
in keeping the European experience in view, we think of war and political
and economic development in terms of linear progression, and consider
therefore that something has gone wrong in Africa – in other words, the
European experience becomes the template. Sometimes, war is simply not a
matter of linear progression (or at least it does not yet appear to be); more-
over, one of the central points of this chapter is that Africa has had its own
particular set of problems and issues, and the African experience needs to
be understood within distinctive parameters.
Finally, it is worth noting that work on African economic development
in historical context frequently neglects the role of war whether as cause
or as consequence.6 his chapter is not about economics – but it is about
the contexts within which economic change takes place, the drivers of such
change, and the economic aspirations of key actors. Suice to note here that
the picture is a patchwork: at times, organized violence is clearly the out-
come of economic decline or catastrophe and seeks to rectify the latter; at
others, war itself is the central driver of economic change, either bringing
about economic growth or being inimical, ultimately, to such growth. It is
certainly evident in nineteenth-century Africa that it could be both simul-
taneously – with war driving certain types of economic expansion, but not
others; facilitating particular trajectories of political development, but not
others.

5
A useful summary can be found in Henniger (2006).
6
Forty years ago, Jack Goody began to consider the comparative history of African and
Eurasian societies with respect to the means of both production and destruction, but little
came of the debate (Goody 1971). In John Ilife’s otherwise superb history of African pov-
erty, the index contains no entry for warfare, violence, conlict, militarization, or any other
term suggesting an analytical link between war and economic depravation (Ilife 1987).
Ralph Austen’s grand survey of African economic history does discuss “warfare and raid-
ing” at various junctures, but in a way that only suggests the need for a great deal more
work (Austen 1987). As Tony Hopkins’s survey article shows, warfare scarcely features in
recent analyses of African economic history (Hopkins 2009).
398 Reid

12.3 NINETEENTH-CENTURY TRANSFORMATIONS

IN VIOLENCE
he beginnings of our putative revolution – or, perhaps more accurately,
revolutions – are not easy to nail down with any great precision. A number
of the key features of the transformations in violence most closely associated
with the nineteenth century were already in place in the eighteenth in some
areas – most notably along those stretches of Atlantic Africa most intensely
involved in the slave trade. Between ca. 1500 and ca. 1800, the violence of
the slave trade drove the formation of new states and societies: the instabil-
ity of the Atlantic zone destroyed some communities – Kongo was an early
victim – yet others thrived on the opportunities for violent entrepreneur-
ialism, for example Dahomey, Oyo, Asante. he human tragedy notwith-
standing, this was an era of remarkable political, economic, and ultimately
military dynamism, leading to the formation of durable states with vigorous
military cultures. Alongside territorial monarchies, there were mobile asso-
ciations of professional warriors, such as the Imbangala, who thrived on –
indeed manufactured – much of the violent disorder around them. hese
would be key features of the nineteenth century, already present in the sev-
enteenth and eighteenth. But the claim made for the revolutionary nature
of the nineteenth century is based on the dramatic proliferation of such
highly militarized states and communities; on the transformative efects of
irearms, which had a muted impact in most places before the early and
mid-eighteenth century; and above all, perhaps, on the widespread emer-
gence of armed entrepreneurs and adventurers responsible for the creation
of an array of new communities – some relatively durable, others less so,
but almost always on the frontiers of contest and exchange – characterized
by military professionalism, material ambition, and, in many cases, ideo-
logical ardor.
his was a long nineteenth century in many places, reaching from the
last third of the eighteenth century to the 1910s and 1920s, when colonial
rule engendered a temporary pax across much of the continent, and change
occurred throughout this era, coming later in some places than in others,
and at times more intense than at others. Nonetheless, the period between
the 1780s and the 1820s can be seen as a transitional moment, when a series
of events and processes signiied dramatic change. he Ngoni conlagration
in southeast Africa, which would give rise to the Zulu state; the disintegra-
tion of the Oyo Empire; the Fulani-Hausa jihad; Buganda’s apogee as driver
of regional military and economic change; the collapse of the Solomonic
state in Ethiopia: these developments were mostly unconnected, but they
 he Fragile Revolution 399

each clearly encapsulated issues germane to much of Africa as a whole in

the century to come. Violence was deployed in ever more innovative and
concentrated ways in pursuit of political and economic objectives. Soldiery
was becoming increasingly professionalized; war provided new opportuni-
ties to young men for the achievement of socioeconomic and indeed polit-
ical status. In this sense, broad comparisons can be drawn with Europe’s
early modernization process from the late iteenth century into the six-
teenth, involving the transition from the feud of chivalry to more profes-
sional, salaried armies and the attendant rise, in the sixteenth century, of
the powerful prince, embodying – and carving out by force – sovereignty.
Africa’s nineteenth-century warlords were doing much the same thing, in
the process frequently articulating ideas about just and legal war – again, as
in sixteenth- and seventeenth-century Europe.
New military organization appeared, including new forms of politico-
military leadership, while irearms became increasingly common and drove
a range of tactical and structural innovations. Guns had been present along
the Atlantic coast, and in the Ethiopia Highlands, since the early sixteenth
century, but their impact on tactics and organization was minimal for at least
two hundred years. Across the West African savannah and in the deeper
sub-Saharan interior, meanwhile, guns had little or no impact whatsoever.
Only from the early eighteenth century, with the arrival in increasing num-
bers of more efective lintlock muskets, did irearms become important,
and only in the course of the nineteenth century did they become weap-
ons of revolutionary change across the continent. In this context, indeed,
a useful comparison can be made with the European experience between
the mid-sixteenth and mid-seventeenth centuries, during which time the
use of portable irearms by the Dutch and the Swedes meant the need for
more professional, better trained soldiers – one of the central planks of the
European military revolution thesis (Black 2009, 64–5, Henniger 2006, 8).
It is impossible to provide detail from across the entire continent; instead,
examples may be selected to illustrate the central idea. To begin with, devel-
opments in coastal West Africa and in the East African interior – among
the Yoruba and the Nyamwezi respectively – encapsulate some of the core
themes. In the irst instance, it was the collapse of an ancien regime – the
Oyo Empire – as well as the local impact of more gradual shits in the global
economy that led to a transformation in the nature and structure of war-
fare. In the second, external commercial pressures likewise played a key
role; but rather than the destruction of an older state system – which had
been notably absent in the early history of the Tanzanian interior – there
were novel attempts at state building and the militarization of political
400 Reid

culture as a result. Oyo’s hegemony in the region of the Benin Gap dated to
the late seventeenth century; by the early years of the nineteenth century,
however, diminishing proits from the slave trade – Oyo’s major source of
revenue for the better part of two centuries – meant the inability to buy
horses from the north, which meant in turn increased internal taxation to
meet the shortfall. But a weakened military and restive centrifugalism fun-
damentally destabilized the Oyo polity, which was dealt a further blow by
the advance of Sokoto’s jihadist forces from the north (Law 1977). he con-
sequence of Oyo’s collapse in the 1820s and 1830s was a century of violent
reform (Smith and Ajayi 1971). he product and then the engine of much
of this change was the Yoruba city-state, self-contained, fortiied, and linked
together by a complex and periodically shiting network of alliances and
antagonisms. Between the 1830s and the 1870s, Ibadan was the dominant
force in Yoruba politics, and its rise came at the expense of Ijaye in particu-
lar, which was largely destroyed in the early 1860s. From the late 1870s, an
alliance crystallized around opposition to Ibadan – now positioning itself
as the successor of Oyo – and included the Egba, centered on Abeokuta,
and the Ijebu, situated close to the coast near Lagos and a key commercial
power in the region. hese prolonged wars facilitated the rise of military
leaders – warlords, in modern parlance – who sometimes had links with
the old Oyo system, but who were oten self-made men who established
new military aristocracies; and they were at the head of increasingly pro-
fessional forces, comprising young men – the “war boys” – equipped with
both matchlocks and more potent lintlocks. Signiicantly, too, these forces
were oten professional mercenaries, hiring themselves out to the highest
bidder and ighting on behalf of their chosen city-state until better material
rewards were ofered elsewhere. War had become a career choice, and oten
a highly lucrative one; and the skills in battleield maneuver and, in par-
ticular, use of irearms were both sharpened and increasingly in demand.
Commercial upheaval – the expansion of the illegal slave trade and the
more general widening of commercial opportunities – loosened the bonds
of older loyalties and conventions and unleashed new generations of armed
youth onto political and economic markets in search of adventure, material
improvement, and social standing.
It was this explosion of novel forces for change – frequently driven by,
and certainly involving, a younger generation eager for new opportunities –
that also deined eastern Africa in the mid- and late nineteenth century.
he expansion of the slave and ivory trades from the early nineteenth cen-
tury onward prompted new ways of organizing violence and heightened
levels of militarization. Between Lake Victoria and the Zambezi valley, the
 he Fragile Revolution 401

Nyamwezi and the Yao were in the vanguard of these violent transforma-
tions, of both political and commercial change, between the 1850s and the
1890s: among these groups, new forms of military leadership emerged,
around which new communities coalesced – oten highly mobile, certainly
luid, and frequently transient.7 As with the Yoruba “war boys,” young men
were becoming an increasingly professionalized soldiery – known among
the Nyamwezi as ruga ruga – ever more adept in the arts of lintlock mus-
ketry and noted for their speed of deployment and maneuver. he ruga ruga
were oten little better than thugs and criminals, but more broadly they
formed part of a self-perpetuating cycle of sociopolitical reform, for mil-
itary states both created the disorder that released young manpower from
more traditional obligations and then utilized that manpower in the quest
for economic proit and political development. hey formed part of a new
generation of aggressive and displaced youth – oten joined by slaves sim-
ilarly armed with guns and with loyalty to a new class of entrepreneurs
of violence with enlarged visions of what was economically and politically
possible. Nyungu-ya-Mawe among the Kimbu; Msiri in Katanga; Tippu
Tip to the west of Lake Tanganyika: these were restlessly brilliant military
leaders who traded and raided simultaneously, overthrowing old politi-
cal structures and processes and seeking to build new polities designed to
take advantage of the commercial opportunities opening up around them
(Bennett 1971; Reid 2007; Roberts 1968). here was no better illustration
of the turbulence of the age than Mirambo, whose Nyamwezi state had
emerged by the end of the 1860s and dominated much of the Tanzanian
interior down to his death in 1884. Mirambo harnessed the youthful, vio-
lent exuberance of his ruga ruga and pursued larger political and economic
objectives, including those of unity and stability. He partially and tempo-
rarily achieved the former, but never the latter, and his state – Urambo –
collapsed soon ater his death. Nonetheless, these experiments in political
and economic reform, highly militarized and underpinned by the novel
utilization of violence, were symptomatic of developmental shits across
nineteenth-century eastern Africa. he fortiied towns that grew up across
the eastern African interior – in themselves a striking innovation, given
the previous pattern of dispersed settlement in an underpopulated region –
were for protection in the irst instance, but they became centers of trade
and industry and political business, too. Even Tippu Tip’s rough-hewn
empire in the eastern Congo – based on slave militias doing the business
of ivory trading and slave trading – represented an advance, if an unsteady

7
Alpers (1975); Reid (2007).
402 Reid

and profoundly unattractive one, in economic organization, and certainly

embodied a new kind of aggressive, dynamic entrepreneurialism.
Such privatized violence – under personal leadership and increasingly
involving mercenary professionals – was in evidence elsewhere, too. In
west-central Africa and along the southern Atlantic coast, the Ovimbundu
and Chokwe created entrepreneurial militarisms similar to those in the
eastern region, for which military professionalism and commercial adven-
turism (the export of slaves) were the twin elements in a process of socio-
political transformation – even if, again, the roots of that process were in
the pre-1800 era. In the Zambezi valley, groups of guns for hire coalesced
into a series of mercenary armies centered on leaders who were merchants
and warlords simultaneously. One common theme of particular signif-
icance was the harnessing and organization of armed youth with larger
aims in view; this was also discernible separately, among, for example, the
Turkana and Maasai (Lamphear 1998; Spear and Waller 1993). Here, age-
grade structures signiied the growing militarization of youth – in the dry
grasslands of the Rit Valley, it may have been driven in the nineteenth cen-
tury by environmental shits – under the direction of new forms of char-
ismatic centralizing leadership, oten provided by prophets and diviners.
Neighboring Somali and Oromo had used age regimentation to good efect
for some time.
Developments in the southeast of the continent also it with the dynamic
model discernible in the Atlantic and eastern zones. During the 1790s and
1800s, competition for land among the Ngoni chiefdoms was escalating fol-
lowing a period of population growth. Increased levels of warfare prompted
the use of age regimentation as a means of both sociopolitical and military
organization among the key centralized states of the region, and by the mid-
1810s, a commander named Shaka – from the comparatively unimportant
Zulu clan – came to the fore as the most prominent exponent of the new
systems and tactical formations. Under Shaka’s leadership, the Zulu state
emerged as the key driver of sociopolitical and above all military change in
the region. he Zulu fought in close formation using short stabbing spears,
deployed with fearsome speed, and were noted for their deadly use of lank-
ing maneuvers to crush opponents. he new state, in place by the beginning
of the 1820s, was based on the skillful and brutally disciplinarian manage-
ment of youthful male aggression; the regiments, the amabutho, were used
to tend cattle while not actively deployed in war. Defeated opponents were
absorbed into the regimental system (or were killed in the attempt), break-
ing down older patterns of loyalty and underpinning a novel style of milita-
rized kingship. he mfecane – a term referring loosely to the era of violent
 he Fragile Revolution 403

upheaval – led to the migration of a number of groups out of the area for
safety, and thus military transformation was exported beyond the imme-
diate area.8 New polities and societies – some at more than one remove
from the original Ngoni upheaval – emulated Zulu organization and tac-
tics: the Ndebele, the Sotho, the Bemba. Other Ngoni groups wandered
north through Malawi and into Tanzania by the 1840s and 1850s, and the
communities they encountered frequently copied their military structures
and ethos: the Hehe, the Bena, the Sangu. he Ngoni – or more accurately
groups that were the product of contact with the Ngoni – almost certainly
inspired the young Mirambo, as such itinerant and armed communities
entered northern Tanzania in the course of the 1850s. Firearms became
increasingly important to these polities ater the mid-nineteenth century.
Some of these groups at length settled down, demilitarized to some
extent, and took up agriculture alongside the keeping of cattle; others main-
tained a more or less predatory way of life down to the 1890s. he Zulu state
itself struggled for stability through much of the nineteenth century follow-
ing Shaka’s assassination in 1828, both in terms of external relations – spe-
ciically with European settlers, present in the area from the second half of
the 1830s, as well as with increasingly well-armed African neighbors – and
internal political arrangements, especially as succession was rarely uncon-
tested, while restless regiments remained a headache for the kingship. here
was therefore enormous variation in terms of local lavor and direction of
travel; nevertheless, taken in the round, a veritable revolution in military
afairs had unfolded across an enormous swathe of southern, central, and
eastern Africa between the 1790s and the 1850s – and this rolling revolu-
tion had intersected with separate, equally explosive, developments in cen-
tral eastern Africa in mid-century. It amounted to a stunning patchwork
of violence across a wide area, creative and brutally destructive in equal
measure. his was a military transformation, moreover, that involved eco-
nomic consolidation and diversiication and social stratiication, as polities
of various hues were driven by material objectives and aimed to maximize
internal productive and external commercial opportunities.
So far, we have dealt – by and large – with wholly novel political forms,
new creations born of transformative violence. In other cases, states of
long standing – that is, with their roots in an earlier era – also experienced
dramatic military change in the course of the nineteenth century, their

8
It might be noted in passing here that some twenty years ago there was debate over whether
the mfecane happened at all, at least in the manner traditionally depicted: see Cobbing
(1988). For a more consensual approach, see Hamilton (1992, 1998), Wylie (2006).
404 Reid

innovations in violence taking place within broadly preexisting political and

territorial parameters. Dahomey, rooted in the seventeenth-century slave
trade, was one such polity – and indeed one of the most successful preco-
lonial militarisms anywhere in Africa (Law 1991; Smith 1976). It expanded
considerably its range of military operations in the course of the nineteenth
century – including westward, into Yoruba country, following the collapse
of Oyo – and was one of the worst ofenders in the illegal slave trade down
to the 1850s, whereupon it turned increasingly to the export of agricultural
produce. Its strikingly martial culture was arguably more monolithic than
that of Asante, another monarchical state with its roots in the seventeenth
century. Asante was more dextrous in its balancing of commerce and war,
and its military success was at least matched by its highly efective diplo-
matic machine – although Europeans, particularly the British, continued to
see Asante, as they did Dahomey, as the embodiment of the archetypically
savage, bloodthirsty African empire, particularly in the later nineteenth
century (Adjaye 1996; Wilks 1975).
On the other side of the continent, Buganda reached the apex of its mil-
itary strength between the 1780s and the 1820s, from which position it
sought to dominate the wider lacustrine region as well as the commercial
system encroaching on it. he Ganda innovated accordingly, militarizing
aspects of the political establishment – explicitly military chietaincies were
created and a martial culture was further underpinned – and importing
irearms in increasing numbers. As the limitations of war on land became
ever clearer, the Ganda developed a leet of canoes with a view to rendering
Lake Victoria a Ganda sea, thus increasing their military and commercial
reach (Reid 2002). In terms of internal military reform, Bunyoro was at
least the equal of its old rival Buganda, for in the 1870s new regiments of
rilemen – the barusura – aided the kingdom’s resurgence in the 1880s and
1890s, a resurgence only crushed by British arms (Doyle 2006). Bunyoro’s
deployment of irepower was more efective than that of the Ganda – the
barusura bear comparison with the ruga ruga and the Yoruba “war boys,”
including their brutality and the license to ravage granted them by their
founder, Kabalega – for in Buganda a gun culture at the royal court served
to undermine the professional military ethos as irearms became more
symbols of sociopolitical standing than indications of new military skill.
Nonetheless, militarization in both Buganda and Bunyoro served to heighten
their mutual antagonism on the eve of colonial rule; this was a conlict into
which the British would be co-opted in the 1890s, with profound implica-
tions for the wider region (Low 2009). Important, too, both Buganda and
Bunyoro sought economic expansion as part of their programs of military
 he Fragile Revolution 405

transformation – internally, through the development of existing resources,

and externally, through the capture of commercial opportunities.
A further variation on the theme of state building through war in the
nineteenth century can be discerned in the western African savannah and
in the Ethiopian Highlands; for here, new states were formed through vio-
lence that laid claim to ancient ideological and religious inheritances, thus
rendering their violence righteous. Revivalist Islam had long facilitated
political and military innovation in the Saharan world; as recently as the
early and mid-eighteenth century, jihads in Futa Toro and Futa Jalon had
demonstrated the possibilities violent theocracy ofered for the creation of
new communities. But the Sokoto Caliphate – founded as a result of jihadist
war among the Hausa states in the irst decade of the nineteenth century –
was the most remarkable manifestation of armed Islam to date (Roberts
1987; Smaldone 1977). Sokoto’s infantry ranks were soon joined by cav-
alry forces to great efect, and by the late 1810s the armies of the faithful
were pressing in on Yoruba territory to the south and exporting both tactics
and ideology across the savannah. Bornu, notably, reformed its military to
check the advance of Hausa and Fulani forces. By the mid-nineteenth cen-
tury, irearms – long eschewed by the highly mobile warriors of savannah
and desert – were common and were deployed with considerable success
by Umar Tal’s armies during the Tukolor jihad of the 1850s. Samori Ture’s
Mandinke forces used both horses and guns in the 1880s and 1890s, and
were serviced by cratsmen who learned how to repair and modify ire-
arms and to make ammunition (Legassick 1966). Sokoto, meanwhile, had
become a major economic power, in addition to a political and ideological
presence, across the wider region.
he violent renewal of Islam drove political, military and economic
change in the region – Sokoto became a major economic power as well as
a religious one – and a broadly comparable set of processes were at work in
the Ethiopian Highlands. Here, however, it was ancient Christian nation-
alism that was at work in the formation of the highly militarized neo-
Solomonic state between the 1850s and the 1890s. he polity claimed to
be the recreation of an older Christian empire in the region, the original
Solomonic state founded in the late thirteenth century that inally collapsed
amidst violent regionalism in the 1760s and 1770s. As with the atermath of
the fall of Oyo among the Yoruba – and indeed the jihadist challenge to the
Hausa states – the disintegration of an ancien regime would produce new
forms of strident militarism and the innovative organization of violence
as ethno-regional units, albeit oten porous, consolidated and competed
over both trade and ideological inheritance. his was the period known in
406 Reid

Ethiopian historiography as the zemene mesaint – the “era of the princes.”

he “new” Ethiopia that eventually emerged from these conlicts was largely
the work of a triumvirate of leaders in the second half of the nineteenth cen-
tury, namely Tewodros, Yohannes, and Menelik, each of whom, in diferent
ways, understood the transformative power of violence and the political
and economic possibilities in its usage. he Ethiopian empire that was cre-
ated by the 1880s and 1890s – encompassing the agricultural and mineral
wealth of newly conquered lands to the south – was the outcome of several
decades of military reform, expansionist violence, and a dynamic military
culture that lauded righteous Christian violence in the face of Muslim ene-
mies who pressed in on the kingdom’s lanks (Reid 2011). Ethiopian armies,
moreover, made arguably more efective use of irearms – some of which
were the best Europe had to ofer, imported via Djibouti – than any on the
continent. he Sokoto Caliphate and Solomonic Ethiopia were the products
of ideological revolutions inspired by ancient faith, but they were under-
pinned by innovative violence and involved transformations in the rather
more secular realms of economy and polity.
Let us pause briely and consider the degree to which comparisons between
Africa and Europe make for something of a mixed bag. Debates are perhaps
to be had over whether Tewodros or Mirambo or Shaka embodied much of
the same skill and ambition of, say, Maurice of Nassau or Gustav Adolphus
or (Shaka’s exact contemporary) Napoleon, in terms of military reorgani-
zation; whether the Zulu or the Amhara in the nineteenth century were
like the Prussians of the seventeenth. Indeed, nineteenth-century European
commentators were keen to establish the same points of reference, witness
Henry Morton Stanley’s description of Mirambo as the region’s Frederick
the Great (Stanley 1899, 386, although he was probably more of an Oliver
Cromwell, militarily if not politically). Whether such debates are ultimately
of any use is another issue, while the question might reasonably be asked
(and indeed has been asked): Why do any of them have to be compared to
any European igure? But, as one of the aims of this chapter is compara-
tive history, the niceties of historical relativism can be dispensed with and
broad contours noted. he African experience of military professionaliza-
tion, not least owing to the widespread adoption of modern irearms, is
broadly comparable to that of Europe, as are the proliferation of mercenary
forces and the rise of military overlords who were armed sovereignty incar-
nate. But while irepower demanded discipline and control, not all African
societies had quite perfected this by the 1880s or 1890s. In Europe, state sys-
tems ultimately provided order, but in Africa, political mobility persisted in
many areas. Much of nineteenth-century Africa looks not unlike Europe
 he Fragile Revolution 407

during the hirty Years’ War in the irst half of the seventeenth century, in
terms of complex regional relations, prolonged and brutal levels of violence,
and attendant political and military innovation. But there had been few
“Westphalia moments” in Africa by the late nineteenth century, even if the
trajectory of violence in many areas suggested that such political settlements
were within reach. And few of Africa’s wars were as crushingly expensive.
In the atermath of Westphalia, Europe’s “wars of mercenaries” became the
“wars of merchants”: funds from business and commerce were needed des-
perately to pay for armies and enlarged military systems (Howard 2009).
While there is evidence of this in Africa – among the Yoruba, the Nyamwezi,
the Amhara, in Sokoto – this process was far from completion by century’s
end, although again there was a clear movement in this direction. Arguably,
too, mobility and instability across Africa – not least in terms of succession
and territory – prevented something similar to Europe’s eighteenth-century
shit from mercenaries and privateers to patriotism and mass revolutionary
war. We return to some of these contrasts a little later. he inal point to
note here, however, is that both Europe and Africa had become highly mili-
tarized by the late nineteenth century. he European “nation” was a military
beast indeed, its roots in war, its outlook bellicose, its sense of self markedly
martial; and it now intersected with not dissimilar, but nonetheless distinc-
tive, processes of militarization further south.

12.4 THE VORTEX OF VIOLENCE: EUROPE

AND THE REVOLUTION
Much of the discussion of late nineteenth-century European imperialism
has rested on technological advances, understandably; but care is needed
to avoid technological determinism, namely the belief that technical inno-
vation in itself profers battleield superiority on the innovator. he second
half of the twentieth century has repeatedly demonstrated the fallacy of
that notion; and it is demonstrable for a much earlier period, too. As Black
points out, for example, there was no fundamental shit in the military bal-
ance between the West and the rest much before the mid-nineteenth cen-
tury; Europe’s innovations in irepower frequently loundered in African
and Asian settings in the seventeenth and eighteenth centuries (Black 2009,
65). Yet Bayly makes the more generally accepted point succinctly: that in
the course of the eighteenth century, Europe became “generally more mobi-
lised for warfare and more ruthless in prosecuting it than Asians, Africans,
and Polynesians.” In part this could be explained by the particular demands
of war in the modern European context – the need to adapt rapidly between
408 Reid

land and sea, cramped spaces requiring ever more investment in techni-
cal innovations to achieve success. More eicient systems of tax, again,
were needed to inance this kind of organized violence, and those systems
underpinned modern state formation (Bayly 2004, 81). Yet it was not that
Africans were any less mobilized for warfare; rather, as Bayly himself puts
it, more efective linkages between war and inance meant that “Europeans
became much better at killing people” (Bayly 2004, 62).
In the latter years of the nineteenth century – and rather earlier in some
areas – the looming external threat drove further military innovation
across the continent. Amhara, Ganda, Nyoro, Nyamwezi, Zulu, Yoruba, and
Fulani leaders saw in the advance of the European imperial frontier both
danger – though some came to this position rather later than others – and
inspiration, in terms of organization of irepower, order, and discipline, and
the economic systems that underpinned the military. North Africa provides
the most dramatic examples of so-called defensive modernization, in which
elites sought to emulate the military-industrial model provided by Europe
in order to ward of the threat posed by it. he perils of succeeding too
well were clear enough, however: witness the case of Muhammad Ali, who
between the 1810s and the 1840s transformed Egypt into a major military
and economic power – using European expertise, inance, and inspiration
as appropriate – but who was subsequently cut down to size by a European
coalition that considered that he had become rather too powerful. Flattery,
in the form of imitation, was not enough to prevent European interven-
tion – as a host of African leaders would discover in the decades that fol-
lowed. Few were as efective as Muhammad Ali, but most would come to
appreciate that the attempt to modernize militarily – in reality, represent-
ing only the latest stage in Africa’s ongoing process of military reform –
only elicited an aggressive response from a Europe determined to manage
Africa’s modernity on its own terms.
he great irony is that the innovation across Africa in the nineteenth
century, as well as the traumas and turbulence of the age, provided Europe
with a critical point of entry and considerable leverage across the continent.
Without Africa’s violent transformation, there would likely have been no
“scramble” – or at least there would have been a greatly modiied form of
European intervention and inluence. As it was, European imperialism was
drawn into a series of vortices of violence; and while the details of the ironic
outcome known as the Scramble for Africa need not detain us here,9 a brief
explanation is required as to how Africa’s military revolution facilitated

9
here is no more stimulating analysis of the era than Lonsdale (1985).
 he Fragile Revolution 409

Europe’s colonial moment. A century of violent upheaval between the 1780s

and the 1880s produced across the continent loating populations – the dis-
placed and the rootless, as well as armed predatory bands representing the
detritus of rapid social change and political reformation – which meant
manpower for European imperial projects. Without this manpower, most of
Europe’s late-century adventures into Africa would have been impossible –
the obvious exceptions being those areas where European troops were used
in large numbers, as in South Africa and Sudan. Further, not only did these
traumatized and militarized communities lend their muscle to a multitude
of European armed incursions, but such incursions were not infrequently
dependent on Africans’ martial skills, as well as their local knowledge. he
European contribution was clearly critical – in terms of modern irepower
(although this is in itself oten exaggerated) and, more important, organi-
zation, discipline, and drill. But European oicers were able to work with
some ine raw material, including a swell of young men well versed in the
martial arts, while the continual lux of sociopolitical conditions in late-
century Africa had produced mercenary cultures – or, perhaps more appro-
priately, mercenary markets. Individuals and entire groups were willing
and able to attach themselves to new leadership in exchange for security,
adventure, and material reward. Loyalty was entirely contingent and nego-
tiable. Many of the new states and societies that had arisen in the course of
the nineteenth century had been founded on the same principles, and thus
were the military foundations of the colonial state laid.
Much has been made, correctly, of the fact that there was no inter-
nal unity within Africa – indeed “Africa” itself did not exist, except as a
European geographical and cultural concept – and that there was therefore
no capacity for more concerted resistance. But this needs to be taken a little
further, for it is in itself neither adequate nor appropriate. he nineteenth
century witnessed a particularly explosive upsurge in interstate and inter-
community violence, which involved the opening up of especially deep is-
sures between states and societies, particularly between centers and their
armed frontier lands. he militarization of African political culture pro-
vided the adventuring agents of European imperialism with ready-made
points of leverage. One interpretation of this would be Europeans skill-
fully (but also rather luckily) positioning themselves between proverbial
petty warring tribes and ighting and overcoming each of these one by one.
here is no doubt that comparatively small European forces – comprising
a handful of European oicers in charge of an African rank and ile – were
able to insert themselves into the contested places between states and soci-
eties, exploiting local antagonisms to their own ends. Yet in fact Europeans
410 Reid

were frequently co-opted into local wars, unwittingly becoming actors in

local struggles, oten of long standing. Africa’s own military revolution in
many ways facilitated the European partition and ushered in the colonial
moment. Struggles for local material and political dominance that had
escalated throughout the nineteenth century now drew external agents into
them; many a beleaguered European company oicial, army oicer, and
missionary in the 1880s and 1890s was uncomfortably aware of the fact,
indeed, and it was only some time ater – certainly by the 1920s – that a
rather grander narrative was constructed around the purposefulness and
inevitable success of the civilizing military mission. he reality had been
altogether messier.
Nonetheless, the “Western way of war” was demonstrated to deadly efect
in the second half of the nineteenth century, when European states engaged
in a process of imperial expansion, which was in many respects the natural
outlow of the state-war vortex inside Europe itself. he wars Europe waged
on Africans in the late nineteenth century were oten more brutally eicient
than anything most of the latter had ever before encountered (Vandervort
1998). Although Douglas Porch is right to remind us that the margin of
victory in many of Europe’s so-called small wars of imperial expansion was
rather narrower than was oten supposed (Porch 2001), there is no ques-
tion that nineteenth-century Europe was armed to the teeth, with greatly
improved military resources and more muscular state apparatuses follow-
ing Europe’s own wars of the 1780–1820 period (Bayly 2004, 266f). he
comparative ease, with one or two exceptions, with which African resis-
tance was crushed in the 1890s and 1900s – especially in eastern and south-
ern Africa – further contributed to the twin notions that, irst, Africa’s
precolonial past had been bloodily tribal and inimical to change, still less to
development; and, second, that a rightful moral monopoly had been estab-
lished on the use of force. Europe mobilized ideas about race and culture to
depict African violence as perpetual and as symptomatic of a deep-rooted
savagery, and in so doing put in place a remarkably enduring set of mis-
conceptions about the role of warfare in Africa’s past. In their essentials,
they endure still, and because precolonial violence had been radicalized, so
now was the development that would supposedly follow it – for the colo-
nial state had stamped out atavistic savagery and would now facilitate, in a
range of ways, Africa’s halting irst steps into modernity. In fact the colo-
nial state represented only the latest stage in an ongoing process of violent
transformation, and had been co-opted – if at times almost impercepti-
bly – into African contours of change. But this is not to dismiss the colonial
moment as largely irrelevant in the grand low of African history – even if
 he Fragile Revolution 411

its signiicance has indeed been greatly exaggerated – for the net outcome,
as far as we can assess it from the standpoint of our own era, has been
long-term destabilization. Europe was indeed drawn into African conlicts,
but the colonial state in many cases crystallized particular patterns of vio-
lence – even if the temporary colonial pax concealed this for a generation
or so – and heightened the potential for continued violence among particu-
lar groups enclosed within and straddling colonial territories. he colonial
era greatly restricted Africans’ room for maneuver in terms of indigenous
resolution and prevention of conlict, and ultimately sustained material
development. his has been the lasting colonial legacy – alongside the per-
petuation of misapprehensions about the role of violence in Africa’s deeper
past, for much contemporary analysis, especially in the world of policy, is
not much more (or perhaps just as) sophisticated as that of the late nine-
teenth century. And so the nineteenth-century revolution has been pro-
longed, although it is manifest in new ways – in modern coups d’état, in
popular ideological insurrection, in the privatized violence of warlordism

12.5 WHAT WAS ACHIEVED, WHAT WAS NOT,

AND “WHAT IF. . .?”
he nineteenth century can be seen as something of a golden age of military
and political creativity, of social mobility, and of economic dynamism. Some
of the perennial limitations on African warfare described earlier began to
be overcome, and the capacity for change was greatly heightened. he rev-
olution proposed here did not happen everywhere (nor did it, indeed, in
Europe), but it was widespread. It unfolded in diferent places at diferent
times, and there was some variation in the drivers involved, but broad,
linking themes are clearly discernible between the late eighteenth and early
twentieth centuries. What we see, in sum, are ever more innovative ways
of organizing and utilizing violence in pursuit of political, economic, and
social objectives.
Important political, economic, and social changes attended military ref-
ormation. Greatly enhanced social mobility was the deining feature of many
of the new communities that came into being; intergenerational struggle,
moreover, involved the rejection of older systems of authority by young
men with access to guns and other commodities. hese were entrepreneurs
of violence, to be sure, but they were also representative of new economies,
closely bound up with long-distance commerce and underpinned by new
and aggressive forms of aspiration. War and material advancement were
indelibly intertwined – as in many ways they would remain – for individuals
412 Reid

as well as larger communities. War itself oten involved the creation of new
economic centers or the reinforcement of existing ones; these were places
of production and exchange that were part of the broader phenomenon of
urbanization across the continent resulting from prolonged violence. As
people moved together for protection, settlements – oten fortiied or with
a fortiied core – became larger, and new ones were carved out of the bush.
Towns themselves meant new socioeconomic and political forms, new iden-
tities, new stresses – and populations that were somewhat easier to con-
trol, vis-à-vis earlier dispersed patterns of settlement, and therefore urban
settlements were in many areas the building blocks of new and enhanced
political systems. More broadly, communities sought to defend or enlarge
their factor endowment bases, very oten at the expense of neighbors and
rivals, and thus involved the further reinement of total war, long a feature of
African warfare. As for those communities themselves, states were oten the
outcome of the violence of the nineteenth century – new forms of political
centralization that sought cohesion and unity. Older states likewise sought
to enhance internal cohesion and pursued external security through cul-
tures of militarism and violence respectively, innovating on the battleield
in the process. In many respects, the nineteenth century saw the culmina-
tion of state-building processes in motion since the fourteenth and iteenth
centuries. Where statehood in the conventionally deined sense was absent,
there were enlarged systems of identity and mobilization – most obviously
among pastoral groups – and many of the same features witnessed in state
creation, namely new forms of leadership, and the use of violence in pursuit
of larger political and economic goals.
And yet this was a fragile, halting revolution in which, for all the advances
made, there were serious failings. his can be explained in several ways.
Certain perennial limitations on the scope and scale of warfare – economic,
environmental, and demographic – lessened its potential to completely
transform African polity and economy. Notably, for example, underpopu-
lation was one of the deining features of African history (Ilife 2007). Low
levels of population governed styles of both war and military organization,
and within the continent itself there was diferentiation between land-rich
and relatively densely populated areas. his clearly had implications for
social stratiication and hierarchy, as well as capacity for mobilization of
available manpower, and, indeed, tactics. In addition, environmental and
climatic challenges were clearly critical. Fragile economies and resource
systems had a direct impact on the nature and meaning of organized vio-
lence, and had implications for the very nature of militarism and military
organization and for the needs of defense and aggression. In sum, states
 he Fragile Revolution 413

and societies continually struggled to maintain stability and cohesion at

the center, and were vulnerable to overextension – in terms of both limited
resources and limited people. hese perennial struggles produced politi-
cally fertile, mobile frontiers10 characterized by the episodic production and
reproduction of societies and cultures, and on the frontier, war represented
part of a creative vortex: in the eastern and central Great Lakes; across the
semiarid grasslands of West Africa; on the escarpments of the Ethiopian
Highlands; on the South African Highveld. his continual process of ission
and fusion – manifest in the armed frontier in opposition to the ambitious
center – facilitated enormous political creativity and military adventurism,
but otherwise prevented long-term growth and even consolidation, politi-
cally as well as economically, although of course signiicant advances were
made in this direction even prior to the nineteenth century.
Contrast this with the European experience: during the early modern
period, the era of mounted nobility was drawing to a close and infantry was
becoming more important. he increasing power of infantry over cavalry
was pushed along with technological innovation, the latter manifest in the
arquebus, the individually portable irearm introduced in the iteenth cen-
tury. his was part of a larger sequence of sociopolitical changes, namely
the increasing incorporation into the political and military system of urban
bourgeoisies and rural peasantries. he arming of infantries was potentially
dangerous, of course, representing as it did the militarization of segments
of the population not previously regarded as part of the political order.
herefore states needed to co-opt and control the process. his could be
done either through the gradual expansion of parliamentary political sys-
tems (as in England), or through the expansion of the centralized monar-
chical system (as in France and Spain). hus the military revolution is seen
to have had a symbiotic relationship with the rise of the European state
system. Only states, moreover, were able to bear the costs of large military
forces, and indeed of that other key aspect of the revolution on land, artil-
lery, which could be deployed against enemies at home as well as further
away. States and enhanced military power, in sum, went hand in hand, and
this was ever more in evidence in the course of the eighteenth century. he
state, in the simplest of terms, emerged from the need on the part of princes
to raise the sums from their subjects to pay for war (Howard 2002, 13–14).
While in Africa there is evidence for what can be described as “cam-
paigning war” – involving structured campaigns and set-piece battles –
“raiding war” was probably more common (Lamphear 2003), as indeed

10
I owe much in the way of inspiration here to Kopytof (1987).
414 Reid

it has remained, in various forms, throughout the modern era across the
continent. he distinction between the two may not, in fact, always have
been particularly pronounced – a single war, as it might be deined in Europe,
could stretch out over several years and involve a combination of the occa-
sional large-scale pitched battle but more commonly cycles of smaller-scale
attacks. he target was frequently the resource base of the enemy – people,
livestock, and materials. his was highly skilled and sophisticated war, if
apparently more limited in its destructive power compared to Europe in the
seventeenth and eighteenth centuries (and beyond); a similar limitation has
been noted for native North American warfare – described as a “skulking
way of war” – with native Americans in the seventeenth century appar-
ently expressing shock and horror at the level of violence European colo-
nists unleashed when they went to war (Malone 1991, 78). In the African
context, the struggle for control of resources – productive and reproduc-
tive – gave rise to, and doubtless was oten driven by, more abstract and
intellectual ideas about identity, territory, and history, and underpinned by
cultures of militarism and violence that framed the righteousness of war
in both spatial and temporal terms. his was cyclical war that was shaped
by relations between shiting frontier and center, with cultures developed
to justify actions and unify communities. But the scale and reach of such
war was limited by the mobility of the frontier itself – facilitated, again,
by relatively low population densities and the availability of land – which
made cultures richly complex but also prevented large-scale political and
economic transformation. In that sense, arguably, the most successful – that
is, enduring and cohesive – states and societies were those that remained
compact, territorially but also culturally and ideologically, for to stretch
beyond approximate boundaries was to risk unleashing and ultimately los-
ing control of frontiers that then became an armed threat. More practically,
terrain remained inhibitive across sub-Saharan Africa, where most African
armies were infantries with no access to pack animals. Movement during
particular seasons and over particular terrain was extremely diicult, and
impossible over large distances. he economic constraints on communities
rendered the transition from what we might broadly term “conservative”
war to revolutionary violence extremely diicult to achieve; it was an elu-
sive take-of point for many.
Africa’s revolution was also rendered fragile by the very nature of its
dependence on imported commodities. War over control of commerce was
frequently self-defeating – presenting us with a kind of “golden goose” sce-
nario – for it led to the destruction of trade itself, as was the case across the
eastern African interior, for example. One of the key elements in military
 he Fragile Revolution 415

change, moreover – the irearm, used increasingly efectively across Africa

in the nineteenth century – was imported, and thus polities ever more
dependent on it were fundamentally weakened by that dependency. More
speciically, exports – used to pay for guns and to meet the costs of enlarged
warfare and political organization – were intrinsically ill suited to develop-
ing internal economies in such a way as to fund further military and polit-
ical transformation. In fact, some societies did remarkably well to use the
proceeds as skillfully as they did, and again advances were made, but export
commodities – slaves, palm oil, groundnuts, ivory – were inimical to long-
term development, as their modern equivalents have generally remained.
he colonial moment exacerbated this trend: the restrictive nature of the
export trade, which in the end beneited particular groups in a narrow
range of ways, locked much of Africa into a spiral of violent competition,
leading to escalating militarization with little or none of the maneuverabil-
ity – in sum, the capacity to ight to the point of local, indigenous resolu-
tion – characteristic of the nineteenth century.
It is also the case that the era of Africa’s revolutionary violence coincided
with that of European imperial expansion. Europe was in a position to take
advantage of turbulence in Africa, inserting itself into heightened com-
petition between societies and states at a critical moment, and harnessing
such violence for its own ends – even if a measure of good fortune was
also frequently involved. Europe willfully or otherwise misread what was
happening in Africa in the nineteenth century, justifying its intervention
through the creation of a series of myths about a universal political, eco-
nomic, and, indeed, moral order. Africa was detained developmentally for
several decades, its revolution becoming diverted, its violence becoming
all the more destabilizing and uncompromising, and therefore its ability to
negotiate and ultimately compete in global economic and political terms
seriously compromised. he resurgence of violence in the mid-twentieth
century is testament to this fact. War in Africa, both past and present, was
racialized; so too was discourse around development, if covertly, which is
profoundly ironic, given that much modern violence in Africa has been
about economic access and opportunity, as indeed it was in the nineteenth
century.
Nevertheless, innovation and development – again, assuming certain
universal criteria, Eurocentric in origin – were much in evidence; and it is
proposed here, counterfactually, that such processes would have produced,
in many respects, a somewhat diferent Africa in the twentieth century. We
must be careful at this juncture, for just as no pristine precolonial Africa is
depicted in this chapter, nor is Africa’s twentieth century imagined as some
416 Reid

glorious and inexorable march toward peace, prosperity, and stability in

the absence of colonial rule. It is quite likely that even without European
colonialism much of Africa would have experienced a high level of violent
instability until well into the twentieth century, probably – if we allow our-
selves a few moments of conjecture – into at least the 1920s and 1930s. And
it would in all likelihood have involved avaricious Western companies com-
peting aggressively for inluence and access, while Western governments
made moral-sounding noises about the need for peace, stability, and good
governance. In other words, it would have looked much like today, except
perhaps – for a while – a little worse. However, the direction of travel would
have been toward local resolution, accommodation, and stabilization, with
a much greater potential for economic and political development and a nat-
ural – if gradual – tendency toward demilitarization. Much greater levels of
violence in the early to mid-twentieth century would have been followed –
more conidently in some areas than in others – by consolidated political
and economic systems, and above all a much greater capacity for maneuver
and for equitable negotiation on the part of African actors.
Of course, struggles over external linkages – and over how to manage
external relationships – would have continued and would probably have
produced new forms of conlict: this had been the case since the early six-
teenth century, and is demonstrable elsewhere in the world, notably in
China and Japan. Low-level conlict would have persisted in many parts of
the continent as state systems struggled to impose themselves on trouble-
some frontiers, although levels of Western concern – moral or otherwise –
over this would clearly have depended on how distracting were Europe’s
own internal problems.11 Slavery, too, would presumably have persisted
across Africa throughout the twentieth century (rather more widespread
and overt, perhaps we should add, than is currently the case), and would,
as in the nineteenth century, have underpinned much of Africa’s economic
activity. But, notwithstanding a continued moral outcry, Western economic
systems in general and consumers in particular would have proven rather
more tolerant of slavery elsewhere in the world than one might imagine –
witness the hegemony of the sweatshop underpinning much of the global
economy today. he key question for our purposes, however, is how preva-
lent, therefore, would slaving violence have been, and how inhibitive would

11
In our imaginary world, for example, would there have been a Second (or even a First)
World War without the Scramble for Africa? How important was the partition and subse-
quent administration of Africa to Europe’s own political, economic, and, indeed, cultural
development in the twentieth century? hese are questions to make the head throb – and
the undergraduate heart sink.
 he Fragile Revolution 417

such low-level war have been regarding African economic growth: one is
tempted to suggest it would have been manageable, and that violence would
have easily coexisted with very successful entrepreneurialism – whether in
state or nonstate settings.
In other words, according to this interpretation, the irst half of the twen-
tieth century would have looked a lot more violent, and the second half a
lot less. Even in the late twentieth century, Africa might well have looked
pretty violent, but there would have been more stable markets, and across
the continent more consolidated political and economic authority at the
local level, as well as more stable indigenous authorities. he latter would
probably have been more accountable, too, as decades of violent upheaval
would have been, in the loosest sense, democratizing: soldiers and their
wars required popular support and popular participation.12 Whether the
particular forms of government that emerged would have won wholesale
approval from Western powers is another question; it seems safe to assume
that many would not, and that some – monarchical autocracies, say, or (at
the other end of the spectrum) pastoral heterarchies – would have been
deemed incompatible with concepts of political modernity. Important,
however, external powers would have had, in the long run, somewhat less
leverage over them. here is every reason to suppose that a map of Africa
in, say, 1920 would have displayed a number of prominent, relatively stable
polities – Sokoto, Dahomey, Asante in West Africa; Buganda, Bunyoro and
neo-Solomonic Ethiopia in the east and northeast of the continent; a Zulu
state of some description in the south. Extrapolating forward from nine-
teenth-century trends, these are polities with the capacity to develop, or
indeed which had already developed, professional administrative bureau-
cracies that would be importers of modern military and economic tech-
nologies with a view to political and economic growth and that would be
seeking – through various external linkages, no doubt – to manage their
economic systems to best advantage. Many African states and societies had
been doing precisely this for the previous century or more.
Again, however, there would have been a fair amount of political and
military instability; in the frontiers between the stable polities, political and
economic conditions would have been rather more in lux. here was clearly
enormous sociopolitical and economic diversity across late nineteenth-
century Africa, and clearly some communities were more likely to succeed
than others: as a rule, compact territorial states as opposed to sprawling

12
here is some evidence to suggest this for nineteenth-century eastern Africa (Reid
2007, 150).
418 Reid

and ineicient ones; efectively centralized hierarchies rather than more

dispersed, stateless groups. But although pastoral-sedentary conlict had
long been a key dynamic, the historic relationship was characterized by a
marked degree of accommodation and reciprocity: more lexible, organic
relationships between socioeconomic forms would have reduced margin-
alization, and indeed as the Somali – and large African diasporas – have
demonstrated, states are not necessary to thriving commerce. Even so,
entire regions would have lacked overarching authority, even in the 1920s
and 1930s – much of west-central Africa, for instance, and on either side
of Lake Tanganyika. Would the Nyamwezi or the Yoruba have had their
“Westphalia moment,” for example? On the evidence of the nineteenth cen-
tury, the answer would seem to be yes – eventually. his was already the
direction of travel in many areas, although it needs to be emphasized that
parts of the continent would have remained fairly violent for quite some
time – and many of the states with an air of permanence in the late nine-
teenth century might not have made it to the mid-twentieth. Buganda was
under great pressure from Bunyoro, as well as undergoing major internal
strains; signiicant internal tensions had also undermined the Zulu state;
Ethiopia’s spectacular expansion might easily have been reversed without
the props of neighboring European colonial administrations. Africa’s alter-
native twentieth century would have been messy – Europe’s was hardly
otherwise – and the map would have been a dazzling and doubtless (to
the European eye) discombobulating arrangement of zones of transition,
and greater or lesser spheres of jurisdiction. But the capacity for indigenous
agency to make mistakes, and then to resolve and build, would have been
all the greater.

12.6 CONCLUSION
We can be sure that the factors prompting a particular path of development
as opposed to another will remain the focus of considerable debate. he
divergent paths of Europe and Africa are the product of countless diferent
determinants – some dramatic, others minuscule, but no less important
for all that. In Europe, there was an intricate and symbiotic relationship
between technology, economic might, military power, and state formation.
As with the “revolution in military afairs,” debates continue on the precise
relationship between war and the emergence of the state in European his-
tory;13 but certainly key themes have endured. Especially germane is the

13
For example, see James (2006).
 he Fragile Revolution 419

thesis William H. McNeill placed on the relationship between economic

and military expansion, and the stabilization (and professionalization) of
the states that were able to facilitate both. As McNeill put it, by the sev-
enteenth century Europe had “launched itself on a self-reinforcing cycle
in which its military organisation sustained, and was sustained by, eco-
nomic and political expansion at the expense of other peoples and poli-
ties of the earth” (McNeill 1982, 143). his was the transformative power
of violence writ large. Paul Kennedy reiterated the idea that “great power”
status depended as much on “the more or less eicient utilisation of the
state’s productive economic resources in wartime” as on events on the bat-
tleield, and indeed on economic performance relative to other powers in
peacetime (Kennedy 1988, xv). Economic might (or at least relative eco-
nomic might) underpinned military success, while military success had the
potential (no more than that) to feed economic performance and enhance
economic capacity. Ultimately, much depended on the powers of coercion
of the state, and the eiciency with which the state could marshal resources.
his was the concern of Charles Tilly, who aimed to explain the various
routes by which the nation-state had become the dominant form of human
organization (Tilly 1992). In essence, he argued that early modern military
innovations (mass infantry, gunpowder, fortiication) were expensive, and
could only be supported by states with large populations and that were able
to accumulate suicient amounts of capital and inancial resources – not
least in terms of tax revenue. he modern state thus came about because of
the needs of waging war.
Ultimately, what diferentiates Africa from Europe in the context of the
nineteenth-century revolution in military afairs proposed here is the nature
of Africa’s internal struggle. his struggle is exempliied by the African
frontier – itself the product of distinctive environmental challenges, limited
and uneven demographic spread, and the ierce introspection demanded by
close economic management. he possibilities for expansion, whether ter-
ritorial or in terms of internal resource bases, were limited, as centrifugal-
ism and strains at the center restricted the reach of even the most successful
states. his meant that African societies were less able to capitalize – liter-
ally and iguratively; to consolidate and accumulate political and economic
resources with a view to enlargement of scale. Timing, moreover, is indeed
critical, and most African societies were certainly unable to achieve a level
of political and economic stability in time to see of (or render unneces-
sary) European intervention; ironically, indeed, the very violence driving
much of the continent’s development in the nineteenth century attracted
detrimental outside interference. Major change was afoot in the nineteenth
420 Reid

century, but it ultimately intersected with a rampant European militarism,

which in many ways prolonged and distorted it.
he issue of cost is ultimately crucial, moreover. African warfare moved
relatively rapidly into the realm of professional entrepreneurialism, even –
arguably – compared to Europe; in so doing it avoided the kind of long,
drawn-out, and enormously expensive wars Europe experienced during the
sixteenth and seventeenth centuries. his meant, in addition to the longer-
term prohibitive dynamics noted earlier, that Africa did not need to develop
comparably elaborate iscal, administrative, and governmental systems to pay
for its wars – at least, again, not on the scale of Europe’s experience. African
warfare contained oten enormous internal costs, as it were – but these were
(outside the horse zone, at any rate) relatively low and required few of the
political as well as technical innovations of Europe’s wars. he African fron-
tier involved highly destructive economic war; but in general, even long wars
allowed for periods of recovery (a key element of the North American “skulk-
ing war” model), and thus these were, as a rule, wars of short, intense bursts
involving comparatively low levels of capital investment, and therefore reduc-
ing the need for the kind of political, economic, and technological innova-
tions of Europe’s later seventeenth and eighteenth centuries.
By this logic, of course, it might be argued that had Africa’s violent trans-
formation run its natural course (in other words, with a minimal level of
outside interference), it would have proceeded in a manner more akin to that
of Europe: war would have involved capital accumulation; political reform
would have followed to relect economic shits; and a clearer path of endog-
enous development would have ensued. Again, this is what the events of the
nineteenth century suggest: Africa’s military revolution – the most visible
manifestation of internal transformation – may be considered the prerequi-
site for the kind of wholesale reform and consolidation witnessed in north-
west Europe, the irst wave of a multifaceted set of changes. his irst wave
of reform was unquestionably violent, but it was also marked by a trajectory
suggestive of an alternative future involving rather more efective economic
management, political restructuring, and ultimately demilitarization. It may
yet prove to be just that, although further violence will certainly attend the
process; there is every reason to hope, given the lessons contained in Africa’s
longue durée, that the colonial experience has merely postponed, not can-
celled, organic solutions to seemingly intractable problems.
he “perennial limitations” noted earlier would have continued to
operate well into our imaginary twentieth century, and would likely still
be operating at the time of writing; and it is foolish to think in terms of
African societies transforming themselves into European clones. Again,
 he Fragile Revolution 421

no “merrie Africa” is proposed here, but rather one characterized by a

great deal of violence with developmental potential. he central point
here is that Africans would have done no worse, and probably altogether
better, in terms of management and distribution of material and political
resources, than recent postcolonial elites. he processes of the nineteenth
century suggest that this was possible, however violent the process was – as
it was, too, in Europe – and might still have been, along the way. As it was,
the notion of Africa’s military defeat in the late nineteenth century, and
the mythology subsequently developed around ideas about development –
principally that violence was inimical to it, when Europe’s own historical
experience demonstrated the exact opposite – generated profound distaste
for what African warfare had supposedly represented, and in turn deep
misunderstanding of it.

References
Adjaye, Joseph K. (1996). Diplomacy and Diplomats in 19th Century Asante. 1st edition
1984. Trenton, NJ: Africa World Press.
Ali, Taisier Mohamed Ahmed and Robert O. Matthews, eds. (1999). Civil Wars in Africa:
Roots and Resolution. Montreal: Montreal & Kingston.
Alpers, Edward A. (1975). Ivory and Slaves in East Central Africa. London: Heinemann.
Austen, Ralph A. (1987). African Economic History: Internal Development and External
Dependency. London: James Currey.
Bayly, C. A. (2004). he Birth of the Modern World 1780–1914: Global Connections and
Comparisons. Oxford: Blackwell Publishers.
Bennett, Norman Robert. (1971). Mirambo of Tanzania, 1840?–1884. New York: Oxford
University Press.
Black, Jeremy. (2009). War: A Short History. London: Continuum.
Boas, Morten and Kevin C. Dunn, eds. (2007). African Guerrillas: Raging against the
Machine. Boulder, CO: Lynne Rienner Publishers.
Clapham, Christopher S. (1998). African Guerrillas. Oxford: James Currey.
Cobbing, Julian. (1988). “he Mfecane as Alibi: houghts on Dithakong and Mbolompo.”
Journal of African History, 29(3): 487–519.
Coquery-Vidrovitch, Catherine. (2009). Africa and the Africans in the Nineteenth
Century: A Turbulent History. London.
Derman, William, R. Odgaard, and E. Sjaastad, eds. (2007). Conlicts over Land and
Water in Africa. Oxford: James Currey.
Doyle, Shane. (2006). Crisis and Decline in Bunyoro: Population and Environment in
Western Uganda, 1860–1955. Oxford: James Currey.
Goody, Jack. (1971). Technology, Tradition and the State in Africa. London: Oxford
University Press.
Hamilton, Carolyn Anne. (1992). “‘he Character and Objects of Chaka’: A
Reconsideration of the Making of Shaka as ‘Mfecane’ Motor.” Journal of African
History, 33(1): 37–63.
422 Reid

(1998). Terriic Majesty: he Powers of Shaka Zulu and the Limits of Historical
Invention. Cape Town: David Philip.
Henniger, Laurent. (2006). “Military Revolutions and Military History.” In Modern
Military History, edited by Matthew Hughes and William James Philpott.
Basingstoke: Palgrave Macmillan.
Hopkins, Anthony G. (2009). “he New Economic History of Africa.” Journal of African
History, 50: 155–77.
Howard, Michael. (2002). he Invention of Peace and the Reinvention of War. London:
Proile.
(2009). War in European History. Oxford: Oxford University Press (new update of 1st
edition, 1976).
Ilife, John. (1987). he African Poor: A History. Cambridge: Cambridge University
Press.
(2007). Africans: he History of a Continent, second edition; irst edition 1995.
Cambridge: Cambridge University Press.
James, A. (2006). “Warfare and the Rise of the State.” In Modern Military History,
edited by Matthew Hughes and William James Philpott. Basingstoke: Palgrave
Macmillan.
Kaarsholm, Preben, ed. (2006). Violence, Political Culture and Development in Africa.
Oxford: James Currey.
Kennedy, Paul Michael. (1988). he Rise and Fall of the Great Powers: Economic Change
and Military Conlict from 1500 to 2000. London: Unwin Hyman.
Kopytof, Igor, ed. (1987). he African Frontier: he Reproduction of Traditional African
Societies. Bloomington & Indianapolis: Indiana University Press.
Lamphear, John. (1998). “Brothers in Arms: Military Aspects of East African Age-Class
Systems in Historical Perspective.” In Conlict, Age and Power in North East Africa,
edited by Eisei Kurimoto and Simon Simonse. Oxford: James Currey.
(2003). “Sub-Saharan African Warfare.” In War in the Modern World, edited by Jeremy
Black. London: Routledge.
Lamphear, John, ed. (2007). African Military History. Aldershot: Ashgate Publishing
Company.
Law, Robin. (1977). he Oyo Empire, c.1600–c.1836: A West African Imperialism in the
Era of the Atlantic Slave Trade. Oxford: Clarendon Press.
(1991). he Slave Coast of West Africa, 1550–1750: he Impact of the Atlantic Slave
Trade on an African Society. Oxford: Clarendon Press.
Legassick, Martin. (1966). “Firearms, Horses and Samorian Army Organisation 1870–
1898.” Journal of African History, 7(1): 95–115.
Lonsdale, J. (1985). “he European Scramble and Conquest in African History.” In
Cambridge History of Africa, Vol 6: From 1870 to 1905, edited by R. Oliver and G.
N. Sanderson. Cambridge: Cambridge University Press.
Low, D. A. (2009). Fabrication of Empire: he British and the Uganda Kingdoms, 1890–
1902. Cambridge: Cambridge University Press.
Malone, Patrick M. (1991). he Skulking Way of War: Technology and Tactics among the
New England Indians. Lanham, MD: Madison Books.
Mazrui, Ali Al’Amin, ed. (1977). he Warrior Tradition in Modern Africa. Leiden: Brill.
McNeill, William Hardy. (1982). he Pursuit of Power: Technology, Armed Force and
Society since AD 1000. Chicago: University of Chicago Press.
 he Fragile Revolution 423

Nhema, Alfred G. and P. Tiyembe Zeleza, eds. (2008a). he Roots of African Conlicts:
he Causes and Costs. Oxford: James Currey.
Nhema, Alfred G. and P. Tiyembe Zeleza, eds. (2008b). he Resolution of African
Conlicts: he Management of Conlict Resolution and Post-conlict Reconstruction
Oxford: James Currey.
Ogot, Bethwell A., ed. (1972). War and Society in Africa. London: F. Cass.
Parker, Geofrey. (1988). he Military Revolution: Military Innovation and the Rise of the
West, 1500–1800. 2nd edition 1996. Cambridge: Cambridge University Press.
Porch, Douglas. (2001). Wars of Empire. London: Cassell Military.
Reid, Richard J. (2002). Political Power in Pre-colonial Buganda: Economy, Society and
Warfare in the Nineteenth Century. Oxford: James Currey.
(2007). War in Pre-colonial Eastern Africa: he Patterns and Meanings of State-level
Conlict in the Nineteenth Century. Oxford: James Currey.
(2011). Frontiers of Violence in Northeast Africa: Genealogies of Conlict since 1800.
Oxford: Oxford University Press.
Reno, William. (1998). Warlord Politics and African States. Boulder, CO: Lynne Rienner
Publishers.
Richards, Paul, ed. (2005). No Peace, No War: An Anthropology of Contemporary Armed
Conlicts. Oxford: James Currey.
Roberts, Andrew D., ed. (1968). Tanzania before c.1900. Nairobi.
Roberts, Richard L. (1987). Warriors, Merchants and Slaves: he State and the Economy
in the Middle Niger Valley, 1700–1914. Stanford. CA: Stanford University Press.
Smaldone, Joseph P. (1977). Warfare in the Sokoto Caliphate: Historical and Sociological
Perspectives. Cambridge: Cambridge University Press.
Smith, Robert Sydney. (1976). Warfare and Diplomacy in Pre-colonial West Africa. 2nd
edition 1989. London: James Currey.
Smith, Robert Sydney and Jacob F. Ade Ajayi. (1971). Yoruba Warfare in the Nineteenth
Century. Cambridge: Cambridge University Press.
Spear, homas T. and Richard Waller, eds. (1993). Being Maasai: Ethnicity and Identity
in East Africa. London: James Currey.
Stanley, Henry M. (1899). hrough the Dark Continent. 1st edition 1878. London: George
Newnes.
hornton, John. (1999). Warfare in Atlantic Africa, 1500–1800. London: University
College of London.
Tilly, Charles. (1992). Coercion, Capital and European States, AD 990–1992. Oxford.
Uzoigwe, G. N. (1977). “he Warrior and the State in Pre-colonial Africa.” In he Warrior
Tradition in Modern Africa, edited by Ali Al’Amin Mazrui. Leiden: Brill.
Vandervort, Bruce. (1998). Wars of Imperial Conquest in Africa, 1830–1914. Bloomington
& Indianapolis: Indiana University Press.
Wilks, Ivor. (1975). Asante in the Nineteenth Century: he Structure and Evolution of a
Political Order. London: Cambridge University Press.
Wylie, Dan. (2006). Myth of Iron: Shaka in History. Scottsville: University of KwaZulu-
Natal Press.
 13

he Imperial Peace

Robert H. Bates

When praised at all, imperialism is most oten commended for the peace it
bestowed. By demobilizing armies, deposing marauding princes, and sub-
duing war-like states, European powers fashioned a half century of political
order. he question nonetheless arises: Should they be lauded for doing so? In
this chapter, I view Africa’s history through the lens of comparative history1
and argue that the imperial peace may have retarded Africa’s development.
hose addressing the rise of the West oten highlight the contribution
of the state. he Western state has inlicted untold misery, they acknowl-
edge; but they also stress its importance for development. An important
school of thought argues that the two faces of power – the one destruc-
tive and the other developmental – share common origin: the Western
state became developmental the better to inance its wars (Brewer 1988;
Gilbert 1975; Roberts 1956; Tilly 1985). Driven by insecurity abroad,
states become developmental at home, these scholars argue: heads of state
employ power to foster economic growth and thereby inance their mili-
taries. If the threat of war thus rendered European states developmental,
then the possibility arises that imperial peace may have contributed to the
underdevelopment of Africa (but see Centeno 1997). Viewed against this
background, the imperial peace assumes a new guise: rather than confer-
ring a blessing, it may have imposed a cost; it may have stilled the impulse
to development that originates from the search for security in a hostile
political environment.
hroughout this chapter, I focus on the relationship between wealth and
power. Power can be used to secure wealth, either by forcefully seizing it or
by safeguarding its creation (see, for example, Hirshleifer 1994). he latter
promotes growth; the former, mere redistribution. In probing the political

1
As was attempted by Warner (1999). For a skeptical response, see Hopkins (2000).

424
 he Imperial Peace 425

foundations of development, it thus becomes critical to understand how power

is exercised and to explore the factors that shape its use. It is my argument that
through purposeful behavior on the one hand and physical extinction on the
other, international insecurity selects for the developmental use of power. It is
for this reason that the imperial peace may have therefore been costly.
I begin this chapter by contending that the state is essential for devel-
opment, an argument I advance by recounting a recent reappraisal of the
logic of Evans-Pritchard’s classic study of the Nuer: a society without a state
(Evans-Pritchard 1940).

13.1 KINSHIP SOCIETIES

Africa is attractive to scholars in part because of the diversity of its political
institutions. Some polities are based on kinship; they are societies without
states. In others, centralization and hierarchy shape political life. In such
polities, kinship ties may remain powerful, but they now lie imbedded
within political structures that they inluence but do not control: politi-
cal hierarchies and bureaucratic oices that may be illed by appointment
rather than succession or descent.
From the study of Africa’s political forms, we learn many lessons. One of
the most important is that political order can prevail even in societies that
lack states. Even in the absence of centralized and hierarchical institutions,
life and property can be secure. A second is that such order is fragile and,
worse yet, susceptible to breakdown in response to development.

13.2 PEACE IN THE FEUD

As Gluckman (1955) argued, order in kinship societies rests on the fear
of revenge. Should a person in one family transgress against a member of
another, then she faces the prospect of retaliation by her victim and his kin.
In such societies, no third party preserves the peace, then. Rather, it is the
“shadow of the future,” in Axelrod’s (1985) phrasing; it is the certainty of
future losses at the hands of a vengeful neighbor that stays the hand of those
who might be tempted to transgress. Even in the absence of formal institu-
tions, property can therefore be secure and people safe in stateless societies;
thus, the argument of Evans-Pritchard (1940), Gluckman (1955), and oth-
ers (e.g., Fallers 1966) who have studied such societies. More recently, how-
ever, political scientists and economists have revisited this argument and
cautioned against its claims. In so doing, they highlight the link between
the state and development.
426 Bates

As noted in Bates, Greif, and Singh (2002), Gluckman’s claim – that

there can be “peace in the feud” – follows from the “folk theorem” of
repeated games, which states (roughly) that the irst best – here, peace
and prosperity – can be sustained as an equilibrium in games wherein
people can adopt punishment strategies, ensuring that future losses ofset
the prospect of short-term gains. Once the parallel to game theory is rec-
ognized, then so too are its implications. Among the most relevant is that
the irst best equilibrium is not robust; in particular, it is vulnerable to the
temptations increased prosperity spawns. As society becomes wealthier
and riches mount, the level of temptation increases as well. As a result,
punishments that once served to underpin peace may no longer deter.

13.3 NO STATE, NO DEVELOPMENT?

Note the implication: if development requires both security and wealth,
then, by requiring that prosperity be sacriiced to secure peace, kinship
societies may remain underdeveloped. Another implication is that while
their polities might be able to protect and sustain some kinds of economic
activities, such as commerce, they might not be able to protect others, such
as those that require ixed investments. Commercial relations are likely to
involve repeated interactions; ixed investments, by contrast, require but a
single move. Unlike merchants, investors therefore lack the ability to make
credible threats. A kinship economy might therefore beneit from markets
and trade; but, lacking investment, it may fail to secure an industrial base.
Because of the nature of their political institutions, such polities may there-
fore achieve but a low level of development.
It is therefore not surprising to encounter evidence that suggests that
stateless societies seek to alter their political institutions, some going so far
as to import political specialists from neighboring societies. Well-known
cases would include the Aro in southern Nigeria (Northrop 1978), the
Atyak in northern Uganda (Southall 1999), and the Lunda in Congo and
Zambia (Vansina 2004), each of whom exported their political services to
neighboring stateless societies, thereby introducing political hierarchy into
previously egalitarian cultures.

13.4 PRECOLONIAL DYNAMICS

In this section, we focus on the formation – and the destruction – of states
in precolonial Africa. Figure 13.1 illustrates the population of states in West
Africa, from 1400 (just before European states began to form empires) to
 he Imperial Peace 427

Number of States by Period

West Africa
40
35
30
25
20 Total
15
10
5
0
00

00

00

00

00

00
14

15

16

17

18

19
Figure 13.1. Number of states by period, West Africa.

1900 (the end of the Scramble for Africa).2 While the scholars from whose
work it was compiled concur in their deinition of West Africa,3 they
count as states a wide variety of political forms, ranging from city-states
to empires. Despite this heterogeneity, where their accounts intersect, they
tend to include the same polities, which increases our conidence in infer-
ences drawn from these data.
Prior to the iteenth century, state formation in West Africa, these
sources remind us, was largely conined to the margins of the Sahel, where
polities formed astride the trade routes linking West Africa’s commercial
centers with North Africa. With the arrival of Europeans along the coast,
commerce shited southward and states began to form in the interstices
between the interior and the coast. he narrative accounts then portray the
rise of Oyo, Bono, Nupe, and the Hausa city-states. Later comes the rise
of states along the coast: Wadyah, Warri, Bonny, Allada, Grand Popo, and
others.
More important for the argument of this chapter, however, is the demise
rather than the creation of states. As shown in Figure 13.2, by the mid-eigh-
teenth century, more states depart than enter the panel, thus making pos-
sible the process of selection. he question is: Did “developmental” states

2
Sources include: Forde and Kaberry (1967), Stevenson (1968), Colson (1969), Gray and
Birmingham (1970), Southall (1970), Fage (1974), Ajayi and Crowder (1976), Oliver (1977),
Law (1978), Fortes and Evans-Pritchard (1987), Southall (1988), Morrison, Mitchell, and
Paden (1989), Curtin et al. (1995), Kwamena-Poh et al. (1995), S. K. McIntosh and R.
McIntosh (1996), McIntosh (1999), hornton (1999), and Banton ([1965] 2004).
3
hat it extends from the Chad basin to the Atlantic coast and from the coast northward to
the Sahel.
428 Bates

Creation and Destruction of States by Period

West Africa: Numbers
10
8
Number

6 Created
4 Destroyed
2
0
00

50

00

50

00

50

00

50

00

50

00
14

14

15

15

16

16

17

17

18

18

19
Figure 13.2. Creation and destruction of states: number by period.

prove more it than those that were not? Did the demise of polities select
for states in which power was used to promote the creation of wealth? If
the response to these questions is yes, then the imperial peace may have
decoupled the search for security from the creation of wealth.

13.5 COLONIAL CROSS-SECTIONS

To address this possibility, we turn our attention to the colonial period.
When doing so, we make use of three cross-sections of data: Ethnographic
Atlas (Murdock 1967), the Standard Cross-Cultural Sample (SCCS) (Divale
2000), and the Human Relations Area Files (HRAF) (Bates 1983). Each
data set exhibits major weaknesses. Both the SCCS and HRAF samples
contain few African cases: twenty-eight in the irst and thirty-six in the
latter. Ethnographic Atlas contains data from many more polities (more
than 400); but its utility is reduced by the limited range of the characteris-
tics it reports, most of which are irrelevant to our purposes. Each provides
insight into the cross-section of cases bequeathed to the colonial order. If
our argument is correct, we should expect this cross-section to suggest a
positive relationship between measures of state formation and indicators
of prosperity.
Addressing the primary question posed in this chapter requires that I
classify the societies according to their political institutions. Scholars such
as Southall (1999), Yofee (1993), and McIntosh (1999) have insisted that
African polities cannot be crudely classiied as stateless or not; nor can they
be arrayed along a scale of “state-likeness.” he irst underestimates their
heterogeneity, while the second opens the door to cultural triumphalism,
they argue, as notions of modernity or civilization might easily inform
 he Imperial Peace 429

Table 13.1. Data from the Human Relations Area Files

Society Pages of documentation Society Pages of documentation

Bambara 1127 Chagga 1986
Dogon 1132 Ngondo 1474
Mossi 942 Pygmies 1350
Mande 605 Azande 3264
Tallensi 964 Mongo 773
Ashanti 3523 Rundi 1314
Katab 352 Mbundu 847
Nupe 856 Bemba 830
Tiv 2891 Ila 998
Yoruba 1637 Lozi 1635
Fang 1117 Tonga 1616
Nuer 1541 Ngoni 1123
Shilluck 1073 honga 1231
Ganda 2261 Yao 555
Dorobo 354 Bushmen 1259
Kikuyu 1950 Hottentot 1339
Luo 463 Lovedu 455
Masai 2085 Tanala 354
Note: he names are those chosen by the Human Relations Area Files.
Source: Bates 1983, 57.

placement on the scale. In conformity with their counsel, I therefore make

but qualitative judgments, determining whether the reports indicated the
existence of chiefs and whether there was a monarch. I also note the exis-
tence of a bureaucracy or a national army (Table 13.1).
Recall the discussion of political order: kinship societies, I argued,
were vulnerable to internal disruption. Suggestive, therefore, is the rela-
tionship between markers of political centralization and incidence of
feuding: it is negative and the relationship is one of the most pronounced
in the data. Whereas the HRAF records attest to feuding in 89 percent
of the societies governed by kinship, they report it in but 23 percent of
those governed by monarchs (Table 13.2). he data thus suggest that pol-
ities based on rulers and bureaucracies are associated with higher levels
of political order.4

4
Supportive are data from the SCCS, which indicate that chietaincies and centralized states
are more likely to have “enforcement specialists,” such as police, than are stateless societies;
the source fails to note the impact of this innovation on the level of conlict, however.
430 Bates

Table 13.2. Feuding and political structure (percent “yes”)

Kinship Chiefs Monarch Central bureaucracy

Present Absent
Report of Feuds? 89 78 23 27 72
Note: N = 28.
Source: Data from Human Relations Area Files (Bates 1983, 45).

Table 13.3. Economic activity and political structure (percent “yes”)

Kinship Chiefs Monarch Central bureaucracy

Presence of: Present Absent
Market Centers? 15 31 54 58 42
Long-distance Trade? 0 45 55 56 44
Crat Specialists? 21 25 54 57 60
Private Investors? 0 33 67 50 48
Note: N = 31–3.
Source: Data from Human Relations Area Files (Bates 1983, 44, 46).

While enhancing the level of political order, do such institutions assure

greater prosperity? Do they thereby provide a corrective to the fragility of
kinship systems? Keeping in mind the poor quality of the data, as seen in
Table 13.3, the answer appears to be yes. Studies of societies with chiefs and
monarchs were far more likely to report the presence of market centers than
were studies of societies ruled through kinship. hey were also more likely
to report participation in long-distance trade, the production of crats, and
the presence of private investors.
Further evidence is contained in Table 13.4. Insofar as possessing a
national army, as opposed to one recruited at the village or household level,
provides an index of military power, the data suggest that wealth and mili-
tary power covary.
Turning briely to Ethnographic Atlas, we note its data on crat special-
ists. According to this source, this kind of economic activity was observed
in but a third of Africa’s stateless societies but in nearly three-quarters of its
centralized states. Similar diferences characterize the incidence of leather
working, ceramics, and metalworking. hese data suggest that occupational
 he Imperial Peace 431

Table 13.4. Military forces and economic activity

Local National
Presence of:
Market Centers?
% Yes 15 31
Long-distance Trade?
% Yes 13 40
Crat Specialists?
% Yes 12 42
Private Investors?
% Yes 15 46
Note: N = 31–3.
Source: Data from Human Relations Area Files (Bates 1983,
44, 46).

Table 13.5. Public goods and political structure (percent “yes”)

Kinship Chiefs Monarch Central Bureaucracy

Presence of: Present Absent
Roads? 0 20 75 88 14
Bridges? 0 0 22 40 0
Pontoons? 0 0 50 50 10
Canals? 0 11 33 40 5
Note: N = 26–30.
Source: Data from Human Relations Area Files (Bates 1983, 50).

specialization more frequently occurs in states than in stateless societ-

ies. he reports archived in the Human Relations Area Files also indicate
(Table 13.5) that when ruled by chiefs, monarchs, and bureaucracies, poli-
ties were more likely to contain roads, bridges, pontoons, and canals. When
suiciently “state-like” to maintain a national army, African polities, once
again, were more likely to invest in such public goods (Table 13.6).
In this instance, neither Ethnographic Atlas nor SCCS provide useful
information; they simply fail to record the relevant data.
While scanty and highly imperfect, the data suggest that the competition
among the states in precolonial Africa may have given rise to a cross-section
of polities in which state formation and economic development covaried.
Either because they responded to external threats by seeking to strengthen
432 Bates

Table 13.6. Military forces and

public goods

Local National
Presence of:
Roads?
% Yes 27 50
Bridges?
% Yes 0 20
Pontoons?
% Yes 7 33
Canals?
% Yes 13 20
Note: N = 29–30.
Source: Data from Human Relations Area.
Files (Bates 1983, 44).

Table 13.7. Political structures and warfare (percent)

External warfare Stateless Chietaincies Centralized states

Yearly 33 60 60
Common/ 33 40 40
Occasional
Rare 33 0 0
N 5 5 5
Source: Standard Cross-Cultural Sample.

their economic base, or because interstate competition winnowed from the

sample polities that were weak and poor, we ind that “state-like” societies
also appear to have been developmental.
Before concluding this section, I turn to two additional pieces of data
from the Standard Cross-Cultural Sample (SCCS) (Divale 2000). While
missingness reduces the sample from twenty-eight to sixteen observations,
the data suggest a relationship between rates of warfare and degrees of
political centralization in Table 13.7.
he data also indicate that more centralized polities were more likely
to ight their neighbors (Table 13.7). Of equal relevance are the data on
increases or decreases in the size of the polities. As seen in Table 13.8,
the data from SCCS suggest that chietaincies and centralized states were
increasing in size whereas most stateless societies were shrinking.
 he Imperial Peace 433

Table 13.8. Political structures and political trajectory (percent)

Boundaries and population Stateless Chietaincies Centralized states

Expanding 20 56 60
Stable 20 11 20
Decreasing 60 33 20
N 10 9 5
Source: Standard Cross-Cultural Sample.

he small size of the sample and uncertainty as to the era to which the
coding applies attenuate the impact of these indings. hey remain sugges-
tive at best.

13.6 POLITICAL RESTRAINT AND POLITICAL STRUCTURE

hus far I have focused on the external relations of African states. I now
turn to their internal politics. Doing so we see that those who governed
states presided over a surplus: they could have extracted the rents that lay
before them. But the data also highlight the presence of internal checks.
hose with power might use it to limit feuding among powerful families
and groups of kin; to construct public works, to ight, and to tax; but given
the institutions that encompassed them, they found it costly to use this
power to seize the assets of those they governed.
he data from the HRAF suggest that states tended to form in locations
that contained relatively rich soils, abundant water supplies, and valuable
resources – gold, copper, salt, and iron. People found such locations attrac-
tive; more important, once settled in them, they were reluctant to depart.
As a result, those who exercised power could extract rents from them with-
out fear that they would defect, as did their counterparts. Diferentials in
productivity thus created the possibility of taxation. hus could states exist
in deserts, if located in the mountains (as were the Fur); or in the Sahel,
if located in inland deltas (as were Ghana, Mali, and Songhay); or in the
savannah, if located in lakes or swamplands (as were the Lozi and Lunda).
In such locations, a surplus existed that could be appropriated without fear
of political breakdown.
If states arose in such environments, why then would not their rulers
exploit their power to secure private advantages? he reports archived by
the HRAF suggest that to some degree they did, thus their possession of
superior homes or clothing and their possession of larger numbers of cattle
434 Bates

Table 13.9. Economic inequality and political structure

(percent “yes”) Do monarchs or chiefs possess?

Superior More numerous

Houses? Clothing? Livestock? Slaves
66 59 58 59
Note: N = 32–5.
Source: Data from Human Relations Area Files (Bates 1983, 56).

Table 13.10. Political checks and political structure (percent)

Are there: Is there an Aristocracy?

Policy Councils? Yes No
Yes 60 15
No 40 15
Independent Courts?
Yes 93 31
No 7 69
Note: N = 28.
Source: Data from Human Relations Area Files (Bates 1983, 57).

or slaves (Table 13.9). But the data also suggest that rulers faced political
checks. Kin groups, they suggest, remained jealous of their power and kept
important oices under their control: that of the “linguist,” among the Akan;
of the war chief and “prime minister,” among the Yoruba; or of the ritualist
and diviner, among the Lunda. As have others, those who were rich or pow-
erful claimed the standing of aristocrats and dominated key councils. he
monarch then had to gain their approval before introducing new policies,
imposing new levies, or declaring new wars. From their position of power,
commoners and aristocrats defended the independence of the courts, thus
imposing yet another limit on the power of the state (Table 13.10).
In the cross-section of societies bequeathed to the colonial era by the pre-
colonial state system, then, those who ruled were oten limited in their abil-
ity to exercise power. Ritualists could call their legitimacy into question and
threaten divine retribution should monarchs abuse their power. Councils
dominated by the heads of powerful families could oppose their policies.
In some, courts could limit their ability to bring the power of the judicial
system to bear on their political enemies. In settings where exit constituted a
costly option, internal checks nonetheless limited the use of power.
 he Imperial Peace 435

13.7 FURTHER REFLECTIONS

It is dangerous to invoke particular cases in support of a general argument.
It is impossible to prove that the cases were not selected simply to bolster
the argument; the selection, moreover, may imply that its author concurs
with the argument when he may not do so. Despite these dangers, I con-
clude by drawing on two works by eminent historian Jan Vansina.
In How Societies are Born, Vansina (2004) studies the history of societies
in south central Africa: a region that runs from the Congo in the north to
Botswana in the south and from the Atlantic in the west to the Lualaba in
the east. hese societies, he contends, initially consisted of small aggregates
of kin who foraged to survive. Over time, their populations increased and
technologies changed. Two changes proved pivotal: the domestication of
cattle and the invention of the iron hoe. Of necessity, foragers travel light;
what they need, they consume. he domestication of plants and animals –
that is, the Neolithic transformation – facilitated accumulation, however,
and notions of private property. Cattle and cultivation, he writes, “wid-
ened the concept of individual ownership, fostered a novel notion of mate-
rial wealth, and made the emergence of classes based on wealth possible”
(Vansina 2004, 99). Economic growth thus gave rise to government, Vansina
argues, by creating private property. Vansina thus sketches out a causal path
that, while difering from the one advanced here, runs parallel to it. To rec-
oncile the two, we need only add the emergence of conlict and the demand
for adjudication and the termination of disputes: conditions that could lead
to the demand for order and the means of enforcing it.
Also suggestive is Vansina’s study of the Kuba,5 a people who dwell in
Congo near the junction of the Kasai and Sankuru. His account begins with
the migration of kin groups from the north, the crossing of the Sankuru,
and their settlement in the lands wherein they now reside. Once again,
Vansina links the growth of population and the transition to permanent
cultivation to the formation of chiefdoms. Distinctive of this work, how-
ever, is his study of the emergence of a monarch capable of collecting taxes,
providing public services, and making war.
Focusing on the most celebrated monarchs – Syaam and his two imme-
diate successors, MboMboosh and Mboong a Leeng – Vansina notes how
they transformed the Kuba into a powerful and prosperous people. In dis-
cussing this transition, Vansina stresses that “the call for a surplus came

5
Also known by the name of their ruling clan, the Bushong. J. Vansina (1978). he Children
of Woot: A History of the Kuba Peoples. Madison, University of Wisconsin Press.
436 Bates

from above, from the political authorities” (Vansina 1978, 184). Syaam
and his royal lineage forged a central bureaucracy consisting of super-
visors of public works; judges, police, and messengers; inspectors and
tax collectors; and attendants to the king, his wives, and ofspring. he
expansive growth of the public sector, in turn, inspired the search for the
resources from which to pay for it. Taking advantage of the arrival of new
crops from the Americas – maize, tubers, and tobacco, for example – the
government promoted their widespread adoption.6 In addition, it pro-
moted multiple cropping. he government then began not only to extract
wealth, but also to create it. he result was the growth of public revenues
and private incomes.
To account for the behavior of this government, we need make use of
hints that lay scattered in Vansina’s account. It is clear that in “the Age of
Chief ” – as he calls the era prior to the formation of the monarchy – the
Kuba competed for territory with others who had crossed the Sankuru such
as the Ngongo, the Ngoonde, the Shoowa, and the Kel. Vansina implies, but
does not boldly claim, that the competition for territory lent impetus to
political centralization and the formation of chietaincies. When turning to
“the Age of Kings,” Vansina notes that MboMboosh and Mboong a Leeng
led the Kuba in wars against the Pyaang, Kote, Cwa, and Luba; he never
places warfare at the center of his analysis, however. A close reading of his
account suggests a possible reason: so rich and powerful were the Kuba that
they had little need to ight.
he Kuba established a “large kingdom that was generally peaceful con-
trasted with the insecurity that became more and more prevalent elsewhere”
(Vansina 1978, 165). he kingdom grew, Vansina notes, not only because its
borders expanded, but also because people migrated to it. he “ascendancy”
of the Kuba, he adds, “was due in large part to the continuing process of
ission in other kingdoms” (Vansina 1978, 164). he Pyaang did not long
remain a serious threat, he notes; rather they “fell apart” (Vansina 1978,
162). Perhaps, one might think, the enemies of the Kuba fractured because
they could not compete.
hose who held power among the Kuba thus devised strategies that
proved winning in the contested terrain of Central Africa. hey secured
themselves politically, in part by promoting their nation’s prosperity. heir
polity endured while others collapsed.

6
While not made explicit in Vansina’s account, it would appear that cultivation was pre-
scribed rather than encouraged. his certainly would be the case in the slave villages that
provisioned the residents of the capital.
 he Imperial Peace 437

While Vansina’s depiction of interstate conlict is muted, his description

of the internal politics of the Kuba is not. He makes clear the extent to which
the monarchs faced internal checks to their use of power. Powerful lineages
lay claim to the status of aristocrats, he states; and the aristocrats dominated
key councils, where they could, if united, check the king’s power and shape
his choice of policies. he monarch, Vansina makes clear, sought to pen-
etrate and control these councils. hrough intermarriage and favoritism,
he sought support from their members. But the aristocrats had interests
that difered from his and were oten unwilling to be suborned. he mon-
arch’s use of power was thus checked. Willing to back policies that provided
security and prosperity, those who populated the state councils were able to
resist any efort by the king to use public power for private gain.
Historians of Europe have noted that interstate conlict selected for soci-
eties in which power was deployed to foster the creation of wealth. We have
seen that in precolonial Africa, interstate conlict also led to the extinction
of polities: between 1700 and 1900, the number of states in West Africa
declined by more than a third. We have also seen that in postcolonial Africa,
societies with states were more likely to foster markets, long-distance trade,
and crat production and more likely as well to create public goods, such
as roads, bridges, and canals than were those that lacked centralized polit-
ical institutions. Taken together, these observations suggest that while tra-
ditional states may well have been violent, they also were developmental.

13.7.1 Discussion
During the nineteenth century, Africa was invaded, conquered, and occu-
pied. he conquerors were few in number and, to a great degree, reliant
on those whom they had conquered for information, cooperation, and
resources. Whatever their theory of rule, most found it useful to rule
through established elites. As in the times before the conquest, Africa’s elites
sought wealth and power. But now they could better satisfy those desires by
providing political service than by preparing for war. As a consequence,
they no longer had to promote development to survive. Following the parti-
tion of Africa, the impulse that earlier and elsewhere had driven those with
power to generate growth had been stilled.

13.8 COUNTERARGUMENTS
But then, why did Africa fail to develop? What checked the developmental
impulse of the African state? Clearly, it was stilled by conquest, something
438 Bates

that was not achieved as readily as is commonly assumed. People appear

to have forgotten the defeat of the Italians by the Ethiopians in East Africa
or of the British by the Zulu in South Africa. hey have forgotten as well
that the British held the Ashanti military in such high regard that the com-
manders who defeated the Ashanti became known as the “Ashanti Ring,”
and moved rapidly into leading positions in Britain’s military. As for their
putative technological superiority, pace Hilaire Belloc, European forces in
Africa oten abandoned their Maxim guns, rather than having to lug them
through Africa’s challenging terrain, and Africans were able to purchase
and import modern weapons – breach loading riles, for example – soon
ater their adoption by armies in Europe. he closeness of the contest pro-
vides an indication of the power of indigenous states.
Other factors were clearly at play, however. One was slavery; a second was
the lack of modern technology. he presence of the irst made it proitable
to use power to compel rather than to encourage efort while the absence of
the second lowered the returns that efort would yield.
Elementary reasoning suggests that under most circumstances, slave
economies will be less productive than those based on free labor.7 Not only
are the incentives to expend efort low, given that labor is highly taxed, so
too are incentives to innovate, given that the majority of those engaged in
production will fail to secure the beneits. Empirical conirmation comes
from the work of economic historians, in particular those who compare
the economies of the northern portions of North America, where labor
was free, and those of the Caribbean, South America, and the southern
states of North America, where slavery prevailed (Acemoglu, Johnson
et al. 2001; Acemoglu, Robinson et al. 2001; Engerman and Sokolof 1997;
Nunn 2008).
As documented by Lovejoy (1983), hornton (1999), Manning (1982),
and others, monarchs, princes, and heads of state in Africa presided over
polities that not only promoted commerce and produced public goods,
but also seized and enslaved numerous human beings. Some they shipped
abroad; others they sold domestically or settled on plantations, compelling
them to feed and provision the royals, oicials, soldiers, bureaucrats, and
political hangers-on. Precolonial African states may have mobilized power
to promote economic development, but, in labor markets at least, they
employed it to extract and forcefully to redistribute.

7
Where teams are more productive than isolated individuals, however, a given quantity of
labor may yield greater output when closely – and coercively – supervised. R. Fogel and S.
Engerman (1974). Time on the Cross: he Economics of American Negro Slavery. New York,
W. W. Norton.
 he Imperial Peace 439

When probing the reasons for the absence of industry, it is useful to turn to
Mokyr’s (2008) discussion of the institutional and cultural setting for scientiic
innovation in eighteenth-century England. In contrast to the patterns of exclu-
sion that characterized many British institutions, those engaged in the natu-
ral sciences created societies that were open; in pursuit of knowledge, people
from diferent social classes met to discuss, to demonstrate, and to appraise.
heir eforts focused on ields that promised human improvement: botany,
soil and animal sciences, navigation, medicine, and astronomy. As important
as the technical innovations that they fostered, Mokyr (2008) argues, were the
principles that they adopted. Knowledge, they recognized, was best pursued
socially. By the eighteenth century, where there were Englishmen, there were
societies that held public meetings where new indings were reported and
correspondence with other societies read and discussed.
In a brilliant essay, Mudimbe and Appiah (1993) probe the creation of
knowledge within African societies. here they too ind skepticism, interro-
gation, and experimentation: key steps in the scrutiny of arguments and the
testing of hypotheses. What is missing, they argue, is veriication. Drawing
on Goody (1986), Mudimbe and Appiah stress the signiicance of literacy;
without writing, they argue, it is impossible to stake a claim or to record a
method such that it could be veriied. In the absence of literacy, it is therefore
diicult to prove oneself right and someone else wrong. Innovation, compe-
tition, and destruction – that is, economic progress – is therefore slowed.
As Mudimbe and Appiah note, some artisans in Africa in fact achieved
high levels of technical proiciency. Among them numbered the ironwork-
ers, who reined and smelted ores to a degree that matched or exceeded
what had been achieved in the West prior to the Industrial Revolution.
Rather than publicly exhibiting their technologies and subjecting them
to investigation, however, the ironworkers jealously hoarded their knowl-
edge. Ritually, they set themselves apart; oten, they married endogamously,
forming a separate caste. Rather than broadcasting their knowledge, they
clothed it in mystery. heirs was the code of the guild rather than of the
scientiic society.
When the Europeans conquered Africa, then, they encountered societies
that, while willing and able to ight, lacked an industrial base and the cul-
tural practices that would might have fostered its creation.

13.9 CONCLUSION
he arguments just discussed resonate with many of the standard accounts
of Africa’s underdevelopment. here is truth in both of them, of course. But
440 Bates

Table 13.11. Slave trading and political structure (percent “yes”)

Kinship Chiefs Monarch Central bureaucracy

Present Absent

Engaged in slave trade? 0 22 78 75 17

Note: N = 31–4.
Source: Data from Human Relations Area Files (Bates 1983, 51).

here I have advanced a new claim. he imperial peace is commonly viewed

as ofering a respite from the misery of warfare and slavery. It is held to
be one of Europe’s principal contributions to Africa’s development. I have
argued the contrary. he imperial peace dismantled relationships that in
other times and places tempered the use of power, rendering those who
possessed it economically productive. Imperial rule let ample opportuni-
ties in place for political elites to employ power for self-aggrandizement.
If power can be employed to protect or to prey on the creation of wealth,
then the imperial peace may have tipped the balance in favor of the latter.
When security does not depend on prosperity, the developmental impulse
is stilled among those who rule (Table 13.11).

References
Acemoglu, Daron., James. A. Robinson et al. (2001). “he Colonial Origins of
Comparative Development.” American Economic Review, 91(1369–1401).
Ajayi, J. F. Ade and Michael Crowder, eds. (1976). History of West Africa. London:
Longman, second edition.
Axelrod, Robert. (1985). he Evolution of Cooperation. New York, Basic Books.
Banton, Michael, ed. (2004). Political Systems and the Distribution of Power. London:
Taylor and Francis, irst published 1965.
Bates, Robert H. (1983). Essays on the Political Economy of Rural Africa. Cambridge:
Cambridge University Press.
Bates, Robert H., Karen Ferree, A. J. Robinson, Smita Singh, and Anne Wren. (1996).
“Toward the Systematic Study of Transitions.” Development Discussion Paper No.
256. Cambridge, MA: Harvard Institute for International Development.
Bates, Robert H., Avner Greif, and Smita Singh. (2002). “Organizing Violence.” Journal
of Conlict Resolution, 46(5): 599–628.
Brewer, John. (1988). he Sinews of Power: War Money and the English State, 1688–1783.
New York: Knopf.
Centeno, Miguel Angel. (1997). “Blood and Debt: War and Taxation in Nineteenth
Century Latin America.” he American Journal of Sociology, 102(6): 1565–1605.
 he Imperial Peace 441

Colson, Elizabeth. (1969). “African Society at the Time of the Scramble.” In Colonialism
in Africa, 1870–1960, edited by Lewis H. Gann and Peter Duignan, 27–65. New
York: Cambridge University Press.
(1974). Tradition and Contract: he Problem of Orer. Chicago: Aldine Publishing
Company.
Crowder, Michael. (1977). West Africa: An Introduction to Its History. London:
Longman.
Curtin, Phillip, Steven Feierman, Leonard hompson, and Jan Vansina. (1995). African
History. Harlow : Longman.
Divale, William. (2000). “Standard Cross-Cultural Sample.” World Cultures, 14: 1–2.
Elvers, T. M. S. and D. Handelman, eds. (2006). Deining the Manchester School. New
York: Berghahn.
Engerman, Stanley and Kenneth Sokolof (1997). “Factor Endowments, Institutions,
and Diferential Paths of Growth among New World Economies.” In How Latin
America Fell Behind, edited by Steven Haber, 260–304. Stanford, CA: Stanford
University Press.
Evans-Pritchard, E. E. (1940). he Nuer: A Description of the Modes of Livelihood and
Political Institutions of a Nilotic People. Oxford: he Clarendon Press.
Fage, J. D. (1974). States and Subjects in Sub-Saharan African History. Johannesburg:
Witwatersrand University Press.
Fallers, Lloyd A. (1966). Bantu Bureaucracy. Chicago: University of Chicago Press.
Fogel, Robert William and Stanley L. Engerman. (1974). Time on the Cross: he
Economics of American Negro Slavery. New York: W. W. Norton.
Forde, Cyril Daryll and Phyllis Mary Kaberry, eds. (1967). West African Kingdoms in the
Nineteenth Century. Oxford: Oxford University Press.
Fortes, Meyer and E. E. Evans-Pritchard, eds. (1987). African Political Systems. New
York: KPI in association with the International African Institute.
Gilbert, F., ed. (1975). he Historical Essays of Otto Hintze. New York, Oxford University
Press.
Gluckman, Max. (1955). Custom and Conlict in Africa. Oxford: Blackwell.
Goody, Jack. (1986). he Logic of Writing and the Organization of Society. New York:
Cambridge University Press.
Gray, Richard and David Birmingham, eds. (1970). Pre-colonial Trade. London: Oxford
University Press.
Hintze, Otto. (1975). he Historical Essays of Otto Hintze, edited by Felix Gilbert. New
York: Oxford University Press.
Hirshleifer, Jack. (1994). “he Dark Side of the Force.” Economic Inquiry, 32(1–10).
(2001). “he Dark Side of the Force: Economic Foundation of Conlict heory.”
Economic Inquiry, 32: 1–10.
Hopkins, Anthony. (2000). “Quasi-states, Weak States and the Partition of Africa.”
Review of International Studies, 26: 311–20.
Kwamena-Poh, Michael, John Tosh, Richard Waller, and Michael Tidy. (1995). African
History in Maps. Harlow : Longman.
Law, Robin. (1978). “Slaves, Trade and Taxes.” In Research in Economic Anthropology,
edited by George Dalton, 37–42. Greenwich, CT: JAI Press.
Lovejoy, Paul. (1983). Transformations in Slavery. New York, Cambridge University
Press.
442 Bates

Manning, Patrick (1982). Slavery, Colonialism, and Economic Growth in Dahomey,

1640–1960. New York, Cambridge University Press.
McIntosh, Susan Keech. (1999). Beyond Chiefdoms: Pathways to Complexity in Africa.
New York: Cambridge University Press.
McIntosh, Susan Keech and Robert J. McIntosh. (1996). “West African Savanna
Kingdoms.” In he Oxford Companion to Archaeology, edited by Brian Fagan, 748–
50. New York: Oxford University Press.
Mokyr, Joel. (2008). “he Institutional Origins of the Industrial Revolution.” In
Institutions and Economic Performance, edited by Elhanan Helpman, 64–119.
Cambridge, MA: Harvard University Press.
Morrison, Donald George, Robert Cameron Mitchell, and John N. Paden. (1989). Black
Africa: A Comparative Handbook. New York: Paragon House, revised edition.
Mudimbe, V. Y. and Kwame A. Appiah. (1993). “he Impact of African Studies on
Philosophy.” In Africa and the Disciplines, edited by Robert H. Bates, V. Y. Mudimbe,
and J. O’Barr, 113–38. Chicago: University of Chicago Press.
Murdock, George Peter. (1967). Ethnographic Atlas. Pittsburgh: University of Pittsburgh
Press.
Northrop, David. (1978). Trade without Rulers. Oxford: he Clarendon Press.
Nunn, Nathan. (2008). “he Long-Term Efects of Africa’s Slave Trades.” he Quarterly
Journal of Economics, 123(1): 139–76.
Oliver, Roland A., ed. (1977). he Cambridge History of Africa. New York: Cambridge
University Press.
Parker, Geofrey. (1991). he Military Revolution: Military Innovation and the Rise of the
West, 1500–1800. New York: Cambridge University Press.
Roberts, Michael. (1956). he Military Revolution 1560–1660. Belfast: M. Boyd.
Southall, Aidan W. (1970). “he Illusion of Tribe.” In he Passing of Tribal Man in Africa,
edited by Peter Claus Wolfgang Gutkind, 28–51. Leiden: Brill.
(1988). “Segmentary State in Africa and Asia.” Comparative Studies in Society and
History, 30(1): 52–82.
(1999). “he Segmentary State and the Ritual Phase in Political Economy.” In Beyond
Chiefdoms: Pathways to Complexity in Africa, edited by Susan Keech McIntosh,
31–8. Cambridge: Cambridge University Press.
Stevenson, Robert F. (1968). Population and Political Systems in Africa. New York:
Columbia University Press.
hornton, John K. (1999). Warfare in Atlantic Africa, 1500–1800. London: UCL Press.
Tilly, C. (1975). “Relections on the History of State Making. In he Formation of
Nation States in Western Europe, edited by C. Tilly, 3–83.Princeton, NJ: Princeton
University Press.
(1985). “War Making and State Making as Organized Crime.” In Bringing the State
Back, edited by P. Evans, D. Rueschemeyer, and T. Skocpol, 169–91. Cambridge:
Cambridge University Press.
Turner, Victor W. (1957). Schism and Continuity in an African Society. Manchester:
Manchester University Press on Behalf of the Rhodes-Livingstone Institute.
Vansina, Jan. (1978). he Children of Woot: A History of the Kuba Peoples, Madison:
University of Wisconsin Press.
 he Imperial Peace 443

(2004). How Societies Are Born: Governance in West Central Africa. Charlottesville
and London: University of Virginia Press.
Warner, Carolyn M. (1999). “he Political-Economy of ‘Quasi-statehood’ and the Demise
of 19th Century African Politics.” Review of International Studies, 25: 233–55.
Yofee, Norman. (1993). “Too Many Chiefs?” In Archeological heory: Who Sets
the Agenda?, edited by Norman Yofee and Andrew Sherratt, 60–78. New York:
Cambridge University Press.
 PART IV

EXTERNAL FORCES
 14

Dahomey in the World: Dahomean Rulers

and European Demands, 1726–1894

John hornton

he Kingdom of Dahomey has a bad reputation. When the debate con-

cerning the slave trade broke out in Europe during the late eighteenth cen-
tury, Dahomey was deined as the classic example of either a pariah state
dedicated to the capture and sale of people to European slavers, or of a
state so addicted to violence that sale of its victims was an act of mercy. Yet
there is always another side to this story, and that is of a Dahomey forced
by one motive or another into the slave trade. One of the earliest writers
on Dahomey’s history, Robert Adams, whose work was published in 1735,
argued that Dahomey opposed the slave trade but that foreign pressure
compelled the kingdom to participate in it. Two inluential books about
Dahomey, however, written by Europeans who claimed to be eyewitnesses,
one by Robert Norris in 1789 and the other by Archibald Dalzel in 1793,
both published in the thick of the debate, presented Dahomey in a diferent
light, and indeed challenged factual data from Adams’s book, to reinforce
the close association between Dahomey and the slave trade (Adams and
Norris 1789; Dalzel 1793).1
When abolitionists won the day, and the British decided that they needed
to end the slave trade, Dahomey was widely regarded as an uncooperative
holdout. Reports of untrammeled slave dealing and deiance of international
pressure to stop enslaving people joined continued reports of Dahomey’s
violence against its neighbors and practice of human sacriice. Ultimately
these claims against Dahomey and its rulers justiied the colonial occupa-
tion of the kingdom by France at the end of the nineteenth century.
Modern historians have also struggled with Dahomey and its histori-
cal legacy. Isaac Akinjogbin strove to reverse the model of Dahomey as a
state devoted entirely to the slave trade by arguing that during its formative

1
See on this general point, Law (1986).

447
448 hornton

years, under King Agaja, Dahomey actually sought very hard to limit and
control the slave trade in the interest of regional harmony, and only became
devoted to the trade when it failed (Akinjogbin 1967). But Akinjogbin’s
work was criticized by a number of other historians, and the most signif-
icant of these, Robin Law, viewed Dahomey as being at the very least an
enthusiastic supporter of the slave trade, a product, he argued of its ori-
gin (according to some traditions) from bandit gangs (Henige and Johnson
1976; Law 1986, 246–7; Ross 1982). Werner Peukert, in contrast, challenged
what he called the “Atlantic theory” of Dahomey’s history, that the state
was largely organized around the slave trade. He pointed out that although
Dahomey had a reputation for military prowess, and repeated successful
military attacks on its neighbors were the source of its slaves, Dahomey was
not, in fact, particularly successful as a military state, losing a good num-
ber of its wars. Instead, he argued that Dahomey’s religious system, which
relied on human sacriice, was more important in explaining Dahomey’s
wars (Peukert 1978).
In addition to wrestling with Dahomey’s reputation as a state devoted
to the slave trade, historians have pointed to it as an example of a state-
dominated society, perhaps because the state grew out of the warfare needs
of the slave trade. K. Madhu Pannikar, an early scholar of dependency the-
ory, argued that Dahomey’s role in the slave trade led to an all-powerful
state, and indeed thwarted growth in more positive directions (Panikkar
1963). Along the same lines, Karl Polanyi used Dahomey as a model of
a state-dominated economy, with administered trade and a classic type of
the nonmarket substantive economy (Polanyi 1968). Here again, Robin Law
has entered the discussion by challenging the idea that Dahomey’s kings
held a monopoly on the slave trade, and indirectly also that Dahomey was
the sort of absolutist state Polanyi and others had argued it was (Law 1977).
At the very least, Law argued, Dahomey had a private sector, both during
the slave trade and aterward. In fact, in my own study of Dahomey’s mil-
itary system, I was able to demonstrate that the army itself was at least in
part privately controlled, though the state’s share in raising and supplying
the army increased considerably over time, especially in the late eighteenth
century and beyond (hornton 1998, 88–93). In fact, one could make a
fairly solid case that the idea of a totally dominant Dahomean state was
based on observations from the middle of the nineteenth century onward,
at the very end of the slave trade and even beyond the end.
It may well be best to think of Dahomey as a iscal-military state, a
term that has recently come into use for early modern Europe. In the
European model, most efectively elaborated by John Brewer for England
 Dahomey in the World 449

ater 1688, the iscal-military state existed to ight (Brewer 1990). It built
its taxation policies and indeed even its commercial policies around max-
imizing the ighting potential of its armies, and then used the armies for
military adventure. Brewer’s model has been subsequently expanded and
employed for a wide variety of other European states in the same period
(Glete 1992).
Dahomey can be seen in the same light. It was a military state, to be
sure, for it fought nearly constant war, and although we know much less
about its internal taxation and expenditure mechanisms than we do for
European states, it clearly sent out large armies virtually every year. But
these armies, which surely captured slaves in large numbers and exported
them in exchange for munitions and luxury goods, may not have simply
been economic instruments destined to shape the development of the state.
In short, the truncation of nineteenth-century development in Africa in
general, and Dahomey in particular, the “reversal of fortune” that moved
Africa back in the pack of countries in the development of the nineteenth
century, may not be so simply blamed on the slave trade, for all its horrible
impact on demography and politics. Africa was not alone in a reversal of
fortune, and the slave trade cannot be so simple an explanation.
In some measure, the nature of the source material that all modern
historians must use to write Dahomey’s history has determined how it
is interpreted. Much of the most available historical material was writ-
ten in the polemic accounts of the late eighteenth century. While modern
historians have had access to and made a lot of use of archival materials,
mostly records of the British, French, and Dutch African Companies and
Portuguese records from Portugal’s fort at Whydah, the most important
port in the area, these records do not give the same sort of information
about political history that the larger narrative histories do. hey are good
for checking and establishing chronology and sometimes for leshing out
details of events mentioned in the narratives, but they provide relatively
little information about Dahomey’s politics.
here are other sources, however, that have not been as systematically
studied that can provide us with much more information about Dahomey
and its politics. hese are letters written by the Dahomean monarchs them-
selves. A few of these have been known for a fairly long time; one letter,
written by Agaja in 1728 and sent to King George II of England, was pub-
lished in 1789, and French researcher Pierre Verger discovered and pub-
lished several others. However, until Robin Law republished the irst letter,
it was largely regarded as a forgery, and the letters published by Verger, while
known since the early 1960s, had not been studied on their own merits,
450 hornton

as important insights into royal thinking as presented by the rulers them-

selves.2 To these I must add the six letters that I uncovered, by singular good
fortune, in the archives of the Instituto Histórico e Geograico Brasileiro in
Rio de Janeiro in 1998. hese letters, written by King Adandozan between
1810 and 1812, are by far the longest and most detailed letters written by
any Dahomean monarch. One of them is eleven folios (22 sides) long and
another is seven folios (14 sides) long. Taken with a few other royal letters
I located in Bahia in 2003, we now have a substantial collection of letters
from Dahomey monarchs.3
his body of letters sheds some light on the single most important source
for the lengthy narrative accounts of Dahomey written in the late eighteenth
century, Norris’s Memoires of Bossa Ahadee and Dalzel’s History of Dahomey,
which declare their dependence for the historical portions of what they
wrote on the diary of Lionel Abson, the English factor at Whydah from 1767
to 1803 (Adams 1789; Dalzel 1793). Abson was close to the Dahomey court,
spoke Fon (Dahomey’s language), and married into the Dahomean elite.
According to another visitor, Robert Adams, Abson was the best informed
European on the coast (Adams 1823, 52). However, Abson’s diary has not
survived aside from its quotation and summarization by Norris and Dalzel,
and it is not always easy to determine what in the diary as Norris and Dalzel
reported it belongs to Abson, and what portion represents commentary by
the Englishmen to prove their polemical points.
he royal letters also help in the critical understanding of Abson’s account.
Abson’s chronicle has many of the same traits and interests as those found
in the long letter that Adandozan wrote to the king of Portugal outlining
his attacks on the coastal kingdoms before 1811. While one cannot be cer-
tain, the lavor of the royal letters and Abson’s accounts matches enough
that we can be somewhat reassured that Abson spoke, at least in some ways,
with the voice that was typical of Dahomean monarchs, or at least the royal
family. While this voice is muted and sometimes overwritten by its English
commentators and editors, it allows us to reconsider the matters dealt with
in his version of Dahomey’s history.

2
he original French version of this book was published in 1966 and is widely cited. In the
Portuguese edition, the texts in Portuguese are transcriptions of the original texts, not
retranslations from the French. See Law (1990) and Verger (1987, 289).
3
hese documents are found in the Instituto Historico e Geograico Brasileiro (hencefor-
ward IHGB), Lata 137, pasta 62, and Arquivo Público do Estado de Bahia (henceforward
APEB), Secção Colonial, Correspondência Recebida de Autoridades Diversas, Maço
197, Cx 76. Now republished in Luis Parés, ed. “Cartas do Daomé,” Afro-Ásia, 47 (2013):
295–395.
 Dahomey in the World 451

hese letters have some important characteristics that also need to be

discussed. It is fairly clear that no Dahomean ruler was personally liter-
ate, and so the letters were written by secretaries in Portuguese, from dic-
tation in Fon. he language of the letters is complex Portuguese, perhaps
the work of a nonnative speaker or even one whose irst language was a
Creole language. One of the authors, who declared himself the “scribe of
this cruel king,” added an unoicial note to the bottom of an 1804 letter that
he had been a prisoner for twenty-three years in Dahomey, was innocent of
any wrongdoing, and was uncomfortable in being forced to write letters.4
Undoubtedly the other letters were written by similar sorts of people, and
in fact, King Agongolo, writing in 1791, mentioned sending “all the neces-
sary governors and commissioners and white boys who know how to read
and write to respond with the captains,” to greet ships.5 It is possible that
these “white boys,” clerks in the employ of the Kingdom of Dahomey, might
subvert letters and thus misrepresent the actual intentions of the authors;
we have to assume as well that the king had reasonable checks in place to
ensure that letters were accurate, even if we do not know what they were.
his seems especially reasonable if we consider that the custom of using
“white boys” as secretaries and literate go-betweens was well established.
A second point to consider is that the letters were written for fairly spe-
ciic reasons to particular parties in Brazil (or the Portuguese court, then
residing in Brazil), and thus cover only the business that needed to be done.
hey might not take too many asides to address questions that would be
of greater interest to modern historians. In his longest letter, however,
Adandozan took the opportunity to console the king of Portugal, then
residing in exile in Rio de Janeiro, about the loss of his country, and to
describe in considerable detail several of his military campaigns as a means
of revealing to his Lusitanian brother that he might have been able to help
him against his enemies. he detailed description of a military campaign
seems to have been an obsession of the Dahomey kings, and not surpris-
ingly is the central theme of Norris and Dalzel’s chronicles as well.
A third and important point is that these letters were not written against
the backdrop of the polemic in Britain at the time concerning the slave trade.
he speeches of Dahomean rulers reported by writers like Norris or Dalzel
have always to be taken with a grain of salt, given their use in the debates

4
Biblioteca Nacional de Rio de Janeiro, (henceforward BNRJ) MS 846 Adandozan to João
Carlos da Bragança, November 20, 1804, published in Verger (1987, 289). Now repub-
lished in Pares, “Cartas do Daomé,” 359.
5
APEB, Secção Colonial, Correspondência Recebida de Autoridades Diversas, Maço 197,
Cx 76, doc. 1, Agonglo to unknown, March 31, 1790. In Pares, “Cartas do Daomé,” 330.
452 hornton

over the slave trade. But Portugal was certainly not interested in suppress-
ing the slave trade, and the Dahomean kings felt no compunction in writing
to its leaders honestly about the role of the trade in their country.
he fact that the language of the letters was Portuguese and that the
addressees were living in Portugal or Brazil and thus not participants in the
polemics of the slave trade allows us to introduce two rarely used sources
of Dahomean history that can provide an independent perspective. he
irst of these is the long report of a more or less failed religious mission
from Brazil to Dahomey led by Vicente Ferreira Pires in 1800 (Pires 1957).
Ferreira Pires, like the kings of Dahomey, was not particularly interested in
the slave trade or its suppression. He did have religious interests of course,
and it is not surprising that his report is the fullest account of Dahomean
religion written during the era of the slave trade. But it is also rich in anec-
dotal information about the functioning of the Dahomean state.
he other source is also in Portuguese and thus equally free from the
polemic of the English sources. Found in the same Brazilian institute where
the cache of royal letters now lies, and published only in 2002, this 1806
document was written by Luís Antonio de Oliveira Mendes, a Brazilian
who had never been to Dahomey. He was, however, a sugar planter and
slave owner, and had learned the Fon language to communicate with his
slaves, and, by his report, had also learned of their country from them. He
also interviewed the Dahomean ambassador to Brazil, and while visiting
Portugal, he happened to meet another long-term resident in Dahomey.
he text is interesting for its insider and occasionally rural details (Oliveira
Mendes 2002).
hese new sources, some insider accounts of Dahomey, others from visi-
tors or students of the kingdom, allow us to move beyond the polemical and
sometimes misleading accounts of its English-speaking historians.
here is no doubt from the body of royal letters that the rulers of Dahomey
considered their trade with Europe as normal and that the slave trade was a
fundamental part of it. Agongolo, in a letter of 1790, noted the care he took
to repopulate and secure his ports and reassured the Portuguese of their
trading status, and then noted in a letter of 1795 that they would be paying
for their tobacco imports in “good captives,” as in ancient times.6 When
writing to Brazil in 1795, Agongolo again reassured his Brazilian customers
that they could trade in precious materials such as silk or worked precious

6
APEB, Secção Colonial, Correspondência Recebida, Maç0 197, Cx 76, doc. 1, Agongolô to
unknown, March 31, 1790 and doc. 2, Agongolô to Fernando José, n. d. ca. 1791 [dated in
the archive]. In Pares, “Cartas do Daomé,” 330–2.
 Dahomey in the World 453

metals in security, “for this there are captives in excess there, and more of
them that can be sold against tobacco and aguardente.”7
here can be little doubt that Dahomey’s commercial interests were not
restricted to the luxury goods the king sought from his Brazilian corre-
spondents. hey were also important components of his military activities,
in that they supplied munitions. he link between warfare and captives was
made very clear in Adandozan’s letter of 1800 in which he asked Prince
João de Bragança to “do me the favor of sending two hundred or three hun-
dred barrels of powder, for which I will make good payment in excellent
captives.”8
While there can be little doubt that Dahomey waged numerous wars,
indeed they can be documented in fair detail in a plethora of sources, the
origin and rationality of these wars is more problematic. Certainly no
Dahomean monarch claimed that his wars were simply to capture people.
Lionel Abson recorded, famously, the statement Agongolo made in 1789
when, as Abson was reading to him about the parliamentary debates on the
slave trade in England, the question of the origin of wars came up. Agongolo
did not accept the idea of war for slaves at all, chiding his English visitor
that he did not understand Africans very well. He noted that England was
protected by the sea and had commerce everywhere, while Dahomey was
on a large continent, hedged in by other nations, “of the same complexion,
but speaking diferent languages,” and in this situation, “we must defend
ourselves by the sharpness of our swords to defend ourselves from their
incursions, and punish the depredations they make on us.” his, he averred,
was the cause of “incessant wars.” He went on to declare, “Your countrymen,
therefore, who allege that we go to war for the purpose of supplying your
ships with slaves, are grossly mistaken.” Continuing in this vein, he asserted
that “in the name of my ancestors and myself I aver, that no Dahoman man
ever embarked in war merely for the sake of procuring wherewithal to pur-
chase your commodities.”
But Agongolo’s wars were not just for defense, they also had judicial
dimensions. He went on to describe how he, who had not ruled the country
for long, had already killed thousands of people whom he could have sold,
but who deserved death for their actions or for other reasons. “When pol-
icy or justice require that men be put to death, neither silk, nor brandy, nor
coral nor cowries, can be substituted for the blood that must be spilt, for

7
BNRJ MS 552, Angongolô to D Maria I, March 20, 1795 in Verger (1987, 287).
8
APEB Ordens Regias, 89, fol. 345, Adandozan to Prince D. João de Bragança, undated, ca.
1800, in Verger (1987).
454 hornton

examples sake.” He continued further that wars would not stop just because
merchants from Europe no longer visited Africa, and moreover that he did
not need guns to ight, pointing out that when he dressed as the third king of
Dahomey, Weebaigh, he carried a bow, the traditional weapon of the time,
and asserting that “Prisoners in those days were killed to a man.” Finally he
challenged assertions made in Europe that Dahomeans sold their wives and
children, but that they only sold “a part of our prisoners.” Ater all, “are not
all prisoners at the disposal of their captors?” (Dalzel 1793, 218–19).
he quotation has been challenged, largely because of the use to which
Dalzel put it to, that is, to assert that in some ways the slave trade rescued
those who would be killed, and that Dahomey was addicted to war. But
in the book, Dalzel set it aside as a direct quotation, and perhaps did so
from Abson’s own notes. While there is no question that it was and remains
potentially a self-serving statement that might support the proslavery posi-
tion in the debate, it is also possible that it relects the genuine attitude of
Agongolo and his colleagues.
Subtle conirmation of this exists in the correspondence with Portugal/
Brazil. Although it is not clear, it would seem that Agongolo waged a war
to secure the trade of the port of Jaquin about 1790; he wrote to Brazil to
inform Portuguese authorities that he had secured the area from wars and
robberies, and that for that reason trade was safe for Portuguese vessels.9
In a letter of 1804, Adandozan gave as his motive for an attack on the
beaches of Allada and Porto Novo “some palavers we had among us,” in this
instance meaning simply disagreements and disputes. He continued that
it was “our custom” to seize everything in such a war, including, it turns
out, three Portuguese ships, hence the need for the letter.10 By far the most
important statement on war from Adandozan, however, came in his long
letter to the king of Portugal in 1810, in which he noted that had he known
in a timely manner, he would have sent an army to aid him in his defeat by
Napoleonic forces. He then went on to say that he had waged a number of
wars himself, “here in my land I have fought many wars in the interior both
against the nation of Magiunas and the Nagos and that in all of these I had
victory, as I also have participated in many wars that I have had to make
because I have a small country,” a sentiment that echoes his predecessor’s
complaint that he was surrounded by enemies.

9
APEB Secção Colonial, Correspondência Recebida de Autoridades Diversas, Maço 197,
Cx 76, doc. 1, Agonglo to unknown, March 31, 1790.
10
BNRJ MS 846 Adandozan to João Carlos de Bragança, November 20, 1804, in Verger
(1987). Adandozan gives the date of the action as September 7, 1804.
 Dahomey in the World 455

Indeed, he mentioned speciically a war waged with Oyo because, in his

opinion, that ruler came against him to take captives.11 He also described
a war against an island on the coast called “aGomça” as a defensive war,
waged to protect his interests, for he took many heads and did many execu-
tions, including some 703 killed or beheaded, to terrorize the ruler so that
he would not attack again.12 He also waged a war of vengeance against a
king “who was an enemy of my father’s” named Sapî Maquino, who had
killed or mistreated a female relative he had married. He intervened in a
war that Sapî Maquino had fought against the country of Aê, and as a result
of his victory, he “destroyed all the land and killed the said king and burned
the place where his sons, daughters and brothers were.”13
hus, the correspondence of Agongolo and Adandozan does in fact
lend support to Agongolo’s description of his motivations for war. Indeed,
from his point of view, the acquisition of slaves was truly a secondary fac-
tor, a given in any war, but rarely the reason for waging it. It also tends to
undermine the contention, made by the English writers, that the kings of
Dahomey waged speciic wars solely to acquire slaves. None of this means
that the slave trade was not a major portion of the income of Dahomey,
though it must also be said that no national income accounting has ever
been done to calculate how much of Dahomey’s national wealth or royal
revenues lowed directly from the slave trade.
here is little doubt that the kings of Dahomey regarded the enslavement
of people captured in war as their right, and they had strict laws against
the enslavement of their own people. But of course, this does not mean
that Dahomey did not sufer from the slave trade. In Dahomey, or rather
the Bight of Benin region of which Dahomey was a part, enslavement took
place as a result of wars more than raids. Dahomey sufered losses from the
incursions of its neighbors, just as Adandozan and Agongolo contended. For
example, the neighboring polities of Porto Novo and other coastal countries
routinely raided the south coast of Dahomey using the intercoastal water-
way. Adandozan’s attacks along those routes were in fact in response to such
raids, and Agongolo’s writings about the disruption to trading ports were
speciic references to this (hornton 1998, 76–9, 94–6).
While the kings’ letters make it appear that they were always victorious
in restoring security and overcoming their enemies, they also reveal that
there had been prior raids and attacks, which presumably had captured

11
IHGB, Lata 137, pasta 62, fol. 1, Adandozan to João Carlos de Bragança, October 9, 1810.
12
Ibid., fols. 1v–2.
13
Ibid., fols. 2v–3.
456 hornton

Dahomean subjects and enslaved them. Similarly, at an earlier period,

Werner Peukert used Abson’s detailed account of Dahomey’s campaigns
against its neighbors during the reigns of Kpengla and Tegbesu, rectiied
by more detailed and chronologically ordered commercial records, to show
that Dahomey waged important wars and sufered defeats (Peukert 1978).
In fact, Dahomey’s armies, in spite of their reputation for being warlike and
eicient, were frequently defeated by their enemies. In these cases, Dahomey
sufered losses and no doubt its people were enslaved as well. here was,
ater all, a suiciently large population of Dahomeans in Brazil that they
dominated the lay brotherhood of Santo Elesbão and Santa Eigênia in Rio
de Janeiro until the 1760s when the fraternity split, in part over their exces-
sive claims over other groups from the region (de Carvalho Soares 2004).
he letters of the kings do not entirely bear out the idea of an overly pow-
erful Dahomean state, created by war and standing in the way of private
enterprise. Adandozan was constantly aware of foreign technology and the
possibility of using it in his country. Of course, he was interested in mili-
tary equipment; in 1811 he asked speciically about diferent types of riles
and explosive shells for his army.14 But his interest went beyond these. He
was also interested in obtaining ireighting equipment for his town, for
example.
Perhaps Adandozan’s most interesting request in this regard came about
because of his observation that the Napoleonic wars had driven both French
and English slave trades out of business. He noted that there was therefore
a shortage of the sorts of commodities that European shippers brought in,
mentioning a wide range of supplies, luxury goods, cloth, and of course
tobacco and alcohol. He suggested that this presented a business oppor-
tunity for Portugal, or rather more exactly for Portugal now in Brazil. He
suggested that the king give him a ship of his own to carry goods, in which
he would be the captain and he would bear the responsibility for proit and
loss. He was conident that he would succeed in shipping and would show
them what blacks could do.15
We do not know if any of Adandozan’s requests were met, or what he
might have done with them. In 1818 he was overthrown by his brother
Guezo, whose letter announcing his usurpation of the throne and declaring
his intention to continue good relations with Brazil is the last one of the Rio
de Janeiro series.16 However, Dahomey did make a transition to addressing

14
IHGB, Lata 137, pasta 62, Adandozan to João Carlos Bragança, October 9, 1810, fol. 5.
15
Ibid., Adandozan to king of Portugal, no date (ca. 1811), fol. 7.
16
Ibid., Guezo to king of Portugal, no date.
 Dahomey in the World 457

the challenges of the nineteenth century. Dahomey was no more free of

warfare than it had been before; indeed, Guezo and his successors contin-
ued to press to expand Dahomean inluence, to shake the country free of the
nominal dependence it had on Oyo, and to ill the vacuum in power created
by the collapse of Oyo ater 1823. As these wars, like the earlier ones, con-
tinued to bring in slaves, Dahomey continued to export them, and likewise
to expand its access to foreign technology. Dahomey never failed to keep up
with the latest developments in weapons technology; in fact, by 1894 when
France conquered the country, its soldiers were equipped with rapid-iring
magazine riles, machine guns, and Krupp artillery (Garcia 1988).
At the same time, Dahomey also moved into alternate products when the
slave trade was clearly ending, but using slaves captured from the wars as
its labor force. State-managed, slave-worked palm oil plantations provided
Dahomey with the foreign exchange it needed to keep the military moving
and to support the management of Dahomey’s expanding state apparatus.
he failure of Dahomey, or indeed of any other African state, to develop
into an industrial society is a problem that needs to be addressed. he par-
ticipation of the Africans in the slave trade provides a useful, but I believe
limited answer. European states also engaged in the sort of destructive mil-
itarism that characterized Dahomey in the same period, and developed
industrial capacity anyway, and at least in some analysts’ opinion, warfare
helped develop technology. It is fairly clear that the Industrial Revolution
was not a state-run enterprise in Europe, but came out of the private sector,
oten in spite of state interference and monopoly concessions.
It is probably safe to say that in the minds of its leaders, the function of
the state in Dahomey was to wage war, in many respects like the European
iscal-military states. he wars need not simply be to feed the slave trade,
but in fact, the revenue of the slave trade was invested in war and perhaps
made more wars or bigger wars possible. he objectives of the wars, how-
ever, were generally not simply to get slaves, but to achieve a variety of other
goals, to centralize Mahi country, to inish the conquest of the coast, to
avenge insults, and to take over strategic positions, oten to further military
objectives.
his is important to keep in mind when considering Dahomey in the
post–slave trade era. he state in nineteenth-century Dahomey did partic-
ipate strongly in the economy, primarily to expand exports of palm oil, but
even then primarily to meet the goals of a iscal-military state, to continue
to ofset the costs of war and material. Furthermore, it is quite clear that
Dahomey did have a private sector, in fact quite a strong one, though pre-
cisely how the private-public competition may have worked is problematic
458 hornton

in the current state of research. Robin Law, in particular, has pointed out
that both small-scale producers and large plantations tapped palm oil for
the international market. he state controlled many of these plantations and
used war captives as labor, but such captives may have also been in private
hands or farmed out by the state to private farmers (Law 1977).
It is diicult to determine if the costs of war, especially the depopulation
elements of the type of war waged in Africa, where casualties were inlated
by exported captives, overwhelmed productive potentials. Demographic
research has not focused on Dahomey, and in any case, nineteenth-century
Dahomey may well have beneited from importing labor forcibly, and ofset
military losses. he internal slave trade of the post-abolition period pre-
sented a transfer of labor within Africa with labor going by force, oten to
the most productive regions.
It is, perhaps more worthwhile to focus on the late nineteenth century
and even the colonial period to see the true locus of the reversal of for-
tune in Dahomey. he nineteenth century was a highly dynamic period,
when improvements in oceanic transportation, a product of the Industrial
Revolution, made low-cost, high-bulk exports possible. It is in the process
of deindustrialization, where, for example, textile producers moved into
importing cloth rather than making it, and attempts to create a local textile
industry, such as that missionaries attempted at Abeokuta, failed, that we
might seek the origins of the reversal, rather than in the earlier slave trade.

References
Adams, Robert. (1823). Remarks on the Country Extending from Cape Palmas to the
River Congo. London. Reprinted, Frank Cass, 1966.
Akinjogbin, I. A. (1967). Dahomey and Its Neighbours 1708–1818. Cambridge:
Cambridge University Press.
Brewer, John. (1990). he Sinews of Power: War, Money, and the English State, 1688–
1783. Cambridge, MA: Harvard University Press.
Dalzel, Archibald. (1793). A History of Dahomey, an Inland Kingdom in Africa. London:
Spilsbury and Son.
de Carvalho Soares, Mariza. (2004). “A nação que se tem e a ‘terra’de onde se vem: cate-
gorias de inserção social de africanos no império português, século XVIII.” Estudos
Afro-Asiáticos, 26: 303–30.
de Oliveira Mendes, Luís Antônio de. (2002). “Memoria histórica sobre os costumes
particulares dos Africanos. . .” in “A Memoria histórica sobre os costumes particu-
lares dos Africanos com relação privitiva ao Reino da Guiné e nele ao respeito ao
Rei de Daomé,” edited by Alberto da Costa e Silva. Afro-Ásia, 28: 253–94.
Garcia, Luc. (1988). Le royaume du Dahomé: Face à la penetration colonial: afrontements
et incomprehension, 1875–1894. Paris: Karthala.
 Dahomey in the World 459

Glete, Jan. (1992). War and the State in Early Modern Europe: Spain, the Dutch Republic
and Sweden as Fiscal-military States, 1500–1660. London: Routledge.
Henige, David and Marion Johnson. (1976). “Agaja and the Slave Trade: Another Look
at the Evidence.” History in Africa, 3: 57–67.
Law, Robin. (1977). “Royal Monopoly and Private Enterprise in the Atlantic Trade: he
Case of Dahomey.” Journal of African History, 18: 555–77.
(1986). “Dahomey and the Slave Trade: Relections on the Historiography of the Rise
of Dahomey.” Journal of African History, 27: 237–67.
(1990). “Further Light on Bulinch Lambe and the ‘Emperor of Pawpaw’: King Agaja
of Dahomey’s Letter to George I of England, 1726.” History in Africa, 17: 211–26.
Norris, Robert. (1789). Memoires of the Reign of Bossa Ahadee, King of Dahomey.
London: Printed for Frank Lowndes, 1789. Reprinted Frank Cass, 1968.
Ross, David. (1982). “he Anti-Slave Trade heme in Dahomean History: An
Examination of the Evidence.” History in Africa, 9: 263–71.
Panikkar, K. Madhu. (1963). he Serpent and the Crescent: A History of the Negro Empires
of Western Africa. Bombay: Asia Publishing.
Parés, Luis ed. (2013). “Cartas do Daomé.” Afro-Ásia, 47: 295–395.
Peukert, Werner. (1978). “Der Atlantische Skavenhandel von Dahomey 1740–1797.”
Wirtschatsanthropologie und Sozialgeschichte. Wiesbaden: Spengler.
Pires, Vicente Ferreira. (1957). Viagem de África em o Reino de Dahomé. Claudo Ribeiro
de Lessa, ed. São Paulo: Editora Nacional.
Polanyi, Karl. (1968). Dahomey and the Slave Trade: An Analysis of an Archaic Economy.
Seattle: University of Washington Press.
hornton, John. (1998). Warfare in Atlantic Africa, 1500–1800. London: University
College of London.
Verger, Pierre. (1987). Fuxo e reluxo do tráico de escravos entre o Golfo de Benin e a
Bahia de Todos os Santos do séculos XVII a XIX. 2nd edition. São Paulo: Curropio.
 15

he Transatlantic Slave Trade and the Evolution

of Political Authority in West Africa

Warren C. Whatley

Since the publication of Douglas North’s Structure and Change in Economic

History (1981), social scientists have accumulated an abundance of new evi-
dence showing how institutions shape the structures and performance of
economies as they move through time. African economies are no exception.
he term patrimonial (or neo-patrimonial) is controversial, but it is oten
used to describe a class of postcolonial African nation-states where a weak
federal authority attempts to rule over ethnically based local authorities.
Political stability oten rests on an unstable coalition of patrons and clients
who extract large amounts of national income through personal networks,
intimidation, corruption, and bribes.1 In these kinds of “failed states,” public
policy is inluenced more by a political calculation to retain the privileges
of power than by the political authorities’ credibly committed to improve
growth and welfare. he recent spread of democratic institutions like mul-
tiparty politics has improved accountability, but stability and violence are
oten more pressing concerns. It is diicult for people to prosper under this
kind of uncertainty, unless of course they have access to a personal network
of patrons and clients that deines the resources available to them.

Department of Economics, University of Michigan, Ann Arbor, 48109. his chapter has
beneited from seminars at Washington University in St. Louis, the University of Michigan,
and Harvard University. I thank Nathan Nunn for very helpful editorial comments and
Rob Gillezeau for research assistance. Remaining errors are mine.
1
See Pitcher, Moran, and Johnston (2009) for a critical discussion of the term patrimonial.
hey argue against the use of the term. I agree with their critique, especially the charge
that as a Weberian concept patrimonialism is inappropriately applied. I use the term to
characterize a historically speciic solution to a political problem faced by many emerg-
ing postcolonial nation-states in Africa. My meaning is similar to the political formations
described in Van de Walle (2001), Boone (2003), Chabal and Daloz (1999), and Bates
(2008), and is similar in many ways to Greif ’s (1994) description of medieval Genoa before
the Podesta, except the clients here are customary local authorities.

460
 he Transatlantic Slave Trade 461

What are the historical origins of this type of political system? he

traditional view links it to the era of colonial rule and how it birthed a dis-
jointed and decentralized nation-state. Mamdani (1996) is perhaps the
best-known proponent of this view. He sees the most important political
outcome of colonial rule to be a kind of dual legal system – European laws
for colonial rulers and a variety of “customary” laws for colonial “subjects.” In
a very real sense, the postcolonial nation-state is the colonial state adjusted
to new political realities like democratic elections, new global economic
forces, stronger internal challenges, and mounting international debt.
In this chapter, I argue that the historical roots go deeper than colonial-
ism and can be found in the absolutist authority structures that emerged
out of the slave trade era. To make my case, I link the data on slave exports
found in the Transatlantic Slave Trade Database with data on authority
structures found in Ethnographic Atlas (Murdock 1967). My purpose is to
see if the transatlantic slave trade transformed local political authority in
Africa prior to colonial occupation. I ind that it did. I then suggest ways in
which indirect colonial rule institutionalized these authority structures and
transformed them into a kind of local rule of law that continues to extract
surplus and challenge federal authority today.

15.1 THE SEARCH FOR ORIGINS AND PATH DEPENDENCE

In his book States and Power in Africa, Jefrey Herbst (2000) advances
the idea that African political authority is ancient in origin and rooted in
Africa’s harsh physical environment, an environment that has always con-
strained population growth and elites’ capacity to broadcast political power.
Herbst argues that low population density constrained the reach of politi-
cal authority long before colonial rule. In fact, colonial authorities wrestled
with the same physical constraints and failed, leaving behind a state appa-
ratus too weak to govern efectively over its exogenously created political
territories.
Herbst is not the only scholar to invoke labor scarcity to explain an African
institution. Anthony Hopkins (1973), in his widely read economic history
of Africa, uses labor scarcity to explain technological choices and the insti-
tution of slavery. John Ilife (2007), the renowned historian of Africa, uses
labor scarcity to explain the institution of polygyny. Gareth Austin (2008)
revises and extends the application of the idea and James Fenske (2010)
applies it to property rights.
I believe there is a fundamental truth in this idea. Following the show-
ing of the ilm Eve, which argues that ancient Africans migrated out of
462 Whatley

Africa and populated the world, I ask my class: “If Africans could pop-
ulate the world, then why couldn’t they populate Africa?” If Africa was
labor scarce, then certainly it was because of the environment and not the
people. he major problem with this view is not consistency but lack of
evidence. here are no census counts of an Africa population prior to the
nineteenth century, and the widely used estimates for earlier times found
in McEvedy and Jones (1978) are educated guesses and the extrapolations
of current estimates back in time. hese types of exercises rely on a vari-
ety of counterfactual assumptions regarding historical rates of population
growth and the impact of the slave trades on African populations (see
Manning 1990).
Here I take a diferent approach. I use the fact that slave exports reduced
African population. I then look to see if cross-sectional variations in slave
exports can explain cross-sectional variations in institutions that others have
attributed to environmentally induced labor scarcity. hanks to the Trans-
Atlantic Slave Trade Database, some slave exports can now be measured
with a high degree of precision. In a very real sense, a focus on the slave
trade opens the door to an empirically grounded assessment of the popu-
lation hypothesis, albeit a door with a slightly diferent view. Rather than
seeing labor scarcity as the result of a steadfast environmental constraint, a
focus on the slave trade emphasizes an economic shock to Africa.2
his shock certainly reduced population, but it also set of a chain reac-
tion of warfare and raiding among African societies that “changed the way
wars were waged at the lowest level” (hornton 1999, 151). Robin Law, in
his study of the Slave Coast, concludes that “by the end of the seventeenth
century, the European demand for slaves had brought about a profound
transformation of African societies of the Slave Coast. Although this was
primarily an economic transformation, it had dramatic efects in the polit-
ical sphere also, in the collapse of political order leading to the rise of the
new state of Dahomey” (Law 1991, 345). One inds similar conclusions in
Barry’s (1998) study of the Senegambia region and Daaku’s (1970) study
of the Gold Coast. Inikori (1982, 2003) argues that the slave trade encour-
aged the formation and spread of banditry and small militarized states.
And in a review of Herbst’s States and Power in Africa, James Robinson
argues that we should “hypothesize that slavery-induced predatory institu-
tions had signiicant adverse inluence on development paths not just in the
Americas where the slaves were used, but also in Africa where the slaves

2
For two estimates of this shock, see Whatley and Gillezeau (2011a) and Gemery and
Hogendorn (1979).
 he Transatlantic Slave Trade 463

originated. . . . Nonetheless, this is ultimately an empirical issue” (Robinson

2002, 517–18).
A decade later, and with a few notable exceptions, the idea that the slave
trade altered African political institutions remains largely untested. Whatley
and Gillezeau (2011a) model both efects mentioned previously. hey treat
the slave trade as an increase in the market value of anyone who can be
enslaved. he shock alters factor prices in a way that mimics labor scarcity,
but it also turns political incentives to build states into economic incen-
tives to raid for slaves. he model predicts decentralized and militarized
absolutist political authorities more interested in raiding for slaves than
growing citizens. Nunn and Wantchekon (2011) present evidence showing
that centuries of slaving in Africa fomented mistrust among people and
toward their institutions, and that this slavery-induced mistrust persists to
today. Whatley and Gillezeau (2011b) present evidence that the slave trade
increased ethnic fragmentation in Africa and that this too persists to today.
Nunn (2008) inds that slaving hampered long-term economic progress.
Among the many possible channels is its impact on economic and political
institutions.
In this chapter, I follow these leads and ask if the slave trade inluenced the
institutions of authority in Africa and whether these inluences persist over
the long term. I am not concerned with assessing the relative importance of
environment versus shock. It could very well be the case that the slave trade
merely tightened a labor constraint that had been binding for millennia. I
do not ofer a way to distinguish between these two views, nor am I partic-
ularly concerned to. My goal is more limited: to trace the long-term efects
of the slave trade on the evolution of political authority in Africa.
As such, the efects, if any, must pass through colonialism. Did slavery-
induced precolonial authority structures survive colonialism? he litera-
ture is surprisingly silent on this question. Herbst devotes an entire chapter
of States and Power to a discussion of authority in precolonial Africa, but he
never mentions the transatlantic slave trade (Herbst 2000, 35–57). Recent
econometric studies of African political economy show how remnants of
precolonial authority continue to inluence local outcomes in Africa, but
none of these studies considers the possibility that the precolonial leg-
acies they identify are traditions of authority that date back to the slave
trade (Englebert 2000; Gennaioli and Rainer 2007; Michalopoulos and
Papaioannou 2010, 2011). Historians consider the link between the slave
trade and African politics, but, with the exception of a few radical histo-
rians like Walter Rodney (1972) and Basil Davidson (1992), they are not
particularly interested in the long-term impact on development.
464 Whatley

Causal history is. It is an empirical project expressly designed to statis-

tically identify and estimate causal relationships between past events and
subsequent outcomes (see Nunn 2009 and Fenske 2010 for recent reviews).
his line of research oten produces the irst salvo in the search for his-
torical legacies, and institutions oten take center stage as carriers of lega-
cies.3 Like a pulling guard on the gridiron, causal history suggests that some
paths forward are more likely than others. he actual play, however, unfolds
on the ground. As Paul David put it in his classic article on the economics
of QWERTY, “it is sometimes not possible to uncover the logic (or illogic)
of the world around us except by understanding how it got that way” (David
1985, 332). Avner Grief (2006) has theorized on the concept “path of play”
in historical settings and has focused attention on how self-enforcing insti-
tutions of social cooperation (the shared playbook in our football analogy)
can order how the play unfolds.
In the remainder of this chapter I follow the leads of this causal history
of Africa and trace the impact of participation in the transatlantic slave
trade on the evolution of political authority. I present econometric evidence
showing how the transatlantic slave trade caused the spread of absolutism
in sub-Saharan West Africa. I argue that slavery-induced absolutism also
inluenced the way politics were played in the colonial era and beyond. I
present aggregate evidence that British colonies that exported more slaves
were ruled less directly by colonial administrations. I argue that indirect
colonial rule instead institutionalized these slavery-induced absolutist
authority structures and transformed them into a kind of rule of law at the
local level. he postcolonial nation-state, like the colonial state before it,
lacks the administrative apparatus or the political clout to govern efectively
over these local authorities, even when it wants to. From this perspective,
the structure of many postcolonial nation-states in Africa is rooted in a
political history that is unique to Africa in many respects, a history that
stretches at least as far back as the era of the transatlantic slave trade.

15.2 THE TRANSATLANTIC SLAVE TRADE AND

POLITICAL AUTHORITY IN AFRICA
Slave trading in Africa has a long and nuanced history. How are we to
think of it as a shock to Africa? It is best to begin with the position of
Africa in medieval Mediterranean civilization. Italian city-states, Arab
caliphates, and European kingdoms all used conquest and trade to obtain

3
On path dependence as legacy in competitive equilibrium, see Arrow (2004).
 he Transatlantic Slave Trade 465

treasure, land, and slaves. Muslim and Christian armies fought “just” war
that justiied the enslavement of nonbelievers. Roman and Islamic law
contained well-developed civil codes regulating the legal status of slaves
(Patterson 1982).
he southern frontier of this system was Africa, down its east coast by
sea and across the Sahara desert by camel. By the middle of the iteenth
century, lateen sails enabled Portuguese caravels to sail down the west coast
of Africa as well. hen came the discovery of the Americas, the decima-
tion of American populations, and the transplantation of slave codes across
the Atlantic. In the hands of the emerging capitalist nations of Europe, the
proximity of cheap African labor was an opportunity to proit from and
defend their American colonies. By the beginning of the eighteenth cen-
tury an old sporadic long-distance African trade in slaves had become the
export engine of growth (Bean 1974; Eltis 1994; Eltis and Jennings 1994;
Gemery and Hogendorn 1990). he real price of slave exports rose 500 per-
cent during the eighteenth century (Eltis and Richardson 2004; Richardson
1991). Exports increased by 1,000 percent (Eltis 2009).
Asking about the efects of these slave trades on the evolution of political
authority in Africa might seem farfetched, but we can focus the question
and break it down into stages. Nunn (2008) asks if there is a correlation
between the numbers of slaves coming out of diferent regions of Africa
and the economic performance of those regions today. His is an ambitious
attempt to locate and count up all of the slaves taken by the trans-Saharan,
Indian Ocean, Red Sea, and transatlantic trades. He then looks to see if slave
exports in the past correlate with GDP levels today. Ater controlling for a
variety of other factors that might explain the correlation, including reverse
causality, he concludes that slaving and slave trading depressed long-term
economic growth. His estimates of slave exports have been used by Nunn
and Wantchekon (2011), Whatley and Gillezeau (2011b), and others.
Here I adopt a slightly diferent empirical approach, one designed to bet-
ter isolate the slave trade as a shock to Africa. First, I limit the analysis to
the transatlantic slave trade because these data are much more reliable than
data on the trans-Saharan, Red Sea, or Indian Ocean trades. he transatlan-
tic slave trade evolved as a core component of European mercantile systems
operating in the Atlantic. Records of ship manifests, government agencies,
insurance companies, and newspapers have been digitized into the widely
used Trans-Atlantic Slave Trade Database. Estimates of the Muslim-based
trades are much less reliable in terms of volume, dates, and places (Austen
1979, 1987).
466 Whatley

Second, I conine my analysis to the transatlantic slave trade on the west

coast of Africa. Ater 1807 the transatlantic slave trade expanded on the
east coast of Africa to avoid British eforts to suppress the trade on the west
coast. I exclude the east coast extension because East African societies had
previously been inluenced by the Indian Ocean and Red Sea trades, for
which slave export data are much less precise.
hird, there is little evidence of slave trading along the west coast of Africa
prior to the transatlantic slave trade. Evidence on West African slave trad-
ing in the iteenth and early sixteenth centuries is very thin (Fage 1980).
Most of the slave trading in the Upper Guinea Coast and in the Western
Sudan were southern extensions of the Muslim-based Mediterranean sys-
tems, and were centered on Lake Chad and the interior delta of the Niger
River (Fisher 2001; Lovejoy 2004). Rodney (1966) surveys evidence for the
Upper Guinea Coast and inds slaves as servants in the households of elites,
as elsewhere in the Mediterranean world, but little evidence of the kind of
systematic slave raiding and slave marketing that would become the norm
centuries later.
True, by the time the Portuguese arrived on the Lower Guinea Coast in
the late iteenth century the Muslim-based network had extended as far
south as the Gold Coast of present-day Ghana. Wilks (1993) documents
how Muslim traders from the north supplied the gold-mining Akan of
Ghana with the slave workforce they needed to clear their rainforests for
agriculture. Not to be outdone, Portuguese traders stimulated a second
slave trade into the Akan gold ields. hey sailed yachts back and forth
along the Lower Guinea Coast, purchasing war captives from Benin’s polit-
ical expansion further down the coast and selling them to the Akan for
their gold. (Graham 1965; Ryder 1969; Wilks 1982). hese are certainly
slave trades and they certainly predate the eighteenth-century explosion
in the transatlantic trade, but they are long-distance slave trades and pre-
cisely because societies along the west coast of Africa were much less pred-
atory than they would become by the eighteenth century (Daaku 1970;
Law 1991; hornton 1999).
Finally, not only is the magnitude of the shock and its timing better mea-
sured on the west coast of Africa, so too is its location. Most of the thirty-four
thousand voyages found in the Trans-Atlantic Slave Trade Database have
some information on the location of slave purchases. Sometimes location is
by port. I use GIS sotware to merge the locations of ports with the locations
of societies found in Murdock’s Ethnographic Atlas. Murdock’s goal was to
code the ethnographic features of societies prior to substantial European
 he Transatlantic Slave Trade 467

inluence. I look to see if slaving inluenced these authority structures, so in

a sense I am testing Murdock too.
Figure 15.1 depicts how I merged the two data sets. he mapping of
subnational societies comes from the digitized version of the map found
in Murdock’s (1967) Ethnographic Atlas. I overlay on top of this the loca-
tions of the slave ports found in the Transatlantic Slave Trade Database.
hese are denoted by the stars along the coast. I then construct two hun-
dred observations from the merged spatial data. I divide the west coast of
Africa into two hundred evenly spaced points, each ity kilometers apart,
running from Tunis, Tunisia to Cape Town, South Africa. he dependent
and independent variables are constructed from spatial data that fall within
predeined bufers around these points. hese are depicted by the circles in
Figure 15.1.
he treatment variable is the numbers of slaves exported from ports in
these bufers (SLAVES). Environmental controls are the average values of
environmental features inside these bufers (X): forest and desert cover-
age, elevation, local agricultural suitability as measured by climate and ter-
rain slope, and population density in 1960 (UNESCO 2010; USGS 2010;
IIASA 2010). he dependent variables are three authority types derived
from variable V72 in Ethnographic Atlas called “Succession to the Oice
of Local Headman.” his variable is dichotomized three diferent ways.
ABSOLUTIST = 1 for Murdock’s code 1 = “patrilineal heir,” and code 2 =
“matrilineal heir;” ABSOLUTIST = 0 otherwise. LIBERAL = 1 for Murdock’s
code 5 = “inluence, wealth or social status, nonhereditary;” LIBERAL = 0
otherwise. DEMOCRATIC = 1 for Murdock’s code 6 = “election or other
formal consensus, nonhereditary;” DEMOCRATIC = 0 otherwise.
Table 15.1 reports unweighted means. he irst three columns report
means for the Western sample, the full 200 bufer sample covering the
western half of the continent of Africa from Tunis, Tunisia to Cape Town,
South Africa. he top panel is for bufers with a 250 km radius. he bot-
tom panel is for bufers with a 500 km radius. On average, 316,050 slaves
were exported from each of the 200 bufers with a 250 km radius.4 One

4
his implies more than 100 million slave exports, but the Transatlantic Slave Trade
Database contains information on approximately 13 million slave exports. he discrep-
ancy is due to the overlap of bufers causing each port’s slave exports to be recorded more
than one time. he center of each bufer is 50 kms away from the next center. With a radius
of 250 km (or a diameter of 500 km) the ports are canvassed once if we take the average
exports of every tenth bufer: 200/10=20, which yields 20 x 316,050 = 6,321,000 slave
exports. he Transatlantic Slave Trade Database has speciic port information for approxi-
mately 6,494,200 West African slave exports, so our ports pick up the vast majority of slave
exports that have information on port.
 GRUNSHI GURENSI MOBA HAUSA HAUSA
DOROSIE DAGARI BUILSA
MAMPRUSI
SOMBA Benin KURAMA, GURE (NE) JARA
CHAKOSS KAMBERI KAMUKU
‘ kina Faso RAUDEBA
BUSA JERAWA CHAWAI (SW)
LOBI BIRITON WABA KADARA AFUSARE
KONKOMBA KABTE BARGU
KATAB
DAGOMBA BIROM
KILINGA GBARI
KORO
ANGAS
BASARI BASILA YESKWA
VAGALA GUANG NUPE MAMA
MADA
TBM
DIULA
KULANGO LIG BI DEGHA(SE) GWANDARA
EGBA GILI
ATYUTI YORUBA ANO
NAFANA ADELE AFO
BUNU ARAGO
KRACHI
BRONG TRIBU
BASA
GAN
Ghana Nigeria
KEBU
e d’ Ivoire ANA IGBIRA NGE
BAULE UEM AKPOSO IFE IDOMA TIV
FON
EKITI IGALA

TOGO
KUKURUKU
IYALA
ASHANTI AVA UEBU
GUN ORRI EGEDE
WE
468

ANYI AKYBM OPO

EDO
ADANGME BOKI
MB BVBE
ATTIE
IRSEKIRI IBO
GA YAKO BOKI
ISOKO EKBI
EBRIEMEKYIBO FANTI
ASSINI
ODODOP
ASSINI
IBIBIO
Camer
UAW
KUNDU

PE

Figure 15.1. Merging the Transatlantic Slave Trade Database and Ethnographic Atlas.
 Table 15.1. Sample means

Western sample Atlantic sample Sub-Saharan sample

N Mean SD N Mean SD N Mean SD

250 km radius bufers
Slave Exports (000) 200 316.05 225.62 169 374.02 238.34 131 482.52 255.32
Absolute Authority 117 0.62 0.41 86 0.84 0.20 48 0.88 0.18
Democratic Authority 117 0.16 0.30 86 0.07 0.17 48 0.04 0.13
Liberal Authority 117 0.03 0.13 86 0.03 0.10 48 0.00 0.00
469

Agricultural Suitability 200 2.89 1.10 169 2.99 1.11 131 2.83 0.99
Population in 1960 200 10.73 15.36 169 9.36 15.81 131 9.91 16.74
Elevation 200 12.56 6.58 169 11.95 6.31 131 12.12 6.18
Forest 200 0.15 0.24 169 0.16 0.26 131 0.19 0.28
Desert 200 0.54 0.65 169 0.53 0.69 131 0.34 0.59
500 km radius bufers
Slave Exports (000) 200 632.1 392.70 169 748.05 407.97 131 965.04 424.58
Absolute Authority 180 0.59 0.39 149 0.71 0.31 111 0.80 0.21
Democratic Authority 180 0.17 0.26 149 0.13 0.24 111 0.06 0.09

(continued)
 Table 15.1. (continued)

Western sample Atlantic sample Sub-Saharan sample

N Mean SD N Mean SD N Mean SD

Liberal Authority 180 0.03 0.11 149 0.04 0.12 111 0.02 0.03
Agricultural Suitability 200 2.89 1.30 169 2.90 1.29 131 2.71 1.17
Population in 1960 200 7.00 8.66 169 6.20 8.56 131 6.56 9.04
Elevation 200 13.09 7.64 169 12.80 7.37 131 12.87 7.54
Forest 200 0.27 0.47 169 0.30 0.50 131 0.36 0.55
470

Desert 200 1.10 1.45 169 1.02 1.51 131 0.60 1.34
Notes: N is the number of observations or bufers with non-missing data. Means are the unweighted averages of observations in the sample. SD is sample standard
deviation.
Sources: I divide the west coast of Africa into 200 evenly spaced points starting at the northernmost point of Tunisia and ending at the middle of South Africa.
Variables are constructed from spatial data falling within a radius of 250 km or 500 km around these points. he Western sample contains all 200 bufers, from
Tunis, Tunisia to Cape Town, South Africa. he Atlantic sample runs from Gibraltar to Cape Town. he sub-Saharan sample runs from Cape Blanc to Cape Town.
SLAVES are in thousands (0,000) of slaves exported from ports that fall within a bufer (Eltis 2009). Environmental controls are forest and desert coverage, eleva-
tion, local agricultural suitability as measured by climate and terrain slope, and population density in 1960 (UNESCO 2010; USGS 2010; IIASA 2010). Authority
code variable V72 in the Ethnographic Atlas (Murdock 1967) called “Succession to the Oice of Local Headman.” V72 is dichotomized three diferent ways.
ABSOLUTIST = 1 for Murdock’s code 1 = “patrilineal heir,” and category 2 = “matrilineal heir;” ABSOLUTIST = 0 otherwise. LIBERAL = 1 for Murdock’s code 5 =
“inluence, wealth or social status, nonhereditary;” LIBERAL = 0 otherwise. DEMOCRATIC = 1 for Murdock’s category 6 = “election or other formal consensus,
nonhereditary;” DEMOCRATIC = 0 otherwise.
 he Transatlantic Slave Trade 471

hundred seventeen of the 200 bufers have at least one ethnic group with
information on authority type. he unweighted average shows 62 per-
cent of the societies being governed by absolutist authority, 16 percent
being governed by democratic authority, and only 3 percent with lib-
eral or wealth-based access to authority. One hundred eighty of the 500
km bufers have information on authority type. he number of hits is
larger because the larger bufers cover more territory and have a higher
probability of containing at least one ethnic group with information on
authority type.
he next three columns report means for the Atlantic sample, which
begins at approximately Gibraltar and runs to Cape Town. his sample cov-
ers the African coastline that faces the Atlantic Ocean and the Americas.
It excludes the African coast that faces the Mediterranean. he number
of observations falls to 169. he average number of slaves per observation
increases to 374,020. Absolutism increases to 84 percent.
he third sample is the sub-Saharan sample, and begins near Cape Blanc
on the southern edge of the Sahara desert. he number of observations
declines to 131, slave exports increase to 482,520, and absolutism increases
to 88 percent. All of the slaves in all of the samples were captured and
exported from sub-Saharan Africa. he sample on authority type gets small
(48 observations), but it serves as a check on the others by making sure
the estimated impact of exports on authority are not being identiied of of
unobserved regional heterogeneity.
he null hypothesis is loosely derived from Whatley and Gillezeau
(2011a), where slaving is modeled as a violent production process – what
Patterson (1982) calls “social death.” he increased violence increases indi-
viduals’ demands for protection – a public good best provided by political
authorities who have the power to authorize and coordinate a defense (or
ofense). Like all individuals in times of war, individuals subject to slave
capture will pay more for protection, including relinquishing freedoms and
rights that might otherwise be cherished in times of peace. It is not incon-
ceivable that centuries of slaving created an environment where peace had
become a distant memory – where a state of war seemed like a natural state
of afairs. A rough calculation reported in Whatley and Gillezeau (2011a)
estimates that between 1700 and 1850 a West African person faced a one
in ive chance of being swept up in the transatlantic slave trade sometime
in his life. Under these conditions it is not inconceivable that individuals
will relinquish political rights and accept absolute rulers (Bates, Grief, and
Singh 2002). It is not inconceivable for political authority to become more
472 Whatley

absolutist in nature where previously there had been some semblance of

democratic or liberal inluence on authority.
I look for evidence of this by performing the following OLS regressions:
AUTHORITY TYPEji = α + β SLAVESi + Ω Xi + vi
Observations are bufers. AUTHORITY TYPEji is the average value of
AUTHORITY TYPEj (ABSOLUTE, LIBERAL OR DEMOCRATIC) for
ethnic groups whose centroid falls in bufer i. SLAVESi is the total number
of slaves exported from the slave ports in bufer i. Environmental variables
are mean values for spaces in bufers. Xi is the vector of environmental con-
trol variables. vi is the error term. he larger circular bufers reach farther
inland, but they overlap more along the coast. I calculate Conley standard
errors to account for spatial correlation in the errors across observations.
OLS results are reported in the columns labeled OLS in Table 15.2. In the
250 km samples, slave exports are positively and statistically signiicantly
correlated with absolutist authority. All of the estimated coeicients are
positive and signiicant at the 99 percent conidence level. he correlation
is also large. Mean slave exports explain between 16 and 40 percent of the
absolutism recorded in Ethnographic Atlas. In the sub-Saharan sample the
correlation is also signiicant and large, so the relationship between slaving
and absolutism is not due to unobserved regional heterogeneity.
I interpret these correlations as if the direction of causality ran from slav-
ing to absolutism, but there could be some reverse causation in the esti-
mated relationship. Absolutist authority, for example, could have been a
more eicient way to organize slave raids, giving rise to a reverse correla-
tion running from absolutism to slave exports. Also, Whatley and Gillezeau
(2011a) discuss conditions under which absolutist rulers might autho-
rize halhearted defenses of their people. Coalition building across ethnic
groups required some degree of power sharing among elites, but if elites can
insulate themselves from capture, and if they care little about the welfare
of their people, then they may prefer to leave their people exposed to cap-
ture so long as they themselves are safe. his outcome is more likely where
authority is inherited and absolute, and less likely in democratic systems
where leaders are held accountable.
In an efort to identify a causal relationship that runs from slaving to
absolutism I perform a two-stage instrumental variables estimation of the
relationship between slaving and absolutism. In the irst stage I seek a set
of exogenous predictors of slave exports that are uncorrelated with the
error terms vi.. Here I make use of the instruments developed by Nunn
 Table 15.2. he impact of the transatlantic slave trade on absolutist authority in West Africa

Western sample Atlantic sample Sub-Saharan sample

OLS IV OLS IV OLS IV

250 km radius for bufer

(β) Est. coeicient on slave exports (000,000) 0.0779*** 0.0590 0.0380*** 0.0755*** 0.0412*** 0.0636***
Conley standard errors 0.0230 0.1367 0.0116 0.0197 0.0133 0.0156
Environmental controls YES YES YES YES YES YES
Number of observations 117 117 86 86 48 48
Adjusted R-squared 0.4393 0.1833 0.4236
Hansen-Sargan test of overid. restrictions 4.07 0.01 4.76
(X̄) Mean slave exports (000,000) 3.260 3.260 3.740 3.740 4.825 4.825
(Ȳ) Mean pct. with Absolute Authority 62.19 62.19 84.02 84.02 88.1 88.1
β*X̄*100 25.40 19.23 14.20 28.24 19.87 30.67
Pct. of Absolutism (Ȳ) explained by (X̄) 40.84 30.92 16.90 33.61 22.56 34.81
473

500 km radius for bufer

(β) Est. coeicient on slave exports (000,000) 0.0225* 0.177*** 0.0040 0.0342 0.0080 0.0134
Conley standard errors 0.0157 0.0079 0.0105 0.0431 0.0082 0.0129
Environmental controls YES YES YES YES YES YES
Number of observations 180 180 149 149 111 111
Adjusted R-squared 0.2234 0.0535 0.3781
Hansen-Sargan test of overid. restrictions 10.03 6.16 1.91
(X̄) Mean slave exports (000,000) 6.321 6.321 7.480 7.480 9.650 9.650
(Ȳ) Mean pct. with Absolute Authority 58.72 58.72 70.94 70.94 79.78 79.78
β*X̄*100 14.21 11.2 3.00 25.56 7.68 12.91
Pct. of Absolutism (Ȳ) explained by (X̄) 24.19 190.82 4.23 36.03 9.62 16.19
Notes: Dependent variable = Pct. of societies with ABSOLUTIST authority; the coeicients in the columns labeled “OLS” are estimated OLS. he two-stage estimates in
the column labeled “IV” were estimated GMM. *** denotes 99 percent conidence level. ** denotes 95 percent conidence level. * denotes 90 percent conidence level.
All two-tail tests. he Hansen-Sargan test statistic of over-identiication restrictions is distributed Chi-Square with two degrees of freedom. he critical value for the
90 percent conidence level is 4.61. Conley standard errors allowed for spatial correlation over ive contiguous bufers for the 250 km bufers and over ten contiguous
bufers for the 500 km bufers.
Sources: See Table 15.1.
474 Whatley

(2008), which in this case are the distances between the center of each
bufer and the nearest destination in the Americas and North Africa.5
hese distances, plus the environmental controls, capture variations in
slave exports that are not likely to be correlated with authority type. In
the second stage I look for correlations between estimated slave exports
and absolutism. In the second stage I also control for the impact of envi-
ronment on the adoption of absolutist authority. I rely on Hansen-Sargan
tests of over-identiication restrictions to test for the absence of correla-
tions with the error terms vi.
In Table 15.3 I report irst-stage regressions of slave exports on distance
and environmental controls. hese are reduced form equations, so it is dii-
cult to interpret coeicients, but the F-statistics and the R-squares show that
the exogenous regressors capture a signiicant amount of the variation in
slave exports. In general, distance to the Americas ofered some protection
against capture. Also, fewer slaves were exported from desert regions and
from higher elevations. his could be due to fewer people living in these
regions, but some of the population efect is already captured by the vari-
ables measuring agricultural suitability and 1960 population, so desert and
elevated regions may have also provided some refuge from slavers (Nunn
and Puga 2012).
he second-stage results are reported in the columns labeled IV in
Table 15.2. In the Western sample of 250 km bufers, the coeicient on slave
exports loses statistical signiicance, but as we move toward sub-Saharan
Africa the estimated coeicients get larger and tighter. In the sub-Saharan
sample, 16.9–34.8 percent of average observed absolutism is explained by
average slave exports. he relationship between slave exports and absolut-
ism is less apparent in the 500 km samples. All of the coeicients are still
positive, but they fail to achieve signiicance, except in the Western sample.
It is not clear why this is the case. One possibility is that the impact of slav-
ing is less direct farther inland. Most of the slaves pulled into the orbit of the
transatlantic slave trade originated within 250–300 kilometers of the coast,
although ater the late eighteenth century higher slave prices covered longer
treks to the coast. here are forty-eight observations in the sub-Saharan
sample that have both 250 km and 500 km bufers. Absolutism is 88 per-
cent in the smaller bufer but declines to 72 percent in the larger bufers that
reach farther inland. he student-t statistic on diference in means is 3.46.

5
he American destinations are Virginia, Havana, Haiti, Kingston, Dominica, Guyana,
Salvador, and Rio de Janeiro; and the North African destinations are Algiers, Tunis,
Tripoli, Benghazi, and Cairo.
 Table 15.3. First-stage IV regressions

250 km radius bufer 500 km radius bufer

Western Atlantic Sub-Saharan Western Atlantic Sub-Saharan
sample sample sample sample sample sample
Distance to American destinations −4.38*** −3.04* −5.74 −7.46*** −5.84** 10.81
(1.23) (2.10) (8.76) (2.26) (3.06) (17.39)
Distance to N. African destinations −2.23 3.80 0.74 −5.20 3.93 30.93
(2.95) (8.39) (16.79) (4.50) (14.68) (35.26)
Agricultural suitability 32.57*** 50.19** 18.97 152.76*** 181.57*** 142.95*
(1.23) (30.71) (37.96) (51.63) (60.40) (82.21)
Population in 1960 4.24*** 3.92 4.74* 10.17** 11.45** 6.94
(1.60) (4.88) (2.47) (5.65) (6.91) (5.97)
475

Elevation −10.01*** −14.19*** −13.23*** −24.45*** −31.75*** −21.75

(3.73) (4.88) (6.44) (7.50) (9.97) (16.00)
Forest 165.77 140.03 164.67 297.00* 275.11* 283.91*
(169.24) (177.82) (182.24) (157.91) (160.36) (159.08)
Desert −88.67** −114.67** −132.72** −105.47*** −113.82** −113.60**
(35.56) (51.61) (56.86) (39.76) (53.62) (53.75)
Constant 540.60** 154.63 538.73 884.87*** 316.41 −2061.30
(223.24) (536.71) (1378.22) (324.80) (881.03) (2792.58)
Number of observations 200 169 131 200 169 131
Adjusted R-Squared 0.3515 0.3367 0.2865 0.6037 0.5941 0.5286
F-statistic 16.41 13.18 8.46 44.31 36.13 21.82
Notes: Dependent variable = slave exports in thousands; coeicients estimated OLS. Conley standard errors in parentheses. *** denotes 99 percent coni-
dence level. ** denotes 95 percent conidence level. * denotes 90 percent conidence level. All two-tail tests. Conley standard errors allow for spatial corre-
lation over ive contiguous bufers for the 250 km bufers and over ten contiguous bufers for the 500 km bufers.
Sources: See Table 15.1.
476 Whatley

Not inding signiicant efects farther inland does not necessarily mean
the impact of the slave trade was limited to the coast, although it does
mean the issue deserves further investigation. he relationship could have
becomes less direct and more complex as the process moved inland. Also,
the inland areas were subject to slave raiding for shorter periods of time,
so inland peasants might have been more successful at resisting extreme
revolutions in authority. he empirical strategy in this essay captures only
revolutionary changes across categories of authority, from non-absolut-
ist to absolutist. It does not capture increases in the degree of absolut-
ism within authority types, something that was undoubtedly occurring
as well.
his interpretation inds some support in the regression for democ-
racy and liberalism. hese are reported in Tables 15.4 and 15.5. First note
that all of the estimated coeicients have negative signs. Slaving tended to
reduce liberalism and democracy. here are no liberal societies in the 250
km sub-Saharan sample, but there are some in the sample that reaches 500
km inland. Slaving reduced liberalism, even in the interior of sub-Saha-
ran Africa. he results for democracy are also negative, but they never
achieve statistical signiicance in the sub-Saharan sample. Perhaps the
designation “democratic” is lexible enough to absorb elements of abso-
lutism while retaining some degree of accountability, like in the case of
the Asante kingdom situated inland from the Gold Coast (Rattray 1929;
Wilks 1975).
Patriarchy is another form of authority, and George Murdock (1949,
1967) claims to have theoretically proven that much of the patriliny in
precolonial West Africa was of recent historical origin. His proof is based
on the observation that many precolonial West African societies with pat-
rilineal rules of descent retained legacy kinship terms more appropriate
for other lineage systems. Not being a historian, and having never set foot
in Africa, Murdock never claimed that the slave trade transformed line-
age rules, but his list of possible causes could have easily included it. Basil
Davidson, the preeminent historian of Africa who chronicled Africa’s his-
tory from slavery to the present, does highlight slaving. He states, “As
the slaving state became increasingly a predator, kinship systems were
strengthened and elaborated as a means of providing protection against
the dangers of the violence created by the slave trade” (Davidson 1992,
266). Predation not only shited more power onto already powerful men
but it may have also, as Ekeh argues, increased the “entrenchment of kin-
ship corporation” (1990, 660).
 Table 15.4. he impact of the transatlantic slave trade on LIBERAL authority in West Africa

Western sample Atlantic sample Sub-Saharan sample

OLS IV OLS IV OLS IV

250 km radius for bufer
(β) Est. coeicient on slave exports −0.0084 −0.0247 −0.0068* −0.0044 NA NA
(000,000)
Conley standard errors 0.0063 0.0269 0.0041 0.0063
Environmental controls YES YES YES YES
Number of observations 117 117 86 86
Adjusted R-squared 0.0313 0.11
Hansen-Sargan test of overid. 0.83 3.363
restrictions
477

500 km radius for bufer

(β) Est. coeicient on slave exports −0.0037* −0.0084 −0.0057** −0.0125** −0.0031** −0.0070**
(000,000)
Conley standard errors 0.0022 0.0087 0.0024 0.0056 0.0016 0.0031
Environmental controls YES YES YES YES YES YES
Number of observations 180 180 149 149 111 111
Adjusted R-squared 0.1671 0.1759 0.2823
Hansen-Sargan test of overid. 1.66 0.0010 2.37
restrictions
Notes: Dependent variable = Pct. of polities with LIBERAL authority; the coeicients in the columns labeled “OLS” are estimated OLS. he two-stage estimates in
the columns labeled “IV” were estimated GMM. *** denotes 99 percent conidence level. ** denotes 95 percent conidence level. * denotes 90 percent conidence
level. All two-tail tests. he Hansen-Sargan test statistic of over-identiication restrictions is distributed Chi-Square with two degrees of freedom. he critical value
for the 90 percent conidence level is 4.61. Conley standard errors allowed for spatial correlation over ive contiguous bufers for the 250 km bufers and over ten
contiguous bufers for the 500 km bufers.
Sources: See Table 15.1.
 Table 15.5. he impact of the transatlantic slave trade on democratic authority in West Africa

Western sample Atlantic sample Sub-Saharan sample

OLS IV OLS IV OLS IV

250 km radius for bufer
(β) Est. coeicient on slave exports −0.0456** −0.0357 −0.0081 −0.0210* −0.0067 −0.0044
(000,000)
Conley standard errors 0.0196 0.1211 0.0057 0.0126 0.0053 0.0071
Environmental controls YES YES YES YES YES YES
Number of observations 117 117 86 86 48 48
Adjusted R-squared 0.1129 0.2174 0.6961
Hansen-Sargan test of overid. restrictions 3.04 0.1483 1.52
478

500 km radius for bufer

(β) Est. coeicient on slave exports −0.0185** −0.0935*** −0.0072 −0.0193 −0.0020 −0.0095
(000,000)
Conley standard errors 0.0083 0.0341 0.0064 0.0259 0.0041 0.0066
Environmental controls YES YES YES YES YES YES
Number of observations 180 180 149 149 111 111
Adjusted R-squared 0.1618 0.1188 0.2238
Hansen-Sargan test of overid. restrictions 3.17 7.83 0.0820
Notes: Dependent variable = Pct. of polities with DEMOCRATIC authority; he coeicients in the columns labeled “OLS” are estimated OLS. he two-stage
estimates in the column labeled “IV” were estimated GMM. *** denotes 99 percent conidence level. ** denotes 95 percent conidence level. * denotes 90 percent
conidence level. All two-tail tests. he Hansen-Sargan test statistic of over-identiication restrictions is distributed Chi-Square with two degrees of freedom. he
critical value for the 90 percent conidence level is 4.61. Conley standard errors allowed for spatial correlation over ive contiguous bufers for the 250 km bufers
and over ten contiguous bufers for the 500 km bufers.
Sources: See Table 15.1.
 he Transatlantic Slave Trade 479

15.3 FROM PREDATION TO RULE OF LAW

What about colonialism? How did colonialism interact with these new slav-
ery-induced predatory political customs? Over the past decade a series of
inluential articles has laid the empirical foundation for approaching this
type of question. One strand of this literature searches for the long-term
impact of legal customs transplanted around the world by European col-
onizing authorities, contrasting primarily the experiences of common law
countries and civil law countries (Glaeser and Shleifer 2002; La Porta 2008).
Another series of articles by Acemoglu, Johnson, and Robinson (2001, 2002)
emphasizes the long-run impact of colonial property rights laws laid down
by European colonial authorities. In colonies where the colonizer’s strategy
(or capability) was to settle and produce, one tends to ind property rights
systems that constrain states’ rights to expropriate surplus from produc-
ers. Where incentives to settle and produce were not so great, the colonial
authority tended to establish institutions that facilitated resource extrac-
tion, and oten by force. Over the long term, settler institutions encouraged
economic development while extractive institutions did not.
hese are useful perspectives, but one needs to be careful when applying
them to Africa. Incorporating the transatlantic slave trade can help. First,
Albouy (2013) argues that thus far the settler data are not good enough
to test for the long-term efects of extraction versus settlement in Africa.
he slave trade was deinitely an extractive institution, and the numbers on
exports are good measure of how extractive it was across time and space.
Second, it is important to note that prior to the Berlin Conference of
1885 the oicial colonization of Africa was conined to a few coastal pro-
tectorates. Most of Africa was oicially colonized for less than seventy-ive
years, contrasting sharply with other areas of the world like the Americas,
the Caribbean, India, and many parts of Asia. Acemoglu, Johnson, and
Robinson (2001, footnote 1) interpret the slave trade as an extractive colo-
nial institution, but others do not. Olsson (2009), for example, presents
evidence that the length of colonial rule encouraged the development of
democratic institutions over the long term, but they deine colonialism in
Africa as formal colonial occupation. Given the impact of the slave trade on
political customs in Africa, incorporating the slave trade as a phase of colo-
nialism might help explain why African democracy difers from democracy
in other former colonies around the world.
Finally, Acemoglu and Johnson (2005) argue that property rights institu-
tions have had a greater impact on structure and performance than have
legal origins because ineicient legal rules can be circumvented by the
480 Whatley

private ordering of contracts (Grief 1989). here are a number of studies,

however, documenting the lasting impact of legal origin in Africa, although
the channels are not always clear. Colonies under British rule, and therefore
administered by common law traditions, tend to exhibit higher postcolo-
nial incomes per capita, more investments in education, and higher rates
of school enrollment than do colonies administered by French civil codes
(Bertocchi and Canova 2002; Grier 1999; Price 2003).
One possible explanation has to do with the way the legal customs of col-
onizer and colonized interacted to shape the institutions of colonial rule. Sir
Frederick Lugard, irst governor-general of the Colony and Protectorate of
Nigeria, is generally considered the architect of the British policy of indirect
rule, but Mamdani and Herbst argue that the policy evolved on the ground,
as a solution to the problem of controlling colonial subjects. Mamdani (1996)
argues that indirect colonial rule grew out of a recognition that direct rule,
a single legal system encompassing colonizer and colonized but peppered
with racial distinctions, was untenable in the long term because it encour-
aged class-based and race-based protests against injustices. Gradually colo-
nial authorities moved toward a sharing of power with real or imagined local
chiefs controlling real or imagined “customary” lands. Herbst (2000) argues
that indirect rule evolved out of a colonial recognition that broadcasting
centralized authority over African territory was too costly to be efective.
Here the legal customs of colonizers can matter in reshaping the politi-
cal landscape. Firmin-Sellers (2000), for example, compares the institutions
of French and English colonial rule among the ethnically similar Akan
peoples who straddled the colonial border between the Ivory Coast and
Ghana. Consistent with civil law tradition, the French colonial system was
designed to achieve a clear directive from above: maximize tax collection.
he French undermined the authority of paramount chiefs, chose lower-
order divisional chiefs as their agents, took periodic censuses of the popu-
lation to estimate potential tax revenue, paid chiefs a share of the taxes they
collected, and monitored chiefs’ performance by checking collected taxes
against expected taxes.
he British, in contrast, careful not to undermine local custom (read
common law), chose the higher-order paramount chiefs as their agents and
allowed them to rule over the divisional chiefs and lineages below them.
hey paid chiefs a ixed salary and deposited tax collections in a public
treasury for the development of public works. he British did not moni-
tor chiefs’ tax collection like the French did, but instead monitored chiefs’
performance as public servants and administrators of justice. Based on this
case study one would conclude that the French system was a more eicient
 he Transatlantic Slave Trade 481

100

90

80

70
Percent indirect rule

60

50

40

30
Indirect = 4.9227*** (SLAVES) + 36.262
R2 = 0.5165
20

10

0
–4 –2 0 2 4 6 8 10
In (slave exports/acre)
Figure 15.2. he slave trade and indirect rule.

system of resource extraction but perhaps less developmental than the

British system.
One could also conclude that the colonial authorities, both French and
British, empowered slavery-induced absolutist power structures. Figure 15.2
presents aggregate evidence that indirect colonial rule relied on the kind of
absolutism that I estimated in the previous section. he igure graphs the
relationship between slave exports and the degree of indirect rule among
British colonies. he unit of analysis is the postcolonial nation-state. he
sample is former British colonies in Africa. he measure of indirect rule is
taken from Lange (2004, 2009) and is the percentage of legal cases adjudi-
cated in African customary courts. Slave exports per acre are taken from
Nunn (2008). he igure shows that indirect rule is positively and signii-
cantly correlated with past slave exports.
his is understandable. he colonial goal is to control population and
extract surplus. he political objective of indirect colonial rule is to take
the overlapping authorities that kin, clans, patriarchs, and elders previously
used to adjudicate local conlict and invest them in a chief who now has the
authority to rule over his people on behalf of the colonizer. Enlisting a pow-
erful local authority was probably easier where the slave trade had already
entrenched absolutist authority structures.
482 Whatley

Bates (2010) contains a concise discussion of how the integration of kin-

ship systems like these into formal state structures transforms local customs
into rule of law. In pre-state systems the heads of kin and clan mete out justice
(also see Murdock 1949). Crimes are crimes against persons, and crimes by a
member of one group against a member of another get settled by intergroup
exchange of some kind. his is what elders and lineage heads do. A rea-
sonable indicator of the incorporation of these types of authority structures
into formal state structures is the appearance of crimes, not against persons,
but against the state, signaling a public monopoly on violence, or rule of
law. Indirect colonial rule essentially legitimized slavery-induced predatory
political customs and institutionalized them as part of the colonial state.6

15.4 CUSTOMARY AUTHORITY AND THE POSTCOLONIAL

NATION-STATE
What about the postcolonial era? In this section I argue that these pred-
atory political customs passed through colonialism into the postcolonial
era and continue to inluence the structure and performance of African
political economies today. At the federal level is the disjointed postcolonial
nation-state facing diiculty with broadcasting power. At the local level is
absolutist customary authority. he poor governance, ethnic conlict, and
corruption observed in many postcolonial African states are unstable equi-
librium outcomes on the path of play among these parties (Bates 2008;
Chabal and Daloz 1999; Herbst 2000; Lange 2009; Van de Walle 2001).
At the federal level, the path begins with the Berlin Conference of 1885
where European powers agreed to cooperate in carving Africa into territo-
ries that would become postcolonial nation-states. he process was swit,
surgical, and haphazard, with colonial powers oten drawing straight-
line borders inland from the coast, and taking possession of real estate
as quickly and as easily as they would eventually relinquish it. Davidson
recounts the random and haphazard nature of colonial occupation on the
Gold Coast. Noting that in the 1880s the king of Asante ofered the British
authorities monopoly control of all Asante trade, Davidson writes, “Now
one might think that the British government, eager to acquire a monop-
oly of commercial control over West African lands, but far less eager to
meet the almost certain high cost of conquest, would have jumped at this

6
Contrast these legal origins with the legal origins of common law in England, where cus-
tom came to rely on a judiciary independent of the king and lords, and based on trial by
juries of peers. See Glaeser and Shleifer (2002) for an insightful discussion.
 he Transatlantic Slave Trade 483

ofer. But the British ministers in charge at home now wanted more than
monopolistic commercial control; they wanted territorial ownership. Partly
in order to keep out the French . . ., but even more, as most evidence seems
to show, because a demand for territorial ownership had become an impe-
rial obsession and even a popular cause . . . ” (Davidson 1992, 70–1). But “no
more than 60 years or so ater preferring to invade . . . British imperial gov-
ernment found that it had little to gain, ater all, from territorial possession
and reverted to the policy of withdrawal” (Davidson 1992, 72).
Lange (2004, 2009) tracks the evolution of federal authority from colo-
nialism to the present among the former British colonies. He focuses on
the disjointed postcolonial nation-state that emerged out of the dual legal
systems of indirect rule. Lange shows how country-level measures of indi-
rect rule (measured by the percent of legal cases adjudicated in custom-
ary courts) are highly correlated with poor governance structures today.
Indirect colonial rule produced a postcolonial nation-state crippled by inef-
fective bureaucracy, restrictions on private contracting, weak rule of law,
and widespread corruption.
Of course, indirect colonial rule could be correlated with poor postco-
lonial outcomes for other reasons. Iyer (2010), for example, shows how in
British India colonial authorities indirectly ruled the poorer regions and
directly ruled the richer ones. his kind of occupation strategy could pro-
duce the same correlation between indirect colonial rule and poor postco-
lonial governance that Lange inds in British Africa. As a preliminary check
on this I take country-level population density in 1900 (from Nunn 2008)
as a proxy for country income in 1900. A simple regression of indirect rule
on population density has an insigniicant and positive coeicient, not the
signiicantly negative coeicient one would expect if indirect rule chased
poor lands in Africa too.7
Given weak federal authority in postcolonial Africa, one should expect
precolonial authorities to continue to assert themselves. Michalopoulos and
Papaioannou (2010) use satellite light density at night to proxy for local
economic development and ind that the borders of precolonial authorities
explain more cross-sectional variations in economic activity than do the
borders of postcolonial nation-states. Gennaioli and Rainer (2007) show that
regions in Africa that had concentrated authority structures in the precolo-
nial era are better able to secure public goods for their constituents today.
Patron-clientism is the glue that holds this system of authority together.
It need not be deadly, but the cost in terms of political stability can be

7
Indirect Rule = 0.9092 * ln(POP DENSITY 1900) + 47.024, SE = .802, R2 = 0.1429.
484 Whatley

extremely high. Greif (1994) attributes the civil wars of medieval Genoa to
this kind of political gamesmanship. Van de Walle (2001) argues that the
iscal crises of the postcolonial nation-state in Africa are not due to exces-
sive government spending, but to states’ inability to collect revenues from
local authorities. Michalopoulos and Papaioannou (2011) ind that much
of the postcolonial political violence and civil war in Africa is located near
precolonial ethnic homelands that were divided by the haphazard drawing
of colonial borders. Divided ethnicities need not lead to conlict, but a his-
tory of slaving and the mistrust it engenders might be a fuel that feeds the
lame (Nunn and Wantchekon 2011).

15.5 CONCLUSION
I have argued in this chapter that the lens of the transatlantic slave trade
provides a useful perspective on contemporary politics in Africa. First,
my main point is that the precolonial authority structures of West Africa
that are recorded in the Ethnographic Atlas are not traditional in the sense
of being old, but are more recent in origins and closely tied to the ways
African societies responded to the shock of the transatlantic slave trade.
he image of a decentralized, absolutist, isolated village exploiting women
is not an image of ancient Africa. It is a modern one.
Second, I have nothing invested in the term patrimonial, but I do want to
suggest that the postcolonial nation-state in Africa is rooted in African his-
tory – a history fundamentally altered by the slave trade and colonialism.
his path dependence is seen in all of the empirical studies linking precolo-
nial authority structures to postcolonial political and economic outcomes.
My contribution to this literature is a reinterpretation of what it means to be
precolonial in the African context.
Finally, I do not mean to suggest that history determined postcolonial
political outcomes in Africa, or that the political problem to be solved was
as simple as I suggested. I do, however, want to suggest that viewing cus-
tomary authority as something of recent origin can encourage change. “You
sometimes cannot understand the logic (or illogic) of the world without
knowing how it got that way.” Radelet (2010) identiies seventeen emerging
countries in Africa. Four of the ive reasons he cites for their success have
to do with improvements in policy and governance. What Ekeh (1975) calls
the “two publics” in Africa, and what Van de Walle (2001) calls “permanent
crisis,” and what Chabal and Daloz (1999) call “disorder as political instru-
ment” need not be permanent nor disorderly. he irst order of business,
however, is to understand how it got that way.
 he Transatlantic Slave Trade 485

References
Acemoglu, Daron and Simon Johnson. (2005). “Unbundling Institutions.” Journal of
Political Economy, 113(5): 949–95.
Acemoglu, Daron, Simon Johnson, and James A. Robinson. (2002). “Reversal of Fortune:
Geography and Institutions in Making of Modern World Income Distribution.”
Quarterly Journal of Economics, 117: 1231–94.
(2001). “he Colonial Origins of Comparative Development: An Empirical
Investigation.” American Economic Review, 91(5): 1369–1401.
Albouy, David. (2013). “he Colonial Origins of Comparative Development: An
Empirical Investigation: A Comment.” American Economic Review, 102(6):
3059–76.
Arrow, Kenneth J. (2004). “Path Dependence and Competitive Equilibrium.” In History
Matters: Essays in Economic Growth, Technology and Demographic Change, edited
by Timothy W. Guinnane, William A. Sundstrom, and Warren C. Whatley, 23–35.
Stanford, CA: Stanford University Press.
Austen, Ralph A. (1979). “he Trans-Saharan Slave Trade: A Tentative Census.” In he
Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade, edited
by Henry. A. Gemery and Jan. S. Hogendorn, 23–75. New York: Academic Press.
(1987). African Economic History: Internal Development and External Dependency.
London: James Curry.
Austin, Gareth. (2008). “Resources, Techniques, and Strategies South of the Sahara:
Revising the Factor Endowments Perspective on African Economic Development,
1500–2000.” Economic History Review, 61: 587–624.
Barry, Boubacar. (1998). Senegambia and the Atlantic Slave Trade. Cambridge:
Cambridge University Press.
Bates, Robert. (2008). When hings Fell Apart: State Failure in Late-Century Africa. New
York: Cambridge University Press.
(2010). Prosperity and Violence: he Political Economy of Development. New York: W.
W. Norton.
Bates, Robert, Avner Greif, and Smita Singh. (2002). “Organizing Violence.” Journal of
Conlict Resolution, 46(5): 599–628.
Bean, Richard. (1974). “A Note on the Relative Importance of Slaves and Gold in West
African Exports.” he Journal of African History, 5(3): 351–56.
Bertocchi, G. and F. Canova (2002). “Did Colonization Matter for Growth? An Empirical
Exploration into the Causes of Africa’s Underdevelopment.” European Economic
Review 46(10): 1851–1871.
Boone, C. (2003). Political Topographies of the African State: Territorial Authority and
Institutional Choice. Cambridge, Cambridge University Press.
Chabal, Patrick and Jean-Pascal Daloz. (1999). Africa Works: Disorder as Political
Instrument. Bloomington: Indiana University Press.
Daaku, Kwame Yeboa. (1970). Trade and Politics on the Gold Coast, 1600–1720: A Study
of the African Reaction to European Trade. London: Clarendon.
David, Paul. (1985). “Clio and the Economics of Qwerty.” American Economic Review,
76: 332–7.
Davidson, Basil. (1992). he Black Man’s Burden: Africa and the Curse of the Nation-
State. New York: hree Rivers Press.
486 Whatley

Ekeh, Peter P. (1975). “Colonialism and the Two Publics in Africa: A heoretical
Statement.” Comparative Studies in Society and History, 17(1): 91–112.
(1990). “Social Anthropology and Two Contrasting Uses of Tribalism in Africa.”
Comparative Studies in Society and History, 32(4): 660–700.
Eltis, David. (1994). “he Relative Importance of Slaves and Commodities in the Atlantic
Trade of Seventeenth-Century Africa.” Journal of African History, 35(2): 237–49.
(2009). Voyages Database. Voyages: he Transatlantic Slave Trade Database. http://
www.slavevoyages.org (accessed October 1, 2010).
Eltis, David, Stephen D. Behrendt, David Richardson, and Herbert S. Klein. (1999).
he Trans-Atlantic Slave Trade: A Database on CD-Rom. London: Cambridge
University Press.
Eltis, David and Lawrence C. Jennings. (1988). “Trade between West Africa and the
Atlantic World in the Pre-Colonial Era.” American Historical Review, 9(4):
936–59.
Eltis, David and David Richardson. (2004). “Prices of African Slaves Newly Arrived
in the Americas, 1673–1865: New Evidence on Long-Run Trends and Regional
Diferentials.” In Slavery in the Development of the Americas, edited by David Eltis,
Frank. D. Lewis, and Kenneth. L. Sockolof, 181–218. Cambridge: Cambridge
University Press.
Englebert, Pierre. (2000). “Pre-Colonial Institutions, Post-Colonial States, and Economic
Development in Tropical Africa.” Political Science Research Quarterly, 53(1): 7–36.
Fage, John. (1980). “Slaves and Society in Western Africa C.1445-C.1700.” Journal of
African History, 21(3): 289–310.
Fenske, James. (2010). “he Causal History of Africa: A Response to Hopkins.” Economic
History of Developing Regions, 25(2): 177–212.
Firmin-Sellers, Kathryn. (2000). “Institutions, Context, and Outcomes: Explaining
French and British Rule in West Africa.” Comparative Politics, 32(3): 253–72.
Fisher, Humphrey. (2001). Slavery in the History of Muslim Black Africa. London: Hurst
and Company.
Gemery, Henry A. and Jan S. Hogendorn. (1979). “he Economic Cost of West African
Participation in the Atlantic Slave Trade: A Preliminary Sampling for the Eighteenth
Century.” In he Uncommon Market: Essays in the Economic History of the Atlantic
Slave Trade, edited by H. A. Gemery and J. S. Hogendorn. New York: Academic
Press.
Gemery, Henry and Jan Hogendorn. (1990). “Evidence on English/African Terms of
Trade and 18th Century.” Explorations in Economic History, 27: 157–77.
Gennaioli, Nicola and Ilia Rainer. (2007). “he Modern Impact of Precolonial
Centralization in Africa.” Journal of Economic Growth, 12: 185–234.
Glaeser, Edward L. and Andrei Shleifer. (2002). “Legal Origins.” Quarterly Journal of
Economics, 117(4): 1193–1229.
Graham, James. (1965). “he Slave Trade, Depopulation and Human Sacriice in Benin
History: he General Approach.” Cahiers d’Etudes Africaines, 5(18): 317–34.
Grief, Avner. (1989). “Reputation and Coalition in Medieval Trade: Evidence on the
Maghribi Traders.” Journal of Economic History, 49(4): 857–82.
(1994). “On the Political Foundations of the Late Medieval Commercial Revolution:
Genoa during the Twelth and hirteenth Centuries.” Journal of Economic History,
54(4): 217–87.
(2006). Institutions and the Path to the Modern Economy: Lessons from Medieval Trade.
Cambridge: Cambridge University Press.
 he Transatlantic Slave Trade 487

Herbst, Jefrey. (2000). States and Power in Africa: Comparative Lessons in Authority and
Control. Princeton, NJ: Princeton University Press.
Hopkins, Anthony. (1973). An Economic History of West Africa. New York: Columbia
University Press.
IIASA. (2010). “Global Agro-Ecological Assessment for Agriculture in the 21st Century:
Methodology and Results.” http://www.iiasa.ac.at/research/luc/gaez/index.html
(Accessed October 1, 2010).
Ilife, John. (2007). Africans: he History of a Continent. New York: Cambridge University
Press.
Inikori, J., Ed. (1982). Forced Migration: he Impact of the Slave Trade on African Societies.
New York, Africana Publishing Company.
Inikori, J. (2003). “he Struggle Against the Transatlantic Slave Trade: he Role of the
State.” In Fighting the Slave Trade: West African Strategies. S. A. Diouf, (ed.). Athens,
Ohio University Press.
Grier, R. M. (1999). “Colonial Legacy and Economic Growth.” Public Choice 98(3):
317–335.
Iyer, Lakshmi. (2010). “Direct Versus Indirect Colonial Rule in India: Long-Term
Consequences.” Review of Economics and Statistics, 90(4): 693–713.
La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer. (2008). “he
Economic Consequences of Legal Origin.” Journal of Economic Literature, 46(2):
285–332.
Lange, Matthew. (2004). “British Colonial Legacy and Political Development.” World
Development, 32(6): 905–22.
(2009). Lineages of Despotism and Development: British Colonialism and State Power.
Chicago: University of Chicago Press.
Law, R. (1991). he Slave Coast of West Africa, 1550–1750 : he Impact of the Atlantic
Slave Trade on an African Society. Oxford, Oxford University Press.
Lovejoy, Paul E. (2004). Slavery on the Frontiers of Islam. Princeton, NJ: Markus Wiener
Publishers.
Mamdani, Mahmood, ed. (1996). Citizen and Subject: Contemporary Africa and the
Legacy of Late Colonialism. Princeton, NJ: Princeton University Press.
Manning, Patrick. (1990). Slavery and African Life: Occidental, Oriental, and African
Slave Trades. Cambridge: Cambridge University Press.
McEvedy, Colin and Richard Jones. (1978). Atlas of World Population History.
Harmondsworth: Penguin.
Michalopoulos, Stelios and Elias Papaioannou. (2010). “Divide and Rule or the Rule of
the Divided? Evidence from Africa.” Discussion Paper Series No. 8088. Center for
Economic Policy Research.
(2011). “he Long-Run Efects of the Scramble for Africa.” NBER Working Paper No.
17620.
Murdock, G. P. (1949). Social Structure. New York, Macmillan Co.
(1967). Ethnographic Atlas. Pittsburgh: University of Pittsburgh Press.
North, Douglas. (1981). Structure and Change in Economic History. New York: Norton.
Nunn, Nathan. (2008). “he Long Term Efects of Africa’s Slave Trades.” Quarterly
Journal of Economics, 123(1): 139–76.
(2009). “he Importance of History for Economic Development.” Annual Review of
Economics, 1: 65–92.
Nunn, Nathan and Diego Puga. (2012). “Ruggedness: he Blessings of Bad Geography
in Africa.” Review of Economics and Statistics, 94(1): 20–36.
488 Whatley

Nunn, Nathan and Leonard Wantchekon. (2011). “he Slave Trade and the Origins of
Mistrust in Africa.” American Economic Review, 111(7): 3221–52.
Olsson, Ola. (2009). “On the Democratic Legacy of Colonialism.” Journal of Comparative
Economics, 37: 534–51.
Patterson, Orlando. (1982). Slavery and Social Death: A Comparative Study. Cambridge,
MA: Harvard University Press.
Pitcher, Anne, Mary H. Moran, and Michael Johnston. (2009). “Rethinking
Patrimonialism and Neopatrimonialism in Africa.” African Studies Review, 52(1):
125–56.
Price, G. N. (2003). “Economic Growth in a Cross-Section of Nonindustrial Countries:
Does Colonial Heritage matter for Africa?” Review of Development Economics 7:
478–495.
Radelet, Steven. (2010). Emerging Africa: How 17 Countries Are Leading the Way.
Washington, DC: Center for Global Development.
Rattray, Capt. R. S. (1929). Ashanti Law and Constitution. Oxford University Press.
Richardson, David. (1991). “Prices of Slaves in West and West Central Africa: Towards
an Annual Series, 1698–1807.” Bulletin of Economic Research, 43(1): 21–56.
Robinson, J. A. (2002). “States and Power in Africa by Jefrey I. Herbst: A Review Essay.”
Journal of Economic Literature 40(2): 510–519.
Rodney, Walter. (1966). “African Slavery and Other Forms of Social Oppression on the
Upper Guinea Coast in the Context of the Atlantic Slave Trade.” Journal of African
History, 7(3): 431–43.
(1972). How Europe Underdeveloped Africa. London: Bogle-L’Ouverture
Publications.
Ryder, Alan F. (1969). Benin and the Europeans, 1485–1897. New York: Humanities
Press.
hornton, John K. (1999). Warfare in Atlantic Africa, 1500–1800. London: UCL Press.
UNESCO. (2010). “Unep Sioux Falls Clearninghouse.” http://Na.Unep.Net/Datasets/
Datalist.Php. (Accessed October 1, 2010).
USGS. (2010). “Usgs Geographic Data Download.” http://Edc2.Usgs.Gov/Geodata/
Index.Php. (Accessed October 1, 2010).
Van de Walle, Nicolas. (2001). African Economies and the Politics of Permanent Crisis,
1979–1999. Cambridge: Cambridge University Press.
Whatley, Warren and Rob Gillezeau. (2011a). “he Fundamental Impact of the Slave
Trade on African Economies.” In Economic Evolution and Revolution in Historical
Time, edited by Paul Rhode, Joshua Rosenbloom, and David Weiman, 86–110 .
Stanford, CA: Stanford University Press.
(2011b). “he Impact of the Slave Trade on Ethnic Stratiication in Africa.” American
Economic Review, 101(3): 571–6.
Wilks, Ivor. (1975). Asante in the Nineteenth Century: he Structure and Evolution of a
Political Order. London, New York: Cambridge University Press.
(1982). “Wangara, Akan and Portuguese in the 15th and 16th Centuries, II: he
Struggle for Trade.” he Journal of African History, 23(4): 463–72.
(1993). Forests of Gold: Essays on the Akan and the Kingdom of Asante. Athens: Ohio
University Press.
 16

Gender and Missionary Inluence in

Colonial Africa

Nathan Nunn

16.1 INTRODUCTION
Increasingly, evidence has emerged showing that the historical presence of
European missionaries is an important factor afecting economic develop-
ment across the globe today. he causal mechanism that has received the
most attention, and the most empirical support, is education. he presence
of Christian missionaries, particularly Protestant missionaries, has been
shown to be strongly correlated with increased educational attainment,
and the efects appear to persist for many generations (Bai and Kung 2011;
Grier 1997; Woodberry 2004, 2009).1 he education-promoting inluence
of Protestant missionaries is not surprising given the evidence that the rise
of Protestantism within Europe had long-term education (and economic
growth) promoting efects (Becker and Woessmann 2009, 2008, 2010;
Schaltegger and Torgler 2009).
his study provides evidence for the long-term impact of colonial mis-
sionary activity within Africa. he analysis uses data on the location of
Protestant and Catholic missions from a map titled “Ethnographic Survey

I thank Robert Barro, Oeindrila Dube, Rachel McCleary, Murat Iyigun, and Robert
Woodberry for valuable comments, as well as Sayon Deb and Eva Ng for excellent research
assistance. I also thank seminar participants at Harvard University, ERSA’s Conference on
Slavery, Colonial History and the New Economic History of Southern Africa, ASSA Annual
Meetings, New Technologies and Interdisciplinary Research in Religion Conference,
African Poverty over the Longue Durée Conference, Mellon Foundation Sawyer Seminar
on Ethnicity in Africa Conference, and Before and Beyond Europe: Economics Change in
Historical Perspective Conference.
Department of Economics, Harvard University, 1805 Cambridge Street, Cambridge, MA
02138, U.S.A. (e-mail: [email protected]; Web site: http://www.economics.harvard
.edu/faculty/nunn). Also ailiated with the NBER and BREAD.
1
Also see Barro and McCleary (2003), as well as the review article by McCleary and Barro
(2006).

489
490 Nunn

of Africa: Showing the Tribes and Languages; also the Stations of Missionary
Societies” published by William Roome (1924). his information, com-
bined with data on the locations of ethnic groups from Murdock (1959),
is used to calculate estimates of the exposure of African ethnic groups to
missionary activity.
I then link this information to the 2005 Afrobarometer survey using
reported information about the ethnicity of each respondent. With this,
I am able to test whether survey respondents with ancestors living closer
to missions during the colonial period are today more educated. I ind a
positive efect of both Protestant and Catholic missions on the educational
attainment of descendants, although the efect of Protestant missions tends
to be stronger and more robust than the efect of Catholic missions. When
I examine the efects by gender, an interesting pattern emerges. While the
impact of Protestant missions on education is strongest for females, the
impact of Catholic missions is strongest for males. his is consistent with
the Protestant belief that both men and women had to read the Bible to
go to heaven. It also supports previous indings that the Protestant reli-
gion had a particularly important impact on female education (Becker
and Woessmann 2008). he inding is also consistent with the arguments
of Woodberry and Shah (2004) and Woodberry (2009), who assert that
because Protestant missionary activity was open to educating minorities
and women, it had a particularly positive efect for these groups.

16.2 HISTORICAL BACKGROUND

A number of factors determined the locations chosen for mission stations.
he three most important appear to have been access to a clean water sup-
ply, a high altitude with as temperate a climate as possible, and the ability
to establish an external trade route with Europe to import needed supplies
(Johnson 1967). Access to water was crucial for the European missionaries,
high altitude and temperate climate reduced the likelihood of disease, and
access to supplies from Europe were necessary for the establishment and
continued functioning of the mission stations.
Other factors like population density also mattered, but the general efect
is ambiguous. Some missionaries and societies intentionally built missions
in more remote locations, where the word of God would not have reached
otherwise, while others recognized the beneits and eiciencies associated
with dense populations, and targeted these groups.
Another potential determinant of the location of missionary activ-
ity is based on path dependence. he routes taken by the irst missionary
 Gender and Missionary Inluence in Colonial Africa 491

explorers determined which areas of Africa became relatively more well

known to Europe, and this may have afected the locations of subsequent
missions.
In the analysis, I control for these determinants. If these factors had a
direct long-term efect on the outcomes of interest, not properly account-
ing for this efect will result in biased estimates of the efects of missionary
activity.
Historical accounts indicate that missionaries had signiicant impacts.
William Brown, writing about missionary eforts in Southern Africa in the
mid-nineteenth century, notes that “the success of the mission was, on the
whole, highly pleasing. he congregations were considerable; numbers of
the natives baptized, many were also admitted as communicants. he inlu-
ence of the mission extended far beyond the stations; it was felt in a great
part of the surrounding country” (Brown 1864, 532).
A large part of this success arose from the provision of education, train-
ing, and health care as incentives for conversion. With increased European
inluence, it became clear that a new era had arrived, and with education
came power and inluence. Robin Horton writes: “With the advent of the
twentieth century . . . Europeans came to be seen as symbols of power, and
Christianity itself came to be seen as part of a larger order, comprising
Western education, colonial administration, commerce and industry, with
which everyone had henceforth to reckon. hese changes created a much
more favourable climate for conversion” (Horton 1971, 86).
William Brown, the secretary of the Scottish Missionary Society, described
the beneits to conversion experienced by the Bechuana of southern Africa:
“he people under the care of the missionaries made considerable advances
in some of the more common and necessary arts of civilized life. Many of
them built themselves convenient houses, some of them of stone, instead
of their old smoky unhealthy huts. In place of the skins of animals which
they used to throw over their bodies, the men adopted in part the European
dress, while the women who had learned to sew made decent clothes for
themselves and their daughters. hough they were previously not simply a
pastoral people, but cultivated millet and other produce, yet now their hus-
bandry was considerably extended. hey obtained ploughs and other agri-
cultural implements, and many of them occupied themselves in the culture
of corn, which they sold to the Dutch farmers for cattle, clothing, soap, salt,
and other useful articles” (Brown 1864, 534).
Education was so central to the conversion strategy of missionaries that
education was almost exclusively provided by missions during the colonial
period. In the 1940s, 97 percent of the student population in Ghana and
492 Nunn

Nigeria were from missionary schools. In South Africa during this period,
there were 5,360 mission-sponsored schools and only 230 state-sponsored
schools (Berman 1974, 527).

16.3 DATA AND THEIR SOURCES

he data used in the analysis are built around the third round (2005) of
the Afrobarometer surveys. hese nationally standardized questionnaires
measure a host of characteristics among a random sample of either twelve
hundred or twenty-four hundred individuals in each country. he ques-
tionnaire provides a variety of information, including individuals’ educa-
tion, religion, age, gender, and their views about female equality.
he 2005 Afrobarometer covers the following seventeen sub-Saharan
African countries: Benin, Botswana, Ghana, Kenya, Lesotho, Madagascar,
Malawi, Mali, Mozambique, Namibia, Nigeria, Senegal, South Africa,
Tanzania, Uganda, Zambia, and Zimbabwe. It is clear that this is not a ran-
dom sample. he countries tend to be concentrated in West Africa, eastern
Africa, and southern Africa. All of west central Africa is not included, as
well as countries inland of the Red Sea. herefore, it is important to keep in
mind that all of the results in this chapter apply only to the seventeen sub-
Saharan African countries included in our sample. he relationships that I
identify in the data may be very diferent within the African countries not
included in our sample.
Information on the historical inluence of Christianity within Africa is
taken from a map created by Roome (1924). he map provides the exact
location of all Protestant and Catholic missions in Africa in 1924. A foreign
mission is deined as a location foreign missionaries established with the
intention to convert, educate, and/or assist in other ways the foreign indig-
enous populations. herefore, churches foreigners developed for their own
populations are not included in the data, as are churches meant to serve
the immigrant (e.g., white) populations living in Africa. he analysis only
includes individuals belonging to ethnic groups indigenous to Africa.
he map Roome produced was highly regarded by experts in the area,
who conirmed that the locations of the mission stations are accurate.2 he
only criticisms of the map involved disagreements about the exact spellings
of the tribes’ names and of their locations on the maps. For my analysis here,
I do not rely on the spellings or locations of ethnic groups from Roome’s
map. Roome’s map is shown in Figure 16.1. Superimposed on the map are

2
See, for example, E. W. S. (1925).
 Gender and Missionary Inluence in Colonial Africa 493

Figure 16.1. Map displaying the location of Catholic missions (triangles) and Protestant
missions (circles) in Africa in 1924.

circles showing the location of each Protestant mission and triangles show-
ing the locations of Catholic missions.
Although there is a clear clustering of Protestant and Catholic missions,
there is also a fair amount of mixing. Many locations contain missions of
both types, closely mixed together. his fact can be explained in part by the
General Act of the Conference of Berlin, which guaranteed freedom of reli-
gion throughout Africa.3
I undertake two strategies to trace the impacts of missions on individu-
als today. he irst is to use an individual’s ethnic identity to construct a

3
Article 6 of Chapter I mandated that each of the European powers “shall, without distinc-
tion of creed or nation, protect and favour all religious, scientiic, or charitable institutions,
and undertakings created and organized for the above ends, or which aim at instructing
the natives and bringing home to them the blessings of civilization. Christian missionar-
ies, scientists, and explorers, with their followers, property, and collections, shall likewise
be the objects of especial protection” (Keith 1919, 304).
494 Nunn

Figure 16.2. Constructing ethnicity-level mission measures.

measure of historical exposure to missionaries. he second is to use an indi-

vidual’s current location to construct a measure of past missionary activity.
In the irst strategy, I examine the reported ethnicities from the
Afrobarometer survey and calculate the intensity of missionary activ-
ity in the 1920s within the area inhabited by the ethnic group at the time.
Information on the location of ethnicities during the colonial period is
taken from a map from Murdock (1959) showing the approximate bor-
ders of ethnic groups in early nineteenth-century Africa. Superimposing
Murdock’s information over Roome’s map, I construct a measure of the
density of Catholic and Protestant mission in the area traditionally inhab-
ited by each ethnic group.
his procedure is illustrated by Figure 16.2. he igure shows the locations
of Catholic and Protestant mission stations in eastern Africa. It also shows,
as black polygons, the boundaries of ethnic groups (in the nineteenth cen-
tury) as mapped by Murdock (1959). For each ethnic group I calculate the
number of missions (total, Protestant, or Catholic) per one thousand kilo-
meters of land area. his variable is my measure of the exposure a respon-
dent’s ethnic group had to religion in the early twentieth century.
 Gender and Missionary Inluence in Colonial Africa 495

Figure 16.3. Constructing village-level mission measures.

In the second strategy, I also construct a geography-based measure of the

historical intensity of missionary activity. Unlike the irst variable, which
is measured at the ethnicity level, this is measured at the village level.4 he
construction of the variable is illustrated in Figure 16.3, which shows the
location of Catholic and Protestant mission stations, as well as a number of
circles, indicating 25 km radii around the locations of each individual from
the Afrobarometer surveys.5
he second measure calculates the number of missions (total, Catholic,
and Protestant) that lie within each circle. herefore, the variable measures
the number of missions that the village where the individual is currently
living was exposed to in 1924.
In the analysis, I control for a number of variables. hese include the
existence of a colonial railway line, and of early explorer contact, both

4
For individuals living in larger cities, the variable is actually measured at the neighbor-
hood level.
5
Twenty-ive kilometers is chosen arbitrarily. he results are similar using larger radii, such
as 50 or 100 kilometers.
496 Nunn

Figure 16.4. Explorer routes.

measured at the ethnicity and village levels. Data on the location of rail lines
(as of 1897) and explorer routes (prior to 1895) are from a map created by
he Century Company (1897). he original map is shown in Figures 16.4
and 16.5. In Figure 16.4, the expedition routes are superimposed over the
map. Figure 16.5 illustrates the colonial railway lines.
he procedure used for the missions is also used to construct analogous
measures for the control variables. For each village and ethnic group, I con-
struct an indicator variable that equals one if a railway contacted the town
or group, and zero otherwise. I also construct ethnicity- and village-level
indicator variables that equal one if a European explorer contacted the eth-
nic group or town.
he railway variable is meant to capture ease of access that villages or
ethnic groups had to supplies from Europe, an important factor afecting
the choice of mission stations. he explorer variable captures the proxim-
ity of villages and ethnic groups to the routes of early missionaries and
other explorers. his may have afected European knowledge about an
area, which may have afected the likelihood of missions being developed
there.
he set of controls also includes three geographic measures: the suit-
ability of climate and land for the cultivation of crops; the amount of land
 Gender and Missionary Inluence in Colonial Africa 497

Figure 16.5. Colonial railway lines.

within ten kilometers of a fresh water source; and average altitude. Again,
the three controls are measured at both the ethnicity and village levels.
Data on suitability for agriculture are from the FAO’s GAEZ 2002 data-
base. he FAO provides a suitability index for ive arc minute by ive arc
minute (approx. 56 km by 56 km) grid cells globally. Data on the location of
fresh water sources and elevation are from Global Mapping International’s
(GMI) Seamless Digit Chart of the World Version 3.2.
he inal control variable included is a measure of the intensity of the
slave trade; given that the goal of many missionaries, particularly Protestant
missionaries, was to end slavery, missions may have been intentionally
located in areas with a history of slavery (Johnson 1967). he measures of
the intensity of the Atlantic and Indian Ocean slave trade are taken from
Nunn and Wantchekon (2011). he study provides both ethnicity- and
village-level measures. he ethnicity-level variable measures the number
of slaves taken from the ethnic group, normalized by the size of the ethnic
group, which is measured by the amount of land it inhabited during the early
twentieth century. To construct a village-level measure, the authors use the
location of each village as well as Murdock’s map to determine which ethnic
group traditionally inhabited the location of the village. he village is then
given the slave export measure of the ethnic group, thereby providing an
estimate of the intensity of the slave trade in that location historically.
498 Nunn

16.4 EMPIRICAL RESULTS: MISSIONARIES AND EDUCATION

I begin by examining the relationship between historical missionary activ-
ity and educational attainment today. I irst examine the average efect of all
Christian mission stations (both Protestant and Catholic) on educational
attainment. he estimating equation is:

Yi ,e ,v ,c α c βM e + γ M v Xi Γ + X e Λ X v Φ + ε i ,e ,v ,c (16.1)

where i indexes individuals, e ethnic groups, v villages, and c countries.

he dependent variable measures individual i’s educational attainment (in
years). he variable Me measures the exposure of ethnic group e to Protestant
and Catholic missions in 1924. It is measured as the log number of missions
per one thousand kilometers of land area historically inhabited by the eth-
nic group. I take the natural log to remove the skew that exists in the dis-
tribution otherwise. Mv is the analogous measure of missionary activity,
but measured at the village level. It is measured by the number of missions
(Catholic and Protestant) in the village of the respondent. More precisely,
the measure is calculated by identifying the number of missions within a
twenty-ive kilometer radius of the geographic location of the respondent.
α c denotes country ixed efects, which are included in all regressions.
hese capture country-level characteristics, including colonial policies,
which may be correlated with average education and missionary activity.
Equation (16.1) also includes a host of control variables that vary at either
the individual, ethnicity, or village levels. Xi denotes a vector of individual-
level control variables that include a gender indicator variable, age, and an
indicator variable that equals one if he or she is living in an urban location.
Xe denotes a vector of ethnicity-level control variables, including an indi-
cator that equals one if the ethnic group was contacted by early European
explorers, an indicator that equals one if a railway was built on the land
inhabited by the ethnic group, the proportion of the ethnic group’s histor-
ical land that is arable, the proportion that was within ten kilometers of a
fresh water source, and the normalized number of slaves exported in the
Atlantic and Indian Ocean slave trades. Xv includes the same covariates, but
measured at the village level, rather than the ethnicity level.
OLS estimates of Equation (16.1) are reported in Table 16.1. Column (1)
reports estimates with the ethnicity-level mission variable and ethnicity-
level controls. Column (2) includes the village-level mission variable and
village-level controls, while column (3) includes the village-level and ethnic-
ity-level variables together. All three speciications also include individual-
 Gender and Missionary Inluence in Colonial Africa 499

Table 16.1. OLS estimates of the relationship between colonial missionary activity
and educational attainment today

Dependent variable: years of education

(1) (2) (3)

Mission stations among ethnic group, Me 0.102** 0.098**
(0.046) (0.044)
Mission stations in village, Mv 0.143*** 0.119***
(0.033) (0.042)
Individual-level controls Yes Yes Yes
Ethnicity-level controls Yes No Yes
Village-level controls No Yes Yes
Country ixed efects Yes Yes Yes
Number of observations 20,914 20,914 20,914
Number of clusters 185 2,693 185 / 2,693
R-squared 0.337 0.336 0.342
Notes: he table reports OLS estimates. he unit of observation is an individual. he individual
control variables are for age, age squared, a gender indicator variable, and an indicator variable
for whether the respondent lives in an urban location. he ethnicity-level controls include the log
number of slaves exported per land area during the Atlantic and Indian Ocean slave trades, an
indicator variable equal to one if the ethnicity was contacted by a European explorer prior to colo-
nial rule (1885), an indicator variable equal to one if a railway line dissected the region in which
the ethnicity was living during colonial rule, the proportion of the ethnic group’s land that is suit-
able for agriculture, the proportion of land that is within ten kilometers of a fresh water source,
and average elevation. he village-level controls include an indicator variable equal to one if the
inhabitant’s current village was contacted by a European explorer prior to colonial rule (1885),
an indicator variable equal to one if a railway line dissected the village, the log number of slaves
exported per land area during the Atlantic and Indian Ocean slave trades among the ethnic group
that historically inhabited the village, the fraction of the land within the village (twenty-ive kilo-
meter radius of the centroid) that is suitable for cultivation, the fraction of the village land that is
within ten kilometers of a fresh water source, and the average elevation of the village. Coeicients
are reported, with clustered standard errors in brackets. ***, ** and * indicates signiicance at the
1 percent, 5 percent, and 10 percent levels.

level controls and country ixed efects. he estimates show a positive and
statistically signiicant efect of missionary activity, measured either at the
ethnicity or village level, on educational attainment.
his inding provides added support for the arguments and cross-country
evidence of Woodberry (2004) that missionary activity (his analysis focuses
on Protestants) had long-term impacts on the level of education.
hese efects could occur through a variety of channels. First, missionar-
ies may have altered people’s views about the importance of education. If
individuals transmitted their new attitudes and beliefs to their children,
500 Nunn

then the descendants of those in contact with Protestant missions would

also value education more and therefore be better educated today. his
explanation is consistent with the positive estimates for the ethnicity-level
missionary variable. he positive and signiicant coeicients in columns (1)
and (3) show that ethnicities that had closer contact to missionaries have
descendants who are more educated today.
An alternative mechanism is because missionaries valued education,
they immediately established schools and universities. In fact, in British
colonies, most of the schools in the colonies were established by missionar-
ies rather than by the colonial administration (e.g., Berman 1974, 527).
hese investments in educational infrastructure tend to persist over time
(e.g., Huillery 2009), so that today, towns and villages that historically had
more missionary activity may have more schools, and therefore the equi-
librium level of education is higher. his channel, which operates through
location and not ethnicity, is captured by the village-level mission variable,
which is positive and signiicant in columns (2) and (3). Villages that had
more missions during the colonial period have more educated inhabitants
today.
Given that the two sets of mission variables – measured at the village
and ethnicity levels – capture diferent transmission mechanisms, a natural
question is which is more important. Because the variables are measured
in diferent units, a direct comparison of their coeicients is inappropri-
ate. Instead, I compare standardized beta coeicients, which are unit-less
coeicients that report the number of standard deviation increases in the
dependent variable per one-standard-deviation increase in the indepen-
dent variable. his comparison suggests that the efect of the ethnicity-level
measure is approximately three times the magnitude of the efect of the
village-level measure. According to the estimates of column (3), the beta
coeicient for the ethnicity-based measure is 0.047 and for the village-level
measure is 0.036. herefore, the evidence suggests that both channels are
important and of a similar magnitude.

16.5 EMPIRICAL RESULTS: MISSIONARIES AND

EDUCATION, DISTINGUISHING BETWEEN CATHOLICS
AND PROTESTANTS
Much of the preexisting qualitative and quantitative literature distinguishes
between the impact of Protestant and Catholic missions, documenting
important diferences between the two (Gallego and Woodberry 2010;
Woodberry 2004; Woodberry and Shah 2004). he most detailed evidence
 Gender and Missionary Inluence in Colonial Africa 501

in the African context is from Gallego and Woodberry (2010), who pro-
vide evidence that Protestant missions, relative to Catholic missions, had
a greater impact on increased education. hey also show that this is most
pronounced in Catholic nation-states where there was less competition
from Protestant missions.
In this section, I examine whether there is evidence of diferential impacts
of Protestant and Catholic missions on religious conversion. I construct
variables that measure the intensity of Catholic missions and of Protestant
missions, at both the ethnicity and the village levels. hese are then included
in the estimating equation, rather than the measures of missionary activ-
ity generally. Doing this allows the impact of missionary activity to difer
depending on whether the mission was Catholic or Protestant. he new
estimating equation is:

Yi ,e ,v ,c αc β p eprot + γ p Mv t βc MeCath + γ c MvCath

+ Xi Γ X e Λ + X v Φ + ε i ,e ,v ,c (16.2)

where i continues to index individuals, e ethnic groups, v villages, and c

countries. he variables MeProt and MeCath are measures of the intensity of
exposure of ethnic group e to Protestant and Catholic missions in 1924.
hey are measured in the same units as Me. Similarly, MvProt and MvCath
measure the intensity of exposure of village v to Protestant and Catholic
missionaries and are given in the same units as Mv. Xe, Xv and Xi are the
same vectors of covariates as in Equation (16.1).
Estimates of Equation (16.2) are reported in Table 16.2. Column (1)
reports estimates with missionary activity measured at the ethnicity level,
while column (2) measures missionary activity at the village level. he esti-
mates show a positive and statistically signiicant impact of Protestant mis-
sions on educational attainment today. he estimates for Catholic missions
are also positive, but smaller in magnitude and statistically insigniicant.
he results are consistent with Gallego and Woodberry’s (2010) inding that
Protestant missionaries had a larger positive impact on long-term educa-
tion than Catholic missionaries. In column (3), I include both the village-
and ethnicity-level variables together. All coeicients remain positive and
approximately the same magnitude as in columns (1) and (2). he coef-
icients for Protestant missions remain statistically signiicant, while the
coeicients for Catholic missions remain insigniicant.
In the previous section, we distinguished between location- and ethnic-
ity-based channels of causality. In this section, we undertake an additional
exercise to help distinguish between mechanisms. It is possible that the
 Table 16.2. OLS estimates of the relationship between colonial missionary activity and education today, allowing for diferences between
Catholic and Protestant missions

Dependent variable: years of education

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Ethnicity-based
measure:
Catholic Missions 0.050 0.036 0.046 0.033
(0.034) (0.031) (0.032) (0.029)
Protestant Missions 0.094* 0.096** 0.068* 0.073*
(0.049) (0.045) (0.040) (0.037)
502

Location-based
measure:
Catholic Missions 0.106 0.124 0.131* 0.141*
(0.072) (0.082) (0.069) (0.080)
Protestant Missions 0.163*** 0.104** 0.105** 0.067
(0.049) (0.048) (0.046) (0.046)
Current Religion
Indicator:
Catholic 1.852*** 1.884*** 1.810*** 1.823*** 1.859*** 1.764***
(0.162) (0.099) (0.159) (0.154) (0.098) (0.150)
Protestant 1.911*** 1.962*** 1.870*** 1.876*** 1.947*** 1.828***
(0.161) (0.100) (0.162) (0.149) (0.098) (0.147)
Other Christian 1.665*** 1.681*** 1.621*** 1.652*** 1.662*** 1.597***
(0.161) (0.102) (0.161) (0.151) (0.101) (0.149)
 Country ixed efects Yes Yes Yes Yes Yes Yes Yes Yes Yes
Ethnicity-level controls Yes No Yes Yes No Yes Yes No Yes
Village-level controls No Yes Yes No Yes Yes No Yes Yes
Number of observations 20,914 20,914 20,914 20,836 20,836 20,836 20,836 20,836 20,836
Number of clusters 185 2,692 185/2,692 185 2,692 185/2,692 185 2,692 185/2,692
R-squared 0.337 0.336 0.343 0.355 0.356 0.359 0.356 0.357 0.361
Notes: he table reports OLS estimates. he unit of observation is an individual. he individual control variables are for age, age squared, a gender indicator
variable, and an indicator variable for whether the respondent lives in an urban location. he ethnicity-level controls include the log number of slaves exported
per land area during the Atlantic and Indian Ocean slave trades, an indicator variable equal to one if the ethnicity was contacted by a European explorer prior to
colonial rule (1885), an indicator variable equal to one if a railway line dissected the region in which the ethnicity was living during colonial rule, the proportion
of the ethnic group’s land that is suitable for agriculture, the proportion of land that is within ten kilometers of a fresh water source, and average elevation. he vil-
lage level controls include an indicator variable equal to one if the inhabitant’s current village was contacted by a European explorer prior to colonial rule (1885),
an indicator variable equal to one if a railway line dissected the village, the log number of slaves exported per land area during the Atlantic and Indian Ocean slave
trades among the ethnic group that historically inhabited the village, the fraction of the land within the village (twenty-ive kilometer radius of the centroid) that
503

is suitable for cultivation, the fraction of the village land that is within ten kilometers of a fresh water source, and the average elevation of the village. Coeicients
are reported, with clustered standard errors in brackets. ***, ** and * indicates signiicance at the 1 percent, 5 percent, and 10 percent levels.
504 Nunn

estimated long-term impacts of missions arise because of changes in the

perception of the value of education, which have been transmitted through
time. Alternatively, if, as documented by Nunn (2010), colonial missions
resulted in conversion, even in the long run, then this may cause increased
education. According to the second explanation, the persistence is through
the transmission of religious values and not through the transmission of
beliefs about the importance of education.
I attempt to distinguish between these two hypotheses in columns (4)–(9)
of Table 16.2. Columns (4)–(6) report estimates of the three speciications
from columns (1)–(3), but with the missionary variables replaced with vari-
ables measuring the religion of the respondent. I include three indicator
variables that equal one if the respondent is: (i) Catholic, (ii) Protestant, or
(iii) another type of Christian. As shown, all three indicator variables are
highly correlated with educational attainment. hose belonging to any of
the three religions, on average, have 1.6–2.0 years more education.
In columns (7)–(8), I reestimate the speciications from columns (1)–
(3), but now include the three religion controls. If the importance of colo-
nial missionary activities, estimated in columns (1)–(3), arises because of
the current religious ailiation of the individual, then we would expect
the estimated impact of historical missionary activity to be much smaller
once we account for the current religion of respondents. Comparing the
estimates in columns (1)–(3) to those in columns (7)–(8), I ind a small
drop in the point estimates of the Protestant mission variables ater the
inclusion of the religion controls. he decrease ranges from 24–35 per-
cent. his suggests that a sizable part of the long-term impact of Protestant
missions on education may arise due to the persistence of religious val-
ues. Interesting, controlling for religion does not attenuate the estimated
impacts of Catholic missions. he point estimates tend to stay constant or
even increase ater the inclusion of the religion controls. his suggests that
the estimated impact of Catholic missions on long-term education is not
due to the persistence of a belief in Catholicism. his is consistent with the
fact that the Catholic religion places much less importance on education
and literacy than Protestantism.

16.6 EMPIRICAL RESULTS: MISSIONARIES, EDUCATION,

AND GENDER
A number of existing studies examine how the Protestant religion may be
responsible for reducing the gender gap in education by emphasizing the
 Gender and Missionary Inluence in Colonial Africa 505

importance of educating women. In Europe, Martin Luther urged the edu-

cation of women so that they too would be able to read the Bible. Becker
and Woessmann (2008) empirically test for this efect of the Protestant
religion. Examining iteen village-level data from the Prussian Census of
1816, they identify a negative relationship between the prevalence of the
Protestant religion and the educational gender gap, measured as average
male education minus average female education.
Motivated by Becker and Woessmann’s inding, I examine whether,
within the African context, the efect of colonial missions on current edu-
cation is diferent for males relative to females. his is done by estimating
a speciication that allows the efect of missions on current education to
difer, depending on the gender of the respondent. In practice, this is done
by interacting the baseline mission variables with an indicator variable that
equals one if the respondent is female, Iifemale . he updated estimating equa-
tion is given by:

Yi ,e ,v ,c α c β p Meprot + θ p Me t Iifemale
f l
+ βc MeC th θc MeCath ⋅ Iifemale
(16.3)
+ Xi Γ X e Λ + X v Φ + ε i ,e ,v ,c

Estimates of Equation (16.3) are reported in columns (1) and (2) of

Table 16.3. Column (1) reports estimates using the ethnicity-level Catholic
and Protestant mission variables and column (2) reports estimates using the
village-level variables. he indings show that for Protestant missions, the
coeicient for the mission variable itself is positive, but not statistically dif-
ferent from zero, while the mission variable interacted with the female indi-
cator variable is positive and highly signiicant. his suggests that historical
missionary activity is associated with no long-term impact on the educa-
tion of men, but a positive and statistically signiicant long-term impact on
female education. As reported in columns (3) and (4), the results are very
similar even when controlling for the current religion of the respondent.
his suggests that the long-term education efect primarily works through
mechanisms other than the current religion of the respondent. Protestant
missions may have created a persistent belief in the importance of female
education.
For Catholic missions, the baseline mission variables in columns (1) and
(2) are both positive and statistically signiicant, while the coeicients for
the mission variables interacted with the female indicator variable are nega-
tive and statistically signiicant. he sum of the two coeicients, which gives
the total efect of Catholic missions on the long-run educational attainment
of women, is very close to zero. Again, the results are similar when religion
 506 Nunn

Table 16.3. OLS estimates of the relationship between missionaries and education,
allowing for diferential efects by gender

Dependent variable: Years of education

(1) (2) (3) (4)

Ethnicity-based measure:
Catholic missions 0.102** 0.095**
(0.042) (0.040)
Catholic missions x Female −0.108* −0.102*
(0.057) (0.055)
Protestant missions 0.016 −0.003
(0.053) (0.046)
Protestant missions x Female 0.158*** 0.145***
(0.052) (0.050)
F-test for Catholic females = 0 (p-value) 0.902 0.885
F-test for Protestant females = 0 (p-value) 0.003 0.004
Location-based measure:
Catholic missions 0.228*** 0.248***
(0.088) (0.086)
Catholic missions x Female −0.244*** −0.234***
(0.087) (0.086)
Protestant missions 0.086 0.022
(0.057) (0.055)
Protestant missions x Female 0.154*** 0.165***
(0.050) (0.050)
F-test for Catholic females = 0 (p-value) 0.841 0.852
F-test for Protestant females = 0 (p-value) 0.000 0.000
Religion indicator variables No No Yes Yes
Individual-level controls Yes Yes Yes Yes
Ethnicity-level controls Yes Yes No No
Village-level controls No No Yes Yes
Country ixed efects Yes Yes Yes Yes
Number of observations 20,914 20,914 20,836 20,836
Number of clusters 185 185 2,692 2,692
R-squared 0.338 0.337 0.357 0.358
Notes: he table reports OLS estimates. he unit of observation is an individual. he individual
control variables are for age, age squared, a gender indicator variable, and an indicator variable for
whether the respondent lives in an urban location. he ethnicity-level controls include the log num-
ber of slaves exported per land area during the Atlantic and Indian Ocean slave trades, an indicator
variable equal to one if the ethnicity was contacted by a European explorer prior to colonial rule
(1885), an indicator variable equal to one if a railway line dissected the region in which the eth-
nicity was living during colonial rule, the proportion of the ethnic group’s land that is suitable for
agriculture, the proportion of land that is within ten kilometers of a fresh water source, and average
elevation. he village-level controls include an indicator variable equal to one if the inhabitant’s
current village was contacted by a European explorer prior to colonial rule (1885), an indicator var-
iable equal to one if a railway line dissected the village, the log number of slaves exported per land
 Gender and Missionary Inluence in Colonial Africa 507

is controlled for (reported columns (3) and (4)). hese indings, which are
in stark contrast to the indings for Protestants, suggest that a history of
Catholic missionary activity increased education in the long run, but only
for males.
he indings of Table 16.3 raise an interesting possibility. It may be that
the historical presence of Protestant missionaries also altered not only indi-
viduals’ beliefs about the importance of female education, but also of the
role of women in society more generally. hat is, the Protestant religion
may have resulted in a stronger belief about the equality of men and women
more generally.
To explore this possibility, I consider two questions about the perceived
role of women relative to men from the Afrobarometer survey. he irst
question asks the respondent whether they agree with the following two
statements: (A) In our country, women should have equal rights and
receive the same treatment as men, (B) Women have always been subject
to traditional laws and customs and should remain so. he respondent
can choose to “agree very strongly” with (A), “agree strongly” with (A),
“agree strongly” with (B), or “agree very strongly” (B). he second ques-
tion asks the respondent their view on the following two statements: (A)
Women should have the same chance of being elected to political oice as
men, (B) Men make better political leaders than women, and should be
elected rather than women. Again, the respondents were given the same
choices.
Using the responses, I code a variable that takes on the value 1, 2, 3, or
4, and is increasing in their agreement with statement (A) over statement
(B). In other words, the variable is increasing in the respondent’s belief that
women and men should have equal rights.
I then estimate variants of Equation (16.2) with the two gender beliefs vari-
ables as dependent variables. he estimates, which are reported in Table 16.4,
fail to provide consistent evidence that historical Protestant missionary
activity promoted attitudes relecting gender equality. For six of the eight
Protestant variables, the coeicient is not statistically diferent from zero. As

Notes: (continued)

area during the Atlantic and Indian Ocean slave trades among the ethnic group that historically
inhabited the village, the fraction of the land within the village (twenty-ive kilometer radius of the
centroid) that is suitable for cultivation, the fraction of the village land that is within ten kilometers
of a fresh water source, and average elevation of the village. Coeicients are reported, with clus-
tered standard errors in brackets. ***, ** and * indicates signiicance at the 1 percent, 5 percent, and
10 percent levels.
508 Nunn

Table 16.4. OLS estimates of the relationship between colonial missionary activity
and gender role attitudes today

Dep var: belief in females Dep var: belief in females as

equality generally equal leaders

(1) (2) (3) (4) (5) (6)

Ethnicity-based measure:
Catholic Missions −0.010 −0.009 0.014** 0.015**
(0.009) (0.009) (0.007) (0.007)
Protestant Missions 0.005 0.004 −0.004 −0.006
(0.009) (0.009) (0.008) (0.008)
Location-based measure:
Catholic Missions −0.025* −0.018 −0.028** −0.032*
(0.015) (0.017) (0.013) (0.017)
Protestant Missions 0.024** 0.016 0.023** 0.014
(0.011) (0.011) (0.010) (0.012)
Country ixed efects Yes Yes Yes Yes Yes Yes
Ethnicity-level controls Yes No Yes Yes No Yes
Village-level controls No Yes Yes No Yes Yes
Number of observations 19,533 19,533 19,533 20,603 20,603 20,603
Number of clusters 185 2,691 185/2,691 185 2,691 185/2,691
R-squared 0.108 0.103 0.109 0.106 0.099 0.107
Notes: he table reports OLS estimates. he unit of observation is an individual. he individual
control variables are for age, age squared, a gender indicator variable, and an indicator variable
for whether the respondent lives in an urban location. he ethnicity-level controls include the log
number of slaves exported per land area during the Atlantic and Indian Ocean slave trades, an
indicator variable equal to one if the ethnicity was contacted by a European explorer prior to colo-
nial rule (1885), an indicator variable equal to one if a railway line dissected the region in which
the ethnicity was living during colonial rule, the proportion of the ethnic group’s land that is suit-
able for agriculture, the proportion of land that is within ten kilometers of a fresh water source,
and average elevation. he village level controls include an indicator variable equal to one if the
inhabitant’s current village was contacted by a European explorer prior to colonial rule (1885),
an indicator variable equal to one if a railway line dissected the village, the log number of slaves
exported per land area during the Atlantic and Indian Ocean slave trades among the ethnic group
that historically inhabited the village, the fraction of the land within the village (twenty-ive kilo-
meter radius of the centroid) that is suitable for cultivation, the fraction of the village land that is
within ten kilometers of a fresh water source, and average elevation of the village. Coeicients are
reported, with clustered standard errors in brackets. ***, ** and * indicates signiicance at the 1 per-
cent, 5 percent, and 10 percent levels.

well, the coeicients for the Catholic mission variables are not consistently
positive or negative and many are not statistically diferent from zero.6

6
It is possible that only the attitudes of males were afected by missionary activity, and this
is the reason for the weak results. However, I have checked for this by restricting the sam-
ple to include males only and continue to ind similar results (not reported in the table).
 Gender and Missionary Inluence in Colonial Africa 509

herefore, although the evidence is consistent with Protestant mission-

ary activity afecting long-term female educational attainment, I do not
ind evidence of an efect on attitudes about gender equality. his inding is
not surprising given the nature of missionary education. Although women
were educated in reading, writing, and even arithmetic, the greatest empha-
sis was typically placed on teaching women practical domestic skills, like
laundry, housework, and cooking, rather than providing them with the
highest levels of education. It was believed that African women should be
taught how to prepare a suitable Christian home for their husbands who
were to be more highly educated and serve in positions of responsibility
and authority.7

16.7 CONCLUSIONS
Combining information on the locations of Catholic and Protestant
missions in colonial Africa, I have examined the long-term impacts of
Protestant and Catholic missionary activity during the colonial period
on educational attainment. Using variation across villages and ethnicities
from seventeen sub-Saharan African countries, I have provided evidence
that Protestant missionary activity had a positive long-term impact on
educational attainment. Catholic missions are estimated to have a quanti-
tatively much smaller impact on education that is not statistically difer-
ent from zero. he inding of a much larger impact of Protestant missions
relative to Catholic missions on education is consistent with the previous
indings of Gallego and Woodberry (2010). Interestingly, these indings,
which are an average across the full Afrobarometer sample, mask stark
diferences of the impacts by gender. he positive impact of Protestant
missions is concentrated almost solely among females. Protestant mis-
sions are estimated to have a small positive and insigniicant efect on
male education, and a large signiicant positive efect on female education.
he positive but insigniicant impact of Catholic missions on education
is composed of a large and signiicant positive impact on male education,
but a precisely estimated zero efect on female education. hese indings
are consistent with the greater importance placed on the education of
women by Protestants relative to Catholics.

7
See Labode (1993) for a discussion of this aspect of missionary education within southern
Africa and Kanogo (1993) for a description for Kenya.
510 Nunn

DATA APPENDIX
Contemporaneous Individual-Level Data
All information at the individual level, such as age, gender, educa-
tion, religion, occupation, and so forth are from the third round of the
Afrobarometer surveys, conducted in 2005. he data are publicly available
and can be downloaded at: www.afrobarometer.org. he Afrobarometer is
an independent and nonpartisan research project conducted by the Center
for Democratic Development (CDD), Institute for Democracy in South
Africa (IDASA), and Michigan State University (MSU).

Data on Missionary Activity

Data on the location of Catholic and Protestant mission stations in the
1920s are from Roome (1924). he mission location data are combined with
information about the boundaries of Africa’s ethnic groups in the late nine-
teenth century from Murdock (1959) to calculate the number of Catholic
and Protestant mission stations in the area inhabited by each ethnic group.
his is the ethnicity-based measure of missionary activity. he informa-
tion is also combined with information about the location of each respon-
dent in the Afrobarometer surveys to construct a measure of the number
of Catholic and Protestant mission stations within a twenty-ive kilometer
radius of the location where the respondent is currently living. his is the
village-based measure of missionary activity.

Ethnicity- and Village-Level Control Variables

he information of the exploration routes of early European explorers is
from he Century Company (1897). he map includes the routes of twenty-
seven expeditions occurring before 1895. Information of the location of
colonial rail lines is also from the same map.
Data on the suitability of a location for agricultural cultivation are from
the FAO’s GAEZ 2002 database. he FAO provides a suitability index for
ive arc minute by ive arc minute (approx. 56 km by 56 km) grid cells glob-
ally. Data on the location of fresh water sources and on elevation is from
Global Mapping International’s (GMI) Seamless Digit Chart of the World
Version 3.2.
he data used to control for the number of slaves taken from each
ethnicity during the transatlantic and Indian Ocean slave trade are from
 Gender and Missionary Inluence in Colonial Africa 511

Nunn and Wantchekon (2011), and the primary sources are described in
Nunn (2008).

References
Bai, Ying and James Kai-sing Kung. (2011). “Difusing Knowledge While Spreading
God’s Message: Protestantism and Economic Prosperity in China, 1840–1920,”
Mimeo, Hong Kong University of Science and Technology.
Barro, Robert J. and Rachel McCleary. (2003). “Religion and Economic Growth across
Countries.” American Sociological Review, October, 68(5): 760–81.
Becker, Sascha O. and Ludger Woessmann. (2008). “Luther and the Girls: Religious
Denomination and the Female Education Gap in 19th Century Prussia.”
Scandinavian Journal of Economics, 110: 777–805.
(2009). “Was Weber Wrong? A Human Capital heory of Protestant Economic
History.” Quarterly Journal of Economics, May, 124(2): 531–96.
(2010). “he Efect of Protestantism on Education before Industrialization: Evidence
from 1816 Prussia.” Economics Letters, 107(2): 224–8.
Berman, Edward H. (1974). “African Responses to Christian Mission Education.” Africa
Studies Review, December, 17(3): 527–40.
Brown, William (1864). he History of Christian Missions, Volume II. London: homas
Baker.
E. W. S. (1925). “Review of: An Ethnographic Map of Africa.” Journal of the Royal African
Society, April, 24(95): 263.
Gallego, Francisco A. and Robert Woodberry. (2010). “Christian Missionaries and
Education in Former Colonies: How Competition Mattered.” Journal of African
Economies, 19(3): 294–329.
Grier, Robin. (1997). “he Efect of Religion on Economic Development: A Cross-
National Study of 63 Former Colonies.” Kyklos, 50(1): 47–62.
Horton, Robin. (1971). “African Conversion.” Africa: Journal of the International African
Institute, 41(2): 85–108.
Huillery, Elise. (2009). “History Matters: he Long-Term Impact of Colonial Public
Investments in French West Africa,” American Economic Journal: Applied
Economics, 1(2): 176–215.
Johnson, Hildegard Binder. (1967). “he Location of Christian Missions in Africa.”
Geographical Review, 57(2): 168–202.
Kanogo, Tabitha. (1993). “Mission Impact on Women in Colonial Kenya.” In Women
and Missions: Past and Present, edited by Fiona Bowie, Deborah Kirkwood, and
Shirley Ardener, 165–86. Oxford: Berg.
Keith, Arthur Berriedale. (1919). he Belgian Congo and the Berlin Act. Oxford:
Clarendon Press.
Labode, Modupe. (1993). “From Heathen Kraal to Christian Home: Anglican Mission
Education and African Christian Girls, 1850–1900.” In Women and Missions: Past
and Present, edited by Fiona Bowie, Deborah Kirkwood, and Shirley Ardener, 126–
44. Oxford: Berg.
McCleary, Rachel and Robert J. Barro. (2006). “Religion and Economy.” Journal of
Economic Perspectives, 20(2): 49–72.
512 Nunn

Murdock, George. (1959). Africa: Its Peoples and heir Cultural History. New York:
McGraw-Hill Book Company.
Nunn, Nathan. (2008). “he Long-Term Efects of Africa’s Slave Trades.” Quarterly
Journal of Economics, 123(1): 139–76.
(2010). “Religious Conversion in Colonial Africa.” American Economic Review Papers
and Proceedings, 100(2): 147–52.
Nunn, Nathan and Leonard Wantchekon. (2011). “he Slave Trade and the Origins of
Mistrust in Africa.” American Economic Review, 101(7): 3221–52.
Roome, William R. M. (1924). “Ethnographic Survey of Africa: Showing the Tribes and
Languages; Also the Stations of Missionary Societies [map],” 1:5,977,382.
Schaltegger, Christoph A. and Benno Torgler. (2009). “Was Weber Wrong? A Human
Capital heory of Protestant Economic History: A Comment on Becker and
Woessmann,” Mimeo, Queensland University of Technology.
he Century Company. (1897). “he Century Atlas, Africa [map].”
Woodberry, Robert D. (2004). “he Shadow of Empire: Christian Missions, Colonial
Policy, and Democracy in PostColonial Societies,” PhD Dissertation in Sociology,
University of North Carolina at Chapel Hill.
(2009). “Dividing Elites: Religious Liberty, Protestant Competition, and Democracy
in the Global South,” Working Paper 002, Project on Religion and Economic
Change. March.
Woodberry, Robert D. and Timothy S. Shah. (2004). “he Pioneering Protestants.”
Journal of Democracy, 15(2): 47–61.
 Index

A type allele, sickle cell disease, 104–6 exports, cash-crop revolution, 298–300,
AA phenotype, sickle cell disease, 104–6 299t
abolition, 233–34 from farming to commerce in southern half
absolutist authority, impact of trans-Atlantic of Africa, 48–50
slave trade on, 467–76, 473t non-use of plows, 97n1
Abuja declaration of 2005, 101–2 overview, 14–15
Adams, Robert, 447 plantation, 71–74
Adandozan (King of Dahomey), 451, ajele system, 213
454–55, 456 Akinjogbin, Isaac, 447–48
Afonso I (Kongo), 369–72 Aksumite kingdom, 46–47
Afrasian (Afroasiatic) peoples, 34, 39 Akyeampong, Emmamuel, ix
Africa Allen, Robert, 76
Colonial Period, 17–20 amabutho, 402
commerce and agriculture, 14–15 Americas
culture, 11–14 commodity production
global historical perspective, 50–53 labor, 70–71
historical perspective, 5–8 plantation economies, 71–76
institutions, 8–11 rice, 69–70
slave trades, 15–17 sugar, 71
theories regarding economic commodity production for export and
development, 1–5 population in Latin America,
he African Poor (Ilife), 144 1850–1912, 67t
African Steam Ship Co., 236–37 comparative examination to West Africa
Africans -he History of a Continent (Ilife), economic development, 58–69
192–93 overview, 56–58
Afroasiatic (Afrasian) peoples, 34, 39 reversal of fortune, 69–81
Afrobarometer surveys, 492, 495, 507–8, 510 population comparison of Africa to, 140
Agbaje, Salami, 225–27 subsistence production, 64
Agongolo (King of Dahomey), 451, 452–54 Aniwura, Efunsetan, 213, 214–15, 216
Agricultural Elaboration Era, 38–39 Anopheles mosquitoes, 102–4
agriculture Ansa, John Owusu, 349–50
in 1500, by country, 124–28 anthropometrics, 313–14
challenges to development, 98–99 artisans, relationship with textile industry, 281,
commerce and, 14–15 284–86
emergence of, 5–6 AS phenotype, sickle cell disease, 104–6

513
514 Index

Asante Brown, William, 491

balance of commerce and warfare, 404 Buganda
European investment in, 18 balance of commerce and warfare, 404–5
obstacles to adopting wheeled cotton industry, 310
transportation, 349–50 Bulawayo Clothing Factory, 275
Ashanti. See Asante Bunyoro, 404–5
asymmetrical git giving, 156–57 Bureau, René, 160–61
Atlantic Creole culture, in Mbanza Kongo/São Burkina Faso
Salvador, 374–75, 379–87 coping strategies for redistributive pressures,
Atlantic theory, Dahomey, 448 171–72, 182
Austen, Ralph, 397n6 malaria mortality rate, 111
Austin, Gareth, ix, 15, 79, 208 Bushong clan, 435n5
Aztec empire, market economy in, 65–66 business networks, 177, 177n15

badingya concept, Gambia, 157n1 Caesar, J.H., 244

Bai Bureh War (Hut Tax Rebellion), 354 Caldwell, John C., 137
Baikie, William, 235–36 Calico Printers of Manchester, 278
balanced polymorphism (heterozygote Cameroon
advantage), 105 coping strategies for redistributive pressures,
banana cultivation, impact on commerce in 172–73
Congo Basin, 49 repressive measures in response to
Barattolo, Roberto, 279 individual success, 198
Bargash, Sultan of Zanzibar, 352–53 witchcrat, 160–61, 161n5
Barth, Frederik, 143–44 campaigning war, 413
barusura, 404 Capitalism and Slavery (Williams), 233–34
Bates, Robert H., ix Capuchins, Mbanza Kongo, 378–78, 384–85
Batwa peoples, 49 Carribean, plantation agriculture, 71–74
Benin Carr-Saunders, A.M., 136
coping strategies for redistributive pressures, cash-crop revolution
173–74 agricultural exports, 298–300
daily collectors, 173–74, 174n12 institutions and welfare outcomes, 311–15
witchcrat, 160–61 overview, 384n33
Berkeley school, 60 Vent-for-Surplus models
Berlin Conference of 1885, 482–83 assumptions regarding, 305–11
Big Man syndrome, 195 general discussion, 300–5
black-eyed peas, 36, 38 Catholics
Bølling-Allerød interstadial, 34–35 location of missions, 493f
Botswana, witchcrat, 160 Mbanza Kongo/São Salvador, 367, 373–75,
Brazil 379–86
Dahomey and, 451–53 missionaries, 500–4
Kongo identity, 388 Cavazzi, Antonio, 380
plantation agriculture, 72–73 Central Africa
population, 1850 and 1870, 67 impact of imperial peace on
social savings of railways, 342–44 development, 435
sugar production, 71 impact of slave trade on population, 145–46
Brewer, John, 448–49
British and African Steam Navigation Co., 237 he Century Company map, 496, 510
British colonial system, 480–81. Chad, malaria mortality rate, 111
See also Colonial Period Chamberlain, Joseph, 18
Brooks, Francis, 60 Chaves, Isaias, x
brotherhoods, Mbanza Kongo, 383–85 Chefneux, Leon, 350, 351
 Index 515

chiefs, role in Euro-African trade, 259–60 private consumption of imported

Christianity, 180, 181–82, 199n27 goods, 357
General Act of the Conference of Berlin, Sierra Leone, 360
493, 493n3 traditional consumption, 359
historical inluence of, 492–93 impact of imperial peace on African
location of Catholic and Protestant development, 428–33
missions, 493f impact on Euro-African trade, 258–60
Mbanza Kongo/São Salvador, 367, 373–75, overview, 17–20
379–86 textile industry, 272–76
missionaries in colonial Africa trans-Atlantic slave trade, 479–82
analysis data, 492–97 Western Nigeria entrepreneurship and,
education, Catholics versus Protestants, 216–27
500–4 commerce. See also economic development;
education, gender, 504–9 Euro-African trade
education, overview, 498–500 city-states and, 47–48
historical background, 490–92 from farming to commerce in southern half
overview, 489–90 of Africa, 48–50
tracing impact of missions on individuals, Gold Coast
493–97 bankruptcies in England and, 252–55
Pentecostalism, 180, 181–82 during economic crisis, 247–52
churches, Mbanza Kongo/São Salvador, William Narh Ocansey, 244–47
369–72, 377–78 merchant-managed, 43–45
cities overview, 14–15
distribution in Africa in 1500, 95f spread of Islam and, 48
emergence of, 47–48 commission houses (consignment houses),
impact of malaria on, 93–95 Euro-African trade, 231–32, 232n1,
world city sizes by region in 1500, 94t 239–40, 249–52
city-states, 47–48 commodity production, Americas
civil law tradition, French colonial system, 480 labor, 70–71
civilizations, deining, 52 plantation economies, 71–76
Clarence-Smith, William G., x rice, 69–70
Clarke, S.H., 212 sugar, 71
cocoa farming, 310–11 common law, 482n6
coercion, lineage-based societies, 158, 158n2 concealing incomes and assets strategy,
Coker, J.K., 219 redistributive pressures, 171–74
collateral kin dependents, 171n10 Congo
collective prosperity versus individual disincentive efects of redistributive
prosperity, 163–68, 163n6 pressures, 170
Colonial Period. See also missionaries in factionalism and tribal patriotism, 195
colonial Africa Congo Basin, 50
estimating GDP, 354–60 Congo Free Trade Zone, 273
consumption of government services, consignment houses (commission houses),
357–58 Euro-African trade, 231–32, 232n1,
consumption of public and related 239–40, 249–52
services, 358 coping strategies, redistributive pressures,
data and methodology, 355–57 171–87
exports, 357 concealing incomes and assets, 171–74
gross capital formation, 358–59 migration and estrangement, 175–79
imports of goods and non-factor religious conversion, 179–87
services, 359 copper currency, 63
Nigeria, 360 Coptic looms, 272
516 Index

cotton industry, 38, 75n11, 266–67, 309–10 factor in missionary activity locations, 490
credit. See also Euro-African trade global comparisons, 138–40
Euro-African trading networks, 238–40 history, 7
in Gold Coast during economic crisis, implications of new estimates on Africa,
247–52 140–147
culture Latin America, 1850–1912, 67t
Atlantic Creole culture in Mbanza Kongo/ Maddison estimates by world region, 139t
São Salvador, 374–75, 379–87 Mexico, 1500, 60n3
Nok, 45 Mexico, 1850 and 1870, 67–68
overview, 11–14 new versus old estimates, 133–37
currency overview, 131–33
1500–1700, 143 population density, 6–7
in emerging market economy of West by country in 1500, 124–28
Africa, 63 cross-country regressions in 1500, 94t
Mbanza Kongo/São Salvador, 375–78 impact of malaria on, 91–92
United States, 1850–1912, 67t
Dahomey, 404 Denevan, William, 60–61
economic development, 456–58 Denevan synthesis, 60–61, 61n4
iscal-military state, 448–49 Dependency theory, 296, 304
reputation, 447 disability
slave trade and, 447–48 disability weights, 120, 120n9
source material regarding, 449–53 years lost to disability due to malaria, 121t,
state-dominated society, 448 119–22
warfare, 453–56 disease environment
daily collectors, Benin, 173–74, 174n12 impact on development, 99–100
Dalzel, Archibald, 447 impact on population, 136
dammar cotton, 266–67 disincentive efects of redistributive pressures,
Dapper, Olifert, 380 166, 168–71
David, Paul, 464 distribution systems, peasant economy, 144
Davies, J.P.L., 218–19 Doherty, J.H., 225
Dawodu, W.A., 220–22 domestic economy, 141–142.
decolonizing warfare, sub-Saharan Africa, See also economic development
394–97 domesticated animals, 38–39
defensive modernization, 408 Dominicans, Mbanza Kongo, 383, 384–85
democratic authority, impact of trans-Atlantic double-heddle looms, 271–72
slave trade on, 476, 478t
demographics East Africa
Africa, 1450–1790, 133 Horn of Africa, 45–47
Africa, 1500–1700, 142–45 impact of slave trade on population, 145–46
Africa, 1500–2000, 137f textile industry
Africa, 1700–2000, 132–39t Colonial period, 272–76
Africa, 1790–1890, 132–33 early technology and production systems,
Africa, 1890–1950, 132 270–72
Africa, 1950–2000, 132 Madagascar, 265–66
Africa, new estimates, 139t mills, 276–84
Africa, social dynamics from 1700–1900, Northeastern Africa, 266–67
145–146 overview, 264–65
Africa, twentieth century, 146–47 survival of artisans in era of
Africa in global context, 147–49 independence, 284–86
comparison of Africa to Eurasia, 138–39, Swahili, 267–70
138n8 Easterlin’s Paradox, 166n7
 Index 517

economic development ethnicity

comparative examination of West Africa approaches to, 143–44
and Americas, 58–69 ethnicity-level variable, missionary activity,
Dahomey, 456–58 494f, 510–11. See also missionaries in
distribution systems, peasant economy, 144 colonial Africa
domestic economy and overseas trade, Ethnographic Atlas, 466–67
141–42, 142n12 colonial cross sections, 428–29, 430–31
economic inequality and political structures, merging with Transatlantic Slave Trade
434t Database, 468f
efect of slave trade on institutions, 16–17 Eurasia, 321
impact of malaria on, 91–100 population comparison of Africa to, 138–39,
land quality, 92–93 138n8
population density, 91–92 population density, 92
production technology, 97 Euro-African trade
state antiquity, 95–96 bankruptcies in England and Gold Coast,
urbanization, 93–95 252–55
informal economy, 144–45 British navy and, 235n3
political structures and, 430t Christian Jacobson, 255–58
theories of, 1–5 commerce and credit in Gold Coast during
education, by missionaries economic crisis, 247–52
Catholics versus Protestants, 500–4 commission houses, 232n1
gender, 504–9 credit and trust in trading networks, 238–40
overview, 498–500 impact of Colonial Period on, 258–60
tracing impacts of missions on individuals, Jaja of Opobo, 258n14
493–97 oils and fats market, collapse of, 241–44
Egypt overview, 231–33
economic nationalism, 228–29 palm oil supplies, 246n10
gloriication of, 52 quinine, 235
Second Great Transition and, 39–42 steamships, 233–38
Ehret, Christopher, x, 5–6 William Narh Ocansey, 244–47
Ekpe secret society, Euro-African Europe
trade, 239 challenges to conquest of Africa, 437, 438
embroidery, textile industry, 275 impact of imperialism on warfare and
empires, West Africa, 47–48 development in sub-Saharan Africa
Engerman, Stanley L., xi and, 407–11, 415
entrepreneurship, Western Nigeria. population comparison of Africa to, 139–40
See also Yorubaland redistributive pressures and witchcrat, 166
Colonial Period, 216–27 European traders, in West Africa, 77–79
Daniel Conrad Taiwo, 210–11 explorer routes, 496f
instability of, 225n36
overview, 208–11 F & A Swanzy company, 247
warfare and political competition in factionalism, 192–96
nineteenth-century, 211–16 factor endowments argument, commodity
estrangement, as coping strategy for production, 74–76
redistributive pressures, 175–79, 197 fadingya concept, Gambia, 157n1
Ethiopia family/kin drag thesis, 177, 177n16
artisanal textile sector, 285 FAO suitability index, 497, 510
obstacles to adopting wheeled female education, in colonial Africa, 504–9
transportation, 350–51 Ferreira Pires, Vicente, 452
Solomonic Ethiopia, 405–6 inger millet, 39
weaving, 274 First Commercial Revolution, 45–47
518 Index

First Great Transition, 33–38, 51 global markets, impact on West Africa, 15.
Firth, Raymond, 164 See also cash-crop revolution
iscal-military state glottochronology, 6n3
Dahomey, 448–49 GMI Seamless Digit Chart of the World,
overview, 8–10 497, 510
forced mutual help, 164 Gold Coast, 78–79, 80–81. See also Euro-
fragile revolution. See warfare African trade
Frame Group, South Africa, 283 Ada, 244–47
French colonial system, 480 bankruptcies in England and, 252–55
French West Africa (FWA), 313t commerce and credit during economic
Friendship Mill, 283 crisis, 247–52
Fulanis, 185, 185n22, 185n23 Ordinance 38, 335n12
FWA (French West Africa), 313t railways, 329–31, 336–38, 337t,
338t, 339t
Gambia social rate of return approach,
badingya concept, 157n1 transportation, 346–47t
fadingya concept, 157n1 William Narh Ocansey, 244–47
Garamantes, 45 Goldie, George, 258
Garrard, Timothy, 45 Goody, Jack, 397n6
GDP (gross domestic product) Great Lakes region, 50
Colonial Period, estimating, 354–60 Great Zimbabwe city, 50
consumption of government services, Grief, Avner, 464
357–58 gross domestic product (GDP)
consumption of public and related Colonial Period, estimating, 354–60
services, 358 consumption of government services,
data and methodology, 355–57 357–58
exports, 357 consumption of public and related
gross capital formation, 358–59 services, 358
imports of goods and non-factor data and methodology, 355–57
services, 359 exports, 357
Nigeria, 360 gross capital formation, 358–59
private consumption of imported imports of goods and non-factor
goods, 357 services, 359
Sierra Leone, 360 Nigeria, 360
traditional consumption, 359 private consumption of imported
transportation costs and, 334 goods, 357
gender, education by missionaries Sierra Leone, 360
and, 504–9 traditional consumption, 359
General Act of the Conference of Berlin, 493, transportation costs and, 334
493n3 ground looms, 272
geography, factor in missionary activity, 490, groundnut oil, 234
496–97 groundnuts, 36, 38
Gesellschat, 197 Guinea yams, 36
Ghana
agricultural exports, 299t Hargreaves, David, 232
cocoa farming, 310–11 Haugerud, Angelique, 162
coping strategies for redistributive head porterage, 327–29, 331–34
pressures, 180–81 health transition, life expectancy and,
factionalism, 195 109–110, 110n6
foreign trade, 313t Herbst, Jefrey, 461
tribal patriotism, 195 Herivel, J.W., 349
git giving, 155–58 heterarchy, 9n8
 Index 519

heterozygote advantage (balanced informal economy, 144–45. See also economic

polymorphism), 105 development
Heywood, Linda, xi Inikori, Joseph, xi, 7–8
Hickson, Robert, 252–53 institutions
hierarchy of rank rationale, redistributive cash-crop revolution and welfare outcomes,
norms, 164–65 311–15
Holl, Augustin, 43–44 overview, 8–11
Holt, John, 258 insurance-based rationale, redistributive
Hopkins, Anthony G., 141–42, 397n6, 461 norms, 163–64
Horn of Africa, 45–47. See also East Africa Invisible Hand doctrine, 157
Horton, Robin, 491 ironworking, 44
Human Relations Area Files (HRAF) Islam, 181, 183–87
colonial cross sections, 429–431, 429t Muslim-based slave network, 466
political structures, 433–34 spread of commerce and, 48
Hunter, Guy, 163 state-building through war, 405–6
Hut Tax Rebellion (Bai Bureh War), 354 trading diasporas and, 183, 183n18
Italy, Zambaiti Group, 284
Ibadan war chiefs, 213, 216
Ife city-state, 48 Jacobson, Christian, 244, 249, 255–58
Igbo Jaja of Opobo, 258n14
Igbo-Ukwu city-state, 48 Jehovah’s Witnesses, 181
Nigerian Youth Movement, 194–95 Jesuits, Mbanza Kongo, 383–85
Ilife, John, 144, 192–93, 397n6 Johnston, Harry, 335–36
imperial peace, impact on African Jones, Alfred, 240
development Juula groups, 184, 184n21
colonial period, 428–33
internal politics, 433–34 Katz, Friedrich, 66
kinship societies, 425 Kennedy, Paul, 175–76, 176n14, 419
Kuba, 435–37 Kenya
overview, 424–25 Mau Mau rebellion, 194
“peace in the feud” concept, 425–26 redistributive pressures
pre-colonial period, 426–28 coping strategies for, 174, 181
slave trade, 438 disincentive efects of, 170–71
South-Central African societies, 435 textile mills, 278
state development and, 426 witchcrat, 162
technological development, 439 Kerma, 41
imperialism, warfare and development in kinship societies, 425. See also lineage-based
sub-Saharan Africa and, 407–11, 415 societies, Sub-Saharan Africa
imports Kirimba islands, 268
of goods and non-factor services in Colonial Kisero, Jaindi, 195–96
Period, 359 Klieman, Kairn, 49
private consumption of imported Knowles, Lillian C. A., 336
goods, 357 Kuba, 435–37
warfare and development in sub-Saharan Kuczynski, R. R., 136
Africa and, 414–15 Kurunmi, 212–13, 215–16
Inca empire, 64–65 Kush kingdom, 42
independence movements, 190–92
indirect colonial rule, 480, 481–83, 481f labor, reversal of fortune of West Africa and,
individual prosperity versus collective 76. See also slave trade
prosperity, 163–68, 163n6 Lace, Ambrose, 239
industrialisation, impact on redistributive Lagos, 329–31. See also Yorubaland
pressures, 188–89 Laird, Macgregor, 235
520 Index

land quality land quality, 92–93

by country, in 1500, 124–28 population density, 91–92
impact of malaria on, 92–93 production technology, 97
Lander, John, 235 state antiquity, 95–96
Lander, Richard, 235 urbanization, 93–95
language distribution, impact on economic efects of malaria morbidity,
population, 136 119–22
Latin America. See Americas malaria episodes, 101t3.3A.
Law, Robin, 366, 448, 458, 462 measuring historical burden of malaria,
Lewis, Arthur, 157–58 107–12
liberal authority, impact of trans-Atlantic slave modern day impact, 100–102
trade on, 467–76, 477t modern mortality rates, 111–12
life expectancy neurological sequelae, 100, 101t
health transition and, 109–110, 110n6 overview, 4
modern malaria mortality rates, 111–12 plasmodium parasite, 102–4
UN Model Life Tables, 110–11 sickle cell disease and, 102–6
Limpopo Valley, 50 sub-Sahara Africa, 1850, 81n15
lineage-based societies, Sub-Saharan Africa years lost to disability per capita, 121t
git giving, 155–58 Mali, 47–48, 111
redistributive norms Malthusian model, assessing malaria mortality,
disincentive efects of, 168–71 117–119, 117f3.8.
rationale behind, 163–68 Manning, Patrick, xii, 7, 60, 79
sanctions in support of, 158–62 Mapungubwe, 50
repressive measures in response to market economy. See also economic
individual success, 160–62, 198 development
Livingston, Charles, 237 development in Americas, 64–68
looms development in West Africa, 61–64
Coptic, 272 Marx, Karl, 188–89
double-heddle, 271–72 Mashariki, 49
ground, 272 Mbanza Kongo/São Salvador
pit, 271 Atlantic Creole culture in, 379–87
Southeast Asian, 272 early transformation, 368–78
vertical, 272 overview, 366–67
luck, lineage-based societies, 158 McIntosh, Roderick, 62
Lugard, Frederick, 327–28, 480 McNeill, William H., 418–19
McPhee, Allan, 336
Machado, Pedro, 265 Menelik II (Ethiopia), 350–51
machira cloth, 268–69, 274–75 merchant-managed commerce, 43–45
Mackinnon, William, 352 Meroe kingdom, 42
Madagascar Mexico
artisanal textile sector, 285–86 population in 1500, 60n3
textile industry, 265–66 population in 1850 and 1870, 67–68
weaving, 274 social savings of railways, 342–44
maize, 136n3 mfecane, 402–3, 403n8
malaria, by country in 1500, 124–28 migration
malaria, impact on African development, 103f, from Africa, 139–40, 140n9
111–12 as coping strategy for redistributive
calculations based on prevalence of sickle pressures, 175–79, 187–88, 188n25, 197
cell trait, 108–121 military power, 431t, 432t
direct efect of malaria mortality, 113–19 military revolution. See warfare
direct efects of, 100–101 Miller, Alexander, 258
economic development, 91–100 mills, textile industry, 276–84
 Index 521

milwani cloth, 268 estimating GDP in Colonial Period, 360

Mirambo, 401 foreign trade, 313t
missionaries in colonial Africa head porterage, 331–33
analysis data, 492–97 malaria mortality rate, 109–111, 112
education and Nigerian Youth Movement, 194–95
Catholics versus Protestants, 500–4, 502t railways, 340–41, 341t, 342–43t, 342t
gender, 504–9, 506t, 508t social rate of return approach,
overview, 498–502, 499t transportation, 346–47t
ethnicity-level variable, 494f, 510–11 Niger-Kordofanian peoples, 34
explorer routes, 496f Nilo-Saharan peoples, 34, 36–37
General Act of the Conference of Berlin, Nok culture, 45
493, 493n3 Norris, Robert, 447
historical background, 490–92 Northeastern Africa, textile industry, 266–67
location of missions, 493f Northern Sudanians, 36–37
overview, 489–90 Nubia, 41
railway lines and, 497f Nunn, Nathan, 12–13
village-level variable, 495f, 510–11 Nzinga Nkuwu, 368
mobility. See Euro-African trade
mortality rates, malaria, 100, 111–12 oblique (vertical) looms, 272
Moxicongo identity, 385 Ocansey, John E., 248–49, 252–54
Mozambique Ocansey, William Narh, 244–47, 252–54
textile mills, 278–79 oils and fats market
witchcrat, 161 collapse of, 241–44
Mwanza Textiles, 283 groundnut oil, 234
Myint, Hlya, 296, 300–1, 305 palm oil, 36, 234, 241–42, 246n10
Oliveira Mendes, Luís Antonio de, 452
Nabta Playa, 40–41 Olukoju, Ayodeji, xii, 11–12
Napatan kingdom, 42 Ormsby-Gore, William, 322–23, 327, 331–33
National Congress of British West Africa other-regarding values, 157–58
(NCBWA), 193 Oyo Empire, collapse of, 399–400
nationalism, 189–96
economic, 228–29 palm oil, 36, 234, 241–42, 246n10
factionalism, 192–96 Pannikar, K. Madhu, 448
independence movements, 190–92 Park, Mungo, 234–35
TANU, 192 path dependence
tribal patriotism, 192–96 factor in missionary activity locations,
NCBWA (National Congress of British West 490–91
Africa), 193 trans-Atlantic slave trade, 461–64
Neolithic revolution, 137–138 patriarchy, slave trade and, 476
nepotism, 169, 169n9 patrimonial, 460, 460n1
neurological sequelae, malaria and, 101t patron-clientism, 483
Ngoni, 403 pawnship, as means of security, 238
Niane, D.T, 59 peace. See imperial peace, impact on African
Niger, malaria mortality rate, 109–111 development
Niger Bend sub-region “peace in the feud” concept, 425–26
market economy and, 61–62 pearl millet, 39
population, 58–59 Pearse, S.H., 222–25
Niger-Congo peoples, 35–36 peasant colonies. See cash-crop revolution
Nigeria. See also entrepreneurship, Western Pentecostalism, 180, 181–82
Nigeria Peukert, Werner, 448
agricultural exports, 299t pit looms, 271
cocoa farming, 310–11 plasmodium parasite, 102–4
522 Index

Platteau, Jean-Philippe, xiii, 12 impact of malaria on, 91–92

Polanyi, Karl, 448 United States, 1850–1912, 67t
political authority, trans-Atlantic slave trade post-colonial period, trans-Atlantic slave
and, 464–76, 469t trade, 482–84
political development, 144 poverty, 144–45, 208n1
in 1500, by country, 124–28 pre-colonial period, impact of imperial peace
efect of slave trade on political on African development, 426–28
centralization, 16 predation, slave trade and, 476
impact of imperial peace on African print revolution, 13
development, 433–34 privatised violence, 400–3, 420
impact of malaria on, 95–96 production technology. See also technology
Western Nigerian entrepreneurship and, impact of malaria on, 97
211–16 textile industry, 270–72
political structures Protestants
economic activity and, 430t location of missions, 493f
economic inequality and, 434t missionaries, 500–4
feuding and, 430t Pryor, Frederic, 144
political checks and, 434t
political trajectory and, 433t quinine, 235
public goods and, 431t Qustul, 40–41
slave trade and, 440t
warfare and, 432t raiding war, 413–14
population railways
Africa, 1450–1790, 133 factor in missionary activity, 495–96, 497f
Africa, 1500–1700, 142–45 Gold Coast, 329–31, 337–38t, 338t, 339t
Africa, 1500–2000, 137f Lagos, 329–31
Africa, 1700–2000, 132–39 Nigeria, 341t, 342–43t, 342t
Africa, 1790–1890, 132–33 overview, 329–31
Africa, 1890–1950, 132 Sierra Leone, 329–31, 339t, 341t
Africa, 1950–2000, 132 social rate of return, 345–48
Africa, in global context, 147–49 social savings approach, 335–44, 340t, 341t
Africa, new estimates, 139t rapid African growth thesis, 146–47
Africa, social dynamics from 1700–1900, Reckless, George, 349
145–46 redistributive pressures, Sub-Saharan Africa
Africa, twentieth century, 146–47 broad-based counteracting forces to, 187–89
comparison of Africa to Eurasia, 138–39, disincentive efects of, 168–71
138n8 git giving, 155–58
factor in missionary activity individual coping strategies, 171–87
locations, 490 concealing incomes and assets, 171–74
global comparisons, 138–40 migration and estrangement, 175–79
history, 7 religious conversion, 179–87
implications of new estimates on Africa, nationalism and, 189–96
140–47 factionalism, 192–96
Latin America, 1850–1912, 67t independence movements, 190–92
Maddison estimates by world region, 139t tribal patriotism, 192–96
Mexico, 1500, 60n3 overview, 153–55
Mexico, 1850 and 1870, 67–68 rationale behind, 163–68
new versus old estimates, 133–137 sanctions in support of, 158–62
overview, 131–133 Reid, Richard, xiii
population density, 6–7 religion, 199n27. See also witchcrat
by country in 1500, 124–128 Catholics
cross-country regressions in 1500, 94t location of missions, 493f
 Index 523

Mbanza Kongo/São Salvador, 367, Shaka, 402–3

373–75, 379–86 sickle cell disease
missionaries, 500–4 AA phenotype, 104–6
ethnicity-level variable, 494f, AS phenotype, 104–6
510–11 A type allele, 104–6
General Act of the Conference of Berlin, calculating carriers in
493, 493n3 population, 105n4
Islam, 181, 183–87 malaria and, 102–106, 108–21
Muslim-based slave network, 466 measuring historical burden of malaria
spread of commerce and, 48 using data on sickle cell trait, 111
state-building through war, 405–6 S type allele, 104–106, 106f3.6.
trading diasporas and, 183n18 Sierra Leone
missionaries in colonial Africa estimating GDP in Colonial Period, 360
analysis data, 492–97 head porterage, 331–33
education and, Catholics versus Hut Tax Rebellion (Bai Bureh War), 354
Protestants, 500–4 obstacles to adopting wheeled
education and, gender, 504–9 transportation, 348
education and, overview, 498–500 railways, 329–31, 339–40, 339t,
historical background, 490–92 341t
location of missions, 493f social rate of return approach,
overview, 489–90 transportation, 346–47t
tracing impacts of missions on slave trade. See also trans-Atlantic slave trade
individuals, 493–97 cash-crop revolution and, 305–8
Pentecostalism, 180, 181–82 Dahomey and, 447–48
Protestants, 493f, 500–4 factor in missionary activity, 497
religious brotherhoods, Mbanza Kongo, impact of imperial peace on African
383–85 development, 438
religious conversion as coping strategy impact on population, 134–36, 137n6,
for redistributive pressures, 179–87, 137–38n7, 145–146
197–99 Mbanza Kongo/São Salvador, 386–87
village-level variable, 495f, 510–11 overview, 15–17
reversal of fortune pawnship, 238
comparative examination of West Africa political structures and, 440t
and Americas, 69–81 reversal of fortune of West Africa and, 76
Dahomey, 449 textile industry and, 270
revolution. See warfare vent-for-surplus theory, 143n13
Robinson, James A., 462–63 violence and, 398
Roome map, 489–90, 492–93, 510 Smith, Adam, 300
royal letters, Dahomey, 449–53 Smith, Shiela, 307n10
ruga ruga, 401 social rate of return approach,
Rwanda, 159, 161 transportation, 345–48, 346n15, 346t,
347t
S type allele, sickle cell disease, 104–106, 106f social savings approach, transportation, 335–44
Salazar, Antonio, 277 by country, 343t
Sarro, Ramon, 162, 165 Gold Coast, 345t
SCCS (Standard Cross-Cultural Sample), other forms of transportation, 344
19–20, 432–33 railways, 341t
Second Great Transition, 39–45, 51 comparative context, 342–44
Senegal Gold Coast, 337–38, 338t, 339t
agricultural exports, 299t Nigeria, 340–41, 342–43t, 342t
disincentive efects of redistributive overview, 335–36
pressures, 169–70 Sierra Leone, 339–40
524 Index

Société Textile Africaine (Texaf), 276–77 Tanganyka African National Union (TANU),
Société Usines Textiles de Léopoldville 191–92, 192n26
(Utexléo), 276–77 Tanzania
socioeconomic development in Atlantic world artisanal textile sector, 285
economic development, 58–69 coping strategies for redistributive
overview, 56–58 pressures, 172
reason for comparison choice, 57n2 independence movements, 191–92
reversal of fortune, 69–81 textile mills, 278, 283
Sokoto Caliphate, 405–6 witchcrat, 159, 161
Solomonic Ethiopia, 405–6 Taussig, Michael, 72
Somalia, 273–74 technology
Songye state, 50, 50n37 in 1500, by country, 124–28
sorghum, 39 impact of imperial peace on African
South Africa development, 439
textile industry, 279–80, 283 impact of malaria on, 97
witchcrat, 158–59, 160–62, 160n4 obstacles to adopting wheeled
South-Central African societies. See Central transportation, 348–53
Africa textile industry, 270–72
Southeast Asian looms, 272 Texaf (Société Textile Africaine), 276–77
Southern Rhodesia (Zimbabwe), 50, textile industry, East Africa
279–80, 283 Colonial Period, 272–76
spinning, 271 early technology and production systems,
SS phenotype, sickle cell disease, 104–6 270–72
Standard Cross-Cultural Sample (SCCS), Madagascar, 265–66
19–20, 432–33 mills, 276–84
state antiquity, 144. See also political Northeastern Africa, 266–67
development overview, 264–65
by country, 1500, 124–28 survival of artisans in era of independence,
impact of malaria on, 95–96 284–86
state development Swahili, 267–70
Horn of Africa, 45–47 thalassemia, 104
impact of imperial peace on African hird Great Transition, 43–45, 52–53
development, 426 hornton, John, xii
state-building through war. See warfare Tichit, 43
state-dominated society, 448 Tilly, Charles, 419
States and Power in Africa (Herbst), 461 Tinubu, Efunroye, 213–14, 213n10
steamships, Euro-African trade and, 233–38 Tippu Tip, 401–2
subsistence innovation, 35 Tongaat Group, South Africa, 283
subsistence production, Americas, 64 Townsend, Henry, 212
Sudan trading diasporas, Islam and, 183, 183n18
artisanal textile sector, 285 trans-Atlantic slave trade
weaving, 274 Colonial Period, 479–82
witchcrat, 159 irst-stage IV regressions on slave exports,
Sultanate of Zanzibar. See Zanzibar 475t
Susu, 47–48 impact on absolutist authority in West
Swahili, 267–70 Africa, 473t
Syaam, 435–37 impact on democratic authority in West
symmetrical git giving, 155–56 Africa, 478t
Szeresewski, Robert, 296, 301–2 impact on liberal authority in West Africa, 477t
indirect colonial rule and, 481f
tailoring, textile industry, 275–76 merging Transatlantic Slave Trade Database
Taiwo, Daniel Conrad, 210–11, 217–18 and Ethnographic Atlas, 468f
 Index 525

origins and path-dependence, 461–64 witchcrat, 162

overview, 336 urbanisation
political authority and, 464–76, 469t emergence in Horn of Africa, 45–47
post-Colonial Period, 482–84 emergence in western Sudan and
Transatlantic Slave Trade Database on slave Sahel, 44–45
exports, 462, 467n4 impact of malaria on, 93–95
transformations in violence, sub-Saharan impact on redistributive pressures, 188–89
Africa, 398–407 Utexléo (Société Usines Textiles de
transportation, West Africa Léopoldville), 276–77
cost comparison, 334
cost estimates, 328–37t Vansina, Jan, 49
head porterage, 327–29, 331–34 Kuba study, 435–37
obstacles to adopting, 348–53 South-Central Africa societies, 435
overview, 327 Vas, George, 379–80
railways Vent-for-Surplus (VFS) models
Gold Coast, 328–37t, 338t assumptions regarding, 305–11
Nigeria, 341t, 342–43t, 342t general discussion, 300–5
overview, 329–31 slave trade, 143n13
Sierra Leone, 339t, 341t vertical (oblique) looms, 272
social rate of return, 345–48 VFS models. See Vent–for-Surplus models
social savings approach, 335–44, 340t, village-level variable, missionary activity,
341t 495f, 510–11. See also missionaries in
roads, 324, 324n5 colonial Africa
social savings approach, Gold violence. See also warfare
Coast, 345t slave trade, 398
Travels in the Interior Districts of Africa (Park), transformations in, 398–407
234–35
tribal patriotism, 192–96 Wagadu (Ghana), 47–48
tribute system, Aztec empire, 65–66 Wahhabism, 186–87, 186n24
trickle-down efect, 167–68, 168n8 warfare
Triggs, Alfred, 246n10 Dahomey, 453–56
Tsetse ly, 323–24 development in sub-Saharan Africa and
Turnbull, Richard, 191, 192 achievements, 411–12
compared to European experience, 413
Uganda decolonizing warfare, 394–97
coping strategies for redistributive dependence on imported commodities,
pressures, 181–82 414–15
foreign trade, 313t European imperialism and, 407–11, 415
Teso rebellion, 194 failures, 412–13
textile mills, 278 overview, 393–94
witchcrat, 161 transformations in violence, 398–407
United Nations (UN) types of warfare, 413–14
Model Life Tables, 110–11 without inluence of European
Population Oice, 133 colonialism, 415–18
United States. See also Americas political structures and, 432t
commodity production for export and Western Nigerian entrepreneurship and,
population, 1850–1912, 67t 211–16
population, 1850 and 1870, 67 weaving, 270, 271–72
Upemba state, 50 Ethiopia, 274
Upper Guinea Coast Madagascar, 274
coping strategies for redistributive pressures, Somalia, 273–74
180n17 Sudan, 274
526 Index

Weil, David N., xii, 6–7 Europe, 166

welfare outcomes, cash-crop revolution and, Kenya, 162
311–15 Mozambique, 161
West Africa South Africa, 158–59, 160–62, 160n4
comparative examination to Americas Sudan, 159
economic development, 58–69 Tanzania, 159, 161
overview, 56–58 Uganda, 161
reason for comparison choice, 57n2 Upper Guinea Coast, 162
reversal of fortune, 69–81 women, education by missionaries, 504–9
coping strategies for redistributive pressures, World Health Organization AFRO region
183–87 modern malaria mortality rates, 100, 109–12
global markets impact on, 15 years lost to disability due to malaria,
impact of slave trade on 119–22
population, 145–46 World Malaria Report 2008, 109–12
wheeled transportation
cost comparison, 334 yams, 36
versus head porterage, 327–29, 331–34 years lost to disability (YLD), malaria and,
obstacles to adopting, 348–53 121t, 119–122
overview, 327 Yoruba, 11–12, 48, 194–95, 209.
railways, 329–31, 335–44, 340t, 341t, See also Yorubaland
342t, 345–48 Yorubaland
Whatley, Warren C., xii, 17 Colonial Period, 216–27
wheeled transportation, West Africa Daniel Conrad Taiwo, 217–18
cost comparison, 334 Efunroye Tinubu, 213–14
versus head porterage, 327–29 Efunsetan Aniwura, 213, 214–15, 216
head porters, 331–34 entrepreneurship in the contexts of warfare
obstacles to adopting, 348–53 and political competition, 211–16
overview, 327 Ibadan war chiefs, 213, 216
railways J.H. Doherty, 225
Nigeria, 341t, 342–43t, J.K. Coker, 219
342t J.P.L. Davies, 218–19
overview, 329–31 Kurunmi, 212–13, 215–16
social rate of return, 345–48 poverty, 208n1
social savings approach, 335–44, 340t, Salami Agbaje, 225–27
341t S.H. Pearse, 222–25
roads, 324, 324n5 W.A. Dawodu, 220–22
Wilkinson, E.F.W., 328 Younger Dryas, 35
Wilks, Ivor, 349–50
Willcox, Walter, 136 Zambaiti Group, 284
Williams, Eric, 233–34 Zambia, 181
witchcrat Zanzibar, 268, 351–53
Benin, 160–61 Zimbabwe (Southern Rhodesia), 50,
Botswana, 160 279–80, 283
Cameroon, 160–61, 161n5 Zulu, 402–3, 435n5